University of Southern California Dissertations and Theses

Mining the felinone A biosynthetic pathway

(USC Thesis Other)

Mining the felinone A biosynthetic pathway

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

MINING THE FELINONE A BIOSYNTHETIC PATHWAY

by

Yi-En Liao

A Thesis Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the Requirements for the Degree
MASTER OF SCIENCE (PHARMACEUTICAL SCIENCES)

May 2018

Copyright 2018 Yi-En Liao
ii

DEDICATION

I dedicate this work to my beloved parents and grandmother, Ching-Jong Liao, Hui-Ling Peng and
Yu-Bao Chen, and all my dear friends in Taiwan and the United States. Thank you for your priceless
love and support.

iii

ACKNOWLEDGEMENTS

First of all, I would like to express my deepest gratitude to my mentor Professor
Clay Wang. Without him, I would not have this precious experience as a master student at
USC school of pharmacy. He warmly welcomed me into his lab in 2016, and inspired me
with his talent, enthusiasm, patience, and confidence. As a successful department chair,
professor and principal investigator, he introduces me into the fascinating fungal world
with his knowledge, teaches me the craft of research, and gives me the opportunity to meet
other researchers around the world. This dissertation would have never been possible
without his guidance. Professor Wang is definitely the best mentor I could ever have, and I
will never forget his great support and encouragement from Taiwan to the United States.
Also, I am sincerely thankful for Professor Wei-Chiang Shen and Professor Curtis
Okamoto, who are willing to serve on my committees. I very appreciate them not only for
lectures but also for their assistance and guidance during my preparation of the dissertation.
In addition, I would like to give my special thanks to our collaborators Professor Berl
Oakley and Cory Benjamin, who taught me fungal biology and gave me valuable and
insightful advices for our collaborative projects.
Last but not the least, it’s my pleasure to work with all the fabulous lab mates in
Wang’s lab. I would like to give my big thanks to Dr. Yi-Ming Chiang and Dr. Kevin Lin,
who generously help me with all my projects. Also, I deeply appreciate Johannes Van Dijk,
Jillian Romsdahl, Michelle Grau and Adriana Blachowicz, who are always supportive in
the lab and help me throughout my graduate studies.

iv

TABLE OF CONTENTS
Dedication ..................................................................................................................... ii
Acknowledgements ...................................................................................................... iii
List of Tables ................................................................................................................ vi
List of Figures.............................................................................................................. vii
Abbreviations ............................................................................................................. viii
Abstract ......................................................................................................................... x
CHAPTER 1: Introduction ........................................................................................ 1
1.1 Natural Product Drug Discovery .......................................................................... 1
1.1.1 Current Status of Natural Product Drug Discovery………………………1
1.1.2 Secondary Metabolites .............................................................................. 2
1.2 Fungal Secondary Metabolites ............................................................................. 3
1.2.1 Advantages of Fungal Secondary Metabolites…………………………...3
1.2.2 Categories of Fungal Secondary Metabolites…………………………….4
1.3 Fungal Polyketide ............................................................................................... 4
1.3.1 Fungal Polyketide Synthases ..................................................................... 4
1.4 Secondary Metabolites in Aspergillus nidulans………………………………....9
1.4.1 Secondary Metabolite clusters…………………………………………..10
1.4.2 Recent Advances in Genome Mining of Aspergillus nidulans………….10
1.4.3 The AN7903 cluster ............................................................................... 12
CHAPTER 2: Materials and Methods .................................................................... 15
2.1 Chemicals and Media ......................................................................................... 15
2.2 Bioinformatic Analysis ........................................................................................... 16
2.2.1 BLAST .................................................................................................. 16
2.2.2 Gene Co-Expression .............................................................................. 16
2.3 Molecular Genetic Manipulations ......................................................................... 17
2.3.1 Fusion PCR and Transformation ........................................................... 17
v

2.3.2 Gene Overexpression ................................................................................. 17
2.3.3 Gene Deletion ......................................................................................... 18
2.3.4 Gene Positive Feedback System……………………………………….19
2.4 Fermentation of Aspergillus nidulans Strains .................................................... 26
2.5 Liquid Chromatography/Mass Spectrometer (LC/MS) ..................................... 26
CHAPTER 3: Results ................................................................................................ 28
3.1 Bioinformatics of the Cluster AN7903 .............................................................. 28
3.1.1 Gene Co-Expression…………………………………………………..28
3.1.2 BLAST analysis……………………………………………………….28
3.2 The Discovery of Genes for the Biosynthesis of Felinone A ............................. 28
3.3 Optimization of the Yield of Felinone A ........................................................... 29
3.4 Proposed Biosynthetic Pathway of Felinone A .................................................. 30
3.5 Determination of the genes responsible for Felinone A Production……………30
CHAPTER 4: Discussion .......................................................................................... 38
REFERENCES .......................................................................................................... 43
vi

LIST OF TABLES

Table 1: The Formulation of Media ........................................................................... 15
Table 2: The Reagents Using for Transformation ..................................................... 16
Table 3: A.nidulans Strains Used in This Study ........................................................ 20
Table 4: Primers Used in This Study ......................................................................... 21
Table 5: Pump Gradient Program .............................................................................. 27
Table 6: The BLAST Search on NCBI of the Genes in AN7903 Predicted Cluster in
A.nidulans ................................................................................................... 32

vii

LIST OF FIGURES
Figure 1: General Biosynthesis Mechanism of Polyketide Elongation Catalyzed by
Fungal Iterative Type I PKS ......................................................................... 6
Figure 2: Fungal PKS domain Structure ..................................................................... 7
Figure 3: Biosynthesis of Fungal HR-PKS and NR-PKS domains…………………13
Figure 4: Intermediate of the Gene Cluster AN7903 ................................................ 14
Figure 5: Single and Duel Promoter Replacement Strategy .................................... 22
Figure 6: Gene Deletion Strategy for Other Genes in the AN9703 cluster............... 24
Figure 7: The Construct of Feedback System ........................................................... 25
Figure 8: Prediction of the AN7903 gene cluster...................................................... 31
Figure 9: Results of Diagnostic PCR ........................................................................ 33
Figure 10: LC/MS Analysis of Promoter Replacement Strains ................................. 34
Figure 11: LC/MS Analysis of Feedback Strains ....................................................... 36
Figure 12: Proposed Biosynthetic Pathway for Felinone A ....................................... 37
Figure 13: The Scaffold of the Azaphilone Family .................................................... 40
Figure 14: Proposed Synthetic Reaction for Compound 1 .......................................... 40

viii

ABBREVIATIONS

HPLC: High Performance Liquid Chromatography
HTS: High-Throughput Screening
MS: Mass Spectrometry
NMR: Nuclear Magnetic Resonance
NRP: Non-Ribosomal Peptide
PKS: Polyketide Synthase
KS: Ketosynthase
FAS: Fatty Acid Synthase
AT: Acyltransferase
ACP: Acyl Carrier Protein
HR-PKS: High Reducing Polyketide Synthase
NR-PKS: Non Reducing Polyketide Synthase
PR-PKS: Partial Reducing Polyketide Synthase
DH: Dehydratase
ER: Enoyl Reductase
CMeT: C-Methyltransferase
SAM: S-Adenosylmethionine
SAT: Starter Unit: ACP transacylase
PT: Product Template
TE/CLC: Thioesterase/ Claisen Cyclase
MSAS: 6-Methylsalicylic Acid Synthase
DTS: Diterpene Synthase/ Cyclase
ix

DMATS: Prenyltransferases/ Dimethylallyltrptophan Synthase
SMURF: Secondary Metabolite Unknown Regions Finder
AntiSMASH: Antibiotics and Secondary Metabolite Analysis Shell
OSMAC: One Strain Many Compounds
GMM: Glucose Minimal Medium
BLAST: Basic Local Alignment Search Tool
CS: Clustering Scores
YAG: Yeast Extract Glucose
CDS: Coding Region
EtOAc: Ethyl Acetate
RT: Retention Time
TQ: terrequinone

ST: sterigmatocystin

LC/MS: Liquid Chromatography/ Mass Spectrometer
CSN: COP9 signalosome complex

x

ABSTRACT

Fungal secondary metabolites are an important source for drug discovery. Whole
genome sequencing has shown that there are thousands of secondary metabolite gene
clusters with unknown products. In this study, we focus on the biosynthetic pathway of
felinone A. Previous study, in the Wang lab the promoter of the nonreducing polyketide
synthase (NR-PKS) AN7903 was replaced and 2,4-dihydroxy-3-methyl-6-(2-oxopropyl)-
benzaldehyde (compound 1) was isolated. However, since we did not activate the
surrounding genes that may modify compound 1, the final product of this gene cluster
remains unknown. BLAST search and gene co-expression analysis show that AN7901 is a
transcription factor and AN7902 is a monooxygenase. We replaced the promoter of
AN7901, AN7902, and AN7903 and identified felinone A and its intermediate. Moreover,
we inserted a feedback system to increase the yield of compounds that we are interested.

