This document is downloaded from DR‑NTU (https://dr.ntu.edu.sg)Nanyang Technological University, Singapore.

Harnessing the Periplasm of Bacterial Cells ToDevelop Biocatalysts for the Biosynthesis ofHighly Pure Chemicals

Yang, Yun; Wu, Yichao; Hu, Yidan; Wang, Hua; Guo, Lin; Fredrickson, James K.; Cao, Bin

2017

Yang, Y., Wu, Y., Hu, Y., Wang, H., Guo, L., Fredrickson, J. K., et al. (2017). Harnessing thePeriplasm of Bacterial Cells To Develop Biocatalysts for the Biosynthesis of Highly PureChemicals. Applied and Environmental Microbiology, 84(1), e01693‑17‑.

https://hdl.handle.net/10356/87209

https://doi.org/10.1128/AEM.01693‑17

© 2017 American Society for Microbiology. This paper was published in Applied andEnvironmental Microbiology and is made available as an electronic reprint (preprint) withpermission of American Society for Microbiology. The published version is available at:[http://dx.doi.org/10.1128/AEM.01693‑17]. One print or electronic copy may be made forpersonal use only. Systematic or multiple reproduction, distribution to multiple locationsvia electronic or other means, duplication of any material in this paper for a fee or forcommercial purposes, or modification of the content of the paper is prohibited and issubject to penalties under law.

Downloaded on 30 Dec 2021 08:36:16 SGT

Harnessing the Periplasm of Bacterial Cells To DevelopBiocatalysts for the Biosynthesis of Highly Pure Chemicals

Yun Yang,a Yichao Wu,b,c* Yidan Hu,b,c Hua Wang,a Lin Guo,a James K. Fredrickson,d Bin Caob,c

aSchool of Chemistry and Beijing Advanced Innovation Center for Biomedical Engineering, Beihang University,Beijing, People's Republic of China

bSchool of Civil and Environmental Engineering, Nanyang Technological University, SingaporecSingapore Centre for Environmental Life Sciences Engineering, Nanyang Technological University, SingaporedPacific Northwest National Laboratory, Richland, Washington, USA

ABSTRACT Although biocatalytic transformation has shown great promise in chem-ical synthesis, there remain significant challenges in controlling high selectivity with-out the formation of undesirable by-products. For instance, few attempts to con-struct biocatalysts for de novo synthesis of pure flavin mononucleotide (FMN) havebeen successful, due to riboflavin (RF) accumulating in the cytoplasm and being se-creted with FMN. To address this problem, we show here a novel biosynthesis strat-egy, compartmentalizing the final FMN biosynthesis step in the periplasm of an en-gineered Escherichia coli strain. This construct is able to overproduce FMN with highspecificity (92.4% of total excreted flavins). Such a biosynthesis approach allows iso-lation of the final biosynthesis step from the cytoplasm to eliminate undesirable by-products, providing a new route to develop biocatalysts for the synthesis of high-purity chemicals.

IMPORTANCE The periplasm of Gram-negative bacterial hosts is engineered to com-partmentalize the final biosynthesis step from the cytoplasm. This strategy is promis-ing for the overproduction of high-value products with high specificity. We demon-strate the successful implementation of this strategy in microbial production ofhighly pure FMN.

KEYWORDS biosynthesis, periplasm, flavin mononucleotide, synthetic biology,metabolic engineering, periplasmic space, riboflavin

Synthetic biology and metabolic engineering approaches have been used success-fully to produce a wide range of value-added products (1–8). Most efforts in

biosynthesis have focused on the establishment and optimization of metabolic fluxestoward the targeted product in the cytoplasm, before excretion or extraction (9–12).The product line of available biocatalysts is limited by current strategies, especiallywhen the targeted compound is structurally similar to other metabolites, which com-plicates the downstream processes for purification.

Flavin mononucleotide (FMN) is a high-value chemical that is widely used as a foodadditive (13), a pharmaceutical agent (14), an energy storage material in rechargeablebatteries (15, 16), and an electron shuttle (17–19) for some exoelectrogens functioningin bioelectrochemical systems. Currently, the industrial production of FMN is achievedby a combination of biotechnological and chemical techniques (see Fig. S1 in thesupplemental material), while the commercial FMN products are extremely expensiveand always of low purity (�70%) (20). High-purity riboflavin (RF), the precursor of FMN,is produced by engineered microbes at low cost (21). RF is then chemically transformedto FMN with phosphorylating reagents (22). Due to the nonspecificity of the phosphor-ylation reaction, several by-products are generated, including isomeric RF monophos-

Received 3 August 2017 Accepted 16October 2017

Accepted manuscript posted online 27October 2017

Citation Yang Y, Wu Y, Hu Y, Wang H, Guo L,Fredrickson JK, Cao B. 2018. Harnessing theperiplasm of bacterial cells to developbiocatalysts for the biosynthesis of highlypure chemicals. Appl Environ Microbiol84:e01693-17. https://doi.org/10.1128/AEM.01693-17.

Editor Haruyuki Atomi, Kyoto University

Copyright © 2017 American Society forMicrobiology. All Rights Reserved.

Address correspondence to Lin Guo,guolin@buaa.edu.cn, or Bin Cao,bincao@ntu.edu.sg.

* Present address: Yichao Wu, College ofResources and Environment, HuazhongAgricultural University, Wuhan, People'sRepublic of China.

Y.Y., Y.W., and Y.H. contributed equally to thiswork.

BIOTECHNOLOGY

crossm

January 2018 Volume 84 Issue 1 e01693-17 aem.asm.org 1Applied and Environmental Microbiology

on February 2, 2018 by N

AN

YA

NG

TE

CH

NO

LOG

ICA

L UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

phates, RF bisphosphates, and RF multiphosphates, as well as residual RF (20, 23).Because of the similarity in chemical structures and properties of FMN and theby-products, FMN purification from the resultant chemical pool has been challengingand costly, which leads to a dramatic price increase and purity decrease from RFprecursor to FMN (Fig. S1).

Biosynthesis based on cellular metabolism has unparalleled advantages over chem-ical processes, in that reactions are highly specific and precise (24). Attempts have beenmade to develop FMN biocatalysts (25–27). However, the ubiquitous presence ofbifunctional RF kinase/flavin adenine dinucleotide (FAD) synthase in prokaryotes, whichcatalyzes the transformation from RF to the FMN intermediate and then to FAD, rendersit very challenging to engineer cells to accumulate FMN in bacterial cytoplasm, withsubsequent excretion (28, 29). To address this issue, a eukaryotic gene encoding amonofunctional RF kinase was overexpressed in Candida famata, resulting in FMNaccumulation in the cytoplasm and excretion (25, 30). However, this effort couldachieve only 60 to 80% FMN purity, with a considerable amount of RF also accumu-lating in the cytoplasm and being excreted along with FMN (25). Since there is only onestep between RF and FMN, attempts to overaccumulate FMN in the cytoplasm willinevitably lead to a relatively small portion of RF being accumulated and excreted alongwith FMN. Therefore, production of high-purity FMN remains challenging. The objectiveof this study is to establish a novel biosynthesis approach by engineering the hoststrain to compartmentalize the final step of FMN biosynthesis in the periplasm so as toenable the production of high-purity FMN.

The coexistence of RF and FMN in the cytoplasm is the main reason for the lowpurity of microbially produced FMN. Herein, we design a novel FMN biosynthesispathway in Escherichia coli, in which FAD rather than FMN is overproduced in thecytoplasm, while the conversion of FAD to FMN occurs in the periplasm (Fig. 1). Basedon a synthetic RF biosynthesis pathway introduced into the E. coli host (31), anexogenous bifunctional RF kinase/FAD synthase is expressed to convert cytoplasmic RFto FAD. A heterologous, highly specific, inner membrane (IM) FAD exporter is used toexport FAD to the periplasm, and then a heterologous periplasmic 5=-nucleotidasecatalyzes the transformation of FAD to FMN, which is excreted to the cell exterior,presumably by diffusion through outer membrane (OM) porins. Ascribed to the uniquearrangement of the final FMN catalytic step in the periplasm instead of the cytoplasm,unprecedented high-purity FMN (�90% of total excreted flavins) in the spent mediumwas achieved prior to any form of downstream processing.

