University of Southern California Dissertations and Theses

Stage-specific transcriptomic analyses of reproductive tissue in the Pacific oyster, Crassostrea gigas

(USC Thesis Other)

Stage-specific transcriptomic analyses of reproductive tissue in the Pacific oyster, Crassostrea gigas

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content STAGE-SPECIFIC TRANSCRIPTOMIC ANALYSES OF REPRODUCTIVE TISSUE
IN THE PACIFIC OYSTER, CRASSOSTREA GIGAS

by
Austin Paul

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
(BIOLOGY)

May 2014

Copyright 2014 Austin Paul

i

Table of Contents

Chapter 1: Introduction 1
References 10

Chapter 2: RNA sequencing-based expression differences between sexes and among
stages of oogenesis in the Pacific oyster Crassostrea gigas 16
Introduction 17
Materials and Methods 21
Results 25
Discussion 29
References 38
Tables 48
Figures 52

Chapter 3: Quantification of total and poly(A) RNA in selected biomarkers during
gonad maturation in the Pacific oyster Crassostrea gigas 58
Introduction 59
Materials and Methods 61
Results 64
Discussion 66
References 71
Tables 75
Figures 76

Chapter 4: Selection of reference genes for normalization of qPCR data from multiple stages of
gonad maturation in the Pacific oyster Crassostrea gigas 79
Introduction 80
Materials and Methods 83
Results 86
Discussion 88
References 92
Tables 96
Figures 97

Chapter 5: Expression of candidate biomarkers exhibit stage-specific expression during
oogenesis in the Pacific oyster Crassostrea gigas 102
Introduction 103
Materials and Methods 106
Results 109
ii

Discussion 112
References 118
Tables 123
Figures 126

Chapter 6: Conclusion 131
References 137

1

Chapter 1: Introduction

Reproduction is one of the most important physiological process in the life cycle of the
Pacific oyster Crassostrea gigas and is tightly coupled with the processes of energy storage and
utilization. Key to these energetic processes in marine bivalves is glycogen and the enzymatic
pathways responsible for the catabolism and anabolism of glucose (Bayne et al. 1982, Ruiz et al.
1992, Mathieu and Lubet 1993). In temperate regions, the annual cycle of reproduction for the
C. gigas begins in winter with the mobilization of glycogen reserves for initiation of
gametogenesis. Spring brings the active phase of gametogenesis, when the maximal growth of
the gonad is observed. The completion of gonad maturation and spawning occur during the
summer. Fall is considered the resting phase of the cycle, and includes resorbtion of remaining
gametocytes and accumulation of glycogen reserves.
-C. gigas is a sequential protrandric hermaphrodite that can reach sexual maturity after a
couple of months but typically, depending on environmental conditions, will not reach maturity
until the second summer of life. Sex-ratios have been shown to be under genetic control with a
strong paternal effect thought to represent two distinct types of sires (Guo et al. 1998; but see
Haley 1977, Hedrick and Hedgecock 2010). In C. virginica, there is evidence that environmental
factors influence the time at which sex-ratios favor females (Needler 1932, Coe 1936). Better
growing conditions tend to favor female sex determination. To date, no specific sex-determining
genes or environmental sex-determining models have been conclusively applied to sex
determination in C. gigas (see Santerre et al. 2013). That females reach market size sooner, on
average, than males makes the issue of sex-determination one of great importance for the
commercial production of C. gigas.
2

Ambient water temperature and photoperiod, along with food availability, are considered the
most important environmental parameters controlling reproduction in C. gigas. In an artificial
setting, alteration of the above parameters can completely modify the timing of gametogenesis in
C. gigas (Lannan et al. 1980, Fabioux et al. 2005). Proliferation of gonial cells is activated by
cold temperatures of between 8 and 11 °C (Fabioux et al. 2005). Active gametogenesis in spring
is triggered by increased water temperatures, while spawning is thought to be controlled by a
combination of the water temperature reaching a threshold of >19 °C, and the presence of food
(Muranaka and Lannan 1984, Starr et al. 1990).
Variance in Reproductive Success
High fecundity is coupled with high early mortality in many marine organisms. In C. gigas,
the combination of high fecundity and high larval mortality creates the potential for a large
variance in reproductive success among individuals of reproductive age in natural populations
(Hedgecock 1992). Factors contributing to differences in reproductive success may either be pre
or post-zygotic in nature. Pre-zygotic effects could be due to “wrong place, wrong time” issues
and may also include gametic effects, such as gamete quality and interactions (Boudry et al.
2002). Post-zygotic contributions to variance in reproductive success include physical and
environmental factors such as water currents and predation may impact cohort success.
Differential viability among particular combinations of parents can also shape patterns of
reproductive success (Boudry et al. 2002).
Although differential viability through parental contributions may affect reproductive success
among families, its effect through expression of deleterious alleles likely impacts all families,
spawned both commercially and in nature. C. gigas possesses a large number of lethal recessive
mutations or genetic load (Launey and Hedgecock 2001). The impact of this load appears highly
significant. The expression of these deleterious mutations was recently tracked during larval
3

development, where it was estimated that 96% of all mortality in the sampled families was due to
genetic load (Plough and Hedgecock 2011).
Artificial Reproduction
C. gigas is an important commercial species that is farmed around the globe. Growers
almost exclusively purchase hatchery produced seed which is later grown out on beaches or
suspended in coastal waters. This is a departure from previous methods of placing artificial
substrates into coastal waters during spawning season to catch the natural set of larvae. Prior to
using artificial substrates to catch settling larvae, the native oyster of the Pacific Northwest coast,
Ostrea lurida, was simply harvested at unsustainable levels (White et al. 2009). And so C. gigas
was introduced, but the lack of established populations ultimately resulted in unpredictable
recruitment. The switch to using hatchery produced seed also met a demand from restaurants of
single oysters for serving on the half-shell. Despite the potential for great output in hatcheries
due to the high fecundities of C. gigas and other bivalve molluscs, production is often stymied by
inconsistently high episodes of larval mortality.
Larvae grown in hatcheries, in addition to running the gauntlet of the aforementioned post-
zygotic causes of mortality, are also subject to additional mortality caused by environmental
factors. Bacterial infections are a major concern in hatcheries. Vibrio sp. have been implicated
in the mass mortality of larvae at several shellfish hatcheries in the Pacific Northwest (Estes et
al. 2004, Elston et al. 2008). The ultimate cause of hatchery-based bacterial infections is unclear,
but increased sea surface temperatures along the Pacific coast in 2007 were correlated with
bacterial mortality that killed an estimated 60% of larvae at an important West coast hatchery
(Elston et al. 2008).
Water chemistry, specifically decreasing pH, has replaced bacterial contamination as the
primary concern among West coast hatcheries. Historically low larval production at the
4

Whiskey Creek shellfish hatchery on the Oregon coast in 2009 was significantly correlated with
aragonite saturation (Barton et al. 2012). The impact of pH on shellfish hatcheries is likely going
to get worse as our oceans continue to absorb CO
2
from the atmosphere at record pace. Potential
exists, however, for curbing the effects of both bacterial contamination and water chemistry by
treating water within hatcheries prior to use in cultivation.

Even prior to the current environmental concerns relating to larval output, considerable
variance in reproductive success existed in hatchery spawns. Success of larval survival spawns
has been observed to follow a bimodal distribution with some spawns that fail entirely and others
that are moderately successful (Lannan 1980). In a report on broodstock management and larval
survival in hatchery populations of C. gigas, Lannan et al. (1980) demonstrated that larval
survival in a pair of factorial crosses was dependent largely on an interaction effect between
parents. In an effort to validate a genetic component in the interaction variance, which was
hypothesized to be due to variation in state of gonad maturation between parents, Lannan
showed that there exists heritable genetic variation in the state of gonad maturation among
experimental lines or families of oysters.
These findings lead naturally to the hypothesis that variation in state of gonad maturation
among adult oysters contributes to high larval mortality. This variance is germane to selection
programs for establishing seasonal hatchery lines and, more fundamentally, is the material, upon
which selection will act in natural populations, as global climate change alters water currents,
chemistry, and temperatures. Later, Lannan showed that an optimal timing of gonad maturation
exists for parents in a similar state of maturation (Lannan et al. 1980), but never proceeded to
demonstrate his expectation that variance in state among parents affects larval survival.
5

Differences in patterns of gonad maturation between families of oysters are created and/or
maintained by causes likely ranging greatly in time and space. Most wild populations in the
Pacific Northwest were transplanted in the last 100 years and the evolutionary histories both
within and between each population are unknown. A transplantation study in the oyster, C.
virginica, found evidence of a genetic effect on the timing of gametogenesis (Barber et al. 1991).
A population from Long Island Sound was transplanted to Delaware Bay, where it was reared
and inbred for six generations. The timing of gametogenesis remained the same as in its native
habitat but was one month prior to the timing of gametogensis in oysters native to Delaware Bay,
which were maintained in the same conditions.
Selection for genetic variance in gonad maturation was made indirectly while selecting for
oysters resistant (R) and susceptible (S) to summer mortality syndrome. In a paper investigating
mechanistic differences between R and S oysters, it was noticed there were differences in their
maturation cycles (Samain et al. 2007). Resistant oysters displayed a lower reproductive effort,
apparently due to a lower germinal investment. Patterns of spawning also differed between the
two groups, with R oysters spawning just once to completion, while S oysters spawned
successively, later into the season.
Additional support of a genetic effect in a different bivalve mollusc comes from the
correlation in the mussel, Mytilus edulis, between the presence or absence of an allele at the Lap
gene, and timing of the initiation of gametogenesis (Hilbish and Zimmerman 1988). The timing
of gametogenesis was correlated with the presence of an allele of the Lap gene, which increased
nitrogen secretion. Consequently, a reduced nitrogen budget resulted for individuals with the
presence of the allele, which delayed initiation of gametogenesis by approximately six weeks.
6

The pervasive role of the environment on the reproductive cycle of the Pacific oyster C.
gigas is easily demonstrated through artificial conditioning of gonad maturation. A recent study
investigated the gametogenic cycle in C. gigas over a full year under three different conditions of
temperature and photoperiod (Fabioux et al. 2005). The results are applicable to year-round
hatchery based production of larvae (most hatcheries produce larvae only half of the year), but
the fundamentals are already common knowledge in commercial hatcheries; warmer water
temperatures accelerate development of the gonad, and vice versa. Hatcheries have long used
elevated water temperatures to condition broodstock for spawning in order to reduce the amount
of time needed to produce mature gametes.
Experimental Methods for Assessing Reproductive Maturation
The gonad in C. gigas is the largest organ at maturation, yet is scarcely visible early in the
winter prior to the initiation of gametogenesis (Galtsoff 1964). Two separate systems of genital
canals wrap around the adductor muscle and fuse on the dorsal side, reaching towards the umbo,
where the organ fully envelopes the digestive tract. At spawning, gametes are moved through
the canals and released through the gonoduct by ciliary motion. The entire soft body of the
oyster is located within two calcium carbonate valves, which make determination of sex and
state of gonad maturation difficult without sacrificing the animals.
Reproductive status is most often judged through microscopic analysis of histological
sections. In this method, the oyster is sacrificed, and a transverse cut is made at a fixed place
(generally at the labial palp) since the gonad is not uniformly thick throughout the length of the
oyster. The section is then fixed in a preservative, sliced thinly, stained, and viewed under a
microscope. Determination of state of gonad maturation is then measured based on cell
composition. Alternatively, the proportion of the surface area of a cross section of the visceral
mass occupied by gonad can be used as a quantitative measure of gonad maturation. This
7

method, though it does not require preserving, staining, slicing, and counting, has its
shortcoming. Namely, the latter stages of maturation are hard to distinguish as the thickness of
the gonad reaches a maximum before maturation is complete. Other methods exist as well,
though most are destructive, which makes the phenotype difficult to measure when asking
questions about the relative roles of the environment and the genotype on maturation.
One nondestructive method currently being developed for assessing state of gonad
maturation at the French research institute IFREMER utilizes magnetic resonance imaging
(MRI) (Davenel et al. 2006, Pouvreau et al. 2006). Using this method, gonad volume measured
by MRI was highly correlated with gonad surface area from histological slides (Flahauw et al.
2012). The advantage of this method, in addition to being nondestructive, is the speed in which
results can be generated and also the relative noninvasiveness. However, the method is
expensive and reproductive maturity does not necessarily vary linearly with gonad volume. It is
more likely that the rate of increase of gonad decreases as maturation increases.
Biological sampling is a more common means of attempting to characterize reproductive
state. In bivalves, the presence of the shell makes tissue sampling a more difficult process.
Oysters may be sampled after being immersed in a heavy salt bath for multiple hours, which
relaxes the adductor muscle and allows direct access to the visceral mass within the shell. The
shell may also be either notched where the valves come together or a hole may be drilled
adjacent the sought after tissue. While not destructive, these approaches are invasive to varying
degrees.
Vitellogenin, the main precursor of vitellin, is a common reproductive target due to it being
the primary yolk protein. Vitellin protein levels were investigated as a potential marker for
reproductive stage in C. corteziensis (Arcos et al. 2009). While vitellin levels found by ELISA
8

were correlated with reproductive stage, due to the amount of input required, the method requires
destructive sampling. Vitellogenin gene expression levels from hemolymph in the shrimp
Penaeus (Litopenaeus) vannamei were shown to increase along with gonad maturation (Arcos et
al. 2011). In P. vannamei and other shrimp, vitellogenin is synthesized both in the ovary and in
the hepatopancreas (Arcos et al. 2011). Heterosynthesis of vitellogenin likely increases the
expression in hemolymph, in which it is transported to the ovary. Where it has been assessed in
mollusks, including bivalves, autosynthesis of vitellogenin in oocytes and surrounding follicle
cells appears to be the only mode of synthesis (Matsumoto et al. 2003, Matsumoto et al. 2008,
Agnese et al. 2012). This pattern likely precludes the use of vitellogenin expression in
hemolymph as an effective indicator of reproductive stage in C. gigas.
Multiple genes have been characterized in C. gigas which exhibit stage-specific expression in
gonad tissue during maturation (Fabioux et al. 2004, Rodet et al. 2005, Fleury et al. 2008,
Meistertzheim et al. 2009, Naimi et al. 2009a,b). These results suggest that gene expression
changes are sufficient for monitoring the progress of gonad maturation, and that commonly used
methods for measuring gene expression, such as qPCR, are sensitive enough to make expression
of maturation specific genes potential indicators of reproductive state. Current technologies for
measuring global changes in gene expression, such as microarrays and RNA-Seq, should assist
in identifying additional genes with stage-specific gene expression. An application of the
microarray method was recently reported in C. gigas, where a combined 2,482 genes were
identified as differentially expressed over a maturation cycle for males and females (Dheilly et
al. 2012). The recently published genome of C. gigas (Zhang et al. 2012) further opens more
sophisticated possibilities of looking at regulation of gene expression such as alternative splicing,
methylation, miRNA and differential polyadenylation profiles.
9

Thesis Objectives
The goal for the work presented here was to gain a greater understanding of processes
involved in stage-specific gonad maturation in female C. gigas, with a potential application of
identifying genes useful as indicators of maturation state. Global expression was measured at
three stages of gonad maturation by RNA-Seq (Chapter 2). The potential impact of
polyadenylation on results of stage-specific expression was also considered (Chapter 3). A
subset of differentially expressed genes was then analyzed by qPCR for predictive ability in
assigning length of artificial gonad conditioning (Chapter 5). An analysis of the stability of
potential housekeeping genes was conducted beforehand to determine their suitability for
normalizing gene expression in developing gonad tissue (Chapter 4).

10

References
Agnese, M., Verderame, M., De Meo, E., Prisco, M., Rosati, L., Limatola, E., del Gaudio, R.,
Aceto, S., Andreuccetti, P., 2013. A network system for vitellogenin synthesis in the mussel
Mytilus galloprovincialis (L.). J. Cell. Physiol., 228, 547-555.
Arcos, F.G., Ibarra, A.M., Racotta, L.S., 2011. Vitellogenin in hemolymph predicts gonad
maturity in adult female Litopenaeus (penaeus) vannamei shrimp. Aquaculture, 316, 93-98.
Arcos, F.G., Maria Ibarra, A., del Carmen Rodriguez-Jaramillo, M., Awilda Garcia-Latorre, E.,
Vazquez-Boucard, C., 2009. Quantification of vitellin/vitellogenin-like proteins in the
oyster Crassostrea corteziensis (hertlein 1951) as a tool to predict the degree of gonad
maturity. Aquacult. Res., 40, 644-655.
Barber, B.J., Ford, S.E., Wargo, R.N., 1991. Genetic variation in the timing of gonadal
maturation and spawning of the Eastern oyster, Crassostrea virginica (gmelin). Biol. Bull.,
181, 216-221.
Barton, A., Hales, B., Waldbusser, G.G., Langdon, C., Feely, R.A., 2012. The Pacific oyster,
Crassostrea gigas, shows negative correlation to naturally elevated carbon dioxide levels:
Implications for near-term ocean acidification effects. Limnol. Oceanogr., 57, 698-710.
Bayne, B.L., Bubel, A., Gabbott, P.A., Livingstone, D.R., Lowe, D.M., Moore, M.N., 1982.
Glycogen utilisation and gametogenesis in Mytilus edulis. Mar. Biol. Lett., 3, 89-105.
Boudry, P., Collet, B., Cornette, F., Hervouet, V., Bonhomme, F., 2002. High variance in
reproductive success of the Pacific oyster (Crassostrea gigas, thunberg) revealed by
microsatellite-based parentage analysis of multifactorial crosses. Aquaculture, 204, 283-296.
Coe, W., 1936. Environment and sex in the oviparous oyster Ostrea virginica. Biol. Bull., 71,
353-359.
11

Davenel, A., Quellec, S., Pouvreau, S., 2006. Noninvasive characterization of gonad maturation
and determination of the sex of Pacific oysters by MRI. Magn. Reson. Imaging, 24, 1103-
1110.
Dheilly, N.M., Lelong, C., Huvet, A., Kellner, K., Dubos, M., Riviere, G., Boudry, P., Favrel, P.,
2012. Gametogenesis in the Pacific oyster Crassostrea gigas: A microarrays-based analysis
identifies sex and stage specific genes. Plos One, 7, e36353.
Elston, R.A., Hasegawa, H., Humphrey, K.L., Polyak, I.K., Hase, C.C., 2008. Re-emergence of
Vibrio tubiashii in bivalve shellfish aquaculture: Severity, environmental drivers,
geographic extent and management. Dis. Aquat. Organ., 82, 119-134.
Estes, R., Friedman, C., Elston, R., Herwig, R., 2004. Pathogenicity testing of shellfish hatchery
bacterial isolates on Pacific oyster Crassostrea gigas larvae. Dis. Aquat. Organ., 58, 223-
230.
Fabioux, C., Pouvreau, S., Roux, F.L., Huvet, A., 2004. The oyster vasa-like gene: A specific
marker of the germline in Crassostrea gigas, Biochem. Bioph. Res. Co., 315, 897-904.
Fabioux, C., Huvet, A., Le Souchu, P., Le Pennec, M., Pouvreau, S., 2005. Temperature and
photoperiod drive Crassostrea gigas reproductive internal clock. Aquaculture, 250, 458-
470.
Flahauw, E., Quellec, S., Davenel, A., Degremont, L., Lapegue, S., Hatt, P., 2012. Gonad
volume assessment in the oyster Crassostrea gigas: Comparison between a histological
method and a magnetic resonance imaging (MRI) method. Aquaculture, 370, 84-89.
Fleury, E., Fabioux, C., Lelong, C., Favrel, P., Huvet, A., 2008. Characterization of a gonad-
specific transforming growth factor-beta superfamily member differentially expressed
during the reproductive cycle of the oyster Crassostrea gigas. Gene, 410, 187-196.
12

Galtsoff, P.F., 1964. The American Oyster, Crassostrea virginica Gmelin. U.S. Fish and
Wildlife Service, 1-480.
Guo, X., Hedgecock, D., Hershberger, W.K., Cooper, K., Allen Jr., S.K., 1998. Genetic
determinants of protandric sex in the Pacific oyster, Crassostrea gigas thunberg. Evolution,
52, 394-399.
Haley, L., 1977. Sex determination in the American oyster. J. Hered., 68, 114-116.
Hedrick, P.W. and Hedgecock, D., 2010. Sex determination: Genetic models for oysters. J.
Hered., 101, 602-611.
Hedgecock, D., Chow, V., Waples, R., 1992. Effective population numbers of shellfish
broodstocks estimated from temporal variance in allelic frequencies. Aquaculture, 108, 215-
232.
Hilbish, T.J. and Zimmerman, K.M., 1988. Genetic and nutritional control of the gametogenic
cycle in Mytilus edulis. Mar. Biol., 98, 223-228.
Lannan, J.E., 1980. Broodstock management of Crassostrea gigas: I. genetic and environmental
variation in survival in the larval rearing system. Aquaculture, 21, 323-336.
Lannan, J.E., Robinson, A., Breese, W.P., 1980. Broodstock management of Crassostrea gigas:
II. broodstock conditioning to maximize larval survival. Aquaculture, 21, 337-345.
Launey, S. and Hedgecock, D., 2001. High genetic load in the Pacific oyster Crassostrea gigas.
Genetics, 159, 255-265.
Mathieu, M. and Lubet, P., 1993. Storage tissue metabolism and reproduction in marine
bivalves-a brief review. Invertebr. Reprod. Dev., 23, 123-129.
13

Matsumoto, T., Nakamura, A., Mori, K., Kayano, T., 2003. Molecular characterization of a
cDNA encoding putative vitellogenin from the Pacific oyster Crassostrea gigas. Zool. Sci.,
20, 37-42.
Matsumoto, T., Yamano, K., Kitamura, M., Hara, A., 2008. Ovarian follicle cells are the site of
vitellogenin synthesis in the Pacific abalone Haliotis discus hannai. Comp. Biochem. Phys.
A, 149, 293-298.
Meistertzheim, A.L., Lejart, M., Le Goïc, N., Thébault, M.T., 2009. Sex-, gametogenesis, and
tidal height-related differences in levels of HSP70 and metallothioneins in the Pacific oyster
Crassostrea gigas. Comp. Biochem. Phys. A, 152, 234-239.
Muranaka, M.S. and Lannan, J.E., 1984. Broodstock management of Crassostrea gigas:
Environmental influences on broodstock conditioning. Aquaculture, 39, 217-228.
Naimi, A., Martinez, A.S., Specq, M.L., Diss, B., Mathieu, M., Sourdaine, P., 2009. Molecular
cloning and gene expression of cg-Foxl2 during the development and the adult
gametogenetic cycle in the oyster Crassostrea gigas. Comp. Biochem. Phys. B, 154, 134-
142.
Naimi, A., Martinez, A.S., Specq, M.L., Mrac, A., Diss, B., Mathieu, M., Sourdaine, P., 2009.
Identification and expression of a factor of the DM family in the oyster crassostrea gigas.
Comp. Biochem. Phys. A, 152, 189-196.
Needler, A.B., 1932. Sex reversal in Ostrea virginica. Contrib. Can. Biol. Fish. 7, 285-294.
Plough, L.V. and Hedgecock, D., 2011. Quantitative trait locus analysis of stage-specific
inbreeding depression in the Pacific oyster Crassostrea gigas. Genetics 189, 1473-1486.
14

Pouvreau, S., Rambeau, M., Cochard, J.C., Robert, R., 2006. Investigation of marine bivalve
morphology by in vivo MR imaging: First anatomical results of a promising technique.
Aquaculture, 259, 415-423.
Rodet, F., Lelong, C., Dubos, M., Costil, K., Favrel, P., 2005. Molecular cloning of a molluscan
gonadotropin-releasing hormone receptor orthologue specifically expressed in the gonad.
BBA-Gene Struct. Expr., 1730, 187-195.
Ruiz, C., Abad, M., Sedano, F., Garcia-Martin, L.O., López, J.L.S., 1992. Influence of seasonal
environmental changes on the gamete production and biochemical composition of
Crassostrea gigas (thunberg) in suspended culture in El Grove, Galicia, Spain. J. Exp. Mar.
Biol. Ecol., 155, 249-262.
Samain, J.F., Dégremont, L., Soletchnik, P., Haure, J., Bédier, E., Ropert, M., Moal, J., Huvet,
A., Bacca, H., Van Wormhoudt, A., Delaporte, M., Costil, K., Pouvreau, S., Lambert, C.,
Boulo, V., Soudant, P., Nicolas, J.L., Le Roux, F., Renault, T., Gagnaire, B., Geret, F.,
Boutet, I., Burgeot, T., Boudry, P., 2007. Genetically based resistance to summer mortality
in the Pacific oyster (Crassostrea gigas) and its relationship with physiological,
immunological characteristics and infection processes. Aquaculture, 268, 227-243.
Santerre, C., Sourdaine, P., Marc, N., Mingant, C., Robert, R., Martinez, A., 2013. Oyster sex
determination is influenced by temperature - first clues in spat during first gonadic
differentiation and gametogenesis. Comp. Biochem. Phys. A, 165, 61-69.
Starr, M., Himmelman, J.H., Therriault, J., 1990. Direct coupling of marine invertebrate
spawning with phytoplankton blooms. Science 247, 1071-1074.
15

White, J., Ruesink, J.L., Trimble, A.C., 2009. The nearly forgotten oyster: Ostrea lurida
carpenter 1864 (olympia oyster) history and management in washington state. J. Shellfish
Res., 28, 43-49.
Zhang, G., Fang, X., Guo, X., Li, L., Luo, R., Xu, F., Yang, P., Zhang, L., Wang, X., Qi, H.,
Xiong, Z., Que, H., Xie, Y., Holland, P.W.H., Paps, J., Zhu, Y., Wu, F., Chen, Y., Wang, J.,
Peng, C., Meng, J., Yang, L., Liu, J., Wen, B., Zhang, N., Huang, Z., Zhu, Q., Feng, Y.,
Mount, A., Hedgecock, D., Xu, Z., Liu, Y., Domazet-Loso, T., Du, Y., Sun, X., Zhang, S.,
Liu, B., Cheng, P., Jiang, X., Li, J., Fan, D., Wang, W., Fu, W., Wang, T., Wang, B., Zhang,
J., Peng, Z., Li, Y., Li, N., Wang, J., Chen, M., He, Y., Tan, F., Song, X., Zheng, Q., Huang,
R., Yang, H., Du, X., Chen, L., Yang, M., Gaffney, P.M., Wang, S., Luo, L., She, Z., Ming,
Y., Huang, W., Zhang, S., Huang, B., Zhang, Y., Qu, T., Ni, P., Miao, G., Wang, J., Wang,
Q., Steinberg, C.E.W., Wang, H., Li, N., Qian, L., Zhang, G., Li, Y., Yang, H., Liu, X.,
Wang, J., Yin, Y., Wang, J., 2012. The oyster genome reveals stress adaptation and
complexity of shell formation. Nature, 490, 49-54.

