Design of a noncovalently linked bifunctional enzyme for whole-cell
biotransformation
Chung-Jen Chianga, Li-Jun Linb, Zei Wen Wangb, Tzu-Tai Leec, Yun-Peng Chaob,d,e,∗
a Department of Medical Laboratory Science and Biotechnology, China Medical University, No. 91, Hsueh-Shih Road, Taichung 40402, Taiwan
b Department of Chemical Engineering, Feng Chia University, 100 Wenhwa Road, Taichung 40724, Taiwan
c Department of Animal Science, National Chung Hsing University, Taichung 402, Taiwan
d Department of Health and Nutrition Biotechnology, Asia University, Taichung 41354, Taiwan
e Department of Medical Research, China Medical University Hospital, Taichung 40447, Taiwan
KEYWORDS: Fusion protein, Biotransformation, Cohesin, Dockerin
Corresponding Authors
*(Y.-P. Chao) E-mail: ypchao@fcu.edu.tw. Tel: 886-4- 24517250, ext. 3677. Fax: 886-4-24510890.
ABSTRACT:
Optical pure d-p-hydroxyphenylglycine (d-HPG) is a precursor for semi-synthetic antibiotics. It can be synthesized from d,l-hydroxyphenyl hydantoin (HPH) by a two-step reaction mediated by dhydantoinase (HDT) and amidohydrolase (AHL). In this study, a bifunctional enzyme was originally created by in-frame fusion of AHL with HDT genes (AHL-HDT). However, the AHL-HDT fusion protein expressed in Escherichia coli was prone to aggregates, recognized as a frequently encountered problem for this conventional method. To address this issue, small interacting motifs, cohesin (Coh) and dockerin (Doc) domains of cellulosomes, were explored and illustrated to interact in vivo. Accordingly, Coh and Doc were fused with AHL and HDT, respectively. After co-expression in E. coli, Coh-tagged AHL and Doc-tagged HDT assembled into a soluble protein complex via the high-affinity interaction of Coh and Doc. Consequently, the protein assembly exhibited both AHL and HDT activities and a higher reaction rate than free counterparts. Whole cells expressing the protein assembly were more stable than ones with free proteins for d-HPG production, and they could be recycled six times with a conversion yield of d-HPG exceeding 90%.
INTRODUCTION
Amino acids of optical purity have a wide application range in industry, and they serve as building components for chemical syntheses of antibiotics, antifungal agents, pesticides, and sweeteners [1]. Of particular importance, d-p-hydroxyphenylglycine (d-HPG) is a precursor for the semi-synthetic antibiotics such as cephalosporin. Optical pure amino acids can be produced from racemic d,l-5-monosubstituted hydantoins by an enzymatic process mediated by stereoselective hydantoinase and amidohydrolase [2]. This production scheme has received attention from industry since a study first reported production of d-HPG from d,lhydroxyphenyl hydantoin (HPH) by a biotransformation reaction based on Agrobacterium radiobacter NRRL B11291 [3]. This bacterial strain displays d-hydantoinase (HDT) and N-carbamoyl-d-amino acid amidohydrolase (AHL) activities. The former catalyzes the hydrolysis of HPH to N-carbamoyl-d-p-hydroxy phenylglycine (CpHPG) while the latter converts CpHPG to d-HPG. We show that the efficiency of this d-HPG transformation reaction is greatly improved using the surrogate strain Escherichia coli that expresses heterologous HDT and AHL [4]. In living cells, metabolic reactions interconnected in a complicated network proceed within the viscous cell cytoplasm that is highly crowed with proteins and metabolite molecules. To ensure effective metabolic activities, living cells have evolved to adopt specialized protein complexes for
designed biological functions [5,6]. This multifunctional enzyme system is commonly conceived as an evolutional advantage for rapid turnover of substrates in a sequential reaction, driving intermediate metabolites away from competing pathways, and timely regulation of the enzyme assembly in response to the physiological status of cells [7]. Inspired by the potential of these nature’s models, many efforts have been devoted toward artificial creation of bifunctional and multifunctional enzymes for various engineering purposes [8]. In general, the results of most cases are encouraging with respect to the kinetic efficiency of enzymes [9,10]. Creation of artificial bifunctional enzymes is not uncommon. This is frequently carried out by in-frame fusion of two distinct structural genes separated by a short linker [11]. As expressed in cells, the fusion gene results in a hybrid protein that exhibits the dual activities of two individual enzymes. Although straightforward, this method becomes problematic to apply when two proteins with multimeric structures are fused together [12]. It leads to the fusion protein with a highly ordered and complex structure, which is poorly expressed and usually misfolds in host bacteria [13]. In this study, the conventional method was originally utilized to create a bifunctional enzyme by end-to-end fusion of AHL with the N-terminus of HDT (AHL-HDT). However, the resulting AHL-HDT fusion protein misfolded and aggregated after expression in E. coli. To address this issue, an alternative approach was implemented by exploration of cohesin (Coh) and dockerin (Doc) domains from cellulosomes. Coh and Doc were fused to AHL and HDT, respectively. Consequently, a soluble protein assembly exhibiting both AHL and HDT activities was obtained in E. coli by means of the high-affinity Coh–Doc interaction. Furthermore, the experiment illustrated that the protein complex gained a kinetic advantage over the free counterparts. For repeated production of d-HPG from HPH, the cells with the protein complex outperformed the ones with free counterpart proteins. Overall, the result indicates the feasibility of this approach in creating a bifunctional enzyme in vivo.
