Modified Deacetylcephalosporin C Synthase for the Biotransformationof Semisynthetic Cephalosporins

Nataraj Balakrishnan, Sadhasivam Ganesan, Padma Rajasekaran, Lingeshwaran Rajendran, Sivaprasad Teddu, Micheal Durairaaj

Biotechnology Division, R&D Centre, Orchid Chemicals and Pharmaceuticals Ltd., Chennai, India

ABSTRACT

Deacetylcephalosporin C synthase (DACS), a 2-oxoglutarate-dependent oxygenase synthesized by Streptomyces clavuligerus,transforms an inert methyl group of deacetoxycephalosporin C (DAOC) into an active hydroxyl group of deacetylcephalosporinC (DAC) during the biosynthesis of cephalosporin. It is a step which is chemically difficult to accomplish, but its development byuse of an enzymatic method with DACS can facilitate a cost-effective technology for the manufacture of semisynthetic cephalo-sporin intermediates such as 7-amino-cephalosporanic acid (7ACA) and hydroxymethyl-7-amino-cephalosporanic acid (HACA)from cephalosporin G. As the native enzyme showed negligible activity toward cephalosporin G, an unnatural and less expensivesubstrate analogue, directed-evolution strategies such as random, semirational, rational, and computational methods were usedfor systematic engineering of DACS for improved activity. In comparison to the native enzyme, several variants with improvedcatalytic efficiency were found. The enzyme was stable for several days and is expressed in soluble form at high levels with signifi-cantly higher kcat/Km values. The efficacy and industrial scalability of one of the selected variants, CefFGOS, were demonstrated ina process showing complete bioconversion of 18 g/liter of cephalosporin G into deacetylcephalosporin G (DAG) in about 80 minand showed reproducible results at higher substrate concentrations as well. DAG could be converted completely into HACA inabout 30 min by a subsequent reaction, thus facilitating scalability toward commercialization. The experimental findings withseveral mutants were also used to rationalize the functional conformation deduced from homology modeling, and this led to thedisclosure of critical regions involved in the catalysis of DACS.

IMPORTANCE

7ACA and HACA serve as core intermediates for the manufacture of several semisynthetic cephalosporins. As they are expensive,a cost-effective enzyme technology for the manufacture of these intermediates is required. Deacetylcephalosporin C synthase(DACS) was identified as a candidate enzyme for the development of technology from cephalosporin G in this study. Directed-evolution strategies were employed to enhance the catalytic efficiency of deacetylcephalosporin C synthase. One of the selectedmutants of deacetylcephalosporin C synthase could convert high concentrations of cephalosporin G into DAG, which subse-quently could be converted into HACA completely. As cephalosporin G is inexpensive and readily available, the technologywould lead to a substantial reduction in the cost for these intermediates upon commercialization.

Semisynthetic cephalosporins, a class of �-lactam antibiotics,have shown remarkable effectiveness in the treatment of infec-

tious diseases. Together with penicillins, they comprise nearly65% of anti-infectives used worldwide. Their high specificity andlow toxicity, coupled with the evolvability of newer generations ofantibiotics, have led to �-lactams being by far the most frequentlyused anti-infectives in clinical medicine (1, 2). The growing inci-dence of resistant isolates and the need for effective broad-spec-trum antibiotics constantly drive the development of semisynthetic�-lactam antibiotics, which are obtained primarily from three coreintermediates, namely, 7-aminodeacetoxy-cephalosporanic acid(7ADCA), 7-amino-cephalosporanic acid (7ACA)/hydroxymethyl-7-amino-cephalosporanic acid (HACA), and 7-amino-3-vinyl-cephalosporanic acid (7AVCA).