16

Chapter 2: RNA sequencing-based expression differences between
sexes and among stages of oogenesis in the Pacific oyster Crassostrea
gigas

Abstract
Reproduction is one of the most energetically demanding process in many marine
bivalves, yet relatively little is known about the molecular biology of gonad maturation. Deep
sequencing of mRNA was used to compare and contrast sex and three distinct stages of
oogenesis in the Pacific oyster Crassostrea gigas. A total of 199 genes were identified as
differentially expressed (DEGs) between sexes, while 237 DEGs were found among the three
stages of oogenesis. The immature stage (stage I – proliferation of gonia) was characterized by
more genes being expressed at a higher level and thus a greater number of genes up-regulated
with respect to the other two stages (stages II & III). In the early-maturation (stage II) and late-
maturation gonad (stage III), oocytes constituted a substantial proportion of the sample, and the
mRNA pool seemed to be reduced. The majority of enriched biological processes among the
three stages were directly linked to reproduction in other organisms. The results add to our
understanding of the molecular biology of reproduction in C. gigas, while identified DEGs could
be useful in the development of sex and female stage-specific molecular biomarkers.

17

Introduction
The Pacific oyster Crassostrea gigas is an important food species with a global
distribution (Mann 1983). Production of seed for farming typically relies on manipulation of the
gametogenic cycle, by means of temperature and diet manipulations within a hatchery setting
(Breese and Malouf 1977). Despite the importance of broodstock conditioning, the molecular
biology of genes involved in the reproductive process of bivalve mollusks has only recently
begun to be investigated (Boutet et al. 2008, Ciocan et al. 2011, Dheilly et al. 2012, Ghiselli et
al. 2012).
Gonad maturation and reproduction is an annual process in C. gigas. Temperate
populations initiate gametogenesis in early spring and spawn towards the end of the summer.
Water temperature is considered the dominant factor affecting the initiation and rate of gonad
maturation (Muranaka and Lannan 1984, Fabioux et al. 2005). The gonad is transient but, when
ripe, is the largest organ in the oyster, comprising up to 40% the mass of an individual (Galtsoff
1964). It is a diffuse organ comprised of two lobes, located between the digestive gland on the
interior and the mantle on the exterior (Galtsoff 1964, Eckelbarger and Davis 1996). Resorption
of remaining reproductive material occurs in the fall followed by accumulation of energy stores
for the subsequent reproductive season.
Cupped oysters, including C. gigas, are protandric hermaphrodites (males becoming
females with increasing age and size), at a population level, but with a complex mode of sex
determination evident when sex ratios of individuals and experimental families are followed,
suggesting a combination of environmental and genetic factors (Coe 1936, Haley 1977, Guo et
al. 1998, Hedrick and Hedgecock 2010). Sex allocation theory suggests that protandric
hermaphroditism is favourable when the expected number of offspring produced increases with
18

body size for females, while remaining relatively constant for males (Ghiselin 1969, Charnov
1979). Specific details of sex determination remain elusive though the recent characterization of
genes homologous to sex determining factors in other organisms is a step in the right direction
(Naimi et al. 2009 a,b). A better understanding of sex determination could have favourable
consequences, such as production of female monocultures, which reach larger sizes in less time.
While originating from Asia, C. gigas is now farmed in all continents except Antarctica,
and is one of the highest produced marine or freshwater organisms in the world (4.2 million
metric tonnes, worth $3.5 billion US, FAO 2005). Farmers have largely abandoned the practice
of catching natural sets in favor of purchasing mass produced hatchery seed. Hatchery
production generates seed in great abundance (a single female may have upwards of 100 million
eggs), but larval survival is highly inconsistent (Lannan 1980, Barton et al. 2012).
Variance in larval survival may be due to genetic predisposition (Launey and Hedgecock
2001) or water chemistry (Barton et al. 2012) issues, but has also been shown to be affected by
the maturation status of spawning parents (Lannan et al. 1980). Unfortunately, methods to assess
broodstock ripeness prior to spawning typically require sacrificing the animals. Assessment of
maturation status is made more difficult by the presence of genetic variation in the rate of timing
of gonad maturation (Lannan 1980, Barber et al. 1991, Samain et al. 2007). Simply calibrating a
degree-day conditioning program for an optimal state of maturation often fails to predict larval
success because of this inherent variation.
Physiological and biochemical characterization of the stages of reproductive maturation
has been performed in relation to the expression of multiple genes involved in processes such as
energy storage and utilization, germline formation, cell signaling, stress response and sex
determination, in C. gigas (Fabioux et al. 2004, Bacca et al. 2005, Rodet et al. 2005, Fleury et al.
19

2008, Le Quere et al. 2009, Meistertzheim et al. 2009, Naimi et al. 2009a, Naimi et al. 2009b).
In the majority of these cases, the gene involved displayed some level of stage-specific
expression, highlighting the large physiological differences between the immature and mature
gonad. Sampling and screening reproductive material for gene expression represents a potential
first step towards establishing state of maturation of potential broodstock in a non-sacrificial
manner.
Knowledge of the transcriptome is essential for understanding changes that occur during
development. Microarrays revolutionized transcriptomic studies over a decade ago by allowing
for the simultaneous screening of thousands of genes at once. Recently emerging high-
throughput sequencing approaches, termed RNA-Seq, offer several benefits over microarrays,
including the ability to survey the entire transcriptome while not requiring any a priori
knowledge of the genome (Wang et al. 2009, Wilhelm and Landry 2009). Such sequencing
approaches have already been used in bivalve mollusks to investigate tissue-specific expression,
sex determination and anthropogenic influence (Craft et al. 2010, Ghiselli et al. 2011, Gavery
and Roberts 2012).
This study took an exploratory approach to identifying genes that are differentially
expressed (DEGs) between sexes and during oogenesis in C. gigas. It was assumed, based on
results seen in the literature referenced above, that reproductive maturation depends on stage-
specific gene expression. Oogenesis was specifically targeted because female parents have been
shown to contribute more than male parents to variance in larval survival (Lannan 1980). In
addition, production spawns have a greater dependency on female broodstock because of the
great number of eggs needed for seed propogation.
20

Libraries prepared from polyadenylated RNA from gonad tissue of three oyster families
were sequenced for males and females at a single gonadal maturation stage, and for females at
three maturation stages, followed by testing for differential expression and subsequent functional
analysis. The results identify sex-specific and oogenesis stage-specific differences in gene
expression. Sex- and reproductive stage-specific biological processes are identified, which
contribute to our understanding of the molecular biology of reproduction in this commercially
important bivalve.

21

Materials and Methods
Artificial conditioning of gonad maturation
Animals from two inbred lines and a hybrid cross of the two parental inbred lines,
produced in 2007, were used for artificial conditioning, which began in early March, 2010, at or
near the timing of natural initiation of gametogenesis. Conditioning was carried out at a constant
25 °C, after increasing the water temperature 2 °C/day from the ambient incurrent water
temperature of 13 °C. Oysters were held in flow through tanks and fed a live mixed algal diet.
Samples were taken after 2 (stage I), 6 (stage II), and 10 (stage III) weeks of conditioning
at 25 °C. Oysters were shucked and sectioned just ventral to the labial palps, after which a gonad
tissue sample was taken from the anterior half of the ventral section and immediately placed on
dry ice. The entire dorsal section was also placed on dry ice for reproductive staging. Total
RNA was extracted using TRIzol® Plus RNA Purification Kit (Invitrogen) and a TissueLyser II
(Qiagen).
Histology and Reproductive Stage Assignment
Histology was performed to assist in identifying the stage-specific nature of the samples
used for deep sequencing. Frozen tissue sections were held overnight at 4 °C in 30% sucrose
PBS solution, followed by mounting in OCT compound. Sections were then frozen in -125 °C
methylbutane for 20 seconds, then cut to 10 µm using a cryostat. Slides were stained with
hemotoxylin and eosin and imaged with an Olympus DP72 digital microscope camera.
Observations made at the time of sampling, including gonadal surface area, tissue
density, and visual inspections of tissue smears, were used in conjunction with histology for
assignment of reproductive stage. Specifically, mature sex cells were not present after two
weeks, were present within organized acini after six weeks, and were found homogenously
22

distributed through the gonad after 10 weeks. Samples were identified as, and assigned the
labels of reproductive stage I (immature), II (early-maturation), and III (late-maturation) based
on whether they were conditioned for 2, 6, and 10 weeks, respectively.
Library preparation
RNA purification, cDNA synthesis, and Illumina library preparation were accomplished
using Illumina’s preparation kit, supplemented at times with analogous NEB library preparation
reagents, and following Illumina’s library construction protocol. Total RNA was quantified on a
NanoDrop® ND-1000, while cDNA libraries were quantified on a Stratagene Mx3005P QPCR
instrument using a Quant-iT™ PicoGreen ® dsDNA kit. Library preparation was initiated with
10 µg RNA. In total, six samples (two per genotype) from each of the three sampling time
points were used for library construction and sequencing. Individual samples were barcoded
using Illumina’s protocol, and sequencing was performed on Genome Analyzer IIx and HiSeq
2000 Illumina sequencing systems. All reads used for mapping were single end and 51-bp in
length.
Mapping and testing of differential expression
Galaxy was used to assess quality and trim reads prior to mapping (Goecks et al. 2010,
Blankenberg et al. 2010). Reads were mapped to the C. gigas genome, v9 (Zhang et al. 2012)
using the splice junction mapping software Tophat2 v0.6 (Trapnell et al. 2009) with an anchor
length of 8 and 2 allowed mismatches. Tophat accepted hit output files were converted to raw
read counts using the htseq-count function in the Python package HTSeq (v0.5.4p2). Only reads
with unique mapping (a single alignment to the genome) were counted and used for downstream
DEG analysis. The GFF file required by HTSeq was created with Cufflinks and Cuffmerge,
tools within the so-called “tuxedo suite”, on the Tophat mapping files (Trapnell et al. 2012).
23

Read counts were then normalized among libraries using the method of Robinson and Oshlack
(2010) and tested for differential expression between sexes and among reproductive stages using
a modified Fisher’s exact test in the Bioconductor R package, EdgeR (Robinson et al. 2010).
Prior to testing for differential expression, EdgeR normalizes sample library count data by
sequencing depth and RNA composition (accounts for effect of possible highly expressed genes
in some samples but not others). Count data is then fit to a negative binomial model prior to
testing for differential expression. Differences among the three oyster families sampled were
also queried. Genes with a p-value ≤ 0.05 after corrected for multiple testing, by method of
Benjamini and Hochberg (1995), were considered differentially expressed.
Functional enrichment analysis
In order to assign function to biological processes associated with differences in
reproductive stage, differentially expressed genes were first compared to the protein sequence
derived from the C. gigas genome, version 9 (gigadb.org/pacific_oyster/). Comparisons were
made with BLASTX (Gish and States 1993) 2.2.25+, using a cutoff E-value of 1E-05. BLASTX
output was parsed using a perl script. DAVID v6.7 was used (Huang et al. 2009) for enrichment
analysis with a query of UniProt accession numbers and the C. gigas genome protein database as
a background. Biological processes for each pairwise sex and reproductive stage comparison
were considered significantly enriched with respect to the C. gigas protein database at the
FDR=0.05 level. DAVID uses a modified Fisher’s exact test in determining the enrichment
score. This simple contingency table test describes the likelihood that the proportion of genes
belonging to a certain biological pathway in a given gene list is the same as the proportion of the
pathway relative to the background genome list.
24

Real time quantitative PCR
Individuals for each of the three developmental stages were screened for gene expression
levels by qPCR in order to validate the RNA sequencing results. Six genes were chosen for
validation of RNA-Seq based on known function, significance level and direction of differential
expression seen by RNA sequencing. Primers were developed using Primer3
(http://bioinfo.ut.ee/primer3-0.4.0/) from cufflinks assembled sequences and checked for
uniqueness and specificity (Table 4).
RNA samples were treated with DNAse I (1 U/µg total RNA, Promega) prior to reverse
transcription. One µg total RNA was reverse transcribed to cDNA. Dilution series were
performed for all primer sets to determine the template range with PCR efficiency of R
2
= 1 +/-
0.05. Amplification reactions contained 5 ng template, 600 nm primer concentration, and 1 X
Brilliant II SYBR green QPCR master mix (Agilent) in a total volume of 15µl. All samples were
run in duplicate. NoRT controls were tested prior to measuring of gene expression, and no-
template controls were included for each gene of each qPCR run. Variation in sample input was
normalized by the geometric mean expression of the reference genes gapdh and hsp70.
Log transformed relative expression values were compared to RNA-Seq RPKM values
through calculation of the coefficients of determination. Differences in expression among stages
were tested by ANOVA. Stage-specific expression differences were compared between RNA-
Seq and qPCR data.

25

Results
Illumina sequencing and aligning to the reference genome
Six gonad samples from each of the three stages were used for total RNA extraction,
cDNA synthesis, and high-throughput Illumina sequencing. Stage I oysters consisted of three
individuals of each sex, while all stage II & III samples were females. Recovered sequences
were filtered for the presence of poor read quality, adapter sequence and uncalled bases. In
total, 136.5 million, 51-bp reads spanning the three stages were utilized for mapping (Table 1).
The relatively low level of average reads mapped per stage (50-53%) was due to the presence of
significant levels of mtDNA in several libraries.
Filtered reads were mapped to the C. gigas genome build v9. Merged cufflinks
assemblies contained 23,215 exons spanning 9,417 transcripts. The average exon length was 202
bp. Read counts were compiled per transcript, which on average contained nearly 2.5 exons.
While 9,358 of the 9,417 transcripts had at least a single read mapped from one library, only
2,216 transcripts averaged over 10 reads mapped per library. The average number of reads
mapped per transcript was 22. The distribution of average reads mapped per transcript is highly
skewed, with a mode of fewer than three (Figure 1).
Stage I had, by far, the greatest number of transcripts averaging 10 or more reads mapped
per sample, followed by stage II, then stage III. There were nearly as many transcripts with this
expression level found only in stage I as there were found shared in all three reproductive stages,
and over 10 times as many as were specific to either stage II or stage III (Figure 2). These
expression results led directly to patterns seen in subsequent stage-specific differential gene
expression analysis.
26

The cause of the elevated stage I expression was investigated by comparing mapping
results between the genome and the genome derived gene annotations, consisting of 28,027
protein coding sequences. Stage I samples had on average 8% more reads mapped to the
genome than to the protein coding sequences, suggesting the majority of mapped reads were
located within coding regions. Stages II and III had 74% and 89% more genes mapped to the
genome than to the protein coding sequences. The overall number of genes mapped to the
genome was similar among stages, but many more reads mapped to noncoding regions or
unannotated regions in stages II and III.
Analysis of stage-specific differential gene expression
The number of reads used for differential expression testing ended up being a small
percent of those initially mapped to the genome (Table 1). This result arose from a great number
of non-uniquely mapped reads as counted by HTSeq-count, which means a read mapped to more
than one location in the genome. Changing mapping parameter settings did not significantly
reduce the number of reads mapping to multiple genomic locations. HTSeq-count recognizes
multiple alignments by the optional NH tag in SAM output mapping files. If the origin of a read
cannot be determined due it mapping to it mapping in multiple location, use of them in DEG
testing may result in spurious DEG reporting. Transcripts may either not receive counts or will
have less power to detect DEGs among stages as a result of the number of ambiguous reads.
A total of 199 DEGs were found between stage I males and females, while a sum of 237
DEGs were identified from the six pairwise comparison outcomes among the three stages of
female gonad maturation. Of the stage-specific DEGs, 39 were identified as differentially
expressed in two comparisons while the remaining 198 were identified in only a single
comparison. The majority of DEGs were the result of higher expression in stage I than in either
27

stage II or stage III (Figure 3). The number of DEGs between sexes was nearly equal to the
number of DEGs found among the three reproductive stage of oogenesis.
Testing of differential expression was also performed among genotypes to assess the
potential variation in expression patterns among the three genotypes sampled. The use of
multiple genotypes was meant in part to account for natural variation in gonad maturation and as
a validation for stage-specific expression. If sufficient conditioning had occurred to isolate
stage-specific expression patterns, then DEGs for reproductive stage should not be found among
genotype specific DEGs. Not a single DEG was found among the three sampled genotypes.
Stage-specific expression patterns of DEGs were further investigated by clustering them
into one of four possible groups (Figure 4). Nearly 75% of the DEGs belonged to clusters 1 & 3,
which begin with stage I expressed at a higher level than stage II. Cluster 4, where stage II
expression is greater than both stage I and III, contains only 10 of the 199 DEGs, or roughly 5%.
Nearly 10% of DEGs were assigned to cluster 2, where expression increased from stage I to
stage II, and from stage II to stage III.
Functional analysis of DEGs
Assignment of function to DEGs was accomplished on an individual basis by blasting
cufflinks output against the C. gigas v9 protein database. The database contains 28,027 genes
50 amino acids or greater in length, predicted by a combination of de novo and evidence-based
search methods. Of the 28,027 predicted proteins, 75.2% matched entries in SWISS-PROT,
InterPro or TrEMBL, and were assigned a putative function (Zhang et al. 2012). When blasted
against the protein database, 65% of the 23,125 exons from the merged cufflinks assembly had
significant hits at an E-value of 1E-05. DEGs from the individual pairwise comparisons among
28

the reproductive stages and between sexes had successful blast results ranging from 48% to 87%
(Table 2).
After assessing function among individual DEGs, an enrichment analysis was performed
to see whether certain types of processes characterized individual reproductive stages and sexes
relative to one another. Comparisons with a higher number of DEGs were of course more likely
to produce positive enrichment results. A list of enriched processes can be seen in Table 3.
Neither stage II or III produced significant results for enrichment of biological processes versus
the other stages. Stage I showed processes relating to cell adhesion enriched relative to stage II,
and processes relating to amino acid transport enriched relative to stage III.
Processes related to cell movement were enriched in male DEGs. The genes contributing
to this enrichment had high similarity to genes involved in the production of sperm flagella.
Specific identified genes contributing included the kinesin-like protein kif-24, and multiple
dynein heavy chain subunits of the DNAH family. In female samples, organic and amino acid
transport proteins were enriched within DEGs. Members of solute carrier protein family 6 (slc6)
were central to the enrichment of the transport processes within female DEGs.
QPCR validation of RNA-Seq expression
Primers for the six genes, as well as genome and genbank accession numbers, can be
found in Table 4. Relative expression values for the six genes were first log
2
transformed prior
to testing of differential expression among stages of maturation by one-way ANOVA. Stage-
specific results between the two methods were largely consistent (Figure 5). Only one of the six
genes screened by qPCR, klf-5, failed to support differential expression identified through RNA-
Seq analysis. The pattern of low expression in stage II for klf-5 was consistent, however.