MATERIALS AND METHODS 2.1. Plasmid construction
The schematic structures of all plasmids applied in this study were summarized in Fig. 1. AHL-HDT was constructed as follows. The gene encoding AHL was first amplified from plasmid pChA203 [14] by polymerase chain reaction (PCR) with the designed primers (tttgtgatatcgctaccaccaccaccggtcaggccaagccaagcttg and taacttctagaaggagatatacatatgac). After digestion with EcoRV and XbaI, the PCR DNA was subcloned into the corresponding sites of plasmid pHDT [15] to produce plasmid pETW-A2HCh. The plasmid carried the AHL-HDT fusion gene under the control of
the T7 promoter (PT7). By PCR, the DNA fragment containing the pSC101 replication origin and the kanamycin-resistant determinant was amplified from plasmid pTH18kr [16] with the designed primers (taaccagatctgattagaaaaactcatcg and agaacctgcagtcagatccttccgtatttagc). The PCR DNA and plasmid pChA203 were both cleaved by BglII and PstI and then spliced together to produce plasmid pTH-ChA203. The resulting plasmid contained the chitin-binding domain (ChBD) of Bacillus circulans WL-12 fused to the C-terminus of AHL (AHL-ChBD). The type I Coh domain was amplified from Clostridium thermocellum DSM1237 by PCR using the designed primers (atcatctcgagttatgcggccgcaagctttggtg and tagaattcagatctcagccaaatgttcc).
After cleavage with EcoRI–XhoI, the PCR DNA was incorporated into plasmid pTH-ChA203 to give plasmid pTH-AL203Coh. The resulting plasmid carried C-terminal fusion of AHL with Coh (AHL-Coh) under the control of PT7. Plasmid pTac-HDTDoCh was constructed in several steps. First, the internal NdeI site of plasmid pET32a (Novagen, USA) was eliminated by site-directed mutagenesis [17] to generate plasmid pET32-N using the designed primers
(ggtgaataattttatcgctcatctgtatatctccttctagaggg and
ccctctagaaggagatatacagatgagcgataaaattattcacc). The DNA bearing HDT-ChBD was then recovered from plasmid pET-TrHDTCh [15] by NdeI–XhoI and ncorporated into the same sites of plasmid pET32-N to give plasmid pET-TrHDTCh. Secondly, the type I Doc was amplified from C. thermocellum DSM1237 by PCR with the designed primers (tgtgactcgagttagagctcgttcttgtacggcaatgtatc and gcatgaagcttaaagtacctggtactccttc). After cleavage by HindIII–XhoI, Doc was incorporated into plasmid TrHDTCh to replace ChBD, resulting in plasmid TrHDTDo. Moreover, the DNA containing ChBD was amplified from plasmid pET-TrHDTCh by PCR with the designed primers (agatgcttttctgtgactgg and agcatgagctcggcctgaccggtctgaac). The PCR DNA was digested by SacI and subsequently spliced into the corresponding site of plasmid pET-TrHDTDo to give plasmid pET-HDTDoCh. The resulting plasmid carried the fusion of HDT-DocI-ChBD. Finally, the DNA containing the tac promoter (Ptac) was amplified from plasmid pJF-Trxfus [18] with the designed primers (gaatttctagacctgtgtgaaattgttatcc and gagtgagatatttatgccag). The PCR DNA was treated with ApaI–XbaI and incorporated into the same sites of plasmid pET32-N to produce pTac-N. By a PstI– XbaI cut, HDT-Doc-ChBD was recovered from plasmid pET-HDTDoCh and ligated into the same sites of plasmid N. This construction gave rise to plasmid pTac-HDTDoCh, which created the C-terminal fusion of trxA (encoding thioredoxin) with HDT-Doc-ChBD (TrxA-HDT-Doc-ChBD). The Ptac-containing DNA was removed from plasmid pTac-HDTDoCh by PstI–XbaI. The DNA was then incorporated into