1

CHAPTER 1
INTRODUCTION
1.1 Natural Product Drug Discovery

1.1.1 Current Status of Natural Product Drug Discovery

Since the first pharmacologically active compound morphine discovered and
purified from a plant, natural products have been served as a valuable resource for drug
discovery for over two centuries (Li and Vederas, 2009). From 1981 to 2014, it is reported
that around 40% of all newly approved drugs are either natural products, natural product
derivatives, or synthetic compounds based on the pharmacophore of a natural product
(Newman and Cragg, 2016). Although the trend of synthetic medicinal chemistry causes
the decline in pharmaceutical research on natural products in the past two decades, there
are some unique properties of natural products that make the study irreplaceable. For
example, natural products bear various pharmacophores and different stereochemistry.
Although they are not applied to first four rules of Lipinski’s fifth rule for druggable
chemical entities, they have provided hits for plentiful targets. Moreover, a recent study
shows that compounds successfully developed for drugs have the property of “metabolite-
likeness” (Harvey et al., 2015). Those properties give natural product additional
advantages over synthetic compounds.
Moreover, recently there are several technological advances revitalize natural
product drug discovery. To begin with, high-performance liquid chromatography (HPLC)
and other novel fractionation strategies enable researchers to isolate pure compounds or
remove artifacts from crude extracts of natural materials. This technique accelerates the
2

following bioactivity analysis by high-throughput screening (HTS) and structural analysis
by mass spectrometry (MS) and multidimensional nuclear magnetic resonance (NMR)
spectrometry. Moreover, applying metabolomics and metagenomics to natural-products
research improve the production of biologically active secondary metabolites (Harvey et
al., 2015). While metabolomics can make the process of profiling and isolation more
effectively by monitoring gene function and biochemical status of an organism,
metagenomics leads to explore new bioactive natural products by manipulating genes to
modify biosynthetic pathways and investigate cryptic gene clusters. Furthermore,
understanding the natural product biogenesis allows the use of combinatorial biosynthesis
or total synthesis to generate analogs, which can further develop to small focused
collection libraries. The process of producing more natural-product-like compounds is
called “diversity-oriented syntheses,” and has been proved working better than total
synthetic compound libraries for screening.
Overall, with the novel technologies, natural products discovered from natural
resources including plants, marine organisms, microorganisms are continuing sources of
novel drug discovery to date. Natural products can be found in the major disease treatment
such as cancer, hypertension, anti-diabetic and other over 50 approved drug therapies,
especially as anticancer and antimicrobial agents.
1.1.2 Secondary Metabolites

Most pharmaceutically relevant natural products are secondary metabolites
secreted by organisms. Unlike primary metabolites, secondary metabolites are not
involved in normal growth, development, and reproduction of a living organism.
Secondary metabolites are not essential for survival, but play an important role in response
3

to the environment, including defense, protection, and competition (Vining, 1992).
Common secondary metabolites that have been used for drugs involve alkaloids,
terpenoids, polyketides and fatty acids. Plants, bacteria, fungi and marine organisms are
the primary sources of these compounds and widely used in pharmaceutical research.

1.2 Fungal Secondary Metabolites

1.2.1 Advantages of Fungal Secondary Metabolites

In this study, we focus on fungi as the source of secondary metabolites. There are
several reasons for the choice. First, the fungal kingdom is a rich source of medically
important compounds. In a literature survey in 2005, more than half molecules of 1,500
compounds isolated and characterized from fungus between 1993 and 2001 have
biological activities such as antibacterial and antitumor (Palaez, 2005). Thus, it is possible
to develop novel therapeutics from those bioactive compounds. Second, there are many
successful examples on the market. Since the first antibiotics penicillin discovered from a
filamentous fungus, Penicillium notatum, by Fleming in 1929, many fungal secondary
metabolites are developed to approved drugs. Successful examples include an
anticholesteremic agent lovastatin, secreted by Aspergillus terreus, and
immunosuppressant cyclosporine, isolated from Trichoderma polysporum. Third, the
structural diversity of these fungal secondary metabolites can further inspire synthetic
chemistry. For example, the antifungal drug anidulafungin is a semisynthetic derivative of
the fungal metabolite echinocandin B (Mishra and Tiwari, 2011). A recently approved
multiple sclerosis therapeutics, fingolimod, is another example of synthetic compound
inspired by fungal secondary metabolites (Chun and Hartung, 2010).
4

In short, fungal secondary metabolites is a quite promising research area because
of the rich source, unique structures, and existing successful examples. With updated
techniques such as gene manipulation, it is possible to develop novel drug candidates from
the pool of fungal secondary metabolites.
1.2.2 Categories of Fungal Secondary Metabolites

The categorization of fungal secondary metabolites varies from researchers to
researchers. Some researchers classified fungal secondary metabolites based on their
biosynthetic origins, such as Turner and Aldridge; some researchers, on the other hand,
suggested a subdivision according to compounds’ medically related function. It is widely
accepted to categorize fungal secondary metabolites into four major classes: polyketides,
terpenes, alkaloids, and nonribosomal peptides (NRP).

1.3 Fungal Polyketide

In this study, we focus on one type of fungal secondary metabolites: polyketides.
Polyketide is the most abundant fungal secondary metabolites (Keller et al., 2005). More
importantly, many polyketides are human-related. There are more than 200 commercial
drugs among 7,000 polyketides with known structures (Li and Vederas, 2009), including
well-known cholesterol-lowering compound lovastatin. On the other hand, there are also
some polyketides cause detrimental health effects, such as the carcinogen aflatoxin B.
Polyketides include secondary metabolites with various structures, such as polyphenols,
polyenes, and macrolides.
1.3.1 Fungal Polyketide Synthases

All of the polyketides are biosynthesized by polyketide synthases (PKSs), which
5

can be categorized into three types based on the structure of enzymes. In addition to the
three subtypes, polyketides can also be classified as iterative or noniterative, depending on
the times of elongation during catalysis by ketosynthase (KS) domain (Hertweck, 2009).
The first two type of PKSs is similar to the classification of fatty acid synthases (FASs):
Type I PKSs refers to covalently linked catalytic domains within large multifunctional
enzymes, while type II PKSs refers to discrete catalytic components and usually
monofunctional enzymes. This two type of PKSs can be found in both bacteria and fungi.
Besides, enzymes belong to type III PKSs are also multifunctional but use different
substrate compared to type I and type II PKSs (Austin and Noel, 2003). A majority of type
III PKSs belongs to plants.
Most of the fungal polyketides are produced by iterative type I PKSs. However,
there are few exceptions. For example, lovastatin diketide synthase (LovF) and its
homolog compactin diketide synthase (MlcB) are noniterative type I PKSs. Furthermore,
some fungal type III PKSs are discovered by fungal genome sequencing projects recently.
A few examples are AnPKS from Aspergillus niger, CsyA and CsyB from Aspergillus
oryzae, and ORAS from Neurospora crassa (Hashimoto et al., 2014).

6

There are various domains in type I PKSs; each domain plays a role on
biosynthesis of polyketides. The minimal domain set of PKSs contain ketosynthase (KS),
acyltransferase (AT) and acyl carrier protein (ACP) (Chiang et al., 2010) (Figure 1). The
formation of a polyketide begins with loading of a starter acyl unit on KS domain. After
loading the starter unit, AT domain specifically select an extender unit, which is usually
malonyl-CoA, and translocate to ACP domain. Following with Claisen condensation
catalyzed by KS domain, the starter unit and extender unit bind and form a diketide
tethered to ACP domain. The diketide can translocate back to KS domain. Thus, ACP
Figure 1. General biosynthesis mechanism of polyketide elongation catalyzed by
fungal iterative type I PKS. Blue orbitals represent the minimal domains of type I PKS.
Dashed orbitals are the optional / non-essential domains. The diketide is formed after
Claisen condensation. Followed by translocating diketide to the KS domain, the reaction
can repeat.
7

domain can accept more malonyl extender units, repeat Claisen condensation, and grow
the polyketide chain. While fungal PKSs are usually iterative, the elongation process is
carried out repeatedly by the same domains.

In addition to the three essential domains in PKS, there are some “optional”
domains contributing to polyketide diversity. Based on the extent of 𝛽 -keto reduction
catalyzed by type I PKS, there are three major classes of type I PKS; and each class is
characterized with a specific combination of domain set (Chooi and Tang, 2012) (Figure
2). To begin with, highly reducing polyketide synthases (HR-PKSs) contain three
additional domains: ketoreductase (KR), dehydratase (DH), and enoyl reductase (ER).
Those domains enable HR-PKSs to generate many diverse linear or cyclic non-aromatic
compounds by undergoing different levels of 𝛽 -keto reduction. Moreover, compound
diversity is further increased by C-methyltransferase (CMeT) domain, which presents in
Figure 2. Fungal PKS domain structure. Blue orbitals represent the minimal domains
of type I PKS. Green, orange and purple orbitals are optional domain unique to a specific
type of PKS.
8

many HR-PKSs and uses S-adenosylmethionine (SAM) to methylate ∝-carbon. On the
other hand, those PKS without the trio 𝛽 -keto processing domains are classified as
nonreducing polyketide synthases (NR-PKSs). Unlike HR-PKSs, NR-PKSs produce
mono- or polycyclic aromatic polyketides by aldol condensation or keto-enol
tautomerization with another domain set. The domains unique to the NR-PKSs include the
starter unit:ACP transacylase (SAT) domain, which afford the PKS selectivity for the acyl
starter unit; the product template (PT) domain, which regulates aldol cyclization and
aromatization of polyketides; and the thioesterase/Claisen cyclase (TE/CLC), which
catalyzes macrolactonization or macrolactamization reactions to release the polyketide.
Besides TE/CLC, some NR-PKSs use nearby 𝛽 -lactamase or a C-terminal reductase (R)
release domain to off-loading the products. Finally, the third class, partial-reducing
polyketide synthases (PR-PKSs) is less abundant than HR-PKSs and NR-PKSs. The only
characterized PR-PKSs is the 6-methylsalicylic acid synthase (MSAS), containing a single
reduction of a specific keto group in a tetraketide backbone. Besides, it is shown that PR-
PKSs is more related to bacterial type I PKSs than HR- or NR-PKS in fungi (Schmitt et
al., 2005). Thus, in this study, we will focus on fungal NR-PKSs and HR-PKSs.
In short, the structural diversity of polyketides is contributed by chain length
control, the level of 𝛽 -keto reduction, ∝-methylation, starter unit choice, cyclization
pattern through different functional domains combination in PKSs (Figure 3). It is also
important to mention that a polyketide can not only be produced by a single PKS, but by
dual PKS systems or by mixing PKSs with other secondary metabolite synthases, such as
fatty acid synthases (FASs) and non-ribosomal peptide synthases (NRPSs). For example,
9

there are two PKSs in asperfuranone pathway: AfoG (HR-PKS) and AfoE (NR-
PKS)(Chiang et al., 2009). By combining more than one secondary metabolite synthases,
the structural diversity is further expanded.
1.4 Secondary metabolites in Aspergillus nidulans

