Title: Efficient whole-cell biocatalyst for 2,3-butanediol/acetoin production with NADH/NAD+ regeneration system in engineered Serratia marcescens

Transcription

1 JMB Papers in Press. First Published online May 24, 2017 DOI: /jmb Manuscript Number: JMB Title: Efficient whole-cell biocatalyst for 2,3-butanediol/acetoin production with NADH/NAD+ regeneration system in engineered Serratia marcescens Article Type: Research article Keywords: 2,3-butanediol, acetoin, Serratia marcescens, auto-inducing expression system, whole-cell biocatalyst

2 Full Title: Efficient whole-cell biocatalyst for 2,3-butanediol/acetoin production with NADH/NAD + regeneration system in engineered Serratia marcescens Authors: Rao Ben Truong Ngoc Tu Fan Jiying Sun Jian an Shen Yaling State Key Laboratory of Bioreactor Engineering, Shanghai Collaborative Innovation Center for Biomanufacturing Technology, East China University of Science and Technology, Shanghai , P.R.China * Corresponding author: Tel: ; fax: (Y.Shen) ylshen@ecust.edu.cn Running title: Efficient whole-cell biocatalyst by using Serratia marcescens 1

3 Abstract Serratia marcescens is reported to possess the potential for industrial 2,3-butanediol or acetoin production by fermentation. But 2,3-butanediol or acetoin are always co-produced and this may make purification process difficult. So biocatalytic technologies may be the appropriate production methods. In this study, we developed an auto-inducing expression system based on pet system and swr quorum sensing system by using S. marcescens. By using this system, S. marcescens could be engineered as whole-cell biocatalyst to obtain 2,3-butanediol or acetoin. In order to convert diacetyl to 2,3-butanediol, formate dehydrogenase (FDH) and 2,3-butanediol dehydrogenase were co-expressed to construct a NADH regeneration system. This whole-cell biocatalyst could efficiently produced 53.6 g/l 2,3-butanediol with a productivity of 3.35 g/lh. Next, in order to convert 2,3-butanediol to acetoin, NADH oxidase and 2,3-butanediol dehydrogenase were co-expressed to construct a NAD + regeneration system. This whole-cell biocatalyst could efficiently produced 59 g/l acetoin with a productivity of 2.95 g/lh. This work indicated this auto-inducing system was a powerful tool to construct whole-cell biocatalyst in S. marcescens in 2,3-butanediol and acetoin production. Keywords 2,3-butanediol; acetoin; Serratia marcescens; auto-inducing expression system; whole-cell biocatalyst 2

4 Introduction As promising bulk chemicals, 2,3-butanediol and acetoin have been given credit for having extensive application in varied fields such as fuel, chemical industry, food industry, and so on [1]. People could produce them by chemical synthetic, fermentative, enzymatic or biocatalytic technologies. However, because of the weaknesses of chemical synthesis, the large-scale production of 2,3-butanediol and acetoin is limited by using this technology. Several microorganisms were proved to be capable of produce 2,3-butanediol or acetoin, such as Klebsiella, Gluconobacter, Bacillus and Serratia, etc al. Morever, in most of these strains, 2,3-butanediol or acetoin are always co-produced and the difficulty of product purification in downstream processes would be increased. So it is difficult to obtain pure 2,3-butanediol or acetoin by fermentation. By contrast, enzymatic or biocatalytic technologies may be the appropriate production methods. As the conversion between 2,3-butanediol and acetoin is reversible and coupled with NADH/NAD + transformation, production of 2,3-butanediol or acetoin could be improved through manipulating the level of intracellular cofactors [2, 3]. Serratia marcescens is a Gram negative, bacillus shaped bacteria belongs to the family Enterobacteriaceae [4,5]. It is reported to possess the potential for industrial 2,3-butanediol or acetoin production which due to the strain s better resistance to bacteria pollution, relatively simple metabolic pathway, broad substrate spectrum and 3