the same sites of plasmid pETTrHDTCh to replace PT7, thus generating plasmid pTac-TrChHDT that carried the fusion of TrxA-HDT-ChBD. In addition, PCR-amplification of Ds-Red was carried out from plasmid pDsRed-Express (Clontech, USA) with the designed primers (gaaacacatatgaccatgattacgccaag and gtcggaattcgtacaggaacaggtggtgg). The PCR DNA was digested by EcoRI–NdeI and then spliced into plasmid pTH-AL203Coh to obtain plasmid pRed- Coh. Meanwhile, the pgpB gene was amplified from strain BL21(DE3) (Novagene, USA) with the designed primers (gaatcatatgcgttcgattgccagacg and aacaaggtaccactttcttgttctcgttgcg). Restricted with KpnI–NdeI, the PCR DNA was ligated into plasmid pET32a to give plasmid pET-Pgp. Doc was amplified from plasmid pET-HDTDoCh with the designed primers (caagtaagcttagccgagctcgttcttgtacg and agcaacccgggaaagtggacatgactc). The PCR DNA was cleaved with HindIII–SmaI and incorporated into plasmid pET-Pgp which was treated with EcoRV–HindIII. The construction resulted in plasmid pGPB-DocI, which created the C-terminal fusion of pgpB with Doc (PgpB-Doc).
2.2. Bacterial culturing and protein analysis
Bacterial growth was monitored using a spectrophotometer at 550 nm (OD550). With the initial OD550 at 0.08, recombinant strains were cultured in shake flasks
containing 20 mL Luria-Bertani (LB) medium [19] supplemented with ampicillin (30 _g/mL) and/or kanamycin (30 _g/mL) at 30 ◦C. To induce protein production, IPTG (100 _M) and/or l-arabinose (50 _M) were added to cell cultures upon OD550
reaching 0.3. Bacterial cultures were harvested by centrifugation after induction for 4 h. Followed by washing with 100 mM sodium phosphate buffer (PBS) at pH 7, the cells were resuspended in the same buffer solution to reach 20 at OD550. The cells were then disrupted by sonication. After centrifugation, the supernatant was recovered as cell-free extract (CFX) and assayed for the protein content using Bio- Rad dye reagent. Each sample containing 20 _g proteins was resolved on dodecyl sulfate-polyacrylamide gel electrophoresis (PAGE). The method for performing SDS-PAGE essentially followed our previous report [20]. In brief, the resolving gel (8%) was prepared and overlaid with a stacking gel (5%). The electric power (100 V followed by 200 V) was then applied to the gel which was mounted in the
electrophoresis apparatus (Bio-Rad). The dismantled gel was finally stained with Coomassie blue for further analyses.
2.3. Analysis of bacteria by confocal microscopy
Bacterial cells were rinsed with PBS and fixed in 3.7% paraformaldehyde (Sigma, USA) for 30 min. Followed by washing three times with 100 mM PBS (pH 7.0), fixed
cells were blocked with 3% BSA (Sigma, USA) in PBS. Bacterial cells were then mounted on glass slides and observed by confocal microscopy (Leica TCS SP2, ermany) according to our reported method [21].
2.4. Affinity adsorption of proteins on chitin beads
The shake-flask cultures were carried out in a similar way. Cells were harvested and washed three times with 100 mM PBS (pH 7.0) containing 1 mM EDTA. CFX was recovered after cell disruption. The protocol for administration of protein adsorption essentially followed the previous report with slight modification [15]. In brief, CFX containing 30 _g proteins was added to the adsorption solution consisting of 20 mM PBS (pH 7.0) and 0.5 g chitin beads. The adsorption reaction was carried out with occasional shaking at 4 ◦C for 12 h. Finally, chitin beads were collected by centrifugation and rinsed twice with PBS.