Among all the fungi producing secondary metabolites, the genus Apsergillus may
be the most human-related species. Some of them are commercially useful, such as A.
niger has been exploited to produce citric acid in industries and A. oryzae is used for the
production of miso and soy sauces. Some of the species are detrimental, like A. fumigatus
(Machida et al., 2005). A. nidulans, on the other hand, is not industrial cell factories nor
pathogens. It is the critical model organism for fungal research for more than half a
century. Compared to other aspergilli, A. nidulans has a well-characterized sexual cycle,
and is productive from aseacual state, sexual to parasexual life cycles (Pontecorvo et al.,
1953). Those features offer an ideal platform for fungal genetic studies. Therefore, A.
nidulans has been widely used for research including cell cycle control, DNA repair and
pathogenesis, and pathogenesis. Also, regulation of fungal secondary metabolites, such as
penicillin and aflatoxin, is elucidated by work on A. nidulans.
A. nidulans itself produce lots of secondary metabolites, too. An investigation
using bioinformatics, manual and experimental methods in 2013 showed that 72 genes
encoding secondary metabolite biosynthetic enzymes in A. nidulans (Inglis et al., 2013).
Most of these enzymes in A. nidulans belong to four classes, including NRPS, PKS,
diterpene synthases/cyclases (DTSs) and prenyltransferases/dimethylallyltrptophan
synthases (DMATSs). Among our interested class, PKS, there are 27 PKS proteins (14
NR-PKSs, and 12 HR-PKSs), a single hybrid PKS/NRPS and 5 PKS-like enzymes.
10

1.4.1 Secondary metabolites clusters

Although PKSs and other secondary metabolite biosynthetic enzymes are essential
to generate natural products, there are also other important genes involved in the
biosynthetic pathways. After the core structure of secondary metabolites synthesized by
PKS or NRPS, other genes encoding tailoring enzymes further modify the structure. Those
enzymes include oxidase, hydroxylase, cyclase, and decarboxylase, just to name a few.
Genes encoding transporters are also needed when the product is completely synthesized,
ready for export from the cell. Furthermore, biosynthesis of some natural products is
controlled by transcription regulator, which encoded by a gene that can activate or silence
the production.
How to investigate the biosynthesis pathways and the final products of secondary
metabolites while there are so many possible genes involved? While it is difficult for
research in other eukaryotes, fungi have a characteristic that highly facilitate the studies of
its secondary metabolites: genes involved in a fungal secondary metabolites biosynthetic
pathway are always clustered (Keller et al., 2005), including A. nidulans. After the
discovery of the core genes like PKS, genes encoding enzymes that further modify or
transport the compound can be identified by manipulate those genes within proximity.
However, another characteristic of fungal secondary metabolites makes the genetic
research challenging. Despite the grouping of biosynthetic pathway genes in a contiguous
cluster, many of the secondary metabolism gene clusters are silent under standard
laboratory conditions. Thus, few compounds can be identified.
1.4.2 Recent Advances in Genome Mining of A. nidulans

In recent year, many strategies with novel technology are developed to advance the
11

research on fungal secondary metabolites. First of all, the genome sequencing of
Aspergillus species reveals abundant information about secondary metabolites genes. The
sequencing project of A. nidulans was done in 2005 (Galagan et al., 2005). Using
sequence similarity to known genes from other species to mine core secondary metabolite
genes in A. nidulans, several useful bioinformatics tools, such as SMURF (Secondary
Metabolite Unknown Regions Finder) and antiSMASH (antibiotics and Secondary
Metabolite Analysis Shell) have been developed (Yaegashi et al., 2014).
Second, various novel approaches have been exploited to activate silent clusters in
A. nidulans and other species. For example, new metabolites can be discovered by altering
the culture condition of A. nidulans, which is the strategy called one strain many
compounds (OSMAC). Besides, interactions between fungi and other microorganisms
such as bacteria by co-incubation can also trigger the activation of silent gene clusters.
Orsellinic acid was identified by this two approaches mentioned above (Sanchez et al.,
2010; Schroeckh et al., 2009).
In addition to changing cultural environment, there are some promising strategies
involving gene manipulation at epigenetic and transcription level. For instance, removal of
genes required for chromatin packing (Bok et al., 2009) to induce expression of gene
clusters. Some approaches target at genes that regulate one or more gene clusters.
Examples include genome-wide analysis of mutant of LaeA and McrA, global regulators
of secondary metabolite clusters (Oakley et al., 2017), and overexpression of a pathway-
specific transcription factor by fusing inducible promoters. If the transcription factor is the
primary regulator of the gene cluster, the promoter replacement can activate the whole
cluster to produce a secondary metabolite final product.
12

However, sometimes induction of the transcription factors did not result in produce
enough secondary metabolites. In this case, promoter replacement can also be used to
activate other genes, such as PKS in the pathway cluster one by one, which enables the
identification of new natural products and its biosynthetic pathways. This strategy has
been proved successful. By replacing promoters of some NR-PKSs in A. nidulans, our lab
determined seven novel compounds in 2012 (Ahuja et al., 2012).
Besides genome sequencing and new strategies to activate silent clusters, the
efficiency of gene targeting system has also been improved recently. First, the
development of fusion PCR facilitates transformation procedure. Compared with the
traditional method, plasmid transformation, fusion PCR saves lots of time and give higher
successful rate (Yang et al., 2004). Second, the deletion of the nkuA gene in the recipient
strain further increase the success rate of transformation. By knocking out nkuA, the
strains are deficient in nonhomologous end joining, thus greatly reduce the frequency of
heterologous integration of transforming DNA during transformation. Third, a procedure
of producing protoplasts from hyphae lower the barrier to insert genes in A.nidulans
through its think cell wall (Szewczyk et al., 2006). All of the three gene engineering
techniques are applied in this study.
1.4.3 The AN7903 cluster

In this study, we focus on one silent secondary metabolite gene cluster in A.
nidulans, which possess an NR-PKS AN7903. In our previous study, we replaced the
promoter of AN7903, and isolated 2,4-dihydroxy-3-methyl-6-(2-oxopropyl)-benzaldehyde
(compound 1, Figure 4) (Ahuja et al., 2012). However, since we did not activate the
surrounding genes that may modify compound 1, the final product of this gene cluster
13

remains unknown. Therefore, we used the approaches mentioned above to activate genes
nearby PKS AN7903, and further identified new compounds and increase their yield from
the cluster.

Figure 3. Biosynthesis of fungal HR-PKS and NR-PKS domains.
14

Figure 4. Intermediate of the Gene Cluster AN7903.
15

CHAPTER 2
MATERIALS AND METHODS
2.1 Chemical and Media

The media used in this project are as follows (Table 1).
Table 1. The Formulation of Media
Media/ Plate Ingredient L
-1
Concentration
GMM Medium
Plate

Dextrose 10g
1%

Agar 15g
1.5%
Liquid LMM
Medium

Lactose 10g
1%
20x salts solution
50mL

NaNO 3
6g

KCl
0.52g

MgSO 4 ・ 7H 2O
0.52g

KH 2PO 4
1.52g

Hunter’s trace element
1mL

Hunter’s trace element
1 mL

ZnSO 4 ・ 7H 2O
2.2g
7.7𝜇 M
H 3BO 3
1.1g
18 𝜇 M
MnCl 2 ・ 4H 2O
0.5g
2.5 𝜇 M
FeSO 4 ・ 7H 2O
0.5g
1.8 𝜇 M
CoCl 2 ・ 5H 2O
0.16g
0.64 𝜇 M
CuSO 4 ・ 5H 2O
0.16g
0.67 𝜇 M
(NH 4) 6Mo 7O 24 ・ 4H 2O
0.11g
0.09 𝜇 M
NaEDTA
5g
13 𝜇 M
KOH
1 mL

ddH 2O
1L

Supplement*

Riboflavin B 2.5mg

Uracil 1g

Uridine (100x) 10mL 10mM

Pyridoxine (1000x) 1mL

KCl 44.7g 0.6 M
*Supplement riboflavin B, uracil and uridine, pyridoxine are added to GMM plates or
LMM medium depending on selective markers in the strain. Uridine should be added after
autoclaving the medium. KCl is added to plates for transformation.
16