The current process for the production of 7ADCA involvesseveral steps, consisting of chemical ring expansion of penicillin G(PenG) to cephalosporin G (CephG) followed by enzymatic hy-drolysis by PenG amidase (3). Although 7ADCA is inexpensivedue to the low cost of penicillin G and is used in the manufactureof active pharmaceutical ingredients (APIs) such as cephalexin,the presence of an inactive methyl group at the third position of itscephem moiety limits its industrial utility. 7ACA is currently man-ufactured by a two-stage enzymatic process from cephalosporin C

(4, 5) (Fig. 1A) and is used for making APIs such as cefalotin,cefaloglycin, etc. Far more significantly, HACA, a deacetylatedderivative of 7ACA, is used for producing prominent APIs such ascefuroxime axetil. There has been progress to simplify the currentprocess for manufacturing 7ACA from a two-step to a single-stepenzymatic process (6). Since cephalosporin C is inherently chem-ically unstable, the need for additional steps in removing the as-sociated impurities during the manufacture of cephalosporin Cleads to the high cost of 7ACA. Despite the high cost, 7ACA and itsderivative HACA are widely used and remain remarkably attrac-

Received 18 January 2016 Accepted 6 April 2016

Accepted manuscript posted online 15 April 2016

Citation Balakrishnan N, Ganesan S, Rajasekaran P, Rajendran L, Teddu S, DurairaajM. 2016. Modified deacetylcephalosporin C synthase for the biotransformation ofsemisynthetic cephalosporins. Appl Environ Microbiol 82:3711–3720.doi:10.1128/AEM.00174-16.

Editor: M. J. Pettinari, University of Buenos Aires

Address correspondence to Micheal Durairaaj, durairaj@orchidpharma.com.

Supplemental material for this article may be found at http://dx.doi.org/10.1128/AEM.00174-16.

Copyright © 2016, American Society for Microbiology. All Rights Reserved.

crossmark

July 2016 Volume 82 Number 13 aem.asm.org 3711Applied and Environmental Microbiology

on July 1, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

tive for the evolution of newer generations of semisynthetic ceph-alosporins due to their versatility for derivatization from thecephem third position in addition to the seventh amino position.Currently, one-third of cephalosporins are manufactured frompenicillins, while the remaining two-thirds are synthesized from7ACA/HACA and similar intermediates. There exists a tremen-dous need for an alternate route for the production of these �-lac-tam bulk intermediates, which needs to be far more cost-effectiveand which will also drastically reduce the negative environmentalimpact.

Native penicillins and cephalosporins are produced by a vari-

ety of bacteria and fungi, and their genetic and biochemical path-ways have been well characterized (7). It is known that 2-oxoglu-tarate (2-OG)-dependent and related oxygenases are involved inthe biosynthesis of at least four families of �-lactam antibioticsand �-lactamase inhibitors, namely, the penicillins, cephalospo-rins, clavams, and carbapenems. They catalyze a wide range ofreactions, including hydroxylations, desaturations, and oxidativering closures (8). Deacetylcephalosporin C synthase (DACS) (en-coded by cefF), one of the 2-OG-dependent oxygenases involvedin the biosynthesis of cephalosporins, carries out a critical trans-formation of an inert cephem methyl group of deacetoxycepha-

FIG 1 Schematics of existing and proposed routes of synthesis of HACA and 7ACA. (A) Current process of synthesis of 7ACA and HACA from cephalosporinC. (B) Biosynthetic reaction catalyzed by DACS. (C) Proposed route of synthesis of 7ACA and HACA from CephG.

Balakrishnan et al.

3712 aem.asm.org July 2016 Volume 82 Number 13Applied and Environmental Microbiology

on July 1, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

losporin C (DAOC) into an active hydroxyl group of deacetyl-cephalosporin C (DAC) in the presence of iron, oxygen, ascorbicacid, and 2-oxoglutarate (9, 10) (Fig. 1B). It is a step which ischemically difficult to accomplish and which offers a second sub-stitution site for derivatization in cephalosporins. For instance,cephalosporin G, alternatively known as CephG or DAOG, is anabundantly available analogue of DAOC. It is less expensive, andits phenylacetyl group can be readily cleaved by PenG amidase.Efficient transformation of CephG by DACS can lead to a novelintermediate, deacetylcephalosporin G (DAG), and upon chemi-cal/enzymatic modification can lead to the formation of 7ACA orHACA (Fig. 1C). Thus, the biocatalytic strategy for the develop-ment of an alternate route of synthesis is focused on DACS andbecame the object for further investigation.