29

Discussion
Summary of stage-specific transcriptomics
Knowledge concerning stage-specific oogenesis in bivalve molluscs has been primarily
limited in source to ultrastructural microscopy investigations (Pipe 1987, Eckelbarger and Davis
1996). In C. gigas, there is a good understanding of cellular composition and organization
during gonad maturation, as well as an understanding of specific gene expression patterns of a
handful of genes considered important to reproduction (Fabioux et al. 2004, Bacca et al. 2005,
Rodet et al. 2005, Fleury et al. 2008, Le Quere et al. 2009, Meistertzheim et al. 2009, Naimi et al.
2009a, Naimi et al. 2009b). The work presented in this report provides a more global view of
expression differences between male and female reproductive tissue, as well as expression
differences among multiple stages of oogenesis.
The immature gonad had a far greater number of reads mapped to the Cufflinks assembly
(Table 1), despite having a similar number of reads overall. This likely contributed to the pattern
of DEGs being expressed at greater levels in stage I. The pattern of up-regulated DEGs in stage
I may reflect a more dynamic developmental environment. The gonad of C. gigas is a transient
organ that is built up in the spring and resorbed after spawning at the end of the summer. Before
eggs can be produced each year, a significant remodelling of the body must take place to
accommodate the developing gonad. Networks of tubules and branching ovarian acini, which
will house the developing eggs, must develop within the outer somatic connective tissue.
Follicle cells, which will surround and contribute to the maturation of eggs, must also be
produced. After the initiation of vitellogenesis (post-stage I), the gonad has largely reached its
mature size and gonad development is focused primarily on oocyte production.
30

The greater expression in Stage I over stages II and III may also reflect changes occurring
within maturing oocytes. Nearly half of stage II and III reads mapped to areas of the genome
outside coding regions. The majority of these mapped to rRNA genes. Ribosomal RNA is
known to increase in quantity in developing oocytes and consists of well over 50% of the total
RNA pool in mature mouse oocytes (Kaplan et al. 1982). Most rRNA lack poly(A) tails, but the
presence of polyadenylated forms is becoming more recognized (Kuai et al. 2004, Slomovic et
al. 2006). The stage-specific nature of the increase in rRNA transcripts makes it unlikely to be a
methodological error.
The 199 DEGs between males and females were more evenly distributed, with 117 genes
expressed at higher levels in males vs. 82 expressed higher in females. While many of the sex-
specific DEGs are functionally consistent with a sex-specific role, it’s possible even more would
have been identified had the analysis included more developmental stages. In the case of the
female transcriptome, given the stage-specific pattern of expression found, this is unlikely. Less
is known about development of the male transcriptome, but the biological processes identified as
enriched in males are fundamentally testis-specific.
Another multi-gene characterization of stage-specific gene expression during gonad
maturation in C. gigas was recently completed using a microarray based approach (Dheilly et al.
2012). Strangely, they found significant levels of increasing gene expression during maturation,
which contrasts strongly with the results presented here. It’s not obvious where this
inconsistency arises from, whether biological or methodological. Comparing expression of
vitellogenin between the two studies provides a direct example of this difference, where Dheilly
and colleagues found expression to decrease over maturation, while here expression was found to
increase over maturation. Vitellogenin is a good gene for comparison as it has been relatively
31

well studied among other bivalve mollusks. Without exception, other studies also showed
expression of vitellogenin to decrease during gonad maturation (Matsumoto et al. 2003, Agnese
et al. 2012, Zheng et al. 2012). Only when measuring protein levels have vitellogenin/vitellin
levels been reported to increase during maturation (Arcos et al. 2009), which aligns well with its
contribution as a stored energy reserve for developing larvae.
The following discussion highlights some stage-specific enriched biological processes as
well as individual DEGs and their potential role in gonad maturation. It is not meant to cover all
DEGs and enriched processes, but rather those previously associated with gonad maturation.
Sex-specific gene gonadal gene expression
Analysis of sex-specific differential expression of reproductive tissue identified enriched
processes characteristic of sexual differentiation. In males, enriched processes unsurprisingly
involved microtubule based movement and motility (Table 3). Contributing to this enrichment
were sequences with high similarity to multiple dynein heavy chains (dnah2, dnah5, dnah8),
kinesin-like protein 24 (kif24), and tubulin polyglutamylase (ttll6), among others. The most
obvious role of microtubules in the testis of C. gigas is to provide flagellar locomotion to mature
spermatozoa, though the epithelium of the gonadal tubules is ciliated as well.
Dyneins are ATP powered motor proteins which bind microtubules and cause them to
bend, creating movement (Summers and Gibbons 1971, Ogawa et al. 1977, Mohri et al. 2012).
These proteins are well conserved among animals and have previously been isolated and
characterized in C. gigas sperm (Wada et al. 1992). Like dyneins, kinesins are microtubule
binding motor proteins. It’s unclear what specific function Kif24 serves, though the protein has
recently been reported to play a role in the regulation of ciliogenesis in humans through the
depolymerization of microtubules (Kobayashi et al. 2011). And tubulin polyglutamylases, which
32

modify alpha tubulin chains, regulate flagellar motility by providing a binding substrate for
dyneins along microtubules (Kubo et al. 2012).
Several DEGs stood out on their own without contributing to the enrichment of biological
processes. Multiple kelch-like proteins (klhl3, klhl7, klhl13, dbo) were found among the 112
DEGs with greater male expression. Kelch proteins are evolutionarily widespread and diverse
(At least 71 kelch proteins exist in the human genome) but share the trait of containing multiple
kelch motifs, which form β-propellers (Adams et al. 2000, Prag and Adams 2003). In
Drosophila, klhl10 is necessary for spermatozoa bulk cytoplasm reduction during terminal
differentiation, and mutated copies cause male sterility (Arama et al. 2007). Other notable DEGs
include the well-known meiosis proteins spo11, which catalyses double-stranded breaks in DNA
during recombination (Keeney et al. 1997), and rec8, which is required for chromosomal
reduction at meiosis I (Watanabe and Nurse 1999). Instead of having a male-specific function,
these meiosis related DEGs likely reflect the greater frequency of meiosis in the testis compared
to the ovary.
Enriched processes in the female gonad involved the transport of amino acids. The genes
which contributed to the enrichment shared sequence similarity with the solute carrier family 6
(slc6a1, slc6a5, slc6a6, slc6a7, slc6a9, slc6a11, and slc6a12). Studies of amino acid transport
genes in bivalves have primarily concerned the impact on osmoregulation (Toyohara et al. 2005,
Hosoi et al. 2005, Hosoi et al. 2007). In addition to their role in osmoregulation, amino acid
transport genes also serve important roles in nutrition and protein synthesis (Christensen 1990).
While details of amino acid transport in developing oocytes have not been investigated in
bivalves, the topic has been well studied in mammals.
33

Uptake of amino acids in mouse oocytes has long been known to be affected by
maturation stage and their close coupling with surrounding cumulus cells (Cross and Brinster
1974, Colonna and Mangia 1983). Cumulus cells were found to increase transport of amino
acids into oocytes through increased expression of transport protein transcripts when in contact
with ooctyes, demonstrating a regulatory role of eggs towards their associated follicle cells
(Eppig et al. 2005). At least six amino acid transport systems are thought to occur in mouse
oocytes, with hypothesized roles in protein synthesis, cell volume regulation, and protection
from oxidative stress (Pelland et al. 2009). In C. gigas, higher expression of amino acid
transporter-like sequences in ovary than in testis suggests the possibility of similar roles existing
for amino acid transport in developing oocytes.
Processes enriched in stage I
The immature female gonad exhibited several enriched biological processes when
compared to stages II & III, where developing and mature oocytes dominated the ovary by
volume, respectively. Processes relating to cell adhesion were enriched with respect to the early-
maturation gonad of stage II. The majority of the sequences contributing to the cell adhesion
enrichment shared high sequence similarity to various integrin subunits (β-1, β-1b, β-3, β-5, β-
ps). Also contributing were a pair of tenascin-like sequences (tnc,tnn) and a collagen-like
sequence (col8a2), which are extracellular matrix proteins (ECM).
Integrins are large membrane-spanning glycoproteins that are grossly responsible for the
joining of cells and the establishment of bidirectional communication between them. In
developing mouse ovaries, they are thought to be important in anchoring stem cells to the
extracellular matrix and to each other (Burns et al. 2002). These cellular functions are likely
occurring in the immature C. gigas ovary, where maturing oocytes gradually lose contact with
34

the surrounding extracellular matrix as maturation progresses (Eckelbarger and Davis 1996).
Integrins are also more widely known to be involved in fertilization, where they span the egg
membrane. The study of integrins in C. gigas has so far been limited to their role in apoptosis
and regulation of the phagocytic immune response of hemocytes (Terahara et al. 2005, Terahara
et al. 2006).
Similar to the up-regulation of amino acid transport in females compared to males, stage I
females had significantly more expression of amino acid transport genes than stage III females.
Results again from mice show the coupling of follicle cells with oocytes increases the uptake of
amino acids into the oocytes (Pelland et al. 2009). By stage III, mature oocytes are de-coupled
from follicle cells in C. gigas, which may explain the enrichment of amino acid transport in stage
I vs. stage III. Immune response processes were also enriched in stage I vs. stage III samples.
Stage II and III female maturation
The overwhelming impression of stage II and III gonad is one or repressed transcription
(Figure 2, 4). Despite the three stages of female maturation having a similar number of reads and
reads mapped, stage I had on average nearly four times as many reads mapped to the cufflinks
assembly (Table 1). Further investigation is needed to understand the cause of this pattern, but
clues may be found from oogenesis and early embryogenesis studies in animal models.
Metazoan eggs are more or less transcriptionally silent prior to egg activation. Maternal
mRNAs, which will be needed to drive early embryogenesis, are created within immature eggs
and transported from adjoining nurse cells. It has been estimated mRNA from half the mouse
genome is present in mature eggs (Wang et al. 2004). The subject of much recent work,
deadenylation of maternal mRNAs is thought to be the primary stabilizing force of these
35

transcripts in arrested oocytes (see Piccioni et al. 2005, Weill et al. 2012, Barckmann and
Simonelig 2013 for reviews).
Multiple studies using in vitro development of bovine oocytes have reported a reduction
in poly(A) RNA in eggs as maturation progresses. The amount of reduction has been estimated
as high as 50%, and not all mRNA species seem to be subject to the same amount of
deadenylation (Lequarre et al. 2004, Thelie et al. 2007). The diluted poly(A) mRNA pool in
developing and mature oocytes, despite library preparation utilizing poly(A) selection, could
cause a relative reduction in the quantification of coding mRNA. Noncoding RNAs can be
polyadenylated as well, for example miRNA precursor molecules in humans (Cai et al. 2004). It
is clear, when comparing expression results among studies of reproductive development, care
must be taken to account for the methods used to procure the data, specifically with regard to
mRNA priming.
Although biological processes enriched in stage II or III were not identified, over 50
transcripts were up-regulated with respect to the immature gonad, the majority of which were
identified in stage III. A pair of genes in particular stood out for being differentially expressed in
stage II and III over stage I, in addition to having possible functional significance. One of them
is a homolog of thioredoxin reductase 2 (tnxrd2), one of two known mammalian thioredoxin
reductases, which are enzymes that reduced oxidized thioredoxin. Tnxrd’s are thought to play a
role in redox homeostasis in mature oocytes, and in bovine oocytes, tnxrd1 shows similar
maturation stage specific expression as was found in this study (Yuan et al. 2012). QPCR
analysis here confirmed lower expression in stage I female gonad.
The second gene up-regulated in both stage II and III over stage I was a synaptotagmin-
15 (syt15) homolog. Synaptotagmins are calcium sensing, membrane bound proteins with
36

multiple proposed functions in oocytes. In mouse oocytes, syt1 is necessary for proper spindle
stability, and the expression of syt1 was shown to increase during oocyte maturation. And in
urchin eggs, syt1 functions in docking cortical granules that, when exocytosed after fertilization,
alter the cell membrane to prevent polyspermy (Leguia et al. 2006). Exocytosis of cortical
granules represents the slow block to polyspermy, which in oysters is thought to occur about 15
minutes after fertilization (Togo and Morisawa 1999). Recent evidence, however, suggests high
variation in gamete recognition protein sequence may play a large role in the prevention of
polyspermy in oysters (Moy et al. 2008).
Summary of RNA-Seq analysis of gonad maturation in C. gigas
There were several factors that impacted the quality of mapping and testing of differential
expression reported here. Multiple libraries contained significant levels of mtDNA, which
reduced the overall number of reads mapping to the genome and available for downstream
analysis. Due to the large numbers of mitochondria in eggs (Matsumoto et al. 1974), an extra
isolation step during library preparation may help to avoid this contamination. Accumulation of
rRNA during gonad maturation also impacted DEG analysis by contributing to the lower read
counts available for DEG testing in stage II and III samples. Because library size is normalized
by EdgeR, the impact of elevated rRNA on stage-specific DEG was lessened, but it certainly
raises uncertainty about the interpretation of DEG results. Lastly, with so many reads mapping
to multiple locations in the genome, the overall number of reads used in DEG testing was greatly
reduced. In spite of these challenges of conducting RNA-Seq analysis with developing
reproductive tissue in C. gigas, meaningful results were attained characterizing sex and stage
specific expression patterns.
37

A better understanding of gonad maturation is important in this commercially important
species. This work is the first global expression study of reproductive material in C. gigas done
by RNA sequencing. Characteristic biological processes were identified of immature male and
female gonad tissue. The immature female gonad undergoes extensive remodeling prior to
oocyte maturation, which likely contributes to the up-regulation of stage I tissue when compared
to the more static environments of stage II and III gonad. Increasing quantities of rRNA likely
reduced the power to detect potential DEGs in stage II and III samples, though the lack of
transcription and possible deadenylation of mRNA transcripts within oocytes may also have
contributed to the pattern of results found. The genomic resources established here also hold
potential for identifying candidate genes for state of gonad maturation, which heretofore has no
molecular biological markers.

38

References
Adams, J., Kelso, R., Cooley, L., 2000. The kelch repeat superfamily of proteins: Propellers of
cell function. Trends Cell Biol., 10, 17-24.
Agnese, M., Verderame, M., De Meo, E., Prisco, M., Rosati, L., Limatola, E., del Gaudio, R.,
Aceto, S., Andreuccetti, P., 2013. A network system for vitellogenin synthesis in the mussel
Mytilus galloprovincialis (L.). J. Cell. Physiol., 228, 547-555.
Arama, E., Bader, M., Rieckhof, G.E., Steller, H., 2007. A ubiquitin ligase complex regulates
caspase activation during sperm differentiation in drosophila. Plos Biol., 5, 2270-2287.
Arcos, F.G., Maria Ibarra, A., del Carmen Rodriguez-Jaramillo, M., Awilda Garcia-Latorre, E.,
Vazquez-Boucard, C., 2009. Quantification of vitellin/vitellogenin-like proteins in the oyster
Crassostrea corteziensis (hertlein 1951) as a tool to predict the degree of gonad maturity.
Aquacult. Res. 40, 644-655.
Bacca, H., Huvet, A., Fabioux, C., Daniel, J.-Y., Delaporte, M., Pouvreau, S., Van Wormhoudt,
A., Moal, J., 2005. Molecular cloning and seasonal expression of oyster glycogen
phosphorylase and glycogen synthase genes. Comp. Biochem. Phys. B, 140, 635-646.
Barber, B.J., Ford, S.E., Wargo, R.N., 1991. Genetic variation in the timing of gonadal
maturation and spawning of the eastern oyster, Crassostrea virginica (gmelin). Biol. Bull.,
181, 216-221.
Barckmann, B. and Simonelig, M., 2013. Control of maternal mRNA stability in germ cells and
early embryos. BBA-Gene Regul. Mech., 1829, 714-724.
Barton, A., Hales, B., Waldbusser, G.G., Langdon, C., Feely, R.A., 2012. The pacific oyster,
crassostrea gigas, shows negative correlation to naturally elevated carbon dioxide levels:
Implications for near-term ocean acidification effects. Limnol. Oceanogr., 57, 698-710.
39

Benjamini, Y. and Hochberg, Y., 1995. Controlling the false discovery rate: A practical and
powerful approach to multiple testing. J. Roy. Stat. Soc. B, 57, 289-300.
Blankenberg, D., Gordon, A., Von Kuster, G., Coraor, N., Taylor, J., Nekrutenko, A., the, G.T.,
2010. Manipulation of FASTQ data with galaxy. Bioinformatics, 26, 1783-1785.
Boutet, I., Moraga, D., Marinovic, L., Obreque, J., Chavez-Crooker, P., 2008. Characterization
of reproduction-specific genes in a marine bivalve mollusc: Influence of maturation stage
and sex on mRNA expression. Gene, 407, 130-138.
Breese, W. and Malouf, R., 1977. Hatchery rearing techniques for oyster Crassostrea rivularis
(Gould, 1861). Aquaculture, 12, 123-126.
Burns, K., Owens, G., Fernandez, J., Nilson, J., Matzuk, M., 2002. Characterization of integrin
expression in the mouse ovary. Biol. Reprod., 67, 743-751.
Cai, X., Hagedorn, C., Cullen, B., 2004. Human microRNAs are processed from capped,
polyadenylated transcripts that can also function as mRNAs. RNA, 10, 1957-1966.
Charnov, E., 1979. Natural-selection and sex change in pandalid shrimp - test of a life-history
theory. Am. Nat., 113, 715-734.
Choo, Y. and Kim, S., 2006. Detection of human adenoviruses and enteroviruses in korean
oysters using cell culture, integrated cell culture-PCR, and direct PCR. Journal of
Microbiology, 44, 162-170.
Christensen, H., 1990. Role of amino-acid-transport and countertransport in nutrition and
metabolism. Physiol. Rev., 70, 43-77.
Ciocan, C.M., Cubero-Leon, E., Minier, C., Rotchell, J.M., 2011. Identification of reproduction-
specific genes associated with maturation and estrogen exposure in a marine bivalve Mytilus
edulis. Plos One, 6, e22326.
40

Coe, W., 1936. Environment and sex in the oviparous oyster Ostrea virginica. Biol. Bull., 71,
353-359.
Colonna, R. and Mangia, F., 1983. Mechanisms of amino-acid-uptake in cumulus-enclosed
mouse oocytes. Biol. Reprod., 28, 797-803.
Craft, J.A., Gilbert, J.A., Temperton, B., Dempsey, K.E., Ashelford, K., Tiwari, B., Hutchinson,
T.H., Chipman, J.K., 2010. Pyrosequencing of Mytilus galloprovincialis cDNAs: Tissue-
specific expression patterns. Plos One, 5, e8875.
Cross, P. and Brinster, R., 1974. Leucine uptake and incorporation at 3 stages of mouse oocyte
maturation. Exp. Cell Res., 86, 43-46.
Dheilly, N.M., Lelong, C., Huvet, A., Favrel, P., 2011. Development of a pacific oyster
(Crassostrea gigas) 31,918-feature microarray: Identification of reference genes and tissue-
enriched expression patterns. BMC Genomics, 12, 468.
Dheilly, N.M., Lelong, C., Huvet, A., Kellner, K., Dubos, M., Riviere, G., Boudry, P., Favrel, P.,
2012. Gametogenesis in the pacific oyster Crassostrea gigas: A microarrays-based analysis
identifies sex and stage specific genes. Plos One, 7, e36353.
Eckelbarger, K.J. and Davis, C.V., 1996. Ultrastructure of the gonad and gametogenesis in the
eastern oyster, Crassostrea virginica. I. ovary and oogenesis. Mar. Biol., 127, 79-87.
Eppig, J., Pendola, F., Wigglesworth, K., Pendola, J., 2005. Mouse oocytes regulate metabolic
cooperativity between granulosa cells and oocytes: Amino acid transport. Biol. Reprod., 73,
351-357.
Fabioux, C., Pouvreau, S., Roux, F.L., Huvet, A., 2004. The oyster vasa-like gene: A specific
marker of the germline in Crassostrea gigas. Biochem. Bioph. Res. Co., 315, 897-904.
41

Fabioux, C., Huvet, A., Le Souchu, P., Le Pennec, M., Pouvreau, S., 2005. Temperature and
photoperiod drive Crassostrea gigas reproductive internal clock. Aquaculture, 250, 458-
470.
Fleury, E., Fabioux, C., Lelong, C., Favrel, P., Huvet, A., 2008. Characterization of a gonad-
specific transforming growth factor-β superfamily member differentially expressed during
the reproductive cycle of the oyster Crassostrea gigas. Gene, 410, 187-196.
Galtsoff, P.F., 1964. The American Oyster, Crassostrea Virginica Gmelin. U.S. Fish and
Wildlife Service, 1-480.
Gavery, M.R. and Roberts, S.B., Characterizing short read sequencing for gene discovery and
RNA-seq analysis in Crassostrea gigas. Comp. Biochem. Phys. D, 7, 94-99.
Ghiselin , M., 1969. Evolution of hermaphroditism among animals. Q. Rev. Biol., 44, 189-208.
Ghiselli, F., Milani, L., Chang, P.L., Hedgecock, D., Davis, J.P., Nuzhdin, S.V., Passamonti, M.,
2012. De novo assembly of the manila clam ruditapes philippinarum transcriptome provides
new insights into expression bias, mitochondrial doubly uniparental inheritance and sex
determination. Mol. Biol. Evol., 29, 771-786.
Gish, W. and States, D.J., 1993. Identification of protein coding regions by database similarity
search. Nat. Genet., 3, 266-272.
Goecks, J., Nekrutenko, A., Taylor, J., Galaxy Team, 2010. Galaxy: A comprehensive approach
for supporting accessible, reproducible, and transparent computational research in the life
sciences. Genome Biol., 11, R86.
Guo, X., Hedgecock, D., Hershberger, W.K., Cooper, K., Allen Jr., S.K., 1998. Genetic
determinants of protandric sex in the pacific oyster, Crassostrea gigas thunberg. Evolution,
52, 394-402.
42

Haley, L., 1977. Sex determination in the American oyster. J. Hered., 68, 114-116.
Hedrick, P.W. and Hedgecock, D., 2010. Sex determination: Genetic models for oysters. J.
Hered., 101, 602-611.
Hosoi, M., Takeuchi, K., Sawada, H., Toyohara, H., 2005. Expression and functional analysis of
mussel taurine transporter, as a key molecule in cellular osmoconforming. J. Exp. Biol., 208,
4203-4211.
Hosoi, M., Shinzato, C., Takagi, M., Hosoi-Tanabe, S., Sawada, H., Terasawa, E., Toyohara, H.,
2007. Taurine transporter from the giant pacific oyster Crassostrea gigas: Function and
expression in response to hyper- and hypo-osmotic stress. Fisheries Sci., 73, 385-394.
Huang, D.W., Sherman, B.T., Lempicki, R.A., 2008. Systematic and integrative analysis of large
gene lists using DAVID bioinformatics resources. Nat. Protocols, 4, 44-57.
Kaplan, G., Abreu, S.L., Bachvarova, R., 1982. rRNA accumulation and protein synthetic
patterns in growing mouse oocytes. J. Exp. Zool., 220, 361-370.
Keeney, S., Giroux, C., Kleckner, N., 1997. Meiosis-specific DNA double-strand breaks are
catalyzed by Spo11, a member of a widely conserved protein family. Cell, 88, 375-384.
Kobayashi, T., Tsang, W.Y., Li, J., Lane, W., Dynlacht, B.D., 2011. Centriolar kinesin Kif24
interacts with CP110 to remodel microtubules and regulate ciliogenesis. Cell, 145, 914-925.
Kuai, L., Fang, F., Butler, J.S., Sherman, F., 2004. Polyadenylation of rRNA in Saccharomyces
cerevisiae. P. Natl. A. Sci. USA, 101, 8581-8586.
Kubo, T., Yagi, T., Kamiya, R., 2012. Tubulin polyglutamylation regulates flagellar motility by
controlling a specific inner-arm dynein that interacts with the dynein regulatory complex.
Cytoskeleton, 69, 1059-1068.
43

Lannan, J.E., 1980. Broodstock management of Crassostrea gigas: I. genetic and environmental
variation in survival in the larval rearing system. Aquaculture, 21, 323-336.
Lannan, J.E., Robinson, A., Breese, W.P., 1980. Broodstock management of Crassostrea gigas:
II. broodstock conditioning to maximize larval survival. Aquaculture, 21, 337-345.
Launey, S. and Hedgecock, D., 2001. High genetic load in the pacific oyster Crassostrea gigas.
Genetics, 159, 255-265.
Le Quere, H., Herpin, A., Huvet, A., Lelong, C., Favrel, P., 2009. Structural and functional
characterizations of an activin type II receptor orthologue from the pacific oyster
Crassostrea gigas. Gene, 436, 101-107.
Lequarre, A., Traverso, J., Marchandise, J., Donnay, I., 2004. Poly(A) RNA is reduced by half
during bovine oocyte maturation but increases when meiotic arrest is maintained with CDK
inhibitors. Biol. Reprod., 71, 425-431.
Mann, R., 1983. The role of introduced bivalve mollusc species in mariculture. J. World
Aquacult. Soc., 14, 546-559.
Matsumoto, L., Kasamatsu, H., Piko, L., Vinograd, J., 1974. Mitochondrial-dna replication in
sea-urchin oocytes. J. Cell Biol., 63, 146-159.
Matsumoto, T., Nakamura, A., Mori, K., Kayano, T., 2003. Molecular characterization of a
cDNA encoding putative vitellogenin from the Pacific oyster Crassostrea gigas. Zool. Sci.,
20, 37-42.
Meistertzheim, A.L., Lejart, M., Le Goïc, N., Thébault, M.T., 2009. Sex-, gametogenesis, and
tidal height-related differences in levels of HSP70 and metallothioneins in the pacific oyster
Crassostrea gigas. Comp. Biochem. Phys. A, 152, 234-239.
44