The reagents using in transformation are described in Table 2:
Table 2. The Reagents using for transformation
Reagent Ingredient Total volume
1.1M KOH solution 6.17g KOH 100mL ddH 2O
KCl, citric acid solution (pH 5.8) 8.2g KCl, 2.1g citric acid 100mL ddH 2O
1.2 M sucrose solution 41.08g sucrose 100mL ddH 2O
0.6 M KCl, 50mM CaCl 2 solution 4.47g KCl, 0.74g CaCl 2 ・ 2H 2O 100mL ddH 2O
PEG solution (pH 7.5)
4.47g KCl, 0.74g CaCl 2 ・ 2H 2O
0.802mL 1M Tris-HCl
0.196mL 1M Tris base
25g PEG
100mL ddH 2O
2x protoplasting solution 2g Vinoflow FCE
10mL KCL, citric acid
solution

2.2 Bioinformatic Analysis
2.2.1. BLAST
BLAST (Basic Local Alignment Search Tool) searches were performed at NCBI
(http://www.ncbi.nlm.nih.gov/) to find regions of similarity between A.nidulans (FGSC
A4) sequence and other biological sequences. In this study, we analyzed sequence in non-
redundant protein sequence database by the algorithm blastp (protein-protein BLAST).
Closest homologs of our interested genes with published functions were described in later
chapters.
2.2.2. Gene Co-Expression
Recently Anderson et al. developed a DNA expression array to identify genes
that were co-expressed with secondary metabolite gene cluster backbone enzymes
(Andersen et al., 2013). Various culture media were used to optimize the expression of
gene clusters. Samples were collected from selected media for transcriptional profiling.
17

To analyze the data, Anderson et al. developed clustering scores (CS) that reflected the
degree to which each gene was co-regulated with its neighbors. This method has been
proved as a novel strategy for the accurate prediction of secondary metabolite gene
cluster boundaries. The algorithm was used to predict the boundaries of 58 secondary
metabolite gene clusters in A. nidulans, including our interested cluster AN7903.
2.3 Molecular Genetic Manipulations
2.3.1. Fusion PCR and Transformation
Construction of fusion PCR products, protoplast production and transformation
were described above (Szewczyk et al., 2006). The fusion PCR cassette design for gene
overexpression and gene deletion is slightly different, as indicated in the following
passage. After obtained the desired PCR fragment, the recipient strain was cultured in
YAG medium (5g yeast extract, 20 g dextrose and 1mL trace element in 1L ddH 2O) for
preparation of making protoplasts. To increase the success rate of transformation, it is
critical to have enough amount of protoplasts and fusion PCR fragments. Usually, we use
5 × 10
6
protoplasts with 2 to 3 𝜇 g DNA for a single transformation. By changing the
supplement in plates and the selectable markers in the construct of PCR fragments, correct
transformants can be selected, grew, and tested their sequence by diagnostic PCR.
All the strains and primers we used in this study are listed below (Table 3 and 4).
2.3.2. Gene Overexpression
In this study, we activated three genes in the cluster by promoter replacement. To
replace the native promoter with ethanol inducible promoter alcA, the fusion PCR cassette
is designed and transformed based on the strategies developed in our lab in 2012 (Figure 5)
(Ahuja et al., 2012). For gene AN7903 and AN7901, the NR-PKS and putative transcription
factor, we use single promoter replacement in figure 5A. First of all, the transforming
18

sequence contains a flanking region (about 1kb), selectable marker AfpyrG, alcA, and our
targeted gene (i). 6 primers are designed and annotated as P1-P6; P3 and P4 have 50%
regions homologous to the selectable marker pyrG or alcA, respectively; and P2 to P3, P4
to P5 should be 1kb apart to generate appropriate homologous flanking regions. P1 and P6
are the next suitable primer sequence nearby P2 and P5. Second, fragments are synthesized
by genomic PCR with primer pair P1/P3 and P4/P6 (ii). The two fragments can be fused to
one piece of transforming sequence by fusion PCR with primers P2/P5 (iii). Once the
fusion PCR product is completed, it can be transformed into the recipient strain. The
targeted gene then become regulatable after homologous recombination of the fragment
with the chromosomal locus (iv).
For gene AN7902 and AN7903, we use dual promoter replacement strategy
described in figure 5B. Because the transcriptional direction of these two nearby genes are
opposite, it is more efficient to design a cassette with one selectable marker and two alcA
controlling two genes respectively (i). Thus, only one transformation is needed to replace
promoters of two genes. While the logic of primer design is similar to the AN7901 cassette,
the fusion PCR is different. Two fragments are generated from fusion PCR, each with a
portion of one targeted gene (about 1kb) fused with alcA and more than half portion of the
selectable marker (iii). During transformation, the overlapping marker sequence from the
two fragments fused by homologous recombination, as well as the transformant and the
chromosomal targeted gene. With the correct transformant, the strain would have the
sequence with two genes driven by two alcA with one selectable marker (iv).
2.3.3. Gene Deletion
Gene deletion uses the same procedure of primer design, fusion PCR, and
transformation. Instead of inserting promoter genes, targeted genes were knock out and
19

replaced with a selectable marker (Figure 6). By gene deletion, we may learn the function
of the targeted gene, the boundary of the secondary metabolite cluster. We may also
identify intermediates or shunt products, which can lead us to discover the final product or
understand its biosynthetic pathway (Chiang et al., 2008).
2.3.4. Gene Positive Feedback System
To increase the yield of the compounds we are interested in the cluster, we
collaborate with Dr. Berl Oakley’s lab at the University of Kansas to insert a gene positive
feedback system in our strain. As indicated in the figure, a cassette including the coding
region (CDS) of alcR, promoter alcA and Tef was put in the strain alcA-AN7903; alcA-
AN7902 (Figure 7). In contrast to alcA promoter, Tef has strong promoter activity and
constitutively expresses genes(Kitamoto et al., 1998). Thus, alcR transcription factor is
constitutively transcribed from alcR coding region (CDS) controlled by Tef. AlcR
transcription factor then binds to the three alcA promoters in this strain, activating the
transcription of AN7902, AN7903 and the second alcR (CDS). Consequently, the second
alcR(CDS) transcribe to the second alcR transcription factor, binding the same three alcA
promoters. In short, alcR transcription factor is amplified by the feedback loop, and turn up
the alcA-AN7902 and alcA-AN7903 even higher. In addition to alcA promoter, we also test
another promoter aldA by replacing alcA. In contrast to alcA, aldA is not repressible by
sugar. While whether alcA or aldA promoter is stronger is still on debate, we made this two
strains and tested them both.
We received the feedback strains from Dr. Berl’s lab, and cultured it with the
parental strain we made for 6 days. Since feedback strains do not grow very well after
induction, we inoculated the strains one more day to grow enough hyphae in order to
generate more compounds.
20

Table 3. A. nidulans Strains Used in This study.

Strain Genotype Source
stcJ∆
LO4389
pyrG89; pyroA4; nkuA::argB; riboB2; stcA-WΔ (Chiang et al., 2008)
stcJ∆,
alcA_ANID_7903
YL021
pyrG89; pyroA4; nkuA::argB; riboB2;
stcJ::AfriboB; ANID_7903::AfpyrG-alcA(p)-
ANID_7903
This study
stcJ∆,
alcA_ANID_7901
YL021, YL024,
YL096
pyrG89; pyroA4; nkuA::argB; riboB2;
stcJ::AfriboB; AN7901.2::AfpyrG-alcA(p)-
AN7901.2
This study
alcA-ANID_7902;
alcA-ANID_7903
YL058, YL059,
YL082
pyrG89; pyroA4; nkuA::argB; riboB2;
stcJ::AfriboB;
ANID_7902::AfpyrG-alcA(p)-ANID_7903;
This study
alcA_alcR_Tef_alcR
LO11334
yA-aldA(p)-alcR(CDS)-Afpyro-Tef(p)-
alcR(CDS)-yA; pyroA4; alcA(p)-AN7902-
AfpyrG-alcA(p)-AN7903; pyrG89; riboB2;
ΔstcJ; ΔnkuA
This study
aldA_alcR_Tef_alcR
LO11338
yA-alcA(p)-alcR(CDS)-Afpyro-Tef(p)-
alcR(CDS)-yA; pyroA4; alcA(p)-AN7902-
AfpyrG-alcA(p)-AN7903; pyrG89; riboB2;
ΔstcJ; ΔnkuA
This study
ANID_7901Δ
YL101, YL102,
YL103
pyrG89; pyroA4; nkuA::argB; riboB2;
stcJ::AfriboB;
ANID_7902::AfpyrG-alcA(p)-ANID_7903;
ANID_7901::AfriboB
This study

21

Table 4. Primers used in this study.