The cefF gene from Streptomyces clavuligerus, coding for DACS,has been cloned and sequenced, and the biochemical parametershave been characterized (9, 10, 11). When cefF from S. clavuligeruswas expressed in Escherichia coli and evaluated for catalytic effi-ciency toward CephG, a commercially available substrate, itshowed negligible activity. Sequence comparison of DACS sug-gests 71% similarity at the nucleotide level and 59% similarity atthe amino acid level to deacetoxycephalosporin C synthase(DAOCS) (also called expandase), another 2-OG-dependent ox-ygenase involved in the ring expansion of penicillin into cephalo-sporin (10). To date, a total of 17 variant crystal structures ofDAOCS have been produced, thus offering greater detail regard-ing its coordination chemistry, the residues involved in bindingthe substrate, and the cosubstrate, which facilitated the develop-ment of a possible catalytic mechanistic model (12, 13, 14, 15, 16).However, rational alterations based on crystallographic data forDAOCS resulted in only limited improvement in the ring expan-sion activity of penicillin G, whereas random mutagenesis usingconventional chemical mutagenesis, error-prone PCRs, shufflingof the family of expandase genes, and directed-evolution studiescould increase expandase activity dramatically (17, 18, 19, 20).The contrasting activity levels observed between rational andrandom mutagenesis studies reflect the limitations of currentunderstanding of catalytic mechanisms derived from static crys-tallographic data and hence the need for adopting random, semi-rational, computational modeling and rational mutagenesis strat-egies for the systematic evolution of DACS as an industrialbiocatalyst. Functional analysis of large numbers of mutants thusaccumulated may provide insights toward fundamental under-standing of the structure-function relationship of DACS and maylead to further refinement of rational engineering strategies.Hence, the objective of this work was to improve the activity ofDACS by using the above-mentioned approach and also to ratio-nalize the functional conformation by homology modeling.

MATERIALS AND METHODSMaterials. All chemicals and reagents were purchased from USB or Sig-ma-Aldrich Chemicals Pvt. Ltd., USA, or from Merck Specialties PrivateLtd., India, unless otherwise specified. Oligonucleotides were synthesizedand supplied by Microsynth GMBH, Switzerland, or Eurofins MWG, In-dia. Restriction enzymes, the pUC19 vector, and strains were obtainedfrom either New England BioLabs Inc., USA, or Gene Technologies, India.The pET24a vector, E. coli strain BL21(DE3), and the Bugbuster reagentwere purchased from Novagen, USA. Kanamycin was supplied by BioBasic Canada Inc., Canada. The deoxynucleoside triphosphate (dNTP)mix was purchased from 5 Prime, Germany. The Kapa polymerase kit waspurchased from Kapa Biosystems, USA. The S. clavuligerus strain was

obtained from the American Type Culture Collection (ATCC), USA.Bradford reagent and protein marker were purchased from Bio-Rad,USA. C18 Xterra columns (50 by 4.6 mm; 5 �m) were obtained fromWaters, USA. The DNeasy plant mini-gDNA isolation kit was supplied byQiagen, Germany, and growth medium components were obtained fromBecton Dickinson, USA. Highly purified cephalosporin G and deacetyl-cephalosporin G were supplied by the Process Development Laboratory,Orchid Chemicals and Pharmaceuticals Ltd., India. 2-Oxoglutarate wasobtained from SD Fine-Chem Ltd., India. DNA sequencing was facilitatedby Eurofins Genomics India Private Ltd., India.