2.5. Enzyme activity and reaction rate assay
Enzyme activities were determined as reported previously [22]. In brief, CFX was added to the reaction solution (1 mL). The content of fusion proteins in CFX was estimated from SDS-PAGE analyses using Image Analyzer (Alpha Innotech, SA) with the build-in software program. The composition of the reaction solution for AHL was 20 mM CpHPG (Widetex, Taiwan) and 100 mM PBS (pH 7.0) while that for HDT was 10 mM HPH (TCI, Japan), 0.5 mM MnCl2 (Sigma, USA), and Tris (pH 8.0). The reaction was conducted at 37 ◦C for 30 min and quenched by heating at 100 ◦C for 10 min. The concentration of CpHPG and d-HPG was measured by highperformance liquid chromatography (HPLC). The unit (U) of the HDT and AHL activity was defined as _mole of CpHPG or d-HPG produced per min. The specific enzyme activity (U/mg) was calculated by dividing U with the fusion protein (mg) in CFX. In addition, the reaction rate (mM/min) based on the produced amount of d-HPG (mM) was determined with CFX in a similar way. The reaction solution was composed of 20 mM HPH and 100 mM PBS (pH 7.5).
2.6. Whole-cell production of d-HPG
Production of d-HPG was conducted with resting cells (OD550 at 70) in 10 mL reaction solution consisting of 100 mM HPH and 100 mM PBS (pH 7.5). The reaction was carried out at 40 ◦C for 6 h with vigorously shaking. At the end of the reaction, entrifugation was applied to recover the supernatant for HPLC analyses and precipitated cells for the next reaction run.
3.1. In vivo interaction of Coh and Doc
Artificial linkage of two or more distinct genetic elements appears to be the most commonly applied strategy to generate chimeric proteins. It is conceived that chimeric proteins exhibit kinetic benefits by brining active sites of enzymes into close
proximity, which enables efficiently relaying intermediate substances in consecutive reaction steps. This phenomenon is known as metabolite (or substrate) channeling
[23]. Therefore, an initial attempt to create bifunctional enzyme was made by construction of plasmid pETW-A2HCh which carried in-frame fusion of AHL and HDT structural genes (Fig. 1). Consequently, the AHL gene is placed in front of the HDT gene with a linker in between. This is based on the fact that AHL has a higher solubility and the N-terminus of HDT is dispensable for the enzyme structure [24]. E. coli strain BL21(DE3) harboring plasmid pETW-A2HCh (BL21/pETW-A2HCh) was then cultured in a shake flask and induced for the protein production. By the SDS-PAGE analysis, the AHL-HDT fusion protein expressed in E. coli was predominantly present in the insoluble part (Fig. 2). Both AHL and HDT are multimeric enzymes
[25,26]. Accordingly, AHL-HDT assumes a large multidomain structure that folds inefficiently. The poor expression of soluble fusion proteins appears to be a problem commonly encountered by the conventional method of in-frame gene fusion.
Therefore, a hybrid enzyme with both AHL and HDT activities was employed as an example to seek an alternative and useful approach to solve this problem. Found in naturally occurring bacteria, cellulosomes are a complex protein assembly that hydrolyzes cellulosic materials in an efficient way [27]. Cellulosomes
are composed of the scaffoldin backbone containing Coh domains and cellulases with Doc domains. Through the strong Fig. 2. SDS-PAGE analysis of the AHL-HDT fusion protein expressed in E. coli. The shake-flask culture of strain BL21/pETW-A2HCh was carried out and then induced with IPTG for protein production. At the end of culturing, bacterial cells were harvested and processed for SDS-PAGE analyses. Keys: lane 1, protein marker; lane 2, the soluble part of uninduced bacteria; lane 3, the insoluble part of uninduced bacteria; lane 4, the soluble part of induced bacteria; lane 5, the insoluble part of induced bacteria. The position of AHL-HDT is indicated by an arrow. interaction of Coh with Doc, cellulases are specifically adhered to the scaffoldin. Both Coh and Doc domains are small peptides and can be utilized as fusion partners. As reported previously, designer cellulosomes could be created in vitro by specific incorporation of Doc-tagged cellulase enzymes into Coh of the designed scaffoldin [28]. This prompted us to investigate the feasibility of applying the Coh–Doc interaction to assemble AHL and HDT. However, it remains unclear that Coh and Doc can interact with each other in vivo. To address this issue, plasmid pGPB-DocI was constructed to have pgpB fused with Doc (Fig. 1). As reported
previously, PgpB is an inner membrane protein with its C-terminal domain exposed in cell cytoplasm [29]. Accordingly, this gene construction is designed to locate Doc all around the inner surface of cell membranes (Fig. 3A). Meanwhile, plasmid pRed-Coh was constructed to carry a variant of red fluorescent protein (DsRed) fused with Coh (Fig. 1). The experiment was then carried out with E. coli strain BAD5 [30] carrying plasmids pRed-Coh and pGPB-DocI (BAD5/pRed-Coh/pGPB-DocI) while the counterpart strain with plasmids pRed-Coh and pETPgp (BAD5/pRed-Coh/pET-Pgp) was used as a control. By confocal microscopy, a halo layer of the red fluorescence signal was detected for strain BAD5/pRed-Coh/pGPB-DocI (Fig. 3A). This result suggests that DsRed-Coh is likely confined along the inner contour face of cell membranes by PgpB-Doc via the Coh-Doc interaction in vivo. In contrast, the control strain displayed freely-distributed red signals because DsRed-Coh without interaction with PgpB is unconstrained inside of the cell (Fig. 3B).