RESULTS AND DISCUSSIONManipulation of the intracellular flavin pool. To integrate the aforementioned

synthetic pathway into E. coli, a two-plasmid system was used. The pSB3C5 vector, witha p15A origin and lac promoter, was applied to overexpress cytoplasmic proteins, andthe pSB4K5 vector with pSC101 origin and tet promoter was adopted to overexpressmembrane and periplasmic enzymes. For ease of description, these two expressionvectors are referred to as lac and tet, respectively. The cytoplasmic flavin pool of E. coliwas manipulated by the heterologous enzymes RibADEH and RibC derived fromBacillus subtilis (Fig. 2). The intracellular flavin content of the wild-type E. coli containingempty vectors (referred to as lac�tet) was at a low level (0.08 � 0.01 �mol/g protein)and was mainly composed of FMN and FAD. A synthetic RF synthesis pathway con-taining ribADEH from B. subtilis (31) was inserted into the lac vector, forming theresultant plasmid A4. Upon introduction of the RF synthesis pathway, the flavin pool ofrecombinant E. coli (referred to as A4�tet) increased �3-fold, and the intracellular RFand FMN levels were also greatly enhanced (from �0 to 0.14 � 0.04 �mol/g proteinand from 0.07 � 0.00 to 0.21 � 0.03 �mol/g protein, respectively). To effectivelychannel intracellular RF toward FMN synthesis, a ribC gene encoding bifunctional RFkinase/FAD synthase derived from B. subtilis was added into the RF synthesis pathway;the lac vector harboring ribADEH and ribC was named A5. Comparing the flavin contentof recombinant E. coli A5�tet with that of A4�tet, the high levels of intracellular flavin

Yang et al. Applied and Environmental Microbiology

January 2018 Volume 84 Issue 1 e01693-17 aem.asm.org 2

on February 2, 2018 by N

AN

YA

NG

TE

CH

NO

LOG

ICA

L UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

remained intact, while RibC successfully converted cytoplasmic RF to FAD. This indi-cates that the heterologous RibC from B. subtilis is effective in controlling the profile ofthe intracellular flavin pool, with minimal RF.

Compartmentalization of the final FMN biosynthesis step in the periplasm. Theperiplasmic catalytic machinery was constructed by introducing the IM FAD transporter

FIG 1 Schematic view showing the design of a novel biosynthesis pathway for flavin mononucleotide (FMN) overproduction with high selectivity. (a) The periplasmof the bacterial host is engineered to compartmentalize the final biosynthesis step from the cytoplasm. This strategy is effective for the overproduction of valuablechemicals with high selectivity. This strategy is implemented in microbial production of highly pure FMN. (b) First, a synthetic riboflavin (RF) biosynthesis pathway,including RibADEH from Bacillus subtilis, that generates RF from GTP and ribulose 5-phosphate (R5P) is introduced into the Escherichia coli host. In order to eliminatethe presence of cytoplasmic RF, the bifunctional RF kinase/flavin adenine dinucleotide (FAD) synthase RibC from B. subtilis is incorporated to convert cytoplasmic RFto FAD. Then, the heterologous inner membrane (IM) FAD exporter SO0702 from Shewanella oneidensis channels cytoplasmic FAD into the periplasmic space. Finally,FMN is overproduced from FAD in the periplasm via the heterologous periplasmic 5=-nucleotidase UshA from Shewanella oneidensis and diffuses to the cell exteriorthrough outer membrane (OM) porins. Highly pure FMN with a minimal presence of RF could be produced with this novel biosynthesis approach by harnessing theperiplasm to catalyze the final step in the biosynthesis of FMN to avoid the generation of by-products.

FIG 2 Description of the double vectors in different engineered E. coli strains and their corresponding intracellular and periplasmic flavin contents. (a) Doubleplasmids were used in E. coli to construct the synthetic pathways. The pSB3C5 vector with p15a origin and lac promoter (referred to as the lac vector) was usedto overexpress cytoplasmic proteins, and the pSB4K5 vector with pSC101 origin and tet promoter (referred to as the tet vector) was used to overexpressmembrane and periplasmic enzymes. The designations for different engineered E. coli strains are shown on the left, and their corresponding vectors areindicated on the right. (b) The intracellular flavin contents, including RF, FMN, and FAD, of different engineered E. coli strains were tested after aerobic growthfor 48 h in modified M9 medium containing 15 g/liter xylose. (c) The periplasmic fraction of E. coli cells was isolated using an osmotic shock method, and thecorresponding flavin contents were normalized to whole-cell total protein levels. Data are presented as means � standard deviations (n � 3).

Harnessing Bacterial Periplasm for Biosynthesis Applied and Environmental Microbiology

January 2018 Volume 84 Issue 1 e01693-17 aem.asm.org 3

on February 2, 2018 by N

AN

YA

NG

TE

CH

NO

LOG

ICA

L UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

SO0702 (32) and the periplasmic 5=-nucleotidase UshA (33), originating from She-wanella oneidensis MR-1, into E. coli A5�tet. The plasmid with SO0702 and ushA genesincorporated into the tet vector is referred to as B2 (Fig. 2a). Recombinant E. colicontaining the heterologous gene ushA exhibited a strong enzyme activity for FADhydrolysis in its periplasm, confirming that the introduced UshA was located in theperiplasm of the E. coli host (see Fig. S2 in the supplemental material) (33). Comparingthe intracellular flavin pattern of recombinant E. coli A4�B2 with that of A4�tet (Fig.2b), no significant changes in flavin content or profile were revealed, confirming thatthe heterologous IM FAD exporter SO0702 was highly specific for FAD. Compared withA4�tet, A4�B2 showed a relatively lower level of RF in the periplasmic space (as shownin Fig. 2c), suggesting that the expression of SO0702 in A4�B2 possibly caused adecrease in RF export from the cytoplasm to the periplasm. When the pathwayregulating the intracellular flavin pool was combined with this periplasmic catalyticmachinery, the intracellular flavin content of E. coli A5�B2 was further elevated by104.5%, in comparison with that of E. coli A5�tet, and was greatly dominated by FMN.Moreover, the significant increase in the intracellular flavin content was mainly attrib-uted to the elevated FMN level in the periplasmic space of E. coli A5�B2 (Fig. 2c). Thisindicates that the cytoplasmic FAD produced by RibADEH and RibC is efficientlyexported by SO0702 into the periplasm, where it is catalyzed to FMN by the down-stream enzyme UshA in E. coli A5�B2. Diffusion of FMN across the OM of E. coli A5�B2could possibly be the rate-limiting step for FMN overproduction, as the ratio ofperiplasmic FMN to excreted FMN in the spent medium of A5�B2 was significantlylower than that of periplasmic RF to excreted RF for A4�tet (Fig. 2c and 3).