Cloning of the cefF gene from S. clavuligerus in E. coli BL21(DE3).The S. clavuligerus strain (ATCC 27064) was inoculated in yeast-malt-glucose (YMG) medium and incubated at 25°C and 180 rpm for 48 h.Once the culture reached an optical density at 600 nm (OD600) of 3, it waspelleted by centrifugation at 16,000 rpm for 15 min, and the genomicDNA was isolated using the DNeasy plant mini-gDNA isolation kit (Qia-gen) as recommended by the supplier. The cefF gene (accession numberM63809) was amplified by PCR using 5=GCATATGGCGGACACGCCCGTACC3= (forward) and 5=CCCGGCTTGAATGCAACGACGAGCAT3=(reverse) primers, 2 U Deep Vent DNA polymerase, 200 �M dNTPs, 10%dimethyl sulfoxide (DMSO), Deep Vent DNA polymerase buffer, 1 mMMgSO4, and water in a final reaction volume of 100 �l. The PCR condi-tions consisted of an initial denaturation for 5 min at 95°C followed by 24cycles consisting of denaturation at 95°C for 40 s, annealing at 60°C for 1min, and extension at 72°C for 5 min, with a final extension at 72°C for 15min. The hydroxylase gene amplicons were purified with the Qiaex II gelextraction kit (Qiagen) and cloned into the pUC19 vector through blunt-end ligation using the SmaI restriction site with a standard cloning pro-tocol (21), resulting in pOBTF. Subsequently, the hydroxylase gene frag-ment was released by restriction digestion with NdeI/EcoRI and ligatedinto similarly digested pET24a(�) vector to give the pOCPLF vector andtransformed into competent E. coli BL21(DE3) cells.

Deacetylcephalosporin C synthase expression in E. coli. A glycerolstock of recombinant E. coli BL21(DE3) was inoculated in 10 ml LB me-dium containing kanamycin (75 �g/ml) for overnight growth at 37°C at220 rpm. The overnight culture was subcultured in 50 ml fresh LB me-dium and further cultivated at 37°C at 220 rpm. When the OD600 reached0.6, 50 �l of 100 mM IPTG (isopropyl-�-D-thiogalactopyranoside) wasadded, and the cultivation was further continued for 4 h at 25°C. Subse-quently, pellets at an OD of 3 were prepared, suspended in 200 �l resus-pension buffer containing 50 mM Tris-HCl (pH 7.5), 0.1 mM dithiothre-itol (DTT), 0.01 mM EDTA, 10% glycerol, and 50 mM glucose, and storedat �80°C until further use in activity measurements, SDS-PAGE analysis,and protein concentration determination.

Preparation of crude extract of deacetylcephalosporin C synthasefrom E. coli. After thawing the expression cell pellets for 10 min at roomtemperature, 40 �l of Bugbuster reagent was added and incubated at 25°Cfor 30 min at 220 rpm in an orbital shaker to facilitate lysis. The crudelysate was centrifuged at 13,000 rpm for 10 min. The supernatant wascollected for catalytic assay, protein concentration determination, andSDS-PAGE analysis of the soluble fraction; the pellet fraction was resus-pended in 240 �l of resuspension buffer, and 30 �l of it was used inSDS-PAGE analysis to determine the extent of inclusion body formation.The protein concentration was determined by the Bradford method (22)using bovine serum albumin (BSA) as the standard. Samples for determi-nation of expression profiles were denatured at 100°C and subjected toSDS-PAGE using standard protocols (21). Upon completion of electro-phoresis, the gel was stained with Coomassie blue for the detection andanalysis of the expression pattern of DACS.

Catalytic assay of deacetylcephalosporin C synthase. A 30-�l aliquotof crude lysate was mixed with 30 �l of assay mix containing 710 mMTris-HCl, 18 mM ascorbic acid, 90 mM 2-oxoglutaric acid, 0.2 mMFeSO4, and 1.125% CephG and incubated in a shaker at 25°C at 220 rpm.After 30 min, the reaction was quenched with 60 �l of absolute methanol

Modified Deacetylcephalosporin C Synthase

July 2016 Volume 82 Number 13 aem.asm.org 3713Applied and Environmental Microbiology

on July 1, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

and centrifuged at 13,000 rpm for 5 min. The supernatant was collectedand used for analysis by high-pressure liquid chromatography (HPLC).

