University of Southern California Dissertations and Theses

Spatial and temporal dynamics of marine microbial communities and their diazotrophs in the Southern California Bight

(USC Thesis Other)

Spatial and temporal dynamics of marine microbial communities and their diazotrophs in the Southern California Bight

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Spatial and Temporal Dynamics of Marine Microbial Communities and Their Diazotrophs in the Southern California Bight by Colette C. Fletcher-Hoppe A Dissertation Presented to the FACULTY OF THE USC GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY MARINE BIOLOGY & BIOLOGICAL OCEANOGRAPHY May 2024 ii Acknowledgements One of the central questions of marine microbial ecology is how microorganisms cope with the stress of nutrient limitation. In studying this question, I have gleaned many insights into how human organisms deal with their own stress. Many people are like the species of Pseudo-nitzchia that produce domoic acid, a potent neurotoxin. When cultured Pseudo-nitzchia strains deplete all the nutrients from their media, they ramp up domoic acid production, prioritizing creation of this pervasive poison over their growth. Still others are like Syndinales, a hypothesized parasite: they compensate for their own lack of nutrients by possessing others and stealing their life-force. Many individuals experiencing stress operate like SAR11. This tiny microbe has evolved to “streamline” its genome, tossing aside any genes that are not essential, in survival mode. Well-intentioned people resemble diazotrophs, rare prokaryotic organisms capable of making nitrogen available for the surrounding community out of thin air. However, because the nutrients are then trapped in diazotrophs’ bodies, the most useful thing a diazotroph can do for other microbes is to die, releasing their fixed nitrogen, the microbial equivalent of setting one’s self on fire to keep others warm. UCYN-A, the obligate symbiont diaozotroph, demonstrates the best way of coping with stress. This organism lives in partnership with a host, Braarudosphaera, in constant communication, each providing the other with the nutrients that it lacks. Such human interactions are as rare as UCYN-A, and deserve as much appreciation and care as this fascinating symbiont. Thank you to each person who has been the Braarudosphaera to my UCYN-A. First, to my advisor, Jed Fuhrman, for giving me an academic home for the last 6.5 years and the freedom to explore my intellectual passions. Jed is someone I aspire to resemble in many ways: he has an informed opinion about every topic, and a perceptive, critical eye tuned towards the pros and cons of any methodological choice. I appreciate all his advice and edits (“Jedits”) on my work over the years. Much appreciation for all the insights on mentorship and people management that I have gleaned. I am indebted for all of Jed’s financial support through grants NSF OCE 1737409, Gordon and Betty Moore Foundation Marine Microbiology Initiative grant 3779, and Simons Foundation Collaboration on Computational Biogeochemical Modeling of Marine Ecosystems (CBIOMES) grant 549943. To my peers in the Fuhrman lab, thank you for providing me with invaluable life lessons, mostly by way of example, some positive and some negative. Everyone can learn something by observing each of you. Thank you to my committee members, Douglas Capone, Sarah Feakins, Dave Hutchins, Jim Moffett, Fengzhu Sun, and Eric Webb, for all their time and investment in my progress through quals and my dissertation defense. It has been a privilege to learn from such academic giants. Decades of appreciation are due to the generations of undergraduates, PhD candidates, and postdoctoral scholars who have contributed to the San Pedro Ocean Time-series. In particular, thanks to research associates Troy Gunderson, Diane Kim, and the crew of the R/V Yellowfin. Thank you to Don Bingham for his tireless support of MEB students. Don, do you sleep? Outside of my department, I have been fortunate to pick the brains of academic legends at conferences and symposia, particularly including members of the Zehr lab. I appreciate all the insights that Kendra Turk-Kubo and Jon Zehr have brought to my work. So many thanks are due to USC alumnus and Zehr lab postdoc Mo Morando, without whom my second main-body chapter would not have been possible. iii The final year of my PhD was supported by the Knauss Fellowship in marine science policy. My mentors at the Department of Energy, including Tyler Christoffel, Jian Fu, and Nate McKenzie, were supportive and understanding as I wrapped up my dissertation. This experience would not have been possible without the administrative support of USC Sea Grant personnel, including Karla Heidleberg, Ruth Dudas, and Linda Chilton. My fellow Fellows kept me focused and on target. I greatly appreciate all the opportunities I have had to build scientific community and conduct scientific outreach through the MEB Graduate Student Government and through the Young Researchers Program. This service kept me motivated, as did the others who participated in it. It’s always refreshing to see people thinking beyond themselves. Special recognition is due to Rebecca Hernandez, USC Class of 2024. For some reason, she believes I have done her a great favor by having her aid my archaeal culture work in the Fuhrman lab for my final two years. In fact, quite the opposite is true. I was lucky to have her enthusiasm and scientific curiosity to aid my daily work. Thank you to my peers in the MEB department, particularly the 2017 cohort – they saved the best for last in our year! Gerid Ollison and Julie Hopper are both excellent examples of positivity and a living reminder to exercise (surprisingly, taking care of your body also takes care of your mind). May all your future endeavors work out. It is a joy to claim Kyla Kelly as my neighbor. Nina Yang and Heidi Aronson are two of the most endlessly supportive and encouraging people I know. The value of two friends who always listen and take my side cannot be measured. Without them, I would not have made it to the finish line with my sanity intact. I know this dissertation will stand the test of time as one of the greatest accomplishments I ever achieve. In the same breath, it took someone both terrible and wonderful to show me that no accomplishment can save you from feeling worthless, sick, or alone, and that the measure of human worth lies not in the lines on their CV, but in the lives they touch. Thank you to my medical team – Drs. Dunlop, Hewus, Katz, Green, and Quaereshi. Being an immunocompromised microbiologist is quite the paradox. Coincidentally, this is what two of us now are. Thank you to Maggie the calico cat, the most benevolent princess there ever was, who graciously allowed me to stop petting her long enough to write up my chapters. I appreciate the support that my family has provided. Just as good is the support of an amazing group of young women I have known since age 12 – watching you grow into the adults you are today has been an irreplaceable gift. Few things bring me as much pride as knowing that the seven of us women collectively hold two PhDs, three M.S. degrees, one MRb, and a J.D. They say that blood is thicker than water, but we know that the blood of the covenant is thicker than the water of the womb. The years 2017-2024 have been a wilder ride than I ever could have imagined. May I always endeavor to feel grateful for the past iterations of myself, who collectively got me through. iv Table of Contents Acknowledgements......................................................................................................................... ii List of Tables................................................................................................................................. vii List of Figures.............................................................................................................................. viii Dissertation Abstract...................................................................................................................... xi Chapter 1– Introduction .................................................................................................................. 1 Background and importance ....................................................................................................... 1 Methodology ............................................................................................................................... 2 Nitrogen fixers ............................................................................................................................ 4 Research questions...................................................................................................................... 6 References................................................................................................................................... 8 Chapter 2–X Marks the SPOT: Spatial and daily variation in marine microbial communities at the San Pedro Ocean Time-series.............................................................................................. 12 Abstract..................................................................................................................................... 12 Introduction............................................................................................................................... 13 Materials and Methods.............................................................................................................. 15 Sample collection and processing......................................................................................... 15 DNA extraction ..................................................................................................................... 17 DNA sequencing and processing .......................................................................................... 17 Data analysis......................................................................................................................... 18 Results....................................................................................................................................... 20 Effects of space vs. time on community stability ................................................................. 20 Community stability of large vs. small size fractions........................................................... 20 Environmental effects on small vs. large communities in spatial transect ........................... 23 Discussion................................................................................................................................. 24 Space & time have similar destabilizing effects on marine microbial communities over small scales ................................................................................................................... 24 Larger marine microbial communities are less homogenous over space & time ................. 27 Environmental distance drives community variation in the larger size fraction of marine microbes, particularly amongst heterotrophic bacteria ......................................................... 30 Conclusions............................................................................................................................... 34 References................................................................................................................................. 36 Chapter 2 Figures...................................................................................................................... 42 Chapter 3–UCYN-A2 dominates high nitrogenase gene expression in autonomously collected samples from the Southern California Bight ................................................................. 59 Abstract..................................................................................................................................... 59 v Introduction............................................................................................................................... 60 Methods..................................................................................................................................... 63 Sample collection.................................................................................................................. 63 Sample preparation ............................................................................................................... 64 Data analysis......................................................................................................................... 65 Results....................................................................................................................................... 66 Relative abundance of nifH is orders of magnitude higher in metatranscriptomes vs metagenomes.................................................................................................................... 66 UCYN-A comprises almost all nifH sequences.................................................................... 67 UCYN-A2 nifH sequences are more expressed in the transcriptomes though less abundant in the metagenomes than UCYN-A1 sequences ............................................ 67 Nitrogenase genes are the most expressed mRNA genes within the UCYN-A genome ...... 68 Discussion................................................................................................................................. 69 nifH is highly expressed in metatranscriptomes, particularly during the day....................... 69 UCYN-A dominated nifH expression and abundance at this site ......................................... 71 UCYN-A2 is more active and less abundant than UCYN-A1.............................................. 74 UCYN-A1 expresses nitrogenase more than any other gene................................................ 76 Conclusions............................................................................................................................... 77 References................................................................................................................................. 79 Chapter 3 Figures...................................................................................................................... 83 Chapter 4–Symbiotic UCYN-A strains co-occurred with El Niño, relaxed upwelling, and varied eukaryotes over 10 years off Southern California....................................................... 92 Abstract..................................................................................................................................... 92 Introduction............................................................................................................................... 93 Materials and Methods.............................................................................................................. 96 Data collection ...................................................................................................................... 96 Phylogenetic tree................................................................................................................... 97 Relationship with abiotic factors .......................................................................................... 97 Network analyses.................................................................................................................. 98 Results....................................................................................................................................... 99 Phylogenetic tree................................................................................................................... 99 Relationship with abiotic factors ........................................................................................ 100 Co-occurrence with non-host 18S taxa ............................................................................... 101 UCYN-A host-symbiont co-occurrences............................................................................ 101 Spatial-temporal distributions of UCYN-A ASVs.............................................................. 102 Discussion............................................................................................................................... 102 Relationship with abiotic factors ........................................................................................ 102 Co-occurrence with non-host 18S taxa ............................................................................... 105 UCYN-A host-symbiont co-occurrences............................................................................ 107 Spatial-temporal distributions of UCYN-A ASVs.............................................................. 109 Conclusions..............................................................................................................................111 vi Acknowledgements..................................................................................................................111 Author contributions................................................................................................................112 Competing interests .................................................................................................................112 Data Availability Statement .....................................................................................................112 References................................................................................................................................113 Chapter 4 Figures.....................................................................................................................119 Chapter 4 Supplementary Information.................................................................................... 126 Materials and Methods........................................................................................................ 126 Results and Discussion ....................................................................................................... 128 Chapter 4 Supplementary References................................................................................. 129 Chapter 4 Supplementary Figures....................................................................................... 131 Chapter 5–Conclusions............................................................................................................... 147 References............................................................................................................................... 151 vii List of Tables Chapter 2–X Marks the SPOT: Spatial and daily variation in marine microbial communities at the San Pedro Ocean Time-series Ch. 2 Table 1–Environmental parameters sampled across the spatial transect............................. 43 Ch. 2 Table 2– Beta diversity between spatial and daily samples is driven primarily by ecological turnover................................................................................................... 47 Chapter 3–UCYN-A2 dominates high nitrogenase gene expression in autonomously collected samples from the Southern California Bight Ch. 3 Table 1–Ratios of nifH BLAST hits in the metatranscriptomes to nifH hits in the metagenomes are almost an order of magnitude higher in UCYN-A2 than in UCYN-A1. ......... 90 Chapter 4–Symbiotic UCYN-A strains co-occurred with El Niño, relaxed upwelling, and varied eukaryotes over 10 years off Southern California Ch. 4 Supplementary Table 1–Identifiers for UCYN-A and Braarudospharea ASVs analyzed in this study. ...................................................................................................... 133 Ch. 4 Supplementary Table 2–Differences in environmental factors on dates UCYN-A1, UCYN-A2, hosts, and potential predator Lepidodinium ASVs are present vs. absent. ................................................................................................................. 134 Ch. 4 Supplementary Table 3–UCYN-A ASVs co-occur with a variety of 18S taxa. ................ 136 viii List of Figures Chapter 2–X Marks the SPOT: Spatial and daily variation in marine microbial communities at the San Pedro Ocean Time-series Ch. 2 Figure 1–Locations (A) and water filtration schematic (B) for the study sites................... 42 Ch. 2 Figure 2–The effects of space and time are comparable within each size fraction of marine microbial communities. ........................................................................... 45 Ch. 2 Figure 3–The smaller size fraction of microbial communities is more stable than the larger size fraction in the surface waters surrounding SPOT. ......................................... 46 Ch. 2 Figure 4–Samples from the smaller size fraction are more homogenous than samples from the larger size fraction of marine microbial communities...................................... 47 Ch. 2 Figure 5– Alpha diversity is higher in the larger size fraction of organisms but comparable across both space and time.................................................................................. 48 Ch. 2 Figure 6––Environmental factors drive dissimilarity in the larger size fraction of marine microbial communities to a stronger degree than in the smaller size fraction, particularly amongst heterotrophic bacteria. .............................................. 53 Ch. 2 Figure 7–Diversity patterns in heterotrophic bacteria and non-photosynthetic eukaryotes are driven by more environmental variables than are their photosynthetic counterparts................................................................................................................................... 54 Ch. 2 Supplementary Figure 1–Rarefaction curves show sequencing depth was sufficient to capture the full diversity of microbial ASVs in most samples from both the smaller size fraction (0.22-1µm, A) and the larger size fraction (1-80µm, B) .................................................................................................................................. 44 Ch. 2 Supplementary Figure 2–The 20 most abundant orders comprise ~95% of the smaller size fraction of marine microbial communities across space and time, but at most 80% of the larger size fraction. .................................................................................. 49 Ch. 2 Supplementary Figure 3 –ASVs from SAR11, Bathycoccus, and Synechococcus drive Bray-Curits dissimilarity in the smaller size fraction (0.22-1µm) of marine microbial communities around SPOT. ...................................................... 50 Ch. 2 Supplementary Figure 4–ASVs from Phaeocystis, Strombidiida, and Synechococcus drive Bray-Curits dissimilarity in the larger size fraction (1-80µm) of marine microbial communities around SPOT. Each ASV contributed to >1% on average of the overall Bray-Curtis dissimilarity across spatial samples, as identified by similarity percentage analysis (SIMPER)...................... 51 Ch. 2 Supplementary Figure 5–Relative abundances of ASVs that drive >1% of Bray-Curtis dissimilarity in daily samples of the larger (A) and smaller (B) size fractions of marine microbial communities. ................................................ 52 Ch. 2 Supplementary Figure 6–– Cyanobacterial relative abundance is highest in the north and northeast sampling sites in the transect. ................................................. 55 Ch. 2 Supplementary Figure 7–– Photosynthetic organisms drive community instability in the larger size fraction of marine microbial communities surrounding SPOT ........................................................................................................................ 56 Ch. 2 Supplementary Figure 8–Bray-Curtis similarity amongst metabolic groups of marine microbes plateaus with increasing distance between sampling sites............................ 57 ix Ch. 2 Supplementary Figure 9–Diversity patterns in genera of heterotrophic bacteria and non-photosynthetic eukaryotes are driven by more environmental variables than are their photosynthetic counterparts............................................. 58 Chapter 3–UCYN-A2 dominates high nitrogenase gene expression in autonomously collected samples from the Southern California Bight Ch. 3 Figure 1–Study location and environmental conditions at the time of ESP sampling. ....... 83 Ch. 3 Figure 2–Although it is orders of magnitude less abundant, nifH is expressed at a higher rate than psbA, a highly expressed diel gene present in cyanobacteria, which comprise a much larger proportion of the population than diazotrophs.......................................................................................................... 84 Ch. 3 Figure 3–UCYN-A dominates expression and abundance of nifH genes in the larger size fraction of ESP samples (1-300μm) .................................................................. 85 Ch. 3 Figure 4–UCYN-A2 comprises a higher proportion of nifH expression than of nifH abundance. ................................................................................................................ 89 Ch. 3 Figure 5–Nitrogenase genes were the most expressed out of all UCYN-A mRNA ........... 91 Ch. 3 Supplementary Figure 1–UCYN-A comprises almost all nifH hits in each metatranscriptomic and metagenomic sample. ............................................................................. 86 Ch. 3 Supplementary Figure 2–UCYN-A dominates nifH expression in the smaller size fraction of ESP samples (0.22-1μm)......................................................................... 87 Ch. 3 Supplementary Figure 3–Non-UCYN-A nifH BLAST hits in the larger size fraction of metagenomes and metatranscriptomes.. .................................................... 88 Ch. 3 Supplementary Figure 4–UCYN-A1 dominates nifH expression in the smaller size fraction of ESP samples (0.22-1μm)......................................................................... 90 Chapter 4–Symbiotic UCYN-A strains co-occurred with El Niño, relaxed upwelling, and varied eukaryotes over 10 years off Southern California Ch. 4 Figure 1–Sampling site location for the San Pedro Ocean Time-series (SPOT). ..............119 Ch. 4 Figure 2–UCYN-A 16S sequences (A) and Braarudosphearea 18S sequences (B) from SPOT are phylogenetically identical to or very close to the reference sequences for UCYN-A1, UCYN-A2, and hosts shown to associate with these symbionts. .................................................................................................. 120 Ch. 4 Figure 3–UCYN-A1 is present at SPOT under El Niño conditions, with relaxed upwelling, higher temperatures, and lower biomass (A-D); these patterns are similar but weaker for UCYN-A2 (note larger p-values and error bars). ............................. 121 Ch. 4 Figure 4–UCYN-A1 relative abundances correlate positively with host organisms, UCYN-A2 host/symbionts, MEI, and SST, and negatively with upwelling indices (BEUTI and CUTI) and indirect indicators of upwelling, including bacterial production, East-West Ekman transport, and chlorophyll concentrations. ............................................................................................................................ 122 Ch. 4 Figure 5–UCYN-A co-occurs with a variety of 18S taxa at the SPOT surface, notably including Lepidodinium, hypothesized to be a predator (pink triangle, center). .......... 123 x Ch. 4 Figure 6–UCYN-A2 is less tightly coupled to its prymnesiophyte host than the UCYN-A1 symbiosis. .................................................................................................. 124 Ch. 4 Figure 7–UCYN-A1 and its prymnesiophyte host are tightly coupled throughout the SPOT dataset. .................................................................................................... 125 Ch. 4 Supplementary Figure 1–Rarefaction curves of AE samples with <100 unique 18S ASVs (species) from the SPOT surface (A) and DCM (B). .................................... 131 Ch. 4 Supplementary Figure 2–eLSA networks constructed using interpolated, non-CLR transformed data from the SPOT surface (5m depth) miss interactions between UCYN-A and other taxa (compare to Figure 5). ......................................................... 132 Ch. 4 Supplementary Figure 3–Principal component analysis of environmental variables at the SPOT surface, overlaid with UCYN-A1 relative abundance and host presence/ absence (A) and UCYN-A2 relative abundance and host presence/ absence (B), show that temperature and MEI associate strongly with high relative abundance of UCYN-A1, but less strongly with that of UCYN-A2............................................................................................................... 135 Ch. 4 Supplementary Figure 4–Schematic of a hypothetical food web between UCYN-A ASVs and the 18S taxa with which they co-occur at the SPOT surface (see Figure 5).................................................................................................................. 136 Ch. 4 Supplementary Figure 5–Relative abundances of UCYN-A 16S sequences linearly correlate with relative abundances of host 18S sequences. ........................................... 139 Ch. 4 Supplementary Figure 6 .................................................................................................... 140 Ch. 4 Supplementary Figure 7 .................................................................................................... 141 Ch. 4 Supplementary Figure 8–Relative abundance of UCYN-A1 (UCYN-A SPOT ASV1) (A), UCYN-A SPOT ASV6 (B), and UCYN-A2 (UCYN-A SPOT ASV5) (C) at the SPOT surface (black) and DCM (green) over time............ 143 Ch. 4 Supplementary Figure 9–Relative abundance of UCYN-A ASVs in the larger size fraction over depth in A) July 2008, the date UCYN-A1 reached its maximum relative abundance, and B) July 2009, the date UCYN-A1 appeared at 890m. ........ 144 Ch. 4 Supplementary Figure 10–UCYN-A ASV6 does not co-occur with any Braarudosphaera ASV across the SPOT dataset at 5m....................................................... 145 Ch. 4 Supplementary Figure 11–Mixed mock communities contain even proportions of 16S (right) and 18S (middle) sequences. ............................................................ 146 xi Dissertation Abstract Marine microbes cycle nutrients on a massive scale. For example, nitrogen is fixed for the community by rare prokaryotes known as diazotrophs. Microbes are frequently studied in ways that miss fine-scale dynamics. Environmental studies of diazotrophs specifically sample the nifH gene, a component of the enzyme responsible for nitrogen fixation. I strive to contextualize marine microbial communities in the Southern California Bight. First, I compare the small-scale effects of space and time on two size fractions of microbes. I show that at scales of kilometers and days, space and time have similar effects on destabilizing the community, that the larger size fraction is patchy over a homogeneous background of smaller microbes, and that changes in environmental parameters more strongly influence the larger size fraction. I then investigate diazotroph diversity and activity. I contextualize nifH expression, showing it is transcribed at higher rates than a protein that is orders of magnitude more abundant and highly expressed. Over 98% of nifH transcripts were from the symbiotic diazotroph UCYN-A; UCYN-A2 is particularly active. Lastly, I contextualize UCYN-A relative abundance within the San Pedro Ocean Timeseries (SPOT). UCYN-A1, the “open ocean” ecotype, is brought to California by tropical waters that are stronger during El Niño and weaker during seasonal upwelling. Both UCYN-A1 and UCYN-A2 co-occur with Lepidodinium, a hypothesized predator. In contrast to UCYN-A1, UCYN-A2 has a weaker relationship with its host. Studying the wider context of microbes in the Southern California Bight is particularly important as climate change alters their dynamics. 1 Chapter 1– Introduction Background and importance Marine microbes – networks of bacteria, viruses, and single-celled plants and animals that live in the ocean – are vitally important to ocean processes (e.g. Fuhrman et al., 2015; Sunagawa et al., 2020). These tiny organisms are microscopic (Sunagawa et al., 2020), and many have short generation times of <24 hours (Long et al., 2021). Yet they are often studied at scales that are disproportionate to their small size, and that might miss fine-scale dynamics (Sunagawa et al., 2020), perhaps due to their global importance. Collectively, microbes have impacts on a massive scale, due to their sheer numbers: a single milliliter of seawater holds ~10^7 viruses, ~10^6 prokaryotes, and ~10^3 protists. Not only do these organisms serve as the base of the marine food web (e.g. Falkowski et al 2008, Sunagawa et al., 2020), but they also convert chemical elements (e.g. carbon, oxygen, nitrogen, and phosphorus) through biogeochemical cycles that span ocean basins. For example, microbes produce about half the oxygen in the atmosphere and absorb carbon dioxide through photosynthesis (Field et al., 1998), thereby mitigating the effects of climate change (e.g. Falkowski et al., 2008). In turn, this process is fertilized by nitrogen, about half of which is made bioavailable–“fixed”, i.e. converted from an abundant, but inert form (N2 gas) into ammonium (NH4)– by rare prokaryotes known as diazotrophs (Galloway et al., 2004; Turk-Kubo et al., 2023; Zehr and Capone, 2023). Much like megafauna, microbes interact with one another in complex, inextricable ways. While some microbes produce their own energy through photosynthesis, others graze on other microbes (e.g., Fuhrman et al., 2015; Sunagawa et al., 2020), while still others, such as many dinoflagellates, can switch between either metabolism (e.g. Millette et al., 2023). Some 2 microorganisms, like the unicellular diazotroph UCYN-A, live in symbiosis with a host organism in a mutually beneficial relationship (Hagino et al., 2009; Thompson et al., 2012). Others, like the order Syndiniales, are parasitic (e.g. Jephcott et al., 2016; Sunagawa et al., 2020). Focusing on one organism, especially in isolation, can miss this greater, interactive network. Methodology Marine microbes are often studied across large spatial transects and coarse temporal resolutions e.g. sampled at stations hundreds of km apart at sea, or in a monthly time series. These scales are often set by practical considerations, and although they can provide insights into the global significance of microbial activity, they miss important community dynamics given the small scales of microbial interactions (Sunagawa et al., 2020) and their short generation times (Long et al., 2021). Samples for large-scale marine microbial studies have been collected across whole ocean basins, famously including the Tara Oceans Project (e.g. Sunagawa et al., 2020). These transects span across ocean provinces and hydrographic features known to shape microbial community composition by changing nutrient availability and water temperature (e.g. Sunagawa et al., 2020; Hörstmann et al., 2022; James et al., 2022), including medium-scale features like fronts and eddies (e.g., Claustre, 1994; Martiny et al., 2006; Follows et al., 2007; McGillicuddy et al., 2007; Lévy et al., 2015; Benavides et al., 2021; Hörstmann et al., 2022). Across time, samples for marine microbial ecology are often collected over a monthly basis for years, or even decades, to understand seasonal variations in communities. The most famous microbial time series include the Bermuda Atlantic Time Series (BATS) (Steinberg et al., 2001), the Hawaii Ocean Time Series (HOTS) (Karl and Church, 2014; Karl and Church, 2017), and our sample site, the San Pedro Ocean Time-series (SPOT) (Cram et al., 2015; Yeh and Fuhrman, 2022). Each 3 sample in these datasets is like a snapshot into what the community is doing at the time the sample was collected, while the time-series as a whole represents a “movie” comprised of the individual shots (Fletcher-Hoppe et al., 2023). Following sample collection, sequencing the DNA or RNA of the microbial community can allow insights into whole community composition and activity. In these techniques, sea water samples are typically sieved through filters with different pore sizes, allowing for organisms to be separated into different size classes (for example, an 80µm pre-filter, followed by a 1µm filter allows collection of all organisms with a diameter from 1-80µm). Nucleic acids (DNA or RNA) are then extracted, and can be processed in different ways. Metagenomics is the study of all of the DNA in all of the microbes collected, while metatranscriptomics looks at all of the genes expressed in the sample by processing the RNA of the whole community (e.g. Sunagawa et al., 2020). In “tag sequencing,” polymerase chain reaction (PCR) is used to target a single gene. For example, nifH, a subunit of the nitrogenase enzyme that fixes nitrogen, is commonly used as a marker gene for diazotrophs, the only organisms that have this gene (e.g. Zehr et al., 1998). A subunit of the ribosomal RNA gene serves as a common marker gene for the whole community, as both prokaryotes and eukaryotes have variants of this gene (SSU rRNA; 16S for prokaryotes, 18S for eukaryotes). Our lab has developed “universal” PCR primers that allows amplification of the 16S and 18S rRNA genes from all three domains of life (Parada et al., 2016), as well as a pipeline to process these sequences into amplicon sequence variants (ASVs), a proxy for microbial species, that differ from one Ir by as little as one base pair (McNichol et al., 2021). 4 Nitrogen fixers Tag sequencing of the nifH gene has greatly expanded our knowledge of diazotrophs, although most studies are limited to just this gene. Traditional paradigms of nitrogen fixation held that this process only occurred in N-limited, tropical gyres and was only carried out by a handful of well-studied prokaryotes, such as Trichodesmium, Crocospharea watsonii, and diatom-diazotroph associations (e.g. Turk-Kubo et al., 2023; Zehr and Capone, 2023). However, tag sequencing of the nifH gene has revealed that nitrogen fixation occurs in N-replete environments, and is carried out by a broader range of prokaryotes than previously studied (e.g. Turk-Kubo et al., 2023). Despite these advances, the relative abundances and transcriptomic activity of diazotrophs compared to whole microbial communities are poorly understood (e.g. Needham et al., 2018; Fletcher-Hoppe et al., 2023), particularly over day-night cycles and in temperate regions, where most nitrogen fixers are not among those tropical cyanobacteria that have been studied extensively (e.g. Zehr and Capone, 2020; Turk-Kubo et al., 2023). For example, a few studies have attempted to link the symbiotic diazotroph UCYN-A to organisms other than its known host, or to environmental parameters in the Southern California Bight (Turk-Kubo et al., 2021; Fletcher-Hoppe et al., 2023), although many studies have focused on this organism. UCYN-A has garnered much attention due to its streamlined metabolism (Tripp et al., 2010), known intimate partnership with a haptophyte host (Thompson et al., 2012), and broad, global distribution (e.g. Martinez-Perez et al., 2016). UCYN-A has previously been detected in the Southern California Bight (Hamersley et al., 2011; Turk-Kubo et al., 2021) and has been shown to be active and abundant in many of the samples analyzed here (Needham et al., 2018; Fletcher-Hoppe et al., 2023). UCYN-A is comprised of at least four clades (Thompson et al., 2014), two of which have been studied extensively. UCYN-A1, the most commonly studied 5 clade, is known as the “open ocean” ecotype due to its ability to proliferate in open waters (e.g. Bombar et al, 2014). In contrast, UCYN-A2, the “coastal ecotype”, prefers waters near shore (Bombar et al., 2014), is larger in size, and hosts contain more symbionts per cell (e.g. Zehr et al., 2016). UCYN-A fixes nitrogen constitutively during the day (Church et al., 2005; Thompson et al., 2014; Muñoz-Marín et al., 2019; Gradoville et al., 2021; Landa et al., 2021), exchanging this fixed N2 for carbon intermediates produced by the host (Thompson et al., 2012). Some have postulated that the microbe is evolving to carry out only this function: UCYN-A may eventually become a “diazoplast” organelle within the host (Zehr et al., 2016). Many environmental studies of UCYN-A and other diazotrophs focus solely on the nifH gene, thereby missing the wider context of these organisms. Further, the PCR primers used to study the nifH gene are biased against non-cyanobacterial diazotrophs (NCDs) (Delmont et al., 2022), which include archaea, alpha-, beta-, and gamma-proteobacteria, and sulfur-reducing organisms (Zehr et al., 1998; Capone and Zehr, 2020; Turk-Kubo et al., 2023; Capone and Zehr, 2023). Although nitrogen fixation by NCDs has not been detected (Turk-Kubo et al., 2023), NCDs may be important contributors to nitrogen fixation, given their high relative abundance in metagenomes and metatranscriptomes (Delmont et al., 2022). Few studies compare cyanobacterial diazotrophs to the relative contributions of these other diazotrophs that PCR primers may miss (e.g. Delmont et al., 2022). Even fewer studies investigate interactions between diazotrophs and 18S taxa (Dugenne et al., Deng et al., 2020), or diazotroph genes other than nifH. 6 Research questions In my dissertation, I strive to contextualize marine microbial communities in the Southern California Bight, particularly diazotrophs, using appropriate sampling scales for these organisms. My first chapter provides context for our San Pedro Ocean Time-series (SPOT). Here, I investigate fine-scale spatial and temporal dynamics that our monthly sampling regime may miss, comparing and contrasting the effects of space and time within two size fractions of marine microbial communities. I attempt to answer: • How does SPOT compare with other sites in the transect we sampled? • How does the community vary over a 11km x 14km spatial transect compared to at the same location over five days? • How do space and time influence the larger size fraction of marine microbes sampled (1-80µm), which contains eukaryotes and particle-attached prokaryotes, vs. the smaller size fraction of organisms (0.22-1µm), which mostly contains free-living prokaryotes? • How do environmental variables combine over our spatial transect to shape the community in each size fraction? I then use an un-targeted method to investigate diazotroph diversity and activity in the context of the wider microbial community, comparing metatranscriptomes to metagenomes to infer relative activity of organisms with different abundances. Avoiding targeted PCR of the nifH gene allows me to both accurately investigate NCD dynamics, which many studies miss due to primer bias, and to contextualize the importance of nifH expression by comparing it to transcription of other genes. I inquire: 7 • How does nifH transcription in our samples compare to that of an abundant, highly expressed cyanobacterial gene that is vital to photosynthesis? • What organisms conduct nifH transcription within the community? • How do the relative abundances of nifH transcription and abundance compare between UCYN-A clades? • How does gene expression vary within the UCYN-A genome? Lastly, I contextualize UCYN-A within the larger marine microbial community over 10 years of 16S/ 18S tag sampling at the San Pedro Ocean Time-series. This large dataset allows greater insight into the UCYN-A autecology and its interactions with other organisms. I used statistical analysis of these data to examine: • What abiotic factors govern UCYN-A presence and relative abundance within the Southern California bight? • What eukaryotic organisms interact with UCYN-A? • How does the host-symbiont relationship differ amongst UCYN-A clades? Studying the wider context of microbes in the Southern California Bight becomes particularly important in light of climate change. As extreme weather events become stronger and more frequent, marine microbial dynamics in the region may change, altering nitrogen inputs to the wider community. 8 References Benavides, M., L. Conradt, S. Bonnet, I. Berman-Frank, A. Petrenko, and A. M. Doglioli. Fine scale sampling unveils diazotroph patchiness in the South Pacific Ocean. ISME Communications 1: 1-3. doi:.1038/s43705-021-00006-2 Bombar, D., P. Heller, P. Sanchez-Baracaldo, B. J. Carter, and J. P. Zehr. 2014. Comparative genomics reveals surprising divergence of two closely related strains of uncultivated UCYN-A cyanobacteria. ISME J 8: 2530–2542. doi:10.1038/ismej.2014.167 Church, M. J., C. M. Short, B. D. Jenkins, D. M. Karl, and J. P. Zehr. 2005. Temporal Patterns of Nitrogenase Gene (nifH) Expression in the Oligotrophic North Pacific Ocean. Appl Environ Microbiol 71: 5362–5370. doi:10.1128/AEM.71.9.5362-5370.2005 Claustre, H., P. Kerhervé, J. C. Marty, L. Prieur, C. Videau, and J.