Mohri, H., Inaba, K., Ishijima, S., Baba, S.A., 2012. Tubulin-dynein system in flagellar and
ciliary movement. P. Jpn. Acad. B-Phys., 88, 397-415.
Moy, G.W. and Vacquier, V.D., 2008. Bindin genes of the pacific oyster Crassostrea gigas.
Gene, 423, 215-220.
Muranaka, M.S. and Lannan, J.E., 1984. Broodstock management of Crassostrea gigas:
Environmental influences on broodstock conditioning. Aquaculture, 39, 217-228.
Naimi, A., Martinez, A.S., Specq, M.L., Diss, B., Mathieu, M., Sourdaine, P., 2009. Molecular
cloning and gene expression of cg-Foxl2 during the development and the adult
gametogenetic cycle in the oyster Crassostrea gigas. Comp. Biochem. Phys. B, 154, 134-
142.
Naimi, A., Martinez, A.S., Specq, M.L., Mrac, A., Diss, B., Mathieu, M., Sourdaine, P., 2009.
Identification and expression of a factor of the DM family in the oyster Crassostrea gigas.
Comp. Biochem. Phys. A, 152, 189-196.
Ogawa, K., Mohri, T., Mohri, H., 1977. Identification of dynein as outer arms of sea-urchin
sperm axonemes. P. Natl. Acad. Sci. USA, 74, 5006-5010.
Pelland, A.M.D., Corbett, H.E., Baltz, J.M., 2009. Amino acid transport mechanisms in mouse
oocytes during growth and meiotic maturation. Biol. Reprod., 81, 1041-1054.
Piccioni, F., Zappavigna, V., Verrotti, A., 2005. Translational regulation during oogenesis and
early development: The cap-poly(A) tail relationship. C. R. Biol., 328, 863-881.
Pipe, R.K., 1987. Oogenesis in the marine mussel Mytilus edulis: An ultrastructural study. Mar.
Biol., 95, 405-414.
Prag, S. and Adams, J., 2003. Molecular phylogeny of the kelch-repeat superfamily reveals an
expansion of BTB/kelch proteins in animals. BMC Bioinformatics, 4, 42.
45

Robinson, M.D. and Oshlack, A., 2010. A scaling normalization method for differential
expression analysis of RNA-seq data. Genome Biol., 11, R25.
Robinson, M.D., McCarthy, D.J., Smyth, G.K., 2010. edgeR: A bioconductor package for
differential expression analysis of digital gene expression data. Bioinformatics, 26, 139-140.
Rodet, F., Lelong, C., Dubos, M., Costil, K., Favrel, P., 2005. Molecular cloning of a molluscan
gonadotropin-releasing hormone receptor orthologue specifically expressed in the gonad.
BBA-Gene Struct. Expr., 1730, 187-195.
Samain, J.F., Dégremont, L., Soletchnik, P., Haure, J., Bédier, E., Ropert, M., Moal, J., Huvet,
A., Bacca, H., Van Wormhoudt, A., Delaporte, M., Costil, K., Pouvreau, S., Lambert, C.,
Boulo, V., Soudant, P., Nicolas, J.L., Le Roux, F., Renault, T., Gagnaire, B., Geret, F.,
Boutet, I., Burgeot, T., Boudry, P., 2007. Genetically based resistance to summer mortality
in the pacific oyster (Crassostrea gigas) and its relationship with physiological,
immunological characteristics and infection processes. Aquaculture, 268, 227-243.
Summers, K. and Gibbons, I., 1971. Adenosine triphosphate-induced sliding of tubules in
trypsin-treated flagella of sea-urchin sperm. P. Natl. Acad. Sci. USA, 68, 3092-&.
Terahara, K., Takahashi, K., Mori, K., 2005. Pacific oyster hemocytes undergo apoptosis
following cell-adhesion mediated by integrin-like molecules. Comp. Biochem. Phys. A, 141,
215-222.
Terahara, K., Takahashi, K., Nakamura, A., Osada, M., Yoda, M., Hiroi, T., Hirasawa, M., Mori,
K., 2006. Differences in integrin-dependent phagocytosis among three hemocyte
subpopulations of the pacific oyster, Crassostrea gigas. Dev. Comp. Immunol., 30, 667-
683.
46

Thelie, A., Papillier, P., Pennetier, S., Perreau, C., Traverso, J.M., Uzbekova, S., Mermillod, P.,
Joly, C., Humblot, P., Dalbies-Tran, R., 2007. Differential regulation of abundance and
deadenylation of maternal transcripts during bovine oocyte maturation in vitro and in vivo.
BMC Dev. Biol., 7, 125.
Togo, T. and Morisawa, M., 1999. Mechanisms for blocking polyspermy in oocytes of the oyster
Crassostrea gigas. J. Exp. Zool., 283, 307-314.
Toyohara, H., Ikeda, M., Goto, C., Sawada, H., Hosoi, M., Takeuchi, K., Hayashi, I., Imamura,
S., Yamashita, M., 2005. Osmo-responsive expression of oyster amino acid transporter gene
and analysis of the regulatory region involved. Fisheries Sci., 71, 465-470.
Trapnell, C., Pachter, L., Salzberg, S.L., 2009. TopHat: Discovering splice junctions with RNA-
seq. Bioinformatics, 25, 1105-1111.
Trapnell, C., Roberts, A., Goff, L., Pertea, G., Kim, D., Kelley, D.R., Pimentel, H., Salzberg,
S.L., Rinn, J.L., Pachter, L., 2012. Differential gene and transcript expression analysis of
RNA-seq experiments with TopHat and cufflinks. Nat. Protoc., 7, 562-578.
Wada, S., Okuno, M., Nakamura, K., Mohri, H., 1992. Dynein of sperm flagella of oyster
belonging to protostomia also has a 2-headed structure. Biol. Cell, 76, 311-317.
Wang, Z., Gerstein, M., Snyder, M., 2009. RNA-seq: A revolutionary tool for transcriptomics.
Nat. Rev. Genet., 10, 57-63.
Wang, Q., Piotrowska, K., Ciemerych, M., Milenkovic, L., Scott, M., Davis, R., Zernicka-Goetz,
M., 2004. A genome-wide study of gene activity reveals developmental signaling pathways
in the preimplantation mouse embryo. Dev. Cell, 6, 133-144.
Watanabe, Y. and Nurse, P., 1999. Cohesin Rec8 is required for reductional chromosome
segregation at meiosis. Nature, 400, 461-464.
47

Wilhelm, B.T. and Landry, J., 2009. RNA-seq-quantitative measurement of expression through
massively parallel RNA-sequencing. Methods, 48, 249-257.
Yuan, Y., Wheeler, M.B., Krisher, R.L., 2012. Disrupted redox homeostasis and aberrant redox
gene expression in porcine oocytes contribute to decreased developmental competence.
Biol. Reprod., 87, 78.
Zhang, G., Fang, X., Guo, X., Li, L., Luo, R., Xu, F., Yang, P., Zhang, L., Wang, X., Qi, H.,
Xiong, Z., Que, H., Xie, Y., Holland, P.W.H., Paps, J., Zhu, Y., Wu, F., Chen, Y., Wang, J.,
Peng, C., Meng, J., Yang, L., Liu, J., Wen, B., Zhang, N., Huang, Z., Zhu, Q., Feng, Y.,
Mount, A., Hedgecock, D., Xu, Z., Liu, Y., Domazet-Loso, T., Du, Y., Sun, X., Zhang, S.,
Liu, B., Cheng, P., Jiang, X., Li, J., Fan, D., Wang, W., Fu, W., Wang, T., Wang, B., Zhang,
J., Peng, Z., Li, Y., Li, N., Wang, J., Chen, M., He, Y., Tan, F., Song, X., Zheng, Q., Huang,
R., Yang, H., Du, X., Chen, L., Yang, M., Gaffney, P.M., Wang, S., Luo, L., She, Z., Ming,
Y., Huang, W., Zhang, S., Huang, B., Zhang, Y., Qu, T., Ni, P., Miao, G., Wang, J., Wang,
Q., Steinberg, C.E.W., Wang, H., Li, N., Qian, L., Zhang, G., Li, Y., Yang, H., Liu, X.,
Wang, J., Yin, Y., Wang, J., 2012. The oyster genome reveals stress adaptation and
complexity of shell formation. Nature, 490, 49-54.
Zheng, H., Zhang, Q., Liu, H., Liu, W., Sun, Z., Li, S., Zhang, T., 2012. Cloning and expression
of vitellogenin (vg) gene and its correlations with total carotenoids content and total
antioxidant capacity in noble scallop Chlamys nobilis (bivalve: Pectinidae). Aquaculture,
366, 46-53.

48

Tables

Table 1
Summary of mapping, assembly, and read counting by reproductive stage. HTSeq counts used
for DEG analysis constituted only reads uniquely mapped to genome.

Immature
(Stage I)
Early-
Maturation
(Stage II)
Late-
Maturation
(Stage III)
Total filtered reads 5.01E+07 4.02E+07 4.62E+07
Reads mapped to genome 53% 52% 50%
Cufflinks assembled transcripts 9417 9417 9417
Genomic scaffolds covered by
cufflinks assembly 45% 45% 45%
HTSeq counts to Cufflinks assembly 2.48E+06 7.18E+05 5.35E+05
49

Table 2
Significant DEGs. Positive blast against C. gigas v9 protein database at E-value of 1E-05

Exons positive blast % with blast hits
Cufflinks exons 23125 15006 0.65
stage I over stage II 287 223 0.78
stage I over stage III 398 288 0.72
stage II over stage I 27 13 0.48
stage II over stage III 15 13 0.87
stage III over stage I 199 133 0.67
stage III over stage II 5 4 0.80
Male over Female 852 460 0.54
Female over Male 202 142 0.70
50

Table 3
Biological processes enriched among sex and stage comparison DGEs relative to cufflinks
assembly. No stage II or III enrichment was found at FDR ≤0.05

Uniprot ID Biological Process FDR
Male over Female
GO:0007018 microtubule-based movement 5.30E-05
GO:0007017 microtubule-based process 2.28E-04
GO:0001539 ciliary or flagellar motility 0.026375
Female over Male
GO:0006836 neurotransmitter transport 5.85E-09
GO:0006865 amino acid transport 2.46E-04
GO:0015837 amine transport 2.46E-04
GO:0015849 organic acid transport 0.001354
GO:0046942 carboxylic acid transport 0.001354
Stage I over Stage II
GO:0007229
integrin-mediated signaling
pathway 1.59E-06
GO:0007160 cell-matrix adhesion 8.01E-06
GO:0031589 cell-substrate adhesion 8.01E-06
GO:0007155 cell adhesion 3.95E-05
GO:0022610 biological adhesion 3.95E-05
Stage I over Stage III
GO:0006836 neurotransmitter transport 0.001898
GO:0006955 immune response 0.002022
GO:0006865 amino acid transport 0.015542
GO:0015837 amine transport 0.015542
51

Table 4
qPCR primer sequences and gene information
Gene Name Abbrev. Genbank OysterDB Primer Sequence
Vitellogenin-6 Vit-6 AB084783 OYG_10021817 5'-GAAGTTTACGTGAGAAGAC-3'
5'-TAGAAGTCTCTGAAGAACAG-3'
Pancreatic lipase-2 Lipase AM857075 OYG_10027499 5'-ATTCACGGCGTTTCCCTGT-3'
5'-CATTCTGTTGCAGCCGTTCT-3'
Hemagglutinin/amebocyte HAAF CU995762 OYG_10000742 5'-TGGTAGTCTCGGGATGGAAG-3'
aggregation factor 5'-CGAAGGACAGGTCTGAAGGA-3'
Forkhead box protein l1 Foxl1 AM865563 OYG_10006159 5'-TGCACATACCACATCCCAAC-3'
5'-TGCGAGTAACTCACGGAAGG-3'
Thioredoxin reductase 2, ThioR2 AM865969 OYG_10000920 5'-GGACTGAAGTCCGTCAAGCA-3'
mitochondrial 5'-AAACACGCCCACCTCAAAG-3'
Krueppel-like factor 5 Klf-5 CU991730 OYG_10000441 5'-AGGATTCGATCTGGGAGGAC-3'
5'-ATGGGATCGTTGGGAGAGTT-3'

52

Figures

Figure 1. Average number of reads mapped to 9,417 cufflinks transcripts across all 18 samples.
Another 622 transcripts averaged greater than 50 reads mapped.
Figure 2. Number of cufflinks transcripts with average expression of 10 or more, unique and
shared, among the three reproductive stages.
Figure 3. Differential expression plots showing DEGs in red. Units are log values of normalized
expression calculated in edgeR. Numbers in boxes indicate the number of DEGs up-regulated in
the condition on the adjacent axis. (a) stage I vs. stage II. (b) stage I vs. stage III. (c) stage II vs.
stage III. (d) males vs. females.
Figure 4. Expression patterns of the 237 DEGs over the course of the three maturation stages.
Cluster 1 was the most abundant where the expression in stage I > stage II > stage III, with 320
DEGs. The reverse pattern in Cluster 2 contained 51 DEGs. 128 profiles fit to Cluster 3, where
expression was at a minimum in Stage II. And the reverse pattern, in Cluster 4 where stage II
had a greater level of expression than Stage I and III, contained 30 profiles.
Figure 5. qPCR and RNA-Seq expression of six genes. Dark bars are log
2
RPKM values, light
bars qPCR expression relative to gapdh and hsp70. Data labels indicate the presence of
significant differences in expression.

53

Figure 1

Reads Mapped
Frequency
0 10 20 30 40 50
0 1000 2000 3000
54

Figure 2

55

Figure 3

56

Figure 4

57

Figure 5

58

Chapter 3: Quantification of total and poly(A) RNA in selected
biomarkers during gonad maturation in the Pacific oyster
Crassostrea gigas

Abstract
Deadenylation of mRNA in developing tissues, such as reproductive tissue, is a means of
repressing translation until transcripts are needed. Interpretation of gene expression data derived
from methodologies that rely on reverse transcription of mRNA, such as qPCR or RNA-Seq,
should therefore consider the priming strategy employed. Expression differences between total
and poly(A) RNA content were assessed for 12 genes at multiple stages of gonad maturation.
Reproduction-specific and housekeeping genes were included to check whether deadenylation of
transcripts during maturation is common among all genes or specific to those related to
reproduction. Deadneylation was not specific to reproduction-specific genes. Half of the 12
genes showed evidence of differing patterns of deadenylation among stages of gonad maturation.
Of these, the resulting pattern of differential expression among stages of reproduction of three
genes differed between total and poly(A) RNA. A reduction in the relative amount of total to
poly(A) RNA throughout maturation was the most common pattern found, suggesting transcripts
are commonly deadenylated early in maturation until utilized through polyadenylation. These
results suggest deadenylation may be common among genes during gonad maturation as a means
of suppressing translation of mRNA. Interpretation of gene expression levels during gonad
maturation should take caution to address results in terms of total or poly(A) RNA levels.

59

Introduction
Zygotic development is initially translationally driven prior to the commencement of
genomic transcription. Embryonic gene activation (EGA) occurs at a time called the midblastula
transition (MBT) in most animals. Prior to EGA, contributions from the egg play critical roles in
setting the stage for future autonomous development (Hara et al. 2005). Maternal mRNAs stored
in the egg are activated through polyadenylation and cap methylation soon after fertilization
occurs. A nearly 20% increase in poly(A) content has been observed between unfertilized eggs
and 1-cell embryos in mice (Piko and Clegg 1982).
According to the closed loop model of translation initiation, protein mediated contact is
established between the 5’ and 3’ ends of the mRNA (Jacobson 1986). One of the proteins is a
poly(A) binding protein (PABP). Through removal of the poly(A) tail, an mRNA molecule may
thus be masked from translation. Other factors are now known to participate in masking
transcripts from translation (Richter and Lasko 2011). Storage of mRNA though deadenylation
in mature oocytes is common among animals. It was first directly measured in the surf clam,
Spisula (Rosenthal et al. 1983), and then subsequently and more extensively in mice and
Xenopus (Paynton et al. 1988, Wickens 1990).
Compared with early embryonic development, relatively little is known about patterns of
polyadenylation and deadenylation during growth and maturation of oocytes. The influence of
these processes contributes to the atypical nature of the female reproductive transcriptome.
Unlike most tissues, protein synthesis does not necessarily reflect concurrent gene transcription.
Polyadenylated transcripts entering the cytoplasm from the nucleus may be subject to various
patterns of deadenylation, degradation and readenylation. In the mouse, the majority of mRNAs
60

are subject to deadenylation, including constitutively expressed “housekeeping” genes (Paynton
et al. 1988).
Crassostrea gigas is a commercially important bivalve mollusk which experiences high
hatchery-based variance in larval survival. Non-optimal reproductive status of broodstock is at
least partially responsible for this drain on efficiency of production (Lannan 1980).
Reproduction follows an annual cycle but is relatively plastic when exposed to artificial variation
in water temperature and food levels. Meiotic maturation of oocytes progresses to metaphase I
after stripping, or release of gametes (Colas and Dube 1998). Characterization of individual and
global gene expression patterns during gonad maturation in C. gigas has begun (Fabioux et al.
2004, Fleury et al. 2008, Naimi et al 2009ab, Dheilly et al. 2012), which should contribute to the
understanding and identification of reproductive stage.
Comparison, among different studies, of gene expression data based on qPCR has proven
difficult (Bustin et al. 2009). Here, I look at the impact of priming strategy (oligo(dT) vs.
random hexamer) during reverse transcription on qPCR generated data. In theory, random
hexamer primed cDNA samples will represent the entire cellular RNA pool, while oligo(dT)
primed cDNA samples will only represent polyadenylated RNA. Specifically, I was curious to
know whether patterns of polyadenylation and deadenylation were detectable during gonad
maturation, and if so, whether they were significant enough to alter interpretation of stage-
specific expression. Expression of 12 genes was analyzed, including 7 known differentially
expressed genes (DEGs) during gonad maturation, and 5 commonly used housekeeping genes.

61

Materials and Methods
Artificial conditioning of gonad maturation and stage-specific assignment
Animals from two inbred lines and the two hybrid cross of the parental inbred lines,
produced in 2007, were used for artificial conditioning, which began in early March, 2010, at or
near the timing of natural initiation of gametogenesis. Multiple families were used to account for
the possibility of poor conditioning in a particular family, and to increase the number of
individuals sampled. Conditioning was carried out at a constant 25 °C, after increasing the water
temperature 2 °C day
-1
from the ambient incurrent water temperature of 13 °C. Conditioning
was halted after 10 weeks, at which point active resorbtion was noticed among sampled female
gonads. Oysters were held in flow through tanks and fed a live mixed algal diet. Oysters were
shucked and sectioned just ventral to the labial palps, after which a gonad tissue sample was
taken from the anterior half of the ventral section and immediately placed on dry ice. Samples
were held at -80 °C until processed.
Samples were grouped into three stages based on the length of their conditioning period.
A subset of samples was subjected to histology. Histology and personal observations during
sampling were used in conjunction to inform the placement of division between reproductive
stages. Stage I (weeks 1-3) were early maturation samples with few if any eggs present. Stage II
(weeks 4-7) were early vitellogenesis samples and stage III (weeks 8-10) were late vitellogenesis
samples. Resorbing gonad was noticed in stage 10 samples, which marked the end of the
conditioning program.
RNA extraction and cDNA synthesis of total and poly(A) RNA
Total RNA was extracted using TRIzol® Plus RNA Purification Kit (Invitrogen) and a
TissueLyser II (Qiagen). RNA samples were treated with DNAse I (1 U/µg total RNA,
62

Promega) to remove genomic DNA prior to reverse transcription. Quantification and purity were
assessed with a NanoDrop® ND-1000. Samples with 260/280 and 260/230 ratios below 1.8
were excluded from further analysis. One µg total RNA was used for synthesis of first strand
cDNA using M-MLV reverse transcriptase (Promega). One pg of luciferase (luc) mRNA was
added to each RT reaction. Reverse transcription reactions were carried out at 42 °C for
oligo(dT)
15
and 37 °C for random hexamer priming. RNasin® (promega) was included in all
RT reactions to inhibit RNase activity. The total volume for RT reactions was 25 µl.
Measurement of gene expression by qPCR
Forty-five samples, evenly distributed over 10 weeks of conditioning, were screened for
gene expression levels by qPCR. The 12 genes consisted of 7 which had previously been found
to be differentially expressed over gonad maturation (forkhead box protein L1, foxl1;
thioredoxin reductase-2, thioR; vitellogenin-6, vit6; pancreatic lipase-2, lipase;
hemagglutinin/amoebocyte aggregation factor, haaf; transforming growth factor-β, tgfb; oyster
vasa-like gene, vasa) and 5 commonly used housekeeping genes (α-actin, actin; β-tubulin,
tubulin; 18s ribosomal RNA, 18s; glyceraldehyde 3-phosphate dehydrogenase, gapdh; heat
shock protein 70, hsp70). Primers were developed using Primer3 (http://bioinfo.ut.ee/primer3-
0.4.0/). Sequences were blasted for uniqueness and specificity. Details of primer set can be seen
in Table 1. Amplified sequences were sized on agarose gels, and homogeneity was assessed by
melt curve analysis following amplification.
Dilution series were performed for all primer sets to determine the template range with
PCR efficiency of R
2
= 1 +/- 0.05. Amplification reactions contained 5 ng template, 600 nm
primer concentration, and 1 X Brilliant II SYBR green QPCR master mix (Agilent) in a total
volume of 15µl. All samples were run in duplicate. NoRT controls were tested prior to
63

measuring of gene expression. IRC and NTC controls were included for each gene of each qPCR
run. Expression of target genes was normalized by luc expression. All statistical analyses were
performed using R v3.0.1. Differences in expression between priming strategies for a single
stage and gene were tested by paired t-test. Differences in the relative change in expression in
progressing reproductive stages between priming strategies were tested by unpaired t-tests.
Stage-specific expression differences for both total and poly(A) RNA were tested by ANOVA
followed by post-huc tukey tests.