5 cultural adaptability [4]. Quorum sensing is a bacterial phenomenon which enables bacteria to communicate and thereby coordinate the expression of specific target genes in response to the population density [6,7, 8, 9, 10]. S. marcescens MG1 employs a AHL-type quorum sensing system, the swr system, to control various functions, including swarming motility, production of extracellular enzymes and 2,3-butanediol [4]. This system depends on two proteins, SwrI, which directs the synthesis of the signal molecules N-butanoyl-L-homoserine lactone and N-hexanoyl-L-homoserine lactone, and SwrR, which, after binding with the signal molecules, is thought to activate or repress transcription of target genes. In S.marcescens MG1 the swr system combined with a LysR-like regulator, SlaR, are in charge of the production of 2,3-butanediol. SlaR is an essential activator for the transcription of genes related with 2,3-butanediol production and the disruption of it will make the strain can t produce acetoin and 2,3-butanediol. The swr system functions in two phases, at logarithmic phase it represses 2,3-butanediol production while this repression is released when the strain enters into stationary phase [5]. In this study, an auto-inducing expression system which could be utilized to express target genes in S.marcescens was developed. 4 By using it, S. marcescens was engineered as whole-cell biocatalyst to obtain 2,3-butanediol or acetoin. In order to convert diacetyl to 2,3-butanediol, formate dehydrogenase (FDH) and 2,3-butanediol

6 dehydrogenase were co-expressed to construct a NADH regeneration system. This whole-cell biocatalyst could efficiently produced 53.6 g/l 2,3-butanediol with a productivity of 3.35 g/lh. Next, in order to convert 2,3-butanediol to acetoin, NADH oxidase and 2,3-butanediol dehydrogenase from B.subtilis were co-expressed to construct a NAD + regeneration system. This whole-cell biocatalyst could efficiently produced 59 g/l acetoin with a productivity of 2.95 g/lh. This work indicated this auto-inducing system was a powerful tool to construct whole-cell biocatalyst in S. marcescens in 2,3-butanediol and acetoin production. Materials and methods Strains The bacterial strains and plasmids used in this study are listed in Table 1. Recombinant plasmids were constructed by standard procedures using E.coli Top10 as host strain. Antibiotics were used at the following concentrations (μg/ml): ampicillin (Amp), 100; kanamycin (Km), 50; chloramphenicol (Cm), 35. Construction of recombinant strain Serratia marcescens MG1ABC and plasmid pswnb By using the method described by Rao Ben [5], we constructed S.marcescens MG1ABC in which genes slaa, slab and slac were all inactivated. The vector pswnb was constructed in following steps: first, we obtained plasmid pbt-1 and 5

7 pbt-2 [5]. The promoter slaa was cloned into the BamI/XhoI sites of pbt-2 before T7 RNA polymerase gene to control its expression to construct target plasmid pswnb (Fig 1) by using primers pswnbf / pswnbr (Table 2). Construction of expression plasmids The fragment of green fluorescent protein gene was amplified from the plasmid pqbi63 (Quantum Biotechnologies Inc.) by primers Pgfpf/Pgfpr and ligated into NdeI-BamHI-linearized vector pet28a resulting in vector pet28a-gfp. The bdha encoding 2,3-butanediol dehydrogenase was amplified from Bacillus subtilis 168 genome DNA by using primers PbdhAf / PbdhAr and ligated into NcoI-SalI-linearized vector petduet-1 resulting in vector petduet-bdha. The yodc encoding NADH oxidase was amplified from Bacillus subtilis 168 genome DNA by using primers Pnoxf / Pnoxr and ligated into BglII-XhoI-linearized vector petduet-bdha resulting in vector petduet-bdha- yodc. The fdh gene fragment was amplified from Candida boidinii NCYC 1513 genome DNA by using primers Pfdhf / Pfdhr and ligated into BglII-XhoI-linearized vector petduet-bdha resulting in vector petduet-bdha-fdh. Culture conditions in flask and bioconversion conditions The auto-inducing vector pswnb and the recombinant plasmid pet-x carrying target genes were co-introduced into S. marcescens by electroporation. Plasmids pswnb 6

8 and pet were transformed into S. marcescens, resulting in the control strain. In order to study the expression of green fluorescent protein, overnight recombinant S.marcescens MG1 (pswnb and pet-gfp) cultures were inoculated in fresh LB media supplemented with 1% glucose to OD 600 =0.1 in flasks and incubated. In order to convert diacetyl to 2,3-butanediol, overnight recombinant S. marcescens (pswnb and petduet-bdha-fdh) cultures were inoculated in LB media supplemented with 1% glucose to OD 600 =0.1 in flasks and incubated. The cells were obtained by centrifugation at rpm for 5 min at 4. After being washed, the cell pellets were resuspended in buffer. The whole-cell biocatalyst reaction was performed as following: 50 ml of mixture was reacted at 30. ph was controlled at g/l of diacetyl and 32 g/l of formate were used as substrates. In order to convert 2,3-butanediol to acetoin, overnight recombinant S. marcescens (pswnb and petduet-bdha-yodc) cultures in LB media were inoculated in LB media supplemented with 1% glucose to OD 600 =0.1 in flasks and incubated at 30 with shaking at 250 rpm. The cells were obtained by centrifugation at rpm for 5 min at 4. After being washed, the cell pellets were resuspended in 50 ml 50 mm sodium phosphate buffer (ph 8.0) containing 40 g/l 2,3-butanediol. The whole-cell biocatalyst reaction was performed at 37 and 220 rpm in 500 ml flasks. 7

