Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Assembling NRPS modules in e. coli to establish a platform for rational design of biologically active compounds
(USC Thesis Other)
Assembling NRPS modules in e. coli to establish a platform for rational design of biologically active compounds
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
ASSEMBLING NRPS MODULES IN E. COLI TO ESTABLISH A PLATFORM
FOR
RATIONAL DESIGN OF BIOLOGICALLY ACTIVE COMPOUNDS
by
Alex Praseuth
____________________________________________________________________
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PHARMACOLOGY AND PHARMACEUTICAL SCIENCES)
August 2008
Copyright 2008 Alex Praseuth
ii
DEDICATION
This work is dedicated to my wife and son, Francine and Mitchell Praseuth, for their
support during my education and allowing me to grow with patience and devotion. Also,
I would like to dedicate this work to my parents, Brian and Diana Praseuth, for their
encouragement to pursue my doctorate and unconditional support throughout my
scientific endeavors. I love you all.
iii
ACKNOWLEDGEMENTS
I’d like to thank my mentor and advisor, Dr. Clay CC Wang, for his invaluable
guidance and support. I joined Dr. Wang’s laboratory at the University of Southern
California about five years ago and have devoted considerable effort toward refining my
scientific and technical knowledge. Dr. Wang continually provided needed
encouragement during the pursuit of my doctorate and was always at hand to support me
during the demanding times. I am also indebted to Dr. Kenji Watanabe for his guidance
and mentorship. His training during my time in Dr. Wang’s laboratory has been
invaluable. He has honed my scientific expertise and strengthened my appreciation of
this boundless discipline.
I would also like to thank my committee members, Dr. Ian Haworth, Dr. Nouri
Neamati, and Dr. Richard Roberts, for their support during my studies. Their assistance
in data interpretation and experimental suggestions opened new doorways for my project
and manuscripts.
I would like to thank my parents, Brian and Diana Praseuth, my brothers Michael,
David and Richard for their love and support during my Ph.D. work and for their
categorical support leading up to it.
iv
TABLE OF CONTENTS
DEDICATION ii
ACKNOWLEDGEMENTS iii
LIST OF TABLES vi
LIST OF FIGURES vii
LIST OF SCHEMES x
ABSTRACT xi
CHAPTER I: INTRODUCTION
1. Importance of natural products and the role of
Escheria coli as a heterologous host 1
2. Deciphering Of Biosynthetic Code Responsible
for Producing Polyketide and Nonribosomal
Peptide Secondary Metabolite 3
3. Echinomycin, the DNA bis-intercalating
Nonribosomal Peptide antibiotic 9
4. Structural studies of the quinoxaline antibiotic 10
5. Proposed Biosynthetic Mechanism for Echinomycin 11
6. Engineering Natural Product Biosynthetic Machinery
and Its Industrial Applications 16
CHAPTER II: DE NOVO PRODUCTION QUINOXALINE
ANTIBIOTICS IN E. COLI
1. Introduction 23
2. Results 25
3. Discussion 41
CHAPTER III: ENHANCED PRODUCTION OF QUINOXALINE ANTIBIOTICS
IN E. COLI
1. Introduction 51
2. Results 54
3. Discussion 69
v
CHAPTER IV: PRODUCTION OF UNNATURAL NATURAL PRODUCT
IN A HETEROLOGOUS HOST
1. Introduction 73
2. Results 75
3. Discussion 90
CHAPTER 5: MATERIALS AND METHODS 97
BIBLIOGRAPHY 110
vi
LIST OF TABLES
Table 2-1 Fermentation conditions for cultivation of engineered
E. coli 33
Table 3-1 Observed maximal titer of 2 as a function of rate
Of nutritional availability 54
Table 3-2 Conditions for maximal production of 2 59
Table 4-1 Fed-batch fermentation conditions for production of
8 using engineered E. coli 80
Table 4-2 Shake flask fermentation condition for production of
8 using engineered E. coli 83
vii
LIST OF FIGURES
Figure 1-1 Chemical structures of natural products in clinical use 4
Figure 1-2 Modular organization of type I PKS and NRPS 6
Figure 1-3 Chemical structures of quinomycin antibiotics 8
Figure 1-4 An echinomycin (GCGTACGS)
2
complex representation 9
Figure 1-5 Complete echinomycin biosynthetic genes 13
Figure 1-6 Chemical structures of marine natural products possessing
Potent antiobitic activity 19
Figure 2-1 Echinomycin biosynthetic cluster isolated from
Streptomyces lasaliensis 27
Figure 2-2 LC-MS analyses of an Ecm18-catalyzed thioacetal bridge
formation 28
Figure 2-3 SDS PAGE analysis of purified echinomycin biosynthetic
proteins 29
Figure 2-4 Mass spectral analysis of broth harvested from a fermentor
And extracted with ethylacetate 31
Figure 2-5 Spectral and chromatographic analysis of authentic
1 and 2 32
Figure 2-6 NMR spectrum of crude ehtylacetate extraction harvested
From a 2L culture 35
Figure 2-7 Mass spectrum analysis of fractionated mycelium 36
Figure 2-8 NMR spectrum of Echinomycin isolated from E. coli
Post 100% 2-butanone PTLC development 38
Figure 2-9 NMR spectrum of 1 isolated from E. coli post 7%
Methanol/chloroform PTLC purificaiton 39
Figure 2-10
1
H NMR spectra of authentic quinomycin antibiotics 40
Figure 2-11 LC-MS spectrograph of 1 isolated from E. coli 45
viii
Figure 2-12
1
H NMR TOCSY spectrum of 1 collected at mixing
Time of 100 ms showing important correlations
(numbered) 46
Figure 2-13 Chemical characteristic spectra of compound 2 produced
By our engineered E. coli strain 47
Figure 2-14
1
H NMR spectrum of 2 48
Figure 3-1 Chemical structures of quinomycin antibiotics 53
Figure 3-2 Production of 2 according to feed rate of nutrition
Following protein expression using a fed-batch
Fermentor over time 55
Figure 3-3 Correlation of OD
600
and pH according to feed rate
Using a fed-batch fermentor over time 58
Figure 3-4 Chromatograms obtained from a UV detector and
mass spectrometer 61
Figure 3-5 Comparison of three 50 mL culture differing in medium
And method in which bicyclic starter unit is furnished 62
Figure 3-6 Correlation of OD
600
and pH of small culture over time 63
Figure 3-7 Production of 2 at varying concentration of DMSO
using shake flask 64
Figure 3-8 Correlation of OD600 and pH according to DMSO in
Shake flask over time 66
Figure 4-1 Chemical structure of quinomycin antibiotics 74
Figure 4-2 Deactivation of M-domains in Ecm7 77
Figure 4-3 Effect of colony size on production of 8 79
Figure 4-4 Chromatographic analysis of 8 obtained from culture 87
Figure 4-5 Chromatographic analysis of 9 obtained from culture 89
Figure 4-6 Proposed allocation of amino acid in culture 93
Figure 4-7
1
H NMR (400 MHz, CDCl
3
) spectrum of 8 95
ix
Figure 4-8
1
H NMR TOCSY (400 MHz, CDCl
3
) spectrum of 8. 96
Figure 4-9
1
H NMR (400 MHz, CDCl
3
) spectrum of 9 96
x
LIST OF SCHEMES
Scheme 1-1 Proposed biosynthetic pathway and modular organization
of echinomycin biosynthesis 12
Scheme 1-2 Proposed mechanism for the echinomycin biosynthesis 15
Scheme 2-1 Quinoxaline antibiotics and its proposed mechanism for
echinomycin biosythesis 24
Scheme 2-2 Proposed mechanism for the Ecm18-catalyzed thioacetal
bridge formation 43
Scheme 3-1 Proposed biosynthetic pathway and modular organization
Of echinomycin biosynthesis 68
xi
ABSTRACT
Nonribosomal peptides (NRPs) are synthesized by modular mega–enzyme called
NRP synthetase (NRPS) that catalyze a peptide bond forming reaction using natural
amino acid as substrate. The majority of compounds from this class exhibit crucial
biological activities such as antibiotic, immunosuppressive, antiviral, and antitumor
activities. However, several of these sought-after natural products are often difficult to
isolate in adequate amounts from its natural sources due to low production levels or the
producing organism’s stringent need for expensive and atypical culture apparatus. As we
know, enzymes capable of synthesizing these natural products including essential or
nonessential secondary metabolites are encoded by biosynthetic genes located on either
chromosomal or plasmid DNA.
An alternative and straightforward approach to obtain these compounds is to
express the entire biosynthetic gene cluster for an NRPS responsible for producing the
intact molecule using E. coli as a heterologous host was developed in our laboratory.
Production of triostin A (2) using small shake flask was optimized to provide a
quantitative yield rivaling levels isolated from S. triostinicus. Addition of QXC, the
priming unit for 2, helped to increase its production and identify the bottleneck in our
system. Knock-down of two M-domain in Ecm7 provided modest amounts of 8, an
unnatural natural product. A novel compound, 9, was also identified from our culture
which will require additional characterization.
xii
Numerous parameters were investigated and optimal conditions were identified.
Ineffective excipients were ruled out and excluded from our fermentation process.
This study has ascertained a novel platform for creating a peptide library by
means of premeditated engineering of biosynthetic domain(s) to provide a customized
macromolecule.
1
CHAPTER I: INTRODUCTION
1. Importance of Natural Products and the Role of Escheria coli as a Heterologous
Host
Natural products remain one of the most important sources of chemotherapeutic
agents currently in clinical use (Newman et al. 2003). They are secondary metabolites
from various organisms that play non-essential roles in the life-sustaining metabolism of
the organisms (Vining 1992). These organisms have developed sophisticated
biosynthetic systems for the production of natural products, partly because these
metabolites provide the producing organism’s survival advantage over their non-
producing counterparts. Many such natural products possess potent biological activities,
such as antimicrobial, antiviral and antitumor properties that make them particularly
valuable as pharmaceutical agents. There are several approaches used at the present to
obtain natural products and their analogs (Newman et al. 2003). Among these are total
chemical synthesis, fermentation of natural product producing organisms, semi-synthesis,
and use of heterologous host systems. Of the four methods, the use of heterologous host
systems has gained ground over the past decade as an effective alternative to preparing
natural products and their analogs (Pfeifer and Khosla 2001). In this approach, the entire
biosynthetic pathway for the production of a target natural product is transferred from the
natural source organism into a well-characterized, readily culturable model organism
such as E. coli.
2
The greatest advantage of using E. coli as a heterologous host is the availability of a
wealth of well-established molecular biological techniques for its genetic and metabolic
manipulation. In addition, ease of fermentation of E. coli makes this organism
particularly suitable for overproduction of natural products. Its robust tolerance toward
exogenous proteins and fast life cycle also promote our engineering efforts. Large-scale
protein production ability will facilitate detailed biochemical characterization of the
biosynthetic pathway, allowing a rational approach to optimizing the pathway and
modifying the production of exogenous natural products for biosynthesis of its analog.
Knowledge of the complete genome (Blattner et al. 1997) and extensive understanding of
many aspects of the metabolic pathways of this organism also greatly reduce the
uncertainty in the process. However, because the biosynthetic pathways targeted are
usually metabolically foreign to E. coli, certain genetic implementations to the bacterium
are necessary to make the heterologous host an efficient secondary metabolite-producing
organism. Also, care must be taken to introduce self-resistance mechanisms into E. coli
if the natural product being produced is likely to be toxic. This toxicity can be avoided
by identifying the mechanisms of self resistance in the original producing host and
supplementing the same or a similar system in E. coli. None-the-less, recent study has
successfully extended the yield of a complex polyketide natural product, 6-
deoxyerythronolide, erythromycin aglicon, to over one gram per liter of E. coli
fermentation (Lau et al. 2004). On the other hand, although there has been a recent report
on partial reconstitution of NRPs activity in E. coli (Gruenewald et al. 2004),
development of an E. coli-based total biosynthesis of nonribosomally synthesized peptide
natural product is yet to be explored in depth.
3
2. Deciphering of Biosynthetic Code Responsible for Producing Polyketide and
Nonribosomal Peptide Secondary Metabolite in E. coli.
Since the 1990s, many gene clusters have been isolated and identified from bacteria
that are responsible for production of secondary metabolites such as polyketides (PKs)
(Cortes et al. 1990; Donadio et al. 1991), nonribosomal peptides (NRPs) (Krause and
Marahiel 1988), terpenes (Armstrong et al. 1989), and aminoglycosides (Ota et al. 2000;
Tamegai et al. 2002). Flanking researchers in this pursuit, numerous yet powerful
molecular biological techniques are also being developed. These findings were a feat of
genetic contribution for next generation’s drug discovery. Recent papers have presented
characteristic mechanisms for biosynthetic proteins coded in gene clusters and elaborate
production of natural products by expression of biosynthetic genes in a heterologous host.
It is highly anticipated that these production systems can offer a more efficient method of
obtaining valuable compounds and displace our dependence on natural products from
their natural hosts.
4
Figure 1-1. Chemical Structures of Natural Products in Clinical Use.
Erythromycin, rifamycin and vancomycin are some examples of clinically
important natural products that are isolated from Streptomyces as secondary metabolites
and have been found to possess noteworthy biological activity (Figure 1-1). Their
notoriety as potent antibiotics has encouraged chemists to chemically synthesize them
along with some of their derivatives for further studies in hopes of improving their
pharmacological properties. In recent years, many more biosynthetic genes have been
isolated and identified from additional strains that are responsible for producing valuable
molecules adding to the list of compounds above. Commonly, these complex natural
products are synthesized by many biosynthetic proteins that are encoded in many genes
clustered together on a chromosomal or linear-plasmid DNA, called “gene cluster”.
N
H
3
C O
OH O
OC H
3
OH
O
O
H 3 C O O
O
O
OH
FK 5 0 6
O
O
HO
O
H
O
H
OA c
O
AcO O OH
H O O
H
O
O
OH
NH O
taxol
O HO
O O
H
N
O
N
H
O
N H
2
O
H
N
O
N H
O
H
N
OH
HO
O
NH
O
N
H
C l
HO
O
O
O
O
OH
HO
OH
NH
2
HO
O
OH
OH
Cl
HO
H
H
H
vancomycin
O
OCH
2
CO
2
H
H
3
C O
OH
OH
AcO
OH
O
NH
OH
O
O
O
O
O
OH
OH
O
O
H O
O
O
H O
N (C H
3
)
2
OH
OC H
3
e ry t h ro my c in
A rifamycin
B
lovastatin
5
Based on sequence analysis, it was revealed that erythromycin and rifamycin are both
synthesized by modular polyketide synthase (PKS) and vancomycin is synthesized by
nonribosomal peptide synthetase (NRPS) (Cane et al. 1998; Walsh 2004). The
mechanism of chain elongation, carbon-carbon bond formation, for backbone
biosynthesis of a compound by modular PKS is similar in fashion to fatty acid synthase.
In general, these gene clusters code for PKS responsible for producing polyketides or
large modularly-organized enzymes (Figure 1-2a). A single module of the enzyme
ranges in size from 120 to 180 kDa and possess the ability to independently extend a
carbon chain using a single building block. Similarly, NRPS is large and genetically
encoded in a similar manner as PKS and is also comprised of multi-domain complexes
(Figure 1-2b). It catalyzes a peptide bond formation reaction to generate the NRP
scaffold. The compulsory biosynthetic pathway components were uncovered and are
known to initiate production of these compounds by manufacturing the starter unit and
building blocks which include propionyl-, methyl- and ethylmalonyl-CoA necessary to a
variety of PKs (Hutchinson 2003). Each PKS or NRPS can then be efficiently converted
from inactive apo- to phosphopantetheinyl-containing holo-acyl carrier protein (ACP)/
thiolation (T) forms for elongation of PK or NRP assembly. Moreover, each module is
able to recognize an appropriate building block, execute a chain elongation reaction, and
control the stereochemistry to biosynthesize the compound’s core structure following a
preprogrammed biosynthetic array.
6
KS AT ACP
SH OH P-pant
SH
KS AT ACP
SH OP-pant
SH
KS AT ACP
S OH P-pant
S
O
KS AT ACP
SH OH P-pant
S
O
KS AT ACP
SH OH P-pant
S
X
O
O downstream
module
SH
KS AT ACP
SH OH P-pant
SH
downstream
module
S
HO
2
C
SCoA
O
R
HSCoA
HO
2
C
O
R
R
HO
2
C
upstream
module
S X
O
R
X
O
X
O O
CO
2
Malonyl-CoA : R = H
Methylmalonyl-CoA : R = CH
3
Ethylmalonyl-CoA : R = C
2
H
5
R R
O
-
O
a
C A T
ATP
P-pant
SH
C A T
P-pant
SH
H
2
N
R
PPi
Amino acid
OH
O
C A T
upstream
module
SH
C A T
H
2
N
R
AMP
O
P-pant
S
O
R
HN
X
O
R'
P-pant
S
upstream
module
S
O
NH
2
O
R
X
R'
H
2
N
R
AMP
O
AMP
b
Figure 1-2. Modular organization of (a) type I PKS and (b) NRPS. Type I PKS and
NRPS modules contain the catalytic domains responsible for acyl and peptide chain
elongation, respectively. Hypothetical type I PKS and NRPS modules are shown with
core domains. KS, ketosynthase; AT, acyltransferase; ACP, acyl carrier protein; C,
condensation; A, adenylation; T, thiolation.
7
Escalating the notion of module swapping between two biosynthetic clusters,
changing the order of the biosynthetic modules, or deleting a domain within the module
will provide a rationally tailored molecule. Possessing this sort of method will allow us
to design and produce them in the same manner as we would via chemical synthesis. In
fact, rational engineering of modular PKSs has successfully provided a divergence of
polyketide scaffolds (Watanabe et al. 2007). Therefore, it is possible to generate more
diverse and novel compounds with desirable pharmacological profiles by means of the
mentioned approach. A method to manipulate genes and assemble biosynthetic clusters
for expression in a heterologous host such as E. coli has been reported (Pfeifer et al.
2001; Watanabe et al. 2006). Using E. coli as a heterologous host will allow researchers
to utilize all of the advantages listed in the previous section. However, producing
biologically active form of complex NRPS and PKS natural products using E. coli by
means of transforming the entire biosynthetic machinery for de novo production possess
inherent challenges. Because such complex natural products are synthesized by
numerous enzymes and the size of each PKS and NRPS module is large, ranging from 4
to 6 kb length, it is very difficult to incorporate essential biosynthetic genes into a
plasmid vector for expression and heterologous production of natural products with an E.
coli-based system.
8
HN
N
N
O
N
N
NH
O
N
H
H
N
N
N
N
N
O
O
O
O
O
O
O
O
S
SMe
O
O
H
Echinomycin (1)
HN
N
N
O
N
N
NH
O
N
H
H
N
N
N
N
N
O
O
O
O
O
O
O
O O
O
Triostin A (2)
S
S
HN
N
N
O
N
N
NH
O
N
H
H
N
N
N
O
O
O
O
O
O
O
O
S
SMe
O
O
H
SW-163D
OH
HO
Figure 1-3. Chemical Structures of Quinomycin Antibiotics.
This dissertation will report on our effort in overcoming these hurdles and
successfully achieving heterologous production of NRP antibiotics, echinomycin and
triostin A, in E. coli (Figure 1-3). Efficacious transplantation of NRP echinomycin gene
cluster from its natural host into E. coli will now allow for development of an E. coli-
based host system for unmitigated engineered biosynthesis of natural product and
unnatural natural product (Watanabe et al. 2006).