64

Results
Differences in efficiency of reverse transcription between oligo(dt) and random hexamer
priming were normalized by spiking reactions with the exogenous polyadenylated luciferase
control RNA. Transcript abundances between priming strategies were then compared at each
stage. The random hexamer primed abundance represents the total RNA abundance of a given
transcript, while the oligo(dt) primed abundance represents only those RNA transcripts with a
poly(A) tail. Expression differences at a single stage (assumed oligo(dt) expression ≤ random
hexamer expression) suggest the presence of deadenylated transcripts, while changes in the ratio
of the two abundances among reproductive stages could be due to either deadenylation of
polyadenylated transcripts or vice versa, depending on the direction of change.
Presence or absence of deadenylation was first investigated by looking at the ratio of
random hexamer primed to oligo(dt) primed expression at each stage, individually. Figure 1
shows such ratios for the 12 genes investigated. A ratio near one suggests a relative lack of
deadenylated transcripts. Two common patterns emerged among the ratios. One group of genes
(lipase, haaf, foxl1, tubulin) had ratios near one for each of the three stages, suggesting a relative
absence of deadenylated transcripts. Another group (vit-6, thioR, vasa, actin, tgf-b) had ratios
that varied between two and five, suggesting the presence of some amount of deadenylation.
Two genes, hsp70 and 18s, possessed noticeably more deadenylated transcripts. 18s by far
displayed the greatest amount of deadenylated transcripts, where the ratio in stage III samples
was nearly 250.
Differences in the level of expression between oligo(dt) and random hexamer primed
cDNA were found for 9 of the 12 genes in at least one stage. The three genes without significant
differences at any stage were lipase, haaf, and foxl1. Of the 9 with differences, three had
65

expression differences at all three stages (vasa, hsp70, 18s). Analysis of variance was followed
by Tukey test was performed for each gene and priming combination, over the three reproductive
stages. Four of the 12 genes (vit, tgf-b, vasa, gapdh) displayed different patterns of stage-
specific expression between oligo(dt) and random hexamer primed sample preparation.
To look more carefully whether differences in the amount of deadenylation/polyadenylation
were occurring during gonad maturation or stable throughout, the expression of stage II and III
samples relative to the stage I gonad samples were compared between priming methods. Stage
III was additionally compared relative to stage II (Figure 2). Significant differences in the
relative expression among stages between the priming methods were found in 6 of the 12 genes
(vit, thioR, vasa, tgf-b, gapdh, hsp70). In 5 of the 6, the change in relative proportions of random
hexamer and oligo(dt) primed expression came between stages I and II. For hsp70, the change in
expression came between stages II and III, which was due to a relative decrease in poly(A)
transcript abundance.
The pattern of change in the 5 genes which differed between stages I and II were of two
types. The reduction in expression of total RNA between stages I and II was greater than the
reduction of poly(A) RNA for gapdh, tgf-b and vit. Alternatively, for thioR and vasa, the
increase in poly(A) RNA between stages I and II was greater than the increase in total RNA
between these stages. Though only marginally significant, a similar pattern was seen for foxl1
expression between stages II and III.

66

Discussion
Differences in the amount of total and poly(A) RNA were compared by normalizing expression
to the exogenous reference gene luciferase. This control RNA comes with a 30 bp poly(A) tail -
and the ratio of luciferase expression between random primed and oligo(dt) primed samples was
not different than one. Of the 12 genes analyzed, only three also had ratios near one, suggesting
the majority of the mRNA present for these three genes was polyadenylated (Figure 1). Ratios
greater than one suggest the presence of active degradation or deadenylation.
Differences in the access of oligo(dT) and random hexamer oligos to the their priming
site could potentially affect interpretation of poly(A) and total RNA levels. However, RNA
samples were denatured prior to reverse transcription, which should provide the two primers
types equal opportunity for access to polyadenylated transcripts. That the ratio of expression
levels for random hexamer and oligo(dT) primed luciferase (all transcripts are polyadenylated)
was not statistically different than one supports the idea that a transcript is equally likely to be
primed by random hexamer or oligo(dt) primers.
The inclusion of 18s in this report acted as a sort of control. It is generally believed that
RNA not destined for translation lack poly(A) tails. And while evidence of the contrary is
increasingly being reported (Kuai et al. 2004, Slomovic et al. 2006), the majority of 18s
transcripts do lack poly(A) tails. Here, I found on average nearly 200 copies of deadenylated 18s
to every one polyadenylated copy (Figure 1). Ribosomal RNA accounts for 65-70% of the total
RNA in mouse oocytes, and increases in quantity as oocyte maturation progresses (Kaplan et al.
1982). The amount of non-poly(A) 18s RNA was similarly found to increase in C. gigas oocytes
from stage II to stage III.
67

The molecular chaperone protein hsp70 also had elevated relative levels of deadenylated
transcripts when compared to genes other than 18s (Figure 1). In Drosophila, it has been
reported that after a heat shock event, rapid deadenylation resulted in roughly 40% of hsp70
transcripts lacking poly(A) tails (Dellavalle et al. 1994). Stage-specific expression was found
here to decrease as maturation increased for both total and poly(A) RNA, which is consistent
with results found in mice amphipods (Manajwala et al. 1991, Schirling et al. 2004). Hsp70
levels were probed by ELISA in C. gigas, where levels of hsp70 were shown to increase during
maturation (Meistertzheim et al. 2009). This discrepancy could be explained by
methodologically-driven specificity differences, given that C. gigas has 88 different hsp70 genes.
Aside from 18s and hsp70, the poly(A) RNA represented between 20-100% of the total
RNA present, and seemed to fall into two groups. Some genes had nearly all transcripts
polyadenylated, while the others had between 20-50% polyadenylated (Figure 1). When
considering changes in the relative proportion of poly(A) RNA between reproductive stages, the
presence of deadenylated transcripts allows for the possibility of changes being due to
deadenylation and readenylation. When the abundance of total and poly(A) mRNA are roughly
equal at a given stage, the only change that could occur is the deadenylation of existing or future
transcribed mRNA.
The 6 genes with significant changes in the relative proportion of poly(A) to total RNA
among reproductive stages all displayed considerable deadenylation in stage I (Figure 1, 2).
Grouping these 6 by the time and direction of change of poly(A) RNA between stages helps to
understand the significance of the changes in the relative proportion of poly(A) RNA. For
example, changes occurring between stages I and II may be related to oocyte growth while
changes between stages II and III may be related to oocyte homeostasis or maternal mRNA
68

storage. And changes in the relative proportion of poly(A) RNA may be due to the mobilization
of deadenylated transcripts or the deadenylation of newly transcribed mRNA entering the
cytoplasm.
Three genes (gapdh, tgfb, vit) displayed a significant decrease in the relative proportion
of total RNA to poly(A) RNA between stages I and II (Figure 2). This pattern could be
explained by the polyadenylation or degradation of existing deadenylated transcripts. In all three
cases, there was a significant decrease in total and poly(A) RNA expression between stages I and
II, but the decrease in poly(A) RNA was significantly less than the decrease in total RNA.
Previous studies of vit and tgfb in C. gigas do shed some light on their expression patterns. Both
genes are expressed primarily in somatic follicle cells adjacent to growing oocytes (Matsumoto
et al. 2003, Fleury et al. 2008). From here they enter oocytes by endocytosis either directly from
the follicle cells or from the surrounding hemolymph. The translated products of both genes
accumulate in growing oocytes roughly at the rate of increasing oocyte cell diameter (Li et al.
1988, Corporeau et al. 2011). It is not clear whether the uptake of vit by oocytes includes mRNA
in addition to synthesized yolk protein. Follicle cells or oocytes may be storing vit and tgfb
deadenylated transcripts during stage I as they accumulate faster than they are able to be
processed.
Two other genes, thioR and vasa, also exhibited a relative increase in the proportion of
poly(A) to total RNA between stages I and II, but in this case the change was coincident with an
overall increase in transcription (Figure 2). Interpretation of this pattern suggests the poly(A)
pool increased both due an increase in transcription and due to polyadenylation of deadenylated
transcripts. And finally, hsp70 showed a decrease in poly(A) RNA relative to total RNA
between stages II and III. This was the only gene of the 6 which displayed significant
69

developmental deadenylation (though not significant, 18s appears to be as well), rather than the
mobilization of previously deadenylated transcripts. In mice, hsp70 is one of the first genes
expressed after EGA (Bensaude et al. 1983). The results here suggest hsp70 transcripts are
stored as deadenylated maternal mRNAs in mature oocytes, which may be important for events
prior to EGA.
Mature oocytes are thought to be largely transcriptionally silent and are known to occupy
the vast majority of volume in mature female C. gigas gonad tissue. Taken together, this leads to
common stage-specific differential gene expression driven by somatic cell expression and oocyte
maturation factors, including stored maternal mRNAs. Even so-called “housekeeping” genes
were shown here to be subject to stage-specific expression patterns of both total and poly(A)
RNA. The seemingly strong influence of oocyte transcription rates and
polyadenylation/deadenylation patterns on stage-specific expression patterns may be reduced if
reverse-transcription used cell numbers rather than RNA quantity as input. The reduction in
expression seen in most genes over gonad maturation may be influenced by competitive
exclusion from accumulating rRNA in maturing oocytes. Deadenylated rRNA transcripts appear
to be accumulating here with oocyte maturation (Figure 2), and have been reported to account for
65-70% of total rRNA in mouse oocytes (Kaplan et al. 1982). This effect of reduced expression
due to competition with rRNAs would be alleviated by loading cell numbers as reverse-
transcription input rather than RNA quantity.
The factors involved in the control of poly(A) tail length are fairly well understood
(Reviewed in Weill et al. 2012). Recent work is beginning to address the regulation of such
factors. Alternative polyadenylation (APA) has recently been discovered to be pervasive,
occurring in 70% of human genes (Shi 2012). APA results in variation in 3’UTRs, a
70

characteristic which has been directly targeted by RNA-Seq due to its observed role in post-
transcriptional gene regulation (Jan et al 2011). Ulitsky and colleagues, working with zebrafish,
observed a direct correlation between 3’UTR length and transcript stability in oocytes (Ulitsky et
al. 2012). They suggest a possible role for APA in essentially assigning the same transcript to
different tasks at alternative developmental processes. Such an analysis holds potential for
learning more about patterns of polyadenylation/deadenylation during C. gigas gonad
maturation.
Consideration of the data generating platform and normalization strategies are important
when interpreting gene expression data from different sources. This becomes even more
paramount when looking at developing tissues, such as gonad maturation and embryonic
development. Here I show the priming strategy used for cDNA generation should be considered
when interpreting gene expression data during gonad maturation. Six of the 12 genes looked at
displayed differences in the relative amount of poly(A) to total RNA among different maturation
stages. Four of the 12 genes would have different stage-specific expression patterns if the data
for each priming strategy were analyzed separately then compared with each other. For example,
vit expression of oligo(dt) primed cDNA was significantly different between stages II and III, but
not between stages I and II. The exact opposite pattern was seen in random hexamer primed
cDNA. Some caution should be taken when considering expression of the two priming strategies
independently rather than relative to each other, because normalization to luc does not account
for variations in sample input. That being said, cDNA was quantified and added uniformly to
reverse transcription reactions. The results here suggest differential deadenylation and
polyadenylation are commonplace during female gonad maturation in C. gigas.

71

References
Bensaude, O., Babinet, C., Morange, M., Jacob, F., 1983. Heat-shock proteins, 1st major
products of zygotic gene activity in mouse embryo. Nature, 305, 331-333.
Colas, P. and Dube, F., 1998. Meiotic maturation in mollusc oocytes. Semin. Cell Dev. Biol., 9,
539-548.
Corporeau, C., Groisillier, A., Jeudy, A., Barbeyron, T., Fleury, E., Fabioux, C., Czjzek, M.,
Huvet, A., 2011. A functional study of transforming growth factor-beta from the gonad of
Pacific oyster Crassostrea gigas. Mar. Biotechnol., 13, 971-980.
Dellavalle R.P., Petersen R., Lindquist S., 1994. Preferential deadenylation of Hsp7O mRNA
plays a key role in regulating Hsp7O expression in Drosophila melanogaster. Mol. Cell.
Biol., 14, 3646-3659.
Fleury, E., Fabioux, C., Lelong, C., Favrel, P., Huvet, A., 2008. Characterization of a gonad-
specific transforming growth factor-beta superfamily member differentially expressed
during the reproductive cycle of the oyster Crassostrea gigas. Gene, 410, 187-196.
Hara, K., Oda, S., Naito, K., Nagata, M., Schultz, R., Aoki, F., 2005. Cyclin A2-CDK2 regulates
embryonic gene activation in 1-cell mouse embryos. Dev. Biol., 286, 102-113.
Jacobson, A., 1986. Translational control. In: Hershey, J.W.B., Mathews, M.B., Sonenberg, N.
(Ed.), Cold Sping Harbor Laboratory, 451-480.
Jan, C.H., Friedman, R.C., Ruby, J.G., Bartel, D.P., 2011. Formation, regulation and evolution of
Caenorhabditis elegans 3 ' UTRs. Nature, 469, 97-U114.
Kaplan, G., Abreu, S.L., Bachvarova, R., 1982. rRNA accumulation and protein synthetic
patterns in growing mouse oocytes. J. Exp. Zool., 220, 361-370.
72

Kuai, L., Fang, F., Butler, J.S., Sherman, F., 2004. Polyadenylation of rRNA in Saccharomyces
cerevisiae. P. Natl. Acad. Sci. USA, 101, 8581-8586.
Li, Q., Osada, M., Suzuki, T., Mori, K., 1998. Changes in vitellin during oogenesis and effect of
estradiol-17 beta on vitellogenesis in the Pacific oyster Crassostrea gigas. Invertebr.
Reprod. Dev., 33, 87-93.
Manejwala, F.M., Logan, C.Y., Schultz, R.M., 1991. Regulation of hsp70 mRNA levels during
oocyte maturation and zygotic gene activation in the mouse. Dev. Biol., 144, 301-308.
Matsumoto, T., Nakamura, A., Mori, K., Kayano, T., 2003. Molecular characterization of a
cDNA encoding putative vitellogenin from the Pacific oyster Crassostrea gigas. Zool. Sci.,
20, 37-42.
Meistertzheim, A.L., Lejart, M., Le Goïc, N., Thébault, M.T., 2009. Sex-, gametogenesis, and
tidal height-related differences in levels of HSP70 and metallothioneins in the Pacific oyster
Crassostrea gigas. Comp. Biochem. Phys. A, 152, 234-239.
Paynton, B., Rempel, R., Bachvarova, R., 1988. Changes in state of adenylation and time course
of degradation of maternal messenger-rnas during oocyte maturation and early embryonic-
development in the mouse. Dev. Biol., 129, 304-314.
Piko, L. and Clegg, K., 1982. Quantitative changes in total rna, total poly(A), and ribosomes in
early mouse embryos. Dev. Biol., 89, 362-378.
Richter, J.D. and Lasko, P., 2011. Translational control in oocyte development. Cold Spring
Harb. Perspect. Biol., 3, a002758.
73

Rosenthal, E., Tansey, T., Ruderman, J., 1983. Sequence-specific adenylations and
deadenylations accompany changes in the translation of maternal messenger-rna after
fertilization of Spisula oocytes. J. Mol. Biol., 166, 309-327.
Schirling, M., Triebskorn, R., Kohler, H., 2004. Variation in stress protein levels (hsp70 and
hsp90) in relation to oocyte development in Gammarus fossarum (koch 1835). Invertebr.
Reprod. Dev., 45, 161-167.
Shi, Y., 2012. Alternative polyadenylation: New insights from global analyses. RNA., 18, 2105-
2117.
Slomovic, S., Laufer, D., Geiger, D., Schuster, G., 2006. Polyadenylation of ribosomal RNA in
human cells. Nucleic Acids Res., 34, 2966-2975.
Ulitsky, I., Shkumatava, A., Jan, C.H., Subtelny, A.O., Koppstein, D., Bell, G.W., Sive, H.,
Bartel, D.P., 2012. Extensive alternative polyadenylation during zebrafish development.
Genome Res., 22, 2054-2066.
Weill, L., Belloc, E., Bava, F., Mendez, R., 2012. Translational control by changes in poly(A)
tail length: Recycling mRNAs. Nat. Struct. Mol. Biol., 19, 577-585.
Wickens, M., 1990. In the beginning is the end - regulation of poly(A) addition and removal
during early development. Trends Biochem. Sci., 15, 320-324.
Zhang, G., Fang, X., Guo, X., Li, L., Luo, R., Xu, F., Yang, P., Zhang, L., Wang, X., Qi, H.,
Xiong, Z., Que, H., Xie, Y., Holland, P.W.H., Paps, J., Zhu, Y., Wu, F., Chen, Y., Wang, J.,
Peng, C., Meng, J., Yang, L., Liu, J., Wen, B., Zhang, N., Huang, Z., Zhu, Q., Feng, Y.,
Mount, A., Hedgecock, D., Xu, Z., Liu, Y., Domazet-Loso, T., Du, Y., Sun, X., Zhang, S.,
Liu, B., Cheng, P., Jiang, X., Li, J., Fan, D., Wang, W., Fu, W., Wang, T., Wang, B., Zhang,
J., Peng, Z., Li, Y., Li, N., Wang, J., Chen, M., He, Y., Tan, F., Song, X., Zheng, Q., Huang,
74

R., Yang, H., Du, X., Chen, L., Yang, M., Gaffney, P.M., Wang, S., Luo, L., She, Z., Ming,
Y., Huang, W., Zhang, S., Huang, B., Zhang, Y., Qu, T., Ni, P., Miao, G., Wang, J., Wang,
Q., Steinberg, C.E.W., Wang, H., Li, N., Qian, L., Zhang, G., Li, Y., Yang, H., Liu, X.,
Wang, J., Yin, Y., Wang, J., 2012. The oyster genome reveals stress adaptation and
complexity of shell formation. Nature, 490, 49-54.

75

Tables

Table 1
Gene names and primer sequences of 10 candidate reference genes used for qPCR
Gene Name Abbreviation Accession number Primer Sequence
Luciferase luc X65316.2 5'-AGAGATACGCCCTGGTTCCT-3'
5'-ATAAATAACGCGCCCAACAC-3'
Vitellogenin-6 vit AB084783 5'-GAAGTTTACGTGAGAAGAC-3'
5'-TAGAAGTCTCTGAAGAACAG-3'
Pancreatic lipase-2 lipase AM857075 5'-ATTCACGGCGTTTCCCTGT-3'
5'-CATTCTGTTGCAGCCGTTCT-3'
Hemagglutinin/amebocyte aggregation factor haaf CU995762 5'-TGGTAGTCTCGGGATGGAAG-3'
5'-CGAAGGACAGGTCTGAAGGA-3'
Forkhead box protein l1 foxl1 AM865563 5'-TGCACATACCACATCCCAAC-3'
5'-TGCGAGTAACTCACGGAAGG-3'
Thioredoxin reductase 2, mitochondrial thioR AM865969 5'-GGACTGAAGTCCGTCAAGCA-3'
5'-AAACACGCCCACCTCAAAG-3'
Oyster vasa-like gene vasa AY423380 5'-GGCAAGGCCATCAGTTTCTA-3'
5'-TAGCCAGAGCACCATCAGTG-3'
Transforming growth factor-β tgfb EF563990 5'-TTGGACATCAGGGAAATTCTG-3'
5'-CCAAACGAAACGACAGGAAC-3'
Heat shock protein 70 hsp70 AB122064 5'-TGTGAAAGGGCAAAGAGGAC-3'
5'-CTCAAACCTAGCCCTGGTGA-3'
β-actin actin AF026063 5'-ATCTCGCTGGACGTGATCTG-3'
5'-GGCTGTGGTGGTGAAAGAGT-3'
α-tubulin tubulin AB196533 5'-GGAGAGGGAATGGAAGAAGG-3'
5'-CAGCCTCTCCCTCAACAGAA-3'
18s rRNA 18s AB064942 5'-GCGTTTAGCCACACGAGATT-3'
5'-ATCCCTAGCACGAAGGAGGT-3'
Glyceraldehyde-3 phosphate-dehydrogenase gapdh AJ544886 5'-CGCCAATCCTTGTTGCTT-3'
5'-TTGTCTTGCCCCTCTTGC-3'

76

Figures

Figure 1. Ratio of random hexamer primed transcripts to oligo(dt) primed transcripts for each of
the three reproductive stages. A ratio greater than one suggests the presence of deadenylated
transcripts. Note y-axes may be different for each gene.
Figure 2. Stage-specific gene expression relative to the mean expression of stage I (stage I –
pink, stage II – green, stage III – blue) for random and oligo(dt) primed rt-qpcr. Error bars +/-
SEM. Stars represent significant differences between oligo(dt) and random primed expression in
the ratio of the stage indicated relative to the previous stage.

77

Figure 1

78

Figure 2

79

Chapter 4: Selection of reference genes for normalization of qPCR
data from multiple stages of gonad maturation in the Pacific oyster
Crassostrea gigas

Abstract
Relative quantification of qPCR data relies upon stable expression of one or more
housekeeping genes. Proper validation of housekeeping gene stability is often not undertaken,
even though it is unlikely that any gene is stably expressed at all times in every tissue. This trend
is especially alarming in developing tissues, where stable expression of any gene is less likely.
The stability of ten commonly used housekeeping genes was screened among male and female
individuals at varying stages of gonad maturation in Crassostrea gigas. The assignment of best
housekeeping gene combinations differed between the different measures of stability employed.
Regardless of the stability measure, unique sex-specific best housekeeping gene combinations
were identified in testis and ovary tissue. Ribosomal and microtubule related genes proved
particularly poorly suited as housekeeping genes in developing C. gigas gonad tissue. Sex-
specific housekeeping genes were identified here. That the best housekeeping genes differed
between sexes highlights the need for a thorough analysis of housekeeping gene stability prior to
normalization of qPCR data in developing reproductive tissue.

80

Introduction
Factors affecting larval survival can be crudely divided between prezygotic and
postzygotic in nature. In the Pacific oyster Crassostrea gigas, in which direct observation of
gonad material is difficult, prezygotic factors such as egg quality or variation in timing of gonad
maturation between parents have been demonstrated (Valdez Ramirez et al. 1999, Boudry et al.
2002, Lannan et al. 1980). The present inability to assess quickly and accurately the
developmental state of gonad and quality of gametes prior to the sacrificing of animals for
spawning has negative consequences in this commercially important species. Screening of
biological material for processes involved in gonad maturation represents a potential mechanism
for assessing reproductive state and or gamete quality.
Reverse transcription quantitative PCR (RT-qPCR) has become a standard method for
assessing the quantity of nucleic acid sequences due to the high sensitivity, relative ease of use,
and speed of generating results. Successful RT-qPCR relies on proper experimental design,
sample quality, assay design, and normalization (Derveaux et al. 2010). Normalization is
necessary to remove variation in gene expression due to differences among samples in reverse
transcription efficiency and in quantity of starting nucleic acid material. Successful
normalization allows observation of true biological variation in the expression of a gene of
interest.
Internal control genes, otherwise called reference genes or housekeeping genes, are most
commonly used to normalize the expression of a gene of interest. In theory, these genes are
stably expressed among samples and treatments, providing normalized expression by taking the
ratio of the gene of interest to the reference gene. Though in practice, reference gene expression
is often not stable, and it is strongly recommended that multiple reference genes be used to
81

reduce normalization error, and that they be experimentally validated prior to use (Bustin et al.
2009). Multiple software applications have been developed to assist in the selection of multiple
reference genes (Vandesompele et al. 2002, Andersen et al. 2004, Pfaffl et al. 2004).
Greater effort has been made of late to validate reference genes for particular
experimental designs as the stability of commonly used candidates and implications of their use
has been called into question (Vandesompele et al. 2002, Dheda et al. 2005), including several
examples in bivalve molluscs (Cubero-Leon et al. 2012, Du et al. 2012, Mauriz et al. 2012). The
selection of reference genes during gonad expression in the scallop Pecten maximus was found to
be sex-specific, with a different combination of reference genes each providing the best
normalization for male and female gonad tissue, respectively (Mauriz et al. 2012). While it may
not be surprising to find differences in expression of reference genes between sex-specific
tissues, it does highlight the need to consider each experimental design individually when
choosing a normalization strategy.
Normalization of expression over multiple developmental stages of oogenesis requires
additional care. Maternal loading of mRNA in a suppressed state into developing oocytes
(Krauchunas and Wolfner 2013) creates the potential for unstable abundances of reference genes
in the developing female gonad. Mature oocytes prior to egg activation are transcriptionally
silent, which creates a similar potential for unstable reference gene abundance since eggs make
up the majority of mature gonad by volume. The use of 18S rRNA, a commonly used
housekeeping gene, has been questioned based on difficulties arising from its high abundance
(Vandesompele et al. 2002), and due to its accumulating presence as maternal mRNA in
developing oocytes (Kroupova et al. 2011, Deloffre et al. 2012). Sex and stage-specific variation
82

in gene expression makes the selection of effective housekeeping genes a necessity when using
qPCR to investigate gonad maturation.
The present study aimed to identify housekeeping genes for normalization of gene
expression during gonad maturation in the commercially important C. gigas. Twenty males and
twenty females, sampled over 8 weeks of a hatchery based conditioning system, were screened
for expression of 10 commonly used housekeeping genes. The best combination of reference
genes was consistent between testis and gonad, while a distinct combination had the greatest
stability in ovary tissue. Reference gene expression was considerably more stable in ovary than
in testis tissue.