9 Analytical procedures Protein expression was analyzed on 12% (w/v) sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) as described by Laemmli (1970). Cell growth was monitored by measuring the absorbance at 600 nm (OD 600 ) with an Ultrospec 3000 spectrophotometer (Pharmacia biotech, USA). The fluorescence intensity was measured using 485nm (excitation) and 515nm (emission) with a fluorometer (Molecular Devices, model: SpectraMax M2), using a 96-well clear bottom black plate (Coaster Co). The products (acetoin and 2,3-butanediol) in the broth were extracted by using ethyl acetate containing n-butanol as the internal standard and then quantified by using a GC system(agilent GC9860; FID detector, DB-5 column). The determination of NADH and NAD + concentrations and assays of 2,3-BDH, NOX and FDH activities were performed according to the studies[2,3]. Results and discussion Construction of the auto-inducing expression system The auto-inducing expression system is based on the pet system and swr quorum sensing system in Serratia marcescens MG1. Here a two-step genetic switching circuit is built and consists of two plasmids: plasmid pswnb (Fig. 1A) is in charge of the expression of T7 RNA polymerase which is in charge of amplifying the transcriptional strength of promoter slaa. The expression plasmid pet-x produces the 8

10 proteins of interest. This system works as following (Fig. 1B): at logarithmic phase, SwrR produced by S. marcescens MG1 represses the expression of the slaa promoter. This effectively prevents the transcription of T7 RNA polymerase. As the cell density increases, AHL is produced by SwrI and alleviates the repression effects of SwrR. In the meanwhile, enough acetate has accumulated and binds with SlaR to trigger a full activation of the slaa promoter and allowing the downstream gene T7 RNA polymerase to be transcript when sugar exists. Then T7 RNA polymerase will amplify the transcription strength of slaa promoter and activate the transcription of genes under the control of T7 promoter. Thus target proteins will be produced. In order to validate the auto-inducing system, several genes were expressed in S. marcescens MG1 by using it and the results were showed in the following sections. Expression profile of GFP by using Serratia marcescens Gene gfp was expressed to investigate the expression profile of proteins and if it works in a population-dependent manner. The recombinant strain S. marcescens MG1 (pswnb+pet28a-gfp) were cultured in LB (1% glucose) media and harvested at different times. The cell growth profile and expression level of GFP were measured as described in section materials and methods. In Fig 2a, compared to the control strain S. marcescens MG1 (pswnb+pet28a), GFP expression levels increased apparently when strain S. marcescens MG1 (pswnb+pet28a-gfp) began to enter late logarithmic phase. As the large amount of exogenous protein expression imposed burden on the host strain, the growth of strain S. marcescens MG1 9

11 (pswnb+pet28a-gfp) was apparently decreasing since then (Fig 2b). In particular, GFP began to accumulate gradually after 4h and reached a maximum at 12h. It apparently revealed that the system could highly express target proteins from the logarithmic phase in S. marcescens MG1. In order to validate that the expression of the target proteins were really responding to the AHL signal and SlaR, the vectors pswnb and pet-gfp were also co-transformed into strain swri mutant S. marcescens MG1I and slar mutant S. marcescens MG1R resulting in strains S. marcescens MG1I (pswnb+ pet28a-gfp) and S. marcescens MG1R (pswnb+pet28a-gfp). These two strains were cultured in LB (1% glucose) media and products were measured and the results were showed in Fig 2a. Strain S. marcescens MG1R (pswnb+pet28a-gfp) completely can t express GFP while strain S. marcescens MG1I (pswnb+pet28a-gfp) could produce a certain amount of the target protein. After 0.5mM AHL purchased from Sigma Aldrich was added to the culture of S. marcescens MG1I (pswnb+pet28a-gfp) in mid-log phase (OD 600 = ), the production ability of S. marcescens MG1I (pswnb+pet28a-gfp) strain was restored (data not shown). These were in agreement with the fact that the slaa promoter is controlled by the quorum sensing system and regulator SlaR. In addition, it was validated that proteins produced by the auto-inducing system were really induced because of AHL signal and SlaR. Overexpression of proteins is one of the effective ways to rebuild the metabolic pathway and perform some useful task. Here we constructed an auto-inducing 10