-H. Hecq. 1994. Phytoplankton dynamics associated with a geostrophic front: Ecological and biogeochemical implications. issn: 0022-2402 52: 711–742. doi:10.1357/0022240943077000 Cram, J. A., L. C. Xia, D. M. Needham, R. Sachdeva, F. Sun, and J. A. Fuhrman. 2015. Crossdepth analysis of marine bacterial networks suggests downward propagation of temporal changes. ISME Journal 9: 2573–2586. doi:10.1038/ismej.2015.76 Delmont, T. O., J. J. Pierella Karlusich, I. Veseli, and others. 2022. Heterotrophic bacterial diazotrophs are more abundant than their cyanobacterial counterparts in metagenomes covering most of the sunlit ocean. The ISME Journal 16: 927–936. doi:10.1038/s41396- 021-01135-1 Deng, L., S. Cheung, and H. Liu. 2020. Protistal Grazers Increase Grazing on Unicellular Cyanobacteria Diazotroph at Night. Frontiers in Marine Science 7. doi:10.3389/fmars.2020.00135 Dugenne, M., F. H. Freitas, S. T. Wilson, D. M. Karl, and A. E. White. 2020. Life and death of Crocosphaera sp. in the Pacific Ocean: Fine scale predator–prey dynamics. Limnology and Oceanography 65: 2603–2617. doi:10.1002/lno.11473 Falkowski, P. G., T. Fenchel, and E. F. Delong. 2008. The Microbial Engines That Drive Earth’s Biogeochemical Cycles. Microbial Ecology 320. 1034-1039. doi: 10.1126/science.1153213 Field, CB, Behrenfeld, MJ, Randerson, JT, and Falkowski, Paul. 1998. Primary production of the biosphere: integrating terrestrial and oceanic components. Science 281: 237–240. doi:10.1126/science.281.5374.237. PMID: 9657713 Fletcher-Hoppe, C., Y.-C. Yeh, Y. Raut, J. L. Weissman, and J. A. Fuhrman. 2023. Symbiotic UCYN-A strains co-occurred with El Niño, relaxed upwelling, and varied eukaryotes over 10 years off Southern California. ISME COMMUN. 3: 63. doi:10.1038/s43705-023- 9 00268-y Follows, M. J., S. Dutkiewicz, S. Grant, and S. W. Chisholm. 2007. Emergent Biogeography of Microbial Communities in a Model Ocean. Science 315: 1843–1846. doi:10.1126/science.1138544 Fuhrman, J. A., J. A. Cram, and D. M. Needham. 2015. Marine microbial community dynamics and their ecological interpretation. Nat Rev Microbiol 13: 133–146. doi:10.1038/nrmicro3417 Galloway, J.N., Dentener, F.J., Capone, D.G., and others. 2004. Nitrogen Cycles: Past, Present, and Future. Biogeochemistry 70: 153–266. Gradoville, M. R., D. Bombar, B. C. Crump, R. M. Letelier, J. P. Zehr, and A. E. White. 2017. Diversity and activity of nitrogen-fixing communities across ocean basins. Limnology and Oceanography 62: 1895–1909. doi:10.1002/lno.10542 Hagino, K., R. Onuma, M. Kawachi, and T. Horiguchi. 2013. Discovery of an Endosymbiotic Nitrogen-Fixing Cyanobacterium UCYN-A in Braarudosphaera bigelowii (Prymnesiophyceae). PloS ONE 8: e81749. doi:10.1371/journal.pone.0081749 Hamersley, M. R., K. A. Turk, A. Leinweber, N. Gruber, J. P. Zehr, T. Gunderson, and D. G. Capone. 2011. Nitrogen fixation within the water column associated with two hypoxic basins in the Southern California Bight. Aquatic Microbial Ecology 63: 193–205. doi:10.3354/ame01494 Hörstmann, C., P. L. Buttigieg, U. John, E. J. Raes, D. Wolf‐Gladrow, A. Bracher, and A. M. Waite. 2022. Microbial diversity through an oceanographic lens: refining the concept of ocean provinces through trophic‐level analysis and productivity‐specific length scales. Environmental Microbiology 24: 404–419. doi:10.1111/1462-2920.15832 James, C. C., A. D. Barton, L. Z. Allen, and others. 2022. Influence of nutrient supply on plankton microbiome biodiversity and distribution in a coastal upwelling region. Nat Commun 13: 2448. doi:10.1038/s41467-022-30139-4 Jephcott, T. G., C. Alves-de-Souza, F. H. Gleason, F. F. van Ogtrop, T. Sime-Ngando, S. A. Karpov, and L. Guillou. 2016. Ecological impacts of parasitic chytrids, syndiniales and perkinsids on populations of marine photosynthetic dinoflagellates. Fungal Ecology 19: 47–58. doi:10.1016/j.funeco.2015.03.007 Karl, D. M., and M. J. Church. 2014. Microbial oceanography and the Hawaii Ocean Time-series programme. Nat Rev Microbiol 12: 699–713. doi:10.1038/nrmicro3333 Karl, D. M., and M. J. Church. 2017. Ecosystem Structure and Dynamics in the North Pacific Subtropical Gyre: New Views of an Old Ocean. Ecosystems 20: 433–457. doi:10.1007/s10021-017-0117-0 Landa, M., K. A. Turk-Kubo, F. M. Cornejo-Castillo, B. A. Henke, and J. P. Zehr. 2021. Critical 10 Role of Light in the Growth and Activity of the Marine N2-Fixing UCYN-A Symbiosis. Front. Microbiol. 12: 666739. doi:10.3389/fmicb.2021.666739 Lévy, M., O. Jahn, S. Dutkiewicz, M. J. Follows, and F. d’Ovidio. 2015. The dynamical landscape of marine phytoplankton diversity. J. R. Soc. Interface. 12: 20150481. doi:10.1098/rsif.2015.0481 Long, A. M., S. Hou, J. C. Ignacio-Espinoza, and J. A. Fuhrman. 2021. Benchmarking microbial growth rate predictions from metagenomes. ISME J 15: 183–195. doi:10.1038/s41396- 020-00773-1 Martinez-Perez, C., W. Mohr, C. R. Loscher, and others. 2016. The small unicellular diazotrophic symbiont, UCYN-A, is a key player in the marine nitrogen cycle. Nat Microbiol 1: 16163. doi:10.1038/nmicrobiol.2016.163 Martiny, J. B. H., B. J. M. Bohannan, J. H. Brown, and others. 2006. Microbial biogeography: putting microorganisms on the map. Nat Rev Microbiol 4: 102–112. doi:10.1038/nrmicro1341 McGillicuddy, D. J., L. A. Anderson, N. R. Bates, and others. 2007. Eddy/Wind Interactions Stimulate Extraordinary Mid-Ocean Plankton Blooms. Science 316: 1021–1026. doi:10.1126/science.1136256 McNichol, J., P. M. Berube, S. J. Biller, and J. A. Fuhrman. 2021. Evaluating and Improving Small Subunit rRNA PCR Primer Coverage for Bacteria, Archaea, and Eukaryotes Using Metagenomes from Global Ocean Surveys. mSystems 6: 10.1128/msystems.00565-21. doi:10.1128/msystems.00565-21 Millette, N. C., Gast, R. J., Luo, J. Y., Moeller, H. V, Stamieszkin, K., Andersen, K. H., Brownlee, E. F., Cohen, N. R., Duhamel, S., Dutkiewicz, S., Glibert, P. M., Johnson, M. D., Leles, S. G., Maloney, A. E., Mcmanus, G. B., Poulton, N., Princiotta, S. D., Sanders, R. W., & Wilken, S. (2023). Mixoplankton and mixotrophy: future research priorities. Journal of Plankton Research, 45, 576–596. https://doi.org/10.1093/plankt/fbad020 Muñoz-Marín, M. del C., I. N. Shilova, T. Shi, H. Farnelid, A. M. Cabello, and J. P. Zehr. 2019. The Transcriptional Cycle Is Suited to Daytime N2 Fixation in the Unicellular Cyanobacterium “Candidatus Atelocyanobacterium thalassa” (UCYN-A). mBio 10: e02495-18. doi:10.1128/mBio.02495-18 Needham, D. M., E. B. Fichot, E. Wang, L. Berdjeb, J. A. Cram, C. G. Fichot, and J. A. Fuhrman. 2018. Dynamics and interactions of highly resolved marine plankton via automated high-frequency sampling. ISME Journal 12: 2417–2432. doi:10.1038/s41396- 018-0169-y Parada, A. E., D. M. Needham, and J. A. Fuhrman. 2016. Every base matters: assessing small subunit rRNA primers for marine microbiomes with mock communities, time series and global field samples. Environmental Microbiology 18: 1403–1414. doi:10.1111/1462- 2920.13023 11 Steinberg, D. K., C. A. Carlson, N. R. Bates, R. J. Johnson, A. F. Michaels, and A. H. Knap. 2001. Overview of the US JGOFS Bermuda Atlantic Time-series Study (BATS): a decade-scale look at ocean biology and biogeochemistry. Deep Sea Research Part II: Topical Studies in Oceanography 48: 1405–1447. doi:10.1016/S0967-0645(00)00148-X Sunagawa, S., S. G. Acinas, P. Bork, and others. 2020. Tara Oceans: towards global ocean ecosystems biology. Nat Rev Microbiol 18: 428–445. doi:10.1038/s41579-020-0364-5 Thompson, A., B. J. Carter, K. Turk-Kubo, F. Malfatti, F. Azam, and J. P. Zehr. 2014. Genetic diversity of the unicellular nitrogen-fixing cyanobacteria UCYN-A and its prymnesiophyte host. Environ Microbiol 16: 3238–3249. doi:10.1111/1462-2920.12490 Thompson, A. W., R. A. Foster, A. Krupke, B. J. Carter, N. Musat, D. Vaulot, M. M. M. Kuypers, and J. P. Zehr. 2012. Unicellular cyanobacterium symbiotic with a single-celled eukaryotic alga. Science (New York, N.Y.) 337: 1546. doi:10.1126/science.1222700 Tripp, H. J., S. R. Bench, K. A. Turk, R. A. Foster, B. A. Desany, F. Niazi, J. P. Affourtit, and J. P. Zehr. 2010. Metabolic streamlining in an open-ocean nitrogen-fixing cyanobacterium. Nature 464: 90–94. doi:10.1038/nature08786 Turk-Kubo, K. A., M. R. Gradoville, S. Cheung, F. M. Cornejo-Castillo, K. J. Harding, M. Morando, M. Mills, and J. P. Zehr. 2023. Non-cyanobacterial diazotrophs: global diversity, distribution, ecophysiology, and activity in marine waters. FEMS Microbiology Reviews 47: fuac046. doi:10.1093/femsre/fuac046 Turk-Kubo, K. A., M. M. Mills, K. R. Arrigo, G. van Dijken, B. A. Henke, B. Stewart, S. T. Wilson, and J. P. Zehr. 2021. UCYN-A/haptophyte symbioses dominate N2 fixation in the Southern California Current System. ISME Communications 1. doi:10.1038/s43705- 021-00039-7 Yeh, Y.-C., and J. A. Fuhrman. 2022. Contrasting diversity patterns of prokaryotes and protists over time and depth at the San-Pedro Ocean Time series. ISME Communications 2. doi:10.1038/s43705-022-00121-8 Zehr, J. P., and D. G. Capone. 2023. Unsolved mysteries in marine nitrogen fixation. Trends in Microbiology S0966842X23002354. doi:10.1016/j.tim.2023.08.004 Zehr, J. P., M. T. Mellon, and S. Zani. 1998. New Nitrogen-Fixing Microorganisms Detected in Oligotrophic Oceans by Amplification of Nitrogenase (nifH) Genes. Applied and Environmental Microbiology 64: 3444. Zehr, J. P., Shilova, I.N, H. M. Farnelid, Muñoz-Maríncarmen Maria Del, and K. A. Turk-Kubo. 2016. Unusual marine unicellular symbiosis with the nitrogen-fixing cyanobacterium UCYN-A. Nature Microbiology 2. doi:10.1038/nmicrobiol.2016.214 12 Chapter 2–X Marks the SPOT: Spatial and daily variation in marine microbial communities at the San Pedro Ocean Time-series Abstract Marine microbial communities are vital to biogeochemical cycling on Earth. These microbes are frequently compared across transects that span ocean basins or at monthly intervals that capture variation over decades, which may obscure dynamics that occur on smaller scales. Few studies compare microbial community variability over space vs. time, and even fewer examine all three domains of life at fine spatio-temporal scales. We sampled marine microbial communities over an X-shaped spatial transect that spanned 14.5km at most and across five consecutive days at the San Pedro Ocean Time-series (SPOT). We used “universal” primers to capture all three domains of life in two size fractions (0.22-1µm and 1-80µm). We observed that space and time have similar effects on community variability within each size fraction on the small scales sampled here. Further, communities in the transect were more similar to one another than not. However, community dissimilarity is higher in the larger size fraction than the smaller: larger and particle-attached marine microbes were patchy over a homogenous background of smaller microbes. Lastly, the larger size fraction of marine microbes, particularly heterotrophic bacteria therein, is more influenced by environmental variation than the smaller. These smallscale patterns provide important context to the entire SPOT time series, whose monthly sampling regime is too coarse to observe them. 13 Introduction Marine microbial communities – networks of bacteria, viruses, and single-celled organisms that live in the oceans – are vital as the base of global food webs due to their capacity to shape biogeochemical cycles on a massive scale (Falkowski et al., 2008). For example, marine microbes carry out over half the primary production on the planet (Field et al., 1998) and mitigate climate change by exporting fixed carbon dioxide to the depths of the oceans (Guidi et al., 2015). Marine microbial communities are frequently sampled on coarse spatial or temporal scales, e.g., kilometers or weeks apart (Martiny et al., 2006). To understand the structure and function of marine microbial communities on a global scale, communities are often probed at regular distances across ocean basins or along longitudinal transects, most notably including the Tara Oceans sampling expedition (e.g., Sunagawa et al., 2019). Communities are quite dissimilar over transects that span entire ocean basins, due to variety in temperatures and nutrient availability (e.g., Sunagawa et al., 2019; Hörstmann et al., 2022; James et al., 2022). Within ocean basins, marine microbial communities may be shaped by hydrographic features such as fronts (e.g., Claustre, 1994; Martiny et al., 2006; Follows et al., 2007; Lévy et al., 2015; Hörstmann et al., 2022) and eddies (e.g., McGillicuddy et al., 2007; Benavides et al., 2021; Follows et al., 2007; Lévy et al., 2015). To understand how community structure and function changes over a seasonal or inter-annual basis, communities are often sampled regularly over the course of several years or even decades (e.g., Steinberg et al., 2001; Karl and Church, 2014; Cram et al., 2015). For example, communities at the San Pedro Ocean Time-series (SPOT) show regular patterns of stability: communities sampled at twelve-month intervals are most similar to one another while communities sampled at six-month intervals are least similar (e.g., Fuhrman et 14 al., 2015). Notably, microbial communities in the smaller size fraction of SPOT samples (0.22- 1µm) are more stable than those from larger size fractions (1-80µm) (Yeh and Fuhrman, 2022). Few papers examine microbial diversity across both space and time (e.g., Hewson et al., 2006), and even fewer examine prokaryotic and eukaryotic communities in tandem (e.g., Brown et al., 2009; Yeh and Fuhrman, 2022) over both space and time (Chénard et al., 2019; James et al., 2022). The spatial and temporal scales at which most studies are conducted (spanning ocean basins, sampling monthly or quarterly) might not be suitable to capture the full dynamics of marine microbes. In general, community dynamics result from reproduction, death, and influx or exodus of community members (Martiny et al., 2006; Fuhrman et al., 2015). Due to the small size of marine microbes (Sunagawa et al., 2019), interactions can take place across fractions of a centimeter (Long and Azam 2001; Seymour et al., 2004). Due to the short generation times of marine microbes, which may be as fast as hours or days (e.g., Long et al., 2020), sampling monthly or quarterly might overlook growth and death dynamics in marine microbial communities. Here, we sampled the marine microbial communities at the San Pedro Ocean Time-series (SPOT) at a small-scale resolution over space and time. Samples from two size fractions (0.22- 1µm and 1-80µm) were collected from an 11km x 14km transect surrounding SPOT and over five consecutive days, spatio-temporal scales that are traditionally considered “small scale” (Martiny et al., 2006; Benavides et al., 2021). Both prokaryotes and eukaryotes were probed using “universal” primers that capture all domains of life (Parada et al., 2016), and resolved into Amplicon Sequence Variants (ASVs) which differ by as little as one base pair (Callahan et al., 2016). The smaller size fraction corresponds to free-living prokaryotes and small eukaryotes, 15 while the larger size fraction includes particle-attached prokaryotes and larger eukaryotes (Yeh and Fuhrman, 2022). Previous studies have observed that the smaller size fraction is more similar over time than the larger (Yeh and Fuhrman, 2022), and that protist community variability was higher over 12 consecutive days of sampling than samples 2-14km apart in the same region (Lie et al., 2013). We therefore hypothesize that community dissimilarity will be highest amongst samples from the larger size fraction collected daily, lowest in samples from the smaller size fraction collected across the spatial transect, and intermediate in samples from the larger size fraction collected over space or samples from the smaller size fraction collected daily. Materials and Methods Sample collection and processing To quantify spatial variation in the SPOT community, eight sites around SPOT were sampled on the 4th day of daily sampling, April 5th, 2018. Including SPOT, nine stations 0.05º latitude or longitude apart from one another were sampled in the order corresponding to station numbers in Figure 1A; note that Station 9 is the location of SPOT. Spatial sampling was completed within 2.5 hours (approximately 7:30-9:50am) to avoid diel influences on community structure. To quantify daily variation in the San Pedro Time Series (SPOT) microbial community, samples were collected from the SPOT location (33.55ºN, 118.4ºW; Figure 1A) between 10:30- 11am on five consecutive days (April 2nd -6 th, 2018). Throughout sample collection, local currents flowed southeast through our study sites, along the coastline, as indicated by the National Ocean and Atmospheric Administration’s (NOAA’s) Ocean Surface Current Simulator (OSCURS) (https://oceanview.pfeg.noaa.gov/oscurs/#). 16 4L of water were collected at each spatial or temporal sampling point. Samples were brought back to the Wrigley Marine Science Center (WMSC) facility near Two Harbors, Catalina. Water from each 4L sample was filtered through two filtering set-ups, such that 2L from each sample went through each. Filtering set-ups included an 80µm mesh, removing mesoplankton, a 1µm, 47mm diameter AE filter (CAT# 61631, Pall, Port Washington, NY; larger size fraction), and finally a 0.22-1µm Durapore filter (Sterivex; CAT#SVGVL10RC, EDM/Millipore, Burlington, MA) (Figure 1B). Filters were immediately preserved in 1-2mL RNALater (CAT#AM7020, Invitrogen/ThermoFisher Scientific) and stored at -80ºC until extraction. Metadata (i.e., temperature, dissolved oxygen, and salinity) were measured with a Profiling Natural Fluorometer (PNF 300, BioSpherical Instruments, San Diego, CA), deployed to 40m depth during daily sample collection. To minimize time spent at each site in the spatial transect, a handheld YSI probe (Xylem, Inc., Yellow Springs, OH) was dipped below the surface during spatial sampling. Inorganic nitrogen ([NO2+NO3]) and phosphate ([PO4]) concentrations were measured via LACHAT spectrophotometer QuickChem 8500 Series 2 at the Marine Science Institute at the University of California, Santa Barbara. However, all nitrogen measurements were below limits of detection of this instrument (0.2µM for [NO2+NO3]), and were excluded from further analyses. (All phosphate measurements were above the limits of detection (0.1µM for [PO4]), and were included.) Bacterial production was measured via incorporation of tritiated thymidine and leucine as described in Fuhrman et al. (2006). Counts of viruses and bacterial cells were enumerated via SYBR slides as described in Patel et al. (2007) (Table 1). 17 DNA extraction DNA was extracted from the larger size fraction (AE filters, 1-80µm) using a modified version of the phenol-chloroform procedure described in Lie et al. (2013). To prevent the possibility of selective loss of DNA in RNALater, the RNALater from each AE filter was rinsed off the filters, collected, and filtered and desalted through a 3K Amicon tube (CAT# UFC500396, Millipore/ Sigma Aldrich, Tullagreen, Ireland). The resulting concentrate was processed in parallel to the filter. Samples and their corresponding RNALater were bead beaten with 100µL of 0.5mm zirconia/ silica beads (CAT#11079105, BioSpec Products) on a vortexer for one minute, then incubated in a 70ºC water bath for five minutes. The process was repeated a total of four times. Samples were centrifuged to separate the lysis buffer from the beads and filter remnants, which were then rinsed with TE to collect any remaining DNA. The supernatant and TE rinses of each wash and RNALater fraction were incubated with 10% CTAB and 2.5M NaCl at 70ºC for 10 minutes prior to phenol chloroform extraction, as described in Lie et al. (2013). DNA was extracted from the smaller size fraction (Durapore filters, 0.22-1µm) via the Qiagen DNA/RNA/Protein AllPrep kit (CAT# 80004), with modifications: samples were bead beaten on a tissue lyser using 0.1mm glass beads and RLT buffer (kit reagent) with beta mercaptoethanol (BME). Lysate was removed via centrifugation. The remaining extraction steps processed according to kit instructions. DNA sequencing and processing The V4-V5 hypervariable region of the 16S/18S rRNA genes was amplified from these extracts using the universal primers (515Y/926R) developed by Parada et al. (2016), which amplify sequences from both eukaryotic and prokaryotic ribosomal RNA genes (Yeh et al., 2021; Yeh and Fuhrman, 2022). Samples were sequenced on either a HiSeq platform to a target depth 18 of 500,000 sequences per sample for Durapore filters or a MiSeq platform to 100,000 sequences per sample for AE filters. Mock marine communities, in which the relative abundances of 16S or 18S sequences are known, were included in the sequencing runs to validate PCR and sequencing (Yeh et al., 2018). Sequences were processed into Amplicon Sequence Variants (ASVs), which differ by as little as a single base pair, using DADA-2 implemented in the QIIME2 pipeline (Callahan et al., 2016; Estaki et al., 2020), with scripts available at <github.com/jcmcnch/eASV-pipeline-for-21 515Y-926R> (McNichol et al., 2021). Prokaryotic and eukaryotic ASVs were taxonomically classified using SILVA 132 in July-August 2022. ASVs corresponding to chloroplasts were subsetted and classified via PhytoRef2. ASVs corresponding to multicellular metazoans were removed for the analyses reported here. Prokaryotic and Eukaryotic ASVs were normalized and combined as described in Fletcher-Hoppe et al. (2023; Supplemental Methods), assuming a 2x bias against longer 18S rRNA sequences (Yeh et al., 2018). One sample from the smaller size fraction and three samples from the larger size fraction were excluded from analyses because they contained too few sequences to capture sufficient diversity of ASVs (rarefaction curves for these samples did not approach an asymptote; Figure S1). Data analysis Statistical analyses and data visualizations were conducted in R v4.2.2 with the packages “ggplot2” (Wickham, 2016), “phyloseq” (McMurdie and Holmes, 2013), and “vegan” v. 2.5-6. Bray-Curtis similarity and dissimilarity, which were most often used to compare pairwise differences between samples, was calculated via vegdist() in “vegan”. Environmental differences between the samples were analyzed using Euclidean distances, which was also calculated via vegdist(), but on data that had been centered and scaled via scale() in base R. Physical distances 19 between the sites were calculated using distHaversine() in the package “geosphere”. To account for the disproportionate influence of rare taxa (and possibly rare PCR or sequencing errors) on alpha diversity (Shannon 1948; Caron and Countway., 2009), ASVs with fewer than 100 counts were removed from Durapore samples and ASVs with fewer than 10 counts were removed from AE samples, such that 0.5-2% of the sequences were removed from each sample before alpha diversity was analyzed. (Note that different numbers of ASVs were removed from Durapore vs. AE samples due to differences in sequencing depth with each sample type.) Non-metric Multidimensional Scaling (NMDS) plots, used to visualize differences in samples, were clustered via k-means partitioning using hclust() and visualized using cascadeKM() in “vegan”. To evaluate the effects of environmental and geographic distance on different metabolic groups, ASVs from the larger size fraction of marine microbial communities were grouped based on the predominant metabolic strategy of each phylum. The five groupings included in this study were Archaea (i.e. ASVs assigned to the domain Archaea), Cyanobacteria (ASVs assigned to the class Cyanobacteria), heterotrophic bacteria (bacterial ASVs not in the class Cyanobacteria), photosynthetic eukaryotes (belonging to the phyla Archaeplastida, Chlorophyta, Choanoflagellida, Cryptophyta, Ochrophyta, Rhodophyta, Stramenopiles , or Streptophyta), or non-photosynthetic eukaryotes (i.e. heterotrophic or mixotrophic eukaryotes, belonging to the phyla Apicomplexa, Centroheliozoa, Cercozoa, Ciliophora, Discoba, Dinoflagellata, Excavata, Fungi, Hacrobia, Haptophyta, Katablepharidophyta, Mesomycetozoa, Opalozoa, Opisthokonta, Picozoa, Pseudofungi, Rhizaria, Sagenista, Radiolaria, or Telonemia). Eukaryotic ASVs were defined as “non-photosynthetic” if most members of the phylum were mixotrophs or 20 heterotrophs (see Materials and Methods). Relative abundances and Bray-Curtis distance were calculated separately for each metabolic group. In addition to the ASV-level analyses described above, ASVs in the larger size fraction were aggregated at the genus level using tax_glom() in PhyloSeq (McMurdie and Holmes, 2013) to avoid the potential of over-splitting eukaryotic ASVs (Caron et al., 2012; Gong and Marchetti, 2019; Schloss, 2021). Select correlations between genera and the environmental parameters were re-examined using genus-level relative abundance data. Results Effects of space vs. time on community stability Distance-decay plots show that Bray-Curtis similarity in the whole community declined to similar values over 14.5km between sample sites and within five days between sample collection in each of the size fractions we sampled (Figure 2). In the smaller size fraction, duplicate samples (collected 0km apart or on the same day) were 85% similar to one another on average; Bray-Curtis similarity decreased to 70% similar as the distance between sampling sites increased to >3.5km or >1 day apart. In the larger size fraction, duplicates were about 75% similar to one another on average. Bray-Curtis similarity decreased to about 50% >5km between sampling sites or >3 days between sample collection (Figure 2). Community stability of large vs. small size fractions In communities from both size fractions, Bray-Curtis similarity remained >0.5 across the spatial transect. Total community Bray-Curtis similarity decreased from ~0.85 between duplicates (samples collected 0km or 0 days apart) to ~0.7 in the smaller community and from 21 ~0.75 between duplicates to ~0.5 in the larger community, as either distance between sampling sites or days between sample collection increased (Figure 2, also see discussion above). In addition, Bray-Curtis similarity compared to SPOT remained higher in the smaller size fraction compared to the smaller. Bray-Curtis similarity to communities at SPOT remained >0.7 up to 4km away from that central location (Figure 3A), with the exception of one station discussed more below. The larger size fraction reached lower levels of Bray-Curtis similarity, ~0.6 compared to SPOT, as distance from SPOT increased to 9.3km (Figure 3B). In ordination plots, the smaller size fraction of marine microbial communities was too homogenous to cluster via kmeans clustering (Figure 4A). In contrast, the larger size fraction clustered neatly by latitude of the sampling site (Figure 4B). Similarly, Jaccard Dissimilarity, a metric based on presence/ absence data, was higher in the larger size fraction over both space and time. In all sample types, turnover comprised a higher proportion of dissimilarity than nestedness (Table 2). Alpha diversity was lower in the smaller size fraction of marine microbial communities over both space and time. Species accumulation curves for the smaller size fraction were nearly asymptotic and show ~750 microbial ASVs were found within five samples collected across space or within five days of sampling (Figure 5A-B). In contrast, these curves continue to rise for the larger size fraction, indicating that the number of sites included in our spatial transect was insufficient to capture the full richness of the larger size fraction (~1300 ASVs). By similar logic, the number of days sampled were even less sufficient to observe the full diversity of ASVs in this size fraction (Figure 5A-B). Shannon Index, a common ecological measure of alpha diversity (Shannon, 1948), was greater in the larger size fraction over both space and time with high statistical significance (p=4.11e-05, p=0.008, respectively). However, Shannon Index was comparable over space vs time within each size fraction (Figure 5C). 22 Similarity percentage (SIMPER) analysis revealed that three ASVs contributed to >1% of Bray-Curtis distance over space in the small size fraction: one SAR11 ASV (discussed below), an ASV from Bathycoccus and a Synechococcus ASV (Figure S3B-C). Over time, the same three ASVs, plus an ASV from Actinobacteria, contributed to >1% of Bray-Curtis distance in the smaller size fraction. (Notably, the same SAR11 ASV that reached 40% of the small size fraction community of Station 1 reached >25% relative abundance on two days of daily sampling (Figure S5A).) In the larger size fraction of marine microbes, four ASVs contributed to >1% of BrayCurtis distance over space, a Phaeocystis ASV, a Strombidiida ASV, and two Synechococcus ASVs, including the same Synechococcus ASV from the smaller size fraction (Figure S4). Over 1% of Bray-Curtis distance in the larger size fraction was driven by an ASV from Ciliophora, two ASVs from Synechococcus (including the same ASVs that drove >1% of Bray-Curtis distance over space, and one that also contributed >1% of Bray-Curtis distance in the smaller size fraction over time), and an ASV from Verrucomicrobia (Figure S5B). At Station 1, the sampling site nearest to Catalina Island, Bray-Curtis similarity to SPOT dropped to 0.6 in the smaller size fraction of marine microbes. All other sites were ~0.7 compared to SPOT (Figure 3A). In ordination plots of the smaller community, Station 1 did not cluster neatly with the other sampling sites (Figure 4A). Stacked barplots of the 20 most abundant microbial orders in each sample show that SAR11 dominates the ASVs in this location more so than at any other location. At the order level, SAR11 comprised ~50% of the sequences at Station 1 and ~30% of the sequences at all other sites (Figure S2). At the ASV level, a single ASV of SAR11 made up ~40% of the whole community at Station 1; this ASV was ~20-35% of the community at the other sites (Figure S3A). 23 Environmental effects on small vs. large communities in spatial transect The extent to which community dissimilarity, represented by Bray-Curtis distance, changed with environmental dissimilarity, represented by Euclidian distance (calculated from the variables measured across the spatial transect, including temperature, salinity, [NO2+NO3], [PO4], bacterial production, and cell counts of bacteria and viruses) differed by size fraction. Partial mantel tests show that when controlling for geographic distance, environmental distance was statistically associated with community dissimilarity, measured by Bray-Curtis distance, in both the larger size fraction (R=0.530, p=0.004) and the smaller size fraction (R=0.541, p=0.002). When controlling for environmental distance between the sites, geographic distance was less strongly associated with community dissimilarity in the larger size fraction (R=0.370, p=0.008) and the smaller (R=-0.225, p=0.854). The smaller size fraction changed less rapidly with environmental distance than the larger size fraction (Figure 6A vs. Figure 6B). Amongst five metabolic groups of microbes (Figure 6CG), the community of heterotrophic bacteria had the strongest relationship with environmental distance (Figure 6E, m=0.092, R2=0.685, p=3-10), while cyanobacteria had the weakest relationship (Figure 6D, m=0.004, R2=0.029, p=0.916). Bray-Curtis distance amongst the community of heterotrophic bacteria correlated with many factors, whereas Bray-Curtis distance amongst cyanobacteria did not corelate with any metabolic groups or any environmental parameter measured here (Figure 7). Relative abundances of cyanobacterial ASVs peaked at Stations 3, 5, and 6, in the northern half of the transect (Figure S6, Figure 1). When compared to SPOT as a central location, community Bray-Curtis distance appeared patchiest amongst cyanobacteria and the most homogeneous amongst archaea (Figure S7). Over 24 geographic distance between sampling sites, Bray-Curtis similarity reached similar levels amongst communities of these same five metabolic groupings, ~0.5 on average (Figure S8). Because the taxonomic resolution at which a marine microbial community is studied has been shown to influence the biogeographic patterns observed (Martiny et al., 2006; Laudau and Eloe-Fadrosh, 2019; Thompson et al., 2017), we also aggregated ASVs to the genus level in PhyloSeq and re-examined these correlations. The same trends held when ASVs were aggregated at the genus level, although a few correlations were weaker (Figure S9). Discussion Space & time have similar destabilizing effects on marine microbial communities over small scales Whole community variability within each size fraction reached the same value over the spatial transect (14.5km or less between sites) and temporal regime (5 days or fewer between sample collection). We measured variability using Bray-Curtis similarity, a beta diversity metric that compares pairs of samples to one another on a scale of 0 (0% similarity; samples share no community members) to 1 (100% similarity; samples are completely identical). Bray-Curtis similarity in the smaller size fraction declined from an average of ~0.85 to an average of ~0.7 within 3.6km or one day between sample collection (Figure 2). In other words, 3.6km distance has the same effect on community differences within free-living marine microbes as one day. Bray-Curtis similarity in the larger size fraction declined from an average of ~0.75 to ~0.5 over 5km or within three days between sample collection (Figure 2). 5km distance between sites has a comparable effect on community distance as three days here. Other studies have observed that the stability of the SPOT protistan community (size class 0.7-80µm) declined less over a 14km spatial transect than over 12 days of consecutive 25 sampling, which contrasts with our observation that the two variables have comparable effects at similar scales. Lie et al. used T-RFLP (terminal restriction length polymorphism) to observe that true replicates were ~71% similar to one another, comparable to our findings of ~75% similarity between pseudo-replicates (two filters from the same water sample, Figure 1B) (Figure 2). They also observed that communities decayed to a stable Bray-Curtis similarity of ~63% over space (n=171, 37-80%) and ~58% over time (n=66, 37-78%), slightly higher than the values observed here (Figure 2; Lie et al., 2013). T-RFLP analysis has lower taxonomic resolution than operational taxonomic units (OTUs; e.g., Needham et al., 2017), which have even lower resolution than ASVs (Callahan et al., 2016; Needham et al., 2017; Needham et al., 2018; Fletcher-Hoppe et al., 2023), used here. The phylogenetic resolution at which this study was conducted may explain this slight discrepancy to our results, given that phylogenetic resolution has been shown to influence the biogeographic patterns of the community studied (e.g., Martiny et al., 2006; Laudau and Eloe-Fadrosh, 2019). Notably, this study was conducted in the absence of small-scale spatial or temporal disturbances. We did not observe any fronts or eddies in our spatial transect based on temperature, salinity, or satellite data (Table 1); the only obvious hydrographic feature was a weak current flowing from northwest to southeast during our sampling dates (black arrow in Figure 3). Marine microbial communities have been shown to differ sharply over small distances that span hydrographic features such as fronts and eddies (e.g., Claustre et al., 1994; Martiny et al., 2006; Follows et al., 2007; McGillicuddy et al., 2007; Lévy et al., 2015; Benavides et al., 2021; Horstmann et al., 2022). Short-term temporal disturbances include upwelling, in which marine microbial communities rapidly respond to a sudden influx of nutrients from the depths (e.g., Bakun et al., 1990; Kudela et al., 2008; Capone and Hutcins, 2013; Jacox et al., 2015; 26 Jacox et al., 2018; Ollison et al., 2022). Although coastal upwelling does not occur at SPOT itself (Yeh et al., 2022; Fletcher-Hoppe et al., 2023), spring blooms nonetheless influence its marine microbial community, as regional upwelling events also affect marine microbial communities far off the coast. For example, Needham et al. found that the most abundant taxon changed almost daily during a spring bloom from March 12-18, 2011 (2016). Similarly, eukaryotic communities sampled 10 days apart were 60% similar to one another during non-bloom conditions, but dropped to <20% similar to one another during a bloom (Needham et al., 2016). Our study occurred in non-bloom conditions: it was conducted just before the regional spring bloom event of 2018, which lasted from April 16-30, 2018 (Ollison et al., 2022). The moderately high levels of Bray-Curtis similarity we observed (>0.5 over space and time) partially reflect the fact that we collected these samples at mesoscale spatial and temporal resolution, defined as within tens of kilometers or over several days (Claustre, 1994; McGillycuddy et al., 2007; Lévy et al., 2015). In contrast, transects that span ocean basins can encompass greater variety in temperatures, nutrient availability, and productivity, which lead to more divergent microbial community compositions (e.g., Cohen et al., 2021; Horstmann et al., 2022; James et al., 2022; Sunagawa et al., 2022). When examined over time intervals ranging from one month to >15 years, Bray-Curtis similarity values at the SPOT time-series were lower than what we observed here. Bray-Curtis similarity ranged from 0.5-0.6 in the smaller size fraction (0.22-1µm), 0.4-0.55 in prokaryotes from the larger size fraction (1-80µm), and 0.15-0.2 in eukaryotes in the larger size fraction (Yeh and Fuhrman, 2022). Similar to larger spatial scales, longer timeframes give microbial communities more opportunities to change by differential growth and allow for influx of new community members (Martiny et al., 2006; Yeh and Fuhrman, 2022). 27 Larger marine microbial communities are less homogenous over space & time Overall, the marine microbial communities sampled across our transect were more similar to each other than not. In both size fractions, Bray-Curtis similarity remained >0.5 (Figure 2), demonstrating that at least on the date we sampled, SPOT was comparable to the surrounding waters within 14x11km, and was representative of this area. Eukaryotes and particle-attached prokaryotes were patchy over a relatively homogeneous background of free-living prokaryotes in our sampling regime. Over both space and time, beta diversity patterns show that the larger size fraction of marine microbes (including eukaryotes and particle-attached prokaryotes) were not as similar as their small, free-living counterparts. As time or distance between sample pairs increased, similarity across the smaller size fraction of marine microbial communities remained about 70%, whereas the large size fraction was only 50% similar (Figure 2). In addition, Bray-Curtis similarity decreases more sharply with distance from SPOT in the larger size fraction than the smaller (Figure 3). Within 3.6km surrounding SPOT, the smaller size fraction remains 70-80% identical to SPOT (Figure 3A), particularly in the sites south of SPOT. The larger size fraction destabilizes to only 60% identical over the same distance (Figure 3B). Similarly, Lie et al. found that the protistan communities destabilized to ~60% identical to SPOT within 2km from the site (2013). The smaller size fraction was so homogenous that k-means clustering could not distinguish groups within the samples collected in this study (Figure 4A). The larger size fraction, however, divided into four clusters along the NMDS1 axis (Figure 4B). Notably, the samples collected on a daily basis did not cluster separately from those collected over the spatial transect, which further reflects that time and distance have similar 28 destabilizing effects. Lie et al. also saw that protistan communities around SPOT did not cluster by spatial vs temporal sampling (2013). These beta diversity patterns were likely driven by the addition of new marine microbes, advected into our study sites by local currents. Beta diversity can be broken down into two component parts: turnover, which occurs when species replace each other between sites, and nestedness, which occurs when sites with fewer species are perfect subsets of sites with more species, with no new species present in the less diverse sites (Balsega et al., 2010). Here, we see that turnover drives beta diversity more so than nestedness (Table 2). These values are based on Jaccard dissimilarity, which only uses the presence or absence of species, not Bray-Curtis similarity or dissimilarity, which takes relative abundances of species into account. In other words, most of the beta diversity between the sites in our transect was due to the influx of new species at different sites. This observation may be due to our use of high-resolution ASVs: previous studies have observed that nestedness decreases and turnover increases with finer taxonomic scale (Thompson et al., 2017). We observed that surface currents generally flowed southeast during sampling, and that sites south of SPOT were more similar to SPOT than sites north of SPOT, particularly in the smaller size fraction (Figure 3). In tandem, these observations are consistent with the possibility that local currents brought a patch of water containing new community members through the northern sites, which had not yet reached the southern stations at the time of sampling, increasing the turnover that drove beta diversity across our transect. We also observed higher alpha diversity in samples from the larger size fraction. The number of samples collected over both space and time was sufficient to capture the regional diversity of ASVs in the smaller size fraction of marine microbes, but not the larger. Species accumulation curves for the smaller size fraction were nearly asymptotic, but continued to rise in 29 the larger size fraction, indicating more ASVs could be found in the larger community if more samples were collected (Figure 5A-B). Our inability to observe the full diversity of ASVs in the larger size fraction was due to insufficient sample collection, but not insufficient sequencing depth within each sample: rarefaction curves show that sequencing depth represents the full number of ASVs present in the samples included in these analyses (Figure S1). Shannon index, which is based on both species richness and relative abundance (Shannon 1948), was significantly higher in the larger size fraction of marine microbial communities than the smaller across both space (p=4.11e-05) and time (p=0.008) (Figure 5C). Previous studies at SPOT have also found that Shannon index was significantly higher in the larger size fraction than the smaller (Yeh and Fuhrman, 2022). Notably, alpha diversity is comparable over space vs time within each size fraction, further evidence that 14.5km of space and five days have similar effects in decreasing community similarity (Figure 2, Figure 5). Stacked barplots of the top 20 orders in each sample (Figure S2) and relative abundance plots of the ASVs that drive Bray-Curtis distance (Figures S3-S5) also show that diversity at the ASV and order levels is higher in the larger size fraction. The 20 most abundant orders of marine microbes comprise ~95% of the ASVs in the smaller size fraction but only 80% of the ASVs in the larger, and their relative proportions look more similar over samples from the smaller size fraction (Figure S2). Similarity percentage analysis (SIMPER) reveals that fewer ASVs drive Bray-Curtis distance in the smaller size fraction of spatial samples, and these ASVs reach higher relative abundances (Figure S3 vs. S4). The ASVs that drive of Bray-Curtis distance over time also reach higher relative abundances in the smaller size fraction (Figure S5). In general, the larger size fraction was more heterogeneous. 