83

Materials and Methods
Sample collection
Animals from two inbred lines and the hybrid crosses of the two parental inbred lines,
produced in 2007, were used for artificial conditioning, which began in early March, 2010, at or
near the timing of natural initiation of gametogenesis. Conditioning was carried out at a constant
25 °C, after increasing the water temperature 2 °C day
-1
from the ambient incurrent water
temperature of 13 °C. Oysters were held in flow-through tanks with bubbled O
2
and fed a live
mixed algal diet.
Tissue samples were taken weekly by shucking and sectioning animals just ventral to the
labial palp. Gonad tissue was then sampled from the anterior half of the ventral section and
immediately placed on dry ice. After collection, samples were transferred to -80 °C until
processed. Twenty males and twenty females spanning 8 weeks of conditioning were used for
screening of gene expression analysis of candidate housekeeping genes.
RNA extraction and cDNA synthesis
Total RNA was extracted using the TRIzol® Plus RNA Purification Kit (Invitrogen) and
a TissueLyser II (Qiagen). RNA samples were treated with DNAse I (1 U/µg total RNA,
Promega) to remove genomic DNA prior to reverse transcription. Quantification and purity were
assessed with a NanoDrop® ND-1000. Samples with 260/280 and 260/230 ratios below 1.8
were excluded from further analysis. One µg total RNA was used for synthesis of first strand
cDNA using M-MLV reverse transcriptase (Promega) with oligo(dT)
15
priming. RNasin®
(promega) was included in all RT reactions to inhibit RNase activity. The total reaction volume
for RT reactions was 25 µl.
84

Selection of candidate housekeeping genes and primer design
Ten candidate reference genes were chosen for screening of gene expression stability.
Actin, Elongation factor-1 (elf-1), glyceraldehyde-3 phosphate-dehydrogenase (gapdh), 18s
rRNA (18s), and 28s rRNA (28s) are commonly used for normalization of gene expression in C.
gigas (Huvet et al. 2003, Fabioux et al. 2004, Herpin et al. 2007, David et al. 2005,
Meistertzheim et al. 2007). Ribosomal protein L7 (rpl7) and ribosomal protein RS18 (rs18)
were recently found to be the most stable reference genes in herpes virus infected C. gigas larvae
(Du et al. 2013). Cytochrome C oxidase subunit 1 and α-tubulin are also commonly used
housekeeping genes, while heat shock protein 70 (hsp70) is known to serve cellular
housekeeping functions, and is occasionally used for gene expression normalization.
Primers for gapdh, rpl7, rs18, and 28s were derived from previous studies (Dheilly et al. 2011,
Du et al. 2013), while primers for the other six genes were designed using primer3 (v. 0.4.0).
Specificity of designed primers was confirmed by BLAST. Primer sequences can be seen in
Table 1.
Real-time RT-PCR and analysis of expression stability
Real time PCR was performed on a Stratagene Mx3005p instrument with MxPro v4.10
software. Dilution series were performed for all primer sets to determine the template range with
PCR efficiency of R
2
= 1 +/- 0.05. Amplification reactions contained 5 ng template, 600 nm
primer concentration, and 1 X Brilliant II SYBR green QPCR master mix (Agilent) in a total
volume of 15µl. All samples were run in duplicate with no template controls in each plate.
After an initial 10 min at 95 °C to activate the polymerase, the cycling conditions were 95 °C for
30 s, 60 °C for 1 min, then 72 °C for 30 s. Following 40 cycles, a melt curve was performed.
85

Threshold cycle (Ct) data were extracted from MxPro software and stability of the 10
candidate reference genes were evaluated with GeNorm v3.5 (Vandesompele et al. 2002) and
Normfinder v0.953 (Andersen et al. 2004). GeNorm selects the best pair of reference genes as
those whose expression ratios are the most stable across samples. Normfinder allows the
incorporation of sample subgroups during the stability analysis and is thought to be less
susceptible to rewarding co-regulation of candidate housekeeping genes (Anderson et al. 2004).
For both methods, the result is a stability value that represents systematic error, where a lower
stability value represents a higher stability of expression.

86

Results
Sex-specific expression levels of housekeeping genes
In general, expression levels of housekeeping genes were greater in ovary tissue than in
testis tissue (Figure 1). The average difference in Ct values between sexes for the 10 candidate
housekeeping genes was 2.22, or roughly a 5-fold difference. Only α-tubulin had a greater
average expression in males than in females. The variance in expression levels was greater on
average in males than in females, with only β-actin being more variable in ovary than in testis
tissue.
Expression of housekeeping genes in ovary
Normfinder and GeNorm identified different pairs of genes as the best for normalization
(Figure 2). Rpl7 and rs18 were found to be the best pair of housekeeping genes by GeNorm
while hsp70 was determined to be the most stable, followed by cox, by NormFinder. The
stability value (M) of a gene calculated by GeNorm is recommended to be under 1.5 for
acceptable normalization. Only 18s and 28s had M values near the 1.5 level, while the other
eight genes have values near or less than 1.0. Though Normfinder and GeNorm produced
inconsistent results in terms of the most stable reference genes, they were consistent in
identifying the least stable genes in ovary tissue as 18s and 28s.
Expression of housekeeping genes in testis
Among male testis tissue, there was greater agreement between GeNorm and Normfinder
for the most stable reference genes (Figure 3). Hsp70 was identified by both methods as the
most stable reference gene. The combination of hsp70 and gapdh was selected by GeNorm as
the most stable pair, while elf1 was second most stable gene as identified by NormFinder.
Similar to the analysis of ovary tissue, there was agreement between the two methods for the
87

least stable reference genes, this time being tubulin and 18s. Only gapdh, hsp70, and elf1 had M
values under the recommended threshold level of 1.5.
Expression of housekeeping genes in gonad
When treating testis and ovary together, hsp70 and gapdh were the best combination of
genes found by GeNorm, while hsp70 followed by cox were the most stable genes found by
NormFinder (Figure 4). Though tubulin was found to be the least stable and hsp70 the most
stable, there otherwise was not a lot of consistency in the stability order of the 10 genes between
GeNorm and NormFinder. Overall, hsp70 was the most stable reference gene. Of the 6
analyses, only the GeNorm method for ovary tissue did not identify hsp70 as the most stable
reference gene. Tubulin along with the ribosomal proteins (rpl7, rs18), and the ribosomal RNA
genes (18s, 28s), were the least suited for use as housekeeping genes. In addition to hsp70,
gapdh and cox appear the best candidates for use as housekeeping genes during gonad
maturation in C. gigas.

88

Discussion
Developing organs have high levels of variation in gene expression, potentially even in so
called “housekeeping” genes, which necessitates validation of candidate normalizing genes prior
to the screening of expression in a gene of interest by qPCR. Differences between male and
female sex organs during gonad maturation further complicate normalization. Significant
differences between sexes in expression levels of commonly used housekeeping genes were
observed here. Previous studies aimed at selecting housekeeping genes for qPCR normalization
in shellfish have identified different reference genes to be best suited for normalization in each
sex (Leelatanawit et al. 2012, Mauriz et al. 2012). Here, the choice of best reference genes for
males and females was dependent on which method of analysis was adopted. While NormFinder
was consistent in selecting hsp70 as the best primary reference gene, GeNorm selected a distinct
pair of reference genes for each sex.
Variation in expression of reference genes was greater on average in testis than in ovary.
This pattern could be due error introduced through differences in mRNA input or reverse
transcription efficiencies. Alternatively, since the mature gonad is composed primary of sex
cells (Eckelbarger and Davis 1996a,b, Franco et al. 2008), the smaller variation in ovary tissue
may be due to maternal loading of mRNA in oocytes (Krauchunas and Wolfner 2013). Along
these lines, the greater variation in tubulin expression in males was not unexpected given the
great amount of sperm cells produced and the role of tubulin in sperm flagella (Kingtong et al.
2013).
The use of rRNA genes for normalization has been questioned due simply to the high
expression levels (Vandesompele et al. 2002). In the roach Rutilus rutilus, where RNA species
were studied over the course of gametogenesis, the presence of 18s and 28s rRNA was found to
89

increase significantly from the primary growth stage to the late vitellogenic stage (Kroupova et
al. 2011). By the late vitellogenic stage, 18s and 28s rRNA dominated the total RNA pool,
which suggests a dynamic role in oocyte maturity and raises questions about the stability of their
expression. In this study, 18s and 28s were found to be the least stable reference genes in ovary
tissue, and among the least stable in testis and combined gonad tissue. These results are not fully
consistent with other reference genes stability analyses of stages of bivalve gonad maturation
(Cubero-Leon et al. 2012, Mauriz et al. 2012), though 18s and 28s rRNA genes were identified
as the least stable of 15 candidates during larval development in C. gigas (Du et al. 2013). In
any event, it seems caution should be taken before using rRNA genes for normalizing expression
in developing tissue.
One of the more interesting results was the identification of rpl7 and rs18 as the best
combination of reference genes for ovary tissue by the GeNorm method. This combination of
reference genes was also identified by GeNorm as the best pair of reference genes in developing
C. gigas larvae (Du et al. 2013). That the best pair of reference genes in developing larvae
would be consistent with ovary tissue but not testis tissue raises the possibility that the result is
due to the maternal mRNA between oocyte and larva. However, it appears more likely the
selection of rpl7 and rs18 is due to the co-regulation of expression between these two genes,
which are both ribosomal proteins with roles in protein translation.
One of the proclaimed strengths of NormFinder is its ability to show less sensitivity to
co-regulation of candidate reference genes (Anderson et al. 2004). NormFinder takes a model-
based approach that can treat sample subgroups (here samples were grouped by length of
reproductive conditioning). Gene stability is then ranked by the minimal combined intra- and
intergroup variation. GeNorm, on the other hand, chooses the best pair of reference genes based
90

on the stability measure M, which is calculated as the lowest average pairwise variation of a
reference gene with all other reference genes (Vandesompele et al. 2002). The difference
between the methods leaves GeNorm more exposed to selecting co-regulated genes, as appears
the case here in developing ovary tissue. Rs18 and rpl7 are both ribosomal proteins (small and
large subunit, respectively) which are likely to experience co-regulated expression, though may
not necessarily be the most stable genes across all sample types. Hsp70 and cox were selected as
the two most stable reference genes by NormFinder in ovary tissue, while rpl7 and rs18 were
ranked 6
th
and 8
th
most stable, respectively. This example highlights the importance of selecting
reference genes that do not share cellular function.
NormFinder and geNorm rank screened reference genes based on based on stability down
to the best single gene (NormFinder) or pair of genes (geNorm) for normalization. The choice of
how many reference genes to use is then left to the discretion of the researcher. Multiple criteria
have been proposed to determine the best number of reference genes to use. If a stability
analysis cannot be performed due to cost or time issues, Vandesompele et al. (2002) suggest at
least three reference genes should be used. GeNorm includes a method for selecting the optimal
number of reference genes to use based on a pairwise variation analysis which assesses the
change in stability value of n reference genes caused by adding additional reference genes
(Vandesompele et al. 2002). The authors selected an arbitrary value of 0.15 for the value where
adding more reference genes was not needed. Another group argued that the minimal pairwise
value represented the most stable normalizing factor, and therefore the number of reference
genes which should be used for normalization (Ling and Salvaterra 2011).
Here, I have identified reference genes suitable for qPCR normalization in developing C.
gigas gonad tissue. There was good correlation in stability analysis between NormFinder and
91

geNorm in male samples and when sexes were combined, but the results were fairly inconsistent
for female samples. Future pPCR expression studies of reproductive tissue should be aware of
the potential effect of sex on normalization. These results will allow for more accurate
normalization of gene expression by qPCR of reproduction-specific genes in the commercially
important C. gigas.

92

References
Andersen, C., Jensen, J., Orntoft, T., 2004. Normalization of real-time quantitative reverse
transcription-PCR data: A model-based variance estimation approach to identify genes
suited for normalization, applied to bladder and colon cancer data sets. Cancer Res., 64,
5245-5250.
Boudry, P., Collet, B., Cornette, F., Hervouet, V., Bonhomme, F., 2002. High variance in
reproductive success of the Pacific oyster (Crassostrea gigas, thunberg) revealed by
microsatellite-based parentage analysis of multifactorial crosses. Aquaculture, 204, 283-296.
Bustin, S.A., Benes, V., Garson, J.A., Hellemans, J., Huggett, J., Kubista, M., Mueller, R.,
Nolan, T., Pfaffl, M.W., Shipley, G.L., Vandesompele, J., Wittwer, C.T., 2009. The MIQE
guidelines: Minimum information for publication of quantitative real-time PCR
experiments. Clin. Chem., 55, 611-622.
Cubero-Leon, E., Ciocan, C., Minier, C., Rotchell, J., 2012. Reference gene selection for qPCR
in mussel, Mytilus edulis, during gametogenesis and exogenous estrogen exposure. Environ.
Sci. Pollut. R., 19, 2728-2733.
David, E., Tanguy, A., Pichavant, K., Moraga, D., 2005. Response of the Pacific oyster
Crassostrea gigas to hypoxia exposure under experimental conditions. Febs J., 272, 5635-
5652.
Deloffre, L.A.M., Andrade, A., Filipe, A.I., Canario, A.V.M., 2012. Reference genes to quantify
gene expression during oogenesis in a teleost fish. Gene, 506, 69-75.
Derveaux, S., Vandesompele, J., Hellemans, J., 2010. How to do successful gene expression
analysis using real-time PCR. Methods, 50, 227-230.
93

Dheda, K., Huggett, J., Chang, J., Kim, L., Bustin, S., Johnson, M., Rook, G., Zumla, A., 2005.
The implications of using an inappropriate reference gene for real-time reverse transcription
PCR data normalization. Anal. Biochem., 344, 141-143.
Dheilly, N.M., Lelong, C., Huvet, A., Favrel, P., 2011. Development of a Pacific oyster
(Crassostrea gigas) 31,918-feature microarray: Identification of reference genes and tissue-
enriched expression patterns. BMC Genomics, 12, 468.
Du, Y., Zhang, L., Xu, F., Huang, B., Zhang, G., Li, L., 2013. Validation of housekeeping genes
as internal controls for studying gene expression during Pacific oyster (Crassostrea gigas)
development by quantitative real-time PCR. Fish Shellfish Immunol., 34, 939-945.
Eckelbarger, K.J. and Davis, C.V., 1996. Ultrastructure of the gonad and gametogenesis in the
Eastern oyster, Crassostrea virginica. I. ovary and oogenesis. Mar. Biol., 127, 79-87.
Eckelbarger, K.J. and Davis, C.V., 1996. Ultrastructure of the gonad and gametogenesis in the
Eastern oyster, Crassostrea virginica. II. testis and spermatogenesis. Mar. Biol., 127, 89-96.
Fabioux, C., Pouvreau, S., Roux, F.L., Huvet, A., 2004. The oyster vasa-like gene: A specific
marker of the germline in Crassostrea gigas, Biochem. Bioph. Res. Co., 315, 897-904.
Franco, A., Heude Berthelin, C., Goux, D., Sourdaine, P., Mathieu, M., 2008. Fine structure of
the early stages of spermatogenesis in the Pacific oyster, Crassostrea gigas (mollusca,
bivalvia). Tissue Cell, 40, 251-260.
Herpin, A., Lelong, C., Becker, T., Favrel, P., Cunningham, C., 2007. A tolloid homologue from
the Pacific oyster Crassostrea gigas . Gene Expr. Patterns, 7, 700-708.
Huvet, A., Daniel, J., Quere, C., Dubois, S., Prudence, M., Van Wormhoudt, A., Sellos, D.,
Samain, J., Moal, J., 2003. Tissue expression of two alpha-amylase genes in the Pacific
oyster Crassostrea gigas. effects of two different food rations. Aquaculture, 228, 321-333.
94

Kingtong, S., Kellner, K., Bernay, B., Goux, D., Sourdaine, P., Berthelin, C.H., 2013. Proteomic
identification of protein associated to mature spermatozoa in the Pacific oyster Crassostrea
gigas. J. Proteomics, 82, 81-91.
Krauchunas, A.R. and Wolfner, M.F., 2013. Molecular changes during egg activation.
Gametogenesis, 102, 267-292.
Kroupova, H., Trubiroha, A., Wuertz, S., Kloas, W., 2011. Stage-dependent differences in RNA
composition and content affect the outcome of expression profiling in roach (Rutilus rutilus)
ovary. Comp. Biochem. Phys., 159, 141-149.
Lannan, J.E., Robinson, A., Breese, W.P., 1980. Broodstock management of Crassostrea gigas:
II. broodstock conditioning to maximize larval survival. Aquaculture, 21, 337-345.
Ling, D. and Salvaterra, P.M., 2011. Robust RT-qPCR data normalization: Validation and
selection of internal reference genes during post-experimental data analysis. Plos One, 6,
e17762.
Leelatanawit, R., Klanchui, A., Uawisetwathana, U., Karoonuthaisiri, N., 2012. Validation of
reference genes for real-time PCR of reproductive system in the black tiger shrimp. Plos
One, 7, e52677.
Mauriz, O., Maneiro, V., Luz Perez-Paralle, M., Luis Sanchez, J., Juan Pazos, A., 2012.
Selection of reference genes for quantitative RT-PCR studies on the gonad of the bivalve
mollusc Pecten maximus L. Aquaculture, 370, 158-165.
Meistertzheim, A., Tanguy, A., Moraga, D., Thebault, M., 2007. Identification of differentially
expressed genes of the Pacific oyster Crassostrea gigas exposed to prolonged thermal stress.
Febs J., 274, 6392-6402.
95

Pfaffl, M., Tichopad, A., Prgomet, C., Neuvians, T., 2004. Determination of stable housekeeping
genes, differentially regulated target genes and sample integrity: BestKeeper - excel-based
tool using pair-wise correlations. Biotechnol. Lett., 26, 509-515.
Ramirez, M., Le Pennec, M., Dorange, G., Devauchelle, N., Nonnotte, G., 1999. Assessment of
female gamete quality in the Pacific oyster Crassostrea gigas. Invertebr. Reprod. Dev., 36,
73-78.
Vandesompele, J., De Preter, K., Pattyn, F., Poppe, B., Van Roy, N., De Paepe, A., Speleman, F.,
2002. Accurate normalization of real-time quantitative RT-PCR data by geometric
averaging of multiple internal control genes. Genome Biol., 3, 0034.1.

96

Tables
Table 1
Gene names and primer sequences of 10 candidate reference genes used for qPCR
Gene Name Abbreviation Accession
number
Primer Sequence
Ribosomal protein L7 RPL7 AJ557884 5'-TCCCAAGCCAAGGAAGGTTATGC-3'
5'-CAAAGCGTCCAAGGTGTTTCTCAA-3'
Ribosomal protein S18 RS18 AB199895 5'-GCCATCAAGGGTATCGGTAGAC-3'
5'-CTGCCTGTTAAGGAACCAGTCAG-3'
Elongation factor 1-α elf1 AB122066 5'-CGTAAAGGAAATCCGTCGTG-3'
5'-GAAGTTCTTGGCGCCCTTT-3'
Heat shock protein 70 hsp70 AB122064 5'-TGTGAAAGGGCAAAGAGGAC-3'
5'-CTCAAACCTAGCCCTGGTGA-3'
Cytochrome c oxidase cox1 AB736486 5'-ACTAGGCATGCGTTGGTTATG-3'
subunit 1 5'-CATCAAAGGGATAAGCCAGTT-3'
β-actin actin AF026063 5'-ATCTCGCTGGACGTGATCTG-3'
5'-GGCTGTGGTGGTGAAAGAGT-3'
α-tubulin tubulin AB196533 5'-GGAGAGGGAATGGAAGAAGG-3'
5'-CAGCCTCTCCCTCAACAGAA-3'
18s rRNA 18s AB064942 5'-GCGTTTAGCCACACGAGATT-3'
5'-ATCCCTAGCACGAAGGAGGT-3'
28s rRNA 28s AY632555 5'-AAGGGCAGGAAAAGAAACTAAC-3'
5'-GTTTCCCTCTAAGTGGTTTCAC-3'
Glyceraldehyde-3 phosphate- gapdh AJ544886 5'-CGCCAATCCTTGTTGCTT-3'
dehydrogenase 5'-TTGTCTTGCCCCTCTTGC-3'

97

Figures

Figure 1. Raw expression cycle data of the 10 candidate reference genes. Genes are graphed in
pairs, by sex.
Figure 2. Stability values of 10 candidate reference genes using Normfinder and geNorm for
female samples. Lower stability values represent increased stability. Note the stability values
are unique for each method.
Figure 3. Stability values of 10 candidate reference genes using Normfinder and geNorm for
male samples. Lower stability values represent increased stability. Note the stability values are
unique for each method.
Figure 4. Stability values of 10 candidate reference genes using Normfinder and geNorm for
male and female samples combined. Lower stability values represent increased stability. Note
the stability values are unique for each method.

98

Figure 1

99

Figure 2

100

Figure 3

101

Figure 4

102

Chapter 5: Expression of candidate biomarkers exhibit stage-
specific expression during oogenesis in the Pacific oyster Crassostrea
gigas

Abstract
Efficiency of hatchery-based larval production suffers from use of broodstock in non-optimal
reproductive state. Access to tissue and resolution of signal hinder current means of identifying
reproductive state. Molecular based indicators, such as protein or gene expression, may offer
better resolution of stage while requiring less biological material than other methods of
identifying maturation stage. Expression of several genes previously found to show stage-
specific expression was used to assess dataset variance and the ability to predict duration of
gonad conditioning. Stage-specific expression was generally confirmed for the genes analyzed
with high statistical significance. Expression of two genes in particular accounted for 42% of the
variance in predicting duration of conditioning by regression analysis. More genes need to be
screened, but given the paucity of tissue needed, gene expression appears to hold potential for
future use as an indicator of reproductive stage.