FMN-producing performance of the recombinant strain. To evaluate the perfor-mance of our recombinant E. coli harboring this novel FMN synthesis pathway, all of theE. coli strains were aerobically cultured for 48 h in shake flasks, in modified M9 mediumcontaining 15 g/liter xylose, and their FMN productivities were compared (Fig. 3). ForE. coli lac�tet, only 2.1 � 0.3 �M flavins were produced, with 24.3% being FMN. Afterintroduction of the RF synthesis pathway, 215.1 � 6.8 �M flavins were excreted into themedium, with 7.2% being FMN. For A4�B2, due to the low cytoplasmic FAD level, therewas little increase in FMN production. After introduction of RibC to shift the cytoplasmicflavin pool to FAD and FMN, the percentage of FMN in the total extracellular flavin poolof E. coli A5�tet was increased to 83.4%, while the overall flavin concentration wasseverely decreased by 53.7%, compared with that of E. coli A4�tet. This result indicatesthat the FMN excretion efficiency of the intrinsic IM FMN exporter (34) in E. coli is low.When both the regulated cytoplasmic flavin profile and the periplasmic catalyticmachinery were present in E. coli A5�B2, the highest FMN yield and purity wereachieved (Fig. 3 and Table 1). The extracellular FMN fraction of the final strain A5�B2

FIG 3 Concentration and purity of FMN overproduced by different engineered E. coli strains. The totalflavin concentrations in the culture medium were assayed after aerobic growth for 48 h in modified M9medium containing 15 g/liter xylose and are shown on the left y axis. The mass percentages of FMNamong total flavins in the culture medium are displayed on the right y axis. Data are presented asmeans � standard deviations (n � 3).

Yang et al. Applied and Environmental Microbiology

January 2018 Volume 84 Issue 1 e01693-17 aem.asm.org 4

on February 2, 2018 by N

AN

YA

NG

TE

CH

NO

LOG

ICA

L UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

was as high as 92.4%, and 148.0 � 1.4 �M FMN, of the 162.7 � 1.4 �M total flavins, wasproduced. These results demonstrate that our design was successful in achievingmicrobial production of high-purity FMN, and we achieved the highest titer of FMN inE. coli that has yet been reported.

Influence of the synthetic FMN pathway on the host cells. In our novel FMNbiosynthesis strategy, the final catalytic step for FMN generation is compartmentalizedinto the periplasm, in order to enhance selectivity in chemical synthesis and productsecretion. The influences of such methods on the cell growth and energy metabolismwere further investigated. As shown in Fig. 4a, the growth rate of the recombinantA5�B2 strain was slightly lower than that of the control lac�tet strain, which isreasonable, considering the greater growth burden brought by the synthetic FMNbiosynthesis pathway (35). In addition, no obvious differences in the xylose consump-tion rates of the FMN-producing strain and wild-type E. coli were observed (Fig. 4b). Asa result of energy dissipation in the biosynthesis of RF or FMN, both A4�tet and A5�B2strains showed decreased levels of intracellular ATP, in comparison with that of thewild-type E. coli. However, the intracellular ATP level of A5�B2 was comparable to thatof A4�tet, indicating that the IM FAD exporter and periplasmic catalytic machinery hadno significant impact on cell respiration (Fig. 4c). These results demonstrate that thisnew FMN biosynthesis approach is superior in selectivity with no significant sacrifice incell growth or respiration of the recombinant cells.

Conclusion. Our work demonstrates microbial production of highly pure FMN,which establishes a platform capable of being easily applied to various industrialRF-producing bacteria to develop more efficient producers for high-purity FMN. Inaddition, the periplasm is engineered to catalyze the final biosynthesis step of a specificchemical to avoid the generation of by-products. It should be pointed out that such adesign concept, which considers the optimal spatial organization of synthetic path-

TABLE 1 FMN productivities of different engineered E. coli strains

StrainFMN concentration(mg/liter)a % FMNb Yield (mg/g)c

Specific productivity(mg FMN/g/h)d

lac�tet 0.2 � 0.1 24.3 � 5.2 0.02 � 0.01 0.01 � 0.00A4�tet 6.1 � 0.1 7.2 � 0.2 0.47 � 0.02 0.13 � 0.00A4�B2 9.5 � 0.1 15.2 � 0.2 0.70 � 0.01 0.21 � 0.00A5�tet 39.6 � 1.1 83.4 � 0.5 2.64 � 0.07 0.86 � 0.04A5�B2 70.8 � 0.6 92.4 � 0.1 4.74 � 0.06 1.54 � 0.01aFMN concentration in the spent medium.bMass percentage of FMN among total flavins in the culture medium.cMilligrams of FMN per gram of consumed xylose.dMilligrams of FMN per gram (dry weight) of cells per hour.

FIG 4 Impact of the synthetic FMN biosynthesis pathway on cell growth and respiration. (a and b) Time course profiles of biomass accumulation (a) and xyloseconsumption (b) of wild-type and recombinant E. coli strains. (c) Intracellular ATP levels of different E. coli strains. Data are presented as means � standarddeviations (n � 3).

Harnessing Bacterial Periplasm for Biosynthesis Applied and Environmental Microbiology

January 2018 Volume 84 Issue 1 e01693-17 aem.asm.org 5

on February 2, 2018 by N

AN

YA

NG

TE

CH

NO

LOG

ICA

L UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

ways, not only is powerful in solving the technical hurdles encountered during theconstruction of high-purity FMN producers but may also greatly facilitate the devel-opment of microbial factories for other high-value chemicals.

MATERIALS AND METHODSIn vitro gene synthesis. The ribADEH genes from B. subtilis, with optimized codons for expression in

E. coli, were developed in our previous work (31). The ribC gene from B. subtilis and SO0702 and ushAfrom S. oneidensis were synthesized in this work with similar methods. All of the genes with codonsadapted for expression in E. coli were synthesized as biobricks, and each biobrick was flanked by anupstream prefix (containing EcoRI and XbaI sites), a ribosome-binding site (BBa_B0034; iGEM) 6 bp aheadof the start codon, and a downstream suffix (containing SpeI and SbfI sites). The primer sets for in vitrogene synthesis of ribC, SO0702, and ushA are presented in Table 2.

Plasmid construction. All plasmid constructions were implemented in E. coli DH5�. The ribA, ribD,ribE, and ribH biobricks were inserted into the vector pSB3C5 sequentially, two Plac promoters(BBa_R0011; iGEM) were placed ahead of ribA and ribE, and the resulting expression plasmid was namedA4. A synthesized ribC biobrick was then incorporated into A4 downstream of ribH, forming theexpression plasmid A5. The control vector with only one Plac promoter inserted into pSB3C5 was namedthe lac vector. One Ptet promoter (BBa_R0040; iGEM) and synthesized SO0702 and ushA were insertedinto pSB4K5 (iGEM) sequentially, forming the plasmid B2. The empty vector with one Ptet promoter inpSB4K5 was named the tet vector. Double plasmids derived from pSB3C5 and pSB4K5 were electropo-rated into the host E. coli BL21-Gold (DE3). E. coli strains were cultured in LB medium at 37°C at 200 rpm.Whenever needed, 50 �g/ml kanamycin or 34 �g/ml chloramphenicol was added to the culture medium.

Culture conditions for FMN production. E. coli strains from �80°C freezer stocks were inoculatedinto 20 ml of LB broth supplemented with 50 �g/ml kanamycin and 34 �g/ml chloramphenicol andgrown aerobically overnight at 37°C, with shaking. Then 0.4 ml E. coli suspension was centrifuged andresuspended in 10 ml of modified M9 medium (12 g/liter Na2HPO4, 6 g/liter KH2PO4, 0.5 g/liter NaCl, 1g/liter NH4Cl, 0.24 g/liter MgSO4, 11.1 mg/liter CaCl2, 2 g/liter urea) supplemented with 15 g/liter xylose,1% vitamin stock solution (36), 1% mineral stock solution (36), 1% amino acid stock solution (36), 50�g/ml kanamycin, 34 �g/ml chloramphenicol, 0.01 mM isopropyl-�-D-thiogalactoside (IPTG), and 10 nManhydrotetracycline (aTc), in a 100-ml shake flask. Xylose was used as the carbon source because ribulose5-phosphate, an intermediate of xylose isomerization, is one of the precursors for flavin synthesis, whichfacilitates flavin production. After growth at 37°C for 48 h at 200 rpm, the supernatant was subjected toflavin and xylose measurement, and the cell pellet was collected for extraction of intracellular flavin andperiplasmic flavin. The biomass concentration was indicated as the total protein concentration.