Primer Sequence (5’  3’)
alcA_ANID_7903_P1
alcA_ANID_7903_P2
alcA_ANID_7903_P3

alcA_ANID_7903_P4

alcA_ANID_7903_P5
alcA_ANID_7903_P6
GCT CCA GAT GAC AAC CAA TG
GTACGACAGACATCAGCAGC
GAA GAG GGT GAA GAG CAT TGT TTC CTG CCT GTT CTA
GTG C
TCC TAT CAC CTC GCC TCA AAA TGC TTG GTC ATC GGG
ACT T
CTCTCCAGTAGCATTGGATC
GAAGCTCTCGAACCACTGAG
alcA_AN7901.2_P1
alcA_ AN7901.2_P2
alcA_ AN7901.2_P3

alcA_ AN7901.2_P4

alcA_ AN7901.2_P5
alcA_ AN7901.2_P6
CGT ACC TGG TAG GAC CAG AA
CTA TGG AAC CAG CAT TGG TC
GAA GAG GGT GAA GAG CAT TGG GTG TCT GTG CCT GGC
ACC G
TCC TAT CAC CTC GCC TCA AAA TGC CTG AGG ACG GAC
CCC C
TAGGTATCGACTTGACCTGG
AGTTCGCGAACGTCGTTGTG
alcA_ANID_7902_P1
alcA_ ANID_7902_P2
alcA_ ANID_7902_P3

alcA_ANID_7903_P4

alcA_ANID_7903_P5
alcA_ANID_7903_P6
TAGCCACTGACTCCTAGTCC
GACCTGCTTCATCAGAGTGG
TCC TAT CAC CTC GCC TCA AAA TGC CAG GTA CCG TGC
GCC C
TCC TAT CAC CTC GCC TCA AAA TGC TTG GTC ATC GGG
ACT T
CTCTCCAGTAGCATTGGATC
GAAGCTCTCGAACCACTGAG
ANID_7901Δ_P1
ANID_7901Δ_P2
ANID_7901Δ_P3

ANID_7901Δ_P4

ANID_7901Δ_P5
ANID_7901Δ_P6
GTA CCT GGT AGG ACC AGA AC
CTA TGG AAC CAG CAT TGG TC
GAA GAG GGT GAA GAG CAT TGG GTG TCT GTG CCT GGC
ACC G
TCA GTG CCT CCT CTC AGA CAG TCC AGC TGA ATT AGC
ACT A
GAT GCT GGT CAG TCC TTC TA
GTA TCG AGG AGC GAG GAC TA

Sequences marked in blue, red and green are tails annealing to A. fumigatus pyrG (AfpyrG) cassettes, A.
nidulans alcA (alcohol dehydrogenase) and A. fumigatus riboB (AfriboB) cassettes that we have
constructed in previous study.

22

Figure 5-1. Single Promoter replacement strategy for AN7901, AN7903.
The procedure is described as above. After replacing the native promoter with alcA, the
promoter of alcohol dehydrogenase, gene expression can be repressed by glucose and
activated by adding inducer such as methylethylketone (MEK). AfpyrG is the selective
marker, which enable the strain with the transforming fragment to grow in the absence of
uridine and uracil.
.
23

Figure 5-2. Duel Promoter replacement strategy for AN7902 and AN7903.
The procedure is described above. With two alcA replacing the native promoters of two
genes, the genes can be activated at once with inducer.
24

Figure 6. Gene deletion strategy for other genes in the AN7903 cluster.
The procedure is described above. Compared to gene replacement, there is no alcA needed to
activate. Besides, because the gene deletion conducted in alcA-AN7902-AfpyrG-alcA-AN7903
strain, we use another selective marker, riboB. The gene sequence riboB enable successful
strains grow without riboflavin B supplement.
25

Figure 7. The construct of feedback system. A. Construct for LO11338: feedback system
with alcR driving by alcA protmoter. B. Construct for LO11334: feedback system with
alcR driving by aldA promoter.

A.
B.
26

2.4 Fermentation of Aspergillus nidulans Strains
The strains with correct transforming sequence, checked by diagnostic PCR, were
cultivated by the following procedure. For gene overexpression, strains were inoculated in
30mL LMM medium with necessary supplements. The concentration is 1 × 10
6

spores/mL, and the shaking speed is 180 rpm at 37℃. 134 𝜇 L methylethylketone (MEK)
(50mM) was added to the culture medium after 42 hours’ inoculation to induce alcA promoter.
After 72 hours’ induction, the culture medium was collected, filtrated, and extracted by an
equivalent volume of EtOAc (ethyl acetate). The medium was later acidified by HCl (pH
2~3) and extracted by EtOAc to collect charged compounds. The mycelium was soaked in
50mL MeOH in the culture flasks for 24 hours, filtrated, and extracted by the same
procedure of EtOAc extraction. Subsequently, the EtOAc layers from medium, acidified
medium and mycelium were evaporated by TurboVap Evaporation System (Caliper
LifeSciences). The dry crude products were collected, and yield was calculated by
measuring the glass tube before and after evaporation. Finally, the products were
redissolved in 0.5 mL or 1mL of DMSO/MeOH (1:4) depending on the yield, and inject
2~20 𝜇 L for LC/MS analysis.
For gene deletion, the fermentation procedure was applied as similar as gene
overexpression. The only difference is that there is no need for induction; therefore, the
strain was cultivated in an incubator for 5 days.
2.5 Liquid Chromatography / Mass Spectrometry (LC/MS)
The detection method used in this study is a single stage LC/MS carried out by a
ThermoFinnigan LCQ Advantage ion trap mass spectrometer with an RP C18 column
(Alltech Prevail C18 3μm 2.1×100 mm).
For the LC gradient program setting, two solvents were used: solvent A (5%
27

MeCN/H2O) and solvent B (95% MeCN/H2O), both with 0.05% formic acid. In 50
minutes, the flow rate is fixed at 125 μL/min under low pressure (0~400 psi). The gradient
was shown below (Table 5).
For the MS, type of ionization techniques is Electrospray Ionization (ESI). The
ionization conditions were adjusted at 5.0 kV and 275℃, for capillary voltage and
temperature, respectively. The sheath gas flow rate and the auxiliary gas flow rate was set
at 50 arbitrary units. The full scan mass was set from m/z 100 to m/z 1500. All MS data
were recorded in both negative and positive mode. Total ion chromatograms were also
recorded during the full scan mass spectra event. LC/MS is a very sensitive detection
method, which can detect samples with only 10 ng, and provide us information including
exact mass weight, approximate polarity, and purity of compounds (Waridel et al., 2004).
In addition, our LC/ mass spectrometer also equipped an in-line diode array
detector (DAD), which allows us to analyze samples with its broad ultraviolet-visible
(UV/Vis) wavelength spectra, from 200 to 600 nm. Because our interested compounds are
usually UV active, the detection results provide us additional structural information, such
as chromophores present in a compound (Hansen et al., 2005).
Table 5. Pump Gradient Program
No. Time (min) Solvent A (%) Solvent B (%)
0 0.0 100.0 0.0
1 5.0 100.0 0.0
2 35.0 0.0 100.0
3 40.0 0.0 100.0
4 45.0 100.0 0.0
5 50.0 100.0 0.0

28

CHAPTER 3
RESULTS
3.1 Bioinformatics Analysis of the Cluster AN7903
3.1.1. Gene Co-Expression
According to the analysis of gene co-regulation mentioned above (Andersen et al.,
2013), a proposed boundary of the cluster AN7903 was made (Figure 8). The prediction
was based on 44 experiments as indicated in X-axis, including cultivation on liquid and
solid medium. 12 genes are expressed concurrently with the core gene PKS AN7903.
3.1.2. BLAST analysis
To predict the boundary of the cluster and functions of the genes involved, we
conducted BLAST analysis to assign genes. Fifteen co-regulated genes, based on the
analysis mentioned above, were examined (Table 6). While we indicate the closest
homologs with published function found in a BLAST search of most genes, some genes
that have homologs with known genes are also described. Moreover, we found that two
genes are homolog to the genes involved in tropolone biosynthetic pathway. Thus, we
designed our promoter replacement experiment according to the BLAST search.
3.2 The Discovery of Gene for Biosynthesis of Felinone A
First of all, we replaced the native promoter of AN7903 with the regulatable
promoter alcA. The selective replacement of the promoter was verified using diagnostic
PCR (Figure 9). To optimize the result, we took a sterigmatocystin deficient (stcJΔ) A.
nidulans strain, LO4389, as our recipient strain. StcJ is the major polyketide produced by
A. nidulans; thus, knocking out its sequence free up materials for polyketide biosynthesis,
such as malonyl-CoA to be used by other secondary metabolite gene cluster. The resulting
HPLC profile showed that there was no stcJ in both parental and alcA strain (Figure 10, i
29

and ii). Furthermore, a new peak appeared in alcA-AN7903 strain, which was identified as
compound 1 by UV/Vis and MS analysis.
Second, we overexpressed AN7901, the predicted transcription factor, in order to
active the whole gene cluster. The promoter replacement was validated by diagnostic PCR
(Figure 9, C). However, compared to alcA-AN7903 and the parental strain, there were no
new compounds generated in HPLC profile. Moreover, the production of compound 1 is
less than alcA-AN7903 strain (Figure 10, iii). As a result, AN7901 is not regarded as the
positive transcription factor of the gene cluster. We then decided to activate a nearby gene,
AN7902, which is a homolog to tropB, to investigate the gene cluster.
The third promoter replacement of the gene in the cluster was carried out on
AN7902 (Figure 9). In contrast to alcA-AN7903, amount of compound 1 decreased and
other two new compounds were produced and detected by HPLC (Figure 10, iv). Based on
UV/Vis and MS analysis, the two peaks at retention time (RT) 13 min and 18 min were
felinone A and 6,7-dihydroxy-3,7-dimethyl-hexahydroisochromenone (assigned as
compound 2), respectively. While compound 2 is an unknown compound, felinone A has
been identified as an antibiotic (Du et al., 2014). Thus, we decided to focus on the study of
felinone A in the AN7903 cluster.
3.3 Optimization of the Yield of Felinone A
Ten strains were received from Dr. Oakley’s lab, including five strains with alcA
promoter the and five strains with the aldA promoter regulating the coding region of alcR.
After cultivation and LC/MS analysis, three strains with aldA promoter and one strain with
alcA promoter showed higher yield of compound 1 and 2 than the parental strain. The
strains gave highest yield are shown (Figure 11).