Monitoring and quantitation of deacetylcephalosporin C synthaseactivity by HPLC. HPLC analysis was performed using either an AgilentTechnologies 1200 series liquid chromatographic system or a HitachiLaChrom Elite series liquid chromatographic system containing a C18

Xterra column (5 �m; 50 by 4.6 mm). The separation was effected underisocratic elution using 30% acetonitrile in 0.25% NaH2PO4 · H2O (pH2.4) at a flow rate of 2.5 ml/min. The peak area was used for quantitativeestimation and comparison studies using a reference standard.

Mutagenic cefF library creation. (i) Random mutagenesis. ThepOBTF vector was used as the template for error-prone PCR mutagenesis.The amplification was carried out with 20 pmol of 5=ATCGGTGCGGGCCTCTTCGCTATT3= and 5=CTCACTCATTAGGCACCCCAGGCT3=primers in a reaction mix containing 10% DMSO, Taq DNA polymerasebuffer, 2.5 U Taq DNA polymerase enzyme, various concentrations ofdNTPs, and water in a final reaction volume of 100 �l and amplified asdescribed above. Three different mutagenesis reactions were employed inwhich dATP, dGTP, and dCTP were biased and maintained at 100 �Mand the remaining complement of dNTPs were maintained at 500 �M,while the fourth variant had a dual bias of 50 �M (each) dATP and dTTPwith their complementing dCTP and dGTP at 500 �M each. The ampli-cons were purified, subjected to restriction enzyme digestion by NdeI andEcoRI, ligated to similarly digested expression vector pET24a, and trans-formed into competent E. coli strain BL21(DE3). Recombinants werescreened for inserts by colony PCR using 20 pmol of 5=GCATATGGCGGACACGCCCGTACC3= and 5=CCCGGCTTGAATGCAACGACGAGCAT3= primers in a reaction mix containing 10% DMSO, 10� Taq DNApolymerase buffer, 1 U Taq DNA polymerase enzyme, and 80 �M dNTPsin a total reaction volume of 50 �l using the PCR conditions describedabove. Alternately, the Genemorph II kit (GMK) was used for randommutagenesis with various template concentrations as recommended bysupplier. Briefly, Mutazyme reaction buffer (5 �l), dNTP mix (suppliedwith the kit) (1 �l), 20 �M T7 forward primer (1.5 �l), 10 �M cefF reverseprimer (5=TCTATGAATTCTCATCCGGCCTGCGGCTC3=, 2.5 �l), 100to 400 ng of template DNA, Mutazyme II polymerase (1 �l), and DMSO(2 �l) were added to a PCR tube, and the reaction volume was made up to50 �l with water. Mutagenic amplification was performed with an initialdenaturation step at 95°C for 5 min followed by 30 cycles consisting ofdenaturation at 95°C for 45 s, annealing at 60°C for 40 s, and extension at72°C for 80 s, with a final extension for 10 min at 72°C. The PCR productwas purified and cloned into the pET24a(�) vector using standard liga-tion procedures and transformed into E. coli BL21(DE3). The recombi-nant clones were grown in LB medium in 96-well microtiter plates andstored as glycerol stocks until further use for expression and screening.

(ii) Rational combination of mutations. Site-directed mutagenesis tocreate variants with combinations of mutations was carried out by usingthe megaprimer PCR strategy (23). Briefly, oligonucleotides carrying mu-tations were added to template DNA along with dNTPs, Vent DNA poly-merase, 10� Vent DNA polymerase buffer, DMSO, and water in a totalreaction volume of 50 �l. The final amounts of dNTPs, primers, DeepVent DNA polymerase, polymerization buffer, and DMSO were 0.15 mM,0.2 pM, 0.02 U, 1�, and 10%, respectively. Thermal cycling of putativemutagenic templates consisted of initial denaturation at 95°C for 5 minfollowed by 30 cycles of denaturation at 95°C for 40 s, annealing at 60°Cfor 30 s, and elongation at 72°C for 2 min, with a final extension at 72°C for15 min. Once the amplification was complete, the PCR product was pu-rified and subjected to megaprimer PCR. The procedure consists of mix-ing 10 �l of template DNA (70 ng), 12 �l of megaprimer (170 ng), 1.5 �lof 5 mM dNTP mix, 3.5 �l of 25 mM MgCl2, 1 �l of Kapa polymerase, 10�l of 5� Kapa GC buffer, and 12 �l of water in a reaction volume of 50 �l.The final concentrations of the dNTPs and MgCl2 are 150 �M and 1.75mM, respectively. Upon mixing of the reaction components, amplifica-tion was carried out by maintaining the mixture at 68°C for 10 s and thendenaturation at 95°C for 5 min, followed by 25 cycles of denaturation at