9
3. Echinomycin, the DNA Bis-intercalating Quinomycin Antibiotic
Echinomycin, a representative NRP that has been isolated from an assortment of
bacteria including Streptomyces lasaliensis and belongs to the large family of quinomycin
antibiotics (Steinerova et al. 1987). As its name implies, this class of natural products
possess two bicyclic aromatic quinoxaline or quinoline chromophores attached to the
dimerized cyclic peptide core structure. Interest in this group of natural products stems
from its potent antibacterial, anticancer and antiviral activities. Nearly all of the
compounds in this class are potent antibiotics against gram-positive bacteria by means of
DNA-directed RNA synthesis inhibition (Waring and Makoff 1974; Takusagawa 1985;
Romero et al. 1997).
Figure 1-4. (a) An echinomycin (GCGTACGC)
2
complex representation (Cuesta-Seijo
and Sheldrick 2005). (b) Diagram representation of chromophore-base stacking. (c)
Rotated 180˚ from (a) showing the chromophores interacting with DNA.
10
Echinomycin has also been shown to exhibit a strong cytotoxic effect against a variety of
cultured tumor cells. Furthermore, a number of these compounds were found to exhibit
potent cytotoxic and antiviral activity at nanomolar concentration in various assays
(Boger et al. 2001). Every one of its observed biological activities, including the
differences in the biological targets, appears to originate from their powerful and
sequence-selective DNA binding capability. As judged by a detailed X-ray and NMR
structural and other biophysical studies of these molecules, the heteroaromatic planar
chromophores, the invariant feature of the quinomycin antibiotics, furnish these
compounds with the ability to tightly bis-intercalate into DNA (Figure 1-4) (Ughetto et al.
1985; Searle et al. 1989; Chen 1995).
4. Structural Studies of the Quinoxaline Antibiotic
Comparative study between echinomycin and SW-163D (Figure 1-3) (Rance et al.
1989; Kurosawa et al. 2001; Takahashi et al. 2001; Nakaya et al. 2007) revealed that the
sequence selectivity of these compounds are influenced by its chromophore (Fox 1990).
However, the cyclic peptide backbone on these molecules, while providing a framework
that holds the chromophores in place for effective DNA intercalation, also plays an
important role in determining the binding sequence specificity for these antibiotics
primarily by forming hydrogen bonds with the base moieties in the minor groove
(Ughetto et al. 1985; Searle et al. 1989; Chen 1995). Furthermore, modification of the
cross-bridge also provides further fine-tuning of the binding site selectivity of these
quinomycin antibiotics. This was confirmed by a DNA-binding study between
11
echinomyicn and its close relative, triostin A. The only structural difference between
them is the cross-bridge: echinomycin boasts a thioacetal bridge, whereas a disulfide
bond interlinks the triostin A backbone. However, the subtle structural differences
between the two compounds result in notable differences in their DNA binding affinity
(Lee and Waring 1978).
A group under the guidance of Waring in the UK at Cambridge University has provided
an abundance of DNA binding studies using analogs of echinomycin and triostin A which
are not found in nature but were synthesized in their laboratory (Lee and Waring 1978).
This has given natural products scientist a clearer picture of the binding characteristics of
these quinoxaline antibiotics to DNA. Of the numerous quinomycin antibiotic analogs
prepared by R. K. Olsen of Utah State University, TANDEM (tetra-N-des methyl triostin
A) and [N-MeCys
3
, N-MeCys
7
] TANDEM were found to favor AT-rich sequence in
comparison to their parent compounds (Viswamitra et al. 1981; Evans et al. 1982; Addess
et al. 1992). These findings have established how a modification as minute as the
removal of two methyl groups from the peptidic backbone can drastically change the
binding preference of these antibiotics.
5. Proposed Biosynthetic Mechanism for Echinomycin
To identify and isolate the echinomycin biosynthetic gene cluster, a cosmid library
was constructed using DNA that was obtained from S. lasaliensis. A sequence of the
echinomycin biosynthetic pathway involved in quinoxaline-2-carboxylic acid (QXC) and
12
peptide backbone biosyntheses is shown in Figure 1-5 and Scheme 1-2. Based on past
findings, L-tryptophan was identified as the precursor to QXC (Reid et al. 1984).
Predicted functions of proteins encoded within the gene cluster allowed us to hypothesize
and construct the QXC biosynthetic pathway illustrated in Scheme 1-1. Moreover, feed
studies were performed using (2S, 3S)- and (2S, 3R)-diastereomers which were
synthesized as the proposed substrate for Ecm11 of the QXC biosynthetic pathway.
NN
OH
O
NN
AMP
O
S
N
O
N HN
N
O
N
S
O
OH
NH
N
O
N
HN
O
HO
O
S
HN
N
O
N
NH
O
OH
O
N
S
O
SH
NH
N
O
N
HN
O
HO
O
N
N
O
HS
O
S
Ecm1 FabC
C AT E C AT C A MT C A MT TE AT
Ecm6 Ecm7
NH
N
N
O
N
N
HN O
HN
NH
Ar
Ar
O
O
O
O
O
O
O
O
O
O
HS
SH
Ecm17
Ecm18
N
H
S
NH
2
O
N
H
S
NH
2
O OH
N
H
OH
OH
NH
2
O
N
H
OH
O
NH
2
Ecm13
L-tryptophan
Ecm13 Ecm12 Ecm2
Ecm8
Ecm13
SH
Ecm13
+
NH
O H
O
OH O
H
2
N
OH
NH
2
H
2
N
O
OH
O OH
β-hydroxykynurenine
Ecm14 Ecm11
AT AT AT
NH
NH
2
OH
HO
O
OH O
NH
2
CO
2
H
HN
O
OH
H
2
O
HO NH
2
NH
OH
HO
O
O N
N
O
OH
O
Quinoxaline-2-
carboxylic acid (QXC)
H
N
N
O
OH
NH
NH
2
OH
O
O
Ecm4
Ecm3
triostin A
echinomycin
Scheme 1-1. Proposed Biosynthetic Pathway and Modular Organization of Echinomycin
Biosynthesis. The catalytic domains found within the enzymes are represented by E,
epimerzation; M, methylation; and TE, thioesterase.
Results from the study found that only one substrate, (2S, 3S)-ß-hydroxytryptophan
afforded QXC when fed to S. lasaliensis (Koketsu, Oguri et al. 2006).
Thus, they were able to establish the absolute configuration of a key intermediate for
13
biosynthesis of QXC and plans of deciphering this pathway in vitro to its entirety are
ongoing.
Figure 1-5. Complete Echinomycin Biosynthetic Genes. (a) Echinomycin biosynthetic
gene cluster from Streptomyces lasaliensis. (b) Predicted fatty acid synthase gene
organization in S. lasaliensis
Analysis into echinomycin’s peptide core revealed that the cluster contained four intact
NRPS modules which are responsible for its assembly. Analysis of the deduced
adenylation (A) domain’s nucleotide sequence using a web-based guide protocol (NCBI)
helped to reveal that the first, third and fourth A domain are capable of recognizing and
ultimately activating L-serine, L-alanine, L-cysteine and L-valine for consecutive addition
downstream, respectively. In addition, an epimerization (E) domain in module 1 of the
four modules catalyzes the conversion of L- to D-serine and a methyltransferase (M)
domain for N-methylcysteine synthesis in module 3 and N-methylvaline synthesis in
module 4 were also identified to provide additional clues to deduce their position in the
assembly line. Furthermore, analysis of reported multi-module NRP systems, such as
gramicidin S and enterobactin suggests that the echinomycin-synthesizing NRPS system
is comprised of four modules capped off with a thioesterase domain capable of peptide
14
chain homodimerization and cyclorelease (Scheme 1-3a). Further evaluation of the gene
cluster failed to reveal an independent aryl carrier protein (ArCP) in the gene cluster.
While waiting for the completion and refinement of the gene cluster sequence, Keller et
al uncovered an adenylation enzyme responsible for transferring QXC to the ArCP of the
echinomycin biosynthetic pathway by directly examining the function of purified protein
isolated from a separate echinomycin producing strain S. echinatus (Schmoock et al.
2005). They demonstrated that the echinomycin biosynthetic pathway recruits an acyl
carrier protein (ACP) from a fatty acid synthesis pathway during its biosynthesis in place
of the putative ArCP. This result parallels our laboratory’s independent sequencing efforts
thereby justifying the absence of an ArCP from the echinomycin biosynthetic cluster
(Schmoock et al. 2005). Following results published by Keller’s group, the fatty acid
ACP was later isolated from S. lasaliensis and ultimately incorporated into our
biosynthesis model. Uncovering a single methyltransferase that is capable of
biotransforming a disulfide bridge into a thioacetal moiety is a significant finding.
15
H
3
CS
B
Ecm18
S S CH
3
S
S
S
S
H
3
C
H
SAM
S
S
H
3
C
NH
N
O
N
HN
O
O
N
N
O
O
O
HO
HS X
SH
NH
N
O
N
HN
O
O
N
N
O
O
S
NH
N
O
N
HN
O
O
N
N
O
O
O
HS
HO HO
HS
NH
N
O
N
HN
O
O
N
N
O
O
O
HN
N
O
N
NH
O
O
N
N
O
HS
O
O
OH
SH
HS
Ecm17
a
b
T
TE
Ecm7
T
TE
Ecm7
T
TE
Ecm7
echinomycin
triostin A
triostin A
Scheme 1-2. Proposed Mechanism for the Echinomycin Biosynthesis. (a) Proposed
mechanism for the peptide chain homodimerization and cyclorelease from Ecm7 and
subsequent modifications for formation of triostin A. (b) Proposed mechanism for
formation of the thioacetal bridge.
Ttriostin A has been determined to be a direct precursor to echinomycin for two
decades. Undergoing a radical producing reaction, triostin A is converted to echinomycin
through the formation of a unique thioacetal bridge, characteristic of echinomycin
(Cornish et al. 1983). Thought to convert the disulfide bridge in triostin A to a thioacetal
bridge in echinomycin via a sulfoniumylide intermidate (Scheme 1-2b), data obtained
from DNA sequence for the gene cluster divulged a single enzyme, Ecm 18, that is highly
homologous to a known S-adenosyl-L-methionine (Sambrook and Russell)-dependent
methyltransferase. To verify this, our laboratory demonstrated, by means of an in vitro
16
assay, that purified Ecm18 was able to catalyze the conversion of triostin A to
echinomycin in the presence of SAM.
6. Engineering Natural Product Biosynthetic Machinery and its Industrial
Applications
It is highly anticipated that biosynthesis of these valuable compounds via
heterologous production may ameliorate their meager levels normally isolated from its
original host. Today, many commercially available antibiotics are obtained by use of
fermentation technology and are isolated from soil microbial organisms. These strains
are frequently amenable to artificial cultivation under laborious optimization of culture
conditions to reach yields that can be isolated at a level of 10-100 mg per one liter of
culture under specified conditions. Although many natural products can only be isolated
in trace amounts, they are sometimes significant and exhibit potent activities and show
promise in the clinical setting. Because of improvements in analytical instrumentation
and purification tools, isolation of diminutive production levels of these molecules has
been possible. These findings combined with technological progression of natural
products chemistry will offer more opportunities for drug discovery. In order to gain
approval for clinical use, potential therapeutic agents must be subjected to numerous
studies that include animal studies and clinical trials which could sometimes exhaust a
kilograms worth of compound. Moreover, once the compounds pass FDA scrutiny and
are approved for market use, pharmaceutical companies are then required to supply
consumers with the product at a competitively low cost. Until recently, there were
limited alternatives to chemical synthesis for obtaining an adequate supply of such
17
molecules required for mandated studies. While chemically synthesizing such complex
natural products, expensive costs for disposing of environmentally harmful byproducts
that are generated during the process, such as organic solvents and heavy metals, on top
of costs for maintaining a staff of highly knowledgeable and skilled synthetic chemists
make this method economically undesirable. Therefore, providing complex molecules by
way of chemical synthesis for commercial use has often been met with much difficulty.
To circumvent limited compound availability and high cost for consumers, many
researchers have shifted their focus to identifying and isolating gene clusters that are
hypothesized to produce compounds of interest. Possessing a system to express essential
gene clusters that code biosynthetic enzymes responsible for synthesizing natural
products in a heterologous host will increase our arsenal of important and useful
compounds by means of a relatively simpler and low cost method. Additionally,
heterologous production may also provide an increase in productivity relative to its
original source by appendaging a promoter to upregulate gene transcription by means of
exploiting routine molecular biological manipulation.
A cornucopia of marine natural products, including those that have yet to be
discovered, may nourish our depleted supply of potential lead compounds that may
exhibit activities against various carcinomas, viral and bacterial assault, and numerous
other ailments. Just as they are alluring, accessibility and isolation of these compounds
are difficult due to their low production and host residency. Also, producers of these
valuable compounds are frequently organisms that are difficult to cultivate with standard
18
laboratory instruments and require special equipment for their proliferation. Ecteinascidin
743 (ET 743), discodermolide, bryostatin 10, arenastatin A, dolastatin 10 are some
examples of marine natural products that are considered last resort lead compounds found
in nature (Figure 1-6). Detailed analysis of their chemical structures has facilitated
prediction of their biosynthetic pathway and discerned enzymes that may be involved. In
fact, a gene segment isolated from a marine organism has recently been deciphered and
reported as a portion of the bryostatin biosynthetic gene cluster coding for a PKS-type
enzyme (Hildebrand et al. 2004; Sudek et al. 2007). Isolation of a complete gene cluster
responsible for biosynthesizing a marine natural product of interest for use with our
heterologous production system would furnish more opportunities to obtaining the
molecule. ET 743, a natural product found only in a marine organism, exhibits
significant antitumor activity drawing much attention from chemists seeking to
chemically synthesize the compound. This antiproliferative entity exhibits activity higher
than most drugs in clinical use, such as Taxol (Rinehart 2000; Schwartsmann et al. 2001;
Simmons et al. 2005). However, because its structure is extremely complex, complete
synthesis of the compound could not afford a yield requisite for mandated studies and
commercial use. Adding to its complex structure, production to host ratio of ET 743 is
roughly 1 milligram per 1 kilogram of tunicate making procurement of this compound
very demanding (Simmons et al. 2005). Incidentally, two varieties of NRPSs encoding
biosynthetic gene clusters were uncovered from soil-borne bacteria which were identical
to genes responsible for synthesizing saframycin A (Li et al. 2008) and safracin (Velasco
et al. 2005), both of which are structurally similar to ET 743. Combinatorial engineering
of these modules using a compatible heterologous expression system may provide some
19
analogs of ET 743 along with various important and effective intermediates. Imparting
such a tactic will alleviate the need to control stereochemistry and evade low yield
reactions that generate the many byproducts during chemical synthesis of ET 743 to
provide a more desirable titer. Once isolation of the complete biosynthetic gene cluster
for ET 743 can be realized, heterologous production of ET 743 is rightfully possible by
expressing the obligatory genes for synthesizing this desirable molecule in our expression
system.
N
N
OCH
3
HO
H
OH
AcO
ecteinascidin 743
H
S
O
O
O
O
NH
HO
H
3
CO
arenastatin A
O O O O
OH
HO
OH NH
2
OH
discodermolide
O O
O
O
H
3
CO
2
C
O
CO
2
CH
3
OH
HO
OH
O
OH
O
bryostatin 10
S
N
N
H
O
H
3
CO
N
H O
N
CH
3
O
O
N
H H
O
N
H
dolastatin 10
O HN
O
O
HN O
O
O
O
OCH
3
Figure 1-6. Chemical Structures of Marine Natural Products Possessing Potent Antibiotic
Activity.
At the present, it is common practice to chemically derivative a precursor that is
abundantly available in order to obtain the resultant target natural product. Alternatively,
desired modifications are synthetically introduced into natural products to improve their
pharmacological properties and avoid any unforeseen adverse biological effects during
commercial distribution.
20
Functional alteration of an enzyme encoded in the biosynthetic gene cluster by
intentionally introducing a genetic mutation may also provide us with several analogs of
the important natural products. A handful of issues must be considered in order to
produce an unnatural analog by means of intentional enzyme mutagenesis: drafting a
proposed biosynthetic pathway diagram for the molecule is indispensable, in vitro
biochemical studies of the enzyme is also essential to understanding protein expression
and functionality, stability of the enzyme, detailed reaction mechanism of the enzyme,
substrate tolerance, and sometimes elucidation of the protein’s three dimensional
structure. To fulfill as many of the above requirements as possible, it would be simpler to
express each biosynthetic gene as a fusion protein in E. coli to facilitate purifying them
almost to homogeneity versus isolating them from its natural source for reasons that will
be discussed below. Some natural sources are difficult to cultivate if not nearly
impossible and production of target compounds are meager at best making it unfeasible
to isolate them (Li et al. 2008). Once fundamental data has been gathered and expression
of functional mutagenic genes with an agreeable host is confirmed, constructing a target
molecule via a plasmid-borne biosynthetic pathway will not only allow production of
natural products but such an approach can also ease assembly of unnatural natural
products. Furthermore, it has been postulated that PKS or NRPS domain swapping of
assorted biosynthetic clusters, rearrangement of its modules, or deleting an explicit
domain within the biosynthetic machinery can provide rationally tailored molecules
thereby offering us the opportunity to design and produce them in the same manner as we
would via chemical synthesis (Watanabe et al. 2007). In addition, alteration of tailoring
enzymes such as hydroxylase, methyltransferase, acyltransferase and glycosyltransferase
21
may afford novel compounds by fusing diverse substituent or changing the position of the
substituent in an aglycon or backbone core structure.
Use of heterologous host system has gained ground over the past decade as an effective
alternative to preparing natural products and their analogs. In this approach, an entire
biosynthetic pathway for the production of a target natural product is transferred from the
natural organism to a well-characterized, readily culturable model organism such as E.
coli (Watanabe and Oikawa 2007). Although identifying and isolating the biosynthetic
pathway from its source organism is an arduous undertaking, recent advancements in
genetic and molecular biological techniques have eased these sorts of endeavors. Once
the initial hurdle of cloning the biosynthetic pathway is overcome, the system can provide
a number of advantages for preparing natural products and their analogs: (1) Because the
process is biosynthetic in nature, it enjoys all the benefits of the fermentation approach to
natural product production; (2) The fact that a well-characterized model organism is used
for the cultivation greatly reduces the difficulty in optimizing culture conditions for
maximizing yield of a target product; (Blattner et al. 1997) Knowledge of a well-
established biosynthetic pathway for a target compound can render itself much more
controllable than the traditional fermentation approach thus making rational remodeling
of the pathway for its analog production feasible; (4) Essentially, the products
biosynthesized by these enzymes are optically active molecules requiring no intervention
to produce its biologically active conformer by means of laboriously controlling its
stereochemistry. Accordingly, heterologous production of compounds can provide
22
valuable insight for future drug discovery in addition to total chemical synthesis and
semi-synthesis approaches thus far.
23
CHAPTER II: de novo PRODUCTION OF QUINOMYCIN ANTIBIOTICS IN E.
coli
1. Introduction:
In recent years, development of E. coli as a vehicle for heterologous metabolite
biosynthesis has seen considerable progress. The greatest advantage of using E. coli is
the wealth of knowledge available on its metabolic pathways and genetic make-up as
well as the availability of well-established techniques for its genetic manipulation.