30

3.4 Proposed Biosynthetic Pathway of Felinone A
With the bioinformatics analysis and compounds identified by our study, we
proposed a biosynthetic pathway of felinone A (Figure 12). First of all, the NR-PKS
AN7903, the homolog to tropA, accepts acetyl-CoA as the starter unit and catalyzes three
iterations of malonyl-CoA extension and four iterations of SAM-dependent methylation.
Next, AN7903, a FAD-dependent monooxygenase homolog to tropB, further hydroxylates
and dearomatizes 3-methylorcinaldehyde. Since we detect felinone A without over-
expressing other genes than AN7902 and AN7903, we proposed that an endogenous
reductase generate felinone A by reduction of the intermediate produced by AN7902. By
our culture LC/MS result, we found three compounds in the medium. On the other hand,
AN7901 is not the main transcription factor regulating the pathway, and is not shown in the
figure.
3.5 Determination of the genes responsible for Felinone A Production
We deleted genes on both side in our alcA-AN7902; alcA-AN7903 strain, from
AN7893 to AN7901, AN7904 to AN7905. The confirmation of one of gene knock-out,
KO-AN7901, is carried out by diagnostic PCR (Figure 9, D). The strains are currently
cultured and examined by LC/MS. The result will be presented in due course.

31

B.
Figure 8. Prediction of the AN7903 gene cluster.
(A) Gene expression profiles from 44 experiment for the 12 genes (ranging from AN7893 to
AN7903) that co-regulated and predicted to be in the AN7903 cluster. The expression profile
of AN7903 is marked in red; other two genes that was manipulated in this study, AN7901 and
AN7902, are marked in yellow and blue, respectively. The figure is reproduced from Anderson
et al. (Andersen et al., 2013) (B) Illustrate of CS scores of the 12 genes in AN7903 cluster.
Genes not included in the predicted cluster are marked in gray.
32

Table 6. The BLAST search on NCBI of the gene in AN7903 predicted cluster in A. nidulans.
ORFs Homologs
Gene Predicted Size
Gene(bp)/ Protein (aa)
Match from BLAST search at NCBI
(accession no.)
Similarity/
Identity
AN7892 555/184 1.

2.
putative 30 kDa heat shock protein
Aspergillus calidoustus
IbpA
Aspergillus parasiticus SU-1
76/68

73/60
AN7893 3031/759 1.

2.

3.
leucoanthocyanidin dioxygenase
Talaromyces stipitatus (ATCC 10500)
Ctn A, citrinin biosynthesis oxygenase
Coccidioides immitis RS
TropC, non-heme iron II
Talaromyces stipitatus (ATCC 10500)
99/94

78/70

72/58
AN7894 863/125 YCII- domain protein
Metarhizium brunneum (ARSEF 329)
65/43
AN7895 1026/314 Cip B zinc-binding oxidoreductase
Talaromyces stipitatus (ATCC 10500)
86/83
AN7896 2045/560 TPA: Putative Zn(II)2Cys6 transcription factor
Aspergillus nidulans (FGSC A4)
96/96
AN7897 1281/393 monooxygenase
Talaromyces stipitatus (ATCC 10500)
84/79
AN7898 1539/444 monocarboxylate transporter
Talaromyces stipitatus (ATCC 10500)
89/81
AN7899 837/278 CtnB, citrinin biosynthesis oxygenase
Coccidioides immitis RS
75/61
AN7900 1404/467 amine oxidase flavin-containing superfamily
Hypoxylon sp. EC38
65/50
AN7901 1502/420 1.

2.
C6 transcription factor
Talaromyces stipitatus (ATCC 10500)
Moc3, transcriptional regulatory protein
Aspergillus fumigatus Z5
84/73

55/37
AN7902 1453/462 1.

2.
monooxygenase
Talaromyces stipitatus (ATCC 10500)
TropB, FAD monooxygenase
Talaromyces stipitatus (ATCC 10500)
92/85

74/57
AN7903
(PkeA)
7818/2605 TropA, Nonreducing Polyketide synthase
Talaromyces stipitatus (ATCC 10500)
89/82
AN7904 1264/325 Methyltransferase domain-containing protein
Cladophialophora carrionii

AN7905 882/293 NAD binding domain of
6-phosphogluconate dehydrogenase
Aspergillus parasiticus SU-1
75/60
33

Closest homologs with putative functions or known genes are listed. Known genes are bolded.

Figure 9. (A)~(C) Results of diagnostic PCR for the alcA promoter replacement strains and
WT (LO4389). (D) Results of diagnostic PCR for the knock-out strains and parental strain
(YL082). PCR Fragments were amplified with two primers, P1 from the chromosomal region
outside of 5’ flanking region of the transformed DNA fragment. P6, similarly, from the
region outside of the 3’ flanking region.

A. alcA-AN7903 B. alcA-AN7901

C. alcA-AN7902, alcA-AN7903 D. KO AN7901

34

Figure 10-1. LC/MS analysis of alcA promoter-replaced strains.
Total scan HPLC profiles of secondary metabolites extracted from the culture medium of A. nidulans.
Identified compounds 1~3 are labeled and shown in figure 10-2. Compounds that are not related to the
cluster AN7903 are indicated with asterisk (*) and cross (+), which represent lumichrome and
emericellamide A, respectively.

35

Figure 10B. LC/MS analysis of alcA promoter-replaced strains.
(B) The structure, UV/Vis and MS profile of the three compounds in the biosynthetic pathway of felinone
A.

36

Figure 11. Total scan HPLC profiles of natural products extracted from the culture medium of
feedback strains. Scale of intensity (y axis) is fixed for comparison of the yield of compounds. Peaks
labeled 1 to 3 are those compounds indicated in figure 10B. Similar to figure 10A, asterisk (*) and cross
(+) represent lumichrome and emericellamide A, respectively. Peaks labeled (1) to (4) are known
compounds not produced by AN7903 cluster, including terrequinone A, aspercrptin, F-9775 A or B, and
10-deoxygerfelin.

37

Figure 12. Proposed biosynthetic pathway for felinone A.

38

CHAPTER 4
DISCUSSION
With recent advances in genome mining of secondary metabolite biosynthetic gene
clusters, many silent clusters have been activated, and lots of new compounds have been
found. However, there are still some silent clusters remain unknown. For example, among
all the 14 NR-PKS in A. nidulans, the cluster AN7903 is one of the least studied one
(Ahuja et al., 2012). While lots of intermediates and final products have been found in
other 11 NR-PKS gene clusters, only one intermediate (compound 1) was identified.
Moreover, there are only a few studies mentioned about the cluster to date, and none of
them focus on the elucidation of its biosynthetic pathways. To begin with, the first study
that mentioned AN7903 was in 2009, while a group of researchers developed the co-
cultivation strategy to activate silent biosynthetic gene clusters(Schroeckh et al., 2009). By
co-cultivation of A.nidulans and bacteria S. hygroscopicus, the expression of two gene
clusters were induced. One of the clusters, characterized by its PKS AN7909, had been
found to produce several compounds including orsellinic acid, F-9775A, and F-9775B.
However, none of the compounds were discovered from the cluster AN7903.
In 2010, a multiprotein, COP9 signalosome complex (CSN) was found to play a
crucial role in secondary metabolism and inspired a study of the AN7903 gene cluster
(Nahlik et al., 2010). CSN regulates various proteins through the ubiquitin system. By
detaching an ubiquitin-like protein Nedd8 from E3 ligases, CSN can activate protein
degradation (Wu et al., 2005). Thus, deletion of CSN, especially the subunit csn5/csnE,
results in many changes in A. nidulans during development, including oxidative stress
protection, formation of asexual spores, and more than 100 secondary metabolites. Most of
the affected secondary metabolite genes were up-regulated, maybe because of the
39

deficiency of protein degradation. Among those genes, three genes in the AN7903 cluster,
AN7893, AN7896, and AN7899 were upregulated, suggested that the cluster was
controlled by csnE.
Gerke and his researchers’ group developed an alternative approach to activate
silent gene cluster by mutation of csn5/csnE in 2012, and successfully activated the gene
cluster AN7903 (Gerke et al., 2012). By genome-wide profiling analysis, AN7893 to
AN7903 were up-regulated. Furthermore, they overexpressed the two putative transcription
factor AN7896 and AN7901. Based on the experimental result, the border of the cluster is
from AN7896 to AN7903, controlled by AN7896 but not AN7901. However, the
compound isolated from overexpression AN7896 was compound 1, which can be produced
by overexpression PKS AN7903 only (Ahuja et al., 2012). Although some yellow pigments
were produced in overexpressing strain and seem like modified by AN7902, no other
compound from the cluster were identified. Therefore, the function of other genes in the
cluster, the biosynthetic pathway, and the final product are remained to be solved.
Our interest in the cluster AN7903 is not only that it is a not fully understand
cluster, but because it belongs to the secondary metabolite family, azaphilone. The
structure core of azaphilone, usually known as isochromene, enables azaphilones to react
with amino group-containing compounds (Figure 13). The reaction can be widely applied
to different fields of drug development, including anticancer, anti-inflammation,
antimicrobial and antiviral (Gao et al., 2013). Moreover, some of azaphilone compounds
are yellow or red, and can be used as natural food colorants (Mapari et al., 2010).
Therefore, it is possible to develop some drug candidates or food colorants from the
products generated from the AN7903 cluster. Also, the possibility for drug discovery from
compounds in the AN7903 cluster can be further elevated by synthesis chemistry. For
40

example, we proposed that the intermediate from the cluster, compound 1, can be
synthesized to antibiotic quinolone-like compounds (Figure 14). In short, the above reasons
urge us to study the AN7903 cluster, especially elucidate potential drug candidates’
biosynthetic pathway.