30 Station 1, the site nearest Catalina Island (Figure 1), is an exception to the general trend of more homogeneity in the smaller size fraction. Approximately 50% of the ASVs at this site were from the order SAR11, compared to ~30% in other sites in the transect (Figure S2). Further, ~40% of the sequences in this site belonged to a single SAR11 ASV (80% of the SAR11 sequences), which drove Bray-Curtis distance across the transect more than any other ASV (Figure S3). This ASV seems better adapted to coastal environments: BLAST results indicate that its representative 16S sequence closely matched two strains from subclade P1a.2, HTCC8045 and HTCC8051 (Brown et al., 2012). These strains were isolated from surface waters within 5 miles of the Oregon coast (Stingl et al., 2007; Giovanoni, 2017), with ITS sequences matching those originally described at SPOT (Brown et al., 2005). It is possible that this coastal SAR11 ASV was able to proliferate due to a nutrient common in coastal waters that was not measured here. Environmental distance drives community variation in the larger size fraction of marine microbes, particularly amongst heterotrophic bacteria Multiple lines of evidence indicate the larger community of marine microbes was more strongly influenced by environmental variation than the smaller. Partial mantel tests show that when controlling for geographic distance, community Bray-Curtis distance in both size fractions was statistically associated with environmental distance. (Here, environmental distance is represented by Euclidean distance of centered and scaled environmental variables from the spatial transect (temperature, salinity, [NO2+NO3], [PO4], bacterial production, and cell counts of bacteria and viruses)). Both the larger size fraction (R=0.530, p=0.004) and the smaller size fraction (R=0.541, p=0.002) showed significant associations with environmental distance. When 31 controlling for environmental distance, geographic distance between the sites was less strongly associated with community distance in the larger size fraction (R=0.370, p=0.008) than the smaller (R=-0.225, p=0.854). Comparing community distance to environmental distance via scatterplot also shows that the larger size fraction of marine microbial communities was more strongly influenced by environmental distance (Figure 6A-B). Note that Bray-Curtis distance spans ~0.1-0.4 in the smaller size fraction compared to ~0.2-0.6 in the larger, further evidence that the larger size fraction is less homogenous. Environmental distance had a particularly strong influence on variation in heterotrophic bacterial communities in the larger size fraction. Dissimilarity in this metabolic group was driven by variation in more metabolic groups and environmental variables than any other metabolic group. Bray-Curtis distance amongst heterotrophic bacteria correlated with environmental variability more strongly (m=0.092, R2=0.685, p=3x10-10) than did Bray-Curtis distance in any of the other groupings (archaea, cyanobacteria, heterotrophic bacteria, photosynthetic eukaryotes, and non-photosynthetic eukaryotes; Figure 6C-G). Similarly, a heterotrophic order, SAR11, drove most of the variation in the smaller size fraction (Figure S2, Figure S3, also see discussion section above). Bray-Curtis distance amongst heterotrophic bacteria also correlated significantly with more component parts of environmental distance, and with dissimilarity amongst more of the other metabolic groups (Figure 7). Variability in heterotrophic bacteria may be driven by nutrient availability, predator-prey interactions, or – more likely – a combination of these bottom-up and top-down factors (e.g., Rodríguez-Gálvez et al., 2023), with the influences of either shifting as the particle is colonized (Datta et al., 2016; Tobias-Hünefeldt et al., 2020). Heterotrophic bacteria have been shown to interact with a variety of marine microbial taxa, many of which belong to one of the metabolic groups investigated here (e.g., Amin et al., 2012; Follett 32 et al., 2022). The high variability in heterotrophic bacterial communities that we observed is consistent with previous reports that particle-attached heterotrophic bacteria are a highly variable group, perhaps due to environmental niches within particles (e.g., Crespo et al., 2013; Mestre et al., 2017; Milici et al., 2017; Reintjes et al., 2023). Environmental similarity between sites that were farther apart drove microbial community similarity more so than geographic distance between sites. K-means clustering of ordinations showed that the large size fraction of communities clustered neatly along latitude: sites north of SPOT (closer to mainland California) had negative NMDS1 values, SPOT had neutral NMDS1 values, and southern sites (furthest from the mainland, closest to Catalina Island) had positive NMDS1 values (Figure 4). Even though sampling sites directly north or south of SPOT were physically closer, they did not cluster together. In addition, Bray-Curtis similarity does not have a linear relationship with geographic distance between sampling sites: the change in community similarity over change over environmental distance is much smoother and more linear (Figure 6). In both size fractions, pairs of samples collected at the same latitude (4.64km and 9.28km apart; Figure 1) were more similar to one another on average than pairs of samples from closer sites (Figure 2). Bray-Curtis similarity jumps at these precise spatial intervals because these are the distances between sampling sites from the same latitude: these pairs had more similar environmental conditions and therefore more similar microbial communities than might be expected based on the distance between them alone. We believe these stations were within the same meso-scale patch of water, which had different environmental conditions and different microbial community members than other sites. Variations in the cyanobacterial community are consistent with the possibility that our meso-scale transect contained different patches of water with different environmental conditions. 33 The sampling sites in the north and northeast section of the transect (Stations 3, and particularly Stations 5 and 6; Figure 1) had high relative abundances of cyanobacterial ASVs (Figure S6). Further, Bray-Curtis similarity to SPOT was lowest in these three sites, about 25-55% similar (Figure S7). (This is the lowest levels of similarity out of the five metabolic groupings: archaea appeared the most similar across the spatial transect, >65% identical to SPOT (Figure S7A), heterotrophic bacteria and both photosynthetic and non-photosynthetic eukaryotes were >45% identical to SPOT (Figure S7B, S7D, and S7E)). Cyanobacteria were the least influenced by environmental distance of any metabolic group (Figure 6), and did not significantly correlate with any environmental parameters measured here (Figure 7). Only 15 ASVs were included in this group, which was dominated by two ASVs closely related to the cultured isolate Synechococcus CC9902 (Figure S6). Variation in the environmental parameters over the small spatial transect sampled here may have been too small to drive notable community differences amongst cyanobacteria. Alternately, cyanobacteria communities may have been shaped by a variable not measured in this study. Although environmental distance affected the five metabolic groups differently, groupings showed similar levels of dissimilarity over geographic distance (Figure S8). As distance between the sampling sites increased, Bray-Curtis distance levels within each of the metabolic groups reached the same level, approximately 0.5 (Figure S8). It is possible that some of our methodology generated artifacts, obscuring true environmental patterns within the communities sampled. We grouped ASVs from the larger size fraction into five subsets by metabolic strategy according to their classification at the phylum level (Figure 6, 7, S6, S8, S9), which may miss differences in metabolic strategies at lower taxonomic levels, particularly amongst eukaryotes. Secondly, observing microbes at the ASV 34 level, the phylogenetic resolution used in this study, might sometimes partition microbial species into multiple ASVs (i.e., “over-splitting”; Schloss 2021), a particular problem for eukaryotic ASVs given the high copy number of the 18S gene and small variations within (Caron et al., 2012; Gong and Marchetti, 2019). We focus analyses on the ASV level data because ASVs typically do represent ecologically different organisms (e.g., Needham et al., 2017) and can be aggregated when desired (Callahan et al., 2016). However, over-splitting ASVs could misrepresent the actual ecological patterns of biological species in marine microbial communities (Martiny et al., 2006; Laudau and Eloe-Fadrosh, 2019; Thompson et al., 2017). To address this, we aggregated ASVs to the genus level and repeated the mantel-like tests. We found the same patterns, although some correlations were slightly weaker (Figure S9). We therefore believe that using ASVs did not lead to artifacts in our data. Conclusions This paper compares and contrasts the influences of space (<14.5km between sites) and time (<5 days between sample collection) on marine microbial communities from two size fractions (0.22-1µm and 1-80µm). We observe that within each size fraction, the space and time scales we sampled have similar effects on community variability: 3.6km distance between sites and one day between sample collection both result in a decrease in Bray-Curtis similarity of ~15% within the smaller size fraction, while similarity in the larger size fraction decreases by ~25% within 5km and 3 days between sample collection. Communities sampled in the transect were more similar to one another than not. However, across both space and time, microbes from the larger size fraction were patchy over a relatively homogeneous background of smaller communities in our sampling regime, with higher Bray-Curtis distance, stronger NMDS 35 clustering, and higher alpha diversity. Lastly, we observed that the larger size fraction, particularly heterotrophic bacteria, was more influenced by environmental variation than was the smaller. In partial agreement and partial contrast to the levels of dissimilarity we hypothesized (highest over time in the larger size fraction, lowest over space in the smaller size fraction, intermediate in other communities), we saw that community variability was higher in the larger size fraction, lower in the smaller size fraction, and comparable across space vs. time. These observations are particularly important to contextualize marine microbial community studies undertaken at much bigger spatial and temporal scales, including monthly samples at the San Pedro Ocean Time-series, which may overlook variations in marine microbes over smaller scales of space and time. Sampling across large-scale spatial and temporal transects misses the patterns of diversity discussed here. 36 References Amin, S. A., M. S. Parker, and E. V. Armbrust. 2012. Interactions between Diatoms and Bacteria. Microbiol Mol Biol Rev 76: 667–684. doi:10.1128/MMBR.00007-12 Bakun, A. 1990. Global Climate Change and Intensification of Coastal Ocean Upwelling. Science 247: 198–201. doi:http://dx.doi.org/10.1126/science.247.4939.198 Baselga, A. 2010. Partitioning the turnover and nestedness components of beta diversity. Global Ecology and Biogeography 19: 134–143. doi:10.1111/j.1466-8238.2009.00490.x Benavides, M., L. Conradt, S. Bonnet, I. Berman-Frank, A. Petrenko, and A. M. Doglioli. 2021. Fine scale sampling unveils diazotroph patchiness in the South Pacific Ocean. ISME Communications 1: 1-3. doi:.1038/s43705-021-00006-2 Brown, M. V., F. M. Lauro, M. Z. DeMaere, and others. 2012. Global biogeography of SAR11 marine bacteria. Molecular Systems Biology 8: 595. doi:10.1038/msb.2012.28 Brown, M. V., G. K. Philip, J. A. Bunge, M. C. Smith, A. Bissett, F. M. Lauro, J. A. Fuhrman, and S. P. Donachie. 2009. Microbial community structure in the North Pacific ocean. ISME J 3: 1374–1386. doi:10.1038/ismej.2009.86 Brown, Mark V., Schwalbach, Michael S, Hewson, Ian, and Fuhrman, Jed A. 2005. Coupling 16S-ITS rDNA clone libraries and automated ribosomal intergenic spacer analysis to show marine microbial diversity: development and application to a time series. Environmental Microbiology 7: 1466–1479. doi: 10.1111/j.1462-2920.2005.00835.x Callahan, B. J., P. J. McMurdie, M. J. Rosen, A. W. Han, A. J. Johnson, and S. P. Holmes. 2016. DADA2: High-resolution sample inference from Illumina amplicon data. Nat Methods 13: 581–583. doi:10.1038/nmeth.3869 Capone, D. G., and D. A. Hutchins. 2013. Microbial biogeochemistry of coastal upwelling regimes in a changing ocean. Nature Geosci 6: 711–717. doi:10.1038/ngeo1916 Caron, D., and P. Countway. 2009. Hypotheses on the role of the protistan rare biosphere in a changing world. Aquat. Microb. Ecol. 57: 227–238. doi:10.3354/ame01352 Caron, D.A., Countway, P.D., Jones, A.C., Kim, D.Y., and Schnetzer, A. 2012. Marine Protistan Diversity. Annual Review of Marine Science 4: 467–493. doi: 10.1146/annurev-marine120709-142802 Chénard, C., W. Wijaya, D. Vaulot, A. Lopes dos Santos, P. Martin, A. Kaur, and F. M. Lauro. 2019. Temporal and spatial dynamics of Bacteria, Archaea and protists in equatorial coastal waters. Sci Rep 9: 16390. doi:10.1038/s41598-019-52648-x Claustre, H., P. Kerhervé, J. C. Marty, L. Prieur, C. Videau, and J.-H. Hecq. 1994. Phytoplankton dynamics associated with a geostrophic front: Ecological and 37 biogeochemical implications. Journal of Marine Research 52: 711–742. doi:10.1357/0022240943077000 Cohen, N. R., M. R. McIlvin, D. M. Moran, and others. 2021. Dinoflagellates alter their carbon and nutrient metabolic strategies across environmental gradients in the central Pacific Ocean. Nat Microbiol 6: 173–186. doi:10.1038/s41564-020-00814-7 Cram, J. A., L. C. Xia, D. M. Needham, R. Sachdeva, F. Sun, and J. A. Fuhrman. 2015. Crossdepth analysis of marine bacterial networks suggests downward propagation of temporal changes. ISME J 9: 2573–2586. doi:10.1038/ismej.2015.76 Crespo, B. G., T. Pommier, B. Fernández‐Gómez, and C. Pedrós‐Alió. 2013. Taxonomic composition of the particle‐attached and free‐living bacterial assemblages in the Northwest Mediterranean Sea analyzed by pyrosequencing of the 16S RRNA. Microbiology Open 2: 541–552. doi:10.1002/mbo3.92 Datta, M. S., E. Sliwerska, J. Gore, M. F. Polz, and O. X. Cordero. 2016. Microbial interactions lead to rapid micro-scale successions on model marine particles. Nat Commun 7: 11965. doi:10.1038/ncomms11965 Estaki, M., L. Jiang, N. A. Bokulich, and others. 2020. QIIME 2 Enables Comprehensive End‐to‐ End Analysis of Diverse Microbiome Data and Comparative Studies with Publicly Available Data. CP in Bioinformatics 70: e100. doi:10.1002/cpbi.100 Falkowski, P. G., T. Fenchel, and E. F. Delong. 2008. The Microbial Engines That Drive Earth’s Biogeochemical Cycles. Microbial Ecology 320. 1034-1039. doi: 10.1126/science.1153213 Field, CB, Behrenfeld, MJ, Randerson, JT, and Falkowski, Paul. 1998. Primary production of the biosphere: integrating terrestrial and oceanic components. Science 281: 237–240. doi:10.1126/science.281.5374.237. PMID: 9657713 Fletcher-Hoppe, C., Y.-C. Yeh, Y. Raut, J. L. Weissman, and J. A. Fuhrman. 2023. Symbiotic UCYN-A strains co-occurred with El Niño, relaxed upwelling, and varied eukaryotes over 10 years off Southern California. ISME Communications. 3: 63. doi:10.1038/s43705-023-00268-y Follett, C. L., S. Dutkiewicz, F. Ribalet, E. Zakem, D. Caron, E. V. Armbrust, and M. J. Follows. 2022. Trophic interactions with heterotrophic bacteria limit the range of Prochlorococcus. Proc. Natl. Acad. Sci. U.S.A. 119: e2110993118. doi:10.1073/pnas.2110993118 Follows, M. J., S. Dutkiewicz, S. Grant, and S. W. Chisholm. 2007. Emergent Biogeography of Microbial Communities in a Model Ocean. Science 315: 1843–1846. doi:10.1126/science.1138544 Fuhrman, J. A., J. A. Cram, and D. M. Needham. 2015. Marine microbial community dynamics and their ecological interpretation. Nat Rev Microbiol 13: 133–146. 38 doi:10.1038/nrmicro3417 Fuhrman, J. A., I. Hewson, M. S. Schwalbach, J. A. Steele, M. V. Brown, and S. Naeem. 2006. Annually reoccurring bacterial communities are predictable from ocean conditions. Proceedings of the National Academy of Sciences of the United States of America 103: 13104–13109. doi:10.1073/pnas.0602399103 Giovannoni, S. J. 2017. SAR11 Bacteria: The Most Abundant Plankton in the Oceans. Annu. Rev. Mar. Sci. 9: 231–255. doi:10.1146/annurev-marine-010814-015934 Gong, W., and A. Marchetti. 2019. Estimation of 18S Gene Copy Number in Marine Eukaryotic Plankton Using a Next-Generation Sequencing Approach. Front. Mar. Sci. 6: 219. doi:10.3389/fmars.2019.00219 Guidi, L., L. Legendre, G. Reygondeau, J. Uitz, L. Stemmann, and S. A. Henson. 2015. A new look at ocean carbon remineralization for estimating deepwater sequestration. Global Biogeochemical Cycles 29: 1044–1059. doi:10.1002/2014GB005063 Hewson, I., J. Steele, D. Capone, and J. Fuhrman. 2006. Temporal and spatial scales of variation in bacterioplankton assemblages of oligotrophic surface waters. Mar. Ecol. Prog. Ser. 311: 67–77. doi:10.3354/meps311067 Hörstmann, C., P. L. Buttigieg, U. John, E. J. Raes, D. Wolf‐Gladrow, A. Bracher, and A. M. Waite. 2022. Microbial diversity through an oceanographic lens: refining the concept of ocean provinces through trophic‐level analysis and productivity‐specific length scales. Environmental Microbiology 24: 404–419. doi:10.1111/1462-2920.15832 Jacox, M. G., C. A. Edwards, E. L. Hazen, and S. J. Bograd. 2018. Coastal Upwelling Revisited: Ekman, Bakun, and Improved Upwelling Indices for the U.S. West Coast. Journal of Geophysical Research: Oceans 123: 7332–7350. doi:10.1029/2018JC014187 Jacox, M. G., J. Fiechter, A. M. Moore, and C. A. Edwards. 2015. ENSO and the California Current coastal upwelling response. J. Geophys. Res. Oceans 120: 1691–1702. doi:10.1002/2014JC010650 James, C. C., A. D. Barton, L. Z. Allen, and others. 2022. Influence of nutrient supply on plankton microbiome biodiversity and distribution in a coastal upwelling region. Nat Commun 13: 2448. doi:10.1038/s41467-022-30139-4 Kudela, Raphael M, Banas, Neil S., Barth, John A., and others. 2008. New Insights into the controls and Mechanisms of Plankton Productivity in coastal Upwelling Waters of the Northern California current system. Oceanography 21: 47–59. doi: 10.5670/oceanog.2008.04 Ladau, J., and E. A. Eloe-Fadrosh. 2019. Spatial, Temporal, and Phylogenetic Scales of Microbial Ecology. Trends in Microbiology 27: 662–669. doi:10.1016/j.tim.2019.03.003 Lévy, M., O. Jahn, S. Dutkiewicz, M. J. Follows, and F. d’Ovidio. 2015. The dynamical 39 landscape of marine phytoplankton diversity. J. R. Soc. Interface. 12: 20150481. doi:10.1098/rsif.2015.0481 Lie, A., D. Kim, A. Schnetzer, and D. Caron. 2013. Small-scale temporal and spatial variations in protistan community composition at the San Pedro Ocean Time-series station off the coast of southern California. Aquatic Microbial Ecology 70. doi:10.3354/ame01652 Long, A. M., S. Hou, J. C. Ignacio-Espinoza, and J. A. Fuhrman. 2021. Benchmarking microbial growth rate predictions from metagenomes. ISME J 15: 183–195. doi:10.1038/s41396- 020-00773-1 Long, R., and F. Azam. 2001. Microscale patchiness of bacterioplankton assemblage richness in seawater. Aquat. Microb. Ecol. 26: 103–113. doi:10.3354/ame026103 Martiny, J. B. H., B. J. M. Bohannan, J. H. Brown, and others. 2006. Microbial biogeography: putting microorganisms on the map. Nat Rev Microbiol 4: 102–112. doi:10.1038/nrmicro1341 McGillicuddy, D. J., L. A. Anderson, N. R. Bates, and others. 2007. Eddy/Wind Interactions Stimulate Extraordinary Mid-Ocean Plankton Blooms. Science 316: 1021–1026. doi:10.1126/science.1136256 McMurdie, P. J., and S. Holmes. 2013. phyloseq: An R Package for Reproducible Interactive Analysis and Graphics of Microbiome Census Data M. Watson. PloS ONE 8: e61217. doi:10.1371/journal.pone.0061217 McNichol, J., P. M. Berube, S. J. Biller, and J. A. Fuhrman. 2021. Evaluating and Improving Small Subunit rRNA PCR Primer Coverage for Bacteria, Archaea, and Eukaryotes Using Metagenomes from Global Ocean Surveys. mSystems 6: 1-13. https://doi.org/10.1128/msystems.00565-21 Mestre, M., E. Borrull, M. M. Sala, and J. M. Gasol. 2017. Patterns of bacterial diversity in the marine planktonic particulate matter continuum. ISME J 11: 999–1010. doi:10.1038/ismej.2016.166 Needham, D. M., E. B. Fichot, E. Wang, L. Berdjeb, J. A. Cram, C. G. Fichot, and J. A. Fuhrman. 2018. Dynamics and interactions of highly resolved marine plankton via automated high-frequency sampling. ISME J 12: 2417–2432. doi:10.1038/s41396-018- 0169-y Needham, D. M., and J. A. Fuhrman. 2016. Pronounced daily succession of phytoplankton, archaea and bacteria following a spring bloom. Nat Microbiol 1: 16005. doi:10.1038/nmicrobiol.2016.5 Needham, D. M., R. Sachdeva, and J. A. Fuhrman. 2017. Ecological dynamics and cooccurrence among marine phytoplankton, bacteria and myoviruses shows microdiversity matters. ISME J 11: 1614–1629. doi:10.1038/ismej.2017.29 40 Ollison, G. A., S. K. Hu, J. V. Hopper, B. P. Stewart, J. Smith, J. L. Beatty, L. K. Rink, and D. A. Caron. 2022. Daily dynamics of contrasting spring algal blooms in Santa Monica Bay (central Southern California Bight). Environmental Microbiology 24: 6033–6051. doi:10.1111/1462-2920.16137 Parada, A. E., D. M. Needham, and J. A. Fuhrman. 2016. Every base matters: assessing small subunit rRNA primers for marine microbiomes with mock communities, time series and global field samples. Environmental Microbiology 18: 1403–1414. doi:10.1111/1462- 2920.13023 Patel, A., R. T. Noble, J. A. Steele, M. S. Schwalbach, I. Hewson, and J. A. Fuhrman. 2007. Virus and prokaryote enumeration from planktonic aquatic environments by epifluorescence microscopy with SYBR Green I. Nat Protoc 2: 269–276. doi:10.1038/nprot.2007.6 Reintjes, G., A. Heins, C. Wang, and R. Amann. 2023. Abundance and composition of particles and their attached microbiomes along an Atlantic Meridional Transect. Front. Mar. Sci. 10: 1051510. doi:10.3389/fmars.2023.1051510 Rodríguez-Gálvez, S., D. Macías, L. Prieto, and J. Ruiz. 2023. Top-down and bottom-up control of phytoplankton in a mid-latitude continental shelf ecosystem. Progress in Oceanography 217: 103083. doi:10.1016/j.pocean.2023.103083 Schloss, P. D. 2021. Amplicon Sequence Variants Artificially Split Bacterial Genomes into Separate Clusters. mSphere 6: e00191-21. doi:10.1128/mSphere.00191-21 Seymour, J., J. Mitchell, and L. Seuront. 2004. Microscale heterogeneity in the activity of coastal bacterioplankton communities. Aquat. Microb. Ecol. 35: 1–16. doi:10.3354/ame035001 Shannon, C.E. 1948. A Mathematical Theory of Communication. The Bell System Technical Journal 27: 379–423. 10.1002/j.1538-7305.1948.tb01338.x Steinberg, D. K., C. A. Carlson, N. R. Bates, R. J. Johnson, A. F. Michaels, and A. H. Knap. 2001. Overview of the US JGOFS Bermuda Atlantic Time-series Study (BATS): a decade-scale look at ocean biology and biogeochemistry. Deep Sea Research Part II: Topical Studies in Oceanography 48: 1405–1447. doi:10.1016/S0967-0645(00)00148-X Stingl, U., H. J. Tripp, and S. J. Giovannoni. 2007. Improvements of high-throughput culturing yielded novel SAR11 strains and other abundant marine bacteria from the Oregon coast and the Bermuda Atlantic Time Series study site. ISME J 1: 361–371. doi:10.1038/ismej.2007.49 Sunagawa, S., S. G. Acinas, P. Bork, and others. 2020. Tara Oceans: towards global ocean ecosystems biology. Nat Rev Microbiol 18: 428–445. doi:10.1038/s41579-020-0364-5 Thompson, L. R., J. G. Sanders, D. McDonald, and others. 2017. A communal catalogue reveals Earth’s multiscale microbial diversity. Nature 551: 457–463. doi:10.1038/nature24621 41 Tobias-Hünefeldt, S. P., J. Wenley, F. Baltar, and S. E. Morales. 2021. Ecological drivers switch from bottom–up to top–down during model microbial community successions. ISME J 15: 1085–1097. doi:10.1038/s41396-020-00833-6 Yeh, Y. C., J. McNichol, D. M. Needham, E. B. Fichot, L. Berdjeb, and J. A. Fuhrman. 2021. Comprehensive single-PCR 16S and 18S rRNA community analysis validated with mock communities, and estimation of sequencing bias against 18S. Environmental Microbiology 23: 3240–3250. doi:10.1111/1462-2920.15553 Yeh, Y.-C., and J. A. Fuhrman. 2022. Contrasting diversity patterns of prokaryotes and protists over time and depth at the San-Pedro Ocean Time series. ISME Communications 2. doi:10.1038/s43705-022-00121-8 Yeh, Y.-C., D. M. Needham, E. T. Sieradzki, and J. A. Fuhrman. 2018. Taxon Disappearance from Microbiome Analysis Reinforces the Value of Mock Communities as a Standard in Every Sequencing Run. mSystems 3. doi:10.1128/mSystems.00023-18 42 Chapter 2 Figures Ch. 2 Figure 1–Locations (A) and water filtration schematic (B) for the study sites. The X-shaped spatial transect was sampled on the fourth day of sampling out of five consecutive days. Station 9 corresponds to the San Pedro Ocean Time-series (SPOT). Santa Catalina Island Port of Los Angeles Stn.1 Stn.2 Stn.3 Stn.4 Stn.5 Stn.6 Stn.7 Stn.8 Stn.9 (SPOT) 33.3°N 33.4°N 33.5°N 33.6°N 33.7°N 118.6°W 118.5°W 118.4°W 118.3°W 118.2°W Longitude Latitude 9.3km 11.1km 43 Ch. 2 Table 1–Environmental parameters sampled across the spatial transect. Bacterial production was measured via incorporation of tritiated Thymidine and Leucine as described in Fuhrman et al. (2006). Counts of viruses and bacterial cells were enumerated via SYBR slides as described in Patel et al. (2007). Location Arrival time (AM) Longitude Latitude Temperature (ºC) Salinity (ppm) DO (mg/L) PAR (mol/m^2/day) [PO4] (µM) Stn1 7:32 -118.450 33.500 15.3 35.5 43.83 135.78 0.26 Stn2 7:47 -118.425 33.525 15.4 35.6 43.93 223.11 0.16 Stn3 8:13 -118.450 33.600 15.1 35.6 43.66 136.62 0.14 Stn4 8:28 -118.425 33.575 15.2 35.6 43.71 184.98 0.19 Stn5 8:47 -118.350 33.600 15.2 35.6 43.71 202.30 0.29 Stn6 9:03 -118.375 33.575 15.2 35.6 43.71 275.60 0.19 Stn7 9:23 -118.350 33.500 15.4 35.6 44.00 432.20 0.25 Stn8 9:35 -118.375 33.525 15.4 35.6 43.96 689.50 0.20 Stn9 (SPOT) 9:50 -118.400 33.550 15.2 35.6 43.79 232.30 0.16 Location [Chl] (mg/mL) Cell counts (cells/mL) Stn1 0.050 1.04E+06 Stn2 0.025 8.94E+05 Stn3 0.085 2.02E+06 Stn4 0.095 2.22E+06 Stn5 0.065 1.89E+06 Stn6 0.140 2.26E+06 Stn7 0.035 4.70E+05 Stn8 0.040 6.88E+05 Stn9 (SPOT) 0.065 2.99E+05 2.94E+07 1.08E+06 Viral counts (viruses/mL) 3.25E+07 3.46E+07 3.70E+07 2.73E+07 2.54E+07 1.79E+07 1.19E+07 1.34E+07 3.92E+05 Bacterial productionThymidine (cells/mL/day) 2.35E+05 1.63E+05 5.65E+05 5.47E+05 3.62E+05 4.92E+05 1.26E+05 5.89E+05 Bacterial productionLeucine (cells/mL/day) 2.97E+05 2.50E+05 4.98E+05 5.05E+05 2.97E+05 5.51E+05 2.22E+05 2.26E+05 44 Ch. 2 Supplementary Figure 1–Rarefaction curves show sequencing depth was sufficient to capture the full diversity of microbial ASVs in most samples from both the smaller size fraction (0.22-1µm, A) and the larger size fraction (1-80µm, B). Inset in B) expands the rarefaction curve of three samples with <200 ASVs (outlined box). 45 Ch. 2 Figure 2–The effects of space and time are comparable within each size fraction of marine microbial communities. Bray-Curtis similarity declines to similar values within each size fraction over distance between sites (A) compared to days between sample collection (B). Distance between sites in Panel A was calculated via the package geosphere in R. Boxplots show the median and interquartile range (IQR) of Bray-Curtis similarity at each distance or day between sampling points; black/ grey circles indicate the mean. Yellow and blue arrows are superimposed along the mean of each box. To better show similarity between replicates (collected 0km apart and on the same day of sampling), replicate samples are treated as individuals here (ASV data are not averaged across duplicates). 46 Ch. 2 Figure 3–The smaller size fraction of microbial communities is more stable than the larger size fraction in the surface waters surrounding SPOT. Bray-Curtis similarity to SPOT is higher in the smaller size fraction of marine microbial communities (0.22-1µm, A) than in the larger (1-80µm, B). Note that SPOT (Station 9, in the center of the X) is identical to itself, indicated by purple (Bray Curits similarity of 1.0). Black superimposed arrows indicate the general direction of current flow on the dates our samples were collected (Southeast, along the Southern California bight, NOAA OSCURS model; https://oceanview.pfeg.noaa.gov/oscurs/#). Figures were visualized via Ocean Data Viewer (ODV). 47 Ch. 2 Figure 4–Samples from the smaller size fraction are more homogenous than samples from the larger size fraction of marine microbial communities. Non-metric multidimensional scaling (NMDS) plots of samples from the smaller size fraction of microbes (0.22-1µm, A) are too homogenous to cluster, whereas samples from larger size fraction (1-80µm, B) clustered by latitude via k-means partitioning. Clusters were assigned via cascadeKM() in “vegan”. Colors correspond to clusters, shapes correspond to latitude. Markers labeled “SPOT” correspond to samples collected daily; note that Station 9 in the spatial transect is also the 4th daily sample collected at SPOT. Ch. 2 Table 2– Beta diversity between spatial and daily samples is driven primarily by ecological turnover. Note that turnover comprises a higher proportion of dissimilarity than nestedness. Overall Jaccard Dissimilarity is higher in the larger size fraction across both space and time. SPOT1 SPOT2 SPOT3 SPOT5 Stn1 Stn2 Stn3 Stn4 Stn5 Stn6 Stn7 Stn8 Stn9_S −0.2 0.0 0.2 −0.2 0.0 0.2 NMDS1 NMDS2 SPOT1 SPOT2 SPOT3 SPOT5 Stn1 Stn2 Stn3 Stn4 Stn5 Stn6 Stn7 Stn8 Stn9_SPOT4 −0.4 −0.2 0.0 0.2 −0.25 0.00 0.25 0.50 NMDS1 NMDS2 A B C D 33.500 33.525 33.550 33.575 33.600 Latitude A Cluster ) B) Stress=0.074 Stress=0.106 Size Type Overall Jaccard Dissimilarity Turnover Nestedness Spatial 0.583 0.496 0.087 Daily 0.529 0.362 0.168 Spatial 0.804 0.699 0.104 Daily 0.672 0.576 0.096 0.22-1µm 1-80µm 48 Ch. 2 Figure 5– Alpha diversity is higher in the larger size fraction of organisms but comparable across both space and time. Species accumulation curves for samples collected over the spatial transect (A) and the daily time-series (B) do not plateau for the larger size fraction, indicating an insufficient number of samples was collected to capture the full diversity of microbial ASVs from 1-80µm. Shading represents standard deviation. Alpha diversity, here measured by the Shannon Index, was higher for the larger size fraction but comparable between spatial and daily samples (C). Double asterisk (**) indicates p<1e-05, single asterisk (*) indicates p<0.01 by Mann-Whitney U test. 49 Ch. 2 Supplementary Figure 2–The 20 most abundant orders comprise ~95% of the smaller size fraction of marine microbial communities across space and time, but at most 80% of the larger size fraction. Notably, >50% of the ASVs in the smaller size fraction from Site 1 belong to the order SAR11. ASVs were grouped at the Order level of phylogeny using PhyloSeq. Note that the same colors correspond to different orders in each plot. 50 Ch. 2 Supplementary Figure 3 –ASVs from SAR11, Bathycoccus, and Synechococcus drive Bray-Curits dissimilarity in the smaller size fraction (0.22-1µm) of marine microbial communities around SPOT. Each ASV contributed to >1% on average of the overall Bray-Curtis dissimilarity across spatial samples, as identified by similarity percentage analysis (SIMPER). (Many more taxa, not shown here, contributed to <1% of overall Bray-Curtis dissimilarity.) The four digits following each hash (#) are the first four digits of the identifying ASV hash generated by QIIME. Note that the scale bars are different in each plot: in Panel A, for example, red represents 40% of the community and blue represents 20%. 51 Ch. 2 Supplementary Figure 4–ASVs from Phaeocystis, Strombidiida, and Synechococcus drive Bray-Curits dissimilarity in the larger size fraction (1-80µm) of marine microbial communities around SPOT. Each ASV contributed to >1% on average of the overall Bray-Curtis dissimilarity across spatial samples, as identified by similarity percentage analysis (SIMPER). (Many more taxa, not shown here, contributed to <1% of overall Bray-Curtis dissimilarity.) The four digits following each hash (#) are the first four digits of the identifying ASV hash generated by QIIME. Note that the scale bars are different in each plot. 52 Ch. 2 Supplementary Figure 5–Relative abundances of ASVs that drive >1% of Bray-Curtis dissimilarity in daily samples of the larger (A) and smaller (B) size fractions of marine microbial communities. Each ASV contributed to >1% on average of the overall Bray-Curtis dissimilarity across daily samples, as identified by similarity percentage analysis (SIMPER). The four digits following each hash (#) are the first four digits of the identifying ASV hash generated by QIIME. 53 Ch. 2 Figure 6––Environmental factors drive dissimilarity in the larger size fraction of marine microbial communities to a stronger degree than in the smaller size fraction, particularly amongst heterotrophic bacteria. Dissimilarity in communities from the smaller size fraction (A) correlates less strongly to environmental distance between the sampling sites than those from the larger size fraction (B). Note the weaker slope (m), lower correlation (R2 ), and lower statistical significance (p) in panel A. The larger size fraction is split into metabolic groups (C-G), revealing that hetrotrophic 54 bacteria are most strongly affected by environmental distance (R2=0.685). Only environmental data from the spatial transect are represented in these plots, including temperature, salinity, [NO2+NO3], [PO4], bacterial production, and cell counts of bacteria and viruses. Ch. 2 Figure 7–Diversity patterns in heterotrophic bacteria and non-photosynthetic eukaryotes are driven by more environmental variables than are their photosynthetic counterparts. Correlation plots show Pearson’s correlation coefficients between Bray-Curtis distances in metabolic groups of ASVs from the larger size fraction and Euclidean distances for each environmental variable measured in the transect (i.e. |value1–value2|), compared pairwise in a Mantel-like test (correlation between two difference matrices). Both the color and size of dots indicate the strength of correlation. All values shown are statistically significant (p<0.05). Note that no negative correlations were statistically significant. Only data from the spatial transect are shown here. 55 Ch. 2 Supplementary Figure 6–– Cyanobacterial relative abundance is highest in the north and northeast sampling sites in the transect. Proportion of cyanobacteria out of all ASVs (16S+ 18S) in the larger size fraction (1-80µm). B) Proportion of ASVs out of all cyanobacteria. The four digits following each underline (_) are the first four digits of the identifying ASV hash generated by QIIME. The majority of cyanobacterial ASVs are most closely related to the cultured isolate Synechococcus CC9902. 56 Ch. 2 Supplementary Figure 7–– Photosynthetic organisms drive community instability in the larger size fraction of marine microbial communities surrounding SPOT. Bray-Curtis similarity was calculated for five metabolic groups of organisms in the larger size fraction (archaea (A), heterotrophic bacteria (B), cyanobacteria (C), mixotrophic eukaryotes (D), photosynthetic eukaryotes I), using SPOT as an anchor point. (Note that SPOT is identical to itself, indicated by purple colors.) Cyanobacteria and photosynthetic eukaryotes are the least stable across the spatial transect, indicated by warmer colors. Organisms were assigned to the five groups based on their taxonomy. BC, Bray-Curtis. 57 Ch. 2 Supplementary Figure 8–Bray-Curtis similarity amongst metabolic groups of marine microbes plateaus with increasing distance between sampling sites. Distance between sites in Panel A was calculated via the package geosphere in R. Boxplots show the median and interquartile range (IQR) of Bray-Curtis similarity at each distance or day between sampling points. To better show similarity between replicates (collected 0km apart and on the same day of sampling), replicate samples are treated as individuals here (ASV data are not averaged across duplicates). 58 Ch. 2 Supplementary Figure 9–Diversity patterns in genera of heterotrophic bacteria and nonphotosynthetic eukaryotes are driven by more environmental variables than are their photosynthetic counterparts. ASVs from the larger size fraction of marine microbes were aggregated via tax_glom() in Phyloseq (McMurdie et al., 2013) and grouped by metabolic activity. Correlation plots show Pearson’s correlation coefficients between Bray-Curtis distances in each metabolic group and Euclidean distances for each environmental variable measured in the transect (i.e. |value1– value2|), compared pairwise in a Mantel-like test (correlation between two difference matrices).. Both the color and size of dots indicate the strength of correlation. All values shown are statistically significant (p<0.05). Note that no negative correlations were statistically significant. Only data from the spatial transect are shown here. 59 Chapter 3–UCYN-A2 dominates high nitrogenase gene expression in autonomously collected samples from the Southern California Bight Abstract Traditional paradigms of marine nitrogen fixation are continuously being re-evaluated. The diazotrophic symbiont UCYN-A is now recognized as a major nitrogen fixer around the globe, and non-cyanobacterial diazotrophs (NCDs) are actively being studied. However, the relative contributions of NCDs to nitrogen fixation are unclear. Most of these studies use nifH PCR, which may miss NCD nifH sequences due to primer bias. To more accurately assess diazotroph diversity and relative activity in the Southern California Bight, we used an untargeted approach to investigate diversity in nifH abundance and expression. The Environmental Sample Processor (ESP) was deployed in surface waters ~1.4 km from Catalina Island and sampled autonomously at 12-hour intervals immediately following a spring bloom in April 2014. We used BLAST to recruit reads from two size fractions of metagenomes and metatranscriptomes against a database of over 8,000 nifH amplicon sequence variants (ASVs). This un-targeted approach allows us to compare expression levels of nifH between different diazotrophs and to compare expression of nifH to other genes, inferring the relative importance of different processes over both day and night. We observed that at the time of sampling, nifH was expressed at higher rates than psbA, which encodes a cyanobacterial protein that is replaced as rapidly as three times per hour. Even with high sampling depth and a database that included >7,900 NCD nifH ASVs, we found that UCYN-A comprised >98% of all the BLAST hits in metatranscriptomes or metagenomes from both size fractions. Although UCYN-A1 expressed >50% of all nifH transcripts, UCYN-A2 nifH was expressed at high levels disproportionate to its low relative abundance. Nitrogenase genes were the most expressed of all UCYN-A mRNA. These findings 60 underscore the importance of UCYN-A in the California Current System and question the relative contributions of other diazotrophs to nitrogen fixation in the region. Introduction Biological nitrogen fixation is a vital source of nutrients to marine microbial communities and may supply up to half of the ocean’s bioavailable nitrogen (Galloway et al., 2004; Zehr and Capone, 2020; Zehr and Capone, 2023). In this process, dinitrogen gas (N2), an abundant but inert form of nitrogen, is converted into ammonium (NH4), a more labile form, by rare prokaryotes known as diazotrophs. This process was once thought to be carried out by only a handful of well-studied, phototrophic organisms that reside in tropical to subtropical waters, including Trichodesmium, Crocosphaera watsonii, and diatom-diazotroph associations (DDAs). However, traditional paradigms of nitrogen fixation have expanded (e.g., Zehr and Capone, 2020; Delmont et al., 2022; Zehr and Capone, 2023). DNA analyses of the marker gene nifH, a subunit of the nitrogenase enzyme, which carries out nitrogen fixation, have enabled the detection of nitrogen fixation in more regions and by more diverse diazotrophs than previously established (e.g. Zehr et al., 1998; Zehr and Capone, 2020; Turk-Kubo et al., 2023; Zehr and Capone, 2023). Despite these advances, the relative abundances and transcriptomic activity of diazotrophs compared to whole microbial communities are poorly understood (e.g. Needham et al., 2018; Fletcher-Hoppe et al., 2023), particularly over day-night cycles and in temperate regions, where most nitrogen fixers are non-phototrophic diazotrophs such as UCYN-A (e.g. Zehr and Capone, 2020; Turk-Kubo et al., 2023). The unicellular diazotroph Candidatus Atleocyanobacterium thalassa, or unicellular cyanobacterium A (UCYN-A), has emerged as a focal point of nitrogen fixation research, due to 61 its streamlined metabolism (Trip et al., 2010) and intimate partnership with a haptophyte host (Thompson et al., 2012). UCYN-A has a global distribution (Martinez-Perez, 2016; Zehr 2016; Turk-Kubo et al., 2017), which spans the tropics (Church et al., 2005, Sohm et al., 2011), temperate regions including the coast of California (Hamersley et al., 2011; Needham et al., 2018; Cabello et al., 2020; Turk-Kubo et al., 2021; Fletcher-Hoppe et al., 2023), and even the Arctic (Harding et al., 2018). In fact, UCYN-A was found in the same samples discussed here, which were previously analyzed by 16S/ 18S tag sequencing (Needham et al., 2018). Based on diversity in its nifH sequences, UCYN-A can be divided into about 40 oligotypes (Turk-Kubo et al., 2017) and at least four clades (Thompson et al., 2014), two of which have been studied extensively. UCYN-A1 is commonly known as the “open ocean” ecotype because it is more abundant in oceanic gyres, while UCYN-A2, the “coastal ecotype,” proliferates nearer shore (e.g. Bombar et al., 2014). UCYN-A fixes nitrogen constitutively during the day (Church et al., 2005; Thompson et al., 2014; Muñoz-Marín et al., 2019; Gradoville et al., 2021; Landa et al., 2021), even in nitrogen-rich environments (Mills et al., 2020), exchanging this fixed N2 for carbon intermediates produced by the host (Thompson et al., 2012). The relative quantities of nitrogen fixation by UCYN-A and well-studied cyanobacterial diazotrophs vs. those of non-cyanobacterial diazotrophs (NCDs) are unclear (e.g. Turk-Kubo et al., 2021; Delmont et al., 2022; Turk-Kubo et al., 2023). NCDs encompass a wide range of organisms that are not photosynthetic, including archaea, alpha-, beta-, and gammaproteobacteria, and sulfur-reducing organisms (Zehr et al., 1998; Capone and Zehr, 2020; TurkKubo et al., 2023; Capone and Zehr, 2023). NCDs are another topic of continued research in marine nitrogen fixation, as they can be highly abundant in marine metagenomes and metatranscriptomes (Delmont et al., 2022), but nitrogen fixation by NCDs has not been measured 62 (Turk-Kubo et al., 2023; Zehr and Capone, 2023). Free-living NCDs have been detected, and some taxa may cycle through free-living and particle-attached life cycles, although they primarily reside on particles (Turk-Kubo et al., 2023). Particles can provide a source of carbon to heterotrophic organisms and likely include micro-environments with low enough oxygen concentration that nitrogenase, which is sensitive to oxygen, is not inhibited (Zehr and Capone, 2020; Turk-Kubo et al., 2023; Zehr and Capone, 2023). NCDs have been shown to colonize particles (Pederson et al., 2018), and have been detected on organic matter as small as 20-200µm in diameter (Farnelid et al., 2019; Harding et al., 2022; Turk-Kubo et al., 2023). It is not known when in the diel cycle these organisms are most active. NCDs have been shown to be highly abundant and active, but may not be detected by nifH PCR primers (Delmont et al., 2022; TurkKubo et al., 2023), which are heavily biased towards certain taxa (Gaby and Buckley, 2011; Turk-Kubo et al., 2023). Here, to more accurately investigate the relative contributions of NCD and cyanobacterial diazotrophs to nifH gene expression, we use an un-targeted approach to study diazotroph abundance and activity off the Southern California Bight. This has the additional benefit of allowing us to compare the relative expression of nifH vs. that of other genes to contextualize the relative importance of different processes. We compare the relative transcriptomic activity of cyanobacteria, UCYN-A, and NCDs within a coastal community of marine microbes by day and night, noting these genes often show diel patterns of expression. Paired samples of metagenomes and metatranscriptomes were collected autonomously and at high temporal resolution off the coast of California via Environmental Sample Processor (ESP). Immediately following a spring bloom, the ESP was deployed 1.4km from the coast of Catalina Island and programmed to collect samples from two size fractions (0.22-1µm and 1-300µm) 12 hours apart. BLAST was 63 used to recruit samples to a custom-made, highly curated database of >8,000 nifH sequences, only ~7,900 of which belonged to cyanobacteria, including UCYN-A. We show that nifH is highly expressed in diel samples, even compared to a much more abundant gene that is known to be highly expressed. UCYN-A dominates nifH abundance and expression, particularly UCYNA2, and nifH is the most transcribed UCYN-A mRNA gene. These findings suggest that at the site and dates of our study, NCDs contributed very little to nitrogen fixation, and support the importance of UCYN-A in the Southern California bight. Methods Sample collection As described in Needham et al. (2018), the Environmental Sample Processor (ESP) was deployed in the Southern California Bight from 13 March to 1 May, 2014. The instrument was anchored approximately 2.4km from the island (Figure 1A), tethered 200m from the ocean floor, such that the sampling depth was between 7-15m from the surface. The ESP collected 1L of seawater at 10am and 10pm from 9 April to 1 May, 2014. Water was autonomously drawn through a 300µm copper mesh, removing most multicellular organisms, then a 1µm AE filter (Pall Gelman), and finally a 0.22µm Durapore filter (Millipore). Thus, the larger size fraction contained microbes from 1-300µm in diameter, and the smaller contained organisms 0.22-1µm in diameter. Salinity and temperature were measured every five minutes by the internal ESP CTD (Figure 1B). 