3.2. Characterization of the protein assembly
According to our previous study, either AHL tagged with ChBD or HDT fused with TrxA and ChBD at two termini enabled adsorption of the proteins onto chitin beads but without affecting the enzyme activities [14,23]. Therefore, these gene constructs for AHL and HDT were used as the backbone for fusion with Coh and Doc, respectively (Fig. 1). First of all, these fusion proteins were separately expressed in strain BAD5 harboring each single plasmid. Meanwhile, two fusion proteins were co-expressed in strain BAD5 bearing either plasmids pTH-ChA203 and pTac-TrChHDT (BAD5/pTH-ChA203/TrChHDT) or plasmids pTH-AL203Coh and pTac-HDTDoCh (BAD5/pTH-AL203Coh/pTac-pTac-HDTDoCh). Consequently, each fusion protein was produced as a soluble form whereas their protein solubility decreased after fusion with CohI or DocI. Nevertheless, the approach still resulted in production of soluble fusion proteins in E. coli (Fig. 4A). The production level for AHL-ChBD, TrxA-HDT-ChBD, AHL-Coh, and TrxA-HDT-Doc-ChBD was estimated to account for around 8%, 5%, 2.5%, and 3% of total proteins, respectively. Moreover, soluble fusion proteins were determined for their respective AHL and HDT activity. As indicted in Table 1, the individual activity of AHL-ChBD and of TrxA-HDT-ChBD were higher than that of AHL-Coh and of TrxA-HDT-Doc-ChBD. Next, in vivo interaction of AHL with HDT via the Coh-Doc affinity was investigated based on the protein absorption experiment. The experiment was preformed with strain BAD5/pTHAL203Coh/ pTac-HDTDoCh and strain BAD5 harboring plasmids pTH-AL203Coh and pTac-TrChHDT (BAD5/pTH-pTH-AL203Coh/pTac- TrChHDT). Both recombinant strains were grown in a similar way. After disruption of cells, the soluble part from each strain was recovered and incubated with chitin beads. At the end of
adsorption, chitin beads were removed and heated to release the adsorbed proteins. As a result, AHL-Coh and TrxA-HDT-Doc-ChBD that were co-produced in strain BAD5/pTH-AL203Coh/pTac-HDTDoCh were released together (Fig. 4B). In contrast, TrxA-HDT-ChBD (lacking Doc) but not AHL-Coh was found in the released part of strain BAD5/pTH-AL203Coh/pTac-TrChHDT (Fig. 4C). Note that AHL-Coh lacks ChBD and is not specifically adsorbed onto chitin beads. Therefore, the result suggests that AHL-Coh is associated with TrxAHDT- Doc-ChB through the Coh-Doc affinity. Tagged with ChBD, the resulting protein assembly is adsorbed onto chitin beads and disintegrates to release TrxA-HDT-Doc-ChBD with AHL-Coh after thermal desorption
3.3. Kinetic advantage of the protein assembly
It was intriguing to investigate the transformation reaction rate catalyzed by the protein assembly. This was carried out by first culturing strains BAD5/pTH-AL203Coh/pTac-HDTDoCh and BAD5/pTH-ChA203/pTac-TrChHDT. Note that the former strain produced the protein assembly consisting of AHL-Coh and TrxAHDT-Doc-ChBD while the latter strain produced free counterpart proteins (i.e., AHL-ChBD and TrxA-HDT-ChBD). At the end of culturing, cells were harvested to prepare CFX and the soluble fraction of proteins from each recombinant strain was added to the reaction solution. It was estimated that the amounts of functional enzymes in the reaction were 0.5-_g AHL-Coh and 0.6-_g TrxAHDT- Doc-ChBD from strain BAD5/pTH-AL203Coh/pTac-HDTDoCh and 1.6-_g AHL-ChBD and 1-_g TrxA-HDT-ChBD from strain BAD5/pTH-ChA203/pTac-TrChHDT. With lower concentrations and individual activities in the reaction, the protein assembly