Extraction of intracellular and periplasmic flavin. The intracellular flavin pool was isolated asdescribed previously (37). Briefly, cell pellets from 25-ml cultures were harvested by centrifugation(5,000 � g for 20 min at 4°C), resuspended in 3 ml of chilled HEPES buffer (100 mM [pH 7.4]), split intoaliquots of 1.0 ml in 1.5-ml tubes, and centrifuged again (10,000 � g for 3 min at 4°C). Then the cell pelletwas resuspended in 100 �l of GES buffer (5.0 M guanidium thiocyanate [GTC], 0.1 M EDTA, 0.5% Sarkosyl[pH 8.0]) and vortex-mixed for 10 s. Finally, the volume of the sample with intracellular extract wasadjusted to 1.0 ml by adding 900 �l of precooled Milli-Q water. Periplasmic flavin content was extractedaccording to previous reports (38, 39). Briefly, cells harvested from 25-ml cultures (by centrifugation at5,000 � g for 20 min at 4°C) were resuspended in 3 ml of osmolysis buffer (18% sucrose, 2.5 mM EDTA,100 mM Tris-HCl [pH 7.8]). The suspension was incubated gently at 50 rpm for 10 min, and the cells wereharvested by centrifugation (13,000 � g for 10 min at 4°C). The supernatant was discarded, and the cellpellet was resuspended in 3 ml of ice-cold Milli-Q water. The cell suspension was incubated in an ice bathfor 10 min at 50 rpm, followed by centrifugation at 13,000 � g for 10 min at 4°C. The supernatantcontained the periplasmic fraction of E. coli cells. For ease of comparison, both the intracellular andperiplasmic flavin levels were normalized to the protein concentrations of whole cells.

FAD hydrolysis assay. Subcellular fractions for enzyme activity assays were isolated and subjectedto FAD hydrolysis assays as reported previously (33). Briefly, cell fractions isolated with the lysozymetreatment method were aliquoted into a 96-well plate. The plate was prewarmed to 37°C, after whichwarm (37°C) FAD was added to each assay mixture to a final concentration of 115 �M. Fluorescence(excitation at 440 nm and emission at 525 nm) was recorded immediately and at 5-min intervals at 37°Cuntil no obvious increase in the fluorescence signals of all samples could be observed. The plate wasgently shaken for 3 s before each data acquisition. Assuming that 1 mol of FAD was hydrolyzed to 1 molof FMN, the increase in the fluorescence signal in response to FAD hydrolysis could be calculated basedon standard solutions of FAD and FMN. The rate of FAD hydrolysis was determined by the slope of theinitial linear segment of the fluorescence curve, which was normalized to the protein concentration ofthe sample.

Quantification of intracellular ATP contents. The intracellular ATP levels were determined by usinga BacTiter-Glo microbial cell viability assay kit (Promega), following the manufacturer’s instructions.

Flavin and xylose measurements. The cell lysate and supernatant samples for flavin quantificationwere filtered and measured using an ultra-performance liquid chromatography (UPLC) system (Shi-madzu, Japan) equipped with a reverse-phase C18 column (10 cm by 2.1 mm, 5 �m; Ascentis) and aSPD-M20A diode array detector. A mixture of methanol and 0.05 M ammonium acetate (pH 6.0) wasapplied as the mobile phase, with controlled ratios as follows. A linear gradient from 25:75 to 70:30(vol/vol) methanol/0.05 M ammonium acetate within 8 min was followed by isocratic elution at 70:30(vol/vol) to 11 min and isocratic elution at 25:75 (vol/vol) to 16 min, at 30°C with a flow rate of 0.2 ml/min.

Yang et al. Applied and Environmental Microbiology

January 2018 Volume 84 Issue 1 e01693-17 aem.asm.org 6

on February 2, 2018 by N

AN

YA

NG

TE

CH

NO

LOG

ICA

L UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

TABLE 2 Oligonucleotide sets for in vitro gene synthesis of ribC, SO0702, and ushA

Gene Primer name Primer sequencea

ribC (1,018 bp)ribC-BB01 (529 bp) ribC-01-Pri_For CGGAATTCGCGGCCGC

ribC-01-Pri_Rev TGTCCTGTTCGGTCAGTTTTTCAACribC-01_F1 CGGAATTCGCGGCCGCTTctagagaaagaggagaaaactagagtgaaaaccribC-01_R1 TGCGGGTGGGTGATGTGGATggttttcactctagttttctcctctttctctagribC-01_F2 ATCCACATCACCCACCCGCAccacctgatcaaagaagaacaggctaaaribC-01_R2 TCGAAGTAACCCAGAGCCATAACAGAtttagcctgttcttctttgatcaggtggribC-01_F3 TCTGTTATGGCTCTGGGTTACTTCGAcggtgttcacctgggtcaccaribC-01_R3 CGATCTGTTTAGCGGTACCGATAACTTTCtggtgacccaggtgaacaccgribC-01_F4 GAAAGTTATCGGTACCGCTAAACAGATCGctgaagaaaaaggtctgaccctggcribC-01_R4 GGTGCGGGTGGAAGGTCATAACAgccagggtcagacctttttcttcagribC-01_F5 TGTTATGACCTTCCACCCGCACCcgtctcacgttctgggtcgtgaribC-01_R5 GGGTGATCAGGTCTTTCGGTTCTTTGtcacgacccagaacgtgagacgribC-01_F6 CAAAGAACCGAAAGACCTGATCACCCcgctggaagacaaaatcaaccagatcgribC-01_R6 ACAGAACTTCGGTACCCAGCTGTTcgatctggttgattttgtcttccagcgribC-01_F7 AACAGCTGGGTACCGAAGTTCTGTacgttgttaaattcaacgaagttttcgcttribC-01_R7 GTCGATGAACTGTTTCGGAGACAGAGaagcgaaaacttcgttgaatttaacaacgtribC-01_F8 CTCTGTCTCCGAAACAGTTCATCGACcagtacatcatcggtctgaacgttcagribC-01_R8 AAGTCGAAACCAGCAACAGCGTGctgaacgttcagaccgatgatgtactgribC-01_F9 CACGCTGTTGCTGGTTTCGACTTcacctacggtaaatacggtaaaggtaccatribC-01_R9 CAGGTCGTCCGGCATGGTTTTCatggtacctttaccgtatttaccgtaggtgribC-01_F10 GAAAACCATGCCGGACGACCTGgacggtaaagctggttgcaccatribC-01_R10 TGTCCTGTTCGGTCAGTTTTTCAACCatggtgcaaccagctttaccgtc

ribC-BB02 (529 bp) ribC-02-Pri_For GCTGGTTGCACCATGGTTGAribC-02-Pri_Rev CGCCTGCAGGCGGCribC-02_F1 GCTGGTTGCACCATGGTTGAAAAACtgaccgaacaggacaaaaaaatctcttctribC-02_R1 TGCAGAGCGGTACGGATGTAAGAagaagagatttttttgtcctgttcggtcaribC-02_F2 TCTTACATCCGTACCGCTCTGCAaaacggtgacgttgaactggctaacribC-02_R2 AGTACGGCTGACCCAGCAGAACgttagccagttcaacgtcaccgtttribC-02_F3 GTTCTGCTGGGTCAGCCGTACTtcatcaaaggtatcgttatccacggtgaribC-02_R3 GAAACCGATGGTACGACCACGTTTGtcaccgtggataacgatacctttgatgaribC-02_F4 CAAACGTGGTCGTACCATCGGTTTCccgaccgctaacgttggtctgaribC-02_R4 TCGGCGGAACGATGTAAGAGTTGTtcagaccaacgttagcggtcggribC-02_F5 ACAACTCTTACATCGTTCCGCCGAccggtgtttacgctgttaaagctgaaribC-02_R5 CAAACACCGTTGTAAACTTCACCGTTAACttcagctttaacagcgtaaacaccggribC-02_F6 GTTAACGGTGAAGTTTACAACGGTGTTTGcaacatcggttacaaaccgaccttctacribC-02_R6 ACGGCTGTTCCGGACGTTTTTCgtagaaggtcggtttgtaaccgatgttgribC-02_F7 GAAAAACGTCCGGAACAGCCGTctatcgaagttaacctgttcgacttcaaccribC-02_R7 GATTTTGATAGCAGCACCGTAAACTTCCTggttgaagtcgaacaggttaacttcgatagribC-02_F8 AGGAAGTTTACGGTGCTGCTATCAAAATCgaatggtacaaacgtatccgttctgaacgribC-02_R8 GTTCGGTCAGTTCTTTGATACCGTTGAATTTAcgttcagaacggatacgtttgtaccattcribC-02_F9 TAAATTCAACGGTATCAAAGAACTGACCGAACagatcgaaaaagacaaacaggaagctatccribC-02_R9 GCTACTAGTATTATTTACGCAGGTTAGAGAAGTAACggatagcttcctgtttgtctttttcgatctribC-02_F10 GTTACTTCTCTAACCTGCGTAAATAATACTAGTAGCggccgcctgcaggcgribC-02_R10 cgcctgcaggcggcc