103

Introduction
Commercial production of shellfish typically relies on hatchery produced larvae or seed,
which are then grown out on beaches or suspended in coastal waters. Despite increased
production of shellfish in recent years, substantial potential exists for continued increased
growth, owing to availability of growth habitat and a naturally replenishing food supply. An
estimated 8% of aquatic primary production is needed to support current combined levels of
capture fisheries and aquaculture (Naylor et al. 2000). For the Pacific oyster Crassostrea gigas,
the production bottleneck exists at the level of larval production, where inconsistency in survival
among spawns is commonplace.
Recent episodes of poor survival of hatchery produced larvae have been directly linked to
increasingly acidic input waters (Barton et al. 2012). While the long terms effects of ocean
acidification are alarming for shellfish production, solace may be found in that adults are not as
susceptible as larvae to the effects of corrosive water. Moreover, current troubles in hatcheries
may be curbed by altering the chemistry of the input water prior to use in larval production. But
inconsistent larval survival is not a new phenomenon; rather, it has a lengthy history (Lannan
1980). Factors leading to inconsistent spawns include water chemistry issues (Eudeline et al.
2009), disease (Elston et al. 2008), food impurities (Lewis et al. 1988), genetic incompatibility
between parents (Gaffney et al. 1993, Boudry et al. 2002), and non-optimal state of gonad
maturation of the parent(s) (Lannan et al. 1980).
The industry-wide impact of variation in state of gonadal maturation between parents on
larval survival is unclear, but accounting for it is difficult. The calcium carbonate shell impedes
assessing the state of maturation of potential broodstock during maturation. In addition, though
the effect of variation between parents has only been directly investigated once, to our
104

knowledge (Lannan et al. 1980), several studies have demonstrated the existence of variation in
state of maturation among potential parents in bivalve molluscs (Lannan 1980, Hershberger et al.
1984, Barber et al. 1991, Samain et al. 2007). Coupling parents that are in an optimal state of
maturation would increase the efficiency of hatchery production.
C. gigas broodstock are generally sacrificed upon spawning. Non-lethal methods capable
of determining the state of maturation of broodstock could assist in the selection of animals for
spawning and prevent the loss of animals chosen for spawning but not in reproductive condition.
Recent attempts to develop such a tool have included magnetic resonance imaging (MRI) of
whole oysters and measuring total protein content of hemolymph (Davenel et al. 2006, Li et al.
2010). These developing techniques lack resolution or entail high cost, in the case of MRI.
The role of gene expression in shaping developing tissue makes it an obvious target for
use as a reproductive biomarker. Several studies on commercially important bivalve molluscs
have demonstrated stage-specific gene expression (Boutet et al. 2008, Ciocan et al. 2011, Dheilly
et al. 2012, Paul, Chapter 2). Reverse transcription quantitative PCR (RT-qPCR) has become the
standard method measuring gene expression when a limited set of genes are queried. Because of
its speed, sensitivity, specificity, and relative ease of use, qPCR is a suitable method for use in
determining reproductive stage.
In this study, gene expression of candidate maturation biomarkers was measured in gonad
tissue over the course of gonad maturation. The candidate biomarkers were previously found to
be differentially expressed by RNA-Seq among multiple stages of gonad maturation within a
limited number of individuals (Paul, Chapter 2). The goal in mind was to see to what degree
gene expression could be used to identify reproductive status. Gonad tissue can be accessed for
biopsy by relaxing animals in magnesium chloride or by drilling a notch or hole in the shell.
105

While hemolymph is a less invasive tissue sample, gonad is likely to carry stronger reproduction-
specific signals. Expression was measured by qPCR on individuals that were conditioned from
one to ten weeks at 25 ºC.

106

Materials and Methods
Artificial conditioning of gonad maturation
Animals from two inbred lines and the two hybrid cross of the parental inbred lines,
produced in 2007, were used for artificial conditioning, which began in early March, 2010, at or
near the timing of natural initiation of gametogenesis. Multiple families were used to account for
the possibility of poor conditioning in a particular family, and to increase the number of
individuals sampled. Conditioning was carried out at a constant 25 °C, after increasing the water
temperature 2 °C day
-1
from the ambient incurrent water temperature of 13 °C. Conditioning
was halted after 10 weeks, at which point active resorption was noticed among sampled female
gonads. Oysters were held in flow through tanks and fed a live mixed algal diet. Five
individuals from each family were sampled weekly for 10 weeks. Oysters were shucked and
sectioned just ventral to the labial palps, after which a gonad tissue sample was taken from the
anterior half of the ventral section and immediately placed on dry ice. Samples were held at -80
°C until processed.
RNA extraction and cDNA synthesis
Total RNA was extracted using TRIzol® Plus RNA Purification Kit (Invitrogen) and a
TissueLyser II (Qiagen). RNA samples were treated with DNAse I (1 U/µg total RNA,
Promega) to remove genomic DNA prior to reverse transcription. Quantification and purity were
assessed with a NanoDrop® ND-1000. Samples with 260/280 and 260/230 ratios below 1.8
were excluded from further analysis. One µg total RNA was used for synthesis of first strand
cDNA using M-MLV reverse transcriptase (Promega) with oligo(dT)
15
priming. RNasin®
(promega) was included in all RT reactions to inhibit RNase activity. The total volume for RT
reactions was 25 µl.
107

Measurement of gene expression by qPCR
A total of 96 females were screened for gene expression levels by qPCR. The six genes
were chosen for measurement had previously been found to increase (forkhead box protein L1,
foxl1; thioredoxin reductase-2, thioR; kruppel like factor-5, klf5) or decrease (vitellogenin-6,
vit6; pancreatic lipase-2, lipase; hemagglutinin/amoebocyte aggregation factor, haaf) over the
course of gonad maturation. Primers were developed using Primer3
(http://bioinfo.ut.ee/primer3-0.4.0/). Sequences were blasted for uniqueness and specificity.
Details of primer set can be seen in Table 1. Amplified sequences were sized on agarose gels,
and homogeneity was assessed by melt curve analysis following amplification.
Dilution series were performed for all primer sets to determine the template range with
PCR efficiency of R
2
= 1 +/- 0.05. Amplification reactions contained 5 ng template, 600 nm
primer concentration, and 1 X Brilliant II SYBR green QPCR master mix (Agilent) in a total
volume of 15µl. All samples were run in duplicate. NoRT controls were tested prior to
measuring of gene expression. IRC and NTC controls were included for each gene of each qPCR
run. Normalization of qPCR expression in developing tissues requires careful selection of
reference genes. Variation in sample input was normalized by the geometric mean expression of
the reference genes gapdh and hsp70. This pair of genes had been shown to be the best pair of
reference genes for normalization in developing C. gigas reproductive tissue among a set of
screened candidates (Author’s unpublished data).
Data Analysis
All statistical analyses were performed using R v3.0.1. Samples were crudely grouped
into three maturation groups based on the length of their conditioning period. Groups were
plotted then tested for differences in mean expression by one-way ANOVA followed by a post-
108

hoc tukey test. Linear discriminant analysis (LDA) was performed in order to see how faithful
sample expression was to the conditioning group within which it was placed. Correlation among
candidate biomarker expression was assessed by principle component analysis (PCA). Finally,
the ability of expression data to predict length of gonad conditioning was modeled by stepwise
multiple regression analysis.

109

Results
As a first step towards analyzing expression dynamics, the ten sampling time points were
grouped into three stages which roughly translate to immature (stage I, weeks 1-3), early
maturation (stage II, weeks 4-7) and late maturation (weeks 8-10). The number of samples
grouped within each stage was 25, 35 and 36 for stages I-III, respectively. Specific maturation
details within the time intervals is less important than determining whether gross differences
exist in expression over the course of maturation. After grouping of samples within stages, one-
way ANOVAs were performed in order to confirm unstable expression levels during female
maturation of the six genes measured herein.
Three of the candidate biomarkers were previously shown (author’s unpublished data) to
increase in expression within the female gonad during maturation (foxl1, thioR, klf5), while the
other three were shown to decrease (vit6, lipase, haaf). With the exception of klf5, the stage-
specific expression patterns are consistent with those previously found (Figure 1). Klf5
expression showed no stage-specific expression here with the chosen artificial divisions among
stage. The other genes all showed highly significant stage specific expression (p < 1 x 10
-7
). In
the case of the three genes with decreasing expression (vit6, lipase, haaf), it was the stage III
samples driving the significant result, while stages I and II were not significantly different from
each other (Table 2). For foxl1 and thioR, it was stage I samples that were significantly different
from stages II and III (Table 2). Foxl1 showed expression increasing from stage I through stage
III, though the statistical difference in expression between stages II and III was only marginally
significant (p = 0.0188). Sample grouping was successfully predicted by LDA for 67% of the 96
samples, suggesting a considerable amount of variation existed either in state of maturation
among samples or in expression of candidate biomarkers at a given maturation state. Still, the
110

expression results described above indicate strong stage-specific patterns among candidate
biomarkers.
PCA was performed as a primary means of assessing the multivariate expression dataset.
Using the Kaiser criterion of dropping factors with eigenvalues < 1 (Kaiser 1960), the first three
principle components were deemed worthy of consideration, and accounted for approximately
85% of the variance. Another way to determine the number of principle components worthy of
consideration is by plotting them against their contributions to the overall variance (Figure 2).
From the scree plot in Figure 2, only the last two principle components are clearly dismissible.
The contributions to each of the first four principle components, or loadings, are plotted in
Figure 3. Vit6, lipase and haaf were the major contributors of PC1, with the other three genes
contributing less. PC1 accounted for 44% of the variance. Foxl1 and thioR contributed most to
PC2, with klf5 contributing a lesser, opposite effect. Klf5 contributed to PC3, while vit was the
major contributor to PC4. The PCA biplot showing the PC1 & PC2, and the correlation among
the genes, can be seen in Figure 4.
Multiple regression models were created in order to view the predictive ability of the
dataset. The primary utility of biomarkers in this report is in providing a signature of a specific
developmental stage. With the number of weeks conditioned as the dependent variable and the
candidate biomarkers as the predictor variables, a minimum adequate model was found to
include just foxl1 and lipase (Table 3). Removal of the other four genes only decreased the R
2

value of the model by 0.03 (0.45 to 0.42). The model constant of 5.83 suggests that when
expression levels of foxl1 and lipase are nearly equal, the animal has been conditioned for
roughly 6 weeks. The predictive power of the model, with R
2
=0.42, is not great. In general,
when foxl1 expression is greater than lipase expression, the predicted time of conditioning will
111

be greater than 6 weeks, and vice versa. Including genotype of samples as a predictor did not
significantly alter the results of the regression model. And running the model separately for each
genotype produced roughly the same results, as well.

112

Discussion
Stage-specific maturation patterns and biomarker function
Stage-specific expression of the six candidate biomarkers largely matched expectation,
with the exception of klf5, when the samples were divided into 3 bins based on length of
conditioning. While klf5 was seemingly identified as a false positive, it is interesting to note that
it did display stage-specific expression among males of the same sampling set (unpublished
data). Among the other five candidates, two consistent patterns emerged. Genes with
decreasing expression during maturation (vit6, lipase, haaf) showed no noticeable difference in
expression in immature and early maturation gonad. It was in the late maturation gonad where
expression levels dropped significantly. Genes with increasing expression (foxl1, thioR) over
maturation displayed little or no significant differences in expression between early and late
maturation samples, while immature samples had significantly lower expression.
The pattern of reduced expression in late maturation samples suggests a possible role in
maturation of the oocytes. And indeed, evidence exists suggesting such a role for vit6 and
lipase. Vitellogenesis is the process of yolk formation which serves as a nutrient store for
developing embryos. Vitellin is the major yolk protein and vitellogenin is the precursor
molecule of vitellin. In bivalves, vitellogenin is produced autosynthetically within oocytes and
heterosynthetically within adjoining cells where it is then transported by endocytosis into the
oocytes (Eckelbarger and Davis 1996, Agnese et al. 2013). When expression of vitellogenin was
previously measured in ovary tissue of bivalves, including C. gigas, a similar reduction in
expression over maturation was reported (Matsumoto et al. 2003, Agnese et al. 2013, Zheng et
al. 2012).
113

Less is known about the role of the pancreatic lipase-2 like gene (called here lipase),
though similar expression patterns were found in the scallop Patinopecten yessoensis, where it
was hypothesized to play a role in delivering metabolites to growing oocytes (Kim et al. 2008).
The role of haaf in oocyte maturation may be less direct as it is an immune-related gene. But
reproduction and immune status are tightly coupled due to the energetic cost associated with
reproduction (Duchemin et al. 2007). Interestingly, an haaf homolog was reported up-regulated
in M. edulis following estradiol exposure (Ciocan et al. 2011).
Greater expression in early and late maturation samples suggests an oocyte-specific
function, and in the case of foxl1 and thioR, may indicate a role in maintaining the integrity of
oocytes. Forkhead box (fox) genes are a diverse group of ancient transcription factors (Shimeld
et al. 2010). The better known foxl2 in involved in sex determination in vertebrates, an ortholog
of which was recently characterized in C. gigas (Naimi et al. 2009). Foxl1 orthologs are
transcriptional repressors that have no known reproduction-specific function, but are necessary
for brain and gut development in vertebrates (Nadaka et al. 2006, Madison et al. 2009). The
potential role of foxl1 in maintaining oocyte integrity may thus involve the repression of
transcription, as mature oocytes are transcriptionally silent. The potential role of thioredoxin
reductase-2 (thioR) is perhaps more obvious. ThioR reduces thioredoxin, allowing it to perform
functions that include protection from oxidant injury, transcription regulation and apoptosis
(Mustacich and Powis 2000). Previous results from bovine showed thioR helped to maintain
redox homeostasis in mature oocytes, while stage-specific expression of thioR was similar to that
reported here (Yuan et al. 2012).

114

Principle component analysis and multiple regression
The PCA analysis provided a clustered view of the expression of the six genes for each of
the 96 female samples. The overall pattern of results was not surprising having knowledge of the
stage-specific expression of each of the six genes individually. There is strong correlation
among vit, lipase, and haaf, as well as between foxl1 and thioR (Figure 4). These two groups of
genes were negatively correlated, as expected, and this correlation contributed most to principle
component 1 (PC1). The stage-specific expression results discussed above would only allow
prediction of PC1. The source of PC2 is less intuitive, and is driven by opposing effects of klf5
with foxl1 and thioR.
Klf5 was originally hypothesized to have increasing expression throughout maturation.
And while this wasn’t seen by stage-specific expression results, it did contribute a small amount
of variance to PC1 in the direction of foxl1 and thioR. Variance in the dataset explained by PC2
was due to the opposing effects of klf5 with foxl1 and thioR. This result, along with its
contribution to PC1 suggests the possibility of two types of mature gonad samples. If foxl1 and
thioR do contribute to maintaining the viability of the mature gonad, perhaps klf5 is involved in
the resorbtion of gonad material after development and/or maintenance is aborted. A role in
gonad resorbtion would not be hard to imagine given the widespread functional processes which
kruppel-like transcription factors are involved in, such as cell proliferation, apoptosis,
differentiation, adipogenesis, inflammation, and development (Pearson et al. 2008, Wu and
Wang 2013). Detailed histological analyses would be needed to confirm any inferences about
the opposing expression results between klf5 and foxl1/thioR and their association with the
underlying physiological condition of oocyte-containing gonad samples.
115

PCA analysis identified correlation in expression of target genes. The analysis extended
what could be learned by looking at stage-specific expression of the target genes individually and
grouping them based on their expression patterns. Multiple regression analysis was used to
investigate correlation between target gene expression and the amount of time conditioned. The
regression analysis clearly identified foxl1 and lipase as the two best predictors of the length of
gonad maturation even though each was highly correlated and showed similar stage-specific
expression with other gene(s) (Figure 1, 4).
As the precursor of the primary yolk storage protein in oocytes, vit is an obvious
candidate for a female maturation stage-specific biomarker. The amount of vitellin/vitellogenin-
like protein in female gonad tissue has been proposed as a tool for reproductive-stage prediction
in C. gigas (Arcos et al. 2009). Not surprisingly, the amount of vitellin protein was found to
increase during gonad maturation. ELISA assays, however, require large amounts of tissue input
and thus require invasive sampling. Expression of vit mRNA during gonad maturation has also
been characterized during gonad maturation, where expression levels decrease over time as the
rate of protein accumulation decreases (Matsumoto et al. 2003). Despite vit expression
displaying stage specific expression here consistent with previous results, it was not highly
correlated with the length of conditioning period.
The lipase measured here has not been formally characterized, but it likely, along with
vit, contributes to yolk accumulation in developing oocytes. A pancreatic lipase with high
sequence similarity to the lipase measured here was recently isolated and characterized in the
scallop Patinopecten yessoensis (Kim et al. 2008). The authors suggest a role in delivering
metabolites to developing oocytes. A similar role has been observed in fish for a lipoprotein
lipase, which helps in delivering neutral lipids to ooyctes (Jose Ibanez et al. 2008). The mRNA
116

of the lipoprotein lipase was localized to the follicle cells surrounding the developing oocytes. In
C. gigas, follicle cells gradually lose their association with oocytes as the oocytes mature. And
mature C. gigas oocytes contain roughly as many lipid droplets as membrane bound vitellin-
containing yolk granules (Eckelbarger and Davis 1996). These details suggest a possible
explanation for why lipase is a better predictor of reproductive stage than vit. Lipase should
steadily decrease as maturation increases, while vit behavior may be subject to unpredictable
variations in storage and mobilization within developing oocytes. The role of the other
significant predictor of time conditioned, the transcriptional repressor foxl1, is unclear due to the
lack of previous association with reproductive processes.
Stage-specific expression was confirmed by qPCR for 5 of the 6 genes tested. Klf5 did
not show differential expression among the three reproductive stages. PCA analysis identified an
opposing effect between genes expressed early or late as the main contributor of variation in the
dataset. Expression of the candidate biomarkers were subjected to a stepwise regression analysis
to check for correlation with the amount of time conditioned. Multiple regression analysis of
gene expression data is common in the literature, where it has been used for prediction of topics
ranging from disease to meat quality to the identification of best embryos produced during
artificial insemination events (Yamaoka et al. 2012, Damon et al. 2013, Wathlet et al. 2013).
This is believed to be the first attempt at correlating gene expression with reproductive maturity.
A pair of genes was identified as being effective predictors of reproductive stage in
female C. gigas. With a coefficient of determination of 0.42, the regression model did not have
high predictive power. The fit of the model may be strengthened if samples were assigned
detailed histological status rather than the length of time conditioned. Non-destructive
biological sampling is easier for hemolymph than for gonad, but gonad offers a stronger, more
117

consistent signal. Considering the small number of genes screened, and the large number of
genes likely differentially expressed during gonad maturation, the method has potential for
accurately staging C. gigas.

118

References
Agnese, M., Verderame, M., De Meo, E., Prisco, M., Rosati, L., Limatola, E., del Gaudio, R.,
Aceto, S., Andreuccetti, P., 2013. A network system for vitellogenin synthesis in the mussel
Mytilus galloprovincialis (L.). J. Cell. Physiol., 228, 547-555.
Arcos, F.G., Maria Ibarra, A., del Carmen Rodriguez-Jaramillo, M., Awilda Garcia-Latorre, E.,
Vazquez-Boucard, C., 2009. Quantification of vitellin/vitellogenin-like proteins in the oyster
Crassostrea corteziensis (hertlein 1951) as a tool to predict the degree of gonad maturity.
Aquacult. Res., 40, 644-655.
Barber, B.J., Ford, S.E., Wargo, R.N., 1991. Genetic variation in the timing of gonadal
maturation and spawning of the eastern oyster, Crassostrea virginica (gmelin). Biol. Bull.,
181, 216-221.
Barton, A., Hales, B., Waldbusser, G.G., Langdon, C., Feely, R.A., 2012. The Pacific oyster,
Crassostrea gigas, shows negative correlation to naturally elevated carbon dioxide levels:
Implications for near-term ocean acidification effects. Limnol. Oceanogr., 57, 698-710.
Boudry, P., Collet, B., Cornette, F., Hervouet, V., Bonhomme, F., 2002. High variance in
reproductive success of the pacific oyster (Crassostrea gigas, thunberg) revealed by
microsatellite-based parentage analysis of multifactorial crosses. Aquaculture, 204, 283-296.
Boutet, I., Moraga, D., Marinovic, L., Obreque, J., Chavez-Crooker, P., 2008. Characterization
of reproduction-specific genes in a marine bivalve mollusc: Influence of maturation stage
and sex on mRNA expression. Gene, 407, 130-138.
Ciocan, C.M., Cubero-Leon, E., Minier, C., Rotchell, J.M., 2011. Identification of reproduction-
specific genes associated with maturation and estrogen exposure in a marine bivalve Mytilus
edulis. Plos One, 6, e22326.
119

Damon, M., Denieul, K., Vincent, A., Bonhomme, N., Wyszynska-Koko, J., Lebret, B., 2013.
Associations between muscle gene expression pattern and technological and sensory meat
traits highlight new biomarkers for pork quality assessment. Meat Sci., 95, 744-754.
Davenel, A., Quellec, S., Pouvreau, S., 2006. Noninvasive characterization of gonad maturation
and determination of the sex of Pacific oysters by MRI. Magn. Reson. Imaging, 24, 1103-
1110.
Dheilly, N.M., Lelong, C., Huvet, A., Kellner, K., Dubos, M., Riviere, G., Boudry, P., Favrel, P.,
2012. Gametogenesis in the Pacific oyster Crassostrea gigas: A microarrays-based analysis
identifies sex and stage specific genes. Plos One, 7, e36353.
Duchemin, M.B., Fournier, M., Auffret, M., 2007. Seasonal variations of immune parameters in
diploid and triploid Pacific oysters, Crassostrea gigas (thunberg). Aquaculture, 264, 73-81.
Eckelbarger, K.J. and Davis, C.V., 1996. Ultrastructure of the gonad and gametogenesis in the
eastern oyster, Crassostrea virginica. I. ovary and oogenesis. Mar. Biol., 127, 79-87.
Elston, R.A., Hasegawa, H., Humphrey, K.L., Polyak, I.K., Hase, C.C., 2008. Re-emergence of
Vibrio tubiashii in bivalve shellfish aquaculture: Severity, environmental drivers,
geographic extent and management. Dis. Aquat. Organ., 82, 119-134.
Eudeline, B., Barton, A., Langdon, C., 2009. Recent struggles of west coast commercial
hatcheries to produce Pacific oyster larvae. J. Shellfish Res., 28, 695.
Gaffney, P.M., Bernat, C.M., Allen Jr., S.K., 1993. Gametic incompatibility in wild and cultured
populations of the Eastern oyster, Crassostrea virginica (gmelin). Aquaculture, 115, 273-
284.
Hershberger, W.K., Perdue, J.A., Beattie, J.H., 1984. Genetic selection and systematic breeding
in Pacific oyster culture. Aquaculture, 39, 237-245.
120

Jose Ibanez, A., Peinado-Onsurbe, J., Sanchez, E., Miguel Cerda-Reverter, J., Prat, F., 2008.
Lipoprotein lipase (LPL) is highly expressed and active in the ovary of european sea bass
(Dicentrarchus labrax L.), during gonadal development. Comp. Biochem. Phys. A, 150,
347-354.
Kaiser, H.F., 1960. The application of electronic computers to factor analysis. Educ. Psychol.
Meas., 20, 141-151.
Kim, K.S., Kim, B.K., Kim, H.J., Yoo, M.S., Mykles, D.L., Kim, H., 2008. Pancreatic lipase-
related protein (PY-PLRP) highly expressed in the vitellogenic ovary of the scallop,
Patinopecten yessoensis. Comp. Biochem. Phys. B, 151, 52-58.
Lannan, J.E., 1980. Broodstock management of Crassostrea gigas: I. genetic and environmental
variation in survival in the larval rearing system. Aquaculture, 21, 323-336.
Lannan, J.E., Robinson, A., Breese, W.P., 1980. Broodstock management of Crassostrea gigas:
II. broodstock conditioning to maximize larval survival. Aquaculture, 21, 337-345.
Lewis, T.E., Garland, C.D., O'Brien, T.D., Fraser, M.I., Tong, P.A., Ward, C., Dix, T.G.,
McMeekin, T.A., 1988. The use of 0.2-µm membrane-filtered seawater for improved control
of bacterial levels in microalgal cultures fed to larval Pacific oysters (Crassostrea gigas).
Aquaculture, 69, 241-251.
Li, Y., Siddiqui, G., Wikfors, G.H., 2010. Non-lethal determination of sex and reproductive
condition of Eastern oysters Crassostrea virginica gmelin using protein profiles of
hemolymph by proteinchip (R) and SELDI-TOF-MS technology. Aquaculture, 309, 258-
264.
121

Madison, B.B., McKenna, L.B., Dolson, D., Epstein, D.J., Kaestner, K.H., 2009. FoxF1 and
FoxL1 link hedgehog signaling and the control of epithelial proliferation in the developing
stomach and intestine. J. Biol. Chem., 284, 5936-5944.
Matsumoto, T., Nakamura, A., Mori, K., Kayano, T., 2003. Molecular characterization of a
cDNA encoding putative vitellogenin from the Pacific oyster Crassostrea gigas. Zool. Sci.,
20, 37-42.
Mustacich, D. and Powis, G., 2000. Thioredoxin reductase. Biochem. J., 346, 1-8.
Naimi, A., Martinez, A.S., Specq, M.L., Diss, B., Mathieu, M., Sourdaine, P., 2009. Molecular
cloning and gene expression of cg-Foxl2 during the development and the adult
gametogenetic cycle in the oyster Crassostrea gigas. Comp. Biochem. Phys. B, 154, 134-
142.
Nakada, C., Satoh, S., Tabata, Y., Arai, K., Watanabe, S., 2006. Transcriptional repressor foxl1
regulates central nervous system development by suppressing shh expression in zebra fish.
Mol. Cell. Biol., 26, 7246-7257.
Naylor, R.L., Goldburg, R.J., Primavera, J.H., Kautsky, N., Beveridge, M.C.M., Clay, J., Folke,
C., Lubchenco, J., Mooney, H., Troell, M., 2000. Effect of aquaculture on world fish
supplies. Nature, 405, 1017-1024.
Pearson, R., Fleetwood, J., Eaton, S., Crossley, M., Bao, S., 2008. Kruppel-like transcription
factors: A functional family. Int. J. Biochem. Cell Biol., 40, 1996-2001.
Samain, J.F., Dégremont, L., Soletchnik, P., Haure, J., Bédier, E., Ropert, M., Moal, J., Huvet,
A., Bacca, H., Van Wormhoudt, A., Delaporte, M., Costil, K., Pouvreau, S., Lambert, C.,
Boulo, V., Soudant, P., Nicolas, J.L., Le Roux, F., Renault, T., Gagnaire, B., Geret, F.,
Boutet, I., Burgeot, T., Boudry, P., 2007. Genetically based resistance to summer mortality
122

in the Pacific oyster (Crassostrea gigas) and its relationship with physiological,
immunological characteristics and infection processes. Aquaculture, 268, 227-243.
Shimeld, S.M., Boyle, M.J., Brunet, T., Luke, G.N., Seaver, E.C., 2010. Clustered fox genes in
lophotrochozoans and the evolution of the bilaterian fox gene cluster. Dev. Biol., 340, 234-
248.
Wathlet, S., Adriaenssens, T., Segers, I., Verheyen, G., Van Landuyt, L., Coucke, W., Devroey,
P., Smitz, J., 2013. Pregnancy prediction in single embryo transfer cycles after ICSI using
QPCR: Validation in oocytes from the same cohort. Plos One, 8, e54226.
Wu, Z. and Wang, S., 2013. Role of kruppel-like transcription factors in adipogenesis. Dev.
Biol., 373, 235-243.
Yamaoka, M., Maeda, N., Nakamura, S., Kashine, S., Nakagawa, Y., Hiuge-Shimizu, A., Okita,
K., Imagawa, A., Matsuzawa, Y., Matsubara, K., Funahashi, T., Shimomura, I., 2012. A
pilot investigation of visceral fat adiposity and gene expression profile in peripheral blood
cells. Plos One, 7, e47377.
Yuan, Y., Wheeler, M.B., Krisher, R.L., 2012. Disrupted redox homeostasis and aberrant redox
gene expression in porcine oocytes contribute to decreased developmental competence.
Biol. Reprod., 87, 78.
Zheng, H., Zhang, Q., Liu, H., Liu, W., Sun, Z., Li, S., Zhang, T., 2012. Cloning and expression
of vitellogenin (vg) gene and its correlations with total carotenoids content and total
antioxidant capacity in noble scallop Chlamys nobilis (bivalve: Pectinidae). Aquaculture,
366, 46-53.