comprising AHL-Coh and TrxA-HDT-Doc-ChBD still exhibited 2.6-folds higher conversion rate than free counterparts containing AHLChBD and TrxA-HDT-ChBD (Table 1). As indicated in Table 1, AHL and HDT fused with Coh and Doc were functionally expressed whereas they exhibited a lower activity than their free counterparts. The fact that proteins post fusion display reduced individual enzyme activities is not uncommon. For instance, two representative examples are the fusion protein resulting from _-galactosidase with galactokinase [31] and thiolase with reductase [32]. It is known that the composition of a linker may affect the property of the resulting fusion protein. In this study, the linker spanning between either AHL and Coh or HDT and Doc was not optimized (Fig. 1). Nevertheless, the protein assembly consisting of Coh-tagged AHL and Doc-tagged HDT displayed a higher overall reaction rate, indicating the synergistic effect of the enzymes (Table 1). This approach is reminiscent of the natural multifunctional enzyme complex in cell metabolism, which is designed to couple activities of related enzymes in the sequential metabolic
reaction.
3.4. Repeated production of d-HPG by whole cells
AHL is known to be liable to oxidative inactivation and susceptible to heat [33,34]. Therefore, production of d-HPG by resting cells was conducted by using strains BAD5/pTH-AL203Coh/pTac- HDTDoCh and BAD5/pTH-ChA203/pTac-TrChHDT. The optimum condition for the conversion reaction by two strains was found as follows: pH 7.5 and 40 ◦C. Each reaction cycle was carried out with the same amount of cells for 6 h. At the end of the reaction, the cells were recovered by centrifugation and used for the next reaction run. As shown in Fig. 5, strain BAD5/pTH-AL203Coh/ pTac-HDTDoCh that expressed the protein assembly exhibited the conversion yield of 100% for the first 4 cycles while the conversion yield decreased to around 90% at the 6th cycle. In sharp contrast, the conversion yield for strain BAD5/pTH-ChA203/ pTac-TrChHDT that expressed free proteins reached 100% for the first 3 cycles but sharply dropped to less than 80% at the 4th cycle. Note that strain BAD5/pTHAL203Coh/ pTac-HDTDoCh exhibits lower levels and activities of enzymes than strain BAD5/pTH-ChA203/pTac-TrChHDT (Fig. 4A and Table 1). This result implies that the protein assembly is more stable than free counterpart proteins. Fusion proteins may suffer the problem of structural instability. A previous study reported that AHL linked to the N-terminus of HDT from Bacillus stearothermophilus SD1 was subjected to extensive proteolysis [35]. By DNA shuffling, the resulting fusion protein was evolved with enhanced structural stability. In contrast, the protein assembly by Coh-tagged AHL and Doc-tagged HDT in this study was not afflicted by structural instability. Fig. 5. Repeated production of d-HPG by whole cells. With the fixed reaction time for each cycle, the reaction was repeated as needed. The conversion yield was defined as the produced amount of d-HPG relative to that of HPH in mM. Keys: d-HPG production by strain AL203Coh/pTac-HDTDoCh (solid bars); d-HPG production by strain BAD5/pTH-ChA203/pTac-TrChHDT (open bars). The experiment was conducted in triplicate. The results of HPLC analyses for d-HPG production at cycles 1 and 6 were also shown in supplementary Fig. S1. assembly for d-HPG production (Fig. 5). Apparently, the success of this approach holds a great promise for industrial development of the d-HPG production process.