64 Sample preparation Briefly, filters were divided in half aseptically; DNA and RNA were extracted from half of each filter. DNA was extracted from the AE filters via bead beating in NaCl/ CTAB (cetyl trimethyl ammonium bromide) as described by Countway et al. (2005). RNA was extracted from both size fractions using the Qiagen RNEasy kit, with a second DNA removal step included for the larger size fraction. As described in Needham et al. (2018), each protocol was modified to include lysis via lysozyme and proteinase K. RNA from the Durapore size fraction was reverse transcribed to cDNA using SuperScript III (Invitrogen). RNA from the AE size fraction was depleted of rRNA via the RiboZero rRNA Removal Kit (Invitrogen) and transcribed to cDNA using the NEB Next mRNA First Strand Synthesis (E7525) and NEB Next mRNA Ultra II Second Strand Synthesis (E7550). Here we only performed metagenomics on the >1µm fraction, while both sizes were processed for metatranscriptomics. Metagenomes and metatranscriptomes were prepared from selected samples using the NuGen Oviation LowFlow kit v2. Samples were sequenced on an Illumina MiSeq or HiSeq in PE2x150 mode or PE2x250 mode. Samples were demultiplexed, FastQC was run to detect sequence quality, and Atropos (Didion et al., 2017) was run to remove poor quality reads. Human and phiX sequences were removed via BBMap. Any remaining rRNA was removed computationally from metatrsanscriptomics samples using SortMeRNA (Kopylova et al., 2012). After these initial steps, the average number of reads kept per sample was approximately 34M/ sample for AE metagenomes, 1M/ sample for AE metatranscriptomes, and 420K/ sample for Durapore metatranscriptomes. 65 Data analysis BLAST was used to recruit reads from metagenomic and metatranscriptomics samples to a custom-made, highly curated database of over 8,000 nifH amplicon sequence variants (ASVs), 102 of which corresponded to UCYN-A. BLAST was run with 70% identity over 66% of the sequence. BLAST output files were filtered by bitscore, then e-value, then percent identity, such that the best nifH hit to each sequence query was included. If multiple nifH ASVs were equally good matches, the hit was weighted based on the number of “best hits.” (e.g. if four nifH ASVs matched a query sequence with the same bitscore, e-value, and percent identity, each ASV received a weighted score of 0.25). The percent of nifH reads in each sample was calculated as the weighted number of BLAST hits to nifH divided by the number of reads in each sample postquality control processing (described above). To compare expression patterns of nifH to an abundant and highly transcribed diel gene, BLAST recruitment and filtering processes were repeated for psbA, a photosystem II gene with high levels of expression and diel variation (Muratore et al., 2022). To create this BLAST database, 557 sequences of psbA were downloaded from NCBI. All cyanobacterial coding sequences were retrieved using the search command “((Cyanobacteria[Organism]) AND psbA[Gene Name]) NOT genome” in December 2023. BLAST analyses were repeated, and output files were filtered as described above. To determine which UCYN-A genes were most expressed, BLAST was used to recruit reads from metagenomes and metatranscriptomes against the UCYN-A1 genome (NCBI RefSeq GCF_000025125.1; Tripp et al., 2010). Read filtering was repeated as described above. 66 Statistical analyses and data visualizations were conducted in R v4.2.2, including the package “ggplot2.” Results Relative abundance of nifH is orders of magnitude higher in metatranscriptomes vs metagenomes nifH was present in very low relative abundances in the metagenomes, but was significantly more abundant in the metatranscriptomes, particularly in samples collected by day (p=3.42e-4 vs. daytime metagenomes, Figure 2A). (Here, the term relative abundance is used to mean the number of BLAST hits to the nifH gene out of the total number of mRNA sequences in each sample.) On average, nifH hits were approximately 10^-7 of all reads in each metagenomic sample, and approximately 10^-4 of all the reads in each metatranscriptomic sample (p=2*10^-7 for metagenomes vs. metatranscriptomes). The relative abundance of nifH was almost two orders of magnitude higher in metatranscriptomes collected by day than in samples collected by night (p=0.016, Figure 2A). To contextualize the expression of nifH, we compare to the relative abundance of psbA, a well-known highly expressed gene which encodes the reaction center for photosystem II in cyanobacteria and chloroplasts (Mulo et al., 2009; Mulo et al., 2011). psbA is approximately three orders of magnitude more abundant in the metagenomes but is only 1-2 orders of magnitude higher in the metatranscriptomes than in metagenomes (p=2.833*10^-8; Figure 2B). BLAST hits to psbA were approximately 10^-4 of all reads in the metagenomes and approximately 10^-2 of all reads in the metatranscriptomes (note that 10^-2 is 1%). Metatranscriptomic samples collected by day also had significantly higher proportions of psbA than samples collected by night (p=0.032; Figure 2B). 67 UCYN-A comprises almost all nifH sequences Although nifH sequences were a small proportion the reads in each sample (Figure 2A), the symbiotic diazotroph UCYN-A comprised nearly all of the nifH sequences in the larger size fraction of metagenomes and metatranscriptomes, as well as in the smaller size fraction of metatranscriptomes (note we do not report the <1m metagenomes here). Over 99% of BLAST hits to nifH in the 1-300µm transcriptomes matched UCYN-A (Figure 3A). On a per-sample basis, non-UCYN-A nifH genes were expressed in only two samples, at <1% of all nifH sequences (Figure 3B, Figure S1A). Over 98% of the BLAST hits to nifH in the larger size fraction of metagenomes matched UCYN-A (Figure 3C); non-UCYN-A nifH sequences were present in one metagenomic sample at low relative abundance (Figure 3D, Figure S1B). In the smaller size fraction of metatranscriptomes (0.22-1µm), UCYN-A comprised over 98% of the nifH sequences expressed; non-UCYN-A nifH sequences were a small percentage of one daytime sample (Figure S2). Of the non-UCYN-A reads, one nifH sequence in the metagenomes matched a gammaproteobacterial, ETSP-1, and cyanobacteria comprised the remainder (Figure S3). UCYN-A2 nifH sequences are more expressed in the transcriptomes though less abundant in the metagenomes than UCYN-A1 sequences As discussed above, UCYN-A comprised nearly all of the nifH sequences in the larger size fraction of metatranscriptomes. UCYN-A1 was just over 50% of nifH sequences, while UCYN-A2 nifH sequences comprised nearly 45% of all nifH sequences expressed (Figure 4A). On a per-sample basis, UCYN-A2 most commonly comprised >50% of the expressed nifH sequences in samples collected by day, and up to 75% of nifH sequences expressed in samples collected by night (Figure 4B, Figure S1C). In metagenomes, less than 25% of BLAST hits to 68 nifH sequences matched UCYN-A2, whereas UCYN-A1 comprised almost 75% of the reads (Figure 4C). UCYN-A2 comprised 0-25% of each metagenomic samples, while UCYN-A1 comprised ~60-100% of the nifH sequences (Figure 4D, Figure S1D). Further, the ratio of UCYN-A2 nifH sequences in the metatranscriptomes to UCYN-A2 nifH sequences in the metagenomes is almost an order of magnitude higher than the same ratio for UCYN-A1 sequences (Table 1). In the smaller size fraction of metatranscriptomes (0.22-1µm), UCYN-A1 comprised over 80% of the nifH sequences expressed, while less than 5% of the sequences matched UCYN-A2 (Figure S4A). On a per-sample basis, UCYN-A1 comprised over 90% of the nifH sequences expressed in each sample from the smaller size fraction of metatranscriptomes (Figure S4C, S4D). nifH expression in all UCYN-A ASVs, and in UCYN-A1 and UCYN-A2 were correlated (Figure S5). Nitrogenase genes are the most expressed mRNA genes within the UCYN-A genome Most BLAST hits to the UCYN-A1 genome (NCBI RefSeq GCF_000025125.1) matched the region between 700,000-800,000bp, which includes the nitrogenase operon (Tripp et al., 2010). In the larger size fraction of metatranscriptomes (1-300µm), this corresponded to 600 CPM (Counts Per Million, i.e. (# of reads/ total # of sequences)*10^6). In the smaller size fraction of metatranscriptomes (0.22-1µm), this corresponded to approximately 150 CPM. 69 Discussion nifH is highly expressed in metatranscriptomes, particularly during the day The sampling approach used in this study has distinct advantages over other methods: it allowed us to compare nifH expression to that of other genes. Because we have both metagenomes and metatranscriptomes from the larger size fraction, we were able to compare expression in the metatranscriptomes per copy in the metagenomes between multiple genes, which we use as a broad proxy for physiological importance to the organisms. In contrast, culture-based studies of diazotrophs frequently compare nifH expression under different conditions, but are limited to only the organisms in culture (e.g. Qu et al., 2022; Yang et al., 2022). Most environmental studies use qPCR of the nifH gene to compare relative rates of nitrogenase expression between different diazotrophs (e.g. Church et al., 2005), including organisms that are not in culture, such as UCYN-A (e.g. Turk-Kubo et al., 2021). This qPCR approach narrowly focuses on only diazotrophs, missing the wider context of other community members, and is biased against NCD nifH sequences (Delmont et al., 2022; Turk-Kubo et al., 2023). Our in-silico approach is similar, in that we report the number of nifH copies in each liter sampled by the ESP but has substantial advantages. Our approach avoided PCR bias of qPCR studies and the lack of context that arises from focusing on nifH expression alone and allowed us to investigate diversity in nifH expression in the context of the wider community by comparing the relative abundance of nifH in the metatranscriptomes vs. metagenomes to the same ratio in other genes. Surprisingly, we found that nifH, a gene present in only very rare diazotrophs, was expressed at a higher rate (proportional to the metagenomes) than a diel gene of vital importance to cyanobacteria, which were much more abundant than diazotrophs. Based on the very low 70 abundances of nifH (approximately 10^-7 of all the genes present in each sample (Figure 2A)), diazotrophs were present, but quite rare in our sample set. However, nifH was present in metatranscriptomics samples at relative abundances three orders of magnitude higher than in daytime metagenomic samples (p=3.24e-4), and approximately one order of magnitude higher than in nighttime metagenomic samples (Figure 2B). In other words, nifH was highly expressed, particularly during the day. In comparison to metatranscriptomics samples collected by night, nifH was expressed two orders of magnitude higher during the day (p=0.016). Further, normalized to the number of copies in the metagenomes, nifH was expressed at higher rates than psbA, which encodes the photosynthetic reaction center of photosystem II (PSII) that is present in organisms much more abundant than diazotrophs at the surface waters of the nearby San-Pedro Ocean Time-series (Fuhrman et al., 2015; Yeh et al., 2022; Fletcher-Hoppe et al., 2023). This protein is prone to photodamage and is constantly degraded, thus is constantly replaced about every 20 minutes in high light conditions (Mulo et al., 2009; Mulo et al., 2011). (Notably, it is most expressed under high light conditions, presumably at 12noon (Mulo et al., 2011), around the same time that nifH expression peaks for UCYN-A (Church et al., 2005)). In fact, it is so necessary to cyanobacterial function that phages encode a copy, presumably to keep their hosts photosynthesizing after early phage infection shuts down host transcription (Sieradzki et al., 2019). psbA was present in our metagenomic samples at approximately 10^-4 of all reads, three orders of magnitude more abundant than nifH (Figure 1). However, psbA was present in metatranscriptomics samples at only one or two orders of magnitude higher than in metagenomics samples (by day vs. by night, respectively), while nifH expression was up to three orders of magnitude more abundant in metatranscriptomics samples (Figure 2): in comparison to relative abundances in the metagenomes, nifH was expressed at higher rates than this abundant, 71 highly expressed PSII gene. This kind of comparison is only possible with our untargeted BLAST approach, as studies that use PCR to quantify nifH expression only focus on one gene. Our observations compare favorably to previous findings in several ways. Although the high rate at which nifH was expressed is surprising, it is not unusual that nitrogen fixation genes are expressed in the nitrogen-replete waters of the Southern California Bight. Although the ESP does not collect nutrient samples (Needham et al., 2018), the nearby San Pedro Ocean Timeseries shows that the average [NO2+NO3] is a high value of 1.613µM in surface waters in April, the month that the ESP was deployed (Yeh et al., 2022). The Southern California Current is generally nitrogen-limited (Deutsch et al., 2021), and nitrogen fixation has been observed at high rates south of our study site, immediately after spring upwelling supported a phytoplankton bloom (Turk-Kubo et al., 2021). Notably, expression of nitrogen fixation genes does not necessarily mean that diazotrophs are actively fixing nitrogen at the time of transcription; nifH genes may be expressed at different times of day than nitrogen fixation occurs (Turk-Kubo et al., 2023). Previous studies have observed that transcription and enzymatic activity of nitrogenase can be de-coupled. For example, nitrogen fixation by the symbiotic diazotroph UCYN-A decreases with depth, as light attenuates, but expression of nifH does not (Gradoville et al., 2021). This indicates that light has a post-transcriptional effect on nitrogen fixation in UCYN-A; nifH expression and nitrogen fixation are not synonymous in this species. UCYN-A dominated nifH expression and abundance at this site The symbiotic diazotroph UCYN-A comprised almost all of the nifH sequences in each sample analyzed. Although our highly curated nifH database included >7,900 NCD ASVs – 72 sufficient ASV diversity to capture NCD nifH sequences present – we observed that BLAST hits to nifH were >98% UCYN-A of nifH sequences expressed in the samples from the larger size fraction of metatranscriptomes (Figure 3, Figure S1). UCYN-A also comprised >98% of BLAST hits to nifH in all but one sample from the smaller size fraction of metatranscriptomes (Figure S2). Previous studies have found that UCYN-A is active and abundant in Southern California. UCYN-A 16S rRNA was detected in the same samples utilized here and peaked in relative abundance at an unexpectedly high 3% of the prokaryotic community on the same days metagenomes and metatranscriptomes were collected for analyses in this study (Needham et al., 2018). UCYN-A was also found in ~40% of the samples collected from 2008-2018 at the nearby San Pedro Ocean Time-series (~12 km away) (Fletcher-Hoppe et al., 2023). Although nifH expression does not necessarily correlate with nitrogen fixation (e.g. Gradoville et al., 2021; Turk-Kubo et al., 2023), we can infer from previous literature that UCYN-A actively fixed nitrogen during our sampling period, which was transferred directly into the haptophyte host (Thompson et al., 2012). We observed only one BLAST hit to an NCD nifH sequence and it was only present in the larger size fraction of metagenomes, which suggests that NCDs have negligible contributions to nitrogen fixation in the region, at least during our time of sampling. Of the sequences that were not UCYN-A, one BLAST hit in the metagenomes matched alpha-ETSP1, a gammaproteobacterium (Zehr et al., 2003; Turk-Kubo et al., 2023). The non-UCYN-A nifH sequences in the metatranscriptomes all matched other cyanobacteria (Figure S3). We expected a higher proportion of BLAST hits to nifH would match NCD ASVs, particularly in the larger size fraction, as NCDs have been shown to colonize particles as small as 5-200µm in diameter (Farnelid et al., 2019; Harding et al., 2022; Turk-Kubo et al., 2023). Further, previous studies 73 have found a higher proportion of NCD nifH sequences than we did (Delmont et al., 2022). However, these studies used metagenomes and metatranscriptomes collected in the Tara Oceans project (Delmont et al., 2022), which were sequenced at a depth over two orders of magnitude higher than samples investigated here (Alberti et al., 2017; Sunagawa et al., 2020), and still found that UCYN-A was the most abundant and expressive nitrogen fixer (Delmont et al., 2018). Further, no Tara Oceans sample was collected at the same time and location as our samples. UCYN-A rRNA was observed to peak in relative abundance within the very samples we analyzed (Needham et al., 2018), so the dominance of UCYN-A nifH sequences in this study is not surprising. We may observe a larger proportion of NCD nifH sequences in samples collected after this peak in UCYN-A relative abundance. UCYN-A’s high relative abundance in our samples explains the diel expression patterns of nifH we observed (Figure 2A). While the timing of nitrogen fixation in NCDs is unclear, many experiments and environmental studies have shown that UCYN-A expresses nifH during the day (Church et al., 2005; Thompson et al., 2014; Muñoz-Marín et al., 2019; Gradoville et al., 2021; Landa et al., 2021; Muñoz-Marín et al., 2023). Both UCYN-A1 and UCYN-A2 are thought to express nifH most near midday (Church et al., 2005; Thompson et al., 2014), thus the 10am sampling time likely caught UCYN-A nifH transcription just before its peak rate. In fact, changing light signals are required to stop nifH transcription in this organism (Landa et al., 2021), which is likely mediated by its haptophyte host (Landa et al., 2021; Muñoz-Marín et al., 2019). Although UCYN-A is more closely related to unicellular diazotrophs, like C. watsonii, its transcription profile is more similar to Trichodesmium, which also fixes nitrogen during the day (Muñoz-Marín et al., 2019). 74 The sample reported to be collected at 10pm on April 23rd, 2014 contains higher nifH expression than any other night time sample (Figure 3B). This could be because UCYN-A is actually present in the sample in relative abundances that are higher than expected: the same sample has the highest relative abundance of UCYN-A nifH out of all metagenomes (Figure 3D). However, the sample also has high relative abundance of cyanobacterial psbA, a Photosystem II gene with strong diel patterns of expression (Figure 2B) (Montoya et al., 2022). Needham et al. observed high relative abundances of rRNA from several taxa that are typically most active in surface waters during the day in this same sample (2018). Because transcription patterns in this sample are more similar to daytime samples, it is possible that the sample was actually collected in the morning, and was mis-labelled. UCYN-A2 is more active and less abundant than UCYN-A1 Although over 50% of the expressed nifH sequences belonged to UCYN-A1, considering that UCYN-A2 was much less abundant in the metagenomes, UCYN-A2 disproportionately expresses nifH. In other words, UCYN-A2 was more active in the larger size fraction of metatranscriptomes and less abundant in the larger size fraction of metagenomes (Figure 1). UCYN-A1 was at least two thirds of all nifH sequences per metagenome, while UCYN-A2 comprised about 10-25% of the nifH hits in each metagenomic sample (Figure 4, Figure S1). Despite comprising <25% of the nifH sequences in each metagenomic sample, UCYN-A2 was about 35-75% of the reads in each metatranscriptomics sample collected by day (Figure 4, Figure S1). Further, the ratio of UCYN-A nifH genes expressed in the metatranscriptomes to nifH genes present in the metagenomes was about 3,286, almost an order of magnitude higher than the same ratio in UCYN-A1 (~428, Table 1). 75 Several other studies have observed that UCYN-A2 is less abundant than UCYN-A1 (e.g. Thompson et al., 2014; Shiozaki et al., 2020), particularly in this region. The same dataset used here has also been processed into rRNA and rDNA tag sequencing (Needham et al., 2018). UCYN-A1 relative abundance reached almost 3% of the 16S community, whereas UCYN-A2 relative abundance remained <1% of the 16S rDNA sequences in these samples (Needham et al., 2018). Incidentally, the peak in UCYN-A1 relative abundance corresponds to a peak in water temperature (Figure 1B), which is conducive to UCYN-A1 proliferation in the region (Cabello et al., 2020; Turk-Kubo et al., 2021; Fletcher-Hoppe et al., 2023). At the nearby San Pedro Ocean Time-series (SPOT), UCYN-A1 was up to 2.5% of the entire microbial community from 1- 80µm, and UCYN-A2 was at most 0.3% (Fletcher-Hoppe et al., 2023). UCYN-A1 was also observed to be the most abundant diazotroph ~300mi south of our study site, off the coast of San Diego (Turk-Kubo et al., 2021). However, previous studies have also observed that UCYN-A2 is a more active nitrogen fixer than UCYN-A1. While nifH expression is not itself evidence of nitrogen fixation (Gradoville et al., 2021; Turk-Kubo et al., 2023; also discussed above), comparisons of UCYNA1 vs UCYN-A2 nifH expression do reflect an accurate picture of nitrogen fixation: UCYN-A2 symbionts fix more nitrogen per cell than UCYN-A1. Martinez-Perez et al. have observed that average nitrogen fixation rates for UCYN-A2 associations are an order of magnitude greater than for UCYN-A1 symbioses (2016). Similarly, Turk-Kubo et al. found that UCYN-A2 contributes most of the nitrogen fixation nearshore in the California bight, and fixes up to 30x more nitrogen per cell than UCYN-A1 (2021). This may be because the larger UCYN-A2 host requires more nitrogen, a need it fulfills by housing more symbionts per cell, each of which are fixing more 76 nitrogen than a UCYN-A1 symbiont (Martinez-Perez et al., 2016; Cornejo-Castillo et al., 2019; Turk-Kubo et al., 2021). The timing of sample collection is unlikely to be a factor In the discrepancy we observed. Although UCYN-A1 and UCYN-A2 do express different genes at different times, transcription of their nitrogen fixation genes is synchronized (Muñoz-Marín et al., 2023). UCYN-A2 is more transcriptionally active than UCYN-A1 overall (Muñoz-Marín et al., 2023), and several studies have found that UCYN-A2 expresses more nifH than UCYN-A1 (Cornejo-Castillo et al., 2019), and has a much stronger diel cycle of nifH expression, which coincides with the timing of our sampling regime (Thompson et al., 2014). Activity of UCYN-A2 correlated with activity of the much rarer UCYN-A4 (Figure S5), as previous studies have also noted (Turk-Kubo et al., 2021). UCYN-A1 expresses nitrogenase more than any other gene Nitrogenase was the most highly expressed mRNA gene in the UCYN-A1 genome, evidence which supports the theory that UCYN-A is evolving to become a nitrogen-fixing organelle of its host. Most BLAST hits to the UCYN-A1 genome corresponded to the region between 700,000-800,000bp (Figure 5), which encompasses the nitrogenase operon (Tripp et al., 2010). This gene was expressed at ~600 CPM in the larger size fraction and ~125 CPM in the smaller size fraction, over twice as much as the next most expressed region (Figure 5). Notably, this may include genes expressed by other clades of UCYN-A, which are likely highly similar to UCYN-A1 and might map to the UCYN-A1 genome. Previous studies have also observed high nitrogenase expression in comparison to other UCYN-A1 genes, and estimated that nifH comprises up to 25% of the mRNA transcripts in UCYN-A1 (Cornejo-Castillo et al., 2019). Such high expression of nitrogenase is one line of evidence that UCYN-A is evolving to become a 77 “diazoplast,” a nitrogen-fixing organelle within its haptophyte host (Zehr, 2016; Cornejo-Castillo et al., 2019). (Other lines of evidence include the fact that UCYN-A shares over 40% of its protein-encoding genes with two extant endosymbionts (Zehr, 2016)). The larger size fraction of metatranscriptomics (1-300µm) contained approximately four times more nifH BLAST hits than the smaller (0.22-1µm) (Figure 5). Similarly, expression of nifH over time reached higher levels in samples from the larger size fraction than the smaller (Figure 3, Figure 4, Figure S2). This is likely because the larger size fraction contained hits to the UCYN-A1 genome from both UCYN-A1, attached to its symbiont, and UCYN-A2, which has a cellular diameter of >1µm (Zehr et al., 2016). UCYN-A1, which has a diameter of <1µm, has been shown to dissociate from its host under the gentle pressures of sample filtration (Tripp et al., 2010), which explains the detectable levels of nifH expression in the smaller size fraction (Figure 5, Figure S2). Conclusions This study examines the diversity and activity of diazotrophs in the context of a marine microbial community sampled at high temporal resolution immediately following a spring bloom in Southern California. We here recruited metagenomes and metatranscriptomes autonomously sampled night and day against a highly curated BLAST database with >8,000 nifH ASVs, including >7,900 NCD nifH sequences. This untargeted approach avoids the primer bias involved in most studies of nitrogen fixation and allowed us to compare transcription of nifH to other genes. We report that, relative to metagenomic abundances, nifH was transcribed at a higher rate than a highly expressed diel gene required by cyanobacteria, which are orders of magnitude more abundant than diazotrophs. The symbiotic diazotroph UCYN-A contributed 78 >98% of nifH sequences in each metagenomic and metatranscriptomics sample; UCYN-A2 was particularly active compared to the other clades, despite being less abundant. Lastly, we show that nifH was the most highly transcribed UCYN-A mRNA gene, lending further evidence to the theory that it is evolving to become a diazoplast. Demonstrating that UCYN-A contributes to >98% of nifH transcription raises important implications for the relative importance of NCDs, or lack thereof, in the region. Given that UCYN-A is the dominant nitrogen fixer in the region, and that climate change is predicted to have strong effects on UCYN-A distribution in Southern California (Fletcher-Hoppe et al., 2023), climate change may substantially alter inputs of fixed nitrogen in the Southern California Bight. 79 References Alberti, A., J. Poulain, S. Engelen, and others. 2017. Viral to metazoan marine plankton nucleotide sequences from the Tara Oceans expedition. Sci Data 4: 170093. doi:10.1038/sdata.2017.93 Bombar, D., P. Heller, P. Sanchez-Baracaldo, B. J. Carter, and J. P. Zehr. 2014. Comparative genomics reveals surprising divergence of two closely related strains of uncultivated UCYN-A cyanobacteria. ISME J 8: 2530–2542. doi:10.1038/ismej.2014.167 Cabello, A. M., K. A. Turk-Kubo, K. Hayashi, L. Jacobs, R. M. Kudela, and J. P. Zehr. 2020. Unexpected presence of the nitrogen-fixing symbiotic cyanobacterium UCYN-A in Monterey Bay, California. J Phycol 56:1521-1533. doi:10.1111/jpy.13045 Church, M. J., C. M. Short, B. D. Jenkins, D. M. Karl, and J. P. Zehr. 2005. Temporal Patterns of Nitrogenase Gene (nifH) Expression in the Oligotrophic North Pacific Ocean. Appl Environ Microbiol 71: 5362–5370. doi:10.1128/AEM.71.9.5362-5370.2005 Countway, P.D., Gast, Rebecca J., Savai, Pratik, and Caron, D.A. 2005. Protistan Diversity Estimates Based on 18S rDNA from Seawater Incubations in the Western North. The Journal of Eukaryotic Microbiology 52: 95–106. doi:https://doi.org/10.1111/j.1550- 7408.2005.05202006.x Delmont, T. O., J. J. Pierella Karlusich, I. Veseli, J., Fuessel, M.A., Eren, R.A., Foster, C., Bowler, P. Wincker, E. Pelletier. 2022. Heterotrophic bacterial diazotrophs are more abundant than their cyanobacterial counterparts in metagenomes covering most of the sunlit ocean. The ISME Journal 16: 927–936. doi:10.1038/s41396-021-01135-1 Didion, J. P., M. Martin, and F. S. Collins. 2017. Atropos: specific, sensitive, and speedy trimming of sequencing reads. PeerJ 5: e3720. doi:10.7717/peerj.3720 Farnelid, H., K. Turk-Kubo, H. Ploug, J. E. Ossolinski, J. R. Collins, B. A. S. Van Mooy, and J. P. Zehr. 2019. Diverse diazotrophs are present on sinking particles in the North Pacific Subtropical Gyre. The ISME Journal 13: 170–182. doi:10.1038/s41396-018-0259-x Fletcher-Hoppe, C., Yeh, Y.C., Y. Raut, J. L. Weissman, and J. A. Fuhrman. 2023. Symbiotic UCYN-A strains co-occurred with El Niño, relaxed upwelling, and varied eukaryotes over 10 years off Southern California. ISME COMMUN. 3: 63. doi:10.1038/s43705-023- 00268-y Fuhrman, J. A., J. A. Cram, and D. M. Needham. 2015. Marine microbial community dynamics and their ecological interpretation. Nature Reviews Microbiology 13: 133. doi:10.1038/nrmicro3417 Gaby, J. C., and D. H. Buckley. 2011. A global census of nitrogenase diversity. Environmental Microbiology 13: 1790–1799. doi:https://doi.org/10.1111/j.1462-2920.2011.02488.x 80 Galloway, J.N., Dentener, F.J., Capone, D.G., and others. 2004. Nitrogen Cycles: Past, Present, and Future. Biogeochemistry 70: 153–266. doi: 10.1007/s10533-004-0370-0 Gradoville, M. R., A. M. Cabello, S. T. Wilson, K. A. Turk‐Kubo, D. M. Karl, and J. P. Zehr. 2021. Light and depth dependency of nitrogen fixation by the non‐photosynthetic, symbiotic cyanobacterium UCYN‐A. Environmental Microbiology 23: 4518–4531. doi:10.1111/1462-2920.15645 Hamersley, M. R., K. A. Turk, A. Leinweber, N. Gruber, J. P. Zehr, T. Gunderson, and D. G. Capone. 2011. Nitrogen fixation within the water column associated with two hypoxic basins in the Southern California Bight. Aquatic Microbial Ecology 63: 193–205. doi:10.3354/ame01494 Harding, K. J., K. A. Turk-Kubo, E. W. K. Mak, P. K. Weber, X. Mayali, and J. P. Zehr. 2022. Cell-specific measurements show nitrogen fixation by particle-attached putative noncyanobacterial diazotrophs in the North Pacific Subtropical Gyre. Nat Commun 13: 6979. doi:10.1038/s41467-022-34585-y Harding, K., K. A. Turk-Kubo, R. E. Sipler, M. M. Mills, D. A. Bronk, and J. P. Zehr. 2018. Symbiotic unicellular cyanobacteria fix nitrogen in the Arctic Ocean. Proc. Natl. Acad. Sci. U.S.A. 115: 13371–13375. doi:10.1073/pnas.1813658115 Kopylova, Evguenia, Noe, Laurent, and Touzet, Helene. 2012. SortMeRNA: fast and accurate filtering of ribosomal RNAs in metatranscritomic data. Bioinformatics 28: 3211–3217. doi: 10.1093/bioinformatics/bts611. Landa, M., K. A. Turk-Kubo, F. M. Cornejo-Castillo, B. A. Henke, and J. P. Zehr. 2021. Critical Role of Light in the Growth and Activity of the Marine N2-Fixing UCYN-A Symbiosis. Front. Microbiol. 12: 666739. doi:10.3389/fmicb.2021.666739 Martinez-Perez, C., W. Mohr, C. R. Loscher, and others. 2016. The small unicellular diazotrophic symbiont, UCYN-A, is a key player in the marine nitrogen cycle. Nat Microbiol 1: 16163. doi:10.1038/nmicrobiol.2016.163 Mills, M. M., K. A. Turk-Kubo, G. L. van Dijken, B. A. Henke, K. Harding, S. T. Wilson, K. R. Arrigo, and J. P. Zehr. 2020. Unusual marine cyanobacteria/haptophyte symbiosis relies on N2 fixation even in N-rich environments. ISME J 14: 2395–2406. doi:10.1038/s41396- 020-0691-6 Mulo, P. 2011. Strategies for psbA gene expression in cyanobacteria, green algae and higher plants: From transcription to PSII repair. Biochimica et Biophysica Acta. 1817: 247-257. doi:10.1016/j.bbabio.2011.04.011 Mulo, P., C. Sicora, and E.-M. Aro. 2009. Cyanobacterial psbA gene family: optimization of oxygenic photosynthesis. Cell. Mol. Life Sci. 66: 3697–3710. doi:10.1007/s00018-009- 0103-6 Muñoz-Marín, M. del C., I. N. Shilova, T. Shi, H. Farnelid, A. M. Cabello, and J. P. Zehr. 2019. 81 The Transcriptional Cycle Is Suited to Daytime N2 Fixation in the Unicellular Cyanobacterium “Candidatus Atelocyanobacterium thalassa” (UCYN-A). mBio 10: e02495-18. doi:10.1128/mBio.02495-18 Muratore, D., A. K. Boysen, M. J. Harke, and others. 2022. Complex marine microbial communities partition metabolism of scarce resources over the diel cycle. Nat Ecol Evol 6: 218–229. doi:10.1038/s41559-021-01606-w Needham, D. M., E. B. Fichot, E. Wang, L. Berdjeb, J. A. Cram, C. G. Fichot, and J. A. Fuhrman. 2018. Dynamics and interactions of highly resolved marine plankton via automated high-frequency sampling. ISME Journal 12: 2417–2432. doi:10.1038/s41396- 018-0169-y Pedersen, J. N., D. Bombar, R. W. Paerl, and L. Riemann. 2018. Diazotrophs and N2-Fixation Associated With Particles in Coastal Estuarine Waters. Front. Microbiol. 9: 2759. doi:10.3389/fmicb.2018.02759 Qu, P., F. Fu, X. Wang, and others. 2022. Two co‐dominant nitrogen‐fixing cyanobacteria demonstrate distinct acclimation and adaptation responses to cope with ocean warming. Environ Microbiol Rep 14: 203–217. doi:10.1111/1758-2229.13041 Shiozaki, T., M. Ijichi, T. Kodama, S. Takeda, and K. Furuya. 2014. Heterotrophic bacteria as major nitrogen fixers in the euphotic zone of the Indian Ocean. Global Biogeochemical Cycles 28: 1096–1110. doi:10.1002/2014GB004886 Sieradzki, E. T., J. C. Ignacio-Espinoza, D. M. Needham, E. B. Fichot, and J. A. Fuhrman. 2019. Dynamic marine viral infections and major contribution to photosynthetic processes shown by spatiotemporal picoplankton metatranscriptomes. Nat Commun 10: 1169. doi:10.1038/s41467-019-09106-z Sohm, J. A., A. Subramaniam, T. E. Gunderson, E. J. Carpenter, and D. G. Capone. 2011. Nitrogen fixation by Trichodesmium spp. and unicellular diazotrophs in the North Pacific Subtropical Gyre. Journal of Geophysical Research: Biogeosciences 116: n/a-n/a. doi:10.1029/2010JG001513 Thompson, A., B. J. Carter, K. Turk-Kubo, F. Malfatti, F. Azam, and J. P. Zehr. 2014. Genetic diversity of the unicellular nitrogen-fixing cyanobacteria UCYN-A and its prymnesiophyte host. Environ Microbiol 16: 3238–3249. doi:10.1111/1462-2920.12490 Thompson, A. W., R. A. Foster, A. Krupke, B. J. Carter, N. Musat, D. Vaulot, M. M. M. Kuypers, and J. P. Zehr. 2012. Unicellular cyanobacterium symbiotic with a single-celled eukaryotic alga. Science (New York, N.Y.) 337: 1546. doi:10.1126/science.1222700 Tripp, H. J., S. R. Bench, K. A. Turk, R. A. Foster, B. A. Desany, F. Niazi, J. P. Affourtit, and J. P. Zehr. 2010. Metabolic streamlining in an open-ocean nitrogen-fixing cyanobacterium. Nature 464: 90–94. doi:10.1038/nature08786 Turk-Kubo, K. A., H. M. Farnelid, I. N. Shilova, B. Henke, and J. P. Zehr. 2017. Distinct 82 ecological niches of marine symbiotic N2-fixing cyanobacterium Candidatus Atelocyanobacterium thalassa sublineages. J Phycol 53: 451–461. doi:10.1111/jpy.12505 Turk-Kubo, K. A., M. R. Gradoville, S. Cheung, F. M. Cornejo-Castillo, K. J. Harding, M. Morando, M. Mills, and J. P. Zehr. 2023. Non-cyanobacterial diazotrophs: global diversity, distribution, ecophysiology, and activity in marine waters. FEMS Microbiology Reviews 47: fuac046. doi:10.1093/femsre/fuac046 Turk-Kubo, K. A., M. M. Mills, K. R. Arrigo, G. van Dijken, B. A. Henke, B. Stewart, S. T. Wilson, and J. P. Zehr. 2021. UCYN-A/haptophyte symbioses dominate N2 fixation in the Southern California Current System. ISME Communications 1. doi:10.1038/s43705- 021-00039-7 Yang, N., Y.-A. Lin, C. A. Merkel, M. A. DeMers, P.-P. Qu, E. A. Webb, F.-X. Fu, and D. A. Hutchins. 2022. Molecular mechanisms underlying iron and phosphorus co-limitation responses in the nitrogen-fixing cyanobacterium Crocosphaera. The ISME Journal 16: 2702–2711. doi:10.1038/s41396-022-01307-7 Yeh, Y.-C., and J. A. Fuhrman. 2022. Contrasting diversity patterns of prokaryotes and protists over time and depth at the San-Pedro Ocean Time series. ISME Communications 2. doi:10.1038/s43705-022-00121-8 Zehr, J. P., B. D. Jenkins, S. M. Short, and G. F. Steward. 2003. Nitrogenase gene diversity and microbial community structure: a cross‐system comparison. Environmental Microbiology 5: 539–554. doi:10.1046/j.1462-2920.2003.00451.x Zehr, J. P. 2011. Nitrogen fixation by marine cyanobacteria. Trends in Microbiology 19: 162– 173. doi:10.1016/j.tim.2010.12.004 Zehr, J. P., S. R. Bench, B. J. Carter, I. Hewson, F. Niazi, T. Shi, H. J. Tripp, and J. P. Affourtit. 2008. Globally distributed uncultivated oceanic N2-fixing cyanobacteria lack oxygenic photosystem II. Science 322: 1110–1112. doi:10.1126/science.1165340 Zehr, J. P., and D. G. Capone. 2020. Changing perspectives in marine nitrogen fixation. Science 368. doi:10.1126/science.aay9514 Zehr, J. P., and D. G. Capone. 2023. Unsolved mysteries in marine nitrogen fixation. Trends in Microbiology S0966842X23002354. doi:10.1016/j.tim.2023.08.004 Zehr, J. P., M. T. Mellon, and S. Zani. 1998. New Nitrogen-Fixing Microorganisms Detected in Oligotrophic Oceans by Amplification of Nitrogenase (nifH) Genes. Applied and Environmental Microbiology 64: 3444-50. doi: 10.1128/AEM.64.9.3444-3450.1998 Zehr, J. P., I. N. Shilova, H. M. Farnelid, M. del C. Muñoz-Marín, and K. A. Turk-Kubo. 2016. Unusual marine unicellular symbiosis with the nitrogen-fixing cyanobacterium UCYN-A. Nat Microbiol 2: 16214-25. doi:10.1038/nmicrobiol.2016.214 83 Chapter 3 Figures Ch. 3 Figure 1–Study location and environmental conditions at the time of ESP sampling. The Environmental Sample Processor (ESP) was deployed 2.4 km from Catalina Island and 12.5 km from the San Pedro Ocean Time-series (SPOT) in April 2014 (A). Temperature was measured via a CTD attached to the ESP (B). 84 Ch. 3 Figure 2–Although it is orders of magnitude less abundant, nifH is expressed at a higher rate than psbA, a highly expressed diel gene present in cyanobacteria, which comprise a much larger proportion of the population than diazotrophs. BLAST hits to nifH are 1-3 orders of magnitude higher in diel metatranscritomic samples than in metagenomic samples (A) (p=2.9e-7). In comparison, psbA, a photosystem II gene, is three orders of magnitude more abundant than nifH in the metagenomes and 1-2 orders of magnitude more abundant than that in the metatranscriptomes (B) (p=2.833e-8). Boxplots show the median and interquartile range (IQR) of the log10 of BLAST hits to nifH (A) and psbA (B) in each sample. p-values were determined via Mann-Whitney U test on non-parametric data; only statistically significant values are reported. MG, metagenome (n=19 samples); MT, metatranscriptomic (n=9 samples). All ESP samples available are included in this figure. 85 Ch. 3 Figure 3–UCYN-A dominates expression and abundance of nifH genes in the larger size fraction of ESP samples (1-300μm). UCYN-A comprises over 99% of the BLAST hits to nifH in the metatranscriptomes (A); nonUCYN-A nifH genes are only expressed in a few samples (B). Similarly, UCYN-A comprises over 98% of BLAST hits to nifH in the metagenomes (C); non-UCYN-A genes are noticeably present in only one sample (D). Note the difference in scale on the y-axis of BLAST hits to transcriptomes (B) and metagenomes (D). ND, no data. 86 Ch. 3 Supplementary Figure 1–UCYN-A comprises almost all nifH hits in each metatranscriptomic and metagenomic sample. UCYN-A2 comprises a larger proportion of the nifH hits in each metatranscriptomic sample than in each metagenomic sample. Note that the navy blue and red color scheme applies to the panels A and B; multi-color scheme applies to panels C and D. MG, metagenome; MT, metatranscriptome. 87 Ch. 3 Supplementary Figure 2–UCYN-A dominates nifH expression in the smaller size fraction of ESP samples (0.22-1μm). UCYN-A comprises over 98% of BLAST hits to nifH in the metatranscriptomes (A). Expression of non-UCYN-A nifH genes was noticeable in only one sample in the time series (B); nifH transcripts in most samples were 100% UCYN-A (C). 88 Ch. 3 Supplementary Figure 3–Non-UCYN-A nifH BLAST hits in the larger size fraction of metagenomes and metatranscriptomes.. One hit in the metagenomes is from alpha1|MH144511|α-ETSP1, a gammaproteobacterium. The algorithm used to filter BLAST output files split reads into equal parts if multiple hits were had an equal percent identity, e-value, and bit score; thus not all taxonomies have a whole number of hits. MG, metagenomes; MT, metatranscriptomes. 0 1 2 MG MT Sample Type Total # of BLAST Hits Taxonomy Crocosphaera watsonii Cyanothece sp. Gammaproteobacteria Nostoc punctiforme 89 Ch. 3 Figure 4–UCYN-A2 comprises a higher proportion of nifH expression than of nifH abundance. Although UCYN-A1 was >50% of the nifH hits in all of the metatranscriptomes, UCYN-A2 nifH comprised about 45% of nifH hits (A), and at times was expressed more than the nifH gene of UCYN-A1 (B). In contrast, UCYN-A2 comprised less than 25% of nifH hits across all the metagenomic samples (C) and UCYN-A1 nifH genes were at least 2x as abundant in each metagenomic sample (D). Note the difference in scale on the y-axis of BLAST hits to transcriptomes (B) and metagenomes (D). ND, no data. 90 Ch. 3 Table 1–Ratios of nifH BLAST hits in the metatranscriptomes to nifH hits in the metagenomes are almost an order of magnitude higher in UCYN-A2 than in UCYN-A1. Also note that UCYN-A4 has a higher ratio than other clades. UCYN-A3 was expressed in two metatranscriptomics samples, but not detectable in the metagenomic samples. Gray shaded boxes have one fewer datapoint than the number reported, i.e. three datapoints for daytime samples and one datapoint for night time samples. ND, No Data; SD, Standard Deviation. Ch. 3 Supplementary Figure 4–UCYN-A1 dominates nifH expression in the smaller size fraction of ESP samples (0.22-1μm). UCYN-A1 comprises more than 80% of BLAST hits to nifH in the metatranscriptomes; UCYNA2 comprises less than 5% (A). Expression of UCYN-A2 and of non-UCYN-A nifH genes was noticeable in only one sample in the time series (B); most nifH transcripts in most samples were UCYN-A1 (C). All UCYN-A UCYN-A1 UCYN-A2 UCYN-A4 UCYN-A5 UCYN-A6 Day (n=4) 718.10 427.874 3285.88 2487.69 884.34 844.81 SD 416.62 157.739 4498.23 777.53 581.12 483.69 Night (n=2) 73.73 39.184 849.66 744.45 139.69 88.28 SD 93.85 54.112 ND ND ND ND 91 Ch. 3 Figure 5–Nitrogenase genes were the most expressed out of all UCYN-A mRNA. All metatrnscriptomic samples from the larger size fraction (A) and smaller size fraction (B) were recruited to the UCYN-A1 genome (Tripp et al., 2010) using BLAST. In each size fraction. BLAST hits were normalized to report counts per million (CPM). Box indicates the location of the nitrogenase genes (approximately 700-800 kbp; Trip et al., 2010). Note the difference in scale of the y-axes of each plot. 92 Chapter 4–Symbiotic UCYN-A strains co-occurred with El Niño, relaxed upwelling, and varied eukaryotes over 10 years off Southern California Fletcher-Hoppe, Colette[1]; Yeh, Yi-Chun[1, 2]; Raut, Yubin[1]; Weissman, J.L.[1, 3]; and Fuhrman, Jed A.[1]* Author affiliations: 1. Marine & Environmental Biology, Department of Biological Sciences, University of Southern California (USC), Los Angeles, CA 2. Department of Global Ecology, Carnegie Institution for Science, Stanford University, Stanford, CA 3. Schmid College of Science and Technology, Chapman University, Orange, CA * Corresponding author, email address: fuhrman@usc.edu This manuscript was published in 2023 as: Fletcher-Hoppe, C., Y.-C. Yeh, Y. Raut, J. L. Weissman, and J. A. Fuhrman. 2023. Symbiotic UCYN-A strains co-occurred with El Niño, relaxed upwelling, and varied eukaryotes over 10 years off Southern California. ISME COMMUN. 3: 63. doi:10.1038/s43705-023- 00268-y Abstract Biological nitrogen fixation, the conversion of N2 gas into a bioavailable form, is vital to sustaining marine primary production. Studies have shifted beyond traditionally studied tropical diazotrophs. Candidatus Atelocyanobacterium thalassa (or UCYN-A) has emerged as a focal point due to its streamlined metabolism, intimate partnership with a haptophyte, and broad distribution. Here, we explore the environmental parameters that govern UCYN-A’s presence at the San Pedro Ocean Time-series (SPOT), its host specificity, and statistically significant interactions with non-host eukaryotes from 2008-2018. 