12 expression system by employing the slaa promoter which is dual-regulated in the process of acetoin biosynthesis in S. marcescens. The same as in Vibrio cholerae, SwrR and SlaR are in charge of this process. SlaR, which belongs to the LysR-type protein family, is an essential activator for the transcription. When SlaR was disrupted, slaa (encoded alpha-acetolactate decarboxylase) and slab (encoded alpha-acetolactate synthase) gene which controlled by the slaa promoter were not expressed. Moreover, SlaR can t activate the transcription of acetoin operon in the absence of acetate. SwrR, which acts as the receptor of the signal molecules, exerts a negative regulatory effect on the transcription of acetoin operon. In detail, SwrR efficiently represses the transcription of slaa promoter at low cell density. As the cell density increases, the repression effects of SwrR are gradually alleviated. At this point, enough acetate has accumulated triggering a full activation of slaa promoter and allowing high levels transcription of genes slaa and slab. Because the transcription of the slaa promoter is directly linked to cell-density, in this study it was used to control the expression of T7 RNA polymerase which activates the transcription of target genes. It showed that the regulation mechanism of 2,3-butanediol production in S. marcescens is very useful to control gene expression. This constructed auto-inducing system does not need specific medium or growth conditions to guarantee auto-induction. It responds directly to cell density and grows to saturation without the need to monitor culture growth or to add an inducer at a specific point in time of the process. In order to validate the auto-inducing system, gene coding for green fluorescent protein was expressed in S. marcescens. The result showed that only when the 11

13 recombinant strain is at high cell density does the GFP expression levels increased significantly. In addition, the S. marcescens mutant that can t synthesize signal molecules or regulator SlaR expressed none or a certain amount of GFP after being transformed with pswnb and pet-gfp. The experiments were not only performed in flasks, but also carried out at fermentor scale. All these clearly indicated that the auto-inducing system works as it s designed. Bioconversion of diacetyl to 2,3-butanediol In strain S. marcescens MG1ABC, genes slaa, slab and slac were all inactivated. So this strain couldn t produce acetoin and 2,3-butanediol itself and suitable to be constructed as whole-cell biocatalyst to achieve bioconversion. 2,3-BDH from Bacillus subilis 168 and FDH from Candida boidnii NCYC 1513 were cloned and recombinant strains were constructed, respectively. Their enzymatic activities were measured and shown in Table 3. After fermentation, the extract of S. marcescens MG1ABC (pswnb+petduet-bdha-fdh) showed a 2,3-BDH activity of 57.8 U/mg and a FDH activity of 0.5 U/mg. Compared to use E.coli as host strain [11], higher enzymatic activities could be obtained by using S. marcescens. SDS-PAGE analyses showed that these two proteins were expressed in S. marcescens MG1ABC (Fig. 3). Obviously these two genes could be successfully co-expressed in S. marcescens which can be used to convert diacetyl to 2,3-butanediol. However, as BDH and FDH have similar molecular weights, they were poorly resolved when being co-expressed in S. marcescens MG1ABC. 12