SO0702 (1,426 bp)SO0702-BB01 (733 bp) SO0702-01-Pri_For CGGAATTCGCGGCCGC

SO0702-01-Pri_Rev TGTCGAAAGCAGCACGTTCCSO0702-01_F1 CGGAATTCGCGGCCGCTTctagagaaagaggagaaactgcatatgaaagaSO0702-01_R1 CAGACAGCAGACCGTGACGGtctttcatatgcagtttctcctctttctctagSO0702-01_F2 CCGTCACGGTCTGCTGTCTGctccgatcggtcgtgttctgcSO0702-01_R2 CAGGTTCGGCAGAGACATGTTCAgcagaacacgaccgatcggagSO0702-01_F3 TGAACATGTCTCTGCCGAACCTGatcggtatcatgaccatcctgggttSO0702-01_R3 ATGAAGAAGGTGTCAGCCAGAGAGAaacccaggatggtcatgataccgatSO0702-01_F4 TCTCTCTGGCTGACACCTTCTTCATctctcagctgggtaccgaagcSO0702-01_R4 GAAGGTGAAAGAGATAGCAGCCAGAgcttcggtacccagctgagagSO0702-01_F5 TCTGGCTGCTATCTCTTTCACCTTCccggttaccctgatcatctcttctatcSO0702-01_R5 ACCAGCACCAACACCGATAGCgatagaagagatgatcagggtaaccggSO0702-01_F6 GCTATCGGTGTTGGTGCTGGTgtttctaccaacctgggtcgtctSO0702-01_R6 CGGAGCGTTACCAGAACCGATCagacgacccaggttggtagaaacSO0702-01_F7 GATCGGTTCTGGTAACGCTCCGcaggctaaagttttcctgcacgacSO0702-01_R7 GGATGAAGGTCAGCAGCAGAGCgtcgtgcaggaaaactttagcctgSO0702-01_F8 GCTCTGCTGCTGACCTTCATCCtgatcgcttctctgtctgctctgSO0702-01_R8 ACAGCGGTTCGATGAAGATAGAACCcagagcagacagagaagcgatcaSO0702-01_F9 GGTTCTATCTTCATCGAACCGCTGTtctctctgctgggtgctaacga

(Continued on next page)

Harnessing Bacterial Periplasm for Biosynthesis Applied and Environmental Microbiology

January 2018 Volume 84 Issue 1 e01693-17 aem.asm.org 7

on February 2, 2018 by N

AN

YA

NG

TE

CH

NO

LOG

ICA

L UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

TABLE 2 (Continued)

Gene Primer name Primer sequencea

SO0702-01_R9 TGGATCAGCGGCAGAGAGGTTtcgttagcacccagcagagagaSO0702-01_F10 AACCTCTCTGCCGCTGATCCAcgactacatgatgtactggtacgttggSO0702-01_R10 CAGAACCAGCAGCGGAGCAccaacgtaccagtacatcatgtagtcgSO0702-01_F11 TGCTCCGCTGCTGGTTCTGctgatggttggtaaccagggtctgSO0702-01_R11 CGGGTGTCACCGGTAGAACGcagaccctggttaccaaccatcagSO0702-01_F12 CGTTCTACCGGTGACACCCGttctccggctatgatcatgaccctSO0702-01_R12 AGGATCAGGTTGATGATAGCAGCCagggtcatgatcatagccggagaaSO0702-01_F13 GGCTGCTATCATCAACCTGATCCTggacccgctgctgatcttcgSO0702-01_R13 CGCGGGAACGGACCGATACcgaagatcagcagcgggtccSO0702-01_F14 GTATCGGTCCGTTCCCGCGtctggaaatccagggtgctgcSO0702-01_R14 AGCCAAGAGAACAGGGTAGCGATAgcagcaccctggatttccagaSO0702-01_F15 TATCGCTACCCTGTTCTCTTGGCTggttgctctgtctctgtctggttSO0702-01_R15 CATACGGTGTTTGATGATCAGCAGGTaaccagacagagacagagcaaccSO0702-01_F16 ACCTGCTGATCATCAAACACCGTATGctggaacgtgctgctttcgacaSO0702-01_R16 tgtcgaaagcagcacgttccag

SO0702-BB02 (733 bp) SO0702-02-Pri_For ATCATCAAACACCGTATGCTGGAASO0702-02-Pri_Rev CGCCTGCAGGCGGCSO0702-02_F1 ATCATCAAACACCGTATGCTGGAACGtgctgctttcgacatcgaccgSO0702-02_R1 CCAGTTTAGACCAGTTAGCACGCATAcggtcgatgtcgaaagcagcaSO0702-02_F2 TATGCGTGCTAACTGGTCTAAACTGGctcacatcgctcagccggcSO0702-02_R2 CGGGTTGATCAGGTTCATCAGAGCAgccggctgagcgatgtgagSO0702-02_F3 TGCTCTGATGAACCTGATCAACCCGctggctaacgctgttatcatggctSO0702-02_R3 GAGTGGTCGATGTGAGCCAGCATagccatgataacagcgttagccagSO0702-02_F4 ATGCTGGCTCACATCGACCACTCtgctgttgctgctttcggtgcSO0702-02_R4 AGAACAGATTCCAGACGGGTACCAgcaccgaaagcagcaacagcaSO0702-02_F5 TGGTACCCGTCTGGAATCTGTTCTgctgatcgttgttatggctctgtctSO0702-02_R5 GAGCGATGAACGGCATCAGAGAAGAagacagagccataacaacgatcagcSO0702-02_F6 TCTTCTCTGATGCCGTTCATCGCTCagaacctgggtgctggtcagcSO0702-02_R6 AGCCTGTTTAGCACGCTGCGgctgaccagcacccaggttctSO0702-02_F7 CGCAGCGTGCTAAACAGGCTctgctgctgtctctgaaattcatcctSO0702-02_R7 GATGTACAGCAGGGTCTGGAAAACCaggatgaatttcagagacagcagcagSO0702-02_F8 GGTTTTCCAGACCCTGCTGTACATCccgctggctttcttcgctcaSO0702-02_R8 GAGAACAGAGAAGCCAGCGGCtgagcgaagaaagccagcggSO0702-02_F9 GCCGCTGGCTTCTCTGTTCTCtaccgacccgcaggttctggSO0702-02_R9 AGAACCAGGATGTAGAAAGACAGCCATTccagaacctgcgggtcggtaSO0702-02_F10 AATGGCTGTCTTTCTACATCCTGGTTCTgccgtgcgcttacggtccSO0702-02_R10 GTAGCGAAGATGATAACGATACCCAGCggaccgtaagcgcacggcSO0702-02_F11 GCTGGGTATCGTTATCATCTTCGCTACcgctctgaacgcttaccaccgSO0702-02_R11 AGGTTGATAACCAGAGAAGACATCGGAcggtggtaagcgttcagagcgSO0702-02_F12 TCCGATGTCTTCTCTGGTTATCAACCTgtgccgtctggttctgctgatgSO0702-02_R12 CAGAGCAGCCAGCGGCAGcatcagcagaaccagacggcacSO0702-02_F13 CTGCCGCTGGCTGCTCTGggttcttacatcgacggtgttaaaggtSO0702-02_R13 CGGCAGAGCCAGCAGCAGacctttaacaccgtcgatgtaagaaccSO0702-02_F14 CTGCTGCTGGCTCTGCCGatcaccaacctgctgatgggtatcgSO0702-02_R14 ACGCTGAGCCAGGTAGTAGCAAGcgatacccatcagcaggttggtgatSO0702-02_F15 CTTGCTACTACCTGGCTCAGCGTatctgcgaaccggttaaagctaccaSO0702-02_R15 GCTACTAGTATTACAGGGTGTCAGCGGtggtagctttaaccggttcgcagatSO0702-02_F16 CCGCTGACACCCTGTAATACTAGTAGCggccgcctgcaggcgSO0702-02_R16 cgcctgcaggcggcc

ushA (1,777 bp)ushA-BB01 (619 bp) ushA-Pri_For CGGAATTCGCGGCCGC

ushA-Pri_Rev CAGAGATGTATTCCGGGTTACCGAushA-01_F1 CGGAATTCGCGGCCGCttctagagaaagaggagaaatctagtatgaccushA-01_R1 CGATCAGACCTTTGATCAGCATGTTggtcatactagatttctcctctttctctagaaushA-01_F2 AACATGCTGATCAAAGGTCTGATCGctaccgctgttctgaccgctushA-01_R2 GTCAGAGTTGCAACCAGCCAGagcggtcagaacagcggtagushA-01_F3 CTGGCTGGTTGCAACTCTGACgacgacaaagttccgaccaccushA-01_R3 AGCACCAGCTTCAGCGCAggtggtcggaactttgtcgtcushA-01_F4 TGCGCTGAAGCTGGTGCTgcttgcaaaaccttcaccatcctgushA-01_R4 CCGTGGTTGTCGTTGGTGTGcaggatggtgaaggttttgcaagcushA-01_F5 CACACCAACGACAACCACGGtcgtttctgggaaaacaaagacggushA-01_R5 TGAGCAGCCAGACCGTATTCAccgtctttgttttcccagaaacgaushA-01_F6 TGAATACGGTCTGGCTGCTCAgaaaaccctggttgaccagatccushA-01_R6 GACCACCTTTTTTAGAAACTTCAGCACggatctggtcaaccagggttttcushA-01_F7 GTGCTGAAGTTTCTAAAAAAGGTGGTCagaccctgctgctgtctgg

(Continued on next page)

Yang et al. Applied and Environmental Microbiology

January 2018 Volume 84 Issue 1 e01693-17 aem.asm.org 8

on February 2, 2018 by N

AN

YA

NG

TE

CH

NO

LOG

ICA

L UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

TABLE 2 (Continued)

Gene Primer name Primer sequencea

ushA-01_R7 ACACCGGTGTTGATGTCACCAccagacagcagcagggtctushA-01_F8 TGGTGACATCAACACCGGTGTtccggaatctgacctgcaagacushA-01_R8 CCGGTGAAGTCCGGGATAGCgtcttgcaggtcagattccggaushA-01_F9 GCTATCCCGGACTTCACCGGtatgaacaaaatcggttacgacgctatgushA-01_R9 TGTCAAATTCGTGGTTACCAACAGCcatagcgtcgtaaccgattttgttcataushA-01_F10 GCTGTTGGTAACCACGAATTTGACAacccgctgtctgttgttgacatushA-01_R10 GAAATTCAGCCAGAGAACGCTGCatgtcaacaacagacagcgggtushA-01_F11 GCAGCGTTCTCTGGCTGAATTTCcgatgctggctgctaacatctacaushA-01_R11 GAAGTAACGGGTACCGTCAGCTTtgtagatgttagcagccagcatcgushA-01_F12 AAGCTGACGGTACCCGTTACTTCgacgcttacaaaatcttcgacgttaacushA-01_R12 CAGACCGATAACAGCGATTTTAACACCgttaacgtcgaagattttgtaagcgtcushA-01_F13 GGTGTTAAAATCGCTGTTATCGGTCTGaccaccgaagacaccgctaaaushA-01_R13 CAGAGATGTATTCCGGGTTACCGATtttagcggtgtcttcggtggt

ushA-BB02 (619 bp) ushA-Pri_For GAAGACACCGCTAAAATCGGTAACCushA-Pri_Rev GAACGTTAGCACGTTCACCGTCushA-02_F1 GAAGACACCGCTAAAATCGGTAACCCggaatacatctctgaactggaatttcgtgacushA-02_R1 TTCTTTGATAACGTTAGCAACTTCTTTTTTCGGgtcacgaaattccagttcagagatgtattccushA-02_F2 CCGAAAAAAGAAGTTGCTAACGTTATCAAAGAAatcaaagaagctaaatctgctgacatcatcttushA-02_R2 GTGACCCATGTGGGTAACAGCGaagatgatgtcagcagatttagcttctttgatushA-02_F3 CGCTGTTACCCACATGGGTCACtacgctgacggtcagaacggttcushA-02_R3 GCAACGTCACCCGGAGCGTTAgaaccgttctgaccgtcagcgtaushA-02_F4 TAACGCTCCGGGTGACGTTGCtctggctcgttctctgaaagaaggtushA-02_R4 GACCACCGATAACAACTTGCAGGTCaccttctttcagagaacgagccagaushA-02_F5 GACCTGCAAGTTGTTATCGGTGGTCactctcagaacccggtttgcatggushA-02_R5 CAGCGTAAGCTTTGTTACCCGGTTccatgcaaaccgggttctgagagtushA-02_F6 AACCGGGTAACAAAGCTTACGCTGctttcaaaccgggtgacgcttgcushA-02_R6 CCGTTCTGCTGGTCCGGAGCgcaagcgtcacccggtttgaaagushA-02_F7 GCTCCGGACCAGCAGAACGGtacctggatcatgcaggctcacgushA-02_R7 AGCACGACCAACGTATTTACCCCATTcgtgagcctgcatgatccaggtaushA-02_F8 AATGGGGTAAATACGTTGGTCGTGCTgacttcgaatacttcaacggtgaactgcushA-02_R8 CGGAACCAGTTTGTAAGAAGCCAGGTgcagttcaccgttgaagtattcgaagtcushA-02_F9 ACCTGGCTTCTTACAAACTGGTTCCGgttaacctggttaaagaagttaccgacgaaushA-02_R9 ACCAACCAGAACTTTTTTTTTGTTACCAGCttcgtcggtaacttctttaaccaggttaacushA-02_F10 GCTGGTAACAAAAAAAAAGTTCTGGTTGGTgaaaaaatcgaaccggacaccgaactushA-02_R10 TTTTCCTGGTAGTAAGACAGCAGTTCTTTCagttcggtgtccggttcgattttttcushA-02_F11 GAAAGAACTGCTGTCTTACTACCAGGAAAAaggtcaggctaaactggacgaagttatushA-02_R11 GCAGAGCGTCGGTGGTAGCGataacttcgtccagtttagcctgacctushA-02_F12 CGCTACCACCGACGCTCTGCtggacggtgaacgtgctaacgttcushA-02_R12 gaacgttagcacgttcaccgtcca

ushA-BB03 (619 bp) ushA-Pri_For ACCACCGACGCTCTGCTushA-Pri_Rev CGCCTGCAGGCGGCushA-03_F1 ACCACCGACGCTCTGCTGGacggtgaacgtgctaacgttcgushA-03_R1 ACGACCCAGGTTGGTCTGTTTGTTAcgaacgttagcacgttcaccgtushA-03_F2 TAACAAACAGACCAACCTGGGTCGTatgctggctatggctcagtctggushA-03_R2 TTCATAACACCGAAGTCAGCAGAAACTTTAccagactgagccatagccagcatushA-03_F3 TAAAGTTTCTGCTGACTTCGGTGTTATGAActctggtggtgttcgtgcttctatcushA-03_R3 TCACGGTAGGTGATGTTACCAGCTTTgatagaagcacgaacaccaccagagushA-03_F4 AAAGCTGGTAACATCACCTACCGTGAcgttctgaccgttcagccgttcushA-03_R4 GTCATTTCGTTCAGGGTAACCATGTTACCgaacggctgaacggtcagaacgushA-03_F5 GGTAACATGGTTACCCTGAACGAAATGACcggtgctgctgctgctgacushA-03_R5 CAGAGAACCAACAGCACCCAGGTAgtcagcagcagcagcaccgushA-03_F6 TACCTGGGTGCTGTTGGTTCTCTGcaaatcggttctggtggttacgctushA-03_R6 ACGGTCATTTTAACACCGGTGATCTGagcgtaaccaccagaaccgatttgushA-03_F7 CAGATCACCGGTGTTAAAATGACCGTtgactgcgttgctaaaaaagctaccgushA-03_R7 CAGAGAAAGCTTTACCGTTGATTTCGTGAAcggtagcttttttagcaacgcagtcaushA-03_F8 TTCACGAAATCAACGGTAAAGCTTTCTCTGctaccgctacctacaaattcaccgttushA-03_R8 CACCAGCAGCGTTGAAAGACGGaacggtgaatttgtaggtagcggtagushA-03_F9 CCGTCTTTCAACGCTGCTGGTGgtgacggttacccgaaactggtttushA-03_R9 ACGTAACCGGTCTGGATCGGAGaaaccagtttcgggtaaccgtcacushA-03_F10 CTCCGATCCAGACCGGTTACGTtgacgctgacctgctgtacaccushA-03_R10 CAGCAACGATAGACTGTTTTTCTTTCAGGAAggtgtacagcaggtcagcgtcaushA-03_F11 TTCCTGAAAGAAAAACAGTCTATCGTTGCTGctgactacaacccggttggtgacushA-03_R11 CTTCAACAGAGTCAGAGTTTTCGTAAACGATgtcaccaaccgggttgtagtcagushA-03_F12 ATCGTTTACGAAAACTCTGACTCTGTTGAAGgttgcaaaatcaccgctaaataatactagtagcushA-03_R12 CGCCTGCAGGCGGCCgctactagtattatttagcggtgattttgcaac

aSegments shown in lowercase letters are complementary to those of the paired primer.

Harnessing Bacterial Periplasm for Biosynthesis Applied and Environmental Microbiology

January 2018 Volume 84 Issue 1 e01693-17 aem.asm.org 9

on February 2, 2018 by N

AN

YA

NG

TE

CH

NO

LOG

ICA

L UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

The elution times for FAD, FMN, and RF were 3.5 min, 4.6 min, and 6.1 min, respectively. Signals at 267nm were used for calculations. For xylose quantification, sulfuric acid (5 mM) was utilized as the mobilephase, flowing at 0.6 ml/min through an Aminex HPX-87H column (Bio-Rad) at 50°C. The elution time forxylose was 9.3 min. The xylose concentration was calculated based on the absorbance signals at 190 nmfor the samples and standard xylose solutions (31).

SUPPLEMENTAL MATERIAL

Supplemental material for this article may be found at https://doi.org/10.1128/AEM.01693-17.

SUPPLEMENTAL FILE 1, PDF file, 0.2 MB.

ACKNOWLEDGMENTSWe thank Yinjin Yuan and Hao Song (Tianjin University) for help in gene synthesis.

We also thank Norazean Zaiden for assistance in ultra-performance liquid chromatog-raphy (UPLC) analyses.

This research was supported by the China Postdoctoral Science Foundation(grant 2016M591043), the Ministry of Education (MOE) Academic Research FundTier 1 (grant M4011622.030), the National Research Foundation, and MOE Singaporeunder its Research Centre of Excellence Programme, Singapore Centre for Environ-mental Life Sciences Engineering, Nanyang Technological University (Singapore)(grant M4330005.C70).

We declare no conflicts of interest.

REFERENCES1. Lee SY, Kim HU. 2015. Systems strategies for developing industrial

microbial strains. Nat Biotechnol 33:1061–1072. https://doi.org/10.1038/nbt.3365.

2. Li JQ, Ballmer SG, Gillis EP, Fujii S, Schmidt MJ, Palazzolo AME, LehmannJW, Morehouse GF, Burke MD. 2015. Synthesis of many different types oforganic small molecules using one automated process. Science 347:1221–1226. https://doi.org/10.1126/science.aaa5414.

3. Wu SG, He L, Wang Q, Tang YJ. 2015. An ancient Chinese wisdom formetabolic engineering: yin-yang. Microb Cell Fact 14:39. https://doi.org/10.1186/s12934-015-0219-3.

4. Liu D, Wan N, Zhang F, Tang YJ, Wu SG. 2017. Enhancing fatty acidproduction in Escherichia coli by Vitreoscilla hemoglobin overexpression.Biotechnol Bioeng 114:463– 467. https://doi.org/10.1002/bit.26067.

5. Wu S, Chen Y, Xu Y, Li A, Xu Q, Glieder A, Li Z. 2014. Enantioselectivetrans-dihydroxylation of aryl olefins by cascade biocatalysis with recom-binant Escherichia coli coexpressing monooxygenase and epoxide hy-drolase. ACS Catal 4:409 – 420. https://doi.org/10.1021/cs400992z.

6. Dong X, Chen X, Qian Y, Wang Y, Wang L, Qiao W, Liu L. 2017. Metabolicengineering of Escherichia coli W3110 to produce L-malate. BiotechnolBioeng 114:656 – 664. https://doi.org/10.1002/bit.26190.

7. Foo JL, Susanto AV, Keasling JD, Leong SS, Chang MW. 2017. Whole-cellbiocatalytic and de novo production of alkanes from free fatty acids inSaccharomyces cerevisiae. Biotechnol Bioeng 114:232–237. https://doi.org/10.1002/bit.25920.

8. Chen GQ, Hajnal I. 2015. The ‘PHAome’. Trends Biotechnol 33:559 –564.https://doi.org/10.1016/j.tibtech.2015.07.006.

9. Wang Z, Cirino PC. 2016. New and improved tools and methods forenhanced biosynthesis of natural products in microorganisms. Curr OpinBiotechnol 42:159 –168. https://doi.org/10.1016/j.copbio.2016.05.003.

10. Lian J, Jin R, Zhao H. 2016. Construction of plasmids with tunable copynumbers in Saccharomyces cerevisiae and their applications in pathwayoptimization and multiplex genome integration. Biotechnol Bioeng 113:2462–2473. https://doi.org/10.1002/bit.26004.

11. Huang JF, Liu ZQ, Jin LQ, Tang XL, Shen ZY, Yin HH, Zheng YG. 2017.Metabolic engineering of Escherichia coli for microbial production ofL-methionine. Biotechnol Bioeng 114:843– 851. https://doi.org/10.1002/bit.26198.

12. Novy V, Brunner B, Muller G, Nidetzky B. 2017. Toward “homolactic” fer-mentation of glucose and xylose by engineered Saccharomyces cerevisiaeharboring a kinetically efficient L-lactate dehydrogenase within pdc1–pdc5deletion background. Biotechnol Bioeng 114:163–171. https://doi.org/10.1002/bit.26048.

13. Joosten V, van Berkel WJH. 2007. Flavoenzymes. Curr Opin Chem Biol11:195–202. https://doi.org/10.1016/j.cbpa.2007.01.010.