Additionally, the ease of E. coli fermentation makes this organism particularly suitable
for metabolite overproduction. Its tolerance toward heterologous protein production and
short doubling time are also vital to the efforts. However, to date, no biologically active
complex natural product synthesized by heterologous polyketide synthase (PKS), NRP
synthetase (NRPS) or mixed PKS/NRPS has been obtained de novo from E. coli.
Therefore, efforts are aimed to establish an E. coli system capable of total biosynthesis of
biologically active forms of NRPs. To this end, echinomycin 1 was chosen as a target
compound (Scheme 2-1). A Proposed pathway for the quinoxaline-2-carboxylic acid
biosynthesis (Scheme 2-1a) and octadepsipeptide core structure (Scheme 2-1b) is
illustrated below along with the proposed mechanism for peptide chain homodimerization
and cyclorelease from Ecm7 and subsequent modifications for the formation of the
quinoxaline antibiotic (Scheme 2-1c).
24
Scheme 2-1. Quinoxaline antibiotics and its proposed mechanism for echinomycin (1) biosynthesis. Enzymes are represented by
A, adenylation; C, condensation; T, thiolation; E, epimerization; M, methyltransferase; and TE, thioesterase. * TANDEM (8)
(Ciardelli, Chkravarty et al. 1978) is a synthetic analog of triostin A (2) . (d) Chemical structures of quinoline-type antibiotics.
25
Echinomycin is a NRP isolated from various bacteria, including Streptomyces
lasaliensis (Steinerova et al. 1987), that belongs to the large family of quinoxaline
antibiotics which boasts two quinoxaline chromophores attached to the C
2
-symmetric
cyclic depsipeptide core structure. Great interest in this group of compounds stems
from its potent antibacterial, anticancer and antiviral activities. Many of the
compounds from this class, including 1 and triostin A (2), exhibit nanomolar potency
(Boger et al. 2001). Also, 1 contains two notable chemical architectures, including its
two quinoxaline-2-carboxylic acid (QXC) moiety and an atypical thioacetal bridge
whose biosynthetic mechanisms remain unknown.
2. Results
The echinomycin biosynthetic gene cluster from the S. lasaliensis linear plasmid
was successfully isolated (Watanabe et al. 2006). DNA sequence analysis of the 36
kilobase-long cluster revealed the presence of eight genes that appeared responsible
for the QXC biosynthesis (ecm2–4,8,11–14), five genes for the peptide backbone
formation and modifications (ecm1,6,7,17,18) and a resistance gene (ecm16) (Figure
2-1).
The final biosynthetic step of echinomycin involves an unusual
transformation of the disulfide bridge of 2 into a thioacetal bridge (Cornish et al.
1983). Ecm18, which is highly homologous to a known S-adenosyl-L-methionine -
26
dependent methyltransferase (Sambrook and Russell), is thought to be responsible
for this transformation (Watanabe et al. 2006). To verify this hypothesis, we
demonstrated in vitro that purified Ecm18 was able to catalyze the transformation of
2 to 1 in the presence of SAM (Figure 2-2a–c).
For E. coli production of 1, following feasibility confirmation of each of the
fifteen gene expression gathered from S. lasaliensis (ecm1–4,6-8,11–14,16–18 and
fabC) in E. coli (Figure 2-3), the fifteen genes were assembled along with the
Bacillus subtilis phosphopantetheine transferase gene sfp, known to efficiently
phosphopantetheinylate heterologous ACPs and T domains (Lambalot et al. 1996),
into three separate plasmids (Figure 2-1) with each gene carrying its own T7
promoter, ribosome binding site and T7 transcriptional terminator. This multi-
monocistronic arrangement was chosen for this multi-gene assembly, not only to
simplify the assembly process but also to minimize potential premature terminations
and mRNA degradation (Sorensen and Mortensen 2005) in transcribing excessively
long polycistronic gene assemblies.
27
Figure 2-1. Echinomycin biosynthetic cluster isolated from Streptomyces lasaliensis. (a) Organization of the echinomycin
biosynthetic gene cluster isolated from S. lasaliensis. (b) Predicted fatty acid synthase gene organization in S. lasaliensis. (c)
Deduced functions of the ORFs of the S. lasaliensis echinomycin biosynthetic gene cluster and fatty acid synthase acyl carrier
protein.
28
Figure 2-2. LC-MS analyses of an Ecm18-catalyzed thioacetal formation. (a) LC-
MS analysis of the crude extract of a reaction mixture for the Ecm18-catalyzed
conversion of 2 to 1. Panels (b-c): UV trace ( λ = 254 nm) is shown on top, while a
MS spectrum is shown beneath. (b) LC-MS analysis of the product 1 isolated from a
crude extract of the reaction mixture. (c) LC-MS analysis of the substrate 2 isolated
from a crude extract of the reaction mixture.
29
Figure 2-3. SDS PAGE analysis of purified echinomycin biosynthetic proteins. (a)
QXC biosynthetic proteins. Lanes 1 and 10: molecular weight marker; lane 2: Ecm13
(63 kDa); lane 3: Ecm2 (27 kDa); lane 4: Ecm3 (36 kDa); lane 5: Ecm4 (52 kDa, red
arrow); lane 6: Ecm8 (9 kDa); lane 7: Ecm11 (25 kDa); lane 8: Ecm12 (44 kDa);
lane 9: Ecm14 (44 kDa). (b) QXC activation and transfer proteins. Lanes 1 and 3:
molecular weight marker; lane 2: Ecm1 (57 kDa); lane 4: FabC (9 kDa). (c) Proteins
responsible for peptide scaffold biosynthesis. Lane 1: molecular weight marker; lane
2: Ecm6 (280 kDa); lane 3: Ecm7 (340 kDa). (d) Modification and self-resistance
protein. Lanes 1 and 4: molecular weight marker; lane 2: Ecm17 (33 kDa); lane 3:
Ecm18 (24 kDa); lane 5: Ecm16 (86kDa).
30
E. coli strain BL21 (DE3) transformed with the three plasmids was subjected
to fed-batch fermentation along with small shake flask cultivation (Table 2-1).
Optimal condition for fed-batch fermentation was achieved using minimal medium
and was allowed to proceed for an eight-day period (Table 2-1, experiment 18).
Isolation of target NRP involved harvesting our 2.0L culture from a New Brunswick
BioFlo 110 fermentor to afford 20.0g of wet cell pellet. Wet cell was extracted with
300mL of equal volume Acetone/chloroform overnight. Supernatant was extracted
twice with equal volumes of ethyl acetate (1.0L), dried with magnesium sulfate and
concentrated under vacuum. Concentrated extract was determined to weigh 73.0mg
with a noticeable greenish hue. Our extract was then diluted with 1.0mL of
methanol and analyzed with an LCQ Deca XP-Plus. Mass spectral analysis (Figure
2-4) confirmed the presence of echinomycin in E. coli culture harvested three days
following protein induction.
31
kw4p49broth #817 RT: 16.75 AV: 1 NL: 3.20E8
T: + c ESI Full ms [ 100.00-1500.00]
1000 1050 1100 1150 1200
m/z
0
10
20
30
40
50
60
70
80
90
100
Relative Abundance
[M+K]+
[M+Na]+
[M+H]+
[M-SCH3]+
1123.4
1101.2
1124.5
1053.4
1102.3
1125.4
1054.4
1103.3
1055.4
1139.3
1104.3
1056.4
1087.3 1141.2
1026.5 1009.4 1173.4 1034.7 1187.3
Figure 2-4. Mass spectral analysis of broth harvested from a fermentor and
extracted with ethylacetate.
Compared to mass spectral analysis of authentic 1 and 2 (Figure 2-5), we observed
four characteristic peaks unique to 1: [M+H]
+
(m/z 1101), [M+K]
+
(m/z 1139), and
[M- SCH
3
]
+
(m/z = 1053) [M+Na]
+
(m/z 1123). We also observed three peaks that
unique to 2; [M+Na]
+
(m/z 1109), [M+H]
+
(m/z 1087), and [M+K]
+
(m/z 1125).
32
Figure 2-5. Spectral and chromatographic analysis of authentic 1 and 2. (a)
Authentic echinomycin (1). (b) Authentic triostin A (2).
33
Table 2-1: Fermentation conditions for cultivation of engineered E. coli.
Experiment Media Supplement DO
2
Agitation
(rpm)
pH Feed Rate
Wet Cell
(g)
Production
of NRP
1 AP2p26 MgSO
4
1mM QXCA
60-
50%
600 7 0.10% n/a
-
2 AP2p30 LB (NH
4
)
2
SO
4
60.0% 600 5.9 0.10% n/a
-
3 AP2p39 2XYT (NH
4
)
2
SO
4
60.0% 600 6.9 0.10% 6.00
-
4 Ap2p44 M9
5g/L casamino
acid Pure, O
2
60.0% 600 6.9 0.10% 51.10
-
5 AP3p4 SB
32g/L
Tryptone,
20g/L yeast
extract 10g/L,
casamino acid
60.0% 600 6.9 0.10% 49.38
-
6 AP3p22 LB QXCA n/a 150 n/a 10mL 30.00
+
7 AP2p27 LB 10 g/L Ala n/a 150 n/a 10mL 13.40
-
8 AP2p28 M9
10g/L Ala,
2nd addition of
IPTG 5 days
later
60.0% 600 6.9 0.10% 124.00
-
9 AP2p34 M9 10g/L Ala 60.0% 600 6.9 0.10% 134.00
-
10 AP2p40 LB 10g/L Ala n/a 150 n/a 10mL 29.00
-
34
Table 2-1: Continued
Experiment Media Supplement DO
2
Agitation
(rpm)
pH Feed Rate
Wet Cell
(g)
Production
of NRP
11 AP2p42 M9 10g/L Ala 60.0% 600 6.9 0.10% 180.00
+
12 AP2p47 M9 10g/L Ala 60.0% 600 6.9 0.10% 52.00
-
13 AP2p50 M9 10g/L Ala 60.0% 600 6.9 0.10% 150.00
+
14 AP2p51 M9 10g/L Ala 60.0% 600 6.9 0.10% 150.00
-
15 AP2p53 M9 10g/L Ala 60.0% 600 6.9 0.10% 250.00
+
16 AP2p56 M9 10g/L Ala 60.0% 600 6.9 0.10% 400.00
-
17 AP2p62 M9 10g/L Ala n/a 150 n/a 10mL n/a
+
18 AP2p67 M9 10g/L Ala 60.0% 600 6.9 0.10% 79.00
+
19 AP2p83 M9 10g/L Ala 73.5% 600 6.9 0.20% n/a
+
Table 2-1: Binary outcome of fermentation process is reported based on detectable production of NRP via LC-MS. Glucose was used
as the carbon source throughout our fermentation study. Aeration rate and cultivation temperature post protein induction was kept
constant at 5 L/min and 15˚C, respectively.
35
With positive results from LC-MS analysis, media extract was then analyzed
with a 400MHz Varian NMR (Figure 2-6). Peaks resolved downfield around 9.6
ppm are characteristic of the proton located on C1 of the pyrazine moiety of
quinoxaline-2-carboxylic acid or more specifically, the chromophore specific to
echinomycin and triostin A. Two peaks were observed around that region and each
was a consequence of the same type of proton but was identifiable to each of the two
quinoxaline chromophores attached to echinomycin.
Figure 2-6. NMR spectrum of crude ethylacetate extraction harvested from a 2L
culture.
It was difficult to assign each proton resolved by the NMR spectra reported in
figure 6 due to detection of protons from contaminants that also migrated to our
organic phase during extraction. Additional steps were required in order to remove
peaks belonging to contaminants in our sample. Medial extract was then applied to a
preparatory thin layer chromatography (PTLC) plate and was developed using 50%
ethylacetate/hexanes to remove less polar contaminants. A proton NMR spectrum
along with an LC-MS analysis was not performed as a consequence of our past
36
experience and knowledge of echinomycin’s resistance to migration on the plate
when developed with this solvent system.
Figure 2-7. Mass spectrum of fractionated mycelium. (a) Mass spectrum of crude
mycelium extracted with acetone and chloroform. (b) Mass spectrum of mycelium
extract fractionated with 50% methanol/chloroform on a column packed with silica.
(c) Mass spectrum of remaining mycelium extract stripped from silica packed
column with 100% methanol.
Therefore, the baseline of our initial PTLC plate was appropriated and our semi-pure
sample was recovered using 20% methanol/chloroform as an eluent. A fraction or
5% (3.8 mg) of the extract was recovered from the application of one PTLC
development. The cell extract was later exposed to additional extraction, twice with
37
an equal volume of chloroform followed by two additional extraction using equal
volumes of acetone. The pooled cell extract was passed through a Buchner funnel
lined with Whatman filter paper to remove all remnants of cell debris. The collected
organic layer was then concentrated under vacuum leaving about 600 mg of a yellow
solid. Crude extract was diluted with methanol and submitted for LC-MS analysis
(Figure 2-7a). From mass spectral analysis, it is clear that 1 was also retained in E.
coli and not all of it was pumped out of its cytosol and into its environment. The cell
extract was then fractionated using silica as the stationary phase. The column loaded
with our crude was later washed with four column volumes of 50% ethyl
acetate/hexanes, fractionated with four column volumes of 50%
methanol/chloroform and later completely stripped with two column volumes of
100% methanol. Fractionated (Figure 2-7b) and stripped (Figure2-7c) samples were
analyzed with the LC-MS to settle on the location of echinomycin post flash
chromatography and to minimize loss of our antibiotic to the column. From figure
2-7b and 2-7c, we were able to conclude that four column volumes of 50%
methanol/chloroform were able to recover echinomycin and leave undetectable
levels of the quinoxaline antibiotic on the column. By weight determination, about
25% (178mg) of semi pure cell extract was recovered from flash column
chromatography and 75% of contaminants were eliminated from our precious
sample. Fractionated cell extract was combined with our semi pure broth extract and
later applied to an additional PTLC plate which was developed with 100% 2-
butanone. A band migrating alongside our reference spot (Rf = 0.50) was observed
and scratched for structural characterization using an NMR with negative results.
38
Peaks downfield at around 9.6 ppm were nonexistent indicating the absence of 1.
Mass spectral data confirmed the absence of our compound retrieved from our
second purification step. Regions half centimeter above our reference spot was
recovered and analyzed with the LC-MS divulging positive results. An NMR
spectrum (Figure 2-8) was then obtained to confirm two things: the presence of 1
from peaks downfield and further removal of unwanted contaminants from our
already complex matrix or disappearance of some major peaks upfield (0-4ppm).
Figure 2-8. NMR spectrum of Echinomycin isolated from E. coli post 100% 2-
butanone PTLC development.
Although development of PTLC plates with 100% 2-butanone helped to resolve
peaks upfield, identification of peaks from 0 to 4 ppm was not obvious and it was not
conclusive proof that the spectra (Figure 2-8) is that of 1. A third and final
purification step should aid in removal of additional unwanted impurity from our
sample and resolve our NMR spectra to provide more valuable information.
39
Figure 2-9. NMR spectrum of 1 isolated from E. coli post 7% methanol/chloroform
PTLC purification.
An amount of about 1.1mg (0.60%) of sample was recovered from our second PTLC
developed with 100% 2-butanone and was applied to a third PTLC plate.
Development of the third plate using 7% methanol/chloroform rendered four bands
with one directly opposite that of our reference spot (Rf = 0.51). Procurement of the
band via elution of scratched silica using 20% methanol/chloroform provided the
spectra displayed in figure 2-9.
40
Figure 2-10.
1
H NMR spectra of authentic quinomycin antibiotics. (a)
Echinomycin 1 (b) Triostin A 2.
.
Although this last step was not able to remove all peaks upfield (0-2ppm), we were
able to resolve a crucial peak at 2.1 ppm. This peak is associated with the S-methyl
moiety of 1’s thioacetal bridge. The culture extract was purified for 1 through a
series of chromatographic steps to give the final yield of 0.3 mg of 1 per liter of
culture. A procedure similar to the one described was implemented to purify 2 from
our engineered bacterium. Solvent system for each step was not modified but
migration differences were anticipated due to chemical differences between 1 and 2.
41
Isolated 2 biosynthesized by our culture was subjected to
1
H NMR analysis (Figure
2-14a) and compared to authentic reference (Figure 2-10).
3. Discussion
Initially, attention was placed on determining the chemical origin of QXC in S.
lasaliensis to decipher gene clusters obtained from the bacterium. Based on previous
findings, L-tryptophan is the precursor of QXCA (Reid, Doddrell et al. 1984).
Because QXCA biosynthesis is similar in nature to the first stage of nikkomycin
biosynthesis (Chen, Hubbard et al. 2002) where Ecm12 hydroxylates L-tryptophan
bound to the thiolation (T) domain of Ecm13, it was possible to predict the functions
for Ecm2–4, Ecm8 and Ecm11–14 (Scheme 2-1a). The product (2S,3S)-β-
hydroxytryptophan 3 (Scheme 2-1a), determined as an intermediate through
substrate feed experiments (unpublished result) is released from Ecm13 by a
thioesterase (TE) , Ecm2. Then, as in the first two steps of the kynurenine
biosynthesis (Kurnasov et al. 2003), oxidative ring-opening of 3 by Ecm11 and
subsequent hydrolysis by Ecm14 can provide β-hydroxykynurenine 4. Subsequently,
oxidative cyclization and hydrolysis of 4 by Ecm4 to form N-(2'-aminophenyl)-β-
hydroxyaspartic acid 5, followed by oxidation of 5 by Ecm3 to yield N-(2'-
aminophenyl)-β-ketoaspartic acid 6. Finally, 6 can undergo spontaneous
decarboxylation, cyclic imine formation and oxidative aromatization to give QXCA.
Through sequence analysis, it became clear that an aryl carrier protein (ArCP)
usually required for incorporating QXCA into 1 was absent from our isolated cluster.
42
However, as proposed for biosynthesis of 2 (Schmoock et al. 2005), it was
speculated that the adenylation (A) domain-containing Ecm1 to activate and transfer
QXCA to the phosphopantetheine arm of FabC, a fatty acid biosynthesis acyl carrier
protein (ACP). The first module of the bimodular NRPS Ecm6 then accepts QC-S-
FabC as a starter unit, while Ecm6 and the second NRPS Ecm7 concludes the
remaining seventeen chemical reaction catalysis to form the peptide core (Scheme 2-
1b). Ecm7 houses a terminal TE domain that appears to homodimerize and
cyclorelease the peptide chain (Scheme 2-1c) (Trauger et al. 2000). The cyclized
product is then accepted as a substrate for an oxidoreductase Ecm17 that is
responsible for catalyzing an oxidation reaction within the reducing cytoplasmic
environment to generate the disulfide bond in 2.
A group of compounds, SW (7,9) (Takahashi et al. 2001) and UK (8,10) (Rance
et al. 1989) isolated from Streptomyces sp. SNA15896 and S. braegensis,
respectively, was found to share an identical backbone while having discrete
modifications at the inter-backbone bridge (Scheme 2-1d). While the conversion of 7
to 8 can proceed in the manner as in the bioconversion of 2 to 1, the formation of 9
and 10 demands attaching alkyl substituents of differing lengths onto the disulfide
bridge. Nonetheless, the reaction mechanism via a sulfonium ylide formation
involving deprotonation and methyl-transfer (Barrett et al. 1997; Van Lanen et al.