123

Tables
Table 1
qPCR primer sequences and gene information
Gene Name Abbrev. Genbank OysterDB Primer Sequence
Vitellogenin-6 Vit-6 AB084783 OYG_10021817 5'-GAAGTTTACGTGAGAAGAC-3'
5'-TAGAAGTCTCTGAAGAACAG-3'
Pancreatic lipase-2 Lipase AM857075 OYG_10027499 5'-ATTCACGGCGTTTCCCTGT-3'
5'-CATTCTGTTGCAGCCGTTCT-3'
Hemagglutinin/amebocyte HAAF CU995762 OYG_10000742 5'-TGGTAGTCTCGGGATGGAAG-3'
aggregation factor 5'-CGAAGGACAGGTCTGAAGGA-3'
Forkhead box protein l1 Foxl1 AM865563 OYG_10006159 5'-TGCACATACCACATCCCAAC-3'
5'-TGCGAGTAACTCACGGAAGG-3'
Thioredoxin reductase 2, ThioR2 AM865969 OYG_10000920 5'-GGACTGAAGTCCGTCAAGCA-3'
mitochondrial 5'-AAACACGCCCACCTCAAAG-3'
Krueppel-like factor 5 Klf-5 CU991730 OYG_10000441 5'-AGGATTCGATCTGGGAGGAC-3'
5'-ATGGGATCGTTGGGAGAGTT-3'

124

Table 2
ANOVA and Tukey test results of potential differences in expression among 3 stages of
maturation. Symbols x, y & z denote significant results among stages as found by Tukey test
Gene df F value p value Stage

I II III
foxl1 2 35.21 4.14e -12 x y Z
klf5 2 1.297 0.27

thioR 2 17.03 4.97e -7 x y Y
vit-6 2 19.69 7.41e -8 x x Y
Lipase 2 16.41 7.88e -7 x x Y
Haaf 2 27.36 4.53e -10 x x Y

125

Table 3
Iterative regression models with number of weeks conditioned as the dependent variable.
Standard errors in parentheses. **,*** indicate 99%, 99.9% significance level, respectively.
model 1 model 2 model 3 model 4 model 5
constant 5.72*** 5.69*** 5.96*** 6.02*** 5.83***

(0.48) (0.47) (0.42) (0.42) (0.39)
lipase -0.16** -0.16** -0.16** -0.12** -0.15***

(0.05) (0.05) (0.05) (0.04) (0.03)
foxl1 0.17*** 0.16*** 0.15*** 0.15*** 0.16***

(0.04) (0.03) (0.03) (0.03) (0.03)
vit6 -0.08 -0.08 -0.08 -0.06

(0.05) (0.05) (0.05) (0.05)

haaf 0.10 0.10 0.10

(0.07) (0.07) (0.07)

klf5 0.17 0.15

(0.13) (0.12)

thioR -0.05

(0.09)

R-squared 0.45 0.45 0.44 0.43 0.42
Adjusted R-squared 0.42 0.42 0.42 0.41 0.41
No. observations 96 p-value 2.48e-11

126

Figures
Figure 1 Box and whisker plot showing stage-specific gene expression of the 6 reproductive-
stage candidate biomarkers. Expression is relative to the geometric mean of hsp70 and gapdh.
Stage I (weeks 1-3) are immature samples, while stage II (weeks 4-7) and stage III (weeks 8-10)
are early and late maturation, respectively. All but klf5 showed significant differences in mean
expression among stages.

Figure 2 Scree plot of component’s contributions to variance. Principle components are shown
on the x-axis, with their contributions to overall variance on the y-axis. The point where the
slope of the line approaches zero roughly indicates which components are worth considering. In
this case, principle components 5 & 6 can be ignored.

Figure 3 Dot plots of loadings for the first four principle components. Each loading indicates the
correlation of each variable with a given principle component. Correlation on the x-axis ranges
from -1 to 1.

Figure 4 PCA biplot showing patterns of expression in the 6 candidate biomarkers. Biplot
symbols indicate reproductive stage of sample. PC1 was responsible for 44% of the variance in
the dataset, while PC2 accounted for 25% of the variance. Vit, lipase and haaf were highly
correlated, as were thioR and foxl1.

127

Figure 1

128

Figure 2

129

Figure 3

pc1
variable loadings
lipase
haaf
vit
klf5
thioR
foxl1
-0.4 -0.2 0.0 0.2
pc2
variable loadings
foxl1
thioR
haaf
lipase
vit
klf5
-0.6 -0.4 -0.2 0.0 0.2
pc3
variable loadings
klf5
thioR
vit
haaf
lipase
foxl1
-0.8 -0.6 -0.4 -0.2 0.0
pc4
variable loadings
vit
thioR
foxl1
klf5
haaf
lipase
-0.5 0.0
pc1
variable loadings
lipase
haaf
vit
klf5
thioR
foxl1
-0.4 -0.2 0.0 0.2
pc2
variable loadings
foxl1
thioR
haaf
lipase
vit
klf5
-0.6 -0.4 -0.2 0.0 0.2
pc3
variable loadings
klf5
thioR
vit
haaf
lipase
foxl1
-0.8 -0.6 -0.4 -0.2 0.0
pc4
variable loadings
vit
thioR
foxl1
klf5
haaf
lipase
-0.5 0.0
130

Figure 4

1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2 2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3 3
3
3
3
3
3
3
3 3
3
3
3
3
3
3
3
3
3
3
foxl1
klf5
haaf
thioR
vit
lipase
PC1 (44%)
PC2 (25%)
131

Chapter 6: Conclusion
The state of gonad maturation in bivalve mollusks is an important for broodstock in
commercial production, yet is difficult to measure in practice. A quick, accurate and
nondestructive method of measuring state of gonad maturation has remained elusive. Even with
the shell removed, the tissue does not provide visual cues more specific than the presence or
absence of reproductive material. Of course, the shell cannot be removed without sacrificing the
animal. The gold standard of reproductive staging is histology, but this technique is slow and
destructive. Potential methods currently being investigated include magnetic resonance imaging
(MRI) (Davenel et al. 2006, Pouvreau et al. 2006) and screening tissue for expression of
biological factors such as proteins or mRNAs (Arcos et al. 2009). Small amounts of tissue can
be sampled through a notch or hole drilled through the shell without causing irreparable harm.
The central goal of this research was to identify potential gene expression biomarkers of
reproductive stage in C. gigas. Females were focused on specifically due to their greater
contribution towards larval survival (Lannan 1980) and their possession of the limiting sex cell.
Identifying reproductive stage in broodstock prior to spawning in commercial hatcheries
is important because use of animals in a non-optimal state has a negative effect on larval survival
(Muranaka and Lannan 1984). Current practices for identifying broodstock for spawning
estimate reproductive stage based on in-hatchery artificial or natural conditioning periods.
Information on conditioning progress of a cohort of animals may be gathered by sacrificing a
few individuals and checking for the presence/absence of gametes and/or relative volume of
gonad present. The success of these methods rely on maturation state being uniformly
distributed within a given population or cohort of animals. Unfortunately, variation in state of
gonad maturation appears to be common in nature (Hilbish and Zimmerman 1988, Barber et al.
132

1991, Samain et al. 2007). This variation results in reduced efficiency in hatchery-based
production of larvae because use of spawning parents in non-optimal states decreases larval
survival.
In order to identify expression biomarkers for state of maturation, global RNA
sequencing (RNA-Seq) was performed at multiple stages of gonad maturation. RNA-Seq has
become a popular method for measuring genome-wide mRNA expression levels due to the ease
of library construction and the great amount of information received quickly at a relatively low
cost. Data analysis is the primary difficulty associated with the method. Specific difficulties that
were realized here working with reproductive tissue in C. gigas will be discussed below. In sum,
237 genes were identified which were differentially expressed (DEGs) between two stages of
reproduction at the p = 0.05 level after correcting for multiple testing. These genes make up the
pool of identified candidate biomarker genes for female reproductive stage.
The use of gene expression as a biomarker in developing gonad tissue requires extra care
in the interpretation and normalization of expression data. Two specific considerations were the
subject of investigations here. The first involves the system in eggs of storing transcripts needed
between the time of meiotic maturation and zygotic genome activation (ZGA). Because mature
eggs are more or less transcriptionally silent, any mRNA needed prior to ZGA must be stored.
Transcripts are stabilized during storage through the removal of the poly(A) tail, which is first
added as a common pre-processing step following transcription in eukaryotes. After transport
into the cytoplasm from the nucleus, the poly(A) tails of transcripts destined for later use are
removed until they are once again added prior to translation.
The reason poly(A) tails are important is that priming of RNA for reverse transcription
(produces cDNA & the most common first step in measuring mRNA abundances) often uses
133

oligo(dT) primers, which bind to and select mRNA which carries a poly(A) tail. The other
commonly used primers are random hexamers, which unlike oligo(dT) primers, will bind to and
reverse transcribe all RNA present. Thus, if a transcript is undergoing active deadenylation, the
choice of priming method will result in differing levels of mRNA expression. Comparing
expression levels between samples which have been primed by alternative methods may result in
spurious conclusions being made. Additionally, if conclusions are being made about the
biological function attached to the expression levels measured, care should be taken to address
whether one is measuring all mRNA or just those transcripts with poly(A) tails.
The extent of deadenylation, including the number of genes and the proportion of each
gene’s expression, is not known. As far as I know, deadenylation of mRNA during gonad
maturation in C. gigas has not been reported. Twelve different genes were investigated here,
including some found to be differentially expressed among maturation stages, as well as some
common housekeeping genes. Deadenylation was found to be common among both
housekeeping genes and maturation stage DEGs. The pattern of results found did not match
expectations, however, with early-stage deadenylation being common, rather than late-stage as
would be expected under the storage hypothesis. Only one of the twelve, heat-shock protein 70
(hsp70), showed evidence of transcript storage in mature eggs. Half of the genes showed
differences in the relative amount of poly(A) to total RNA among stages, confirming the
importance of specifying total or poly(A) RNA levels when reporting gene expression levels in
developing gonad tissue. And of even greater importance, three genes had different patterns of
differential expression among reproductive stages for total and poly(A) RNA. These results
show clearly when reporting gene expression levels that the method of priming must be
considered when analyzing expression data.
134

Were expression data to be used for a small set of genes for identifying reproductive
stage, quantitative PCR (qPCR) would currently be the best choice for generating expression
data. The benefits to qPCR include high sensitivity and specificity, low sample input required,
easy sample preparation and quick generation of data. Relative quantification of qPCR data
requires normalization to one or several stably expressed housekeeping genes. In theory, these
housekeeping genes should be expressed uniformly in all tissues/stages/conditions being
measured. Unfortunately few if any genes meet this requirement in any given experimental
design.
Many studies have measured the expression of a given gene during gonad maturation in
C. gigas. None of them have provided direct support for their choice of housekeeping genes
despite working with a developing tissue which is less likely to have stably expressed
housekeeping genes. Several methods have been developed and are freely available for
analyzing the stability of candidate housekeeping genes. Here, ten commonly used
housekeeping genes were analyzed in male and female gonad tissue representing multiple stages
of gonad maturation. The best pair of housekeeping genes far males samples was different that
for female samples. When samples were combined, however, the pair of glyceraldehyde 3-
phosphate dehydrogenase (gapdh) and hsp70 met proposed criteria of stability to adequately
normalize expression data in developing gonad tissue samples. Commonly used ribosomal RNA
housekeeping genes, such as 18s, and microtubule-related housekeeping genes, such as β-actin
or α-tubulin, were particularly ineffective as housekeeping genes.
Equipped with housekeeping genes adequate for normalizing gene expression data, a set
of DEGs was selected for screening as potential indicators of reproductive stage. DEGs were
selected that both increased in expression and decreased in expression. In this way, when put on
135

a relative scale, the expression of each should cross at a particular maturation stage. Ninety-six
female samples spread over 10 weeks of artificial conditioning were subjected to biomarker
screening.
For the most part, stage-specific expression by qPCR matched that of RNA-Seq with high
statistical significance. Biomarkers were modeled by multiple regression analysis in order to
identify the candidate genes which best predicted the length of artificial conditioning period each
individual was subjected to. Although there was high correlation among the two groups of genes
chosen (those increasing or decreasing during maturation), one gene from each was identified as
significant indicators of maturation stage. A pancreatic-lipase related protein (lipase) and
forkhead box-like protein 1 (foxl1) together were highly correlated with artificial conditioning
period. The other candidate genes screened, which were highly correlated with lipase and foxl1,
did not contribute to the significance of the regression model.
The R
2
of the model, which indicates the predictive ability of the regression model, was
relatively low at 0.42. Essentially, the model has the predictive power to tell whether a sample is
in an early or late stage of maturation. When expression of foxl1 is greater than lipase, the
animal is past mid-maturation, and vice versa. Taking a step back, the results are encouraging
given the few number of genes initially screened. Even more significant is that samples were
simply grouped by the length of conditioning rather than by, for example, histological staging.
Variation in state of gonad maturation within cohorts of individuals is common, and this likely
contributed to the lower R
2
value. The statistical significance of the model tells that the
combined gene expression of lipase and foxl1 do provide meaningful information about the state
of maturation of any individual.
136

The direction taken with this research to identify candidate biomarkers for state of
maturation was met with success overall. This work would have benefited from increased RNA-
Seq read depth in mid and late maturation stage samples to account for the accumulation of
rRNA in these samples. Because of the nature of deadenylation and lack of transcription in
mature eggs, selecting against rRNA, or at least not selecting for poly(A) RNA during library
preparation, would provide more robust libraries for the testing and identification of stage
specific DEGs. Identification and screening of additional candidates would only increase the
predictive ability of gene expression to identify maturation state. While state of maturation has
been correlated with larval survival, using larval survival rather than state of maturation or
conditioning period as the response variable of the regression model would highlight the
significance of the biomarkers’ potential utility.
Gene expression has potential as an indicator of reproductive stage. Accessing and
sampling tissue would require some level of training. And development of additional biomarkers
would serve to increase the precision of the staging. Compared with other non-destructive
staging methods in development, such as MRI, gene expression is more accessible and offers
greater potential resolution. Current troubles with hatchery-based larval survival related to
climate change associated water quality issues have the full attention of the industry for good
reason. Should in-house treatment be effective in stabilizing changing input waters,
hatcherymen will once again be faced with inefficiencies in larval survival primarily related to
variation in state of gonad maturation among broodstock. Development of a method for the
identification and tracking of maturation stage in potential broodstock would help to increase
larval survival and the efficiency of larval production. Gene expression stands as a promising
candidate indicator for the state of gonad maturation in C. gigas and other bivalve mollusks.
137

References
Arcos, F.G., Maria Ibarra, A., del Carmen Rodriguez-Jaramillo, M., Awilda Garcia-Latorre, E.,
Vazquez-Boucard, C., 2009. Quantification of vitellin/vitellogenin-like proteins in the oyster
Crassostrea corteziensis (hertlein 1951) as a tool to predict the degree of gonad maturity.
Aquacult. Res. 40, 644-655.
Barber, B.J., Ford, S.E., Wargo, R.N., 1991. Genetic variation in the timing of gonadal
maturation and spawning of the Eastern oyster, Crassostrea virginica (gmelin). Biol Bull
181, 216-221.
Davenel, A., Quellec, S., Pouvreau, S., 2006. Noninvasive characterization of gonad maturation
and determination of the sex of Pacific oysters by MRI. Magn. Reson. Imaging 24, 1103-1110.
Hilbish, T.J. and Zimmerman, K.M., 1988. Genetic and nutritional control of the gametogenic
cycle in Mytilus edulis. Mar. Biol. 98, 223-228.
Lannan, J.E., 1980. Broodstock management of Crassostrea gigas: I. genetic and environmental
variation in survival in the larval rearing system. Aquaculture, 21, 323-336.
Pouvreau, S., Rambeau, M., Cochard, J.C., Robert, R., 2006. Investigation of marine bivalve
morphology by in vivo MR imaging: First anatomical results of a promising technique.
Aquaculture 259, 415-423.
Samain, J.F., Dégremont, L., Soletchnik, P., Haure, J., Bédier, E., Ropert, M., Moal, J., Huvet,
A., Bacca, H., Van Wormhoudt, A., Delaporte, M., Costil, K., Pouvreau, S., Lambert, C.,
Boulo, V., Soudant, P., Nicolas, J.L., Le Roux, F., Renault, T., Gagnaire, B., Geret, F.,
Boutet, I., Burgeot, T., Boudry, P., 2007. Genetically based resistance to summer mortality
in the Pacific oyster (Crassostrea gigas) and its relationship with physiological,
immunological characteristics and infection processes. Aquaculture 268, 227-243.

Abstract (if available)

Abstract The Pacific oyster is an important aquaculture species which suffers from reduced larval survival related to variation in state of gonad maturation among broodstock used for larval production. Gene expression was investigated as a potential indicator of reproductive stage. Stage-specific libraries were created by RNA sequencing, which were used for testing and identification of stage-specific, differentially expressed genes. Multiple factors related to the normalization and interpretation of gene expression data were queried, including the role of deadenylation of mRNA in developing oocytes. A set of candidate biomarker genes were screened for their correlation with and ability to predict maturation stage in female oysters. The results justify further development and use of expression data to identify maturation state.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Genetic architecture underlying variation in different traits in the Pacific oyster Crassostrea gigas

PDF

Genome-wide analysis of genetic load and larval mortality in a highly fecund marine invertebrate, the Pacific Oyster Crassostrea gigas

PDF

Simulated and field environmental effects on the transcriptome and metabolome of mussel Mytilus californianus

PDF

Genome-wide studies reveal the isoform selection of genes

PDF

Global analysis of the molecular activities defining maturation-stage amelogenesis

PDF

Studies in bivalve aquaculture: metallotoxicity, microbiome manipulations, and genomics & breeding programs with a focus on mutation rate

PDF

Phylogeography, reproductive isolation, and the evolution of sex determination mechanisms in the copepod Tigriopus californicus

PDF

Applications of next generation sequencing in sessile marine invertabrates

PDF

Molecular and behavioral mechanisms of circatidal biological rhythms in intertidal mollusks

PDF

Probing the genetic basis of gene expression variation through Bayesian analysis of allelic imbalance and transcriptome studies of oil palm interspecies hybrids

PDF

Transcriptional and morphological impacts of copper on Mytilus californianus larval development in current and future ocean conditions

PDF

Discovery of mature microRNA sequences within the protein- coding regions of global HIV-1 genomes: Predictions of novel mechanisms for viral infection and pathogenicity

PDF

Innovative sequencing techniques elucidate gene regulatory evolution in Drosophila

PDF

Transgenerational inheritance of thermal tolerance in two coral species in the Florida Keys

PDF

The evolution of pollution tolerance in the marine copepod Tigriopus

PDF

Probabilistic modeling and data integration to examine RNA-protein interactions

PDF

Transcriptomic maturation of developing human cone precursors in fetal and 3D hESC-derived tissues

PDF

Peripheral blood mononuclear cell capture and sequencing: optimization of a droplet based capture method and its applications

PDF

Using novel small molecule modulators as a tool to elucidate the role of the Myocyte Enhancer Factor 2 (MEF2) family of transcription factors in leukemia

PDF

Placenta growth factor-miRNAs-lncRNAs axis in the regulation of ET-1 gene involved in pulmonary hypertension in sickle cell disease

Asset Metadata

Creator Paul, Austin M. (author)

Core Title Stage-specific transcriptomic analyses of reproductive tissue in the Pacific oyster, Crassostrea gigas

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Biology

Publication Date 02/24/2014

Defense Date 12/13/2013

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

Tag biomarker,OAI-PMH Harvest,Pacific oyster,polyadenylated RNA,qPCR,reproduction,RNA-seq

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Hedgecock, Dennis (committee chair), Dean, Matthew D. (committee member), Edmands, Suzanne (committee member), Gracey, Andrew (committee member)

Creator Email austinpa@usc.edu,olymarmot@yahoo.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-365000

Unique identifier UC11296292

Identifier etd-PaulAustin-2265.pdf (filename),usctheses-c3-365000 (legacy record id)

Legacy Identifier etd-PaulAustin-2265.pdf

Dmrecord 365000

Document Type Dissertation

Format application/pdf (imt)

Rights Paul, Austin M.

Type texts

Source 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 a...

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