16S and 18S rRNA gene sequences were amplified by “universal primers” from monthly samples and resolved into Amplicon Sequence 93 Variants, allowing us to observe multiple UCYN-A symbioses. UCYN-A1 relative abundances increased following the 2015-2016 El Niño event. This “open ocean ecotype” was present when coastal upwelling declined, and Ekman transport brought tropical waters into the region. Network analyses reveal all strains of UCYN-A co-occur with dinoflagellates including Lepidodinium, a potential predator, and parasitic Syndiniales. UCYN-A2 appeared to pair with multiple hosts and was not tightly coupled to its predominant host, while UCYN-A1 maintained a strong host-symbiont relationship. These biological relationships are particularly important to study in the context of climate change, which will alter UCYN-A distribution at regional and global scales. Introduction Biological nitrogen fixation sustains primary production in much of the ocean. In this process, rare prokaryotes known as diazotrophs convert inert dinitrogen gas (N2) to ammonia (NH3). For many years, only a handful of well-characterized photosynthetic bacteria were thought to be capable of marine nitrogen fixation: Trichodesmium, Crocospharea watsonii, and symbionts in diatom-diazotroph associations (DDAs) were considered the dominant diazotrophs (e.g. [1, 2]). However, traditional paradigms of biological nitrogen fixation are continuously being challenged (e.g. [2]). Studies that used amplicon sequencing to target the gene nifH, which encodes a subunit of the nitrogenase enzyme that conducts nitrogen fixation, have revealed that diazotrophs are a more diverse group than previously recognized (e.g. [3–6]). The first study that applied this technique to marine organisms observed a cluster of nifH sequences belonging to a clade termed “UCYN-A”, for “unicellular cyanobacterial group A” [7]. This clade of organisms 94 has been tentatively named Candidatus Atelocyanobacterium thalassa [8], and is now recognized as a major contributor to biological nitrogen fixation (e.g. [9, 10]). UCYN-A is an aberrant cyanobacterium, lacking Photosystem II and key components of cellular pathways, such as the Krebs cycle [11]. Its metabolism is streamlined because it lives in symbiosis with a photosynthetic haptophyte host, exchanging fixed nitrogen for carbon compounds [8, 12]. Four clades of UCYN-A are currently recognized based on their nifH sequences, although more may exist [13]. UCYN-A1, the most extensively studied type of UCYN-A, associates with a coccolith-forming member of the genus Braarudosphaera [8], and is found primarily in open-ocean regions [14, 15]. UCYN-A1 is <1 m in diameter while host cells have a diameter of 1-3 m and can house 1-2 symbionts each [13, 16]. UCYN-A2, a coastal ecotype [14, 15], associates with Braarudosphaera bigelowii, also coccolith-forming [12, 17], and is larger than 1m, while host cells are 4-10 m in diameter (e.g. [18]). The UCYN-A2 host likely has more symbiont per cell: although microscopy suggests UCYN-A2 has one symbiont per host cell [19], DNA sequencing shows the UCYN-A2 host can house 4-10 symbionts per cell [14, 16]. UCYN-A has a broad, global distribution [9, 13], including nitrogen-rich systems such as coastal and equatorial upwelling systems (e.g. [2, 10, 20]). UCYN-A1 and UCYN-A2 have been found previously in the Southern California Current System and Monterey Bay (primarily UCYN-A2) [22, 23]. Notably, UCYN-A has been observed at our study site, the San Pedro Ocean Time-series (SPOT), as deep as 890m [24], and at a nearby daily time series off the coast of Catalina Island [25]. Furthermore, UCYN-A was found to comprise up to 95% of the diazotroph population sampled from San Diego to Sebastian Vizcaino Bay (Baja, CA) and within the period of our 10-year timeseries (2008-2018) [22]. 95 Nitrogen fixation by many species has been observed in coastal ecosystems in surprisingly high rates, often despite high concentrations of available nitrogen. Nitrogen fixation off the New Jersey Shore may support up to 100% of primary production in this ecosystem, with some of the highest reported UCYN-A abundances [26]. Coastal nitrogen fixation in Southern California has been measured at lower rates (e.g. [22, 23]). However, few studies have attempted to link diazotrophs to potential predators, and these reports have focused on diazotrophs confined to the tropics [27, 28]. In addition, many questions remain about the UCYN-A symbiosis, including the specificity of host-symbiont partnerships (e.g., UCYN-A has been reported without a host [18]). The question of UCYN-A host specificity is especially important because ocean acidification may degrade the calcareous shells of the established hosts during the coccolith-bearing phases of their life cycles [23, 29]. This in turn could alter the global distribution of the symbiont and alter its global contributions to biological nitrogen fixation. DNA sequencing in the context of a long time series project, as we report here, is a particularly valuable method for studying microbial interactions and changes therein. In our time series, genes encoding the small subunit of ribosomal RNA (i.e. 16S rRNA genes (“16S”) for prokaryotes and 18S rRNA genes (“18S”) for eukaryotes) of the entire microbial community were processed into Amplicon Sequence Variants (ASVs) using “universal” primers that capture all three domains of life [30–32]. Because we did not design the study with a particular set of organisms in mind, we can examine these data for co-occurrences, which may suggest biological interactions, between any microbes. Each set of DNA sequences from a community provides a single snapshot into its structure at the time of sampling. Time series projects, in which communities are sampled on a regular basis over years, allow researchers to assemble a “movie” 96 of what entire microbial communities are actually doing over time [33], including looking at events that occurred only rarely. In this study, we sought to characterize the abiotic niche of UCYN-A, its potential predators, and its host specificity, using 16S & 18S ASVs over a decade+- long time series in coastal, temperate waters. Materials and Methods Data collection Seawater was collected monthly from 5m (surface) and deep chlorophyll maxima (DCM) from 2000-2018 at the San Pedro Ocean Time-series (33.55ºN, 118.4ºW; Figure 1) (although statistical analyses represent data from 2008-2018). Samples were collected and processed into 16S &18S ASVs, which differ by as little as one base pair, using a wrapper of the software Quantitative Insights Into Microbial Ecology v2 (QIIME2) ([30, 34], also see supplemental methods for more details). Because we only wished to examine interactions between whole, single-celled organisms, ASVs corresponding to chloroplasts and multicellular metazoans were removed from the analyses reported here. To investigate host specificity, relative abundances of UCYN-A ASVs and their known hosts were normalized to a common denominator and compared (see supplemental methods). Upwelling intensity at 33ºN, measured by the Biologically Efficient Upwelling Transport Index (BEUTI) and Coastal Upwelling Transport Index (CUTI) [35], were downloaded from the National Ocean and Atmospheric Administration (NOAA’s) Pacific Fisheries Environmental Laboratory (https://oceanview.pfeg.noaa.gov/products/upwelling/cutibeuti). Components of the Bakun Index for upwelling at the nearest available point (33.5ºN, -118.5ºW) were also downloaded from NOAA (https://coastwatch.pfeg.noaa.gov/erddap/griddap/erdlasFnWPr.html). 97 Multivariate ENSO Index (MEI) data, indicative of El Niño (positive MEI)/ La Niña (negative MEI), were obtained from the NOAA Physical Sciences Laboratory (https://psl.noaa.gov/enso/). As recommended, each bimonthly sliding window was used to represent the latter of the two months. Bacterial production was measured by incorporation of tritiated leucine (e.g. [36]). Inorganic nitrogen ([NO2 -+NO3 - ]) and phosphate ([PO4 3- ]) concentrations at the time of sampling were measured via a Lachat spectrophotometer QuickChem 8500 Series 2 at the Marine Science Institute at the University of California, Santa Barbara (concentration range 0.2-300µM for nitrogen, 0.1-200µM for phosphate). Phylogenetic tree The QIIME2 classifier identified six 16S ASVs as UCYN-A with >99.9% confidence and identified seven 18S ASVs were classified as Braarudospharea with >93% confidence. The sequences for these ASVs, along with published 16S sequences of UCYN-A and 18S sequences of Braarudospharea, were assembled into phylogenetic trees for host and symbiont. Trees included all UCYN-A and Braarudospharea sequences publicly available as of January 2022. Sequences were aligned via maaft with default settings [37]. The alignment was trimmed via trimAl, also with default settings [38]. Trees were constructed via Randomized Accelerated Maximum Likelihood (RaxML) with rapid bootstrap analysis and 100 bootstraps [39], and visualized via the interactive Tree of Life (iTOL; [40]). Relationship with abiotic factors Logistic regression was used to evaluate effects of each environmental parameter on UCYN-A symbiont and host presence at SPOT. This process was repeated for a Lepidodinium ASV. The mean and standard error of each parameter were plotted on days that each organism 98 was present/ absent (defined as >0.01% of the 16S community) via ggplot 2 [41]. Statistical significance of these differences was corrected for multiple testing via Benjamini-Hochberg correction. For all other analyses, relative abundance data were prepared as follows. Five samples from the 5m depth and two samples from the DCM were excluded because they contained too few sequences to capture the diversity of 18S ASVs at SPOT (Figure S1). Abundance data from missing dates were linearly interpolated via na.approx() from the R package zoo [42]. To avoid common problems with compositional data [43, 44], relative abundances were centered log-ratio (CLR) transformed with the mclr() function from the SPRING package [45]. Spearman’s rank-order correlation was used to evaluate the monotonic relationship between environmental variables and CLR-transformed host/ symbiont relative abundances. Spearman’s correlation was performed using the rcor() function with type=”spearman” from the Hmisc package in R. Additional analyses are described in Supplementary Methods. Network analyses Co-occurrence networks were generated on interpolated, CLR-transformed data with extended local similarity analysis (eLSA; [46, 47]). Networks constructed using non-CLR transformed data missed several correlations detectable from transformed data (Figure S2). This study reports on associations between UCYN-A and 18S taxa at 5m from March 2008–July 2018. CLR-transformed UCYN-A1 relative abundances from the smaller size fraction were included on all dates. On several dates that UCYN-A1 and UCYN-A2 relative abundances peaked in the larger size fraction prokaryotic community, the 100 most abundant eukaryotic taxa were selected for inclusion in network analyses. eLSAs were run with 1000 permutations and 99 default normalization off. Q values were calculated from P-values using the qvalues() package (e.g. [48]). Correlations at the 5m depth that were highly statistically significant via both Pearson’s correlation and Spearman’s correlation (all P values < 0.005 and all Q values <0.01) were visualized using Cytoscape v3.5 [49]. Relative abundances of the 18S taxa included in these networks were visualized via Krona plots. Results Phylogenetic tree Six 16S ASVs were classified as UCYN-A with >99.99% confidence according to QIIME2. (QIIME2 classifies ASVs based on SILVA 132 and assigns each a unique identifier or hash (#); ASV numbers reflect the alphabetical order of the hashes QIIME2 generated). One ASV (#3d852410f44d21c92c9c55fbbb25187e) matched the published genome of the UCYN-A1 sublineage perfectly (100% BLAST identity, also see Figure 2A, Table S1; [50]), and will be referred to as UCYN-A1. Another 16S ASV (#af1bb1f9fb1c3f3d18571e711df407bb) matched the published genome of the UCYN-A2 sublineage (100% BLAST identity, also see Figure 2A, Table S1; [14]). In addition, this ASV never appeared in the smaller size fraction of filters (0.22-1m), which is consistent with the reported larger diameter of UCYN-A2 (>1m) [51]. This ASV will be referred to as UCYN-A2. QIIME2 classified seven of the 18S ASVs found at SPOT as Braarudosphaera, the genus that contains previously established hosts of UCYN-A symbionts, with >93% confidence. Four of these seven ASVs were classified as B. bigelowii, the putative host of UCYN-A2. B. bigelowii ASV1, ASV3, and ASV4 (#04926e2fd1b8706b4866c02650f702dd, #70a5283da28db501a349c5beb22881e7, #8c144683114fbb1ad2e9425f7dcd1b02, respectively) 100 closely matched strains of B. bigelowii shown to associate with UCYN-A2 (Figure 2B, Table S1; [12]). The other three Braarudospharea ASVs did not match a formally named species, and are referred to as 5x, 6x, and 7x. Of these, Braarudosopharea ASV7x matched isolate BIOSOPE T60 (100% BLAST identity, also see Figure 2B, Table S1), the established host of the published UCYN-A1 genome [8]. Relationship with abiotic factors Regional coastal upwelling, measured by BEUTI, was significantly weaker on dates that UCYN-A1 was present (Figure 3, Table S2). In addition, chlorophyll-a concentrations and bacterial production were lower when UCYN-A1 was present, while sea surface temperature (SST) and MEI were higher (Figure 3). Ekman transport moved surface water north and east during months UCYN-A1 was present at SPOT (Table S2). UCYN-A1 relative abundances correlated positively with MEI and SST, and negatively with bacterial production, upwelling, and chlorophyll-a (Figure 4). Higher than average relative abundances of UCYN-A1 coincided with low upwelling indices, positive MEI, and SST >19ºC (Figure S3). On dates UCYN-A2 was present, the BEUTI index was slightly lower (Figure 3). Relative abundances of UCYN-A2 showed a weak positive correlation with SST and a weak negative correlation with BEUTI (Figure 4), but no significant correlation to chlorophyll concentrations, bacterial production, SST, and MEI (Table S2). Neither nitrogen nor phosphorus concentrations differed significantly on dates that UCYN-A ASVs were present vs. absent (Table S2). 101 Co-occurrence with non-host 18S taxa With high statistical significance (P<0.005 and Q<0.01 by Local Similarity, Pearson’s and Spearman’s correlation), UCYN-A co-occurred with numerous 18S ASVs, such as prymnesiophytes and Dinoflagellates, notably including parasitic Syndiniales and an ASV from the genus Lepidodinium. UCYN-A1 and UCYN-A2 shared several of these taxa in common (Figure 5, Figure S4, Table S3). The Lepidodinium ASV was present at SPOT under environmental conditions similar to those found when UCYN-A ASVs were present (Table S2). UCYN-A host-symbiont co-occurrences UCYN-A2 co-occurred with B. bigelowii ASV3 on ~60% of the dates it was present at the SPOT surface (Figure 6A-C, S5, S6). On average, the ratio of UCYN-A2 16S: B. bigelowii ASV3 18S was about 2:1 at the SPOT surface (Figure 6C). Another ASV of B. bigelowii, ASV1, peaked in relative abundance on dates where UCYN-A2 was present, but B. bigelowii ASV3 was absent (Figure 6D). At the DCM, Braarudospharea ASV3 was present on ~60% of the dates UCYN-A2 was present, but the ratio of their 16S: 18S genes of these organisms was about 1:1 on average (Figure S6). The relative abundance of UCYN-A1 closely mirrored that of Braarudosphaera ASV7x, the closest match for the known host of UCYN-A1 (BIOSOPE T60; Figure 2B), at the SPOT surface (Figure 7A-C, S5, S7). The 16S gene of UCYN-A1 and 18S gene of Braarudosphaera ASV7x consistently co-occurred in nearly a 2:1 ratio at 5m depth (Figure 7C). In addition, UCYN-A1 co-occurred with Braarudosphaera ASV7x in eLSA networks with high statistical 102 significance at the SPOT surface (Figure 5). These organisms mirror each other less strongly at the DCM, where their ratio of 16S: 18S genes is also approximately 2:1 (Figure S7). Spatial-temporal distributions of UCYN-A ASVs UCYN-A ASVs were primarily found in the larger size fraction in the euphotic zone and were also detected once at 890m at about 0.5% of the 16S community (Figure S8, S9). UCYNA1 was found on 44% of sampling dates, while UCYN-A2 was present at the SPOT surface on 38.4% of all sampling dates (Figure S3, 6A, 7A). Relative abundances of all Braarudosphaera ASVs were highest in the euphotic zone, and they appeared at depths greater than the DCM even more rarely than UCYN-A. Discussion Relationship with abiotic factors Currents flowing from the tropical Pacific to the north and east likely brought the UCYNA1 symbiosis, the “open ocean” ecotype, into our study system. These currents are particularly strong during El Niño conditions and less so during regional upwelling. Positive Multivariate ENSO Index (MEI) values indicate El Niño conditions, in which warm, tropical waters flow west to east across the equator, altering climate patterns throughout the Pacific Ocean and weakening upwelling (e.g. [52, 53]). UCYN-A1 presence and relative abundance at our study site are positively correlated with higher MEI, i.e. El Niño conditions (Figure 3, Table S2, Figure 4, Figure S4). Our data show a sudden increase in the relative abundances of UCYN-A1 and host following the 2015-2016 El Niño event, which were sustained in the subsequent years (Figure 103 7A). Similarly, other time-series data show offshore ASVs were advected into the Southern California Bight after the same El Niño event [54], including warm ecotypes of Prochlorococcus at our study site [55]. Prior to this event, e.g. in 2010-2011, the UCYN-A1 symbiosis was infrequently detected at our sampling site and was similarly absent from other locations within the Southern California Current during overlapping periods of sampling [20, 22]. In addition, multiple lines of evidence suggest that UCYN-A1 presence and relative abundance are negatively influenced by seasonal upwelling. In this annual, coastal event, Ekman transport displaces surface waters west and south from SPOT, and nutrient-rich waters rise from the deep to replace them, stimulating increased bacterial chlorophyll concentrations and productivity (e.g. [35, 56]). Upwelling is here measured by the BEUTI and CUTI indices [35], with lower BEUTI and CUTI values indicating weaker upwelling. Lower BEUTI and CUTI are correlated with UCYN-A1 presence (Figure 3, Table S2), while sparse binomial regression indicated that the CUTI index negatively affected UCYN-A1 presence at SPOT (see Supplemental Information). Indirect indicators of upwelling, including Ekman transport to the south and west, lower SST, higher bacterial production rates and [Chl-a], were also negatively correlated with UCYN-A1 presence and abundance (Figure 3, Table S2, Figure 4, Figure S4). In Monterey Bay, north of our study site, both UCYN-A1 and UCYN-A2 (the dominant sublineage at this site), also regularly appeared in August-October, after regional upwelling had ceased [23]. Previous studies have also related UCYN-A1 abundances to temperature: although UCYN-A distribution may not be directly influenced by temperature (e.g. [9, 18]), warmer waters are favorable to increased UCYN-A1 relative abundance [22, 23]. In tandem, these observations indicate that the “open ocean” ecotype of the UCYN-A symbiosis is advected into our study system by warm, tropical waters during El Niño and periods of relaxed upwelling, which foster 104 lower biomass and coincide with higher temperatures, providing ideal conditions for the openocean ecotype. UCYN-A1 relative abundances and nitrogen fixation rates were observed to increase days after upwelling in nearshore samples from the Scripps Pier and the Alaskan Beaufort shelf [22, 57]. The daily sampling resolution of these cruises or their proximity to shore may explain discrepancies with our study. Upwelling may stimulate UCYN-A seed populations at a daily timescale, perhaps due to community production of a micronutrient not measured here that is essential to the UCYN-A1 symbiosis [22, 58, 59]. SPOT data, which is collected monthly, may not capture these short-term dynamics. Instead, we observe that upwelling hinders UCYN-A proliferation. Notably, both previous studies collected samples solely in 2017, soon after the 2015-2016 El Niño event. This could have brought UCYN-A seed populations to the Scripps Pier study, which was concurrent with an influx of warm, low-density waters from the tropics. Our study shows that over longer time scales, ocean currents including this El Niño event and relaxed upwelling created warmer, low biomass conditions which were conducive to UCYN-A1. Differences between UCYN-A1 and UCYN-A2 symbioses regarding abiotic niches support the hypothesis partitioning them into coastal and open-ocean ecotypes (e.g. [14]). UCYN-A2 presence was not strongly affected by seasonal upwelling or El Niño (Figure 3), and its relative abundance was only correlated with two abiotic factors: BEUTI index (negative) and SST (positive) (Figure 3, Table S2, Figure S4). Given the paradigm presented here, climate change will likely have a strong but mixed effect on UCYN-A1 at SPOT. Intense El Niño events are expected to become more frequent with climate change [58], advecting warm, tropical waters, along with UCYN-A1, into Southern California more often. Simultaneously, upwelling events along the California Current System are 105 expected to intensify, as warming on land increases air pressure differences between land and sea, strengthening coastal winds and upwelling [56, 60, 61]. Notably, the degree to which upwelling may intensify is unclear [61], partially due to mitigative effects of El Niño events, which weaken upwelling [53]. Although upwelling deters UCYN-A1 proliferation at SPOT, the symbiosis was able to recover from upwelling events following the 2015-2016 El Niño event (Figure 7). Future populations of UCYN-A1 may adhere to this trend, such that the symbiosis reaches greater abundances following future El Niño events. Because UCYN-A symbioses comprise up to 95% of the nitrogen fixing population in Southern California, this could substantially increase inputs of fixed nitrogen [22]. Notably, inputs of fixed nitrogen are small in this region [22], but this may be due to eukaryotic grazing that rapidly consumes diazotrophs as they fix nitrogen [10]. Co-occurrence with non-host 18S taxa Based on co-occurrence data with high statistical significance (Figure 5, S4, Table S3), we hypothesize a predator-prey relationship between the UCYN-A symbiosis and Lepidodinium, which in turn may be parasitized by Syndiniales (Figure S4). The dinoflagellate genus Lepidodinium has been shown to prey on the tropical diazotroph Crocosphearea watsonii through Lotka-Volterra modeling of environmental data [28], and doubles its grazing rates at night, when C. watsonii organism fixes nitrogen [27]. Furthermore, Lepidodinium has been hypothesized as a grazer of UCYN-A symbioses, as the diameter of UCYN-A symbioses (1-3µm for UCYN-A1 symbioses and 4-10µm for UCYN-A2 symbioses) are similar to that of C. watsonii (3.9µm for C. watsonii [27]). Its appearance in our dataset supports the hypothesis that this genus could prey on UCYN-A and other nitrogen fixers. The order Syndiniales is primarily 106 known to parasitize dinoflagellates (e.g. [62-65]. Syndiniales ASVs most likely co-occur with UCYN-A indirectly, due to the presence of their dinoflagellate hosts (see Figure S4). However, it is possible that Syndiniales are parasitizing Braarudospharea directly, as some Syndiniales are able to parasitize non-dinoflagellate hosts (e.g. [66]). Our co-occurrence data and others show statistically significant co-occurrence links between haptophytes and Syndiniales (Figure 5, Figure S4, Table S3 [66]). Many of the associations discussed here have been reported previously. The paper that announced Braarudosphaera as the putative host of UCYN-A1 used a UCYN-A nifH probe and FACS (Fluorescent-Activated Cell Sorting) to show that UCYN-A physically associates with many families, including Dinophyceae, the family that contains Dinoflagellates [8]. Using 18S primers, Krupke et al. found that UCYN-A co-occurs with Rhizaria, Dinoflagellates, including Lepidodinium and Syndiniales, and prymnesiophytes including Chrysochromulina, which is taxonomically very close to (and in some cases may overlap with) Braarudospharea [67]. One recent study proposed that the prymnesiophyte called Chrysochromulina parkaea is a strain of B. bigelowii (99.87% identical 18S sequence to a Braarudosphaera sequence) and is also able to associate with UCYN-A [19]. It should be noted that co-occurrence of taxa does not imply a biological association [68]. It is possible that UCYN-A and Braarudosphaera do not interact with any of the taxa reported here, but happen to co-occur due to a mutually favorable environmental niche. Notably, the Lepidodinium ASV is present at SPOT under similar conditions as UCYN-A1 and UCYN-A2 (Table S2). However, few of these parameters differed significantly on dates Lepidodinium was present vs. absent, and it is unclear if they influence the organism’s relative abundance. Additional studies should be performed to elucidate true interactions. For example, CARD FISH 107 with labelled probes for Lepidodinium and for UCYN-A could reveal physical evidence of predation. To tease apart other potential interactions, multi-stressor grow-out experiments could be conducted, in which temperature and availability of key nutrients are each varied along a gradient (e.g. [69]). UCYN-A host-symbiont co-occurrences The relationship between UCYN-A2 and host appears inconsistent from ASV cooccurrences. UCYN-A2 was present at the SPOT surface more frequently and in higher relative abundances than its presumed host, B. bigelowii ASV3 (38.4% vs. 28% of sampling dates, Figure 6, S5). The expected ratio of symbiont 16S: host 18S rRNA genes is unclear. Microscopy images show one symbiont per cell [19]. Based on DNA sequencing data, the UCYN-A2 host is thought to have 30-40 copies of the 18S rRNA gene [14], and 7-10 symbionts with two rRNA genes each [16], such that the ratio of UCYN-A 16S rRNA genes: 18S Braarudosphaera rRNA genes should be about 0.66 at most. We do not expect a constant ratio or a linear relationship between host 18S and symbiont 16S genes, due to polyploidy in the host or symbiont, differences in copy number of host rRNA genes, or different stages of cell division for symbiont or host. However, the average ratio of sequences was higher than expected (average=2.729), and their genes showed poor correlation (Figure 6B-C). We also observe that the same UCYN-A2 symbiont co-occurs with another potential host, B. bigelowii ASV1. On two days in particular, UCYN-A2 peaked in relative abundance when the abundance of its presumed host, B. bigelowii ASV3, was lower than expected. On both of these days, and on no other days, B. bigelowii ASV1 also peaked in relative abundance (Figure 6C). Both B. bigelowii ASV1 and ASV3 are closely related to genotypes of Braarudosphaera found 108 to associate with UCYN-A2 in the coastal waters of Japan. Hagino et al. used full length 18S sequencing, CARD FISH, and electron microscopy to observe that multiple cryptic species of B. bigelowii from distinct habitats amid the Japanese islands possessed UCYN-A symbionts [12, 17]. Notably, B. bigelowii ASV1 and ASV3 cluster with different pseudo-species: B. bigelowii ASV1 clusters with Intermediate form 1A, Genotype I, while B. bigelowii ASV3 clusters with Large form, Genotype IV [17] (Figure 2B). This shows that the 18S gene fragments used here can differentiate between cryptic species, and suggests that the UCYN-A2 ASV we observed may have more than one host. It may be that the UCYN-A2-B. bigelowii ASV1 partnership is the dominant symbiosis at a different location–consistent with the observation that different B. bigelowii cryptic species dominate different local environments in Japan [12]–and was brought to SPOT certain dates. It is also possible that these two hosts have different UCYN-A symbionts that are identical over the portion of 16S that we sequenced: full length sequencing of the symbiont 16S rRNA gene may differentiate two closely related symbionts. Whether or not the SPOT UCYN-A2 symbiont has two partners, UCYN-A2 and its established host are present at SPOT under different environmental conditions. The UCYN-A2 symbiont was present on days with higher concentrations of phosphate and increased upwelling (currents flowing to the south and west) compared to its host, although these differences are not statistically significant (Table S2). Culture-based studies also support the idea of weak coupling between UCYN-A2 symbiont and host: the UCYN-A2 host appears able to dissociate from its N2-fixing symbiont, and transcriptomics suggest it may instead rely on predation as a nitrogen source [19]. In contrast, UCYN-A1 presence more closely paralleled that of its established Braarudosphaera host. On >75% of days UCYN-A1 or Braarudosphaera ASV7x were present 109 at SPOT, the two organisms were present in an average ratio of 2.176 (Figure 7, S5), close to the expected 2:1 ratio of UCYN-A1 16S: Braarudosphaera 18S rRNA genes (UCYN-A1 has two 16S rRNA genes and its host is thought to have one copy of the 18S rRNA gene based on its small biomass [14, 16]). Each organism peaked in relative abundance only on days that its partner was also present. Conversely, on dates that one half of the pair was present but the other absent, the organism was only present in low abundances (Figure 7). These two organisms were present under similar environmental conditions (Table S2), and co-occurred with high levels of statistical significance (Figure 5). Many studies have shown a tight metabolic partnership between UCYN-A1 and its haptophyte host (e.g. [51, 70]), evidence that this host-symbiont pair are unable to survive without each other. This study shows that UCYN-A1 and its symbiont are seldom found apart, adding to this existing evidence. Another ASV from the UCYN-A1 sublineage, SPOT UCYN-A ASV6, was also present at the SPOT surface (Figure S3), but did not co-occur with any Braarudosphaera ASVs at any level of statistical significance (Figure S7, S8, S10). Other studies have reported free-living UCYN-A1 symbionts in open-ocean regions but attributed this phenomenon to host-symbiont dissociation during sample collection [18]. However, these studies used CARD-FISH and clustered 16S sequences into operational taxonomic units (OTUs) rather than ASVs, methods that miss the high-resolution differences between 16S ASVs from the same clade of organisms. This further illustrates the utility of high-resolution 16S ASVs (e.g. [25]). Spatial-temporal distributions of UCYN-A ASVs As expected, UCYN-A ASVs were primarily found at 5m and the DCM throughout the time series (Figure S3, S9), consistent with reports that its host is photosynthetic [8]. However, 110 UCYN-A1 was found as deep as 890m depth (Figure S9) at around the same time prior studies noted the UCYN-A nifH gene at the bottom of the San Pedro channel [25]. This strengthens existing evidence of the host-symbiont organisms’ capacity to export calcareous carbonate and fixed nitrogen from the euphotic zone [8], although this study does not address whether nitrogen fixation occurs at these depths. Braarudosphaera ASVs were found almost exclusively in the euphotic zone; the limited depth range of this genus is unique amongst Prymnesiophytes [9, 18]. Both the coastal and open ocean ecotypes of UCYN-A were present at SPOT, reflective of the fact that SPOT is an open ocean sampling site but is close enough to shore to be influenced by coastal dynamics [71]. UCYN-A1 appeared in 44% of all large size fraction samples collected from 5m depth at SPOT (1-80m), and often in relative abundances as high as 2.5% of the whole community (Figure S3, 7). On several occasions, UCYN-A1 was found in very low relative abundances in the smaller size class of organisms (0.22-1µm) (Figure 7A, inset). The UCYN-A1 symbiont is known to dissociate from its host, most likely due to the gentle pressures of sample filtration, which explains why it is found in multiple size fractions [11]. UCYN-A2 appeared in almost as many SPOT surface samples as UCYN-A1 (38.4%), but in lower relative abundances, at most 0.3% of the community (Figure S3, 6). This ASV was never found in the smaller size fraction, which further indicates that the ASV is from UCYN-A2: UCYN-A2 organisms are reported to have a diameter >1µm [51] and should consistently remain in the larger size fraction (1-80µm). The UCYN-A2 symbiosis was outnumbered by its Clade 1 counterpart at our study location: both were present on ~40% of sampling dates (Figure S3, 6A, 7A) and co-occurred with one another (Figure 4, Figure S6), but UCYN-A1 was present at almost an order of magnitude higher relative abundance than UCYN-A2 (at most 2.5% of the 111 community vs. 0.3%; Figure 7A, 6A). Similar patterns were seen in the same current system, but south of our study site: UCYN-A2 symbioses were present at consistently low abundances, while UCYN-A1 symbioses had well defined seasonal patterns [22]. This may be because the “coastal ecotype” is accustomed to living in environments with more biomass than is typically present at SPOT, and the study location is more comparable to environments preferred by the “open ocean ecotype.” Conclusions This paper reports trends in the spatio-temporal dynamics of the symbiotic diazotroph UCYN-A, its haptophyte hosts, and other associated taxa over ten years off the California Coast. We present important differences between two highly studied clades of UCYN-A with regards to host-symbiont relationships, co-occurrences with potential predators, and abiotic parameters. Studying these changes is particularly important as climate change continues to alter the distributions of UCYN-A, its hosts, and other associated 18S taxa [2]. Acknowledgements The SPOT dataset is the collective effort of generations of undergraduate, graduate student, and postdoctoral researchers. We particularly wish to thank Dave Caron, Jed Fuhrman, Troy Gunderson, Diane Kim, and the crew of the R/V Yellowfin for their support. We gratefully acknowledge invaluable conversations with Ana María Cabello, Virginia Edgcomb, Kendra-Turk Kubo, Jon Zehr, and others at the ASLO Ocean Sciences Meetings in 2020 and 2022. Additional comments from two anonymous peer reviewers helped improve the manuscript. This work was supported by NSF OCE 1737409, Gordon and Betty Moore Foundation Marine Microbiology 112 Initiative grant 3779, and Simons Foundation Collaboration on Computational Biogeochemical Modeling of Marine Ecosystems (CBIOMES) grant 549943. Research was conducted on land and waters traditionally stewarded by the Tongva people. Author contributions CFH devised the research questions, led data analysis, data visualization, and interpretation of results, and drafted the manuscript. YCY carried out sample processing and sequencing. YR contributed to data analysis and advised on manuscript writing. JLW advised on statistical analyses. JAF supervised the work and advised on data analysis and manuscript drafting. All authors edited the manuscript. Competing interests The authors declare no competing interests. Data Availability Statement Forward and reverse reads from each sample in the San Pedro Ocean Time-series (SPOT) are available at EMBL under accession number PRJEB48162 and PRJEB35673, as described by Yeh and Fuhrman [34]. Scripts necessary to reproduce the analysis are available at https://github.com/jcmcnch/eASV-pipeline-for-515Y-926R [31]. ASV tables generated from these files are available in /OriginalFiles at https://osf.io/6ku49/. Input files for the analyses presented here are available in /ModifiedFiles at the same link. Scripts necessary to reproduce these analyses are available at https://github.com/fletchec99/UCYNA_at_SPOT. 113 References 1. Zehr JP. Nitrogen fixation by marine cyanobacteria. Trends in Microbiology 2011; 19: 162–173. 2. Zehr JP, Capone DG. Changing perspectives in marine nitrogen fixation. Science 2020; 368. 3. Bombar D, Paerl RW, Riemann L. Marine Non-Cyanobacterial Diazotrophs: Moving beyond Molecular Detection. Trends in Microbiology 2016; 24: 916–927. 4. Delmont TO, Pierella Karlusich JJ, Veseli I, Fuessel J, Eren AM, Foster RA, Bowler C, Wincker P, Pelletier E. Heterotrophic bacterial diazotrophs are more abundant than their cyanobacterial counterparts in metagenomes covering most of the sunlit ocean. ISME J. 2022; 16(4):927-936. doi: 10.1038/s41396-021-01135-1. 5. Gradoville MR, Bombar D, Crump BC, Letelier RM, Zehr JP, White AE. Diversity and activity of nitrogen-fixing communities across ocean basins. Limnology and Oceanography 2017; 62: 1895–1909. 6. Karlusich JJP, Pelletier E, Lombard F, Carsique M, Dvorak E, Colin S, et al. Global distribution patterns of marine nitrogen-fixers by imaging and molecular methods. Nature Communications 2021; 12. 7. Zehr JP, Mellon MT, Zani S. New Nitrogen-Fixing Microorganisms Detected in Oligotrophic Oceans by Amplification of Nitrogenase (nifH) Genes. Applied and Environmental Microbiology 1998; 64: 3444. 8. Thompson AW, Foster RA, Krupke A, Carter BJ, Musat N, Vaulot D, et al. Unicellular cyanobacterium symbiotic with a single-celled eukaryotic alga. Science (New York, NY) 2012; 337: 1546. 9. Martinez-Perez C, Mohr W, Loscher CR, Dekaezemacker J, Littmann S, Yilmaz P, et al. The small unicellular diazotrophic symbiont, UCYN-A, is a key player in the marine nitrogen cycle. Nat Microbiol 2016; 1: 16163. 10. Wang W-L, Moore JK, Martiny AC, Primeau FW. Convergent estimates of marine nitrogen fixation. Nature 2019; 566: 205. 11. Tripp HJ, Bench SR, Turk KA, Foster RA, Desany BA, Niazi F, et al. Metabolic streamlining in an open-ocean nitrogen-fixing cyanobacterium. Nature 2010; 464: 90– 94. 12. Hagino K, Takano Y, Horiguchi T. Pseudo-cryptic speciation in Braarudosphaera bigelowii (Gran and Braarud) Deflandre. Marine Micropaleontology 2009; 72: 210– 221. 114 13. Thompson A, Carter BJ, Turk-Kubo K, Malfatti F, Azam F, Zehr JP. Genetic diversity of the unicellular nitrogen-fixing cyanobacteria UCYN-A and its prymnesiophyte host. Environ Microbiol 2014; 16: 3238–3249. 14. Bombar D, Heller P, Sanchez-Baracaldo P, Carter BJ, Zehr JP. Comparative genomics reveals surprising divergence of two closely related strains of uncultivated UCYN-A cyanobacteria. ISME J 2014; 8: 2530–2542. 15. Turk-Kubo KA, Farnelid HM, Shilova IN, Henke B, Zehr JP. Distinct ecological niches of marine symbiotic N2-fixing cyanobacterium Candidatus Atelocyanobacterium thalassa sublineages. J Phycol 2017; 53: 451–461. 16. Cornejo-Castillo FM, Muñoz-Marín M del C, Turk-Kubo KA, Royo-Llonch M, Farnelid H, Acinas SG, et al. UCYN-A3, a newly characterized open ocean sublineage of the symbiotic N2 -fixing cyanobacterium Candidatus Atelocyanobacterium thalassa. Environmental Microbiology 2019; 21: 111–124. 17. Hagino K, Onuma R, Kawachi M, Horiguchi T. Discovery of an Endosymbiotic Nitrogen-Fixing Cyanobacterium UCYN-A in Braarudosphaera bigelowii (Prymnesiophyceae). PloS ONE 2013; 8: e81749. 18. Cabello AM, Cornejo-Castillo FM, Raho N, Blasco D, Vidal M, Audic S, et al. Global distribution and vertical patterns of a prymnesiophyte-cyanobacteria obligate symbiosis. ISME Journal 2016; 10: 693–706. 19. Suzuki S, Kawachi M, Tsukakoshi C, Nakamura A, Hagino K, Inouye I, et al. Unstable Relationship Between Braarudosphaera bigelowii (= Chrysochromulina parkeae) and Its Nitrogen-Fixing Endosymbiont. Frontiers in Plant Science 2021; 12. 20. Cornejo-Castillo FM, Cabello AM, Salazar G, Sanchez-Baracaldo P, Lima-Mendez G, Hingamp P, et al. Cyanobacterial symbionts diverged in the late Cretaceous towards lineage-specific nitrogen fixation factories in single-celled phytoplankton. Nat Commun 2016; 7: 11071. 21. Sohm JA, Subramaniam A, Gunderson TE, Carpenter EJ, Capone DG. Nitrogen fixation by Trichodesmium spp. and unicellular diazotrophs in the North Pacific Subtropical Gyre. Journal of Geophysical Research: Biogeosciences 2011; 116. 22. Turk-Kubo KA, Mills MM, Arrigo KR, Dijken G van, Henke BA, Stewart B, et al. UCYN-A/haptophyte symbioses dominate N2 fixation in the Southern California Current System. ISME Communications 2021; 1. 23. Cabello AM, Turk-Kubo KA, Hayashi K, Jacobs L, Kudela RM, Zehr JP. Unexpected presence of the nitrogen-fixing symbiotic cyanobacterium UCYN-A in Monterey Bay, California. J Phycol 2020. 24. Hamersley MR, Turk KA, Leinweber A, Gruber N, Zehr JP, Gunderson T, et al. Nitrogen fixation within the water column associated with two hypoxic basins in the 115 Southern California Bight. Aquatic Microbial Ecology 2011; 63: 193–205. 25. Needham DM, Fichot EB, Wang E, Berdjeb L, Cram JA, Fichot CG, et al. Dynamics and interactions of highly resolved marine plankton via automated high-frequency sampling. ISME Journal 2018; 12: 2417–2432. 26. Tang W, Wang S, Fonseca-Batista D, Dehairs F, Gifford S, Gonzalez AG, et al. Revisiting the distribution of oceanic N2 fixation and estimating diazotrophic contribution to marine production. Nat Commun 2019; 10: 831. 27. Deng L, Cheung S, Liu H. Protistal Grazers Increase Grazing on Unicellular Cyanobacteria Diazotroph at Night. Frontiers in Marine Science 2020; 7. 28. Dugenne M, Freitas FH, Wilson ST, Karl DM, White AE. Life and death of Crocosphaera sp. in the Pacific Ocean: Fine scale predator–prey dynamics. Limnology and Oceanography 2020; 65: 2603–2617. 29. Hagino K, Young JR. Marine Protists. 2015. Springer Japan. 30. Parada AE, Needham DM, Fuhrman JA. Every base matters: assessing small subunit rRNA primers for marine microbiomes with mock communities, time series and global field samples. Environmental Microbiology 2016; 18: 1403–1414. 31. McNichol J, Berube PM, Biller SJ, Fuhrman JA. Evaluating and Improving Small Subunit rRNA PCR Primer Coverage for Bacteria, Archaea, and Eukaryotes Using Metagenomes from Global Ocean Surveys. 2021. 32. Yeh Y-C, McNichol J, Needham DM, Fichot EB, Berdjeb L, Fuhrman JA. Comprehensive single-PCR 16S and 18S rRNA community analysis validated with mock communities, and estimation of sequencing bias against 18S. Environmental Microbiology 2021; 23: 3240–3250. 33. Fuhrman JA, Cram JA, Needham DM. Marine microbial community dynamics and their ecological interpretation. Nature Reviews Microbiology 2015; 13: 133. 34. Yeh Y-C, Fuhrman JA. Contrasting diversity patterns of prokaryotes and protists over time and depth at the San-Pedro Ocean Time series. ISME Communications 2022; 2. 35. Jacox MG, Edwards CA, Hazen EL, Bograd SJ. Coastal Upwelling Revisited: Ekman, Bakun, and Improved Upwelling Indices for the U.S. West Coast. Journal of Geophysical Research: Oceans 2018; 123: 7332–7350. 36. Fuhrman JA, Hewson I, Schwalbach MS, Steele JA, Brown MV, Naeem S. Annually reoccurring bacterial communities are predictable from ocean conditions. Proceedings of the National Academy of Sciences of the United States of America 2006; 103: 13104– 13109. 37. Katoh K, Standley DM. MAFFT multiple sequence alignment software version 7: 116 Improvements in performance and usability. Molecular Biology and Evolution 2013; 30: 772–780. 38. Capella-Gutiérrez S, Silla-Martínez JM, Gabaldón T. trimAl: A tool for automated alignment trimming in large-scale phylogenetic analyses. Bioinformatics 2009; 25: 1972–1973. 39. Stamatakis A. RaxML version 8: A tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics 2014; 30: 1312–1313. 40. Letunic I, Bork P. Interactive Tree of Life (iTOL) v4: Recent updates and new developments. Nucleic Acids Research 2019; 47. 41. Wickham H. ggplot2: Elegant Graphics for Data Analysis. 2016. Springer-Verlag. 42. Zeileis A, Wien W, Grothendieck G. Journal of Statistical Software zoo: S3 Infrastructure for Regular and Irregular Time Series. 2005. 43. Gloor GB, Macklaim JM, Pawlowsky-Glahn V, Egozcue JJ. Microbiome Datasets Are Compositional: And This Is Not Optional. Front Microbiol 2017; 8: 2224. 44. Coenen AR, Hu SK, Luo E, Muratore D, Weitz JS. A Primer for Microbiome TimeSeries Analysis. Frontiers in Genetics 2020; 11. 45. Yoon G, Gaynanova I, Müller CL. Microbial networks in SPRING – Semi-parametric rank-based correlation and partial correlation estimation for quantitative microbiome data. Frontiers in Genetics 2019; 10. 46. Xia LC, Steele JA, Cram JA, Cardon ZG, Simmons SL, Vallino JJ, et al. Extended local similarity analysis (eLSA) of microbial community and other time series data with replicates. BMC Systems Biology 2011; 5. 47. Xia LC, Ai D, Cram J, Fuhrman JA, Sun F. Efficient statistical significance approximation for local similarity analysis of high-throughput time series data. Bioinformatics 2013; 29: 230–237. 48. Cram JA, Xia LC, Needham DM, Sachdeva R, Sun F, Fuhrman JA. Cross-depth analysis of marine bacterial networks suggests downward propagation of temporal changes. ISME Journal 2015; 9: 2573–2586. 49. Franz M, Lopes CT, Huck G, Dong Y, Sumer O, Bader GD. Cytoscape.js: A graph theory library for visulaization and analysis. Bioinformatics 2016; 32: 309–311. 50. Zehr JP, Bench SR, Carter BJ, Hewson I, Niazi F, Shi T, et al. Globally distributed uncultivated oceanic N2-fixing cyanobacteria lack oxygenic photosystem II. Science 2008; 322: 1110–1112. 