4. Conclusion
As an alternative to the pathway engineering approach [36], our proposed method by compartmenting the spatial enzymes could be utilized to improve the efficiency of whole-cell biotransformation. This method is likely to circumvent the problems
frequently encountered by the approach of protein fusion, including poor expression of high ordered protein structures, protein misfolding and aggregates, and structural instability. Note that one should take into linker optimization into account when further improvement of this approach is concerned. This method may be also useful for a rational combination of natural proteins, which enables to broaden the application range of enzymes.
Notes
The authors declare no competing financial interest. ACKNOWLEDGMENTS
This work is supported by Ministry of Science and Technology, Taiwan (NSC 101-2221-E-035-057-MY3 and NSC
103-2622-E-035-008-CC1).
References
[1] Syldatk C, Laufer A, Muller R, Hoke H. Production of optically pure dand l-_-amino acids by Bioconversion of d,l-5-monosubstituted hydantoin
derivatives. In: Fiechter A, editor. Advances in biochemical engineering/ biotechnology, vol. 41. Berlin: Springer; 1990. p. 29–75.
[2] Chao YP, Chiang CJ, Chern JT, Tzen JTC. Hydantoinase. In: Polaina J, MacCabe AP, editors. Industrial enzymes. The Netherlands: Springer; 2007. p. 599–606.
[3] Olivieri R, Fasctti E, Angelini L, Degen L. Microbial transformation of racemic hydantoins to d-amino acids. Biotechnol Bioeng 1981;23:2173–83.
[4] Chao YP, Fu H, Lo TE, Chen PT, Wang JJ. One-step production of
d-p-hydroxyphenylglycine by recombinant Escherichia coli strains. Biotechnol Prog 1999;15:1039–45.
[5] Charles IG, Keyte JW, Brammar WJ, Smith M, Hawkins AR. The isolation and nucleotide sequence of the complex AROM locus of Aspergillus nidulans. Nucleic Acids Res 1986;14:2201–13.
[6] Momoyo I, Daisuke T, Takuji O, Yasuo TMK. Structural basis for channelling mechanism of a fatty acid _-oxidation multienzyme complex. EMBO
J 2004;23:2745–54.
[7] Cascante MA, Sorribas A, Canela EI. Enzyme–enzyme interactions and metabolite
channelling: alternative mechanisms and their evolutionary significance. Biochem J 1994;298:313–20.
of metabolic enzymes: mimicking nature’s synergy. Curr Opin Biotechnol 2008;19:492–9.
[9] Ljungcrantz P, Carlsson H, Mansson MO, Buckel P, Mosbach K, Bulow L. Construction of an artificial bifunctional enzyme, _-galactosidase/galactose dehydrogenase, exhibiting efficient galactose channeling. Biochemistry 1989;28:8786–92.
[10] Hak SS, Yeon JK, Jae YL, Jong TS, Chung HK, Ju KK, et al. Characterization of a bifunctional enzyme fusion of trehalose-6-phosphate synthetase and
trehalose-6-phosphate phosphatase of Escherichia coli. Appl Environ Microbiol 2000;66:2484–90.
[11] Arai R, Ueda H, Kitayama A, Kamiya N, Nagamune T. Design of the linkers which effectively separate domains of a bifunctional fusion protein. Protein Eng 2001;14:529–32.
[12] Bulow L, Mosbach K. Multienzyme systems obtained by gene fusion. Trends Biotechnol 1991;9:226–31.
[13] Chang HC, Kaiser CM, Hartl FU, Barral JM. De novo folding of GFP fusion proteins:
high efficiency in eukaryotes but not in bacteria. J Mol Biol 2005;353:397–409.
[14] Chiang CJ, Chern JT, Wang JY, Chao YP. Facile immobilization of evolved Agrobacterium radiobacter carbamoylase with high thermal and oxidative stability. J Agric Food Chem 2008;56:6348–54.
[15] Chern JT, Chao YP. Chitin-binding domain based immobilization of d-hydantoinase. J Biotechnol 2005;117:267–75.
[16] Hashimoto-Gotoh T, Yamaguchi M, Yasojima K, Tsujimura A, Wakabayashi Y, Watanabe Y. A set of temperature sensitive-replication/-segregation and temperature resistant plasmid vectors with different copy numbers and in an
isogenic background (chloramphenicol, kanamycin, lacZ, repA, par, polA). Gene 2000;241:185–91.
[17] Chao YP, Chern JT, Wen CS, Fu H. Construction and characterization of thermo-inducible vectors derived from heat-sensitive lacI genes in combination with the T7 A1 promoter. Biotechnol Bioeng 2002;79:1–8.
[18] Chao YP, Law WS, Chen PT, Hung WB. High production of heterologous proteins
in Escherichia coli using the thermo-regulated T7 expression system. Appl Microbiol Biotechnol 2002;58:446–53.
[19] Miller JH. Experiments in molecular genetics. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory; 1972.
immobilization process for recombinant endoglucanase from Clostridium thermocellum.
Process Biochem 2013;48:1886–92.
[21] Chiang CJ, Lin SC, Lin LJ, Chen CJ, Chao YP. Caleosin-assembled oil bodies as a potential delivery nanocarrier. Appl Microbiol Biotechnol 2012;93:
1905–15.
[22] Chao YP, Juang TY, Chern JT, Lee CK. Production of d-p-hydroxy-phenylglycine
by N-carbamoyl-d-amino acid amidohydrolase-overproducing Escherichia coli strains. Biotechnol Prog 1999;15:603–7.
[23] Zhang YH. Substrate channeling and enzyme complexes for biotechnological applications. Biotechnol Adv 2011;29:715–25.
[24] Kim GJ, Kim HS. C-terminal regions of d-hydantoinases are nonessential for catalysis, but affect the oligomeric structure. Biochem Biophys Res Commun 1998;243:96–100.
[25] Cheon YH, Kim HS, Han KH, Abendroth J, Niefind K, Schomburg D, et al. Crystal structure of d-hydnatoinase from Bacillus stearothermophilus: insight into the stereochemistry of enantioselectivity. Biochemistry 2002;41: 9410–7.
[26] Wang WC, Hsu WH, Chien FT, Chen CY. Crystal structure and site-directed mutagenesis studies of N-carbamoyl-d-amino-acid amidohydrolase from
Agrobacterium radiobacter reveals a homotetramer and insight into a catalytic cleft. J Mol Biol 2001;306:251–61.
[27] Fontes CM, Gilbert HJ. Cellulosomes: highly efficient nanomachines designed to deconstruct plant cell wall complex carbohydrates. Annu Rev Biochem
2010;79:655–81.
[28] Fierobe HP, Mechaly A, Tardif C, Belaich A, Lamed R, Shoham Y, et al. Design and production of active cellulosome chimeras. Selective incorporation of
dockerin-containing enzymes into defined functional complexes. J Biol Chem 2001;276:21257–61.
[29] Touze T, Blanot D, Mengin-Lecreulx D. Substrate specificity and membrane topology of Escherichia coli PgpB, an undecaprenyl pyrophosphate phosphatase. J Biol Chem 2008;283:16573–83.
[30] Wang ZW, Lai CB, Chang CH, Chiang CJ, Chao YP. A glucose-insensitive T7 expression system for fully-induced expression of proteins at a subsaturating
level of l-arabinose. J Agric Food Chem 2011;59:6534–42.
[31] Bulow L, Ljungcrantz P, Mosbach K. Preparation of a soluble bifunctional enzyme by gene fusion. Nat Biotechnol 1985;3:821–3.
[32] Kourtz L, Dillon K, Daughtry S, Madison LL, Peoples O, Snell KD. A novel thiolase-reductase gene fusion promotes the production of polyhydroxybutyrate in Arabidopsis. Plant Biotechnol J 2005;3:435–47.
[33] Buson A, Negro A, Grassato L, Tagliaro M, Basaglia M, Grandi C, et al. Identification,
sequencing and mutagenesis of the gene for a d-carbamoylase from Agrobacterium radiobacter. FEMS Microbiol Lett 1996;145:55–62.
[34] Grifantini R, Pratesi C, Galli G, Grandi G. Topological mapping of the cysteine residues of N-carbamyl-d-amino-acid amidohydrolase and their role in enzymatic activity. J Biol Chem 1996;271:9326–31.
[35] Kim GJ, Cheon YH, Kim HS. Directed evolution of a novel
N-carbamylase/d-hydantoinase fusion enzyme for functional expression with enhanced stability. Biotechnol Bioeng 2000;68:211–7.
[36] Siedler S, Lindner SN, Bringer S, Wendisch VF, Bott M. Reductive whole-cell biotransformation with Corynebacterium glutamicum: improvement of NADPH generation from glucose by a cyclized pentose phosphate pathway using
pfkA and gapA deletion mutants. Appl Microbiol Biotechnol 2013;97: 143–52.