Figure 13. The scaffold of azaphilone family

Figure 14. Proposed synthetic reaction for compound 1.

In our study, we confirmed the production of compound 1 from the alcA-AN7903
strain we made. Moreover, we overexpressed the putative transcription factor AN7901 with
alcA promoter replacement. However, the overexpression results in no azaphilones
produced but a side product lumichrome, which is a derivative of the supplement riboB.
There are several possible reasons for the result. First, AN7901 may not be the
transcription factor that controls the cluster, as mentioned in Gerke and his research group
(Gerke et al., 2012). Second, instead of up-regulation of the cluster, AN7901 may be the
negative transcription factor. Based on the northern hybridization analysis published in
2012, 7 out of 12 genes in the cluster were expressed while AN7901 overexpression
41

repressed, and inhibited by induction of AN7901 overexpression. Specifically, two genes
AN7893 and AN7898 are strongly affected by AN7901 gene manipulation. Moreover, as
we implemented an in-depth BLAST analysis on AN7901, we found that AN7901 is
homolog to a protein Moc3 (Table 5). Moc3 is a negative containing Zn(II)2-Cys6
transcription factor involved in stress response, maintenance DNA integrity and sexual
development (Goldar et al., 2005). While the cluster AN7903 is regulated by csnE, which
play an important role on sexual development in A.nidulans, it is possible that AN7901 and
AN7903 are both transcription factor controlling the cluster, by down-regulation and up-
regulation, respectively. Transcriptional activator-repressor module has been identified in
other species such as A.niger, but not in A. nidulans (Niu et al., 2017). To test our
hypothesis, we plan to manipulate AN7901 and AN7896 by knock-out and activation in the
future.
On the other hand, we determined the product felinone A from the cluster
AN7903. Felinone A was first isolated from marine fungus Beauveria feline EN-135, as a
yellow solid with antimicrobial activity (Du et al., 2014). The absolute configuration of
felinone A was just revised by a group in Japan by total synthesis very recently (Abe et al.,
2017). Here, we discovered two genes responsible for felinone A production, AN7902 and
AN7903, and proposed a biosynthetic pathway. Felinone A might be the yellow pigment
that seen from Gerke’s group research on the cluster, as mentioned above, but not isolated
from A. nidulans yet. Thus, we also plan to scale-up the strain and isolate felinone A to
further confirm our finding.
As for optimizing the yield of felinone A, we collaborated with Dr. Oakley’s lab
and insert a construct for the positive feedback system. Interestingly, only three strain with
aldA promoter has significantly higher production of compounds produced by the cluster
42

AN7903. We suppose that aldA, which is not repressed by sugar, can produce compounds
earlier than alcA. Moreover, F-9775A, F9775B and other compounds were found in the
feedback strain. Those compounds are known to be produced by other clusters, such as
AN7909, a cluster nearby the cluster AN7903 (Sanchez et al., 2010). It is possible that the
feedback system is too strong, thus activate other clusters as the off-target effect. To
eliminate the effect, we can delete more clusters responsible for main secondary
metabolites produced in A. nidulans, just as our lab created the LO4389 strain with the
deletion of sterigmatocystin cluster.
In conclusion, AN7903 is viewed as a possible source for drug discovery. In this
study, we used promoter replacement strategy to made three strains, and identified two
genes that responsible for the production of felinone A and its intermediates from the
cluster. Moreover, the production was further increased by inserting feedback system. In
the future, we plan to do LC/MS analysis on gene deletion strain to determine the boundary
of felinone A biosynthesis cluster, and isolate compounds for further NMR analysis for
structure identification.
43

REFERENCES
Abe, H., Tango, H., Kobayashi, T., and Ito, H. (2017). Asymmetric total synthesis and
revision of absolute configurations of azaphilone derivative felinone A. Tetrahedron Lett 58,
4296-4298.

Ahuja, M., Chiang, Y.M., Chang, S.L., Praseuth, M.B., Entwistle, R., Sanchez, J.F., Lo,
H.C., Yeh, H.H., Oakley, B.R., and Wang, C.C. (2012). Illuminating the diversity of aromatic
polyketide synthases in Aspergillus nidulans. J Am Chem Soc 134, 8212-8221.

Andersen, M.R., Nielsen, J.B., Klitgaard, A., Petersen, L.M., Zachariasen, M., Hansen,
T.J., Blicher, L.H., Gotfredsen, C.H., Larsen, T.O., Nielsen, K.F., et al. (2013). Accurate
prediction of secondary metabolite gene clusters in filamentous fungi. Proc Natl Acad Sci U S A
110, E99-107.

Austin, M.B., and Noel, J.P. (2003). The chalcone synthase superfamily of type III
polyketide synthases. Nat Prod Rep 20, 79-110.

Bok, J.W., Chiang, Y.M., Szewczyk, E., Reyes-Dominguez, Y., Davidson, A.D.,
Sanchez, J.F., Lo, H.C., Watanabe, K., Strauss, J., Oakley, B.R., et al. (2009). Chromatin-level
regulation of biosynthetic gene clusters. Nat Chem Biol 5, 462-464.

Chiang, Y.M., Oakley, B.R., Keller, N.P., and Wang, C.C. (2010). Unraveling polyketide
synthesis in members of the genus Aspergillus. Appl Microbiol Biotechnol 86, 1719-1736.

Chiang, Y.M., Szewczyk, E., Davidson, A.D., Keller, N., Oakley, B.R., and Wang, C.C.
(2009). A gene cluster containing two fungal polyketide synthases encodes the biosynthetic
pathway for a polyketide, asperfuranone, in Aspergillus nidulans. J Am Chem Soc 131, 2965-
2970.

Chiang, Y.M., Szewczyk, E., Nayak, T., Davidson, A.D., Sanchez, J.F., Lo, H.C., Ho,
W.Y., Simityan, H., Kuo, E., Praseuth, A., et al. (2008). Molecular genetic mining of the
Aspergillus secondary metabolome: discovery of the emericellamide biosynthetic pathway. Chem
Biol 15, 527-532.

Chooi, Y.H., and Tang, Y. (2012). Navigating the fungal polyketide chemical space:
from genes to molecules. J Org Chem 77, 9933-9953.

Chun, J., and Hartung, H.P. (2010). Mechanism of action of oral fingolimod (FTY720) in
multiple sclerosis. Clin Neuropharmacol 33, 91-101.

Du, F.Y., Li, X.M., Zhang, P., Li, C.S., and Wang, B.G. (2014). Cyclodepsipeptides and
other O-containing heterocyclic metabolites from Beauveria felina EN-135, a marine-derived
entomopathogenic fungus. Mar Drugs 12, 2816-2826.

Galagan, J.E., Calvo, S.E., Cuomo, C., Ma, L.J., Wortman, J.R., Batzoglou, S., Lee, S.I.,
Basturkmen, M., Spevak, C.C., Clutterbuck, J., et al. (2005). Sequencing of Aspergillus nidulans
and comparative analysis with A. fumigatus and A. oryzae. Nature 438, 1105-1115.

44

Gao, J.M., Yang, S.X., and Qin, J.C. (2013). Azaphilones: chemistry and biology. Chem
Rev 113, 4755-4811.

Gerke, J., Bayram, O., Feussner, K., Landesfeind, M., Shelest, E., Feussner, I., and Braus,
G.H. (2012). Breaking the silence: protein stabilization uncovers silenced biosynthetic gene
clusters in the fungus Aspergillus nidulans. Appl Environ Microbiol 78, 8234-8244.

Goldar, M.M., Jeong, H.T., Tanaka, K., Matsuda, H., and Kawamukai, M. (2005). Moc3,
a novel Zn finger type protein involved in sexual development, ascus formation, and stress
response of Schizosaccharomyces pombe. Curr Genet 48, 345-355.

Hansen, M.E., Smedsgaard, J., and Larsen, T.O. (2005). X-Hitting: an algorithm for
novelty detection and dereplication by UV spectra of complex mixtures of natural products. Anal
Chem 77, 6805-6817.

Harvey, A.L., Edrada-Ebel, R., and Quinn, R.J. (2015). The re-emergence of natural
products for drug discovery in the genomics era. Nat Rev Drug Discov 14, 111-129.

Hashimoto, M., Nonaka, T., and Fujii, I. (2014). Fungal type III polyketide synthases.
Nat Prod Rep 31, 1306-1317.

Hertweck, C. (2009). The biosynthetic logic of polyketide diversity. Angew Chem Int Ed
Engl 48, 4688-4716.