51. Zehr JP, Shilova, I.N, Farnelid HM, Muñoz-Maríncarmen M.D., Turk-Kubo KA. 117 Unusual marine unicellular symbiosis with the nitrogen-fixing cyanobacterium UCYNA. Nature Microbiology 2016; 2. 52. Trenberth KE. The Definition of El Niño. Bull Amer Meteor Soc 1997; 78: 2771–2777. 53. Jacox MG, Fiechter J, Moore AM, Edwards CA. ENSO and the California Current coastal upwelling response. J Geophys Res Oceans 2015; 120: 1691–1702. 54. James CC, Barton AD, Allen LZ, Lampe RH, Rabines A, Schulberg A, et al. Influence of nutrient supply on plankton microbiome biodiversity and distribution in a coastal upwelling region. Nat Commun 2022; 13: 2448. 55. Yeh Y-C, Fuhrman JA. Contrasting diversity patterns of prokaryotes and protists over time and depth at the San-Pedro Ocean Time series. ISME Communications 2022; 2. 56. Bakun A. Global Climate Change and Intensification of Coastal Ocean Upwelling. Science 1990; 247: 198–201. 57. Selden CR, Einarsson SV, Lowry KE, Crider KE, Pickart RS, Lin P, et al. Coastal upwelling enhances abundance of a symbiotic diazotroph (UCYN-A) and its haptophyte host in the Arctic Ocean. Front Mar Sci 2022; 9: 877562. 58. Seneviratne, S.I., Zhang, X., Adnan, M., Badi, W. Weather and Climate Extreme Events in a Changing Climate. Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. 2021. Cambridge University Press, Cambridge, pp 1513–1766. 59. Bakun A, Black BA, Bograd SJ, García-Reyes M, Miller AJ, Rykaczewski RR, et al. Anticipated Effects of Climate Change on Coastal Upwelling Ecosystems. Curr Clim Change Rep 2015; 1: 85–93. 60. Arellano B, Rivas D. Coastal upwelling will intensify along the Baja California coast under climate change by mid-21st century: Insights from a GCM-nested physical-NPZD coupled numerical ocean model. Journal of Marine Systems 2019; 199: 103207. 61. García-Reyes M, Sydeman WJ, Schoeman DS, Rykaczewski RR, Black BA, Smit AJ, et al. Under Pressure: Climate Change, Upwelling, and Eastern Boundary Upwelling Ecosystems. Front Mar Sci 2015; 2. 62. Jephcott TG, Alves-de-Souza C, Gleason FH, Ogtrop FF van, Sime-Ngando T, Karpov SA, et al. Ecological impacts of parasitic chytrids, syndiniales and perkinsids on populations of marine photosynthetic dinoflagellates. Fungal Ecology 2016; 19: 47–58. 63. Anderson SR, Harvey EL. Temporal Variability and Ecological Interactions of Parasitic Marine Syndiniales in Coastal Protist Communities. mSphere 2020; 5. 64. Mazzillo FFM, Ryan JP, Silver MW. Parasitism as a biological control agent of 118 dinoflagellate blooms in the California Current System. Harmful Algae 2011; 10: 763– 773. 65. Clarke LJ, Bestley S, Bissett A, Deagle BE. A globally distributed Syndiniales parasite dominates the Southern Ocean micro-eukaryote community near the sea-ice edge. ISME J 2019; 13: 734–737. 66. Berdjeb L, Parada A, Needham DM, Fuhrman JA. Short-term dynamics and interactions of marine protist communities during the spring-summer transition. ISME J 2018; 12: 1907–1917. 67. Krupke A, Mohr W, LaRoche J, Fuchs BM, Amann RI, Kuypers MM. The effect of nutrients on carbon and nitrogen fixation by the UCYN-A-haptophyte symbiosis. ISME J 2015; 9: 1635–1647. 68. Blanchet FG, Cazelles K, Gravel D. Co-occurrence is not evidence of ecological interactions. Ecol Lett 2020; 23: 1050–1063. 69. Kling JD, Lee MD, Fu F, Phan MD, Wang X, Qu P, et al. Transient exposure to novel high temperatures reshapes coastal phytoplankton communities. ISME J 2020; 14: 413– 424. 70. Mills MM, Turk-Kubo KA, Dijken GL van, Henke BA, Harding K, Wilson ST, et al. Unusual marine cyanobacteria/haptophyte symbiosis relies on N2 fixation even in Nrich environments. ISME J 2020; 14: 2395–2406. 71. Teel EN, Liu X, Seegers BN, Ragan MA, Haskell WZ, Jones BH, et al. Contextualizing time-series data: quantification of short-term regional variability in the San Pedro Channel using high-resolution in situ glider data. Biogeosciences 2018; 15: 6151. 119 Chapter 4 Figures Ch. 4 Figure 1–Sampling site location for the San Pedro Ocean Time-series (SPOT). SPOT is ~16km from the Port of Los Angeles (33.55ºN, 118.4ºW). USC, University of Southern California. Sample collection is described in detail in Yeh et al. [33]. Santa Catalina Island USC SPOT 33.0°N 33.2°N 33.4°N 33.6°N 33.8°N 34.0°N 119.0°W 118.8°W 118.6°W 118.4°W 118.2°W 118.0°W 117.8°W Longitude Latitude 120 Ch. 4 Figure 2–UCYN-A 16S sequences (A) and Braarudosphearea 18S sequences (B) from SPOT are phylogenetically identical to or very close to the reference sequences for UCYN-A1, UCYNA2, and hosts shown to associate with these symbionts. Bolded names indicate UCYN-A sequences from SPOT; yellow highlighted names indicate 16S sequences from the published genomes of symbionts. Braarudospharea ASVs that associate with UCYN-A are bolded and highlighted in green, purple, and blue. Trees were generated via RaxML with 100 bootstraps; numbers indicate bootstrap values. UCYN- A SPOT ASV2 Cyanobacterium endosymbiont 1 B.bigelowii- - Hagino et al 2013 Cyanobacterium endosymbiont 2 B.bigelowii- - Hagino et al 2013 Cyanobacterium endosymbiont 3 B.bigelowii- - Hagino et al 2013 UCYN- A SPOT ASV5 JPSP01000022.1 UCYN- A2 isolate- - Bombar et al 2014 JPSP01000003.1 UCYN- A2 isolate- - Bombar et al 2014 UCYN- A 7656BH923 SP6- - Thompson et al 2012 UCYN- A 7656BH902 SP6- - Thompson et al 2012 UCYN- A 7656BH927 SP6- - Thompson et al 2012 UCYN- A 7656BH943 SP6- - Thompson et al 2012 UCYN- A SPOT ASV3 UCYN- A SPOT ASV1 NC 013771.1UCYN- A1 isolate ALOHA- - Zehr et al. 2008 CP001842.1 UCYN- A1 isolate ALOHA- - Zehr et al. 2008 UCYN- A 7656BH984 SP6- - Thompson et al 2012 UCYN- A 7656BH884 SP6- - Thompson et al 2012 UCYN- A 7656BH911 SP6- - Thompson et al 2012 UCYN- A 7656BH863 SP6- - Thompson et al 2012 UCYN- A JC142 R11- 110- - Cornejo Castillo et al 2018 UCYN- A SPOT ASV4 Crocosphaera watsonii- - Mazard et al 2004 Tr ci hodesmium erythraeum IMS101 -- Tomitaniet al 2001 UCYN- A SAGAD- 638J10- - Cornejo Castillo et al 2018 UCYN- A SPOT ASV6 4 47 63 16 59 55 12 13 91 66 21 7 75 5 99 4 100 69 95 24 61 62 Treescale: 0.1 Chrysochromulina SPOT ASV2 Chrysochromulina SPOT ASV3 Chrysochromulina SPOT ASV1 Braarudosphaera sp.X SPOT ASV6x Braarudosphaera sp.X SPOT ASV5x Braarudosphaera sp.X SPOT ASV7x BIOSOPET60.034- - 99%IDto Thompson et al 2012 Yatsushiro- 1- - Hagino et al 2009 SIOJp1D8BH SP6- - Thompson et al 2014 SIOJp3E5BH SP6- - Thompson et al 2014 SIOJp3B4BH SP6- - Thompson et al 2014 Braarudosphaera bigelowii SPOT ASV4 Braarudosphaera bigelowii SPOT ASV3 SIOJp3E12BH SP6- - Thompson et al 2014 SIOJp3D5BH SP6- - Thompson et al 2014 SIOJp3E1BH SP6- - Thompson et al 2014 SIOJp3B7BH SP6- - Thompson et al 2014 Furue16- - Hagino et al 2009 Furue- 15- - Hagino et al 2009 Isolate KC15.24 NIES.4442- - Suzuki et al 2021 Tosa5- - Hagino et al 2009 Tosa14- - Hagino et al 2013 TMRscBb8- - Hagino et al 2013 TP05- 6- a- - Haginoet al 2006 TMRscBb7- - Hagino et al 2013 Tosa8- - Hagino et al 2009 Tosa6- - Hagino et al 2009 Isolate KC1.P2 NIES.3865- - Suzuki et al 2021 AB778293.1- - Hagino et al 2013 Tosa2- - Hagino et al 2009 Uska15.2.6- - Hagino et al 2009 TP05- 6- b- - Hagino et al 2006 Shukutsu19- - Hagino et al 2009 Shukutsu27- - Hagino et al 2009 Shukutsu22- - Hagino et al 2009 Braarudosphaera bigelowii SPOT ASV2 Braarudosphaera bigelowii SPOT ASV1 Funahama- T3- - Hagino et al 2009 100 84 100 33 4 6 56 1 66 37 81 20 54 28 98 88 14 99 78 3 78 90 58 1 73 7 65 Tree scale: 0.01 A) B) 121 Ch. 4 Figure 3–UCYN-A1 is present at SPOT under El Niño conditions, with relaxed upwelling, higher temperatures, and lower biomass (A-D); these patterns are similar but weaker for UCYNA2 (note larger p-values and error bars). Differences in the mean of each value when ASVs were present/ absent were evaluated via logistic regression. P-values reported were corrected for multiple testing via BenjaminiHochberg correction. Error bars represent standard error. Higher Multivariate ENSO Index (MEI) indicates El Niño conditions (more stratified); higher Biologically Effective Upwelling Transport Index (BEUTI) indicates stronger upwelling. Additional variables are shown in Supplementary Table 1, including statistically significant differences in upwelling (measured by Coastal Upwelling Transport Index (CUTI), bacterial production, and currents on days UCYN-A were present vs. absent. 122 Ch. 4 Figure 4–UCYN-A1 relative abundances correlate positively with host organisms, UCYNA2 host/symbionts, MEI, and SST, and negatively with upwelling indices (BEUTI and CUTI) and indirect indicators of upwelling, including bacterial production, East-West Ekman transport, and chlorophyll concentrations. UCYN-A symbiont/host ASVs were CLR-transformed and correlated with environmental parameters via Spearman’s rank-order correlation. Dot size indicates the strength of correlation, while dot color represents positive or negative associations. Only statistically significant correlations (p<0.05) are shown via dots; correlations between variable pairs that are insignificant or that do not correlate are represented by blank squares. BEUTI, Biologically Effective Upwelling Transport Index; CUTI, coastal upwelling transport index; EW, East-West; NS, North-South; MEI, Multivariate ENSO Index. 123 Ch. 4 Figure 5–UCYN-A co-occurs with a variety of 18S taxa at the SPOT surface, notably including Lepidodinium, hypothesized to be a predator (pink triangle, center). Networks were generated via eLSA and visualized in Cytoscape 3.5. Each node represents one ASV; only ASVs that co-occurred with P<0.005 and Q<0.01 by both Pearson’s and Spearman’s correlation are shown using easily recognizable names. QIIME-generated hashes of each node are presented in Table S3; 16S & 18S sequences of each node, including UCYN-A symbionts, are publicly available (see Data Availability Statement). 124 Ch. 4 Figure 6–UCYN-A2 is less tightly coupled to its prymnesiophyte host than the UCYN-A1 symbiosis. UCYN-A2 often, but not always, co-occurs with its established host at 5m depth across the SPOT time series. B) UCYN-A2 relative abundance correlates with relative abundance of its most common host more weakly than the UCYN-A1 symbiosis. C) The ratio of 16S: 18S genes of these organisms is lower than expected. Boxplot values indicate the median and interquartile range values of this ratio; the red line indicates the average (2.139). 29 sampling dates, on which host and symbiont are present, are included. D) UCYN-A2 occasionally co-occurs with Braarudospharea ASV1. The relative abundances of 16S and 18S ASVs were normalized as described in Supplementary Methods to allow for direct comparison. 125 Ch. 4 Figure 7–UCYN-A1 and its prymnesiophyte host are tightly coupled throughout the SPOT dataset. UCYN-A1 consistently co-occurs with its putative host at 5m depth across the SPOT time series. Panel inset indicates the relative abundance of UCYN-A1 in the smaller size fraction. B) UCYNA1 relative abundance strongly correlates with its putative host relative abundance. C) The ratio of 16S: 18S genes of these organisms is, on average, as expected. Boxplot values indicate the median and interquartile range (IQR) values of this ratio; the red line indicates the average (1.991). One outlier (ratio >50) was excluded. 43 sampling dates are included. As with UCYNA2, the relative abundances of 16S and 18S ASVs were normalized to allow for direct comparison. 126 Chapter 4 Supplementary Information Materials and Methods DNA SEQUENCING AND PROCESSING: 15-20L of seawater was filtered sequentially through an 80µm mesh (removing mesoplankton) a 1µm glass AE filter (Pall, Port Washington, NY) (collecting the larger 1-80µm size fraction), and finally a 0.2µm Durapore filter (ED Millipore, Billerica, MA) (collecting the 0.2-1µm smaller size fraction). DNA was extracted from Durapore filters using the phenol-chloroform method described by Fuhrman et al. [72]. DNA was extracted from AE filters via bead beating, followed by a phenol-chloroform protocol, as described in Lie et al. [73]. The V4-V5 hypervariable region of the 16S and 18S rRNA gene was amplified from these extracts using the universal primers (515Y/926R) as reported by Parada et al. [30], which amplify sequences from both eukaryotic and prokaryotic ribosomal RNA genes [31, 33]. Samples were sequenced on either a HiSeq 2500 in PE250 mode or MiSeq PE300 platform at the USC and UC Davis Genome Core facilities at a target depth of 100,000 sequences/ sample. Per-sample sequence files were submitted to the EMBL database under accession number PRJEB48162 and PRJEB35673. Sequences were processed into Amplicon Sequence Variants (ASVs), which differ by as little as a single base pair, using Divisive Amplicon Denoising Algorithm v2 (DADA2) implemented in Quantitative Insights Into Microbial Ecology v2 (QIIME2) [74] with scripts available at github.com/jcmcnch/eASVpipeline-for-515Y-926R. [31]. Prokaryotic and eukaryotic ASVs were taxonomically classified using SILVA 132 in May and July of 2020, respectively. PRINCIPLE COMPONENT ANALYSES: Principle component analysis (PCA) was used to visualize differences in environmental parameters on dates UCYN-A ASVs were absent (<0.01% 127 of the 16S community) vs. present in relative abundances higher or lower than average. PCA was conducted on abiotic data, which was centered and scaled, using prcomp() in R. Ordinations were plotted using autoplot() from the “ggfortify” package in conjunction with ggplot2 [41]. MODELING EFFECTS OF ENVIRONMENTAL PARAMETERS: Sparse binomial regression was used to resolve which environmental parameters best predicted whether UCYN-A1 and UCYN-A2 would be present at SPOT. Model input data consisted of bacteria production rates, nutrient availability, upwelling indices, and other environmental variables. Data from missing dates were linearly interpolated via na.approx() from the R package zoo [42]. UCYN-A ASVs were considered “present” on dates that they were over 0.01% of the 16S community, and “absent” when their relative abundances were lower than 0.01%. For each ASV, a sparse binomial logistic regression model was constructed via the glmnet package in R [75, 76]. 80% of the data was used as training data for the model, and 20% was used as the test set. Variable selection was performed using lasso regression, and the appropriate lambda was selected using 10-fold cross validation on the training set. F1, sensitivity, specificity, and accuracy of each model were calculated on the test set using the caret package [76]. DATA NORMALIZATION: Our DADA2 pipeline splits 16S and 18S sequences, generating separate tables of ASVs for prokaryotes and eukaryotes. In order to plot UCYN-A and associated eukaryotes with the same denominator, sequencing data was normalized as follows. Sequences from chloroplasts and metazoans were removed, leaving only SSU sequences from prokaryotes and single-celled eukaryotes in the dataset. Raw sequencing counts of prokaryotic and eukaryotic ASVs were divided by the percent of sequences passing quality control in DADA2. Because 128 HiSeq and MiSeq platforms have been shown to discriminate against the 18S rRNA sequences, favoring the shorter 16S rRNA sequences with a two-fold bias [77], sequence counts from eukaryotic ASVs were then doubled. Normalized sequencing counts of prokaryotic and eukaryotic ASVs were combined and converted to proportions, representing the relative abundances of taxa out of the entire microbial community (16S+18S sequences). This method was developed and successfully tested on mixed mock communities, which contain 16S and 18S rRNA sequences in equal concentrations [77], that were sequenced via HiSeq or MiSeq. Following normalization, communities contained equal proportions of each of the organisms in the sequenced sample, as expected (Figure S11). Code normalizing the 16S/ 18S ASV tables of this QIIME-2 pipeline [31] is available at https://github.com/fletchec99/normalizing_16S_18S_tags. Results and Discussion PRINCIPLE COMPONENT ANALYSES: Principle component analyses (PCA) indicate that temperature drove variation in the abiotic factors along PC1, which was generally associated with UCYN-A1 and host presence. Higher MEI (Multivariate ENSO Index) was also associated with UCYN-A1 and host presence in PCA. Upwelling indices, as well as indirect indicators of upwelling such as increased nutrient concentration, bacterial production, and chlorophyll concentration, drove variation along PC2 and were associated with UCYN-A1 and host absence. These trends were not as obvious for UCYN-A2 (Figure S3). MODELING EFFECTS OF ENVIRONMENTAL PARAMETERS: Sparse binomial logistic regression indicated that UCYN-A1 presence was negatively affected by upwelling and no other 129 variables (coefficient= -0.570, sensitivity=0.667, specificity=0.333, accuracy=0.542, F1=0.645). Models were not able to reliably predict the presence of UCYN-A2 (sensitivity=1.00, specificity=0.00, accuracy= 0.5, F1=0.667). RELATIONSHIP WITH INORGANIC NUTRIENTS: UCYN-A ASVs were not significantly correlated with inorganic nitrogen and phosphorus concentrations (Table S1, Figure 4). Others have observed UCYN-A abundances and activity have no strong relationships with nitrogen concentrations (e.g. [18, 70, 78]). Due to the well-established link between upwelling and increased inorganic nutrient concentrations (e.g. [35]), the strong influence of upwelling on UCYN-A1 might seem incongruous with the weak influence of inorganic nutrients. It is important to note that upwelling indices are aggregated across latitude on a monthly basis, whereas the inorganic nutrients were measured at SPOT on the day of sampling. Upwelling in Southern California is generally coastal, and it is likely that coastal phytoplankton close to the sites of upwelling consumed the upwelled nitrogen and phosphate, before these compounds could reach our study site, ~16km from the coast (Figure 1). Supplementary References 72. Fuhrman JA, Comeau DE, Hagstrom A, Chan AM. Extraction from Natural Planktonic Microorganisms of DNA Suitable for Molecular Biological Studies. Applied and Environmental Microbiology 1988; 54: 1426. 73. Lie A, Kim D, Schnetzer A, Caron D. Small-scale temporal and spatial variations in protistan community composition at the San Pedro Ocean Time-series station off the coast of southern California. Aquatic Microbial Ecology 2013; 70. 74. Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJ, Holmes SP. DADA2: High-resolution sample inference from Illumina amplicon data. Nat Methods 2016; 13: 581–583. 75. Friedman J, Hastie T, Tibshirani R. Regularization Paths for Generalized Linear Models 130 via Coordinate Descent. Journal of Statistical Software 2010; 33: 1–22. 76. Simon N, Friedman J, Hastie T, Tibshirani R. Regularization Paths for Cox’s Proportional Hazards Model via Coordinate Descent. Journal of Statistical Software 2011; 39: 1–13. 77. Katoh K, Standley DM. MAFFT multiple sequence alignment software version 7: Improvements in performance and usability. Molecular Biology and Evolution 2013; 30: 772–780. 78. Yeh Y-C, Needham DM, Sieradzki ET, Fuhrman JA. Taxon Disappearance from Microbiome Analysis Reinforces the Value of Mock Communities as a Standard in Every Sequencing Run. mSystems 2018; 3. 79. Turk-Kubo KA, Achilles KM, Serros TR, Ochiai M, Montoya JP, Zehr JP. Nitrogenase (nifH) gene expression in diazotrophic cyanobacteria in the Tropical North Atlantic in response to nutrient amendments. Front Microbiol 2012; 3: 386. 131 Chapter 4 Supplementary Figures Ch. 4 Supplementary Figure 1–Rarefaction curves of AE samples with <100 unique 18S ASVs (species) from the SPOT surface (A) and DCM (B). Five AE samples from the surface and two AE samples from the DCM were excluded from analyses due to insufficient diversity of 18S sequences. All samples with >100 unique 18S ASVs sufficiently captured the diversity of 18S organisms. Rarefaction curves were generated with the R package vegan version 2.5-6. 132 Ch. 4 Supplementary Figure 2–eLSA networks constructed using interpolated, non-CLR transformed data from the SPOT surface (5m depth) miss interactions between UCYN-A and other taxa (compare to Figure 5). Networks were generated via eLSA and visualized in Cytoscape 3.5. 133 Ch. 4 Supplementary Table 1–Identifiers for UCYN-A and Braarudospharea ASVs analyzed in this study. Hashes were generated by QIIME2 and were named in alphabetical order (e.g. UCYN-A ASV1 is first alphabetically in the list of six UCYN-A ASVs); representative sequences were classified using SILVA132. BLAST identity is given for the sequences that 100% match an established reference sequence for UCYN-A or its host. Full 16S and 18S sequences are publicly available (see Data Availability Statement). QIIME2 generated hash ASV name BLAST Identity (if applicable) 3d852410f44d21c92c9c55fbbb25187e UCYN-A SPOT ASV1 or UCYN-A1 100% identical to established UCYN-A1 16S sequence 42c9bf8576acc275f3c9281e6b24f5a3 UCYN-A SPOT ASV2 6115eab19c52bc45c6ba11d72ec88031 UCYN-A SPOT ASV3 a641110da9fb0da8f68143b5a79ba5d1 UCYN-A SPOT ASV4 af1bb1f9fb1c3f3d18571e711df407bb UCYN-A SPOT ASV5 or UCYN-A2 100% identical to established UCYN-A2 16S sequence e6f42c535cf3849e1f1e12e7575561b7 UCYN-A SPOT ASV6 04926e2fd1b8706b4866c02650f702dd Braarudospharea bigelowii SPOT ASV1 100% identical to established UCYN-A1 host 18S sequence 529269deeb5fb7fbf0d0ebda989d9d82 Braarudospharea bigelowii SPOT ASV2 70a5283da28db501a349c5beb22881e7 Braarudospharea bigelowii SPOT ASV3 8c144683114fbb1ad2e9425f7dcd1b02 Braarudospharea bigelowii SPOT ASV4 324627f7f367298bbb5692fc5038e680 Braarudosphaera SPOT ASV5x ab5338a49f7e9307027c50b3256a7f59 Braarudospharea SPOT ASV6x be3cdecefbceb0d8b25a2e42ed058b50 Braarudospharea SPOT ASV7x 134 Ch. 4 Supplementary Table 2–Differences in environmental factors on dates UCYN-A1, UCYNA2, hosts, and potential predator Lepidodinium ASVs are present vs. absent. Ekman transport is measured on the Cartesian coordinate system, such that movement Northwards and Eastwards is recorded as a more positive value. Positive MEI indicates El Niño events, while negative MEI represents La Nina conditions. P values were corrected for multiple testing via Benjamini-Hochberg correction. Boldface text indicates p<0.05, boldface, underlined text indicates p<0.01. Mean Standard Error P-values Mean Standard Error P-values Mean Standard Error P-values Present 0.471 0.288 0.514 0.327 0.626 0.333 Absent 0.562 0.164 0.531 0.152 0.462 0.135 Present 0.253 0.042 0.247 0.045 0.222 0.023 Absent 0.200 0.014 0.208 0.017 0.221 0.028 Present 284052.200 27744.452 305718.400 30237.819 365066.200 44642.812 Absent 481256.900 46408.102 453865.400 44129.048 422600.000 41673.916 Present 0.507 0.070 0.451 0.046 0.509 0.051 Absent 0.901 0.286 0.902 0.266 0.883 0.277 Present 19.090 0.400 19.079 0.431 18.878 0.347 Absent 17.604 0.247 17.724 0.247 17.796 0.292 Present 0.326 0.019 0.345 0.021 0.416 0.026 Absent 0.458 0.023 0.437 0.023 0.396 0.022 Present 0.378 0.098 0.368 0.102 0.680 0.159 Absent 1.021 0.130 0.977 0.124 0.806 0.110 Present -0.024 0.133 -0.027 0.152 -0.204 0.145 Absent -0.467 0.118 -0.431 0.110 -0.336 0.116 Present 6.850 0.255 6.881 0.253 7.468 0.314 Absent 8.094 0.244 7.980 0.243 7.657 0.231 Present -675.851 61.569 -712.083 58.921 -897.779 83.521 Absent -1039.232 71.392 -990.841 71.243 -885.622 65.968 Present -512.139 84.007 -536.981 80.845 -767.567 101.137 Absent -919.728 82.514 -874.361 83.054 -743.356 79.260 Mean Standard Error P-values Mean Standard Error P-values Present 0.603 0.217 0.620 0.355 Absent 0.494 0.260 0.477 0.144 Present 0.179 0.041 0.246 0.048 Absent 0.239 0.011 0.209 0.016 Present 329817.400 63965.889 298758.800 23202.408 Absent 428844.800 22388.304 451901.200 43959.033 Present 0.444 0.377 0.566 0.080 Absent 0.858 0.030 0.827 0.255 Present 19.242 0.415 19.197 0.451 Absent 17.799 0.260 17.714 0.241 Present 0.360 0.033 0.397 0.026 Absent 0.421 0.016 0.407 0.022 Present 0.321 0.186 0.630 0.163 Absent 0.933 0.057 0.821 0.110 Present -0.034 0.162 -0.183 0.160 Absent -0.386 0.114 -0.337 0.110 Present 7.057 0.354 7.284 0.297 Absent 7.796 0.199 7.736 0.235 Present -774.193 100.728 -815.777 70.719 Absent -937.017 52.125 -928.030 68.823 Present -583.699 119.591 -708.573 92.885 Absent -820.665 65.127 -775.010 81.202 North-South Component of Ekman Transport (kg/m/s) East-West Component of Ekman Transport (kg/m/s) Lepidodinium ASV MODIS [Chl] (mg * m-3) Bacterial Production (cells/mL/day) [PO4] (uM) MODIS SST (ºC) CUTI index BEUTI index MEI Magnitude of Surface Wind (m/s) [NO2+NO3] (uM) UCYN-A2 (ASV5) B. bigelowii ASV3 0.957 0.417 0.746 0.197 MODIS [Chl] (mg * m-3) MODIS SST (ºC) CUTI index 0.025 0.050 0.124 0.025 0.025 [NO2+NO3] (uM) [PO4] (uM) UCYN-A1 (ASV1) Braarudosphaera ASV7x Bacterial Production (cells/mL/day) East-West Component of Ekman Transport (kg/m/s) North-South Component of Ekman Transport (kg/m/s) 0.809 0.282 0.006 0.350 0.004 MEI Magnitude of Surface Wind (m/s) 0.004 0.004 0.004 0.025 0.004 0.004 BEUTI index 0.025 0.025 0.988 0.988 0.988 0.988 0.446 0.049 0.025 0.988 0.988 0.988 0.988 0.988 0.988 0.197 0.197 0.059 0.197 0.197 0.197 0.197 0.197 0.059 0.754 0.754 0.106 0.754 0.029 0.802 0.719 0.754 0.718 0.719 0.754 135 Ch. 4 Supplementary Figure 3–Principal component analysis of environmental variables at the SPOT surface, overlaid with UCYN-A1 relative abundance and host presence/ absence (A) and UCYN-A2 relative abundance and host presence/ absence (B), show that temperature and MEI associate strongly with high relative abundance of UCYN-A1, but less strongly with that of UCYN-A2. UCYN-A ASVs were considered “absent” if they were <0.01% of the 16S community, “low abundance” if they were present in abundances lower than average (0.099% for ASV1, UCYNA1, and 0.026% for ASV5, UCYN-A2), and “high abundance” if their relative abundance was greater than average. This is indicated by white, grey, and black points, respectively; host absence/ presence is indicated by squares vs. circles on both panels. 136 Ch. 4 Supplementary Figure 4–Schematic of a hypothetical food web between UCYN-A ASVs and the 18S taxa with which they co-occur at the SPOT surface (see Figure 5). Arrows indicate predation/ parasitism; lines indicate symbiosis. Ch. 4 Supplementary Table 3–UCYN-A ASVs co-occur with a variety of 18S taxa. Hashes were generated via QIIME2; representative sequences were classified using SILVA132. Full 16S and 18S sequences are publicly available (see Data Availability Statement). Also see Supplementary Table 2 for hashes associated with UCYN-A and Braarudosphaera ASVs. 137 Taxonomy ASV Hash Co-occurs with Alveolata 636f2ad7bbc9624ed6d8bd438d96f7b7 UCYN-A1_0.22-1um Alveolata b930d2802540ff419424186646952b7d UCYN-A1_0.22-1um Alveolata 3d3aaf7802f4db6f8bc887bb9e7774a8 UCYN-A1_1-80um Alveolata 4b7935f69f43802ce9ea4b65a741845d UCYN-A1_1-80um Alveolata 8e0bc96f64c6b0ca6e4aead874edee2a UCYN-A1_1-80um Alveolata b6d8ad4a0a89749e0427b20d028f317d UCYN-A1_1-80um Alveolata cccd609465491f29a68b22cdb1506556 UCYN-A1_1-80um Alveolata 8a5ef38b1ab34ac0febbd7832a3ed0d1 UCYN-A1_1-80um, UCYN-A1_0.22-1um Alveolata 2a77e539d5c3d4878fb0cc41c84a0fa3 UCYN-A1_1-80um, UCYN-A1_0.22-1um, UCYN-A2 Alveolata a9666cf5e782f71ad780c73f0f59c119 UCYN-A2 Braarudosphaera_sp_7x be3cdecefbceb0d8b25a2e42ed058b50 UCYN-A1_1-80um, UCYN-A1_0.22-1um Braarudospharea_bigelowii_3 70a5283da28db501a349c5beb22881e7 UCYN-A1_1-80um, UCYN-A2 Chlorophyta fd7bca02dc31e29cb36624e39cbcc27a UCYN-A1_1-80um, UCYN-A1_0.22-1um Chlorophyta c3b3a5a8e14ea2c9cdb58028bc55bca0 UCYN-A1_1-80um, UCYN-A1_0.22-1um, UCYN-A2 Chrysochromulina d55992e6da65321a9b3c0ce3426e73ac UCYN-A1_1-80um, UCYN-A2 Chrysochromulina eaaf40a3c970e0ec2167de48c4b001eb UCYN-A2 Dinoflagellate 26cc4edeb2a1788fc8fdf147923c06b4 UCYN-A1_0.22-1um Dinoflagellate 997d8cc0ee56b8f185884eb47f7bbbd6 UCYN-A1_0.22-1um Dinoflagellate 4bb7bd73a0c0f6a55e0c4acecdbfd099 UCYN-A1_1-80um Dinoflagellate 70241d94140d1eaea54e19bc52b1bdd1 UCYN-A1_1-80um Dinoflagellate 791b0b5d59992db5031a18510b325ced UCYN-A1_1-80um Dinoflagellate ead0d51affbef7dd17d17e483eb0f244 UCYN-A1_1-80um Dinoflagellate ef4dcdee685b3fd3427dff66b3a9e379 UCYN-A1_1-80um Dinoflagellate f383ee229e9cae5c76c76062e8a0f99f UCYN-A1_1-80um Dinoflagellate a2bcca80cacedb46e9e55cd424022684 UCYN-A1_1-80um, UCYN-A2 Dinoflagellate dc228f26735fb8483e2e7a8f07994c23 UCYN-A1_1-80um, UCYN-A2 Dinoflagellate 23f2bbf2ecbae94d2b57c207aa8abc62 UCYN-A2 Dinoflagellate 95e7881e06d651005303f80a0824f7f9 UCYN-A2 Dinoflagellate eaee3f386d9d2d2c4ea3d90ccf81f48a UCYN-A2 Dinoflagellate ec296e5b3180d5c65f9e81e7165a4763 UCYN-A2 Eukaryote 0edea9a57e97085c27800904bc55437d UCYN-A1_1-80um Eukaryote af7721ace95e845243e2b715dbcca683 UCYN-A1_1-80um Eukaryote 72ad5bae980d6f3a205494139435e4e5 UCYN-A1_1-80um, UCYN-A1_0.22-1um Lepidodinium c114523e0bef5840b096693e46f441a2 UCYN-A1_1-80um, UCYN-A1_0.22-1um, UCYN-A2 Prymnesiophyte 23206df8663c546b75a92fadb5de9b33 UCYN-A1_0.22-1um Prymnesiophyte 79c87d3c4954a1dec1972fc679befcb5 UCYN-A1_1-80um Prymnesiophyte 98d1369a93242f295deda7ee996b9886 UCYN-A1_1-80um Prymnesiophyte 803aa95deeb6167672a24a67029f2b82 UCYN-A1_1-80um, UCYN-A1_0.22-1um Prymnesiophyte ffa92571884bcbe0193ce4a1e4e843be UCYN-A1_1-80um, UCYN-A1_0.22-1um Prymnesiophyte 55bc25d102de6079cc48ba48515e72e2 UCYN-A2 Rhizaria 307e50f0be54fdc6ac0ea8183ff77d0e UCYN-A1_0.22-1um Rhizaria a615f49d7d2fbed5b85069fb94087fa3 UCYN-A1_1-80um Rhizaria ded52caa99b40106f449ecc43dbbeee0 UCYN-A1_1-80um Rhizaria 4f2711ffbb0611ea359dd96680df1c4f UCYN-A2 Rhizaria afacfa595bc8d424f2a648990538e0d4 UCYN-A2 Stramenopile eac29d65650c1a316e01b9fa8a5a9038 UCYN-A1_0.22-1um Stramenopile 25e0c3351c5b93e80e0f02e6ba23c077 UCYN-A1_1-80um Stramenopile 3465aaadf57e9cf04c930da3beb3deef UCYN-A1_1-80um Stramenopile 5ca13089ea80484bd62871929d00bf95 UCYN-A1_1-80um Stramenopile 6f39e2bbe0e4c6fa1e2eb082348de7c5 UCYN-A1_1-80um Stramenopile 8940d219ac059e1421b8910843b4fd82 UCYN-A1_1-80um Stramenopile 9ec148397bce05c40a8d97a91574a1ef UCYN-A1_1-80um Stramenopile ddaade5c44a6bbd3a74f879c85b5faf3 UCYN-A1_1-80um Stramenopile e67e2f1a5016412cf9826b35523352c4 UCYN-A1_1-80um Stramenopile e966740cb469db7493a9b384262bce67 UCYN-A1_1-80um Stramenopile a94291b249e004b97f399d41c5cc4b82 UCYN-A1_1-80um, UCYN-A1_0.22-1um Stramenopile 5d86bdd765a6ce520791f28dda5819ed UCYN-A1_1-80um, UCYN-A2 Stramenopile 3d03bbd7eb3a06f19d2f377b5eb2efb5 UCYN-A2 Stramenopile e09f07091bd33255025bd0a1cec414ca UCYN-A2 Syndiniales 00399bcfe64d51dac93b62a40832b514 UCYN-A1_0.22-1um Syndiniales 10657154c17634d9006a00243a3736b1 UCYN-A1_0.22-1um Syndiniales 38caa165f588a30638f74c2edb93662c UCYN-A1_0.22-1um Syndiniales 69e6b69c7b0addcc5c9c44d08015e45b UCYN-A1_0.22-1um Syndiniales bce9eb5a6281547d9fffcd95d9156913 UCYN-A1_0.22-1um Syndiniales 1082ec476ee33ebf0800a137e9dd6730 UCYN-A1_1-80um Syndiniales 12516469ae16dbe47afd1383f1a38d29 UCYN-A1_1-80um Syndiniales 43d72e28390647a43a479322275ebabc UCYN-A1_1-80um Syndiniales 4f0a7afd789e6f14b745c17acad17068 UCYN-A1_1-80um Syndiniales 6fa9b449505f9b00f18edbedca877d17 UCYN-A1_1-80um Syndiniales 80d903e0cb7271e69b2ec8bd1f87f45c UCYN-A1_1-80um Syndiniales b22089acf5e02fe3477be0ae8fe10cb5 UCYN-A1_1-80um Syndiniales bf238b09c39ad60be96eb93f30ac6915 UCYN-A1_1-80um Syndiniales eefc3152825b60051b3f78c504aca2a9 UCYN-A1_1-80um Syndiniales f14396a0e7efc97700da47bcb853bfad UCYN-A1_1-80um Syndiniales 365026beab40dc3dacd12678ac56cfc2 UCYN-A1_1-80um, UCYN-A1_0.22-1um Syndiniales 6bbbadb60359057c81abc84e7d0b797a UCYN-A1_1-80um, UCYN-A1_0.22-1um Syndiniales b3a109d089b15d7d926497af27f7fd2c UCYN-A1_1-80um, UCYN-A1_0.22-1um Syndiniales 624af6018837349e853bc2e772461e80 UCYN-A1_1-80um, UCYN-A1_0.22-1um, UCYN-A2 Syndiniales 68216c70c4cc7ae340a7f17cb6973f5b UCYN-A1_1-80um, UCYN-A1_0.22-1um, UCYN-A2 Syndiniales 6a30306b1212b48152ee097506fedad2 UCYN-A2 Syndiniales 6c3205e48cb5c1cc405609e73063607d UCYN-A2 Syndiniales 6cf6cfba7e7b46309c2fc38b44a28064 UCYN-A2 138 139 B) Ch. 4 Supplementary Figure 5–Relative abundances of UCYN-A 16S sequences linearly correlate with relative abundances of host 18S sequences. Pairwise comparison of all UCYN-A vs. all Braarudospharea ASVs in the SPOT dataset. B) UCYN-A1 (UCYNA_1) and UCYN-A2 (UCYNA_5) relative abundances correlate with one another, hence these symbionts appear to correlate with one another’s host organisms. Relative abundances are given out of the whole microbial community (16S + 18S) in the 1-80µm size fraction at 5m depth. See Supplementary Table 2 for hashes associated with UCYN-A and Braarudosphaera ASVs. 140 Ch. 4 Supplementary Figure 6 A) UCYN-A2 co-occurs with its most common host at the DCM across the SPOT time series. B) UCYN-A2 relative abundance correlates with its putative host relative abundance at the DCM. C) The ratio of 16S: 18S genes of these organisms is, on average, as expected. Boxplot values indicate the median and IQR values of this ratio; the red line indicates the average (0.703). 141 Ch. 4 Supplementary Figure 7 A) UCYN-A1 co-occurs with its putative host at the DCM across the SPOT time series. Panel inset indicates the relative abundance of UCYN-A1 in the smaller size fraction. B) UCYN-A1 relative abundance correlates with its putative host relative abundance. C) The ratio of 16S: 18S genes of these organisms is, on average, as expected. Boxplot values indicate the median and IQR values of this ratio; the red line indicates the average (1.674). 142 143 Ch. 4 Supplementary Figure 8–Relative abundance of UCYN-A1 (UCYN-A SPOT ASV1) (A), UCYN-A SPOT ASV6 (B), and UCYN-A2 (UCYN-A SPOT ASV5) (C) at the SPOT surface (black) and DCM (green) over time. Relative abundances are given out of the 16S community in the larger size fraction (AE filters; 1- 80µm). 144 Ch. 4 Supplementary Figure 9–Relative abundance of UCYN-A ASVs in the larger size fraction over depth in A) July 2008, the date UCYN-A1 reached its maximum relative abundance, and B) July 2009, the date UCYN-A1 appeared at 890m. 145 Ch. 4 Supplementary Figure 10–UCYN-A ASV6 does not co-occur with any Braarudosphaera ASV across the SPOT dataset at 5m. See Supplementary Table 2 for hashes associated with UCYN-A and Braarudosphaera ASVs. 146 Ch. 4 Supplementary Figure 11–Mixed mock communities contain even proportions of 16S (right) and 18S (middle) sequences. DNA from the small subunit of the rRNA gene of 21 organisms was pooled and sequenced in equal concentrations on an Illumina HiSeq or MiSeq. Normalized mixed mock communities match the proportions of sequenced DNA, indicating normalization was successful (left). 147 Chapter 5–Conclusions This dissertation provides context to better understand marine microbial communities in the Southern California Bight. In the three previous chapters, I contextualized the San Pedro Ocean Time-series (SPOT) dataset by sampling the surrounding community on small spatial scales, examined diazotroph activity in the context of the whole community, and investigated the symbiotic diazotroph UCYN-A over a decades’ worth of samples collected at SPOT. My first main body chapter provided perspective on the San Pedro Ocean Time-series (SPOT) by examining the surrounding waters and daily fluctuations in the community, which our monthly sampling regime might miss. I compared and contrasted the effects of space and time on two size fractions of marine microbes using tag sequencing of the 16S/ 18S rRNA gene. I showed that at scales of several kilometers and days, SPOT is generally representative of the surrounding waters of the Southern California Bight. At the scales we sampled, space and time had similar effects on variations in community composition. Additionally, over both space and time, the larger size fraction, which consists of larger and particle-attached organisms, was patchy over a much more homogeneous background of smaller, free-living microbes. Finally, changes in abiotic parameters–not geographic distance between the sites– more strongly influenced the larger size fraction than the smaller. Of all the metabolic groups observed in our transect, heterotrophic bacteria showed the strongest correlation with an aggregation of the environmental parameters sampled, as well as with individual measurements. We inferred that our sampling transect contained multiple patches of water, probably moving south (based on regional hydrography), which contained slightly different abiotic environments and microbial communities, particularly particle-attached heterotrophic bacteria. However, these differences were not enough to substantially differentiate SPOT from the surrounding waters. 148 I then investigated nifH expression by diazotrophs within the context of the larger community, using metagenomes as a reference point to compare expression of different genes and of nifH by different diazotrophs. Samples were collected autonomously via Environmental Sample Processor (ESP) and sequenced into metagenomes and metatranscriptomes. A custommade BLAST database of >8,000 nifH ASVs was used to identify which ASVs were active and abundant across the time-series. When compared to their relative abundances in the metagenomes, nifH was more highly expressed than a cyanobacterial photosystem II protein that is orders of magnitude more abundant in the community and known to be rapidly replaced (Mulo et al., 2009). Even though metagenomes and metatranscriptomes were sequenced to a high depth coverage and the nifH database I used contained >7,900 ASVs that did not belong to UCYN-A, I found very little evidence for nitrogen fixation by organisms other than UCYN-A. 98% of nifH transcripts belonged to UCYN-A, contrary to our expectations. Furthermore, UCYN-A2 was disproportionately active in the metatranscriptomes compared to its low relative abundance in the metagenomes. Of all the UCYN-A mRNA genes, nifH was the most highly expressed, which highlights the paramount significance of nitrogen fixation to this organism. In my last main body chapter, I used the SPOT dataset to contextualize UCYN-A relative abundance within inter-annual climate oscillations and within the microbial community over a decade’s worth of samples. UCYN-A symbioses were present more often in years following the 2015-2016 El Niño event, and their relative abundances correlated with warmer waters, lower bacterial biomass, and ocean currents flowing to the north and east of our study site, consistent with El Niño conditions. Further, UCYN-A1 presence and relative abundances were negatively correlated with upwelling conditions, which were both directly measured by upwelling indices, and indirectly indicated by lower temperatures, higher nutrient concentrations, and increased 149 bacterial production and chlorophyll concentrations. The same trends were also true of UCYNA2, but less statistically significant. We hypothesized that currents advected UCYN-A symbioses into our study site from more southern and western areas including the fringes of the central Pacific Gyre, and may have increased “seed populations” of the UCYN-A symbioses already there. Next, I showed that UCYN-A co-occurs with an ASV from the dinoflagellate genus Lepidodinium, a hypothesized predator of diazotrophs (Deng et al., 2020; Dugenne et al., 2020), and with ASVs from the group Syndiniales, which have been shown to parasitize dinoflagellates (e.g. Jephcott et al., 2016). Lastly, I showed that UCYN-A1 has a tighter partnership with its haptophyte host than did UCYN-A2. In particular, on two days, UCYN-A2 was present when the established host was absent. On those days, and on no other days, an 18S ASV closely related to the host peaked in relative abundance. This suggests that one ASV of UCYN-A2 may have two closely related hosts. This observation of a rare host-switching event was only made possible by our long-term time-series of prokaryotic and eukaryotic microorganisms from SPOT, finely resolved into ASVs. It is particularly important to contextualize marine microbial communities of the Southern California Bight in light of climate change. We anticipate that human-caused global warming will cause a suite of changes in coastal upwelling zones such as in Southern California (Capone and Hutchins, 2013), which could influence our long-term San Pedro Ocean Timeseries. For example, climate change may intensify the California current and resulting eddies and fronts (e.g., Martiny et al., 2006; McGillicuddy et al., 2007; Benavides et al., 2021; Hörstmann et al., 2022), which have been shown to shape microbial communities at the same sampling scales we used. Thus, the microbial community surrounding SPOT may become more “patchy,” such that 10 years from now, we would observe greater variation in samples collected over the 150 transect that we sampled in my first main body chapter. Some models predict that climate change will intensify upwelling (Bakun, 1990). This in turn could inhibit proliferation of UCYN-A in the region, particularly of UCYN-A1, as my fourth chapter shows UCYN-A is negatively correlated with upwelling conditions at SPOT. Other models, however, predict that climate change will intensify the strength and duration of El Niño events, which has been shown to mitigate seasonal upwelling (Jacox et al., 2015). These conditions could be conducive to UCYN-A1 in the future. Given that UCYN-A comprises >98% of the nitrogen fixing organisms in the region, as shown in Chapter 3, climate change could have a strong influence on fixed nitrogen inputs in the Southern California bight, Notably, human-induced efforts to mitigate climate change through offshore wind farms are also predicted to weaken upwelling (Raghakumar et al., 2023). Studies are underway to determine if these potential changes could influence the phytoplankton community of Southern California; UCYN-A1 relative abundances could be a useful marker of these changes. Whether these hypothesized effects of climate change and offshore wind do occur, this dissertation provides important perspective on marine microbial communities in the Southern California Bight. Here, I contextualize the San Pedro Ocean Time-series dataset through spatial and temporal sampling at fine scales, I show that diazotrophs are a rare but active component of the community, and that UCYN-A, the most abundant diazotroph in the region, is interconnected with a variety of 18S organisms and is influenced by abiotic conditions susceptible to climate change. Although few other studies address the knowledge gaps discussed here, diazotroph autecology cannot be fully understood without considering the wider community with which they co-exist. 151 References Bakun, A. 1990. Global Climate Change and Intensification of Coastal Ocean Upwelling. Science 247: 198–201. doi:http://dx.doi.org/10.1126/science.247.4939.198 Benavides, M., L. Conradt, S. Bonnet, I. Berman-Frank, A. Petrenko, and A. M. Doglioli. 2021. Fine scale sampling unveils diazotroph patchiness in the South Pacific Ocean. ISME Communications 1: 1-3. doi:.1038/s43705-021-00006-2 Capone, D. G., and D. A. Hutchins. 2013. Microbial biogeochemistry of coastal upwelling regimes in a changing ocean. Nature Geosci 6: 711–717. doi:10.1038/ngeo1916 Deng, L., S. Cheung, and H. Liu. 2020. Protistal Grazers Increase Grazing on Unicellular Cyanobacteria Diazotroph at Night. Frontiers in Marine Science 7. doi:10.3389/fmars.2020.00135 Dugenne, M., F. H. Freitas, S. T. Wilson, D. M. Karl, and A. E. White. 2020. Life and death of Crocosphaera sp. in the Pacific Ocean: Fine scale predator–prey dynamics. Limnology and Oceanography 65: 2603–2617. doi:10.1002/lno.11473 Hörstmann, C., P. L. Buttigieg, U. John, E. J. Raes, D. Wolf‐Gladrow, A. Bracher, and A. M. Waite. 2022. Microbial diversity through an oceanographic lens: refining the concept of ocean provinces through trophic‐level analysis and productivity‐specific length scales. Environmental Microbiology 24: 404–419. doi:10.1111/1462-2920.15832 Jacox, M. G., J. Fiechter, A. M. Moore, and C. A. Edwards. 2015. ENSO and the California Current coastal upwelling response. J. Geophys. Res. Oceans 120: 1691–1702. doi:10.1002/2014JC010650 Jephcott, T. G., C. Alves-de-Souza, F. H. Gleason, F. F. van Ogtrop, T. Sime-Ngando, S. A. Karpov, and L. Guillou. 2016. Ecological impacts of parasitic chytrids, syndiniales and perkinsids on populations of marine photosynthetic dinoflagellates. Fungal Ecology 19: 47–58. doi:10.1016/j.funeco.2015.03.007 Martiny, J. B. H., B. J. M. Bohannan, J. H. Brown, and others. 2006. Microbial biogeography: putting microorganisms on the map. Nat Rev Microbiol 4: 102–112. doi:10.1038/nrmicro1341 McGillicuddy, D. J., L. A. Anderson, N. R. Bates, and others. 2007. Eddy/Wind Interactions Stimulate Extraordinary Mid-Ocean Plankton Blooms. Science 316: 1021–1026. doi:10.1126/science.1136256 Mulo, P., C. Sicora, and E.-M. Aro. 2009. Cyanobacterial psbA gene family: optimization of oxygenic photosynthesis. Cell. Mol. Life Sci. 66: 3697–3710. doi:10.1007/s00018-009- 0103-6 Raghukumar, K., T. Nelson, M. Jacox, C. Chartrand, J. Fiechter, G. Chang, L. Cheung, and J. 152 Roberts. 2023. Projected cross-shore changes in upwelling induced by offshore wind farm development along the California coast. Commun Earth Environ 4: 116. doi:10.1038/s43247-023-00780-y