14 The bioconversion experiments were first performed in flasks under the conditions in section materials and methods. The initial DA concentration was 30 g/l. Cells containing plasmid petduet-bdha and petduet-bdha-fdh were used as whole cell biocatalysts to achieve the bioconversion. As shown in Table 4, we can obtain g/l of 2,3-BD with a productivity of 4.66 g/lh by using S. marcescens MG1ABC (pswnb+petduet-bdha-fdh) after 6 h. And the productivity of 4.66 g/lh was 1.3-fold than that by using E.coli as whole cell biocatalysts. It was obviously higher 2,3-BD concentration, productivity and yield could be obtained by using S. marcescens. It means S. marcescens was appropriately used as whole cell biocatalysts to convert DA to 2,3-BD. The introduction of FDH into S. marcescens was mean to regenerate cofactors. We examined the intracellular concentrations of NADH and NAD + in recombinant strains to confirm this (Fig. 4A, 4B, 4C). Compared to the strain S. marcescens MG1ABC (pswnb+petduet-bdha), the intracellular NADH/NAD + was higher in the strain S. marcescens MG1ABC (pswnb+petduet-bdha+fdh). It showed a relatively high level of NADH and a low level of NAD +. After several hours, the NADH level and ratio of NADH to NAD + increased continuously and meant that NADH was indeed reproduced in the strain S. marcescens MG1ABC (pswnb+petduet-bdha+fdh). It means the NADH regeneration efficiency in strain S. marcescens MG1ABC (pswnb+petduet-bdha+fdh) was higher than in S. marcescens MG1ABC (pswnb+petduet-bdh). This could accelerate the production of 2,3-BD. In order to improve the concentrations of 2,3-BD, fed-batch conversion experiments 13

15 were performed (Fig 5). The initial DA concentration was 30 g/l and 10 g/l of it was added at 2, 4 and 6 h. Fig 5 showed 53.6 g/l of 2,3-BD was produced with a productivity of 3.35 g/(l;h). The yield of 2,3-BD on DA was high and only 1.3 g/l of AC was accumulated. Almost no formate was detected at the end of the reaction, it means it was used up. Unlike several microorganisms which are susceptible to the toxicity of many organic chemicals, S. marcescens exhibited remarkable tolerance to 2,3-butanediol and acetoin. In previous studies, S. marcescens was reported to possess the potential for industrial acetoin and 2,3-butanediol production. However, these studies all focused on fermentative technologies by using S. marcescens and it was difficult using fermentation parameter optimization to improve production of the target products. Here we tried to produce acetoin and 2,3-butanediol by using S. marcescens as whole cell biocatalysts. The bioconversion reaction between 2,3-BD and acetoin is coupled with the NADH/NAD + transformation. So it s necessary to manipulate the level of intracellular cofactors. Considering the internal cofactor regeneration is always not efficient enough, researchers introduce other enzyme to accelerate the cofactor regeneration. We co-expressed 2,3-BDH and FDH to convert diacetyl to 2,3-butanediol in S. marcescens. By using the auto-inducing system, it was easy to co-express two proteins. 2,3-BDH from B.subtilis and FDH from Candida boidinii NCYC 1513 were subcloned into the expression plasmid petduet resulting in recombinant plasmid petduet-bdha-fdh. Results showed that S. marcescens MG1ABC 14

16 (pswnb+petduet-bdha) could convert diacetyl to 2,3-butanediol. However, as the NADH was consumed gradually, it was difficult to maintain the intracellular NADH concentration. In order to improve NADH regeneration efficiency, FDH from Candida boidinii NCYC 1513 was co-expressed in S. marcescens. It showed introduction of FDH could apparently improve intracellular NADH/NAD + level. After culturing, the enzyme activities of 2,3-BDH and FDH produced by S. marcescens were all higher than by E.coli. Corresponding with this, higher 2,3-butanediol concentration, productivity and yield could be obtained by using S. marcescens. It was obviously S. marcescens is very suitable to produce 2,3-butanediol. Bioconversion of 2,3- butanediol to acetoin The bdha and yodc from Bacillus subilis 168 were cloned and recombinant strains were constructed, respectively. Their enzymatic activities were measured and shown (Table 5). After fermentation, strain S. marcescens MG1ABC (pswnb+petduet-bdha-yodc) showed an AR activity of U/mg and a NOX activity of U/mg. SDS-PAGE analyses of the recombinant proteins were shown in Figure. 6. Obviously the results of enzyme activity assay and SDS-PAGE indicated these two genes could be successfully co-expressed in S. marcescens which can be used to convert 2,3-butanediol to acetoin. The bioconversion experiments were performed in flasks under the conditions in 15

17 previous section. The initial 2,3-BD concentration was 50 g/l. We directly used the cells containing plasmid petduet-bdha-yodc as whole cell biocatalysts to achieve the bioconversion. As shown in Table 6, we can obtain 44.9 g/l of acetoin with a productivity of 3.74 g/(lh) by using S. marcescens MG1ABC (pswnb+petduet-bdha-yodc). The conversion yield of 2,3-butanediol on acetoin was 89.8%. It was obviously high acetoin concentration, productivity and yield could be obtained by using S. marcescens. It means S. marcescens could be appropriately used as whole cell biocatalysts to convert 2,3-butanediol to acetoin. The introduction of NOX into S. marcescens MG1ABC was mean to increase the overall intracellular NAD + pool. We examined the intracellular concentrations of NADH and NAD + in recombinant strains to confirm this (Fig. 7A, 7B, 7C). Compared to the strain S. marcescens MG1ABC (pswnb+petduet-bdha), the intracellular NADH/NAD + was significantly lower in strain S. marcescens MG1ABC (pswnb+petduet-bdha-yodc). It means the NAD + regeneration efficiency in strain S. marcescens MG1ABC (pswnb+petduet-bdha-yodc) containing yodc gene was higher than in S. marcescens MG1ABC (pswnb+petduet-bdha). This could accelerate the production of acetoin. In order to improve the concentrations of acetoin, fed-batch conversion experiments were performed (Fig. 8). The initial 2,3-BD concentration was 50 g/l and 30 g/l of it was added at 8 h. Fig 8 showed 59 g/l of acetoin was produced with a productivity of 2.95 g/(l;h). The yield of acetoin on 2,3-BD was high. In previous section, we used S. marcescens to convert diacetyl to 2,3-butanediol. 16

18 Since BDH was a reversible enzyme, it can also be used to convert 2,3-butanediol to acetoin. However, as the reduced NAD + pool may restrict the process for the bioconversion, here NOX was expressed to maintain low NADH/NAD +. The following experiments showed introduction of yodc could indeed reduce intracellular NADH/NAD + level. After culturing, higher acetoin concentration, productivity and yield could be obtained by using S. marcescens. It was obvious S. marcescens was also suitable to produce acetoin. In this study, we developed an auto-inducing expression system which could be utilized to construct various whole-cell biocatalyst systems. As the quorum sensing was generally occurred response to high cell density at stationary growth phase which may be not ideal for the production of bio-based chemicals or compounds, we produced enzymes by using the auto-inducing expression system. Then we produced 2,3-butanediol and acetoin by using the constructed whole-cell biocatalyst systems. By building different cofactor regeneration systems, S. marcescens could be used in different bioconversion processes. The results showed it was suitable to produce 2,3-butanediol and acetoin. Abbreviations 17

19 AHL, N-acyl homoserine lactones; OOHL, N-3-oxooctanoyl-L-homoserine; SEAP, secreted alkaline phosphatase; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; GFP,green fluorescent protein; slaa, acetoactate decarboxylase ; slab, acetolactate synthase; XylA, xylose isomerase; XylB, xylulokinase; X5P, xylulose-5-phosphate; BHL, N-butanoyl-L-homoserine lactone ; HHL, N-hexanoyl-L-homoserine lactone. Acknowledgements This work was supported by Shanghai Leading Academic Discipline Project (Project: B505) and National Special Fund for State Key Laboratory of Bioreactor Engineering, (No ). References 1. Xiu Z, Zeng A Present state and perspective of downstream processing of biologically produced 1,3-propanediol and 2,3-butanediol. Appl. Microbiol. Biotechnol. 78: Bao T, Zang X, Rao Z, Zhao X, Zhang R, Yang T, Xu Z, Yang S Efficient Whole-Cell Biocatalyst for Acetoin Production with NAD + Regeneration System through Homologous Co-Expression of 2,3-Butanediol Dehydrogenase and NADH Oxidase in Engineered Bacillus subtilis. PLoS. One. 18; 9(7): e Wang Y, Li L, Ma C, Gao C, Tao F, Xu P Engineering of cofactor 18

20 regeneration enhances (2S,3S)-2,3-butanediol production from diacetyl. Sci. Rep. 3: Hejazi A, Falkiner F Serratia marcescens. J. Med. Microbiol. 46: Rao B, Zhang L, Sun J, Su G, Wei D, Chu J, Zhu J, Shen Y Characterization and regulation of the 2,3-butanediol pathway in Serratia marcescens. Appl. Microbiol. Biotechnol. 93: Bainton N, Bycroft B, Chhabra S, Stead P, Glehill L, Hill P, Rees C, Winson M, Salmond G, Stewat G, Williams P A general role for the lux autoinducer in bacterial cell signalling: control of antibiotic biosynthesis in Erwinia. Gene. 116: Basu S, Gerchman Y, Collins C, Arnold F, Weiss R A synthetic multicellular system for programmed pattern formation. Nature. 434: Derzelle S, Duchaud E, Kunst F, Danchin A, Bertin P Identification, characterization, and regulation of a cluster of genes involved in carbapenem biosynthesis in Photorhabdus luminescens. Appl. Environ. Microbiol. 68: Miller M, Bassler B Quorum sensing in bacteria. Annu. Rev. Microbiol. 55: Taga M, Bassler B Chemical communication among bacteria. Proc. Natl. Acad. Sci. 100: Rao B, Fan J, Sun J, Tu T N, Sun J, Zhou J, Qiu Y, Shen Y An auto-inducible expression system based on the rhli-rhlr quorum-sensing regulon 19

21 for recombinant protein production in E. coli. Biotechnol. Bioproc. E. 21(1): Table 1 Bacterial strains and plasmids used in this study Strain or plasmid Relevant characteristics Reference or source Serratia marcescens The source of other strains in this study Rao et al. (2012) MG1 Serratia marcescens Serratia marcescens MG1 swri - mutant Rao et al. (2012) MG1I Serratia marcescens Serratia marcescens MG1 slar - mutant Rao et al. (2012) MG1R Serratia marcescens Serratia marcescens MG1 slaabc - In this study MG1ABC mutant pbt-1 pbt vector without λci gene Rao et al. (2012) pbt-2 pbt-1 vector carrying T7 RNA polymerase Rao et al. (2012) gene pswnb Newly constructed switch vector In this study pet28a T7-based expression vector Invitrogen petduet T7-based expression vector Invitrogen pet28a-gfp pet28a vector containing gfp expression In this study cassette petduet-bdha petduet vector containing bdha expression In this study cassette petduet-bdha- yodc petduet vector containing bdha and yodc In this study expression cassettes petduet-bdha-fdh petduet vector containing bdha and fdh In this study expression cassettes Table 2 Primers used in this study Primer pswnbf pswnbr 20 Sequence attggatccgagccgcctgcggagttgat attctcgagggcaatggtggtttcaccctc

22 Pgfpf Pgfpr PbdhAf PbdhAr Pnoxf Pnoxr Pfdhf Pfdhr attcatatgatgagtaaaggagaagaactttcactgg attggatcctcacttgtacagctcgtccatgcc attagatctatgaaggcagcaagatgg attctcgagatgacgaata CTCTGGAT attccatgggttaggtctaacaagg attgtcgaccagccaa GTTGATAC attagatctaagatcgttttagtcttatatgatgctggta attctcgagttatttcttatcgtgtttaccgtaagctttg Table 3 Activities of 2,3-BDH and FDH Strain 2,3-BDH activity (U/mg) FDH activity (U/mg) S. MG1ABC (petduet) ND ND S. MG1ABC (petduet-bdha) 108.1±2.1 ND S.MG1ABC (petduet-bdha- 57.8± ±0.03 fdh) Table 4 The products of batch bioconversion with different cofactor regeneration systems Strain 2,3-BD(g/L) AC(g/L) Productivity of S. MG1ABC (petduet) ND ND S. MG1ABC (petduet-bdha) S. MG1ABC (petduet-bdha- fdh) 2,3-BD (g/(lh)) ND ND ND 21 Yield of 2,3-BD (%) 22.7± ± ± ± ± ± ± ±1.45 Table 5. Enzyme activity and the whole-cell biocatalytic activity of different strains Strains AR activity (U/mg) BDH activity (U/mg) NOX activity (U/mg) S. MG1ABC (petduet)

23 S. MG1ABC (petduet-bdha) S. MG1ABC (petduet-bdhayodc) 221.3± ±1.7 ND 212.4± ± ±2.3 Table 6 The products of batch bioconversion with different cofactor regeneration systems Strain 2,3-BD(g/L) AC(g/L) Productivity of S. MG1ABC (petduet) S. MG1ABC (petduet-bdha) S. MG1ABC (petduet-bdhayodc) ND 22 ND AC (g/(lh)) Yield of AC (%) 10.9± ± ± ± ± ± ± ±1.45 Fig 1 (A)Structure of the plasmid pswnb. (B) The mechanism of the auto-inducing system. Fig 2 Comparison of specific GFP levels (a) and cell growth (b) and in LB medium. Diamond, Serratia marcescens MG1 (pswnb+pet28a-gfp); triangle, Serratia marcescens MG1I (pswnb+pet28a-gfp); square, Serratia marcescens MG1R (pswnb+pet28a-gfp); circle, Serratia marcescens MG1 (pswnb+pet28a). Fig 3 Validation of the expression of 2,3-BDH and FDH in Serratia marcescens

24 through SDS-PAGE. M, maker; lane 1, Serratia marcescens MG1ABC (pswnb+petduet); lane 2, Serratia marcescens MG1ABC (pswnb+petduet-bdha); lane 3, Serratia marcescens MG1ABC (pswnb+petduet-fdh); lane 4, Serratia marcescens MG1ABC (pswnb+petduet-bdha+fdh). Fig 4 Effects of NADH regeneration system on levels of intracellular NADH, NAD + and ratios of NADH to NAD +. Circular represents Serratia marcescens MG1ABC (pswnb+petduet-bdha-fdh). Square represents Serratia marcescens MG1ABC (pswnb+petduet-bdha). Fig 5 Time course of fed-batch bioconversion by Serratia marcescens MG1ABC (pswnb+petduet-bdha-fdh). Circular represents acetoin. Square represents 2,3-butanediol. Fig 6 Validation of the expression of 2,3-BDH and NOX in Serratia marcescens through SDS-PAGE. M, maker; lane 1, Serratia marcescens MG1ABC (pswnb+petduet); lane 2, Serratia marcescens MG1ABC (pswnb+petduet-bdha); lane 3, Serratia marcescens MG1ABC (pswnb+petduet-yodc); lane 4, Serratia marcescens MG1ABC (pswnb+petduet-bdha+yodc). 23

25 Fig 7 Effects of NAD + regeneration system on levels of intracellular NADH, NAD + and ratios of NADH to NAD +. Circular represents Serratia marcescens MG1ABC (pswnb+petduet-bdha-yodc). Square represents Serratia marcescens MG1ABC (pswnb+petduet- yodc). Fig 8 Time course of fed-batch bioconversion by Serratia marcescens MG1ABC (pswnb+petduet- bdha-yodc). Circular represents 2,3-butanediol. Square represents acetoin. 24

26 Fig. 1 (A)Structure of the plasmid pswnb. (B) The mechanism of the auto-inducing system.

27 Fig. 2 Comparison of specific GFP levels (a) and cell growth (b) and in LB medium. Diamond, Serratia marcescens MG1 (pswnb+pet28a-gfp); triangle, Serratia marcescens MG1I (pswnb+pet28a-gfp); square, Serratia marcescens MG1R (pswnb+pet28a-gfp); circle, Serratia marcescens MG1 (pswnb+pet28a).

28 Fig. 3 Validation of the expression of 2,3-BDH and FDH in Serratia marcescens through SDS-PAGE. M, maker; lane 1, Serratia marcescens MG1ABC (pswnb+petduet); lane 2, Serratia marcescens MG1ABC (pswnb+petduet-bdha); lane 3, Serratia marcescens MG1ABC (pswnb+petduet-fdh); lane 4, Serratia marcescens MG1ABC (pswnb+petduet-bdha+fdh).

29 Fig. 4 Effects of NADH regeneration system on levels of intracellular NADH, NAD+ and ratios of NADH to NAD+. Circular represents Serratia marcescens MG1ABC (pswnb+petduet-bdha-fdh). Square represents Serratia marcescens MG1ABC (pswnb+petduet-bdha).

30 Fig. 5 Time course of fed-batch bioconversion by Serratia marcescens MG1ABC (pswnb+petduet-bdhafdh).

31 Fig. 6 Validation of the expression of 2,3-BDH and NOX in Serratia marcescens through SDS-PAGE. M, maker; lane 1, Serratia marcescens MG1ABC (pswnb+petduet); lane 2, Serratia marcescens MG1ABC (pswnb+petduet-bdha); lane 3, Serratia marcescens MG1ABC (pswnb+petduet-yodc); lane 4, Serratia marcescens MG1ABC (pswnb+petduet-bdha+yodc).

32 Fig. 7 Effects of NAD+ regeneration system on levels of intracellular NADH, NAD+ and ratios of NADH to NAD+. Circular represents Serratia marcescens MG1ABC (pswnb+petduet-bdha-yodc). Square represents Serratia marcescens MG1ABC (pswnb+petduet- yodc).

33 Fig. 8 Time course of fed-batch bioconversion by Serratia marcescens MG1ABC (pswnb+petduet- bdhayodc).