Inglis, D.O., Binkley, J., Skrzypek, M.S., Arnaud, M.B., Cerqueira, G.C., Shah, P.,
Wymore, F., Wortman, J.R., and Sherlock, G. (2013). Comprehensive annotation of secondary
metabolite biosynthetic genes and gene clusters of Aspergillus nidulans, A. fumigatus, A. niger
and A. oryzae. BMC Microbiol 13, 91.

Keller, N.P., Turner, G., and Bennett, J.W. (2005). Fungal secondary metabolism - from
biochemistry to genomics. Nat Rev Microbiol 3, 937-947.

Kitamoto, N., Matsui, J., Kawai, Y., Kato, A., Yoshino, S., Ohmiya, K., and Tsukagoshi,
N. (1998). Utilization of the TEF1-alpha gene (TEF1) promoter for expression of
polygalacturonase genes, pgaA and pgaB, in Aspergillus oryzae. Appl Microbiol Biotechnol 50,
85-92.

Li, J.W., and Vederas, J.C. (2009). Drug discovery and natural products: end of an era or
an endless frontier? Science 325, 161-165.

Machida, M., Asai, K., Sano, M., Tanaka, T., Kumagai, T., Terai, G., Kusumoto, K.,
Arima, T., Akita, O., Kashiwagi, Y., et al. (2005). Genome sequencing and analysis of
Aspergillus oryzae. Nature 438, 1157-1161.

Mapari, S.A., Thrane, U., and Meyer, A.S. (2010). Fungal polyketide azaphilone
pigments as future natural food colorants? Trends Biotechnol 28, 300-307.

Mishra, B.B., and Tiwari, V.K. (2011). Natural products: an evolving role in future drug
discovery. Eur J Med Chem 46, 4769-4807.
45

Nahlik, K., Dumkow, M., Bayram, O., Helmstaedt, K., Busch, S., Valerius, O., Gerke, J.,
Hoppert, M., Schwier, E., Opitz, L., et al. (2010). The COP9 signalosome mediates
transcriptional and metabolic response to hormones, oxidative stress protection and cell wall
rearrangement during fungal development. Mol Microbiol 78, 964-979.

Newman, D.J., and Cragg, G.M. (2016). Natural Products as Sources of New Drugs from
1981 to 2014. J Nat Prod 79, 629-661.

Niu, J., Alazi, E., Reid, I.D., Arentshorst, M., Punt, P.J., Visser, J., Tsang, A., and Ram,
A.F. (2017). An Evolutionarily Conserved Transcriptional Activator-Repressor Module Controls
Expression of Genes for D-Galacturonic Acid Utilization in Aspergillus niger. Genetics 205, 169-
183.

Oakley, C.E., Ahuja, M., Sun, W.W., Entwistle, R., Akashi, T., Yaegashi, J., Guo, C.J.,
Cerqueira, G.C., Russo Wortman, J., Wang, C.C., et al. (2017). Discovery of McrA, a master
regulator of Aspergillus secondary metabolism. Mol Microbiol 103, 347-365.

Pontecorvo, G., Roper, J.A., Hemmons, L.M., Macdonald, K.D., and Bufton, A.W.
(1953). The genetics of Aspergillus nidulans. Adv Genet 5, 141-238.

Sanchez, J.F., Chiang, Y.M., Szewczyk, E., Davidson, A.D., Ahuja, M., Elizabeth
Oakley, C., Woo Bok, J., Keller, N., Oakley, B.R., and Wang, C.C. (2010). Molecular genetic
analysis of the orsellinic acid/F9775 gene cluster of Aspergillus nidulans. Mol Biosyst 6, 587-
593.

Schmitt, I., Martin, M.P., Kautz, S., and Lumbsch, H.T. (2005). Diversity of non-
reducing polyketide synthase genes in the Pertusariales (lichenized Ascomycota): a phylogenetic
perspective. Phytochemistry 66, 1241-1253.

Schroeckh, V., Scherlach, K., Nutzmann, H.W., Shelest, E., Schmidt-Heck, W.,
Schuemann, J., Martin, K., Hertweck, C., and Brakhage, A.A. (2009). Intimate bacterial-fungal
interaction triggers biosynthesis of archetypal polyketides in Aspergillus nidulans. Proc Natl
Acad Sci U S A 106, 14558-14563.

Szewczyk, E., Nayak, T., Oakley, C.E., Edgerton, H., Xiong, Y., Taheri-Talesh, N.,
Osmani, S.A., and Oakley, B.R. (2006). Fusion PCR and gene targeting in Aspergillus nidulans.
Nat Protoc 1, 3111-3120.

Vining, L.C. (1992). Secondary metabolism, inventive evolution and biochemical
diversity--a review. Gene 115, 135-140.

Waridel, P., Wolfender, J.L., Lachavanne, J.B., and Hostettmann, K. (2004).
Identification of the polar constituents of Potamogeton species by HPLC-UV with post-column
derivatization, HPLC-MSn and HPLC-NMR, and isolation of a new ent-labdane diglycoside.
Phytochemistry 65, 2401-2410.

Wu, J.T., Lin, H.C., Hu, Y.C., and Chien, C.T. (2005). Neddylation and deneddylation
regulate Cul1 and Cul3 protein accumulation. Nat Cell Biol 7, 1014-1020.
46

Yaegashi, J., Oakley, B.R., and Wang, C.C. (2014). Recent advances in genome mining
of secondary metabolite biosynthetic gene clusters and the development of heterologous
expression systems in Aspergillus nidulans. J Ind Microbiol Biotechnol 41, 433-442.

Yang, L., Ukil, L., Osmani, A., Nahm, F., Davies, J., De Souza, C.P., Dou, X., Perez-
Balaguer, A., and Osmani, S.A. (2004). Rapid production of gene replacement constructs and
generation of a green fluorescent protein-tagged centromeric marker in Aspergillus nidulans.
Eukaryot Cell 3, 1359-1362.

Abstract (if available)

Abstract Fungal secondary metabolites are an important source for drug discovery. Whole genome sequencing has shown that there are thousands of secondary metabolite gene clusters with unknown products. In this study, we focus on the biosynthetic pathway of felinone A. Previous study, in the Wang lab the promoter of the nonreducing polyketide synthase (NR-PKS) AN7903 was replaced and 2,4-dihydroxy-3-methyl-6-(2-oxopropyl)-benzaldehyde (compound 1) was isolated. However, since we did not activate the surrounding genes that may modify compound 1, the final product of this gene cluster remains unknown. BLAST search and gene co-expression analysis show that AN7901 is a transcription factor and AN7902 is a monooxygenase. We replaced the promoter of AN7901, AN7902, and AN7903 and identified felinone A and its intermediate. Moreover, we inserted a feedback system to increase the yield of compounds that we are interested.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Genome mining of natural product biosynthesis pathways in filamentous fungi for novel drug discovery and production

PDF

Genetic engineering of fungi to enhance the production and elucidate the biosynthesis of bioactive secondary metabolites

PDF

Secondary metabolites of Aspergillus nidulans

PDF

Application of genome-wide strategies for the mining of secondary metabolite biosynthesis pathways in filamentous fungi

PDF

Genetic and chemical characterization of two highly-reducing polyketide synthase clusters from Aspergillus species

PDF

Genome manipulation of filamentous fungi for upregulating the production and illustrating the biosynthesis of valuable secondary metabolites using CRISPR-Cas9

PDF

Examinations of secondary metabolites and their production in fungi and bacteria

PDF

Secondary metabolites biosynthesis in Aspergillus species revealed by fungal genome mining

PDF

Molecular genetic mining of the secondary metabolome in Aspergillus nidulans

PDF

Fungal polyketides -- Review of recent findings

PDF

Comparison of the pharmacokinetic profiles between monomeric and oligomeric transferrin

PDF

Optimization of nanomedicine based drug delivery systems for the treatment of solid tumors

PDF

Development of new approaches for antibody modification

PDF

Genome mining of secondary metabolites in Aspergillus terreus isolated from the Chernobyl Collection

PDF

Genome mining of secondary metabolites in Scedosporium apiospermum and Paecilomyces variotii using CRISPR-Cas9 technology

PDF

Genetic engineering of non-reducing polyketide synthases by domain swap strategy in Aspergillus nidulans

PDF

Comparison of cellular uptake of arginine-rich cell-penetrating peptides

PDF

Genotoxic impurity analysis in pharmaceuticals

PDF

Harnessing environmental and culture conditions to alter fungal ‘omics’

PDF

Computer modeling of protein-peptide interface solvation

Asset Metadata

Creator Liao, Yi-En (author)

Core Title Mining the felinone A biosynthetic pathway

School School of Pharmacy

Degree Master of Science

Degree Program Pharmaceutical Sciences

Publication Date 04/30/2020

Defense Date 04/30/2018

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

Tag Aspergillus nidulans,biosynthetic pathway,drug discovery,felinone A,Fungi,natural product,OAI-PMH Harvest

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Wang, Clay (committee chair), Okamoto, Curtis (committee member), Shen, Wei-Chiang (committee member)

Creator Email mandy40433@gmail.com,yienliao@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-497755

Unique identifier UC11267149

Identifier etd-LiaoYiEn-6289.pdf (filename),usctheses-c40-497755 (legacy record id)

Legacy Identifier etd-LiaoYiEn-6289.pdf

Dmrecord 497755

Document Type Thesis

Format application/pdf (imt)

Rights Liao, Yi-En

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