Abstract (if available)

Abstract Marine microbes cycle nutrients on a massive scale. For example, nitrogen is fixed for the community by rare prokaryotes known as diazotrophs. Microbes are frequently studied in ways that miss fine-scale dynamics. Environmental studies of diazotrophs specifically sample the nifH gene, a component of the enzyme responsible for nitrogen fixation. I strive to contextualize marine microbial communities in the Southern California Bight. First, I compare the small-scale effects of space and time on two size fractions of microbes. I show that at scales of kilometers and days, space and time have similar effects on destabilizing the community, that the larger size fraction is patchy over a homogeneous background of smaller microbes, and that changes in environmental parameters more strongly influence the larger size fraction. I then investigate diazotroph diversity and activity. I contextualize nifH expression, showing it is transcribed at higher rates than a protein that is orders of magnitude more abundant and highly expressed. Over 98% of nifH transcripts were from the symbiotic diazotroph UCYN-A; UCYN-A2 is particularly active. Lastly, I contextualize UCYN-A relative abundance within the San Pedro Ocean Time-series (SPOT). UCYN-A1, the “open ocean” ecotype, is brought to California by tropical waters that are stronger during El Niño and weaker during seasonal upwelling. Both UCYN-A1 and UCYN-A2 co-occur with Lepidodinium, a hypothesized predator. In contrast to UCYN-A1, UCYN-A2 has a weaker relationship with its host. Studying the wider context of microbes in the Southern California Bight is particularly important as climate change alters their dynamics.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Marine bacterioplankton biogeography over short to medium spatio-temporal scales