14. Medina M, Ferreira P, Martinez-Julvez M. 2013. Flavoproteins and fla-voenzymes with biomedical and therapeutic impact. Curr Pharm Des19:2497–2498. https://doi.org/10.2174/1381612811319140001.

15. Lee M, Hong J, Seo DH, Nam DH, Nam KT, Kang K, Park CB. 2013. Redoxcofactor from biological energy transduction as molecularly tunableenergy-storage compound. Angew Chem Int Ed Engl 52:8322– 8328.https://doi.org/10.1002/anie.201301850.

16. Orita A, Verde MG, Sakai M, Meng YS. 2016. A biomimetic redox flowbattery based on flavin mononucleotide. Nat Commun 7:13230. https://doi.org/10.1038/ncomms13230.

17. Okamoto A, Nakamura R, Nealson KH, Hashimoto K. 2014. Bound flavinmodel suggests similar electron-transfer mechanisms in Shewanella andGeobacter. ChemElectroChem 1:1808 –1812. https://doi.org/10.1002/celc.201402151.

18. Shi L, Rosso KM, Clarke TA, Richardson DJ, Zachara JM, Fredrickson JK. 2012.Molecular underpinnings of Fe(III) oxide reduction by Shewanella oneidensisMR-1. Front Microbiol 3:50. https://doi.org/10.3389/fmicb.2012.00050.

19. Pinchuk GE, Rodionov DA, Yang C, Li XQ, Osterman AL, Dervyn E,Geydebrekht OV, Reed SB, Romine MF, Collart FR, Scott JH, FredricksonJK, Beliaev AS. 2009. Genomic reconstruction of Shewanella oneidensisMR-1 metabolism reveals a previously uncharacterized machinery forlactate utilization. Proc Natl Acad Sci U S A 106:2874 –2879. https://doi.org/10.1073/pnas.0806798106.

20. Nielsen P, Rauschenbach P, Bacher A. 1983. Phosphates of riboflavin andriboflavin analogs: a reinvestigation by high-performance liquid chro-matography. Anal Biochem 130:359 –368. https://doi.org/10.1016/0003-2697(83)90600-0.

21. Abbas CA, Sibirny AA. 2011. Genetic control of biosynthesis and trans-port of riboflavin and flavin nucleotides and construction of robustbiotechnological producers. Microbiol Mol Biol Rev 75:321–360. https://doi.org/10.1128/MMBR.00030-10.

22. Scola-Nagelschneider G, Hemmerich P. 1976. Synthesis, separation, iden-tification and interconversion of riboflavin phosphates and their acetylderivatives: a reinvestigation. Eur J Biochem 66:567–577. https://doi.org/10.1111/j.1432-1033.1976.tb10583.x.

23. Nielsen P, Harksen J, Bacher A. 1985. Hydrolysis and rearrangement reac-tions of riboflavin phosphates: an explicit kinetic study. Eur J Biochem152:465–473. https://doi.org/10.1111/j.1432-1033.1985.tb09220.x.

24. Lee JW, Na D, Park JM, Lee J, Choi S, Lee SY. 2012. Systems metabolic

Yang et al. Applied and Environmental Microbiology

January 2018 Volume 84 Issue 1 e01693-17 aem.asm.org 10

on February 2, 2018 by N

AN

YA

NG

TE

CH

NO

LOG

ICA

L UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

engineering of microorganisms for natural and non-natural chemicals.Nat Chem Biol 8:536 –546. https://doi.org/10.1038/nchembio.970.

25. Yatsyshyn VY, Ishchuk OP, Voronovsky AY, Fedorovych DV, Sibirny AA.2009. Production of flavin mononucleotide by metabolically engineeredyeast Candida famata. Metab Eng 11:163–167. https://doi.org/10.1016/j.ymben.2009.01.004.

26. Yatsyshyn VY, Fedorovych DV, Sibirny AA. 2010. Medium optimizationfor production of flavin mononucleotide by the recombinant strain ofthe yeast Candida famata using statistical designs. Biochem Eng J49:52– 60. https://doi.org/10.1016/j.bej.2009.11.010.

27. Yatsyshyn VY, Fedorovych DV, Sibirny AA. 2009. The microbial synthesisof flavin nucleotides: a review. Appl Biochem Microbiol 45:115–124.https://doi.org/10.1134/S000368380902001X.

28. Mashhadi Z, Zhang H, Xu H, White RH. 2008. Identification and charac-terization of an archaeon-specific riboflavin kinase. J Bacteriol 190:2615–2618. https://doi.org/10.1128/JB.01900-07.

29. Frago S, Martinez-Julvez M, Serrano A, Medina M. 2008. Structuralanalysis of FAD synthetase from Corynebacterium ammoniagenes. BMCMicrobiol 8:160. https://doi.org/10.1186/1471-2180-8-160.

30. Santos MA, Jimenez A, Revuelta JL. 2000. Molecular characterization ofFMN1, the structural gene for the monofunctional flavokinase of Sac-charomyces cerevisiae. J Biol Chem 275:28618 –28624. https://doi.org/10.1074/jbc.M004621200.

31. Yang Y, Wu YC, Hu YD, Cao YX, Poh CL, Cao B, Song H. 2015. Engineeringelectrode-attached microbial consortia for high-performance xylose-fedmicrobial fuel cell. ACS Catal 5:6937– 6945. https://doi.org/10.1021/acscatal.5b01733.

32. Kotloski NJ, Gralnick JA. 2013. Flavin electron shuttles dominate extra-

cellular electron transfer by Shewanella oneidensis. mBio 4:e00553–12.https://doi.org/10.1128/mBio.00553-12.

33. Covington ED, Gelbmann CB, Kotloski NJ, Gralnick JA. 2010. An essentialrole for UshA in processing of extracellular flavin electron shuttles byShewanella oneidensis. Mol Microbiol 78:519 –532. https://doi.org/10.1111/j.1365-2958.2010.07353.x.

34. McAnulty MJ, Wood TK. 2014. YeeO from Escherichia coli exports flavins.Bioengineered 5:386–392. https://doi.org/10.4161/21655979.2014.969173.

35. Wu G, Yan Q, Jones JA, Tang YJ, Fong SS, Koffas MA. 2016. Metabolicburden: cornerstones in synthetic biology and metabolic engineeringapplications. Trends Biotechnol 34:652– 664. https://doi.org/10.1016/j.tibtech.2016.02.010.

36. Baron D, LaBelle E, Coursolle D, Gralnick JA, Bond DR. 2009. Electro-chemical measurement of electron transfer kinetics by Shewanella one-idensis MR-1. J Biol Chem 284:28865–28873. https://doi.org/10.1074/jbc.M109.043455.

37. von Canstein H, Ogawa J, Shimizu S, Lloyd JR. 2008. Secretion offlavins by Shewanella species and their role in extracellular electrontransfer. Appl Environ Microbiol 74:615– 623. https://doi.org/10.1128/AEM.01387-07.

38. Ramanan RN, Tan JS, Mohamed MS, Ling TC, Tey BT, Ariff AB. 2010.Optimization of osmotic shock process variables for enhancement of therelease of periplasmic interferon-�2b from Escherichia coli using re-sponse surface method. Process Biochem 45:196 –202. https://doi.org/10.1016/j.procbio.2009.09.007.

39. Paal M, Heel T, Schneider R, Auer B. 2009. A novel ecotin-ubiquitin-tag (ECUT) for efficient, soluble peptide production in the periplasmof Escherichia coli. Microb Cell Fact 8:7. https://doi.org/10.1186/1475-2859-8-7.

Harnessing Bacterial Periplasm for Biosynthesis Applied and Environmental Microbiology

January 2018 Volume 84 Issue 1 e01693-17 aem.asm.org 11

on February 2, 2018 by N

AN

YA

NG

TE

CH

NO

LOG

ICA

L UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from