2003), similar to the proposed mechanism for the Ecm18-catalyzed reaction, can
account for a single SAM-dependent methyltransferase’s ability to perform an
iterative methylation of a disulfide bridge and subsequent deprotonation and
43
rearrangement to provide 9 and 10 (Scheme 2-2). Because Ecm18 mediated neither
further methylation of 1 nor installation of a thioacetal bridge in TANDEM
11, a
synthetic des-N-methyl derivative of 2 (Ciardelli et al. 1978) (data not shown), it is
suspected that the presence of an additional related methyltransferase that can
iteratively methylate a disulfide bond.
Scheme 2-2. Proposed mechanism for the Ecm18-catalyzed thioacetal formation.
Black arrows represent the proposed steps associated with ECM18-catalyzed
conversion of 2 to 1 in the echinomycin biosynthetic pathway, while red and green
arrows represent the proposed pathways for the biosynthesis of the differently
alkylated thioacetal bridge in the antibiotics from Streptomyces sp. SNA15896 and S.
braegensis, respectively.
44
Since methyltransferases capable of performing multiple rounds of methylation have
not been reported to date, attempts to extract and identify this enigmatic enzyme
from the biosynthetic cluster of 7 and 9 from the SNA15896 genome is ongoing.
Locating the associated methyltransferase(s) should provide more insight into the
mechanism of this unique enzymatic thioacetal bridge formation reaction.
Upon identifying the function of each gene from the cluster to effectively
assemble a construct to house each of the obligatory genes for production of 1,
attention and focus was required to optimize fermentation conditions. A series of
fermentation attempts using both a bench top fed-batch fermentor and small shake
flask were made to cultivate E. coli carrying the entire biosynthetic gene cluster for
production of 1 (Table 2-1). More than halfway through the 18 experiments, a
consensus of media, broth supplement, pH and DO
2
level were reached. M9
Minimal media appears to provide more precise control of culture or bacterial
growth when supplemented with 10 g/L of alanine, pH maintained at 6.9, and a
dissolved oxygen level of 60%. Although a trend was not obvious, there was a
higher probability of success or observed production of 1 when these three
parameters were employed. It was apparent that shake flask cultivation of
metabolically engineered E. coli produced a higher frequency of desirable outcome
versus high density fermentation. Because the high density fermentor was allowed to
continue for a longer period relative to traditional shake flask fermentation and the
vessel was not a closed system, there may be some unrecorded output that we may
have missed during cultivation. A software package that can modulate analogue
45
output from our fermentors into digital data on a computer will be necessary to
determine any variation in culture conditions during our absence in the process.
Supplemental chemical analysis was employed for purified 1 by electrospray
ionization-mass spectrometry (ESI-MS) (Figure 2-10a) and MS/MS (Figure 2-11b).
Comparison of mass spectrum and fragmentation patter of 1 from E. coli is
consistent with the chemical structure of 1.
Figure 2-11. (a) LC-MS spectrograph of 1 isolated from E. coli. (b) Fragment
pattern of 1 isolated from E. coli.
Also, the
1
H NMR spectrum (Figure 2-9) revealed the presence of four
characteristic resonances at δ 9.66 (s, 1 H, QC H-3), 9.64 (s, 1 H, QC H-3), and 2.10
(s, 3 H, -SCH
3
). These peaks were observed for authentic 1 (Figure 2-10a) which
helps to verify to a degree that our purified sample is indeed our expected compound.
Values for identifiable peaks of 1 obtained from figure 2-10a are listed.
46
Additionally, totally correlated spectroscopy (TOCSY) NMR data (Figure 2-12)
provided a more conclusive confirmation that our engineered E. coli possessed the
ability to produce 1. The observed correlation helped to uncover upfield regions of
our
1
H NMR spectra which was masked by an aliphatic contaminant.
Figure 2-12.
1
H NMR TOCSY spectrum of 1 collected at mixing time of 100 ms
showing important correlations (numbered).
47
These findings corroborates for the first time, the identity of obligatory genes to
biosynthesize 1. More importantly, however, the result serves as a first example of
de novo production of a bioactive form of NRP in a heterologous host, E. coli.
Figure 2-13. Chemical characteristic spectra of compound 2 produced by our
engineered E. coli strain. (a) LC-MS spectrum of 2. (b) MS/MS spectra of 2
collected at a collision energy of 4.5 eV, (c) 4.0 eV and (d) 3.0 eV. Important
fragment ions (Dell, Williams et al. 1975) in the MS/MS spectra that are assigned.
Furthermore, to demonstrate the ease and effectiveness of modifying our E.
coli-based heterologous biosynthetic system we chose to convert the echinomycin
biosynthetic pathway into a triostin A biosynthetic pathway. Simple modification of
the plasmid by merely removing ecm18 afforded a strain capable of producing
compound 2 as expected. Confirmed by LC-MS, MS/MS (Figure 2-13) as compared
to its authentic chromatogram and mass spectra (Figure 2-4b) it was determined that
48
the resultant strain produced 2 at a yield of 0.6 mg per liter of culture. Similarly,
1
H
NMR and TOCSY
1
H NMR (Figure 2-14) analyses was acquired as evidence to
corroborate our system’s ability to biosynthesize 2.
Figure 2-14. (a)
1
H NMR spectrum of 2. (b)
1
H NMR TOCSY spectrum of 2
collected with a mixing time of 100 ms, showing important correlations (numbered).
49
Toxicity issues were considered when introducing any exogenous
biosynthetic pathway into E. coli, neglecting this crucial detail can impair the host
and its ability to produce the desired bioactive compound. This problem was
circumvented by introducing a self-resistance mechanism into E. coli that confers
resistance to the host without destroying the product. For echinomycin biosynthesis,
the homology between Ecm16 and daunorubicin resistance-conferring factor DrrC
(Furuya and Hutchinson 1998) and the similarity between the mode of action of 1
and daunorubicin (Neidle et al. 1987) suggested that Ecm16 may be capable of
endowing a non-destructive resistance against 1 in S. lasaliensis. Subsequently, to
demonstrate that ecm16 can in fact confer echinomycin resistance in BL21 (DE3) an
agarose plate assay was performed (data not shown). In the absence of ecm16 from
our system, hampered growth of the host was observed suggesting that sufficient
levels of 1 and 2 would have been unattainable devoid of a self-resistance
mechanism.
For the first time, viability of an E. coli-based biosynthetic system was
successfully demonstrated. Total biosynthesis of a bioactive form of heterologous
complex NRPs from simple carbon and nitrogen sources will pave the way to
developing an economical and general platform for a one-pot mass-production of
natural products and their analogues. This system showed that using a multi-plasmid,
multi-monocistronic gene assembly is a straightforward, highly stable and easily
modifiable approach for establishing and engineering exogenous biosynthetic
pathways in E. coli. With the use of appropriate orthogonal selection markers and
50
origins of replication, in combination with other potential approaches, such as
chromosome integration (Royo et al. 2005), introducing even larger, more complex
biosynthetic pathways seems attainable. Combining our current efforts with the
successes in introducing other PKS (Watanabe et al. 2003; Peiru et al. 2005) and
mixed PKS-NRPS (Pfeifer et al. 2003) pathways into E. coli and engineering of PKS
(Watanabe et al. 2003) and NRPS (Gruenewald et al. 2004) should help broaden the
scope of E. coli-based heterologous mass-production of a wider range of natural
products and their analogues.
51
CHAPTER III: ENHANCED PRODUCTION OF QUINOMYCIN
ANTIBIOTICS IN E coli
1. Introduction
Nonribosomal peptides which include vancomycin, cyclosporine A,
echinomycin and triostin A are celebrated components of a variety of microbial
secondary metabolites possessing biological activities such as antibiotics,
immunosuppressants, and antitumor agents immensely important for clinical use
(Steinerova et al. 1987; Shuler 1994; Boger et al. 2001; Schwarzer et al. 2003). This
category’s broad spectrum of natural products and its structural complexity are a
result of them being biosynthesized by NRP synthetases (NRPSs) encoded in a
single modulated megaenzyme ranging in size of 120 to 180 kDa. Each NRPS
consists of three essential functioning domains: condensation, adenylation and
thiolation. These domains boast the ability to catalyze an amide bond formation
using amino acids as building units for their peptide architecture (Marahiel et al.
1997; Cane et al. 1998; Schwarzer et al. 2003). A single module of this
megasynthetase may also carry a methylation domain for N–methylation of the
peptide backbone and/or epimerization domain for altering an amino acid’s
stereochemistry during the peptide elongation process. Their structural complexity
and diversity are largely attributed to these modules and methods in which they are
synthesized.
52
Nature has provided a generous assortment of NRPs, however, our artillery of
natural products as potential drugs or seeds of promising medication have their
limitations due to indigent material for investigative clinical trial. This obstacle can
be attributed to the original host’s low productivity or unavoidably expensive cost of
multi-step chemical syntheses to avoid undesirable by-products and impart a
favorable pharmacokinetic profile. Many notable NRPs have been isolated from
streptomycetes during the past decade and many biosynthetic genes encoding NRPSs
have been identified and thoroughly sequenced from these bacteria. To understand
the biosynthetic mechanisms, with aims for increasing the yield of production and
intention of pursuing desired analogs, two agreeable expression systems using
Streptomyces lividans and S. coelicolor as hosts with amplifiable expression vectors
were developed by Hopwood et al (Kieser et al. 2000). These appropriated systems
have presented substantial results to provide a more lucid biological understanding of
streptomycetes’ proteins and the production of secondary metabolites, more
markedly, polyketides (PKs) a category of natural products similar to NRPs in terms
of how they are biosynthesized by modular macroenzymes (Cane et al. 1998).
Although the contributions by the model hosts are influential to this field of research,
construction of plasmid vectors and the cultivation of these cells are both time
consuming and problematic.
In recent years, progress has been made using E. coli as a surrogate host for
gene expression of NRPSs and PK synthases (PKSs) (Watanabe and Oikawa 2007).
Despite the complexity of Khosla’s innovative but elaborate multiple–plasmid
53
expression system (Pfeifer et al. 2001; Pfeifer et al. 2003; Watanabe et al. 2003)
positive results for heterologous production of anticipated compounds in E. coli were
achievable.
Figure 3-1. Chemical structures of quinomycin antibiotics.
Use of a heterologous host, E. coli for de novo production of echinomycin (1) and
triostin A (2) (Figure 3-1) (Watanabe et al. 2006) was recently established.
In this chapter, we investigate various parameters to optimize culture conditions
in a fed-batch fermentor and small scale culture. Estimation of triostin A (2)
production is obtained by use of a liquid chromatography–mass spectrometer (LC–
MS) for quantitative analysis of the product to screen for optimal culture conditions.
Production by means of furnishing quinoxaline–2–carboxylic acid (QXC) is
54
evaluated relative to its de novo production to corroborate this chromophore’s role as
the priming unit for biosynthesis of the quinomycin antibiotics (Reid et al. 1984;
Koketsu, Oguri et al. 2006).
2. Results
In our previous report (chapter 2), E. coli expression of a multiple-plasmid
system in M9 minimal medium containing the complete biosynthetic pathway of 2
produced 0.6 mg/L of isolated compound via fed-batch fermentation. Although de
novo production of NRP, 2 was a successful achievement, its titer was modest in
amount despite the use of fed-batch fermentation process. To address this issue, a
series of fermentation studies were implemented.
Table 3-1: Maximal titer of 2.
Feed Rate [2] μg/L Day Observed
0.1% 1.0 1
0.2% 78 3
0.4% 19 4
1.0% 5.0 1
Table 3-1. Observed maximal titer of 2 as a function of rate of nutritional
availability.
Optimal carbon source, pH, temperature, and induction point were determined in the
previous chapter. Here, we begin by analyzing the effects of nutrition feed rate on
55
production of 2. A range of nutrition provided to our fed-batch fermentor range from
0.1% per minute to 1.0% per minute. Daily samples of our culture was harvested
and quantitated. Observed maximal titer and level of 2 appears to depend on
nutrition availability (Table 3-1).
Figure 3-2. Production of 2 according to feed rate of nutrition following protein
expression using a fed-batch fermentor over time. 0.1% feed media per minute ( ♦),
0.2% feed media per minute ( ■), 0.4% feed media per minute ( ▲), 1.0% feed
media per minute ( ●).
A trend is more apparent upon plotting daily production of 2 (Figure 3-2) and we can
extrapolate a desirable range from our data. Although decelerated feed rate of 0.1%
produced 2 on the day following protein expression, such a sluggish rate was not
56
able to increase the titer thereafter. Similarly, a relatively extreme tempo was not
able to provide a desirable titer beyond the observed maximal production level a day
after protein expression. From figure 3-2, decay or disappearance of 2 was also
recorded the second day. The most desirable quatitative level of 2 is accomplished
when a feed tempo of 0.2% is implemented. At a rate of 0.4%, twice that of the
superlative observed titer, the presence of 2 was decreased by more than four folds.
Two parameters of interest, pH and OD
600
, were recorded in order to obtain a
more in-depth representation of culture activity (Figure 3-3). Our fed-batch
fermentor is equipped with a pH control module that is able to sustain a preferred pH
during bacterial cultivation. For our study, the optimal pH was decided on 6.9 and a
dead band was set to 0.2 pH units. For feed tempo of 0.1% and 0.4%, negligible
volume of 1M NH
4
OH was required to maintain the environmental pH. Conversely,
nearly 500 mL of the base was utilized to maintain the same environmental pH at a
feed tempo of 1.0%. When the pH regulator on our fed-batch fermentor was
disabled for our 0.2% feed rate study, a drastic decrease is pH initiated on the third
day of cultivation was observed and recorded during this study. To obtain a clearer
understanding of how feed rate affects pH which ultimately affects bacterial growth
and how all three parameters are interrelated, a pH and OD
600
correlation is
presented below (Figure 3-3). From the figure below, it is apparent that feed rate
affects bacterial growth rate. A two-fold increase in feed tempo resulted in a three-
fold increase in OD
600
which directly relates to growth conditions in our vessel.
There were no apparent affects of pH on culture growth and no obvious affects on
57
biosynthetic activity. For all four conditions, there was a spike in production of 2 in
the vessel followed by its disappearance thereafter. This was observed across the
board which indicated that we must rule out pH as an effecter on biosynthetic
activity.
58
Figure 3-3. Correlation of OD
600
( ) and pH ( ) according to feed rate using a fed-batch fermentor over time. (a) 0.1% feed
rate. (b) 0.2% feed media per minute. (c) 0.4 % feed media per minute. (d) 1.0% feed media per minute. M9 minimal medium was
used throughout this study. Regulation of pH for culture supplied with nutrients at 0.2% per minute was not used.
59
To address the low productivity by means of a simple and quick procedure,
production of 2 was analyzed using small-scale culture (Figure 3-5). Stemming
from previous studies and gathered observations, triostin A’s intact biosynthetic
pathway including QXC’s gene cluster when transformed into E. coli readily
accepted QXC as the starting unit. However, lacking speculative QXC biosynthetic
genes, ecm2, ecm3, ecm4, ecm8, ecm11, ecm12, ecm13, and ecm14 (Scheme 3-1),
the heterologous host could not provide measurable levels of 2 even with exogenous
QXC supplementation. Past experiments directed our titer assessment to focus on
the availability of QXC, the speculative starting unit. It was observed that levels of
the bicyclic chromophore were miniscule and undetectable when its intact
biosynthetic pathway was independently expressed in E. coli.
Table 3-2. Conditions for Maximal Production of 2
Shake Flask Fed–Batch Fermentor
M9 LB M9
Daily feed
13 3 NA
Single dose
7 NA NA
de novo
0.1 NA 0.6 *
Table 3-2. Production of 2 as a result of varied QXC feed and culture medium.
Reported level of compound 2 is shown in milligram of product per liter of culture. *
Isolated quantity of 2 from fed–batch fermentation (Watanabe et al. 2006). NA, Not
applicable; M9, M9 minimal medium; LB, Luria-Bertani medium.
60
Suspecting this step as the encumbrance for desirable antibiotic production,
we investigated three conditions: supply culture daily with QXC exclusive of the
QXC biosynthetic gene cluster (pKW532), single dose of priming unit, and daily
supply of QXC to recombinant bacteria carrying triostin A’s complete biosynthetic
gene cluster. Results are tabulated below (Table 3-2). The first scenario abolished
production of triostin A while a single feed of QXC for a final concentration of 0.1
mg/L at point of gene induction conferred a maximal titer of 7 mg/L of 2 following 7
days of incubation (Figure 3-4C, 3-5). Alternatively, daily supply of QXC initiated
during the point of induction produced a notably higher titer of 2 reaching a
threshold concentration of 13 mg/L when subjected to the same culture duration of 7
days post–induction (Figure 3-4D, 3-5). Moreover, we evaluated the productivity
dependence on medium by comparing NRP production in M9 versus LB media
under similar conditions described for day by day feeding of QXC. LB medium
provided subpar titer of compound 2, reaching maximal concentration of 3 mg/L
after 3–days post–induction (Figure 3-4E, 3-5). Relatively, deplorable de novo
production of compound 2 by using shake flask and M9 medium reached a meager
titer of only 0.1 mg/L (Figure 3-4F, 3-5). In terms of shake flask experiments,
transformants carrying triostin A’s complete biosynthetic pathway with a daily
supply of QXC can produce a titer that is elevated more than 13 orders of magnitude
relative to de novo production of the antibiotic. Noteworthy production under
continuous feed of QXC revealed a titer approximately 22–folds beyond that of de
novo fed–batch fermentation.
61
Figure 3-4. Chromatograms obtained from a UV detector (a) and mass spectrometer
(c–f). A mass spectrum (b) scanning for masses ranging from m/z = 600 to 1500 for
retention time (*) 17.55 min reveal two major ions (m/z = 1087 and m/z = 1109).
Each centroid represents triostin A associated with a proton adduct or sodium adduct,
respectively. Mass–chromatogram was obtained from 1 mL samples for titer
quatitation at each time point according to varying culture conditions: M9 minimal
medium with a single dose of QXC (c), M9 minimal medium with daily dose of
QXC (d), LB medium with daily dose of QXC (e), and M9 minimal medium devoid
of QXC dosing (f).
62
Figure 3-5. Comparison of three 50 mL culture differing in medium and method in
which bicyclic starter unit is furnished. ( ), M9 minimal medium culture with a
single dose of QXC at point of induction; ( z ), M9 minimal medium culture supplied
daily with 5 mg of QXC; ( ), LB medium culture supplied daily with 5 mg of QXC.
The data points are averaged values of three runs.
A pH and cell growth profile is shown in Figure 3-6, negligible influence of
adding QXC, an acidic compound, was observed. Interestingly, using LB medium
for the cultivation of our recombinant E. coli instigated an experimental spike in pH
reaching its peak on the third day after induction. Illustrated in Figure 3-6c,
productivity of 2 was thwarted followed by the degradation of residual antibiotic.
This observation was also documented for production of echinomycin (1) by S.
echinatus, the original producing host of compound 1 (Gauvreau and Waring 1984).
63
In contrast, when using M9 minimal medium, such correlation between compound
productivity and its pH profile was veiled unlike cultures grown in LB medium.
a
b
c
Figure 3-6. Correlation of OD
600
( ) and pH ( ) of small culture over time. (a)
Culture in M9 minimal medium furnished with QXC at point of induction for a final
concentration of 0.1 mg/mL. (b) Culture in M9 minimal medium furnished daily
with 5 mg of QXC. (c) Culture in LB medium furnished daily with 5mg of QXC.
The data points are averaged values of three runs.
64
Effects of DMSO on biosynthetic activity was investigated. The final
concentration of DMSO added to our small culture was, 0%, 0.1%, 0.5%, 2.5%,
5.0% and to 10% (v/v). There was not much difference in production of 2 between
the negative control culture and 0.1% DMSO. A modest decline in titer of 2 was
observed as DMSO concentration is increased to 0.5% and 1.0% respectively. More
drastically, production of 2 was obstructed when the concentration of a permeability
enhancing solvent, DMSO, was escalated to 2.5% and 5.0%. Production of 2 was
negligible at 10% (data not shown). From the data gathered (Figure 3-7), addition
of DMSO did not provide a significant effect on biosynthetic activity.
Figure 3-7. Production of 2 at varying concentration of DMSO using shake flask.
fermentor. 0% DMSO ( ), 0.1% DMSO ( ■), 0.5% DMSO ( ▲), and 1.0% DMSO
( ●), 2.5% DMSO ( ), and 5.0% DMSO ( ○).
65
Unfortunately, the intended effect of DMSO was not observed. A pH and
OD
600
profile is presented below (Figure 3-8). As anticipated, the aprotic chemical
had no obvious influence on culture pH. Conversely, increasing concentration of
DMSO exhibited an undesireable effect on bacterial growth. This trend was
correlated from each culture’s recorded optical density as a function of time.
Optical density observed at 600nm exhibited zero growth velocity with the addition
of 5.0% to 10% of DMSO (Figure 3-8e-g). This suggests that the bacteria in our
culture is capable of biosynthesizing 2 but at a subtle level. From figure 3-8a-d,
increasing DMSO levels in culutre hinders bacterial population which directly
decreases the observed level of 2. The data gathered from this study reveals a
dependence of titer on the success of bacterial growth. This nullifies previous
observation suggesting little dependence of compound production on bacterial
population.
66
a
c
b
d
Figure 3-8. Correlation of OD
600
( ) and pH ( ) according to concentration of DMSO in shakeflask over time. (a) 0% DMSO,
(b) 0.1% DMSO, (c) 0.5 % DMSO, (d) 1.0% DMSO.
67
e
g
f
Figure 3-8 (continued). Correlation of OD
600
( ) and pH ( ) according to concentration of DMSO in shakeflask over time. (e)
2.5% DMSO, (f) 5.0% DMSO, (g) 10% DMSO.
68
NN
OH
O
NN
AMP
O
S
N
O
N HN
N
O
N
S
O
OH
NH
N
O
N
HN
O
HO
O
S
HN
N
O
N
NH
O
OH
O
N
S
O
SH
NH
N
O
N
HN
O
HO
O
N
N
O
HS
O
S
Ecm1 FabC
C AT E C AT C A MT C A MT TE AT
Ecm6 Ecm7
NH
N
N
O
N
N
HN O
HN
NH
Ar
Ar
O
O
O
O
O
O
O
O
O
O
HS
SH
2 1
Ecm17 Ecm18
N
H
S
NH
2
O
N
H
S
NH
2
O OH
N
H
OH
OH
NH
2
O
N
H
OH
O
NH
2
Ecm13
L-tryptophan
Ecm13 Ecm12 Ecm2
Ecm8
Ecm13
SH
Ecm13
+
NH
O H
O
OH O
H
2
N
OH
NH
2
H
2
N
O
OH
O OH
β-hydroxykynurenine
Ecm14 Ecm11
AT AT AT
NH
NH
2
OH
HO
O
OH O
NH
2
CO
2
H
HN
O
OH
H
2
O
HO NH
2
NH
OH
HO
O
O N
N
O
OH
O
Quinoxaline-2-
carboxylic acid (QXC)
H
N
N
O
OH
NH
NH
2
OH
O
O
Ecm4
Ecm3
Scheme 3-1. Proposed biosynthetic pathway and modular organization of echinomycin biosynthesis. The catalytic domains found
within the enzymes are represented by A, adenylation; C, condensation; T, thiolation; E, epimerzation; M, methylation; and TE,
thioesterase.
69
3. Discussion
A noteworthy platform was developed where biosynthetic genes from a complex
matrix was transplanted to a simpler more widely used host, E. coli. Success in such
a feat also requires that these encoded compounds be produced at useable levels.
Mounting evidence has revealed how high cell–density fed–batch bioprocess
technology for metabolic engineering may possess the capability of producing a
more substantial level of 2 by merely supplementing exogenous starting unit during
biosynthesis of 2. Cultivation of E. coli bearing exogenous biosynthetic gene cluster
in a fed-batch fermentor requires finesse. It is a fine balance of nutritional utilization
for growth and significant production of bioactive compounds. Production of these
desirable compounds can be viewed as a biphasic process. Making the bacterium’s
primary growth dormant in order to allow exogenous enzymes the opportunity to
perform their function and ultimately afford our expected compound, 2 in this case.
As our results point out, there is not a linear correlation between nutritional
supplementation and product yield. The observed decrease in pH as feed tempo is
increased may be due to uncontrolled glycolysis and production of acetic acid as a
byproduct thereby influencing the physiology of the bacterium (Kleman and Strohl
1994). Such changes can affect production of recombinant protein (Bech-Jensen and
Carlsen 1990) which explains the disappearance or decrease in production of 2. This
also corroborates the lack of or miniscule detection of 2 along with the precipitous
decline in DO
2
level at higher feed rate. A clearer understanding of this process was
70
attained from our study but an alternate and less tedious approach must be
considered to achieve our desired titer.
One such method that is commonly used in a microbiology laboratory and also
by biochemists is small-scale shake flask. Effects of various parameters on titer of 2
can be explored much more expediently and easily by use of a readily available
technology. In addition to vessel affects, medium, and starter unit influence on titer
of 2 was also explored. This adjustment afforded 2 in amounts closely resembling
that obtained from its natural host, S. triostinicus. Under optimal conditions, the
cultures were then subsidized with QXC to afford 13 mg/L of triostin A, an increase
of more than 130-folds relative to its de novo production in shake flask fermentation.
Increase in biosynthetic activity may is attributed to biomass or oxygen transfer
differences in vessels of choice. Mass transfer coefficient (k
L
a) was determined to
affect metabolic activity of E. coli (Gupta and Rao 2003; Losen et al. 2004). The
reported effect of dissolved oxygen on organism metabolic activity is explained by a
characteristic critical DO
2
level that is organism dependent (Cutayar and Poillon
1989). It is obvious from this set of experiment that the conditions obtained from
this set of shake flask studies are optimal for preferable production of 2. The
challenge lies in scaling up production of 2 and to mimic these conditions in our fed-
batch bioreactor.
Tansporters or pumps are means by which Strepotmyces expell bioactive
compounds, that they produce, both for self-resistence and continued existence
against competing organisms (Linton et al. 1994; Mendez and Salas 1998). This
mechanism is to reduce cytosol accumulation of these toxic compounds. Failed
71
attempts were made to include such a transporter into the plasmid-borne system
which may be blamed on the differences in cell membrane composition of E. coli
and S. lasaliensis. Alternatively, addition of DMSO to aid expulsion of 2 from the
cytosol of E coli was evaluated using shake flask. It has been shown that antibiotics
not effective against certain bacteria such as penicillin to E. coli showed growth
inhibitory effects when the medium contained DMSO (Kamiya et al. 1966; Ghajar
and Harmon 1968). Effective drug action for resistant strains imply an unregulated
affect of DMSO on the cell membrane of the bacterium by decreasing the organism’s
cell impermeability. Intended use of DMSO was not discernible based on gathered
data reported in the previous section.
These findings suggest an indispensable interaction between QXC’s biosynthetic
protein or proteins with either Ecm1 or Ecm1 and FabC. Alleviating modest yields
from de novo production of 2 by simply adding commercial QXC has confirmed its
assembly as the bottleneck. These findings have provided further evidence and
narrowed QXC’s role as the priming unit for biosytnthesis of triostin A. Also,
further studies are necessitated to scaling up production of 2 from shake flask to fed-
batch fermentation.
Facile expression of triostin A’s intact biosynthetic gene cluster in a heterologous
host, E. coli has eased our attempts at garnering data to provide a more
comprehensive picture of triostin A’s dependence on QXC. Although we have
identified and presented the stereochemical assignment for the β–hydroxytryptophan
intermediate in the QXC biosynthetic pathway (Koketsu et al. 2006), it will require
additional investigation in order to aid and circumvent the inherent challenges of
72
biological studies and metabolic engineering for de novo production of 2 in E. coli to
ultimately increase its yield. Further elucidation of the pathway is ongoing.
This observation substantiated QXC as the primer unit of 2 by using a
heterologous host for its biosynthesis while supplying the culture with the
chromophore. By using LC–MS to analyze our small–scale culture and fed-batch
fermentor we obtained indispensable information and better our understanding of the
machinery involved in assembling NRPs using a heterologous host, E. coli.
Equally, testing of precursor–directed biosynthesis by means of feeding an
assortment of chromophore to the heterologous expression system may provide us
with a combinatorialy engineered biosynthesis of unnatural quinomycin antibiotics.
This tolerance is indispensable to the biosynthetic research field because of its
simplicity and speedy assembly line.
73
CHAPTER IV: PRODUCTION OF UNNATURAL NATURAL PRODUCTS
IN A HETEROLOGOUS HOST
1. Introduction
Many biosynthetic genes encoding PKSs, NRPSs and mixed PKS–NRPSs
have been identified and thoroughly sequenced from a variety species. We can take
advantage of these findings and develop an innovative system for heterologous
production of significant secondary metabolites by employing a host that can
suitably express exogenous genes such as streptomycete, yeast or E. coli (McDaniel
et al. 1993). E. coli is an obvious candidate as a heterologous host because they
possess a handful of advantages over their counterparts. Appealingly, there are an
abundance of commercially available expression plasmids that are tolerable for this
bacterium and genetic manipulation of these vectors is relatively straightforward.
Additionally, expression of these genes is reliable and culture doubling time is short.
Successful production of these compounds in E. coli has been presented thereby
generating prospects of exploiting this system for compiling an assortment of
analogs with a higher potency to combat a variety of clinical indications. Because E.
coli features genetic material that is easy to manipulate and assembly of these
macromolecules is linear, a rational approach to designing novel derivatives can be
employed by simple mutagenesis of the biosynthetic genes or reconstituting gene
clusters of mixed biosynthetic pathways. However, there are modest reports of using
E. coli as a heterologous host to synthesize biologically active PKs, NRPs, or mixed
PK–NRPs analog in vivo to date.
In chapter 2, total biosynthesis of two antitumor NRPs, echinomycin (1) and
its intermediate triostin A (2) in E. coli was reported and discussed in more detail.
Originally identified and isolated from streptomycete, both compounds were
74
categorized as quinomycin antibiotics because of their quinoxaline or quinoline
chromophores attached to the C
2
–symmetric cyclic peptide core. This family of
antibiotics exhibit potent antibacterial, anticancer and antiviral activities. In this
chapter, we report an E. coli system that was rationally engineered to synthesize a
biologically active unnatural NRP, TANDEM (8) and a novel echinomycin like
TANDEM christened compound 9 (Figure 4-1).
HN
N
N
O
N
N
NH
O
N
H
H
N
N
N
N
N
O
O
O
O
O
O
O
O
S
SMe
O
O
H
Echinomycin (1)
HN
N
N
O
N
N
NH
O
N
H
H
N
N
N
N
N
O
O
O
O
O
O
O
O O
O
Triostin A (2)
S
S
HN
H
N
N
H
O
H
N
N
H
NH
O
N
H
H
N
N
N
N
N
O
O
O
O
O
O
O
O O
O
TANDEM (8)
S
S
HN
H
N
N
H
O
H
N
N
H
NH
O
N
H
H
N
N
N
N
N
O
O
O
O
O
O
O
O
S
SMe
O
O
H
Compound 9
Figure 4-1. Chemical structure of quinomycin antibiotics.
To demonstrate the versatility and ease of our laboratory’s plasmid–based approach
for establishing and engineering a heterologous biosynthetic pathway in E. coli, we
have chosen to biosynthesize 8. Fundamentally, we conserved the peptide–scaffold’s
integrity by making modest alterations. This rational yet simple approach involves a
six point–mutation in the biosynthetic pathway for compound 1 to afford 8 with
75
structural differences that dramatically influence their binding preference. While
both bis–intercalates into DNA, 1 and 2 exhibits a distinct preference for GC–rich
sequences and 8 on the other hand is more inclined to bind selectively into
alternating AT sequences (Schmoock et al. 2005). A systematic approach taken to
improve titer of 8 will also be discussed in this chapter. This forthright yet
economical approach to assembling a more diverse library of cyclic peptide
antibiotics bearing little variance in its chemical structure can benefit scientists in the
future.
2. Results
Rationally engineering the NRPS system developed in our laboratory to
produce an expected compound was achieved by means of deactivating two
methylation domains (M-domains) of Ecm7 (Figure 4-2c). Of three conserved
motifs (Cornish et al. 1983; Trauger et al. 2000; Takahashi et al. 2001), our target
was the S-adenosylmethionine (Sambrook and Russell) binding site(Cornish et al.
1983). The three point mutation was accomplished using primers presented below
(Figure 4-2b). Protein expression was analyzed and expression of Ecm7* in the
soluble fraction was slightly reduced relative to Ecm7 (Chapter 2).
Results from fed-batch fermentation experiments (Table 4-1) in hopes of
obtaining a generous level of 8 did not provide an obvious trend based on biomass,
DO2 levels or pH. Deactivation of M-domain in Ecm7 by mutating GXGXG to
GXRXG as reported (Cornish et al. 1983) was not able to completely deactivate the
domain. Production of echinomycin (1) was detected in culture to confirm an
insufficient knockdown of M-domain. Mutant constructs were later made to
76
knockdown the M-domain from GXGXG to GXSXG. Feed rate was also
investigated only to confirm our reported outcome in Chapter 3. Much like the
previous chapter, high cell weight or biomass was not an indication of a successful
fermentation run nor does it indicate otherwise.
77
a
b
Primers used to generate Ecm7*
Mutants generated Sequence
G1000S/G1002S/G1004S 5’-CCCGAGCGCGTCCTCGAGCTCAGTGCTAGCACCAGCCTGCTGCTCGCG-3’
G2439S/G2441S/G2443S 5’-CGGGTGCTGGAGATCTCCGTGAGCTCCAGCCTGCTGCTCGGCCCC-3’
c
C A M T C A M T TE
Ecm7*
HN
H
N
N
H
O
H
N
N
H
NH
O
N
H
H
N
N
N
N
N
O
O
O
O
O
O
O
O O
O
TANDEM (8)
S
S
Figure 4-2. Deactivation of M-domains in Ecm7. (a) SDS-PAGE analysis of ecm7*
expression in E. coli.: lane 1, protein maker; lane 2, soluble fraction prepared from
the lysate of E. coli harboring pKW606 carrying the gene for Ecm7* (340 kDa). (b)
Primers used to introduce a six-point mutation in Ecm7. (c) Schematic of domains
targeted (underlined) for point mutation.
78
In addition to fed-batch studies, temperature effect on production of 8 (post-
induction), induction point, colony size and screening experiments were also
performed. Customarily, cultures have been allowed to proceed at 15˚C post
induction. To determine temperature effects on NRP production, culture
environment was cooled from incubation temperature of 37˚C to 22˚C for one study
and later 8˚C for a subsequent study (data not shown). Culture temperature
maintained above and below 15˚C impeded production of 8. Screening experiments
were implemented to secure the selection of a positive colony. Using small shake
flasks, 125mL Erlenmeyer flasks, transformation of our multi-plasmid system
provided a ten percent success rate. Cultures were analyzed using LC-MS and
positive colonies were used to make glycerol stock for subsequent fermentation
experiments. Dependence of induction point was also investigated based on OD
600
reading for fed-batch fermentors and small shake flasks. Evidently, it was observed
for both reaction vessels that addition of IPTG between OD
600
readings of 0.4-1.00
provided more robust production of 8. More importantly, an OD
600
reading of 0.3
demonstrated a more desirable fermentation outcome that was observed repeatedly.
This study complemented our attempts at developing a robust fermentation process
for production of 8. Thinking outside conventional molecular biological technique,
we investigated the correlation of colony size with repeatable fermentation (Figure 4-
3a). There is a direct correlation between colony size and production of 8. Larger
colonies that are picked from the agar plate and used to generate seed cultures
produced 8 at almost twice that of smaller colonies (Figure 4-3b).
79
a
b
Figure 4-3. Effect of colony size on production of 8. (a) Colony on agar plate that is
picked for seed culture: large colony is circled and small colony is enclosed in a
square. (b) Culture production of 8 relative to colony size: small culture ( ■) and
large culture ( □).
80
Table 4-1: Fed-batch fermentation conditions for production of 8 using engineered E. coli.
Experiment Note Supplement DO
2
Agitation
(rpm)
pH Feed Rate
Wet Cell
(g/L)
Production
of 8
AP5p4 n/a
1mM QXC
10 g/L Alanine
60.0% 600 6.9 0.10% 24
-
AP5p5 n/a 10 g/L Alanine 60.0% 600 6.9 0.10% 10
+
AP5p9 n/a 10 g/L Alanine 60.0% 600 6.9 0.10% 9
-
AP5p11**
Ecm7*:
GXRXG
10 g/L Alanine 60.0% 600 6.9 0.10% 13
-
AP5p14
Ecm7*:
SXSXS
10 g/L Alanine 60.0% 600 6.9 0.10% 19
+
AP5p15
Detected 8
only in cell.
10 g/L Alanine 60.0% 600 6.9 0.10% 25
+
AP5p16
Detected 8
only in cell.
10 g/L Alanine 60.0% 600 6.9 0.10% 23
+
AP5p23
Culture time:
4 days
10 g/L Alanine 60.0% 600 6.9 0.10% 14
+
AP5p24
Detected 8
only in cell.
10 g/L Alanine 60.0% 600 6.9 0.10% 22
+
AP5p27
Detected 8
only in cell.
10 g/L Alanine 60.0% 600 6.9 0.10% 25
+
81
Table 4-1 (continued):
Experiment Note Supplement DO
2
Agitation
(rpm)
pH Feed Rate
Wet Cell
(g/L)
Production
of 8
AP5p29 n/a 10 g/L Alanine 60.0% 600 6.9 0.10% 18
+
AP5p32 n/a 10 g/L Alanine 60.0% 600 6.9 0.10% 18
+
AP5p34
Change feed
rate.
10 g/L Alanine 60.0% 600 6.9 0.05% 19
-
AP5p35
Change feed
rate.
10 g/L Alanine 60.0% 600 6.9 0.20% 31
+
AP5p37_a
Use Cascade:
Agitation (50-
900)
10 g/L Alanine 60.0% Cascade 6.9 0.10% 14
-
AP5p37_b
Use Cascade:
Agitation (50-
900)
10 g/L Alanine 60.0% 600 6.9 0.4% 16
-
AP5p38_a n/a 10 g/L Alanine 60.0% 600 6.9 0.10% 21
-
AP5p38_b
Inoculate at
15C.
Induced 3
rd
day
10 g/L Alanine 60.0% 600 6.9 0.10% 26
-
82
Table 4-1 (continued):
Experiment Note Supplement DO
2
Agitation
(rpm)
pH Feed Rate
Wet Cell
(g/L)
Production
of 8
AP5p41
Inoculate at
15˚C.
10 g/L Alanine 60.0% 600 6.9 0.10% 44
-
AP5p44 14 day culture 10 g/L Alanine 60.0% 600 6.9 0.10% 57
+
AP5p46
13 day culture
Base: NaOH
10 g/L Alanine
,
2
)SO
4
60.0% 600 6.9 0.10% 39
-
AP5p47 13 day culture
40% Glycerol
10 g/L Alanine
((NH
4
)
2
)SO
4
60.0% 600 6.9 0.10% 40
-
AP5p50 8 day culture 10 g/L Alanine 60.0% 600 6.9 0.10% 40
+
AP5p52 5 day culture 10 g/L Alanine 60.0% 600 6.9 0.10% 23
+
AP5p54 7 day culture 10 g/L Alanine 60.0% 600 6.9 0.10% 18
+
AP5p60
7 day culture
Discard Cell
10 g/L Alanine 60.0% 600 6.9 0.10% 18
+
Table 4-1. Binary outcome for fed-batch fermentation experiments. M9 minimal medium was used throughout study. **M-
domain of Ecm7 mutated from GXGXG to RXRXR produced detectable levels of echinomycin.
83
Table 4-2: Shake flask fermentation condition for production of 8 using engineered E. coli.
Experiment Note Supplement
Agitation
(rpm)
Flask
Size
Feed media
Wet Cell
(g/L)
Production
of 8
AP5p63
7 day culture
10X500mL
10 g/L Alanine 250 1000mL 50mL 7
+
AP5p65
7 day culture
10X500mL
10 g/L Alanine 250 1000mL 50mL 16
+
AP5p77
7 day culture
10X500mL
10 g/L Alanine 250 1000mL 50mL 7
+
AP5p78
7 day culture
6X500mL
10 g/L Alanine 250 1000mL 50mL 6
+
AP5p81
7 day culture
10X500mL
10 g/L Alanine 250 1000mL 50mL 13
+
AP5p83
7 day culture
10X500mL
10 g/L Alanine 250 1000mL 50mL 12
+
AP5p85
7 day culture
20X60mL
10 g/L Alanine 250 1000mL 10mL 1
+
AP5p104
8 day culture
20X60mL
0.5 g/L Glycine
5mg QXC daily
150 125mL 10mL 9
+
AP5p114
8 day culture
20X60mL
0.5 g/L Glycine
5mg QXC daily
150 125mL 10mL 8
+
AP5p119
8 day culture
20X60mL
0.5 g/L Glycine 150 125mL 10mL 8
+
84
Table 4-2 (continued):
Experiment Note Supplement
Agitation
(rpm)
Flask
Size
Feed media
Wet Cell
(g/L)
Production
of 8
AP5p123
8 day culture
20X60mL
0.5 g/L Glycine
5mg QXC daily
150 125mL 10mL 8
+
AP5p124
7 day culture
20X60mL
0.5 g/L Glycine
5mg QXC daily
150 125mL 10mL 8
+
AP5p126
7 day culture
28X60mL
0.5 g/L Glycine
5mg QXC daily
150 125mL 10mL 7
+
AP5p127
7 day culture
28X60mL
0.5 g/L Glycine
0.5 g/L Cysteine
0.5 g/L Serine
5mg QXC daily
150 125mL 10mL 7
-
AP5p129
7 day culture
28X60mL
0.5 g/L Glycine
5mg QXC daily
150 125mL 10mL 7
+
AP5p133
7 day culture
28X60mL
0.5 g/L Glycine
5mg QXC daily
150 125mL 10mL 8
+
AP5p135
7 day culture
28X60mL
0.5 g/L Isoleucine
5mg QXC daily
150 125mL 10mL 7
+
AP5p136
6 day culture
28X60mL
0.5 g/L Isoleucine
5mg QXC daily
150 125mL 10mL 7
+
AP5p138
7 day culture
28X60mL
0.5 g/L Glycine
0.5 g/L Isoleucine
5mg QXC daily
150 125mL 10mL 7
+
85
Table 4-2 (continued):
Experiment Note Supplement
Agitation
(rpm)
Flask
Size
Feed media
Wet Cell
(g/L)
Production
of 8
AP5p139
6 day culture
36X60mL
New QXC
0.5 g/L Glycine
5mg Isoleucine
5mg QXC daily
150 125mL 10mL 8
-
AP5p140
5day culture
36X60mL
New QXC
0.5 g/L Glycine
5mg Isoleucine
5mg QXC daily
150 125mL 10mL 7
-
AP5p142
6 day culture
28X60mL
0.5 g/L Glycine
5mg QXC daily
150 125mL 10mL 7
+
Table 4-2. Binary outcome for Shake flask fermentation experiments. M9 minimal medium was used throughout study.
86
Following numerous fed-batch fermentation studies, a change in reaction
vessel selection, shake flask versus a bioreactor, was later performed. Although fed-
batch fermentation provided high cell density, production of NRP did not seem
optimal and was not dependent on biomass. Screening experiments described above
revealed the repeatability of shake flasks to provide 8 almost 90% of the time (Table
4-2). A higher abundance of 8 was detected when culture was initiated in a 125 mL
shake flask versus a 1000 mL flask. Replacement of alanine with glycine provided a
more desirable titer of 8. Negative detection of 8 when culture was supplemented
with isoleucine along with glycine was due to change in 3 vendor. Biomass from
shake flask experiments was consistent from batch to batch.
Purification of 8 was achieved as described in experimental section. NMR
analysis of band (R
f
= 0.25) that was developed using an 8% MeOH/CHCl
3
revealed
a contaminant that resolved peaks masking regions upfield. A final purification step
using an HPLC was required in order to remove this contaminant (Figure 4-4a). The
sample was fractionated by minute and identification of the residence time of 8 was
obtained with the aid of a LC-MS (Figure 4-4b-c). From LC-MS analysis, the final
purification step was able to provide 8 at high purity.
87
a
b
c
Figure 4-4. Chromatographic analysis of 8 obtained from culture. (a) Chromatogram
obtained from HPLC following PTLC purification. (b) LC-MS chromatogram of peak
(*) fractionated from HPLC. (c) Mass spectrum of (*).
88
Purification of 9 was accomplished as described in the experimental section.
Separation of 8 and 9 required the change in chromatographic phases. Reverse phase
could not resolve the two compounds that we speculate to be similar in structure,
chemically. Using PTLC which was developed with an 8% MeOH/CHCl
3
solution, the
two compounds were resolved and 9 (R
f
= 0.35) was observed to migrate higher on the
plate. Similarly, PTLC could not provide a high-quality NMR spectra and an additional
step was required to clean up peaks upfield. Using RP-HPLC, our sample was
fractionated (Figure 4-5a) to provide a relatively cleaner chromatogram. LC-MS analysis
confirmed the removal of impurities (Figure 4-5b-c).
89
a
b
c
Figure 4-5. Chromatographic analysis of 9 obtained from culture. (a) Chromatogram
obtained from HPLC following PTLC purification. (b) LC-MS chromatogram of peak
(*) fractionated from HPLC. (c) Mass spectrum of (*).
90
3. Discussion
Bench top fed-batch fermentors have been the vessel of choice for high density cell
fermentation and procurement of secondary metabolites in E. coli (Lin and Swartz 1992;
Lau et al. 2004). Attempts were made by our laboratory to utilize the advantages offered
by this heavily employed mode of natural products production using E. coli. Results
varied per experiment and a trend was not clear from the numerous fermentation attempts
(Table 4-1). As attention shifted from fed-batch to shake flask, positive results and
reiterative detection of 8 indicated that aeration or mass transfer in fed-batch fermentation
as the culprit. A non-invasive study of oxygenation in shake flask and the effects of mass
transfer on metabolic tendencies of E. coli were reported (Shin et al. 1996; Gupta and
Rao 2003; Losen et al. 2004). From our own study, we see a continually favorable
outcome from shake flask versus a fed-batch fermentor. The report also demonstrated the
capacity for fed-batch fermentors to mimic shake flask conditions. Transfer of our
culture process in shake flask to a fed-batch fermentor will alleviate the tedious daily
ritual of feeding 3 to improve titer of quinomycin antibiotics.
Many of the crucial hurdles innate in fermentation of E. coli in hopes of producing
NRPs at a desirable yield were addressed and the process was optimized in this chapter.
Availability of building blocks, acetate level, and culture period were examined and steps
to address these issues were reported in this chapter. As the name for this class of
compounds suggests, they are structurally comprised of amino acids that must be
distributed efficiently in E. coli for propagation and various other cellular activities
(Figure 4-6ai). Although this bacterium is capable of synthesizing amino acids (Dunlop
1949), supplementation amino acids used in biosynthesis of 8 may quench this demand
and ultimately increase its titer. An obvious choice of amino acids to supplement would
be serine, alanine, cysteine, and valine. In fact, the addition of these amino acids
91
hindered growth of the bacterium thereby abolishing any production of 8. Our
observation was explained by the fact that valine plays an important role in the branch
chain amino acid pathway and illicit a negative feedback response in high levels
(Andersen et al. 2001). This challenge can be bypassed through the addition of
isoleucine, a component of the same pathway that is able upregulate production of valine.
Regrettably, supplementation of isoleucine to increase the yield of 8 was inconclusive.
Levels of 8 were not raised significantly with increased supply of the branch chain amino
acid.
Along with the advantages of using E. coli, there are disadvantages of using these
bacteria for fermentation studies. Under aerobic conditions, partial oxidation of glucose
impart acetic acid as a major by-product (Reiling et al. 1985; Kleman and Strohl 1994)
that can affect cell growth (Phue and Shiloach 2005). In addition to increasing the
availability of building units, amino acids are also supplemented to alleviate the
detrimental effects of acetic acid (Han et al. 1993). Two of the twenty amino acids have
shown to suppress the negative effects of acetic acid, methionine and glycine. Our
previous fermentation work has also included alanine which had shown little to no
beneficial consequence in E. coli fermentation. Glycine and L-methionine exhibited the
most attractive outcome making them a logical selection for process optimization.
Increase in production of 8 was observed at a minimal. More important, cultivation of
our E. coli strain and anticipated outcome was robust.
Culture conditions for production of 1 and 2 were established in preceding chapters,
more specifically cultivation period. It was ascertained that optimal culture duration of
culture was eight days. Data obtained during our attempts to optimize production of 8
implicate that seven days is optimal (Figure 4-6b). An unforeseen sink was unraveled
during our studies (Figure 4-6aii) making fermentation time even more critical. The ratio
92
of 8 to 9 increased when our culture was allowed to proceed beyond 6 days. Increased
levels of 9 toughened our attempts at isolating 8 from our culture because reverse-phase
HPLC could not resolve these two compounds. From figure 4-6b, six days would
facilitate attempts to purify 8 because both compounds are present at a one-to-one ratio
but it is more important that we allow our culture to produce 8 at its highest level which
would require an additional day of cultivation (Figure 4-6b, line plot). We are
speculating that an unidentified enzyme encoded in E. coli, similar to Ecm18 that may be
responsible for this bioconversion. This alteration was realized while screening our
culture for production of 8. A peak trailing our anticipated fraction was observed with an
exact mass of m/z 1045 (Figure 4-5). This exact mass accounts the assembly of a
thioacetal bridge positioned at the disulfide bridge of 8 or an increase of 14 amu (Figure
4-5c). Additionally, m/z 1067 or [M+Na]
+
is also observed during our LC-MS analysis.
Small amounts of this compound were isolated for NMR analysis but the absolute
structure for this compound could not be established because of its novelty.
93
a
NH
2
HO
R
O
Culture Growth
NRPS
TANDEM (8) Compound 9
Amino Acid (s)
?
i
ii
b
Figure 4-6. (a) Proposed allocation of amino acid in culture (i). Proposed consumption
pathway of 8 (ii). (b) Ratio of 8 and 9 according to cultivation period. Observation is
reported as a percentage of each quinomycin antibiotic relative to both compounds in
each culture: % TANDEM( ■), % E-TANDEM ( □). Observed level of TANDEM relative
to 5 days culture ( ♦).
94
A purification method was developed for our culture (Method 2 of Materials and
Methods) that afforded 100 μg of 8. A simplified method relative to method 1 was
developed to decrease the steps necessary for purifying 8. A normal phase or TLC step
was essential to separate 8 and 9 from our extract due to the inability of C
18
column to
resolve the two compounds.
1
H NMR Spectral analysis of 8 from our culture (Figure 4-
7b) when compared to authentic reference (Figure 4-7a) confirmed the identity of our
compound. Homonuclear correlation analysis,
1
H NMR TOCSY, was also implemented
to verify the authenticity of our compound. Five correlations (Figure 4-8, numbered)
revealed important peaks upfield that were partially masked by water.
1
H NMR Spectral
analysis of 9 from our culture (Figure 4-9) provides some indication that this compound
is a quinomycin derivative. This assumption is made based on peaks resolved from 7.5
ppm to 10.0 ppm that are indicative of protons characteristic of QXC. More importantly,
a peak is resolved at 9.6 ppm which is indicative of the proton located on C1 of the
pyrazine moiety of quinoxaline-2-carboxylic acid. A peak is also resolved that is
characteristic of the methyl moiety of the thioacetal bridge that was observed for 1
(Chapter 2). This phenomenon is not a result of remnant 8 from our purification step
because our LC-MS data did not detect its presence in our sample (Figure 4-5b). These
results are only speculation and additional characterization is needed to fully corroborate
our observation.
95
a
b
Figure 4-7.
1
H NMR (400 MHz, CDCl
3
) spectrum of 8. (a) Authentic reference. (b)
Isolated from culture.
96
Figure 4-8.
1
H NMR TOCSY (400 MHz, CDCl
3
) spectrum of 8. The spectrum was
collected at 100 ms mixing time resolving correlations that were partially masked by a
water peak (numbered).
Figure 4-9.
1
H NMR (400 MHz, CDCl
3
) spectrum of 9.
97
CHAPTER V: MATERIALS AND METHODS
Chemicals.
Antibiotics were used at the following concentrations: carbenicillin 100 µg/mL,
kanamycin 50 µg/mL, and spectinomycin 50 µg/mL. Authentic echinomycin was
purchased from Sigma–Aldrich. Authentic triostin A was kindly provided to our
laboratory by Dr. Michael Waring. QXC was purchased from Sigma–Aldrich. Authentic
TANDEM was kindly provided to our laboratory by Dr. Searcey. Other chemical
reagents were purchased at the highest commercial quality and used without further
purification.
Culture and Feed Media.
M9 minimal medium used was taken from Sambrook and Russell (Sambrook and
Russell 2001). Feed media used for fermentation contained 430 g/L of glucose, 3.90 g/L
of MgSO
4
, 10 g/L of alanine, trace metal (0.278 g/L of FeCl
3
y 6H
2
O, 0.130 g/L of ZnCl
2
,
0.013 g/L of CaCl
2
y 2 H
2
O, 0.021 g/L of NaMoO
4
y 2 H
2
O, 0.190 g/L of CuSO
4
y 5 H
2
O,
0.024 g/L of H
3
BO
3
), and vitamins (0.00420 g/L of riboflavin, 0.05456 g/L of
pantothenic acid, 0.06078 g/L of nicotinic acid, 0.014140 g/L of pyridoxin, 0.00062 g/L
of biotin, and 0.00048 g/L of folic acid).
98
Thioacetal formation assay.
The assay mixture containing 10 µM Ecm18 and 10 mM SAM in 0.1 M Tris-HCl
pH 7.2 was pre-incubated at 30 ˚C for 5 min. After addition of 2 to the final concentration
of 1 mM, the reaction was incubated at 30 ˚C for five minutes before being terminated by
addition of 10% (w/v) SDS. The reaction mixture was extracted with ethyl acetate and
the extract was concentrated in vacuo to give a white residue. A mixture of the substrate
2 and the product 1 was isolated from the extracts using preparative thin-layer
chromatography (PTLC) (5% MeOH/CHCl
3
). The isolated samples of 1 and residual
unreacted substrate 2, along with the mixture were analyzed independently by LC-MS
using Alltech 2.1 x 100 mm C
18
reverse-phase column. Samples were subjected to a
linear gradient of 5 to 95% CH
3
CN (v/v) in H
2
O (Cragg et al. 1997) supplemented with
0.05% (v/v) formic acid over 60 min at a flow rate of 0.15 mL/min at room temperature.
Overexpression and protein purification of the echinomycin biosynthetic genes
exclusive of two NRPS genes ecm6 and ecm7.
BL21 (DE3) harboring the desired plasmid was grown overnight in 10 mL of
2xYTmedium with 50 μg/mL kanamycin at 37 °C. Each liter of fresh 2xYT medium with
50 μg/mL kanamycin was inoculated with 5 mL of an overnight culture and incubated at
37 °C until the optical density at 600 nm (OD
600
) reached 0.7. Then, expression of each
gene was induced with 100 μM isopropylthio- β-D-galactoside (IPTG) at 15 °C.
Incubation was continued for additional 12 h, after which cells were harvested by
centrifugation at 2,500 x g. All subsequent procedures were performed at 4 °C or on ice.
Harvested cells were resuspended in disruption buffer [0.1 M Tris-HCl (pH 7.2), 0.3 M
NaCl, 10 mM imidazole, 10 mM MgCl
2
, 1 mg/mL lysozyme, 10 μg/mL
99
deoxyribonuclease I and 10 % (v/v) glycerol]. Cells were disrupted via sonication, and
the lysate was clarified by centrifugation at 40,000 x g. The supernatant and precipitate
were recovered as a soluble and insoluble fraction, respectively. The soluble fraction
containing protein was applied to a HisTrap HP column (5 mL; Amersham Biosciences)
that was previously equilibrated with binding buffer [0.1 M Tris-HCl (pH 7.2), 0.3 M
NaCl and 10% (v/v) glycerol] supplemented with 10 mM imidazole, at a flow rate of 1
ml/min. The column was washed with the binding buffer and 10 mM imidazole. Loaded
proteins were eluted with a gradient of 10 to 500 mM imidazole over 100 mL of the
binding buffer. Fractions containing protein with target molecular weight were pooled
and dialyzed against 0.1 M Tris-HCl (pH 7.2), 1 mM dithiothreitol (DTT), 1 mM
ethylenediaminetetraacetic acid (EDTA), and 25% (v/v) glycerol. Protein concentration
was estimated using the Bio-Rad protein assay kit with bovine immunoglobulin G as a
standard. Purified protein samples were analyzed by sodium dodecyl sulfate (SDS)-
polyacrylamide gel electrophoresis (PAGE) using Tris-HCl 4–15% gradient gel (Bio-
Rad) stained with Coomassie Brilliant Blue R-250 stain solution (CBB; Bio-Rad).
100
Overexpression and purification of Ecm6 and Ecm7.
Procedure for production of the two modules is follow those described for E. coli
expression of epoC of the epothilone biosynthetic gene cluster (O'Connor et al. 2002).
Cells were harvested by centrifugation at 2,500 x g. Cells resuspended in the disruption
buffer were disrupted by sonication and the lysate was clarified by centrifugation at
40,000 x g. The supernatant was brought to 60% (w/v) saturation with (NH
4
)
2
SO
4
and
precipitate was allowed to form overnight at 4 °C. After centrifugation at 40,000 x g, the
pellet was dissolved in resuspension buffer [0.1 M Tris-HCl (pH 7.2) and 10% (v/v)
glycerol]. The solution was exchanged into fresh resuspension buffer using PD-10
columns (Amersham Biosciences) in order to remove (NH
4
)
2
SO
4
. The sample was loaded
onto a Ni-NTA (QIAGEN) column. After washing the column with 10 mM imidazole in
0.1 M Tris-HCl (pH 7.2), target protein was eluted with 0.1 M imidazole in 0.1 M Tris-
HCl (pH 7.2). Further purification was carried out on an anion-exchange column (HiTrap
Q, Amersham Biosciences). A gradient of 0–1 M NaCl in 0.1 M Tris-HCl (pH 7.2), 1
mM DTT, 1 mM EDTA and 10% (v/v) glycerol was applied at a flow rate of 3 ml/min
over 10 column volumes. Fractions of 3 mL were collected and fractions with desired
protein, which typically eluted at 0.4 M NaCl, were pooled and further concentrated with
Amicon Ultra centrifugal concentrator (Millipore). Purified proteins were analyzed by
SDS PAGE using a 4–15% linear gradient acrylamide gel stained with CBB.
101
E. coli Production of 1 and 2.
BL21 (DE3) transformed with pKW532/pKW538/pKW541 and
pKW532/pKW539/pKW541 for the production of 1 and 2, respectively (Watanabe et al.
2006), was incubated at 37 ˚C overnight in 2 mL LB medium and subsequently in 100
mL M9 minimal medium. The entire culture was used to inoculate 1.5 liters of M9
minimal medium kept at 37 °C, pH 7.0 by a BioFlo110 fermentor system (New
Brunswick Scientific). Once the glucose level from the medium was exhausted as
indicated by a sudden increase of the dissolved oxygen level, feeding of the feed media
(Pfeifer et al. 2003) was initiated. Simultaneously, the temperature was reduced to 15 ˚C
and IPTG was added to the final concentration of 200 µM. The bioreactor was allowed to
progress for duration of eight days.
Analysis of 1 and 2.
Upon induction of gene expression, 1 mL of the culture was collected and
centrifuged to pellet the cell and harvest the supernatant for analysis. The supernatant was
extracted with ethyl acetate (2 x 1 mL) and concentrated in vacuo. The resultant residue
was dissolved with 100 µL of methanol and 15 µL of the resulting mixture was analyzed
by LC–MS using an Alltech 2.1 x 100 mm C
18
reverse-phase column. Samples were
separated on a linear gradient of 5 to 95% CH
3
CN (v/v) in H
2
O supplemented with 0.05%
(v/v) formic acid over 40 min at room temperature and flow rate of 0.1 mL/min. Mass
spectra collected in normal mode were obtained by monitoring ions ranging from m/z =
900 to m/z = 1500. The following optimized values were employed for data acquisition:
capillary temperature 275°C; spray voltage 5 kV; source current 80 µA; capillary voltage
102
12 V; sheath gas N
2
flow 60 (arbitrary units). Accumulated compound 2 in culture was
then extrapolated using our standard curve following steps described above.
Isolation and characterization of 1 and 2 from E. coli.
Following eight days of incubation, the culture was centrifuged to separate the
supernatant and cells. The supernatant and cell pellet were extracted with ethylacetate
and acetone, respectively. The extracts were combined and concentrated in vacuo to give
an oily residue, which was fractionated by silica gel flash column chromatography with
50% MeOH/CHCl
3
. The fractions containing the target compound were collected and
further purified by a series of PTLC (i. 50% EtOAc/hexane, ii. 100% 2-butanone, iii. 5%
MeOH/CHCl
3
) to afford purified samples 1 (Cheung et al. 1978) and 2 (Blake et al.
1977).
1
H NMR (400 MHz, CDCl
3
) of Echinomycin (1).
δ 9.66 (s, 1 H), 9.64 (s, 1 H), 8.85 2 H), 8.68 (d, J = 7.2 Hz, 1 H), 8.19 (t, J = 7.8
Hz, 2 H), 8.05-7.80 (m, 6 H), 6.96 (d, J = 7.2 Hz, 1 H), 6.81 (d, J = 6 Hz, 1 H), 6.48 (d, J
= 8.8 Hz, 1 H), 6.14 (d, J = 10.8 Hz, 1 H), 5.21 (d, J = 10.4 Hz, 1 H), 5.15 (d, J = 9.6 Hz,
1 H), 5.00-4.90 (m, 3 H), 4.85-4.60 (m, 6 H), 3.44 (d, J = 17.2 Hz, 1 H), 3.18 (s, 3 H),
3.10 (s, 3 H), 3.01 (s, 3 H), 3.00 (s, 3 H), 2.87 (dd, J = 16.4 Hz, J = 11.8 Hz, 1 H), 2.36
(m, 2 H), 2.10 (s, 3 H), 1.45-1.35 (m, 6 H), 1.15-1.05 (m, 6 H), 0.95-0.75 (m, 6 H).
103
1
H NMR TOCSY (400 MHz, CDCl
3
) of Echinomycin (1).
1: δ 0.95—0.75 (Val—CH
3
) and 5.15 (Val—H α); 2: δ1.15—1.05 (Val—CH
3
) and
5.21 (Val—H α); 3: δ 1.45—1.35 (Ala—CH
3
) and 5.00—4.90 (Ala—H α); 4: δ 1.45—1.35
(Ala—CH
3
) and 6.96 (Ala—NH), and δ 1.45 (Ala—CH
3
) and 6.81 (Ala—NH)
1
H NMR (400 MHz, CDCl
3
) Triostin A (2).
δ 9.67 (s, 1 H), 9.60 (s, 1 H), 8.98 (d, J = 8.0 Hz, 1 H), 8.85 (d, J = 6.4 Hz, 1 H),
8.39 (d, J = 9.6 Hz, 1 H), 8.25-7.82 (m, 8 H), 7.21 (bs, 1 H), 6.83 (bs, 1 H), 5.73 (t, J =
7.4 Hz, 1 H), 5.22 (d, J = 10.4 Hz, 1 H), 5.15-4.90 (m, 3 H), 4.85-4.70 (m, 1 H), 4.65-
4.35 (m, 4 H), 4.27 (d, J = 10.0 Hz, 1 H), 3.50-3.25 (m, 4 H), 3.32 (s, 3 H), 3.09 (s, 3 H),
3.04 (s, 3 H), 2.99 (s, 3 H), 2.24 (m, 2 H), 1.46 (d, J = 6.8 Hz, 3 H), 1.12 (d, J = 6.4 Hz, 3
H), 1.10 (d, J = 6.8 Hz, 3 H), 1.06 (d, J = 6.8 Hz, 3 H), 0.87 (d, J = 6.8 Hz, 3 H), 0.72 (d,
J = 7.2 Hz, 3 H).
1
H NMR TOCSY (400 MHz, CDCl
3
) of Triostin A (2).
1: δ 0.72 (Ala—CH
3
) and 5.15— 4.90 (Ala—H α); 2: δ 0.72—1.05 (Ala—CH
3
)
and 8.39 (Ala—NH); 3: δ 1.46 (Ala—CH
3
) and 4.85—4.70 (Ala—H α); 4: δ 1.46 (Ala—
CH
3
) and 7.21 (Ala—NH), 5: δ 0.87 (Val—CH
3
) and 4.27 (Val—H α); 6: δ 0.87(Val—
CH
3
) and 5.22 (Val— H α); 7: δ 1.06 (Val—CH
3
) and 4.27 (Val—H α); 8: δ 1.06 (Val—
CH
3
) and 5.22 (Val— H α), 9: δ 1.10 (Val—CH
3
) and 4.27(Val— H α), δ 1.12 (Val—CH
3
)
and 4.27(Val— H α); 10: δ 1.10 (Val—CH
3
) and 5.22 (Val—H α) and δ 1.12 (Val—CH
3
)
and 5.22 (Val— H α).
104
Echinomycin resistance assay.
E. coli echinomycin resistance was determined using BL21 (DE3) transformed
with pKW409 carrying ecm16. The transformant was incubated in 3 mL LB medium at
37 ˚C for 5 hours. The culture was spread on LB agar plates containing two different
concentrations of echinomycin (10 and 100 µg/mL) supplemented either with or without
300 µM IPTG. The plates were incubated at 37 ˚C overnight to determine colony
formation.
E. coli Production of Triostin A (2) in Shake Flask.
For the de novo production of triostin A, pKW541, along with pKW532 and
pKW539 carrying the remainder of the triostin A biosynthetic genes (Watanabe et al.
2006) were introduced into E. coli BL21 (DE3). The transformed cell carrying
pKW532/pKW539/pKW541 was incubated at 37˚C overnight in 2 mL LB medium
supplemented with carbenicillin, spectinomycin, and kanamycin. Subsequently, 0.5 mL
of the culture was used to inoculate 50 mL of M9 minimal medium with the antibiotics
described above. The culture was grown at 37˚C until its OD
600
reached 0.3–0.6, at this
point, the temperature was reduced to 15˚C and IPTG was added to a final concentration
of 200 µM. Simultaneously, the culture was accompanied by the addition of 10 mL of a
feed medium (Pfeifer et al. 2003) and continuously shaken for an additional 8 days at
15˚C at 150 rpm. For precursor fed experiments, we tested two varieties of transformants:
one that included the QXC biosynthetic gene cluster and another carrying only pKW539
(ecm1, 16, 17, fabC, and sfp) and pKW541, the triostin A biosynthetic gene cluster
(Lambalot et al. 1996; Schmoock et al. 2005). The described constructs were introduced
105
into BL21 (DE3) and cultivated for production of 2. Culture conditions for BL21 (DE3)
transformed with pKW532/pKW539/pKW541 are described above together with daily
supplementation of commercially available QXC. The putative starter unit, QXC was
added to our small culture in two manners: a single dose at point of protein expression for
a final concentration of 0.1 mg/mL or daily dose of the bicyclic chomophore was
supplied in the amount of 5 mg per feed initiated at point of protein expression.
Recombinant bacteria carrying all three plasmids were also tested for medium influenced
productivity using LB and M9 media. BL21 (DE3) transformed with pKW539/pKW541
was incubated at 37˚C overnight in 2 mL LB medium supplemented with spectinomycin
and kanamycin. Subsequently, 0.5 mL of the culture was used to inoculate 50 mL of M9
minimal medium with added antibiotics as described above. All other culture conditions
matched those described for de novo production of 2. Daily feed of QXC in the amount
of 5 mg per feed was initiated at induction point. Addition of DMSO at the appropriate
concentration was initiated at point of culture induction simultaneously with the
introduction of QXC.
Quantitative Analysis of Triostin A (2) Production.
Upon induction of gene expression, 1 mL and 100 mL of the culture was collected
from shake flask or fed-batch fermentor, respectively, and centrifuged to pellet the cell
and harvest the supernatant for analysis. The supernatant was extracted twice with ethyl
acetate at equal volume and concentrated in vacuo. The resultant residue was dissolved
with methanol and 15 µL of the resulting mixture was analyzed by LC–MS using an
106
Alltech 2.1 x 100 mm C
18
reverse-phase column. Samples were separated on a linear
gradient of 5 to 95% CH
3
CN (v/v) in H
2
O supplemented with 0.05% (v/v) formic acid
over 40 min at room temperature and flow rate of 0.1 mL/min. To quantitate the
production of 2, a Finnigan LCQ Deca XP mass-spectrometer equipped with an
electrospray probe operating on positive mode was used. We prepared a standard curve
with a range between 1 ng and 10 µg of reference compound 2. Mass spectra collected in
single ion monitoring mode were obtained by monitoring two ions ([M+H]
+
and
[M+Na]
+
, m/z 1087 and 1109, respectively) observed at retention time 17.5 min of the
LC. The following optimized values employed for data acquisition: capillary temperature
275°C; spray voltage 5 kV; source current 80 µA; capillary voltage 12 V; sheath gas N
2
flow 60 (arbitrary units). Accumulated compound 2 in culture was then extrapolated
using our standard curve following steps described above.
Analysis of TANDEM (8) and compound 9 from Shake Flask.
Two liters of the supernatant was extracted with ethyl acetate (3 x 2 liters). Ethyl
acetate extracts of the culture filtrate were combined and concentrated in vacuo to give an
oily material. To ensure the presence of 8 in culture, resultant residue was dissolved with
methanol to an appropriate dilution and 15 µL of the resulting mixture was analyzed by
LC–MS using an Alltech 2.1 x 100 mm C
18
reverse-phase column. Samples were
separated on a linear gradient of 5 to 95% CH
3
CN (v/v) in H
2
O supplemented with 0.05%
(v/v) formic acid over 40 min at room temperature and flow rate of 0.1 mL/min. Mass
spectra collected in normal mode were obtained by monitoring ions ranging from m/z =
400 to 1400. Collected mass spectra was processed to look for 8 and 9 ions (For 8:
[M+H]
+
and [M+Na]
+
, m/z 1031 and 1053, respectively) (For 9: [M+H]
+
and [M+Na]
+
,
107
m/z 1045 and 1067, respectively) observed at retention time 24.6 min and 25.2 min of
the LC.
Purification of TANDEM (8) and compound 9.
Method 1: Following eight days of incubation, the culture was centrifuged to
separate the supernatant and cells. The supernatant and cell pellet were extracted with
ethylacetate and acetone, respectively. The extracts were combined and concentrated in
vacuo to give an oily residue, which was fractionated by silica gel flash column
chromatography with 75% EtOAC/cyclohexane, 2.5% MeOH/CHCl
3
, 5.0%
MeOH/CHCl
3
, and 10% MeOH/CHCl
3
. Fractions containing the target compound were
collected and further purified with an HPLC using a Wako 20 x 250 mm C
18
reverse–
phase column. Samples were separated on a linear gradient of 0 to 100% CH
3
CN (v/v) in
H
2
O supplemented with 0.05% (v/v) formic acid over 45 minutes at a flow rate of 5.5
mL/min
at room temperature. A PTLC (5 X 10 cm plate) was also utilized to further
purify 8 (8% MeOH/CHCl
3
) to afford semi-purified 8 (R
f
= 0.25) and 9 (R
f
= 0.35).
Peaks downfield were observed from 0-5ppm using a 400MHz Varian NMR suggesting
the need for additional purification. Band containing the target compound was further
purified with an HPLC using a Wako 20 x 250 mm C
18
reverse–phase column. Samples
were separated on a linear gradient of 0 to 100% CH
3
CN (v/v) in H
2
O supplemented with
0.05% (v/v) formic acid over 45 minutes at a flow rate of 5.5 mL/min
at room
temperature.
108
Method 2: Oily extract was loaded onto a preparative Waters 25 x 100 mm C
18
reverse–
phase column for purification of the target compound performed with an HPLC. Samples
were separated on a linear gradient of 0 to 100% CH
3
CN (v/v) in H
2
O supplemented with
0.05% (v/v) formic acid over 45 minutes at a flow rate of 15 mL/min at room
temperature. The fractions containing them were collected and further purified by PTLC
(8% MeOH/CHCl
3
). The bands containing our product were collected and eluted with
20% MeOH/CHCl
3
. The eluate fractions containing the target compound were
concentrated and another round of purification was performed with an HPLC using a
Wako 20 x 250 mm C
18
reverse–phase column. Samples were separated on the same
solvent system as described above over 45 minutes at a flow rate of 5.5 mL/min at room
temperature.
1
H NMR (400 MHz, CDCl
3
) Triostin A (2).
δ 9.68 (s, 2 H, QXC H-3), 8.81 (d, 2 H, DSer–NH), 8.58 (d, 2 H, NH Val), 8.25
(d, 2 H, QXC H-8), 8.11 (d, 2 H, QXC H-5), 7.95-7.92 (m, 4 H, QXC-H), 6.43 (d, 2 H,
Cys–NH), 6.28 (d, 2 H, Ala– NH), 5.68 (dd, 2 H, Cys–H α), 5.00 (dd, 2 H, DSer–H β),
4.89 (m, 2 H, DSer–H α), 4.84 (dd, 2 H, Val–H α), 4.66 (dd, 2 H, DSer–H β), 4.43 (quin, 2
H, Ala–H α), 2.94 (d, 4 H, Cys–H β), 2.55 (m, 2 H, Val–H β), 1.38 (d, 6 H, Ala–H β), 1.16
(d, 6 H, Val–Hγ), 1.12 (d, 6 H, Val–Hγ)
109
1
H NMR TOCSY (400 MHz, CDCl
3
) of TANDEM (8).
1: δ 1.17–1.13 (Val–CH
3
) and 2.58–2.50 (Val–Hβ); 2: δ 1.39 (Ala–CH
3
) and 4.43
(Ala–Hα); 3: δ 1.39 (Ala–CH
3
) and 6.27 (Ala–NH); 4: δ 2.99–2.88 (Cys–Hβ) and 5.69
(Cys–Hα); 5: δ 2.99–2.88 (Cys–Hβ) and 6.42 (Cys–NH).
110
BIBLIOGRAPHY
Addess, K. J., D. E. Gilbert, et al. (1992). "Proton NMR studies of [N-MeCys3,N-
MeCys7]TANDEM binding to DNA oligonucleotides: sequence-specific binding at the
TpA site." Biochemistry 31(2): 339-50.
Andersen, D. C., J. Swartz, et al. (2001). "Metabolic oscillations in an E. coli
fermentation." Biotechnol Bioeng 75(2): 212-8.
Armstrong, G. A., M. Alberti, et al. (1989). "Nucleotide sequence, organization, and
nature of the protein products of the carotenoid biosynthesis gene cluster of Rhodobacter
capsulatus." Mol Gen Genet 216(2-3): 254-68.
Barrett, A. G. M., D. Hamprecht, et al. (1997). "Iterative Cyclopropanation: A Concise
Strategy for the Total Synthesis of the Hexacyclopropane Cholesteryl Ester Transfer
Protein Inhibitor U-106305." J Am Chem Soc 119: 8608-8615.
Bech-Jensen, E. and S. Carlsen (1990). "Production of recombinant human growth
hormone in Escherichia coli: expression of different precursors and physiological effects
of glucose, acetate, and salts." Biotechnol. Bioeng. 36: 1-11.
Blake, T. J., J. R. Kalman, et al. (1977). "Two symmetrical conformations of the triostin
antibiotics in solution." Tetrahedron Lett 18: 2821-2624.
Blattner, F. R., G. Plunkett, 3rd, et al. (1997). "The complete genome sequence of
Escherichia coli K-12." Science 277(5331): 1453-74.
Boger, D. L., S. Ichikawa, et al. (2001). "Total syntheses of thiocoraline and BE-22179
and assessment of their DNA binding and biological properties." J Am Chem Soc 123(4):
561-8.
Cane, D. E., C. T. Walsh, et al. (1998). "Harnessing the biosynthetic code: combinations,
permutations, and mutations." Science 282(5386): 63-8.
Chen, H., B. K. Hubbard, et al. (2002). "Formation of beta-hydroxy histidine in the
biosynthesis of nikkomycin antibiotics." Chem Biol 9(1): 103-12.
111
Chen, H. and D. J. Patel (1995). "Solution structure of a quinomycin bisintercalator-DNA
complex." J Mol Biol 246(1): 164-79.
Cheung, H. T., J. Feeney, et al. (1978). "The conformation of echinomycin in solution." J
Am Chem Soc 100: 46-54.
Ciardelli, T. L., P. K. Chkravarty, et al. (1978). "Des-N-tetramethyltriostin A and bis-L-
seryldes-N-tetramethyltriostin A, synthetic analogues of the quinoxaline antibiotics." J
Am Chem Soc 100: 7684-7690.
Cornish, A., M. J. Waring, et al. (1983). "Conversion of triostins to quinomycins by
protoplasts of Streptomyces echinatus." J Antibiot (Tokyo) 36(12): 1664-70.
Cortes, J., S. F. Haydock, et al. (1990). "An unusually large multifunctional polypeptide
in the erythromycin-producing polyketide synthase of Saccharopolyspora erythraea."
Nature 348(6297): 176-8.
Cragg, G. M., D. J. Newman, et al. (1997). "Natural products in drug discovery and
development." J Nat Prod 60(1): 52-60.
Cuesta-Seijo, J. A. and G. M. Sheldrick (2005). "Structures of complexes between
echinomycin and duplex DNA." Acta Crystallogr D Biol Crystallogr 61(Pt 4): 442-8.
Cutayar, J. M. and D. Poillon (1989). "High cell density culture of E. coli in a fed-batch
system with dissolved oxygen as substrate feed indicator." Biotechnology Letters 11(3).
Dell, A., D. H. Williams, et al. (1975). "Structure revision of the antibiotic echinomycin."
J Am Chem Soc 97(9): 2497-502.
Donadio, S., M. J. Staver, et al. (1991). "Modular organization of genes required for
complex polyketide biosynthesis." Science 252(5006): 675-9.
Dunlop, S. G. (1949). "The synthesis of amino acids by Escherichia coli in pure
cultures." J Bacteriol 58(4): 457-62.
Evans, D. H., J. S. Lee, et al. (1982). "A method for the specific inhibition of poly[d(A-
T)] synthesis using the A-T specific quinoxaline antibiotic TANDEM." Can J Biochem
60(2): 131-6.
112
Fox, K. R. (1990). "Footprinting studies of the interaction of quinomycin antibiotic
UK63052 with DNA: comparison with echinomycin." J Antibiot (Tokyo) 43(10): 1307-
15.
Furuya, K. and C. R. Hutchinson (1998). "The DrrC protein of Streptomyces peucetius, a
UvrA-like protein, is a DNA-binding protein whose gene is induced by daunorubicin."
FEMS Microbiol Lett 168(2): 243-9.
Gauvreau, D. and M. J. Waring (1984). "Directed biosynthesis of novel derivatives of
echinomycin. II. Purification and structure elucidation." Can J Microbiol 30(6): 730-8.
Ghajar, B. M. and S. A. Harmon (1968). "The effect of dimethyl sulfoxide (DMSO) on
permeability of Staphylococcus aureus." Biochem Biophys Res Commun 32(6): 940-4.
Gruenewald, S., H. D. Mootz, et al. (2004). "In vivo production of artificial nonribosomal
peptide products in the heterologous host Escherichia coli." Appl Environ Microbiol
70(6): 3282-91.
Gupta, A. and G. Rao (2003). "A study of oxygen transfer in shake flasks using a non-
invasive oxygen sensor." Biotechnology and Bioengineering 84(3): 351-358.
Han, K., J. Hong, et al. (1993). "Relieving Effects of Glycine and Methionine from
Acetic-Acid Inhibition in Escherichia-Coli Fermentation." Biotechnology and
Bioengineering 41(3): 316-324.
Hildebrand, M., L. E. Waggoner, et al. (2004). "bryA: an unusual modular polyketide
synthase gene from the uncultivated bacterial symbiont of the marine bryozoan Bugula
neritina." Chem Biol 11(11): 1543-52.
Hutchinson, C. R. (2003). "Polyketide and non-ribosomal peptide synthases: falling
together by coming apart." Proc Natl Acad Sci U S A 100(6): 3010-2.
Kamiya, S., T. Wakao, et al. (1966). "[On the potentiating effect of DMSO on the
anesthetic eye drops]." Nippon Ganka Kiyo 17(1): 48-54.
Kieser, T., M. J. Bibb, et al. (2000). Practical Streptomyces Genetics,
The John Innes Foundation.
113
Kleman, G. L. and W. R. Strohl (1994). "Acetate metabolism by Escherichia coli in high-
cell-density fermentation." Appl Environ Microbiol 60(11): 3952-8.
Koketsu, K., H. Oguri, et al. (2006). "Identification and stereochemical assignment of the
beta-hydroxytryptophan intermediate in the echinomycin biosynthetic pathway." Org Lett
8(21): 4719-22.
Krause, M. and M. A. Marahiel (1988). "Organization of the biosynthesis genes for the
peptide antibiotic gramicidin S." J Bacteriol 170(10): 4669-74.
Kurnasov, O., L. Jablonski, et al. (2003). "Aerobic tryptophan degradation pathway in
bacteria: novel kynurenine formamidase." FEMS Microbiol Lett 227(2): 219-27.
Kurosawa, K., K. Takahashi, et al. (2001). "SW-163C and E, novel antitumor
depsipeptides produced by Streptomyces sp. I. Taxonomy, fermentation, isolation and
biological activities." J Antibiot (Tokyo) 54(8): 615-21.
Lambalot, R. H., A. M. Gehring, et al. (1996). "A new enzyme superfamily - the
phosphopantetheinyl transferases." Chem Biol 3(11): 923-36.
Lau, J., C. Tran, et al. (2004). "Development of a high cell-density fed-batch bioprocess
for the heterologous production of 6-deoxyerythronolide B in Escherichia coli." J
Biotechnol 110(1): 95-103.
Lee, J. S. and M. J. Waring (1978). "Bifunctional intercalation and sequence specificity
in the binding of quinomycin and triostin antibiotics to deoxyribonucleic acid." Biochem
J 173(1): 115-28.
Lee, J. S. and M. J. Waring (1978). "Interaction between synthetic analogues of
quinoxaline antibiotics and nucleic acids. Changes in mechanism and specificity related
to structural alterations." Biochem J 173(1): 129-44.
Li, L., W. Deng, et al. (2008). "Characterization of the saframycin A gene cluster from
Streptomyces lavendulae NRRL 11002 revealing a nonribosomal peptide synthetase
system for assembling the unusual tetrapeptidyl skeleton in an iterative manner." J
Bacteriol 190(1): 251-63.
114
Lin, N. S. and J. R. Swartz (1992). "Production of Heterologous Proteins from
Recombinant DNA Escherichia coli in Bench Fermentors." Methods: A Companion to
Methods in Enzymology 4: 159-168.
Linton, K. J., H. N. Cooper, et al. (1994). "An ABC-transporter from Streptomyces
longisporoflavus confers resistance to the polyether-ionophore antibiotic tetronasin." Mol
Microbiol 11(4): 777-85.
Losen, M., B. Frolich, et al. (2004). "Effect of oxygen limitation and medium
composition on Escherichia coli fermentation in shake-flask cultures." Biotechnology
Progress 20(4): 1062-1068.
Marahiel, M. A., T. Stachelhaus, et al. (1997). "Modular Peptide Synthetases Involved in
Nonribosomal Peptide Synthesis." Chem Rev 97(7): 2651-2674.
McDaniel, R., S. Ebert-Khosla, et al. (1993). "Engineered biosynthesis of novel
polyketides." Science 262(5139): 1546-50.
Mendez, C. and J. A. Salas (1998). "ABC transporters in antibiotic-producing
actinomycetes." FEMS Microbiol Lett 158(1): 1-8.
Nakaya, M., H. Oguri, et al. (2007). "Relative and absolute configuration of antitumor
agent SW-163D." Biosci Biotechnol Biochem 71(12): 2969-76.
Neidle, S., L. H. Pearl, et al. (1987). "DNA structure and perturbation by drug binding."
Biochem J 243(1): 1-13.
Newman, D. J., G. M. Cragg, et al. (2003). "Natural products as sources of new drugs
over the period 1981-2002." J Nat Prod 66(7): 1022-37.
O'Connor, S. E., H. Chen, et al. (2002). "Enzymatic assembly of epothilones: the EpoC
subunit and reconstitution of the EpoA-ACP/B/C polyketide and nonribosomal peptide
interfaces." Biochemistry 41(17): 5685-94.
Ota, Y., H. Tamegai, et al. (2000). "Butirosin-biosynthetic gene cluster from Bacillus
circulans." J Antibiot (Tokyo) 53(10): 1158-67.
115
Peiru, S., H. G. Menzella, et al. (2005). "Production of the potent antibacterial polyketide
erythromycin C in Escherichia coli." Appl Environ Microbiol 71(5): 2539-47.
Pfeifer, B. A., S. J. Admiraal, et al. (2001). "Biosynthesis of complex polyketides in a
metabolically engineered strain of E. coli." Science 291(5509): 1790-2.
Pfeifer, B. A. and C. Khosla (2001). "Biosynthesis of polyketides in heterologous hosts."
Microbiol Mol Biol Rev 65(1): 106-18.
Pfeifer, B. A., C. C. Wang, et al. (2003). "Biosynthesis of Yersiniabactin, a complex
polyketide-nonribosomal peptide, using Escherichia coli as a heterologous host." Appl
Environ Microbiol 69(11): 6698-702.
Phue, J. N. and J. Shiloach (2005). "Impact of dissolved oxygen concentration on acetate
accumulation and physiology of E. coli BL21, evaluating transcription levels of key
genes at different dissolved oxygen conditions." Metab Eng 7(5-6): 353-63.
Rance, M. J., J. C. Ruddock, et al. (1989). "UK-63,052 complex, new quinomycin
antibiotics from Streptomyces braegensis subsp. japonicus; taxonomy, fermentation,
isolation, characterisation and antimicrobial activity." J Antibiot (Tokyo) 42(2): 206-17.
Reid, D. G., D. M. Doddrell, et al. (1984). "A 15N nuclear magnetic resonance study of
the biosynthesis of quinoxaline antibiotics." Biochim Biophys Acta 798(1): 111-4.
Reiling, H. E., H. Laurila, et al. (1985). "Mass-Culture of Escherichia-Coli - Medium
Development for Low and High-Density Cultivation of Escherichia Coli-B/R in Minimal
and Complex Media." Journal of Biotechnology 2(3-4): 191-206.
Rinehart, K. L. (2000). "Antitumor compounds from tunicates." Med Res Rev 20(1): 1-
27.
Romero, F., F. Espliego, et al. (1997). "Thiocoraline, a new depsipeptide with antitumor
activity produced by a marine Micromonospora. I. Taxonomy, fermentation, isolation,
and biological activities." J Antibiot (Tokyo) 50(9): 734-7.
Royo, J. L., E. Moreno-Ruiz, et al. (2005). "Stable long-term indigo production by
overexpression of dioxygenase genes using a chromosomal integrated cascade expression
circuit." J Biotechnol 116(2): 113-24.
116
Sambrook, J. and D. W. Russell (2001). Molecular Cloning. New York, Cold Spring
Harbor Laboratory Press.
Schmoock, G., F. Pfennig, et al. (2005). "Functional cross-talk between fatty acid
synthesis and nonribosomal peptide synthesis in quinoxaline antibiotic-producing
streptomycetes." J Biol Chem 280(6): 4339-49.
Schwartsmann, G., A. Brondani da Rocha, et al. (2001). "Marine organisms as a source
of new anticancer agents." Lancet Oncol 2(4): 221-5.
Schwarzer, D., R. Finking, et al. (2003). "Nonribosomal peptides: from genes to
products." Nat Prod Rep 20(3): 275-87.
Searle, M. S., J. G. Hall, et al. (1989). "Interaction of the antitumour antibiotic luzopeptin
with the hexanucleotide duplex d(5'-GCATGC)2. One-dimensional and two-dimensional
n.m.r. studies." Biochem J 259(2): 433-41.
Shin, C. S., M. S. Hong, et al. (1996). "Oxygen Transfer Correlation In High Cell Density
Culture of Recombinant E. Coli." Biotechnology Techniques 10(9): 679-682.
Shuler, M. L. (1994). "Bioreactor engineering as an enabling technology to tap
biodiversity. The case of taxol." Ann N Y Acad Sci 745: 455-61.
Simmons, T. L., E. Andrianasolo, et al. (2005). "Marine natural products as anticancer
drugs." Mol Cancer Ther 4(2): 333-42.
Sorensen, H. P. and K. K. Mortensen (2005). "Advanced genetic strategies for
recombinant protein expression in Escherichia coli." J Biotechnol 115(2): 113-28.
Steinerova, N., H. Lipavska, et al. (1987). "Production of quinomycin A in Streptomyces
lasaliensis." Folia Microbiol (Praha) 32(1): 1-5.
Sudek, S., N. B. Lopanik, et al. (2007). "Identification of the putative bryostatin
polyketide synthase gene cluster from "Candidatus Endobugula sertula", the uncultivated
microbial symbiont of the marine bryozoan Bugula neritina." J Nat Prod 70(1): 67-74.
117
Takahashi, K., H. Koshino, et al. (2001). "SW-163C and E, novel antitumor
depsipeptides produced by Streptomyces sp. II. Structure elucidation." J Antibiot (Tokyo)
54(8): 622-7.
Takusagawa, F. (1985). "The role of the cyclic depsipeptide rings in antibiotics." J
Antibiot (Tokyo) 38(11): 1596-604.
Tamegai, H., E. Nango, et al. (2002). "Significance of the 20-kDa subunit of
heterodimeric 2-deoxy-scyllo-inosose synthase for the biosynthesis of butirosin
antibiotics in Bacillus circulans." Biosci Biotechnol Biochem 66(7): 1538-45.
Trauger, J. W., R. M. Kohli, et al. (2000). "Peptide cyclization catalysed by the
thioesterase domain of tyrocidine synthetase." Nature 407(6801): 215-8.
Ughetto, G., A. H. Wang, et al. (1985). "A comparison of the structure of echinomycin
and triostin A complexed to a DNA fragment." Nucleic Acids Res 13(7): 2305-23.
Van Lanen, S. G., S. D. Kinzie, et al. (2003). "tRNA modification by S-
adenosylmethionine:tRNA ribosyltransferase-isomerase. Assay development and
characterization of the recombinant enzyme." J Biol Chem 278(12): 10491-9.
Velasco, A., P. Acebo, et al. (2005). "Molecular characterization of the safracin
biosynthetic pathway from Pseudomonas fluorescens A2-2: designing new cytotoxic
compounds." Mol Microbiol 56(1): 144-54.
Vining, L. C. (1992). "Secondary metabolism, inventive evolution and biochemical
diversity--a review." Gene 115(1-2): 135-40.
Viswamitra, M. A., O. Kennard, et al. (1981). "Structure of TANDEM and its implication
for bifunctional intercalation into DNA." Nature 289(5800): 817-9.
Walsh, C. T. (2004). "Polyketide and nonribosomal peptide antibiotics: modularity and
versatility." Science 303(5665): 1805-10.
Waring, M. and A. Makoff (1974). "Breakdown of pulse-labeled ribonucleic acid and
polysomes in Bacillus megaterium: actions of streptolydigin, echinomycin, and triostins."
Mol Pharmacol 10(2): 214-24.
118
Watanabe, K., K. Hotta, et al. (2006). "Total biosynthesis of antitumor nonribosomal
peptides in Escherichia coli." Nat Chem Biol 2(8): 423-8.
Watanabe, K. and H. Oikawa (2007). "Robust platform for de novo production of
heterologous polyketides and nonribosomal peptides in Escherichia coli." Org Biomol
Chem 5(4): 593-602.
Watanabe, K., A. P. Praseuth, et al. (2007). "A comprehensive and engaging overview of
the type III family of polyketide synthases." Curr Opin Chem Biol 11(3): 279-86.
Watanabe, K., M. A. Rude, et al. (2003). "Engineered biosynthesis of an ansamycin
polyketide precursor in Escherichia coli." Proc Natl Acad Sci U S A 100(17): 9774-8.
Abstract (if available)
Abstract
Nonribosomal peptides (NRPs) are synthesized by modular mega–enzyme called NRP synthetase (NRPS) that catalyze a peptide bond forming reaction using natural amino acid as substrate. The majority of compounds from this class exhibit crucial biological activities such as antibiotic, immunosuppressive, antiviral, and antitumor activities. However, several of these sought-after natural products are often difficult to isolate in adequate amounts from its natural sources due to low production levels or the producing organism's stringent need for expensive and atypical culture apparatus. As we know, enzymes capable of synthesizing these natural products including essential or nonessential secondary metabolites are encoded by biosynthetic genes located on either chromosomal or plasmid DNA.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Targeting human base excision repair as a novel strategy in cancer therapeutics
PDF
Discovery of a novel HIV-1 integrase inhibitor binding site: insight into enzyme structure/function and inhibitor design
PDF
Discovery of new HIV-1 integrase inhibitors
PDF
Genome engineering of filamentous fungi for efficient novel molecule production
PDF
Cytochrome c cationic oligopeptide conjugates: their cellular uptake, intracellular processing and biological activity
PDF
Examinations of secondary metabolites and their production in fungi and bacteria
PDF
Discovery of novel small-molecule compounds with antiviral and anticancer properties
PDF
Targeting cellular redox modulations for pancreatic cancer treatment
PDF
Algorithm development for modeling protein assemblies
PDF
An organic chemistry approach toward the synthesis of valuable biological compounds: synthetic progress toward the Palmerolide A subunits, expeditious enyne coupling via alkynes, and development ...
PDF
Proinsulin-transferrin recombinant fusion protein: mechanism of activation and potential application in diabetes treatment
PDF
Genome mining of secondary metabolites in Aspergillus terreus isolated from the Chernobyl Collection
PDF
Discovery of small molecules for brain cancer treatment
PDF
Fungal polyketides -- Review of recent findings
PDF
Genetic engineering of fungi to enhance the production and elucidate the biosynthesis of bioactive secondary metabolites
PDF
Toward a new (Au)ra of gold-based therapeutics: thio-reactive probes in autoimmunity
PDF
Semi-synthesis of sclerotiorin derivatives for tau aggregate inhibition and antifungal activity
PDF
Application of genome-wide strategies for the mining of secondary metabolite biosynthesis pathways in filamentous fungi
PDF
Harnessing environmental and culture conditions to alter fungal ‘omics’
PDF
Genome mining of natural product biosynthesis pathways in filamentous fungi for novel drug discovery and production
Asset Metadata
Creator
Praseuth, Alex (author)
Core Title
Assembling NRPS modules in e. coli to establish a platform for rational design of biologically active compounds
School
School of Pharmacy
Degree
Doctor of Philosophy
Degree Program
Pharmaceutical Sciences
Publication Date
07/02/2008
Defense Date
06/06/2008
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
NRPS,OAI-PMH Harvest
Language
English
Advisor
Wang, Clay C. C. (
committee chair
), Haworth, Ian S. (
committee member
), Neamati, Nouri (
committee member
), Roberts, Richard W. (
committee member
)
Creator Email
praseuth@gmail.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-m1317
Unique identifier
UC1187574
Identifier
etd-Praseuth-20080702 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-85841 (legacy record id),usctheses-m1317 (legacy record id)
Legacy Identifier
etd-Praseuth-20080702.pdf
Dmrecord
85841
Document Type
Dissertation
Rights
Praseuth, Alex
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Repository Name
Libraries, University of Southern California
Repository Location
Los Angeles, California
Repository Email
cisadmin@lib.usc.edu
Tags
NRPS