PDF

Spatial and temporal investigations of protistan grazing impact on microbial communities in marine ecosystems

PDF

Characterizing protistan diversity and quantifying protistan grazing in the North Pacific Subtropical Gyre

PDF

Investigations of animal-sediment-microbe interactions at two different environments: coastal lagoons and methane seeps

PDF

Plant-microbial interactions in mangrove sediments under different nutrient conditions

PDF

Biological nitrogen fixation associated with living and decomposing macroalgae

PDF

Microbe to microbe: monthly microbial community dynamics and interactions at the San Pedro Ocean Time-series

PDF

Ecological patterns of free-living and particle-associated prokaryotes, protists, and viruses at the San Pedro Ocean Time-series between 2005 and 2018

PDF

Patterns of molecular microbial activity across time and biomes

PDF

Dynamics of marine bacterial communities from surface to bottom and the factors controling them

PDF

Molecular ecology of marine cyanobacteria: microbial assemblages as units of natural selection

PDF

Temporal variability of marine archaea across the water column at SPOT

PDF

Changes in the community composition of marine microbial eukaryotes across multiple temporal scales of measurement

PDF

Iron-dependent response mechanisms of the nitrogen-fixing cyanobacterium Crocosphaera to climate change

PDF

Marine protistan diversity, spatiotemporal dynamics, and physiological responses to environmental cues

PDF

The development of novel measures of landscape diversity in assessing the biotic integrity of lotic communities

PDF

Enhancing recovery of understudied and uncultured lineages from metagenomes

PDF

Genetic characterization of microbial eukaryotic diversity and metabolic potential

PDF

The molecular adaptation of Trichodesmium to long-term CO₂-selection under multiple nutrient limitation regimes

PDF

Ecological implications of daily-to-weekly dynamics of marine bacteria, archaea, viruses, and phytoplankton

Asset Metadata

Creator Fletcher-Hoppe, Colette (author)

Core Title Spatial and temporal dynamics of marine microbial communities and their diazotrophs in the Southern California Bight

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Biology (Marine Biology and Biological Oceanography)

Degree Conferral Date 2024-05

Publication Date 06/05/2024

Defense Date 06/04/2024

Publisher Los Angeles, California (original), University of Southern California (original), University of Southern California. Libraries (digital)

Tag diazotrophy,environmental DNA,environmental microbiology,marine microbial ecology,microbes,nifH,nitrogen fixation,OAI-PMH Harvest,spatial transect,symbiosis,time-series,transcription,UCYN-A

Format theses (aat)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Fuhrman, Jed (committee chair), Capone, Douglas (committee member), Hutchins, David (committee member), Sun, Fengzhu (committee member), Webb, Eric Alva (committee member)

Creator Email c.fletcher.hoppe@gmail.com,colettef@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC113987390

Unique identifier UC113987390

Identifier etd-FletcherHo-13063.pdf (filename)

Legacy Identifier etd-FletcherHo-13063

Document Type Dissertation

Format theses (aat)

Rights Fletcher-Hoppe, Colette

Internet Media Type application/pdf

Type texts

Source 20240605-usctheses-batch-1165 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu