Metabolic Engineering of Escherichia coli for Efficient Production of 2

Subscriber access provided by READING UNIV

Article

Metabolic Engineering of Escherichia coli for Efficient Production of 2-Pyrone-4,6-dicarboxylic Acid from Glucose Zi Wei Luo, Won Jun Kim, and Sang Yup Lee ACS Synth. Biol., Just Accepted Manuscript • DOI: 10.1021/acssynbio.8b00281 • Publication Date (Web): 10 Aug 2018 Downloaded from http://pubs.acs.org on August 12, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Synthetic Biology

1

Metabolic Engineering of Escherichia coli for Efficient Production

2

of 2-Pyrone-4,6-dicarboxylic Acid from Glucose

3 4

Zi Wei Luo,†,‡ Won Jun Kim,†,‡ and Sang Yup Lee*,†,‡,§

5 6

†

7

Chemical and Biomolecular Engineering (BK21 Plus Program), Institute for the BioCentury,

8

Korea Advanced Institute of Science and Technology (KAIST), Daejeon 34141, Republic of

9

Korea

Metabolic and Biomolecular Engineering National Research Laboratory, Department of

10

‡

11

Laboratory, KAIST, Daejeon 34141, Republic of Korea

12

§

13

34141, Republic of Korea

Systems Metabolic Engineering and Systems Healthcare Cross-Generation Collaborative

BioProcess Engineering Research Center and BioInformatics Research Center, KAIST, Daejeon

14 15

*Correspondence to: Prof. Sang Yup Lee (Email: [email protected])

16 17

Address: Department of Chemical and Biomolecular Engineering, KAIST, Daejeon 34141,

18

Republic of Korea

19

Tel.: +82-42-350-3930; Fax: +82-42-350-3910

20

1 ACS Paragon Plus Environment

ACS Synthetic Biology 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1

ABSTRACT: 2-Pyrone-4,6-dicarboxylic acid (PDC) is a pseudo-aromatic dicarboxylic acid and

2

is a promising bio-based building block chemical that can be used to make diverse polyesters

3

with novel functionalities. In this study, Escherichia coli was metabolically engineered to

4

produce PDC from glucose. First, an efficient biosynthetic pathway for PDC production from

5

glucose was suggested by in silico metabolic flux simulation. This best pathway employs a

6

single-step biosynthetic route to protocatechuic acid (PCA), a metabolic precursor for PDC

7

biosynthesis. Based on the selected PDC biosynthetic pathway, a shikimate dehydrogenase

8

(encoded by aroE)-deficient E. coli strain was engineered by introducing heterologous genes of

9

different microbial origin encoding enzymes responsible for converting 3-dehydroshikimate

10

(DHS) to PDC, which allowed de novo biosynthesis of PDC from glucose. Next, production of

11

PDC was further improved by applying stepwise rational metabolic engineering strategies. These

12

include elimination of feedback inhibition on 3-deoxy-D-arabino-heptulosonate-7-phosphate

13

synthase (encoded by aroG) by overexpressing a feedback-resistant variant, enhancement of the

14

precursor phosphoenolpyruvate supply by changing the native promoter of the ppsA gene with

15

the strong trc promoter, and reducing accumulation of the major byproduct DHS by

16

overexpression of a DHS importer (encoded by shiA). Furthermore, cofactor (NADP+/NADPH)

17

utilization was manipulated through genetic modifications of the E. coli soluble pyridine

18

nucleotide transhydrogenase (encoded by sthA), and the resultant impact on PDC production was

19

investigated. Fed-batch fermentation of the final engineered E. coli strain allowed production of

20

16.72 g/L of PDC from glucose with the yield and productivity of 0.201 g/g and 0.172 g/L/h,

21

respectively, representing the highest PDC production performance indices reported to date.

22

KEYWORDS: 2-pyrone-4,6-dicarboxylic acid, protocatechuic acid, biomonomer, metabolic

23

engineering, Escherichia coli 2 ACS Paragon Plus Environment

Page 2 of 43

Page 3 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

2-Pyrone-4,6-dicarboxylic acid (PDC) is a dicarboxylic acid with a polar pseudo-aromatic

2

moiety, which naturally occurs during microbial degradation of lignin-derived aromatic

3

compounds by Sphingobium sp. SYK-6 (previously characterized as Shingomonas paucimobilis

4

SYK-6).1 Given the structural similarity between PDC and terephthalic acid (TPA), which is an

5

industrially important platform chemical for the production of various polyesters, e.g.,

6

polyethylene terephthalate (PET),2 PDC has been successfully polymerized with various diols or

7

hydroxyacids by direct dehydration poly-condensation reactions, generating diverse polymers of

8

a novel category possessing unique and promising characteristics.3,4 Due to the pseudo-aromatic

9

nature of PDC, the PDC-derived polymers are thermally stable, possess higher Young’s modulus

10

and undergo easier degradation in the environment compared to PET.3 For these reasons, PDC

11

has been considered as a potential replacement of TPA in aromatic polyesters.

12

To the best of our knowledge, synthesis of PDC by means of chemical methods has never

13

been reported. Over the past decades, a wide range of chemicals and materials have been

14

produced from renewable resources by metabolically engineered microorganisms.5,6 Production

15

of aromatic compounds including aromatic amino acids has been well documented.7,8 However,

16

microbial production of PDC has rarely been reported except for the following two studies. In

17

the first study, whole-cell biotransformation of protocatechuic acid (PCA) into PDC was

18

achieved in a recombinant Pseudomonas putida strain by introducing PCA 4,5-dioxygenase

19

(encoded by ligA and ligB) and 4-carboxy-2-hydroxymuconate-6-semialdehyde (CHMS)

20

dehydrogenase (encoded by ligC) from S. paucimobilis SYK-6.9 In the second study more

21

recently reported, de novo PDC biosynthesis from glucose was realized by constructing a PDC

22

biosynthetic pathway that stems from the shikimate (SHK) pathway.10 However, the PDC titer

23

obtained was rather low (350 mg/L). The authors in this report simultaneously adopted two 3 ACS Paragon Plus Environment


1

different pathway routes to obtain the precursor PCA, which might be less than optimal due to

2

the potential carbon flux loss at multiple nodes. Additionally, all the genes in the above study

3

were overexpressed using the strong T7 promoter from plasmid, which is not necessarily optimal

4

for the production of metabolites including PDC.11,12

5

To develop a bacterial strain capable of efficiently producing PDC from simple carbon

6

sources, we aimed at establishing a synthetic pathway allowing efficient carbon flux throughout

7

the pathway from glucose to PDC (Figure 1) aided by in silico metabolic flux simulation studies.

8

After successful de novo biosynthesis of PDC from glucose by introducing the necessary and

9

screened heterologous enzymes, PDC production was further improved through elimination of

10

feedback repression on the key enzyme in the aromatic pathway, increase of precursors

11

availability and importer engineering for byproduct reduction. In addition, the redox cofactor

12

(NADP+/NADPH) regeneration for optimal PDC biosynthesis was considered, which revealed

13

an important role of redox cofactor balance on PDC production. Ultimately, fed-batch culture of

14

the best-performing engineered strain was performed to demonstrate its potential for the

15

production of PDC from glucose.

16

17

RESULTS AND DISCUSSION

18

Comparison of Different PDC Biosynthetic Pathways via In Silico Flux Response

19

Analysis. The biosynthetic pathway from glucose to PDC comprises two parts (Figure 1). At

20

first, carbon flow is directed from glucose to PCA, the first heterologous metabolite in the PDC

21

biosynthetic pathway. Next, PCA is converted to PDC by an aromatic ring cleavage operon. In E.

22

coli, PCA can be produced through two different routes (Figure 2A). One route is from 34 ACS Paragon Plus Environment

Page 4 of 43

Page 5 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

dehydroshikimate (DHS) to PCA via a single heterologous enzymatic reaction catalyzed by DHS

2

dehydratase, which is designated as the single-step route in this study. Several previous studies

3

have employed this single-step PCA route for building synthetic metabolic pathways leading to

4

production of various PCA-derived chemicals, including adipic acid,13 cis,cis-muconic acid and

5

catechol.14-16 To promote the availability of DHS to be converted into PCA using this single-step

6

route, formation of SHK from DHS was prevented by the inactivation of SHK dehydrogenase.13

7

The other route to PCA synthesis is from 4-hydroxybenzoate (4HBA), which is derived from the

8

end product of SHK pathway and converted into PCA by 4HBA hydroxylase. Combing the

9

conversions of DHS to 4HBA and 4HBA to PCA, it takes totally six enzymatic steps to obtain

10

PCA from DHS through 4HBA, which is designated as the six-step route. Recently, this six-step

11

PCA route has been utilized in a number of studies for constructing novel biosynthesis pathways

12

for value-added aromatic products.17,18

13

Despite both the PCA-supplying routes described above have been reported and used for

14

production of various aromatic chemicals of interest, a systematic comparison between the two

15

has not been made. In order to select more efficient biosynthetic pathway for PDC production,

16

we first performed in silico flux analyses on both the genome-scale metabolic networks

17

employing the single-step or six-step pathway for PDC production from glucose in E. coli. The

18

simulation results indicated that the single-step PDC pathway was more efficient for PDC

19

production than the six-step pathway, as reflected by the higher specific PDC production rate

20

(Figure 2B). The theoretical maximum PDC yields obtainable by employing the single-step

21

pathway and the six-step pathway were calculated to be 0.906 and 0.738 g/g glucose,

22

respectively, suggesting again that the single-step pathway is more efficient for PDC production.

23

It was hypothesized that the six-step pathway is less efficient due to the requirement for 5 ACS Paragon Plus Environment


1

cofactors (e.g., 1 molecule of ATP and 2 molecules of NADPH) (Figure 2A). To verify this

2

assumption, another simulation was performed for the six-step pathway where cofactor

3

requirement was artificially removed (Figure 2A). This artificial six-step pathway without

4

cofactor requirement showed increased specific PDC production rate (Figure 2B), which proved

5

our hypothesis. Based on these results, the single-step pathway was selected for constructing the

6

PDC biosynthetic pathway from glucose in E. coli.

7

Biosynthesis of PDC from Glucose. To construct the PDC biosynthetic pathway based

8

on the single-step PCA route, we first focused on constructing the upstream pathway module

9

from glucose to PCA. As described earlier, PCA can be produced from glucose via the single-

10

step conversion of DHS to PCA catalyzed by DHS dehydratase. Naturally occurring and

11

structurally diverse DHS dehydratases have been discovered in different microbial species

12

(Supplementary Table 1),19-23 and employed in many reports for microbial production of PCA

13

and PCA-derived products.10,13,14,16

14

To identify a suitable DHS dehydratase candidate for efficient production of PCA from

15

glucose, three DHS dehydratases including AroZ of Klebsiella pneumonia KCTC 2208, AsbF of

16

Bacillus thuringiensis ATCC 10792 and QsuB of Corynebacterium glutamicum ATCC 13032

17

were examined for the following reasons. The DHS dehydratase AroZ (encoded by aroZ) has

18

been used in an earlier study to construct a cis,cis-muconic acid biosynthetic pathway, which

19

produced as high as 36.8 g/L of cis,cis-muconic acid in fed-batch culture.13 In a more recent

20

study, the DHS dehydratase AsbF (encoded by asbF) was compared with two other DHS

21

dehydratases for providing PCA as a precursor, showing better performance.14 The DHS

22

dehydratase QsuB (encoded by qsuB) was found in the SHK/quinate degradation pathway of C.

23

glutamicum,24 and it has been expressed in a recombinant plant Arabidopsis to engineer the 6 ACS Paragon Plus Environment

Page 6 of 43

Page 7 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

lignin deposition in cell walls.25 However, the application potential of QsuB for production of

2

PCA and PCA derivatives in E. coli host has not been examined to date.

3

To compare these three DHS dehydratases, the aroZ, asbF and qsuB genes were cloned

4

into plasmid pTrc99A under trc promoter, resulting in plasmids pTrcZ, pTrcF and pTrcB,

5

respectively. These plasmids were transformed individually into E. coli GYT0, which is DHS

6

dehydrogenase-deficient strain of E. coli W3110 constructed by disrupting the aroE gene to

7

block the downstream SHK pathway. Flask cultures showed that the E. coli GYT0 strain

8

harboring pTrcF produced the highest titer of PCA (1.14 g/L), which was 4.4-fold and 1.2-fold

9

that of PCA produced by the recombinant GYT0 strains harboring pTrcZ and pTrcB,

10

respectively (Figure 3A). Thus, the DHS dehydratase AsbF was selected for further experiments

11

that include assembling with the downstream pathway module for converting PCA to PDC.

12

After the successful validation of upstream pathway conversion from glucose to PCA, we

13

then set out to introduce the downstream pathway module from PCA to PDC. As mentioned, this

14

aromatic ring opening process is catalyzed by PCA 4,5-dioxygenase and CHMS dehydrogenase

15

along with a non-enzymatic, spontaneous conversion step. In previous studies on microbial PDC

16

production,9,10 a gene operon containing ligA and ligB genes for PCA 4,5-dioxygenase and ligC

17

for CHMS dehydrogenase from S. paucimobilis SYK-6 were employed. In this study, we

18

employed an alternative operon comprising pmdA, pmdB and pmdC genes from Comamonas

19

testosteroni ATCC 11996 for converting PCA to PDC. The pmdA, pmdB and pmdC genes are

20

functionally homologous to ligA, ligB and ligC,26 with amino acid sequence identities of 48.4%,

21

57.1% and 75.0%, respectively. It was reported that the PCA 4,5-dioxygenase encoded by

22

pmdAB shows less promiscuous activity compared with the one encoded by ligAB.27 Thus, it is

23

of interest to examine the performance of pmdABC operon in biocatalysis of PCA into PDC. To 7 ACS Paragon Plus Environment


1

investigate the proper expression levels of pmdA, pmdB and pmdC for PDC production, three

2

genes were cloned as the native operon into plasmids pTac15K and pET-22b(+) under the tac

3

and T7 promoters, giving plasmids pTacABC and pETABC, respectively. Two recombinant

4

strains, W3110 harboring pTacABC and BL21(DE3) harboring pETABC, were cultured in flasks

5

supplemented with 2.50 g/L of PCA. The results showed that PCA was successfully converted

6

into PDC by both the E. coli transformants (Supplementary Figure 1). No bioconversion was

7

detected in the recombinant E. coli cells harboring the respective empty vectors. Furthermore,

8

when induction levels were varied by applying different concentrations (0.1, 0.5 and 1 mM) of

9

the inducer isopropyl β-D-1-thiogalactopyranoside (IPTG), distinct profiles of the bioconversion

10

from PCA to PDC were observed for the two transformants (Supplementary Figure 1). The E.

11

coli W3110 harboring pTacABC achieved complete conversion of PCA under conditions of all

12

the three IPTG concentrations tested. Particularly under the condition of 1 mM IPTG, the highest

13

PDC titer of 2.67 g/L was obtained from 2.50 g/L of PCA supplemented, which corresponded to

14

a conversion yield of 89.4 mol%. In contrast, the E. coli BL21(DE3) harboring pETABC

15

achieved only partial conversion of PCA, and the final PDC titers were declined as the IPTG

16

concentrations increased. The highest PDC titer of 2.32 g/L at conversion yield of 78.8 mol% for

17

the E. coli BL21(DE3) harboring pETABC was obtained at the IPTG concentration of 0.1 mM.

18

These PCA-feeding experiments indicated that the tac promoter was performing better in driving

19

expression of the pmdABC operon for an optimal PCA to PDC conversion compared to T7

20

promoter. Also, it is noteworthy that intermediary metabolites (e.g., CHMS isoforms) might have

21

also been produced over the course of PCA conversion, as can be seen from the fact that

22

complete PCA conversion did not lead to a PDC yield of 100 mol%. However, the intermediates

23

were not investigated in this study due to analytical difficulties.


Page 8 of 43

Page 9 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

Having constructed both pathway modules from glucose to PCA and from PCA to PDC,

2

the complete biosynthesis pathway from glucose to PDC was then assembled by constructing the

3

plasmid pTacFABC. The E. coli GYT0 strain harboring pTacFABC successfully produced 0.97

4

g/L of PDC in flask culture using glucose as the sole carbon source (Figure 3B). Moreover, no

5

PCA formation was detected in the culture broth, which suggested that the pmdABC operon

6

bears sufficient capability for catalyzing the downstream pathway module from PCA to PDC

7

(Figure 3B). The PDC production was also confirmed by GC-MS analysis (Supplementary

8

Figure 2). Since the key enzyme DHS dehydratase, AsbF, was of heterologous origin, it was

9

examined whether an E. coli-codon optimized asbF gene, designated as asbFopt (Supplementary

10

Table 2), would enhance PDC production; there was a report showing that a codon-optimized

11

asbF gene exhibited higher AsbF activity.14 The asbFopt gene obtained by gene synthesis service

12

was cloned in the same configuration as for the asbF gene, giving pTrcFopt and pTrcFoptABC. In

13

flask culture, however, the E. coli GYT0 strain harboring pTrcFopt produced 0.32 g/L of PCA

14

(Figure 3A), which was less by 71.9% than that obtained with the wild-type asbF; the E. coli

15

GYT0 strain harboring pTrcFoptABC produced 0.55 g/L of PDC (Figure 3B), less by 43.3% than

16

that obtained with the wild-type asbF. The low activity of asbFopt might be due to the formation

17

of inclusion body, which requires further investigation to confirm. Overall, the de novo

18

production of PDC from glucose was successfully accomplished through the use of the B.

19

thuringiensis asbF gene and C. testosteroni pmdABC operon based on the single-step pathway.

20

Redirecting Flux into the Truncated SHK Pathway for Improved PDC Production.

21

After the single-step PDC biosynthetic pathway was functionally established in E. coli, we next

22

focused on engineering the cellular targets that could redirect central metabolic flux towards

23

PDC biosynthetic pathway, aiming at higher PDC productivity. 9 ACS Paragon Plus Environment


1

As the first step in SHK pathway, condensation of erythrose 4-phosphate (E4P) and

2

phosphoenolpyruvate (PEP) generates 3-deoxy-D-arabinoheptulosonate 7-phosphate (DAHP) by

3

the action of DAHP synthase (Figure 1), which dictates the carbon flux directed into SHK

4

pathway. In E. coli, the total activity of DAHP synthase is comprised of three isoenzymes, AorG

5

(accounting for 80%), AroF (15%) and AroH (5%), which are subjected to tight regulation

6

imposed by allosteric control and transcriptional repression.28 Hence, in order to increase cellular

7

DAHP synthase activity, we overexpressed aroGfbr gene encoding a feedback-inhibition resistant

8

mutant of AroG,29 by generating plasmid pBBR1Gfbr. E. coli GYT0 strain harboring pTacFABC

9

and pBBR1Gfbr produced dramatically increased PDC titer (2.07 g/L) (Figure 4), 1.1-fold higher

10

than that obtained by the parental strain in flask culture.

11

As the elevated DAHP synthase activity significantly raised PDC titer, it was speculated

12

that this might in turn cause a higher demand for the precursor substrates E4P and PEP.

13

Therefore, it is necessary to also enhance the intracellular availability of these two key aromatic

14

substrates. To enhance supply of E4P, a commonly utilized approach is to overexpress the

15

transketolase encoded by tktA, which is a key enzyme responsible for E4P synthesis.30,31 The E.

16

coli tktA gene was thus cloned into plasmid pBBR1Gfbr to generate pBBR1Gfbr-A. When

17

pBBR1Gfbr-A was transformed into E. coli GYT0 strain harboring pTacFABC, however, only

18

1.97 g/L of PDC was produced (Figure 4), which was 4.8% lower than that obtained by GYT0

19

strain harboring pTacFABC and pBBR1Gfbr. On the other hand, for enrichment of PEP,

20

strategies including amplification of PEP synthetase (encoded by ppsA) for improved formation

21

of PEP from pyruvate (PYR),31 and inactivation of PYR kinase I (encoded by pykF) and PYR

22

kinase II (encoded by pykA) for preventing consumption of PEP to PYR in glycolysis,32,33 have

23

previously proven to be effective. Based on these validated strategies, we constructed the 10 ACS Paragon Plus Environment

Page 10 of 43

Page 11 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

following engineered strains, GYT1 (carrying ppsA overexpression through exchange of its

2

native promoter with the strong trc promoter), GYT2 (carrying pykF deletion), GYT3 (carrying

3

pykF and pykA double-deletions) and GYT4 (carrying both ppsA overexpression by trc promotor

4

exchange and pykF deletion). When each of these four strains was transformed with pTacFABC

5

and pBBR1Gfbr-A, and tested in flask cultures, it was found that only GYT1 strain harboring

6

pTacFABC and pBBR1Gfbr-A produced slightly higher PDC titer (2.12 g/L) (Figure 4) than that

7

obtained by the parental strain GYT0 harboring pTacFABC and pBBR1Gfbr-A. All of the other

8

three strains harboring pTacFABC and pBBR1Gfbr-A produced less PDC (Figure 4). These

9

results suggest that the availability of precursors E4P and PEP might not be the limiting factor

10

for PDC production from glucose in our engineered strain under the tested culture condition.

11

Although not performed in this study, in silico genome-scale metabolic simulations can be

12

performed for identifying potential bottlenecks and also new engineering targets in the future.

13

To sum up, among all the gene targets examined above, only aroGfbr and ppsA

14

overexpression displayed improved PDC production. Thus, we combined these two positive

15

targets by generating GYT1 strain harboring pTacFABC and pBBR1Gfbr, which produced 2.28

16

g/L of PDC in flask culture (Figure 5), showing a cumulatively positive effect on PDC

17

production. Also, this PDC titer represented a 1.4-fold increase over the preliminary strain GYT0

18

harboring pTacFABC, suggesting that carbon flux was successfully redirected towards PDC

19

biosynthesis through the above efforts.

20

Importer Engineering for Reduced Byproduct DHS Accumulation. During analysis

21

of the culture supernatants of PDC production strains described above, we observed that

22

concentrations of PCA remained very low (0~10 mg/L) (data not shown), but large amounts of



1

DHS were accumulated. For instance, GYT1 strain harboring pTacFABC and pBBR1Gfbr, which

2

produced the highest titer of PDC so far, accumulated 2.52 g/L of DHS (Figure 5).

3

To avoid carbon loss via such a surplus extracellular formation of DHS and, in turn, to

4

reuse it for boosting PDC production, we targeted on a membrane-bound transporter called ShiA

5

in E. coli, which reportedly could assimilate extracellular DHS into the cytosol of cells.34,15 In

6

the attempt to enhance DHS import and thus its intracellular availability for improved PDC

7

production, we tested E. coli native ShiA (EcShiA) as well as a number of ShiA proteins

8

originated from other different microbial species and showing various amino acid sequence

9

similarities relative to EcShiA (Supplementary Table 3). Genes encoding K. pneumonia ShiA

10

(KpShiA), Rhodococcus opacus PD630 ShiA (RoShiA) and Acinetobacter sp. ADP1 ShiA

11

(AsShiA) along with EcShiA, were inserted into plasmid pBBR1Gfbr to generate pBBR1Gfbr-

12

KpA, pBBR1Gfbr-RoA, pBBR1Gfbr-AsA and pBBR1Gfbr-EcA, respectively. Introduction of

13

these four plasmids individually into E. coli GYT1 strain harboring pTacFABC led to distinct

14

profiles of DHS formation and PDC production in flask cultures (Figure 5). Among these four

15

ShiA importers tested, AsShiA and EcShiA showed reduced DHS formation, whereas KpShiA

16

and RoShiA resulted in elevated DHS accumulation (Figure 5). Particularly, GYT1 strain

17

harboring pTacFABC and pBBR1Gfbr-EcA produced the least DHS (0.42 g/L), 83.3% lower than

18

that obtained by the parental strain GYT1 harboring pTacFABC and pBBR1Gfbr. However, such

19

a considerable decrease in DHS accumulation improved PDC production by only 5.7%, reaching

20

2.41 g/L. On the other hand, co-expression of AsShiA also lowered DHS formation (by 40.5%)

21

compared to the control strain without expressing AsShiA, probably thanks to its relatively high

22

amino acid sequence similarity to EcShiA (65.3%) (Supplementary Table 3), but the PDC titer

23

was also decreased (1.97 g/L) surprisingly. 12 ACS Paragon Plus Environment

Page 12 of 43

Page 13 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

Given that overexpression of EcShiA had the best effect on reducing byproduct DHS

2

formation and improving PDC production, we sought to further increase the expression level of

3

EcShiA, in order to further reduce extracellular DHS accumulation thereby increasing PDC

4

production. It was reported that although its chromosomal transcription is constitutive, E. coli

5

shiA gene is repressed under normal growth conditions via a small RNA-mediated translation

6

regulation mechanism, and the small RNA base-pairing interaction region is located upstream of

7

the coding sequence in the mRNA of shiA gene.35 As such, we determined to replace the native

8

promoter of E. coli shiA gene with the strong trc promoter. By doing so, not only could the

9

transcription of shiA gene be strengthened, but also its translation could be deregulated as the

10

small RNA binding region would be eliminated after the promoter exchange. GYT5 strain was

11

thus generated, which carried trc promoter-driven shiA gene in the chromosome. At first, we

12

examined effect of the promoter exchange of shiA alone on reducing DHS formation. As

13

expected, flask culture of GYT5 strain harboring pTacFABC and pBBR1Gfbr showed reduced

14

accumulation of DHS (1.86 g/L) by 26.2% compared to the parental strain GYT1 harboring

15

pTacFABC and pBBR1Gfbr (Figure 5). However, GYT5 strain harboring pTacFABC and

16

pBBR1Gfbr produced less PDC (2.18 g/L). Furthermore, combined effect of the plasmid-borne

17

and chromosomal trc promoter-driven overexpression of E. coli shiA was investigated. Flask

18

culture of GYT5 strain harboring pTacFABC and pBBR1Gfbr-EcA showed further reduced

19

accumulation of DHS (0.37 g/L) by 11.9% compared to GYT1 strain harboring pTacFABC and

20

pBBR1Gfbr-EcA (Figure 5). But similarly, only 2.16 g/L of PDC was produced, which was less

21

than that by GYT1 strain harboring pTacFABC and pBBR1Gfbr-EcA.

22

Taken together, we successfully reduced accumulation of the major byproduct DHS to a

23

minimum level through overexpression of DHS importer. And GYT1 strain harboring 13 ACS Paragon Plus Environment


1

pTacFABC and pBBR1Gfbr-EcA, which produced the highest PDC titer (2.41 g/L) along with

2

the second least DHS formation (0.42 g/L), was selected as base strain for further engineering.

3

Effect of Genetic Manipulations of the Soluble Transhydrogenase (SthA) on PDC

4

Production. In order to further optimize PDC production, it is necessary to consider the cofactor

5

utilization for the following reason. In the last step of PDC pathway (Figure 1), one molecule of

6

the cofactor NADP+ is converted to NADPH accompanying one molecule of PDC produced; this

7

might interfere with cellular redox cofactor balance. Thus, we targeted on E. coli soluble

8

pyridine nucleotide transhydrogenase (encoded by sthA) as SthA has been reported to be

9

responsible for reoxidation of NADPH to NADP+, particularly under metabolic conditions with

10

excess NADPH formation.36,37

11

To enhance SthA activity, we decided to overexpress sthA gene through two layers of

12

manipulations. First, we overexpressed it by promoter exchange with the strong trc promoter in

13

GYT1 strain, generating GYT6 strain. Second, based on GYT6, we additionally constructed

14

GYT7 strain, in which three silent point mutations (C15T, T18C, C21T) were introduced into

15

sthA gene in the chromosome, for the following reason. In E. coli, a small regulatory RNA called

16

Spot 42 (encoded by spf), highly abundant in the presence of glucose, was reported to repress

17

sthA expression by translational regulation through inhibitory base-pairing with the sthA mRNA,

18

and the binding site also localized within the beginning region of the coding sequence of sthA.38

19

Thus, we attempted to deregulate this repression mechanism by introducing the above three

20

silent point mutations that would affect the complementary base-pairing interaction as seen from

21

the binding free energy increase from -14.52 to -1.57 kcal/mol (Figure 6).


Page 14 of 43

Page 15 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

These two newly constructed strains GYT6 and GYT7 were examined in flask cultures

2

using GYT1 strain as a control. Among these three strains tested, distinct time profiles of cell

3

growth, PDC production and byproducts (i.e., PCA and DHS) formation were observed (Figure

4

7). Throughout the course of flask culture, GYT6 strain harboring pTacFABC and pBBR1Gfbr-

5

EcA produced less PDC than the control strain GYT1 harboring same plasmids (Figure 7B). In

6

the case of GYT7 strain harboring pTacFABC and pBBR1Gfbr-EcA, also less amount of PDC

7

was produced than the control as seen from the final titer (2.21 g/L) (Figure 7B). However, at

8

time point of 36 h, GYT7 strain harboring pTacFABC and pBBR1Gfbr-EcA produced higher titer

9

of PDC (2.05 g/L) versus that by the control (1.82 g/L) (Figure 7B), which was accomplished

10

apparently due to the higher maximum PDC productivity of GYT7 strain harboring pTacFABC

11

and pBBR1Gfbr-EcA (0.103 g/L/h) over that of the control (0.082 g/L/h). Moreover, it was

12

interestingly noticeable that GYT7 strain harboring pTacFABC and pBBR1Gfbr-EcA had higher

13

byproducts formation (Figure 7C and D), particularly DHS, than the control, which indicated a

14

stronger total flux redirected into the aromatic biosynthetic pathway. According to these results,

15

GYT7 strain harboring pTacFABC and pBBR1Gfbr-EcA, and GYT1 strain harboring pTacFABC

16

and pBBR1Gfbr-EcA, were selected and further evaluated in bioreactor fermentations as

17

discussed below.

18

Bioreactor Fermentations. Following development of engineered E. coli strains for

19

PDC production, bioreactor fermentations were conducted to demonstrate their potentials for

20

scale-up production.

21

Batch fermentations were first carried out to optimize cultivation conditions and select

22

the best-performing PDC production strain for subsequent fed-batch culture study. As oxygen is

23

required by PmdAB in PDC biosynthesis (Figure 1), we thus tested effects of different dissolved 15 ACS Paragon Plus Environment


1

oxygen (DO) levels (25%, 40% and 80% of air saturation) on PDC fermentation, using GYT1

2

strain harboring pTacFABC and pBBR1Gfbr-EcA. Batch fermentation profiles under the three

3

DO conditions are depicted (Supplementary Figure 3) and the results are summarized (Table 2).

4

It was observed that the highest PDC production (4.78 g/L) was obtained under DO level of 40%

5

(Supplementary Figure 3B and Table 2). When DO level was set as 25% or 80%, the PDC titer

6

was slightly decreased (Supplementary Figure 3A and 3C, Table 2).

7

Using the selected DO setting of 40% of air saturation, GYT7 strain harboring

8

pTacFABC and pBBR1Gfbr-EcA was examined in batch fermentation with all other conditions

9

same as for GYT1 strain harboring pTacFABC and pBBR1Gfbr-EcA. Unexpectedly, GYT7 strain

10

harboring pTacFABC and pBBR1Gfbr-EcA produced only 0.38 g/L of PDC with accumulations

11

of PCA and DHS as high as 4.74 g/L and 9.37 g/L, respectively (Supplementary Figure 4A and

12

Table 2). Out of curiosity, we further examined GYT6 strain harboring pTacFABC and

13

pBBR1Gfbr-EcA in batch fermentation under the same conditions, which produced 2.32 g/L of

14

PDC with formations of 1.86 g/L of PCA and 4.57 g/L of DHS (Supplementary Figure 4B and

15

Table 2).

16

Comparing the results in bioreactor batch fermentations with those in flask cultures for

17

GYT1, GYT6 and GYT7 strains harboring pTacFABC and pBBR1Gfbr-EcA, it was observed that

18

GYT1 and GYT6 strains harboring pTacFABC and pBBR1Gfbr-EcA had relatively consistent

19

performances in PDC production when they were transferred from flask cultivation to bioreactor

20

cultivation. By contrast, GYT7 strain harboring pTacFABC and pBBR1Gfbr-EcA exhibited a

21

dramatic decay in PDC production capability when moved from flask culture to bioreactor

22

culture. Such a huge change in PDC production performance suggested an intriguing insight into

23

the important role of cellular redox cofactor in affecting PDC production and the flux towards 16 ACS Paragon Plus Environment

Page 16 of 43

Page 17 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

DHS in the context of different cultivation conditions. Yet, further detailed investigation is

2

needed to clarify the underlying mechanisms that have resulted in such outcomes. Although the

3

three silent point mutations of sthA gene in GYT7 strain failed to increase PDC production in

4

this study, this genetic modification strategy might potentially serve as a useful approach to

5

perturb the SthA activity in E. coli when needed.

6

Based on batch fermentation results above, GYT1 strain harboring pTacFABC and

7

pBBR1Gfbr-EcA was selected as the final strain and examined in a fed-batch culture under DO

8

level of 40% of air saturation and with a pH-stat nutrient feeding strategy. The fed-batch

9

fermentation resulted in production of 16.72 g/L of PDC with the productivity of 0.172 g/L/h and

10

the yield of 0.201 g/g glucose (Figure 8). At the same time, 9.76 g/L of DHS was accumulated,

11

while only 0.36 g/L of PCA was formed. It is speculated that the high accumulation of DHS

12

during fed-batch culture might be caused by relatively limited activity of DHS dehydratase

13

and/or decay of DHS importer activity over the course of fermentation. This suggests potential

14

room for further enhancement of PDC production level. Thus, future studies should be focused

15

on improving the activity of DHS dehydratase and/or maintaining of DHS importer activity

16

during large-scale fermentation processes.

17

18

CONCLUSIONS

19

In this paper, we developed metabolically engineered E. coli platform strains capable of

20

producing PDC using glucose as a carbon source. An efficient metabolic pathway towards PDC

21

was suggested on the basis of in silico flux simulation, and the selected PDC pathway was

22

reconstructed through introduction of screened, efficient heterologous enzymes. Removal of 17 ACS Paragon Plus Environment


1

feedback inhibition of a key enzyme, elevated precursor supply and overexpression of DHS

2

importer were important in improving PDC production and reducing byproduct formation.

3

Furthermore, manipulation of the cofactor (NADPH/NADP+) utilization through perturbing

4

small regulatory RNA-based soluble pyridine nucleotide transhydrogenase gene expression

5

provided insight into the impact of redox cofactor balancing on PDC production. Finally, fed-

6

batch culture of the best-performing engineered strain produced 16.72 g/L of PDC with an

7

overall productivity of 0.172 g/L/h and yield of 0.201 g/g glucose, which represented the highest

8

titer for PDC production from glucose reported to date. Further improvement in PDC production

9

would be possible by taking advantage of various systems metabolic engineering tools,39 such as

10

systematic flux optimization, synthetic regulatory sRNA-mediated large-scale gene target

11

identification coupled with high-throughput techniques,40 as well as bioprocess optimization.

12

13

METHODS

14

Bacterial Strains and Media. The bacterial strains used in this study are listed in Table

15

1. E. coli DH5α was employed for gene cloning and plasmid propagation, and E. coli W3110 and

16

its derivatives were used for PDC production as host strains. For general cultures during plasmid

17

construction and bacterial genome manipulation, E. coli cells were cultured in Luria-Bertani (LB)

18

broth or on LB plates (1.5%, w/v, agar) with proper antibiotics supplemented as follows: 100

19

µg/mL of ampicillin, 50 µg/mL of kanamycin, and/or 34 µg/mL of chloramphenical. Bacterial

20

species Comamonas testosteroni ATCC 11996, Bacillus thuringiensis ATCC 10792, Klebsiella

21

pneumonia KCTC 2208, Corynebacterium glutamicum ATCC 13032, Rhodococcus opacus

22

PD630 and Acinetobacter sp. ADP1 served as genetic sources for heterologous enzymes or


Page 18 of 43

Page 19 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

transporters employed in this study. Cultivation of these wild-type strains was carried out under

2

the conditions specified by the corresponding suppliers.

3

Plasmid Construction. All plasmids used in this study are listed in Table 1. The basic

4

molecular biology experiments including PCR, gel electrophoresis, and bacterial transformation

5

for plasmid and strain construction were conducted according to standard procedures.41

6

Commercial kits were used for preparation of bacterial genomic DNA, isolation of plasmid DNA

7

and purification of DNA fragments from agarose gels following the supplier’s instructions

8

(Genotech, Daejeon, South Korea). Services of primer synthesis and DNA sequencing were also

9

provided by Genotech. All the oligonucleotide primers utilized in this work are listed in

10

Supplementary Table 4. To construct pETABC, the pmdABC operon from the genome of C.

11

testosteroni ATCC 11996 was amplified using pmdABC(pET)-f and pmdABC(pET)-r primers,

12

and cloned into pET-22b(+) vector amplified using pET(pmdABC)-f and pET(pmdABC)-r

13

primers by Gibson assembly method.42 To construct pTacABC, the pmdABC operon was

14

amplified using pmdABC(pTac)-f and pmdABC(pTac)-r primers, and cloned into pTac15K

15

vector amplified using pTac(pmdABC)-f and pTac(pmdABC)-r primers by Gibson assembly. To

16

construct pTrcZ, the aroZ gene from K. pneumonia KCTC 2208 was amplified using

17

aroZ(pTrc)-f and aroZ(pTrc)-r primers, and cloned into pTrc99A vector amplified using

18

pTrc(aroZ)-f and pTrc(aroZ)-r primers by Gibson assembly. To construct pTrcB, the qsuB gene

19

from C. glutamicum ATCC 13032 was amplified using qsuB-f and qsuB-r primers, and cloned

20

into pTrc99A vector at SacI and XbaI sites. To construct pTrcF, the asbF gene from B.

21

thuringiensis ATCC 10792 was amplified using asbF-f and asbF-r primers, and cloned into

22

pTrc99A vector at EcoRI and PstI sites. To construct pTrcFopt, an E. coli codon-optimized

23

version of asbF gene (synthesized by GenScript) was amplified using asbFopt-f and asbFopt-r



1

primers, and cloned into pTrc99A vector at EcoRI and PstI sites. To construct pTacFABC, the

2

trc-asbF-rrnBT1T2 cassette was amplified from pTrcF using trc-asbF-f and trc-asbF-r primers,

3

and assembled by Gibson assembly with pTacABC linearized by enzyme digest at NheI site. To

4

construct pTacFoptABC, the trc-asbFopt-rrnBT1T2 cassette was amplified from pTrcFopt using

5

trc-asbF-f and trc-asbF-r primers, and assembled by Gibson assembly with pTacABC linearized

6

by enzyme digest at NheI site. To construct pBBR1Gfbr, the feedback inhibition-resistant version

7

of the E. coli aroG gene was amplified using aroGfbr(pBBR1)-f and aroGfbr(pBBR1)-f primers,

8

and cloned into pBBR1MCS vector amplified using pBBR1(aroGfbr)-f and pBBR1(aroGfbr)-r

9

primers by Gibson assembly. To construct pBBR1Gfbr-A, the RBS-tktA cassette was amplified

10

using RBS-tktA-f and RBS-tktA-r primers, and assembled by Gibson assembly with pBBR1Gfbr

11

linearized by enzyme digest at BamHI site. To construct pBBR1Gfbr-EcA, the RBS-shiA cassette

12

was amplified from E. coli using RBS-EcA-f and EcA-r primers, and cloned into pBBR1Gfbr at

13

HindIII and BamHI sites. To construct pBBR1Gfbr-KpA, the RBS-shiA cassette was amplified

14

from K. pneumonia KCTC 2208 using RBS-KpA-f and RBS-KpA-r primers, and cloned into

15

pBBR1Gfbr at HindIII and SpeI sites. To construct pBBR1Gfbr-RoA, the RBS-shiA cassette was

16

amplified from R. opacus PD630 using RBS-RoA-f and RBS-RoA-r primers, and cloned into

17

pBBR1Gfbr at HindIII and BamHI sites. To construct pBBR1Gfbr-AsA, the RBS-shiA cassette

18

was amplified from A. sp. ADP1 using RBS-AsA-f and RBS-AsA-r primers, and cloned into

19

pBBR1Gfbr at HindIII and BamHI sites. All the recombinant plasmids were confirmed by colony

20

PCR and DNA sequencing.

21

Genome Manipulation. One-step homologous recombination-based inactivation

22

method43 was used for gene deletion and promoter replacement experiments. For deleting aroE,

23

the first knockout PCR product was amplified using aroE-KO-f and aroE-KO-r primers having


Page 20 of 43

Page 21 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

50 bp nucleotide extension homologous to the upstream and downstream regions of aroE within

2

the genome. The plasmid pECmulox bearing a lox71-chloramphenical resistance gene (cat)-

3

lox66 cassette was used as a template for PCR amplifying the first knockout fragment. The

4

homology sequence was elongated to 100 bp via a second PCR using the first PCR product as

5

template and using aroE-KOEX-f and aroE-KOEX-r primers. For the promoter change of ppsA,

6

the pMtrc9 plasmid containing the strong trc promoter and the upstream lox71-cat-lox66 cassette

7

was employed as a template in the PCR reactions using primer pairs of Ptrc(ppsA)-f/ Ptrc(ppsA)-

8

r and Ptrc(ppsA)-EX-f/ Ptrc(ppsA)-EX-r, instead of pECmulox. PCR products for deletion or

9

promoter replacement of the remaining genes were prepared in the same manner using primers

10

listed in Supplementary Table 4. For construction of the GYT7 strain, the three point mutations

11

(C15T, T18C, C21T) were introduced into the promoter replacement primers. Regardless of gene

12

deletion or promoter replacement, the amplified knockout PCR products were introduced into E.

13

coli cells expressing the Red recombinase from pKD46 induced by 10 mM L-arabinose.

14

Colonies were selected on LB agar plates supplemented with chloramphenical, and the

15

successful genome manipulation was confirmed by colony PCR. Subsequently, the helper

16

plasmid pJW168 that expressed the Cre recombinase by 1 mM IPTG was introduced to remove

17

the chloramphenical marker gene.44 With the temperature-sensitive replication origins, both

18

pKD46 and pJW168 were easily cured through temperature shift between 30 °C and 42 °C. The

19

excision of the marker gene was further confirmed by performing colony PCR. For promoter

20

replacement, the positive clones were further sequenced to validate the correct manipulation.

21

In Silico Flux Simulation Experiments. To perform in silico flux response analysis for

22

comparing two biosynthetic pathway routes (single-step and six-step pathways) for PDC, we

23

employed the genome-scale metabolic model EcoMBEL979,45 which is a modified model of



1

iJR9043046 and comprises 979 metabolic reactions and 814 metabolites. As PDC biosynthetic

2

pathways are not native to E. coli, the heterologous pathway reactions are required to be

3

additionally recruited to the EcoMBEL979 model. To obtain PCA, the following reaction “DHS

4

↔ PCA + H2O” was added to the EcoMBEL979 model for the single-step biosynthetic pathway,

5

whereas the following reaction “4HBA + NADPH + O2 + H+ ↔ PCA + NADP+ + H2O” was

6

added for the six-step biosynthetic pathway. From PCA to PDC, the subsequent reactions were

7

introduced into the EcoMBEL979 model for both pathways: “PCA + O2 ↔ CHMS + H+”,

8

“CHMS ↔ CHMS (hemiacetal form)”, and “CHMS (hemiacetal form) + NADP+ ↔ PDC +

9

NADPH + H+”. Next, we investigated the theoretical impacts of the two different biosynthetic

10

pathways (single-step and six-step) on PDC production, through constraint-based flux simulation

11

with the hypothesis of pseudo steady-state. Cell growth rate was maximized as an objective

12

function as the PDC biosynthetic flux was gradually increased from minimum to maximum

13

values. Over the course of simulation, the glucose uptake rate was set at 10 mM/gDCW/h.

14

Cultivation. Shake-flask cultivations were performed in the baffled flasks (300 mL),

15

which contained a working volume of 50 mL MR minimal salts medium (pH 7.0) supplemented

16

with 10 g/L glucose. The MR medium (1L) contained: 6.67 g KH2PO4, 4 g (NH4)2HPO4, 0.8 g

17

MgSO4·7H2O, 0.8 g citric acid, and 5 mL trace metal solution.47 Glucose, MgSO4·7H2O, MR

18

salts and trace metal solution were prepared and autoclaved separately. To deal with the

19

auxotroph caused by the disruption of aroE gene, the MR medium (1 L) was also supplemented

20

with L-phenylalanine (40 mg), L-tyrosine (40 mg), L-tryptophan (40 mg), 4-hydroxybenzoic

21

acid (10 mg), 4-aminobenzoic acid (10 mg), 2,3-dihydroxybenzoic acid (10 mg) and thiamine

22

hydrochloride (10 mg),13 stock solutions of which were separately sterilized through 0.22 µm

23

membrane. To prepare the inoculums for flask cultivation, glycerol stocks of the engineered


Page 22 of 43

Page 23 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

strains were first inoculated into a 25 mL test tube containing 5 mL of LB medium and cultured

2

at 200 rpm and 37 °C for 12 h. Aliquots of 1 mL of the preculture were used for inoculation.

3

After inoculation, flasks were placed in a rotary incubator at 200 rpm and 37 °C for initial

4

growth. After 6 h of cultivation when cells grew up to the OD600 value of 0.6~0.8, 1 mM IPTG

5

was added for induction, and then cell cultures were immediately transferred to a rotary

6

incubator at 200 rpm and 30 °C for additional 54 h of cultivation. Appropriate antibiotics were

7

added to the medium when necessary. All flask cultivations described in this study were

8

conducted in biological duplicates, and the results were depicted as mean values plus standard

9

deviations in the graph.

10

Batch reactor fermentations were conducted using a 6.6-L jar fermentor (Bioflo 3000;

11

New Brunswick Scientific Co., Edison, NJ) containing 1.8 L of the MR medium supplemented

12

with 20 g/L of glucose, 3 g/L yeast extract, 40 mg/L of L-phenylalanine, 40 mg/L of L-tyrosine,

13

40 mg/L of L-tryptophan and 10 mg/L of 4-hydroxybenzoic acid under 30 °C. To prepare seed

14

cultures, 2 mL of overnight LB test-tube cultures were transferred into the Erlenmeyer non-

15

baffled flasks (300 mL) containing 100 mL of the same medium and cultivated in a shaking

16

incubator for 10 h at 200 rpm and 37 °C. A total volume of 200 mL of seed cultures were used to

17

inoculate the fermentor to make a total working volume of 2 L. The fermentation culture pH was

18

maintained at 7.0 with the addition of ammonia solution (28%, v/v). The dissolved oxygen (DO)

19

was maintained at 25%, 40% or 80% of air saturation as indicated by automatically changing the

20

agitation speed from 200 to 1000 rpm and additional supply of pure oxygen when necessary with

21

a constant air flow of 2.0 L/min. Cells were induced with 1 mM IPTG at the OD600 value of

22

2.0~3.0. All batch fermentations with each of engineered strains were performed twice with

23

reproducible results, and the results of one representative batch fermentation are presented.



1

Fed-batch fermentation was performed under the same settings except the followings.

2

The initial fermentation medium contained 1.8 L of the MR medium supplemented with 20 g/L

3

of glucose, 3 g/L yeast extract, 100 mg/L of L-phenylalanine, 100 mg/L of L-tyrosine, 100 mg/L

4

of L-tryptophan and 25 mg/L of 4-hydroxybenzoic acid. Cells were induced with 1 mM IPTG at

5

the OD600 value of 4.0~5.0. When glucose was depleted at the end of batch culture, fed-batch

6

mode was initiated by a pH-stat nutrient feeding strategy. The pH-stat loop was whenever the

7

medium pH >= 7.02, the nutrient solution was pumped into the fermentor at the rate of 100%.

8

The nutrient solution comprised 700 g/L of glucose, 8 g/L of MgSO4·7H2O, 5 mL trace metal

9

solution, 3 g/L of yeast extract, 100 mg/L of L-phenylalanine, 100 mg/L of L-tyrosine, 100 mg/L

10

of L-tryptophan, 25 mg/L of 4-hydroxybenzoic acid, 1 mM IPTG and appropriate antibiotics.

11

Foam formation was repressed by manually adding Antifoam 289 (Sigma Chemical Co., St.

12

Louis, MO, USA). Fed-batch fermentation was performed three times, and the result of one

13

representative fed-batch culture is presented.

14

Analytical Procedures. Cell growth was monitored by measuring the optical density at

15

the wavelength of 600 nm (OD600) using the Ultrospec 3100 spectrophotometer (Amersham

16

Biosciences, Uppsala, Sweden). The residual glucose concentration in culture broth was

17

measured using high-performance liquid chromatography (HPLC) (Waters 1515/2414/2707,

18

Waters, Milford, MA). The concentrations of PDC, DHS and PCA were analyzed according to a

19

modified method,48 which employed a HPLC (1100 Series, Agilent) equipped with a Zorbax SB-

20

Aq column (4.6 x 250 mm, Agilent) operating at 30 °C. The mobile phase consisting of buffer A

21

(25 mM potassium phosphate buffer, pH 2.0) and buffer B (acetonitrile), flowed at 0.8 mL/min

22

according to the following program: 0-1 min, 0% B; 1-8 min, a linear gradient of B from 0% to

23

70%; 8-11 min, 70% B; 11-18 min, a linear gradient of B from 70% to 0%; 18-20 min, 0% B.


Page 24 of 43

Page 25 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

The injection volume of 10 µL was used and photodiode array detector was used to monitor the

2

signal at 280 nm for PDC, DHS and PCA. Samples for HPLC analysis were prepared as follows:

3

cells and debris in the fermentation broth were first pelleted by centrifugation at 13,200 g for 10

4

min, and then the supernatant was diluted properly and filtered through a 0.22 µm membrane.

5

The quantification was made according to a standard calibration curve established with each

6

authentic compound. To further confirm the production of PDC, culture sample was centrifuged

7

to remove cells and the supernatant was acidified to pH 1.0 by concentrated HCl. An aliquot (1

8

mL) of the resulting solution was extracted twice with the same volume of ethyl acetate. The

9

collected extract was vacuum-dried on a rotary evaporator. The residue was dissolved in pyridine

10

and incubated with bistrimethylsilyl-trifluoroacetamide to prepare trimethylsilyl derivatives, and

11

then subjected to gas chromatography (GC)-mass spectrometry as described previously.9

12 13

ASSOCIATED CONTENT

14

Supporting Information

15

The Supporting Information is available free of charge on the ACS Publications website at DOI:

16

17

AUTHOR INFORMATION

18

Corresponding Author

19

*Tel: +82-42-350-3930. Fax: +82-42-350-3910. Email: [email protected]

20

Author Contributions



1

Z.W.L. and S.Y.L. designed the research; Z.W.L. performed the experiments; W.J.K. performed

2

in silico simulation experiments; and Z.W.L. and S.Y.L. wrote the manuscript.

3

Notes

4

The authors declare no competing financial interest.

5

6

ACKNOWLEDGEMENTS

7

We thank Dr. Masaya Nakamura at Forestry and Forest Products Research Institute (FFPRI) in

8

Japan for generously providing us the authentic PDC compound used as an analytical standard in

9

this study. This work was supported by the Intelligent Synthetic Biology Center through the

10

Global Frontier Project (2011-0031963) and also by the Technology Development Program to

11

Solve Climate Changes on Systems Metabolic Engineering for Biorefineries (NRF-

12

2012M1A2A2026556 and NRF-2012M1A2A2026557) from the Ministry of Science and ICT

13

through the National Research Foundation of Korea.

14 15

REFERENCES

16 17 18 19 20 21 22 23 24 25 26

(1) Masai, E., Katayama, Y., and Fukuda, M. (2007) Genetic and biochemical investigations on bacterial catabolic pathways for lignin-derived aromatic compounds. Biosci., Biotechnol., Biochem. 71, 1-15. (2) Shigehara, K., Katayama, Y., Nishikawa, S., and Hotta, Y. (2001) Polyester and process for producing the same. US Patent 6303745B1. (3) Michinobu, T., Hishida, M., Sato, M., Katayama, Y., Masai, E., Nakamura, M., Otsuka, Y., Ohara, S., and Shigehara, K. (2008) Polyesters of 2-pyrone-4,6-dicarboxylic acid obtained from a metabolic intermediate of lignin. Polym. J. 40, 68-75. (4) Michinobu, T., Bito, M., Tanimura, M., Katayama, Y., Masai, E., Nakamura, M., Otsuka, Y., Ohara, S., and Shigehara, K. (2010) Synthesis and characterization of hybrid biopolymers of L-lactic acid and 2-pyrone-4,6-dicarboxylic Acid. J. Macromol. Sci. A. 47, 564-70.


Page 26 of 43

Page 27 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46


(5) Choi, S., Song, C. W., Shin, J. H., and Lee, S. Y. (2015) Biorefineries for the production of top building block chemicals and their derivatives. Metab. Eng. 28, 223-39. (6) Park, S. H., Kim, H. U., Kim, T. Y., Park, J. S., Kim, S.-S., and Lee, S. Y. (2014) Metabolic engineering of Corynebacterium glutamicum for L-arginine production. Nat. Commun. 5, 4618. (7) Rodriguez, A., Martnez, J. A., Flores, N., Escalante, A., Gosset, G., and Bolivar, F. (2014) Engineering Escherichia coli to overproduce aromatic amino acids and derived compounds. Microb. Cell Fact. 13, 126. (8) Santos, C. N. S., Xiao, W., and Stephanopoulos, G. (2012) Rational, combinatorial, and genomic approaches for engineering L-tyrosine production in Escherichia coli. Proc. Natl. Acad. Sci. U. S. A. 109, 13538-43. (9) Otsuka, Y., Nakamura, M., Shigehara, K., Sugimura, K., Masai, E., Ohara, S., and Katayama, Y. (2006) Efficient production of 2-pyrone 4,6-dicarboxylic acid as a novel polymerbased material from protocatechuate by microbial function. Appl. Microbiol. Biotechnol. 71, 608-14. (10) Nakajima, M., Nishino, Y., Tamura, M., Mase, K., Masai, E., Otsuka, Y., Nakamura, M., Sato, K., Fukuda, M., Shigehara, K., Ohara, S., Katayama, Y., and Kajita, S. (2009) Microbial conversion of glucose to a novel chemical building block, 2-pyrone-4,6-dicarboxylic acid. Metab. Eng. 11, 213-20. (11) Jones, J. A., Toparlak, Ö. D., and Koffas, M. A. G. (2015) Metabolic pathway balancing and its role in the production of biofuels and chemicals. Curr. Opin. Biotechnol. 33, 52-9. (12) Blazeck, J., and Alper, H. S. (2013) Promoter engineering: recent advances in controlling transcription at the most fundamental level. Biotechnol. J. 8, 46-58. (13) Niu, W., Draths, K. M., and Frost, J. W. (2002) Benzene-free synthesis of adipic acid. Biotechnol. Prog. 18, 201-11. (14) Weber, C., Brückner, C., Weinreb, S., Lehr, C., Essl, C., and Boles, E. (2012) Biosynthesis of cis,cis-muconic acid and its aromatic precursors, catechol and protocatechuic acid, from renewable feedstocks by Saccharomyces cerevisiae. Appl. Environ. Microbiol. 78, 8421-30. (15) Zhang, H., Pereira, B., Li, Z., and Stephanopoulos, G. (2015) Engineering Escherichia coli coculture systems for the production of biochemical products. Proc. Natl. Acad. Sci. U. S. A. 112, 8266-71. (16) Curran, K. A., Leavitt, J. M., Karim, A. S., and Alper, H. S. (2013) Metabolic engineering of muconic acid production in Saccharomyces cerevisiae. Metab. Eng. 15, 55-66. (17) Sengupta, S., Jonnalagadda, S., Goonewardena, L., and Juturu, V. (2015) Metabolic engineering of a novel muconic acid biosynthesis pathway via 4-hydroxybenzoic acid in Escherichia coli. Appl. Environ. Microbiol. 81, 8037-43. (18) Chen, Z., Shen, X., Wang, J., Wang, J., Zhang, R., Rey, J. F., Yuan, Q., and Yan, Y. (2017) Establishing an Artificial Pathway for De Novo Biosynthesis of Vanillyl Alcohol in Escherichia coli. ACS Synth. Biol. 6, 1784-92. (19) Elsemore, D. A., and Ornston, L. N. (1995) Unusual ancestry of dehydratases associated with quinate catabolism in Acinetobacter calcoaceticus. J. Bacteriol. 177, 5971-8. (20) Jiménez, J. I., Miñambres, B., García, J. L., and Díaz, E. (2002) Genomic analysis of the aromatic catabolic pathways from Pseudomonas putida KT2440. Environ. Microbiol. 4, 824-41. (21) Lamb, H. K., Hawkins Ar Fau - Smith, M., Smith M Fau - Harvey, I. J., Harvey Ij Fau Brown, J., Brown J Fau - Turner, G., Turner G Fau - Roberts, C. F., and Roberts, C. F. (1990) 27 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44

Spatial and biological characterisation of the complete quinic acid utilisation gene cluster in Aspergillus nidulans. Mol. Gen. Genet. 223, 17-23. (22) Fox, D. T., Hotta, K., Kim, C.-Y., and Koppisch, A. T. (2008) The missing link in petrobactin biosynthesis: asbF encodes a (−)-3-dehydroshikimate dehydratase. Biochemistry 47, 12251-3. (23) Peek, J., Roman, J., Moran, G. R., and Christendat, D. (2017) Structurally diverse dehydroshikimate dehydratase variants participate in microbial quinate catabolism. Mol. Microbiol. 103, 39-54. (24) Teramoto, H., Inui, M., and Yukawa, H. (2009) Regulation of expression of genes involved in quinate and shikimate utilization in Corynebacterium glutamicum. Appl. Environ. Microbiol. 75, 3461-8. (25) Eudes, A., Sathitsuksanoh, N., Baidoo, E. E. K., George, A., Liang, Y., Yang, F., Singh, S., Keasling, J. D., Simmons, B. A., and Loqué, D. (2015) Expression of a bacterial 3dehydroshikimate dehydratase reduces lignin content and improves biomass saccharification efficiency. Plant Biotechnol. J. 13, 1241-50. (26) Mampel, J., Providenti, M. A., and Cook, A. M. (2005) Protocatechuate 4,5-dioxygenase from Comamonas testosteroni T-2: biochemical and molecular properties of a new subgroup within class III of extradiol dioxygenases. Arch. Microbiol. 183, 130-9. (27) Barry, K. P., and Taylor, E. A. (2013) Characterizing the promiscuity of LigAB, a lignin catabolite degrading extradiol dioxygenase from Sphingomonas paucimobilis SYK-6. Biochemistry 52, 6724-36. (28) Sprenger, G. A. (2007) From scratch to value: engineering Escherichia coli wild type cells to the production of L-phenylalanine and other fine chemicals derived from chorismate. Appl. Microbiol. Biotechnol. 75, 739-49. (29) Báez-Viveros, J. L., Osuna, J., Hernández-Chávez, G., Soberón, X., Bolívar, F., and Gosset, G. (2004) Metabolic engineering and protein directed evolution increase the yield of Lphenylalanine synthesized from glucose in Escherichia coli. Biotechnol. Bioeng. 87, 516-24. (30) Liu, S.-P., Xiao, M.-R., Zhang, L., Xu, J., Ding, Z.-Y., Gu, Z.-H., and Shi, G.-Y. (2013) Production of L-phenylalanine from glucose by metabolic engineering of wild type Escherichia coli W3110. Process Biochem. 48, 413-9. (31) Shen, T., Liu, Q., Xie, X., Xu, Q., and Chen, N. (2012) Improved production of tryptophan in genetically engineered Escherichia coli with TktA and PpsA overexpression. J. Biomed. Biotechnol. 2012, 8. (32) Weiner, M., Albermann, C., Gottlieb, K., Sprenger, G. A., and Weuster-Botz, D. (2014) Fed-batch production of L-phenylalanine from glycerol and ammonia with recombinant Escherichia coli. Biochem. Eng. J. 83, 62-9. (33) Yao, Y.-F., Wang, C.-S., Qiao, J., and Zhao, G.-R. (2013) Metabolic engineering of Escherichia coli for production of salvianic acid A via an artificial biosynthetic pathway. Metab. Eng. 19, 79-87. (34) Whipp, M. J., Camakaris, H., and Pittard, A. J. (1998) Cloning and analysis of the shiA gene, which encodes the shikimate transport system of Escherichia coli K-12. Gene 209, 185-92. (35) Prévost, K., Salvail, H., Desnoyers, G., Jacques, J.-F., Phaneuf, É., and Massé, E. (2007) The small RNA RyhB activates the translation of shiA mRNA encoding a permease of shikimate, a compound involved in siderophore synthesis. Mol. Microbiol. 64, 1260-73.


Page 28 of 43

Page 29 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41


(36) Sauer, U., Canonaco, F., Heri, S., Perrenoud, A., and Fischer, E. (2004) The soluble and membrane-bound transhydrogenases UdhA and PntAB have divergent functions in NADPH metabolism of Escherichia coli. J. Biol. Chem. 279, 6613-9. (37) Canonaco, F., Hess, T. A., Heri, S., Wang, T., Szyperski, T., and Sauer, U. (2001) Metabolic flux response to phosphoglucose isomerase knock-out in Escherichia coli and impact of overexpression of the soluble transhydrogenase UdhA. FEMS Microbiol. Lett. 204, 247-52. (38) Beisel, C. L., and Storz, G. (2011) The base-pairing RNA Spot 42 participates in a multioutput feedforward loop to help enact catabolite repression in Escherichia coli. Mol. Cell 41, 286-97. (39) Lee, S. Y., and Kim, H. U. (2015) Systems strategies for developing industrial microbial strains. Nat. Biotechnol. 33, 1061. (40) Na, D., Yoo, S. M., Chung, H., Park, H., Park, J. H., and Lee, S. Y. (2013) Metabolic engineering of Escherichia coli using synthetic small regulatory RNAs. Nat. Biotechnol. 31, 1704. (41) Sambrook, J., and Russell, D. W. (2001) Molecular Cloning: A Laboratory Manual, 3rd ed., Cold Spring Harbor Lab Press. (42) Gibson, D. G., Young, L., Chuang, R.-Y., Venter, J. C., Hutchison Iii, C. A., and Smith, H. O. (2009) Enzymatic assembly of DNA molecules up to several hundred kilobases. Nat. Methods 6, 343. (43) Datsenko, K. A., and Wanner, B. L. (2000) One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci. U. S. A. 97, 6640-5. (44) Palmeros, B. z., Wild, J., Szybalski, W., Le Borgne, S., Hernández-Chávez, G., Gosset, G., Valle, F., and Bolivar, F. (2000) A family of removable cassettes designed to obtain antibiotic-resistance-free genomic modifications of Escherichia coli and other bacteria. Gene 247, 255-64. (45) Lee, S. Y., Woo, H. M., Lee, D.-Y., Choi, H. S., Kim, T. Y., and Yun, H. (2005) Systems-level analysis of genome-scalein silico metabolic models using MetaFluxNet. Biotechnol. Bioprocess Eng. 10, 425. (46) Reed, J. L., Vo, T. D., Schilling, C. H., and Palsson, B. O. (2003) An expanded genomescale model of Escherichia coli K-12 (iJR904 GSM/GPR). Genome Biol. 4, R54. (47) Lee, Y., and Lee, S. (1996) Enhanced production of poly(3-hydroxybutyrate) by filamentation-suppressed recombinant Escherichia coli in a defined medium. J. Environ. Polymer Degradation 4, 131-4. (48) Luo, Z. W., and Lee, S. Y. (2017) Biotransformation of p-xylene into terephthalic acid by engineered Escherichia coli. Nat. Commun. 8, 15689. (49) Elzer, P. H., Kovach, M. E., Phillips, R. W., Robertson, G. T., Peterson, K. M., and Roop, R. M. (1995) In vivo and in vitro stability of the broad-host-range cloning vector pBBR1MCS in six Brucella species. Plasmid 33, 51-7. (50) Kim, J. M., Lee, K. H., and Lee, S. Y. (2008) Development of a markerless gene knockout system for Mannheimia succiniciproducens using a temperature-sensitive plasmid. FEMS Microbiol. Lett. 278, 78-85.

42

43



Page 30 of 43

1

FIGURE LEGENDS

2

Figure 1. Biosynthetic pathway of PDC from glucose and the metabolic engineering strategies

3

used in this study. Abbreviations: PTS, phosphotransferase system; TCA, tricarboxylic acid; PP,

4

pentose phosphate; G6P, glucose 6-phosphate; G3P, glyceraldehydes 3-phosphate; PEP,

5

phosphoenolpyruvate; PYR, pyruvate; OAA, oxaloacetate; E4P, erythrose 4-phosphate; DAHP,

6

3-deoxy-D-arabinoheptulosonate

7

dehydroshikimate; SHK, shikimate; S3P, shikimate-3-phosphate; EPSP, 5-enolpyruvyl-

8

shikimate 3-phosphate; CHA, chorismate; 4HBA, 4-hydroxybenzoate; PCA, protocatechuate;

9

CHMS, 4-carboxy-2-hydroxymuconate-6-semialdehyde; PDC, 2-pyrone-4,6-dicarboxylic acid.

10

Genes that encode enzymes: ppsA, PEP synthetase; pykF, pyruvate kinase I; pykA, pyruvate

11

kinase II; pckA, PEP carboxykinase; ppc, PEP carboxylase; tktA, transketolase I; aroGfbr,

12

feedback-inhibition resistant mutant of DAHP synthase; aroB, DHQ synthase; aroD, DHQ

13

dehydratase; aroE, SHK dehydrogenase; aroK, SHK kinase I; aroL, SHK kinase II; aroA, 3-

14

phosphoshikimate-1-carboxyvinyltransferase; aroC, CHA synthase; ubiC, CHA lyase; asbF,

15

dehydroshikimate dehydratase; pmdAB, PCA 4,5-dioxygenase; pmdC, CHMS dehydrogenase;

16

shiA, SHK:H+ symporter. Overexpressed genes are marked in blue, and red fork indicates gene

17

deletion. SHK:H+ symporter, two recombinant plasmids and genome manipulations for promoter

18

replacement employed in this study are illustrated.

19

Figure 2. Comparison of two distinct pathways for PDC biosynthesis from glucose via in silico

20

flux analysis. (A) Schematic for the two distinct PDC pathways: single-step and six-step, as well

21

as the six-step pathway with cofactors intentionally removed. Cofactors (NADPH and ATP) are

22

indicated. Abbreviations: DHS, 3-dehydroshikimate; SHK, shikimate; S3P, shikimate-3-

23

phosphate; EPSP, 5-enolpyruvyl-shikimate 3-phosphate; CHA, chorismate; 4HBA, 4-

7-phosphate;

DHQ,


3-dehydroquinate;

DHS,

3-

Page 31 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

hydroxybenzoate; PCA, protocatechuate; PDC, 2-pyrone-4,6-dicarboxylic acid. Black arrows

2

indicate E. coli native metabolic reactions, red arrows indicate heterologous enzymatic reactions,

3

and dotted arrows indicate multiple steps of enzymatic reactions. (B) Performance of the

4

different PDC pathways during in silico simulation experiments. The orange line corresponds to

5

the result of the single-step pathway, the magenta line corresponds to the result of the six-step

6

pathway, and the blue line corresponds to the result of the six-step pathway without cofactors

7

requirement.

8

Figure 3. Flask cultures for PCA production from glucose by introducing different DHS

9

dehydratases (A) and for PDC production from glucose using the selected DHS dehydratase (B).

10

Symbols are: gray box, cell growth (OD600); blue box, PCA concentration (g/L); red box, PDC

11

concentration (g/L).

12

Figure 4. Effects of overexpression of the aroGfbr, tktA and ppsA genes, and deletion of the pykF

13

and/or pykA genes on PDC production. Symbols are: gray box, cell growth (OD600); red box,

14

PDC concentration (g/L).

15

Figure 5. Effect of overexpression of DHS importers from different microbial origins on PDC

16

production. Engineered strains indicated are: Control, GYT1 harboring pTacFABC and

17

pBBR1Gfbr; KpShiA, GYT1 harboring pTacFABC and pBBR1Gfbr-KpA; RoShiA, GYT1

18

harboring pTacFABC and pBBR1Gfbr-RoA; AsShiA, GYT1 harboring pTacFABC and

19

pBBR1Gfbr-AsA; EcShiA, GYT1 harboring pTacFABC and pBBR1Gfbr-EcA; PtrcShiA, GYT5

20

harboring pTacFABC and pBBR1Gfbr; EcShiA+PtrcShiA, GYT5 harboring pTacFABC and

21

pBBR1Gfbr-EcA. Symbols are: gray box, cell growth (OD600); red box, PDC concentration (g/L);

22

yellow box, DHS concentration (g/L).



1

Figure 6. Schematic representation of the chromosomal manipulations of the sthA gene: (1)

2

promoter replacement with the strong trc promoter and (2) further introduction of three silent

3

point mutations (C15T, T18C, C21T) in the coding sequence of sthA, which leads to weakened

4

base-pairing interaction of the sthA mRNA with the small regulatory RNA spf, as seen from the

5

binding free energy increase from -14.52 to -1.57 kcal/mol.

6

Figure 7. Time profiles of cell growth (A), PDC (B), PCA (C) and DHS (D) of engineered E.

7

coli strains in flask cultures. Symbols are: blue square, GYT1 harboring pTacFABC and

8

pBBR1Gfbr-EcA; red circle, GYT6 harboring pTacFABC and pBBR1Gfbr-EcA; green triangle,

9

GYT7 harboring pTacFABC and pBBR1Gfbr-EcA.

10

Figure 8. Fed-batch fermentation profile of GYT1 harboring pTacFABC and pBBR1Gfbr-EcA.

11

Symbols are: blue circle, cell growth (OD600); red square, residual glucose concentration (g/L);

12

green diamond, PDC concentration (g/L); magenta triangle, PCA concentration (g/L); orange

13

inverted triangle, DHS concentration (g/L).

14

15

16

17

18

19

20


Page 32 of 43

Page 33 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1


Table 1. Strains and Plasmids Used in This Studya Name Strains DH5α

BL21(DE3) W3110 GYT0 GYT1 GYT2 GYT3 GYT4 GYT5 GYT6 GYT7 Plasmids pTrc99A pTac15K pET-22b(+) pBBR1MCS pTrcZ pTrcF pTrcFopt pTrcB pTacABC pETABC pTacFABC pTacFoptABC pBBR1Gfbr pBBR1Gfbr-A pBBR1Gfbr-EcA pBBR1Gfbr-KpA pBBR1Gfbr-RoA

Relevant genotype

Reference

F- endA1 glnV44 thi-1 recA1 relA1 gyrA96 deoR nupG purB20 φ80dlacZ∆M15 ∆(lacZYA-argF)U169, hsdR17(rK– mK+), λ– F- ompT gal dcm lon hsdSB (rB-mB-) (DE3) Coli genetic stock center strain No. 4474 W3110 ∆aroE W3110 ∆aroE PppsA::Ptrc W3110 ∆aroE ∆pykF W3110 ∆aroE ∆pykF ∆pykA W3110 ∆aroE ∆pykF PppsA::Ptrc W3110 ∆aroE PppsA::Ptrc PshiA::Ptrc W3110 ∆aroE PppsA::Ptrc PsthA::Ptrc W3110 ∆aroE PppsA::Ptrc PsthA::Ptrc (C15T, T18C, C21T)

Invitrogen

ApR, trc promoter, pBR322 origin, lacIq, 4.2 kb KmR, tac promoter, p15A origin, 4.0 kb ApR, T7 promoter, pBR322 origin, 5.5 kb CmR, lac, T3 and T7 promoters, 4.7 kb pTrc99A derivative containing aroZ gene from K. pneumonia pTrc99A derivative containing asbF gene from B. thuringiensis pTrc99A derivative containing E. coli-codon optimized asbF gene pTrc99A derivative containing qsuB gene from C. glutamicum ATCC13032 pTac15K derivative containing pmdABC operon from C. testosteroni pET-22b(+) derivative containing pmdABC operon from C. testosteroni pTacABC derivative containing trc-asbF-rrnBT1T2 cassette pTacABC derivative containing trc-asbFopt-rrnBT1T2 cassette pBBR1CS derivative containing aroGfbr gene under the lac promoter pBBR1Gfbr derivative containing RBS-tktA cassette pBBR1Gfbr derivative containing RBS-shiA cassette from E. coli pBBR1Gfbr derivative containing RBS-shiA cassette from K. pneumoniae pBBR1Gfbr derivative containing RBS-shiA cassette from R. opacus PD630

Lab stock Lab stock Novagen


Invitrogen CGSCb This study This study This study This study This study This study This study This study

49

This study This study This study This study This study This study This study This study This study This study This study This study This study


pBBR1Gfbr-AsA

1 2 3 4 5

pBBR1Gfbr derivative containing RBS-shiA cassette from A. This study sp. ADP1 pKD46 ApR, λ-Red recombinase under arabinose-inducible araBAD 43 promoter, ts origin, 6.3-kb pJW168 ApR, Cre-recombinase under IPTG-inducible lacUV5 44 promoter, ts origin, 5.5-kb 50 pECmulox ApR, CmR, lox66-cat-lox71, 3.5-kb pMtrc9 Modified pECmulox containing trc promoter downstream of Lab stock lox66-cat-lox71 cassette a Abbreviations: Ap, ampicillin; Km, kanamycin; Cm, chloramphenicol; R, resistance; ts, temperature sensitive. b Coli Genetic Stock Center, New Haven, CT.

6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 34 ACS Paragon Plus Environment

Page 34 of 43

Page 35 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1


Table 2. Batch Fermentation Performances of the Engineered Strains for PDC Production Strain

2 3 4

Maximum OD600 11.90

PDC (g/L) 4.28

PCA (g/L) 0.14

DHS (g/L) 1.71

Productivity (g/L/h) 0.104

Yield (g/g) 0.209

GYT1/pTacFABC +pBBR1Gfbr-EcAa 2.79 0.118 0.235 4.78 0.18 GYT1/pTacFABC 12.45 fbr b +pBBR1G -EcA 4.27 0.70 1.70 0.101 0.221 GYT1/pTacFABC 12.25 fbr c +pBBR1G -EcA 0.38 4.74 9.37 0.007 0.016 GYT7/pTacFABC 12.85 fbr b +pBBR1G -EcA 2.32 1.86 4.57 0.052 0.107 GYT6/pTacFABC 11.90 fbr b +pBBR1G -EcA a DO level was set as 25% of air saturation during fermentation. b DO level was set as 40% of air saturation during fermentation. c DO level was initially set as 40% of air saturation and then shifted to 80% of upon induction.



Figure 1. Biosynthetic pathway of PDC from glucose and the metabolic engineering strategies used in this study. Abbreviations: PTS, phosphotransferase system; TCA, tricarboxylic acid; PP, pentose phosphate; G6P, glucose 6-phosphate; G3P, glyceraldehydes 3-phosphate; PEP, phosphoenolpyruvate; PYR, pyruvate; OAA, oxaloacetate; E4P, erythrose 4-phosphate; DAHP, 3-deoxy-D-arabinoheptulosonate 7-phosphate; DHQ, 3-dehydroquinate; DHS, 3-dehydroshikimate; SHK, shikimate; S3P, shikimate-3-phosphate; EPSP, 5enolpyruvyl-shikimate 3-phosphate; CHA, chorismate; 4HBA, 4-hydroxybenzoate; PCA, protocatechuate; CHMS, 4-carboxy-2-hydroxymuconate-6-semialdehyde; PDC, 2-pyrone-4,6-dicarboxylic acid. Genes that encode enzymes: ppsA, PEP synthetase; pykF, pyruvate kinase I; pykA, pyruvate kinase II; pckA, PEP carboxykinase; ppc, PEP carboxylase; tktA, transketolase I; aroGfbr, feedback-inhibition resistant mutant of DAHP synthase; aroB, DHQ synthase; aroD, DHQ dehydratase; aroE, SHK dehydrogenase; aroK, SHK kinase I; aroL, SHK kinase II; aroA, 3-phosphoshikimate-1-carboxyvinyltransferase; aroC, CHA synthase; ubiC, CHA lyase; asbF, dehydroshikimate dehydratase; pmdAB, PCA 4,5-dioxygenase; pmdC, CHMS dehydrogenase; shiA, SHK:H+ symporter. Overexpressed genes are marked in blue, and red fork indicates gene deletion. SHK:H+ symporter, two recombinant plasmids and genome manipulations for promoter replacement employed in this study are illustrated. 160x106mm (300 x 300 DPI)

ACS Paragon Plus Environment

Page 36 of 43

Page 37 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 2. Comparison of two distinct pathways for PDC biosynthesis from glucose via in silico flux analysis. (A) Schematic for the two distinct PDC pathways: single-step and six-step, as well as the six-step pathway with cofactors intentionally removed. Cofactors (NADPH and ATP) are indicated. Abbreviations: DHS, 3dehydroshikimate; SHK, shikimate; S3P, shikimate-3-phosphate; EPSP, 5-enolpyruvyl-shikimate 3phosphate; CHA, chorismate; 4HBA, 4-hydroxybenzoate; PCA, protocatechuate; PDC, 2-pyrone-4,6dicarboxylic acid. Black arrows indicate E. coli native metabolic reactions, red arrows indicate heterologous enzymatic reactions, and dotted arrows indicate multiple steps of enzymatic reactions. (B) Performance of the different PDC pathways during in silico simulation experiments. The orange line corresponds to the result of the single-step pathway, the magenta line corresponds to the result of the six-step pathway, and the blue line corresponds to the result of the six-step pathway without cofactors requirement. 129x62mm (300 x 300 DPI)



Figure 3. Flask cultures for PCA production from glucose by introducing different DHS dehydratases (A) and for PDC production from glucose using the selected DHS dehydratase (B). Symbols are: gray box, cell growth (OD600); blue box, PCA concentration (g/L); red box, PDC concentration (g/L). 129x53mm (300 x 300 DPI)


Page 38 of 43

Page 39 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 4. Effects of overexpression of the aroGfbr, tktA and ppsA genes, and deletion of the pykF and/or pykA genes on PDC production. Symbols are: gray box, cell growth (OD600); red box, PDC concentration (g/L). 90x57mm (300 x 300 DPI)



Figure 5. Effect of overexpression of DHS importers from different microbial origins on PDC production. Engineered strains indicated are: Control, GYT1 harboring pTacFABC and pBBR1Gfbr; KpShiA, GYT1 harboring pTacFABC and pBBR1Gfbr-KpA; RoShiA, GYT1 harboring pTacFABC and pBBR1Gfbr-RoA; AsShiA, GYT1 harboring pTacFABC and pBBR1Gfbr-AsA; EcShiA, GYT1 harboring pTacFABC and pBBR1Gfbr-EcA; PtrcShiA, GYT5 harboring pTacFABC and pBBR1Gfbr; EcShiA+PtrcShiA, GYT5 harboring pTacFABC and pBBR1Gfbr-EcA. Symbols are: gray box, cell growth (OD600); red box, PDC concentration (g/L); yellow box, DHS concentration (g/L). 89x68mm (300 x 300 DPI)


Page 40 of 43

Page 41 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 6. Schematic representation of the chromosomal manipulations of the sthA gene: (1) promoter replacement with the strong trc promoter and (2) further introduction of three silent point mutations (C15T, T18C, C21T) in the coding sequence of sthA, which leads to weakened base-pairing interaction of the sthA mRNA with the small regulatory RNA spf, as seen from the binding free energy increase from -14.52 to 1.57 kcal/mol. 70x52mm (300 x 300 DPI)



Figure 7. Time profiles of cell growth (A), PDC (B), PCA (C) and DHS (D) of engineered E. coli strains in flask cultures. Symbols are: blue square, GYT1 harboring pTacFABC and pBBR1Gfbr-EcA; red circle, GYT6 harboring pTacFABC and pBBR1Gfbr-EcA; green triangle, GYT7 harboring pTacFABC and pBBR1Gfbr-EcA. 179x146mm (300 x 300 DPI)


Page 42 of 43

Page 43 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 8. Fed-batch fermentation profile of GYT1 harboring pTacFABC and pBBR1Gfbr-EcA. Symbols are: blue circle, cell growth (OD600); red square, residual glucose concentration (g/L); green diamond, PDC concentration (g/L); magenta triangle, PCA concentration (g/L); orange inverted triangle, DHS concentration (g/L). 89x71mm (300 x 300 DPI)


Metabolic Engineering of Escherichia coli for Efficient Production of 2

Recommend Documents