98°C for 20 s, annealing for 30 s at 65°C, and elongation at 68°C for 6 min,with a final extension at 68°C for 20 min. Subsequently, the PCR productwas purified, digested with DpnI restriction enzyme for 2 h at 37°C, andused for transforming competent E. coli BL21(DE3). After overnight in-cubation at 37°C, putative mutant clones were inoculated in LB, plasmidswere isolated from overnight culture and subjected to restriction digestionusing standard molecular biology protocols to identify the mutant clones,and the positive clones were stored as glycerol stocks at �80°C.

Expression of the mutagenic cefF library of clones. Glycerol stocks ofthe putative mutant cefF library of clones were inoculated in 96-well platesand expressed in deep-well blocks as described earlier. Upon completionof expression, the cell pellets were harvested by centrifugation in an Ep-pendorf microplate centrifuge at 4,000 rpm for 10 min at 4°C and werestored at �80°C until further use for activity measurements.

Screening of mutant deacetylcephalosporin C synthase enzyme li-brary. After thawing expression pellets at room temperature for 10 min,150 �l of 50 mM Tris (pH 7.5) and 15 �l of Bugbuster reagent were addedto each well, and the block was left in an orbital shaker for 20 min at roomtemperature at 260 rpm to facilitate the release of enzyme. After cell lysis,30 �l of the lysate from each well of the deep-well block was transferred toa new plate, and 30 �l of assay mixture containing 90 mM CephG, 137mM 2-oxoglutaric acid, 68 mM ascorbic acid, and 5 mM FeSO4 wasadded. After mixing the reaction components thoroughly, the assay platewas left in an orbital shaker for 3 h at room temperature at 260 rpm. Thereaction was quenched with the addition of 60 ml of 100% methanol ineach well, the mixture was centrifuged for 10 min at 4,000 rpm in a mi-croplate centrifuge, and the supernatant was used for estimation of prod-uct concentration by HPLC. The mutant clones with improved activitywere subjected to reconfirmation by activity measurements from shakeflask expression experiments, and the identities of the mutations wereconfirmed by DNA sequencing.

Preparation of variants of deacetylcephalosporin C synthase en-zyme for process reaction. Native and selected mutant variants of DACSwere expressed in LB medium at larger volumes as described above. About20 g of cell pellets of variants of DACS, including native hydroxylase, inseparate experiments, was suspended in 50 ml of buffer containing 50 mMphosphate and 0.6 M NaCl. After stirring the suspension for 30 min at4°C, 40 mg of lysozyme was added and stirred for another 30 min at 4°C.The contents were sonicated with a Labsonic M ultrasonic processor,using a titanium probe of 10 mm in diameter, with a cycle of 0.6 s andamplitude of 70% for 30 min in 3 cycles of 10 min each with constantstirring at 4°C. After sonication, 2.5 ml of 10% polyethyleneimine (PEI)was added and stirred for 2 h at 4°C, and the mixture was centrifuged at13,000 rpm for 30 min. The protein concentration in the supernatant wasestimated using the Bradford method, and the soluble enzyme was storedat �80°C until further use in process reactions.

Purification of mutants of deacetylcephalosporin C synthase en-zyme. About 8 g of recombinant E. coli BL21(DE3) cell pellets of variantsof DACS, in separate experiments, were suspended in 80 ml of resuspen-sion buffer containing 50 mM Tris-HCl (pH 8.0), 0.01 mM EDTA, 1 mMDTT, 10% glucose, and 10% glycerol. After complete suspension, 10 mgof lysozyme was added and stirred for a further 30 min. The contents weresonicated as described above for 20 min in 2 cycles of 10 min each withconstant stirring at 4 to 8°C. After sonication, 400 �l of 10% streptomycinsulfate and 80 �l of 10% PEI were added to the sonicated mix and stirredfor another 20 min. Subsequently, the entire mixture was centrifuged at10,000 rpm for 30 min at 4°C. The supernatant obtained from the earlierstep was transferred to an XK16/20 DEAE-Sepharose FF column equili-brated with buffer A containing 50 mM Tris-HCl (pH 8.0), 0.01 mMEDTA, and 1 mM DTT. After washing, bound protein was eluted withbuffer B containing 50 mM Tris-HCl (pH 8.0), 0.01 mM EDTA, 1 mMDTT, and 1 M NaCl using stepwise gradients of 0 to 12%, 12 to 18%, 18 to25%, 25 to 40%, and 40 to 100%. Fractions collected from the 12 to 18%and 18 to 25% gradients were pooled, diluted four times and loaded ontoan XK16/20 Q Sepharose FF column equilibrated with buffer A. The

Balakrishnan et al.

3714 aem.asm.org July 2016 Volume 82 Number 13Applied and Environmental Microbiology

on July 1, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

bound protein was subsequently eluted using buffer B with stepwise gra-dients as described above. The DACS enzyme eluted from the 18 to 25%gradient was concentrated using Amicon Ultra-15 centrifugal filter units,and the concentrated protein solution was injected into a Superdex 75pggel filtration column equilibrated with buffer A. Fractions containing ac-tive protein were pooled, concentrated using Amicon Ultra-15 centrifugalfilter units, and subjected to gel filtration again using the Superdex 75pgcolumn. To this, glycerol was added to a final concentration of 10% andstored.

Determination of kinetic parameters of mutants of deacetylcepha-losporin C synthase. The kinetics parameters (Km and kcat) were deter-mined using three independent experiments for both wild-type deacetyl-cephalosporin C synthase and selected mutants. Each experiment wasconducted at substrate concentrations of 0.5, 1, 2, 3, 4, 5, 6, 8, and 10 mMusing quantities of proteins specific to each variant enzyme in a totalreaction volume of 300 �l. The consumption of substrate at all concen-trations was less than 10% and displayed a linear reaction profile. Allmeasurements were performed using 50 �l of a 6� stock solution ofcofactors (30 mM 2-oxoglutarate, 15 mM ascorbate, 1.5 mM FeSO4, and200 mM Tris-HCl [pH 7.8]) mixed with 100 mM CephG in volumesappropriate to obtain the desired substrate concentration, and water wasadded to make up the volume to 250 �l. The reaction was initiated byadding 50 �l of the enzyme solution and was carried out at 25°C bycontinuously mixing the reactants at 1,000 rpm using an Eppendorf Mix-mate. The reaction was terminated at 15 min by the addition of an equalvolume of 100% methanol. The reaction mix was vortexed, incubated for5 min, and centrifuged at 13,000 rpm, and the supernatant was used forquantitative estimation by HPLC. The amount of product formed wasestimated using a standard linear plot created using a DAG standard. Thekinetic parameters were estimated by fitting experimental data using theLineweaver-Burk double-reciprocal method.

Process-level biotransformation reaction of CephG to DAG by mu-tants of deacetylcephalosporin C synthase. Amounts of 360 mg ofCephG and 120 mg of NaHCO3 were mixed in 6 ml of demineralized(DM) water (pH �6.9) in a beaker. Upon complete dissolution of CephG,the cofactors 2-oxoglutaric acid, ascorbic acid, and FeSO4 were added togive final concentrations of 84 mM, 27 mM, and 0.3 mM, respectively,followed by the addition of a 1% concentration of the antifoaming agentpoly propylene glycol (PPG) to control frothing. After stirring the mixturethoroughly, 158 mg of selected mutants of DACS enzyme were added tobeakers, and the reaction volume was made up to 20 ml using DM water.The reaction pH was maintained at 6.8 at room temperature, and theprogress of the reaction was monitored by HPLC at regular intervals untilcompletion.

Process-level biotransformation of deacetylcephalosporin G toHACA by CefFGOS. Upon completion of the biotransformation reactionof CephG to DAG by CefFGOS as described above, the pH of the solutionwas adjusted to 7.5 using 13% ammonia, and 60 units of recombinantPenG amidase from Alcaligenes faecalis was added to facilitate the hydro-lysis of the phenylacetyl moiety. Quantitative monitoring and analysis ofHACA formation were performed using HPLC. HACA was found to elutewith a retention time of 0.3 min. Further refinement of quantitation ofHACA was performed by a modification of the elution profile with an SSIntersil ODS-3V column(250 mm by 4.6 mm; 5 �m), using an elutionbuffer containing 0.14% tetra-n-butyl-ammonium hydrogen sulfate in30% acetonitrile with a flow rate of 1 ml/min. DAG and HACA werefound to elute at 12.6 and 3.2 min, respectively.

RESULTS AND DISCUSSIONBiotransformation of CephG by native deacetylcephalosporin Csynthase. The decaetylcephalosporin C synthase gene (cefF) fromS. clauvuligerus was cloned in E. coli, and the recombinant clonewas confirmed by restriction digestion of the isolated plasmidand by DNA sequencing. The enzyme was expressed in E. coliBL21(DE3), and the crude extract was assayed for its activity un-

der standard assay conditions, except 1 mM CephG was used asthe substrate (11). When the HPLC chromatogram was examined,no detectable peak could be observed, and this was further con-firmed by spiking with the reference standard. As CephG was anunnatural substrate, the assay was repeated under higher substrateconcentrationss of 2.5 mM, 5 mM, and 10 mM. Examination ofchromatograms revealed the presence of a small peak correspond-ing to the retention time of the product (DAG) in assay samplescontaining 10 mM CephG. The observation of a barely detectablepeak indicated negligible catalytic activity and hence poor affinityand activity toward the unnatural substrate CephG. Analysis bySDS-PAGE revealed that the native enzyme was found predomi-nantly in inclusion bodies. The wild-type enzyme was found to belabile and lost activity rapidly, as no activity could be detected after24 h when stored at 4°C.

High-throughput screening of the mutagenic enzyme libraryby HPLC. The challenge in modifying multiple parameters ofDACS, such as catalytic efficiency, stability, and solubilization,required the generation and screening of a large library of putativemutant clones. This necessitated first the development of a rapidmethod for the detection of DAG in mutant clones. The initialmethod of detecting the formation of DAG at 260 nm involved alinear gradient elution profile consisting of mobile phase A (6.25mM ammonium acetate [pH 3.2]) and mobile phase B (80%methanol, 20% mobile phase A) at a flow rate of 1 ml/min in aSymmetry C18 column (75 by 4.6 mm; 3.5 �m) with a run time of30 min. DAG and CephG were found to elute with retention timesof 8.2 and 12.5 min, respectively. In order to reduce the run timeand facilitate rapid screening of a large number of clones, thecolumn length was reduced to 50 mm, elution profiles were al-tered to isocratic conditions with a modified buffer based on ace-tonitrile and phosphate, and the flow rate was increased to 2.5ml/min as described in Materials and Methods. This led to a re-duction in the run time from 30 min to 1.8 min, with substantialresolution between DAG and CephG, eluting at 0.55 and 1.1 min,respectively (see Fig. S1 in the supplemental material). This rapidmethod was found to be effective in screening the large number ofclones generated.

Evolution of mutant deacetylcephalosporin C synthase en-zymes. (i) Random mutagenesis by biasing dNTPs concentra-tions. cefF was subjected to random mutagenesis by biasing dNTPconcentrations and using error-prone Taq DNA polymerase inthe first round of evolution as described in Materials and Meth-ods. The assay duration was extended for 3 h in order to detectputative mutant enzymes with enhanced stability. Screening ofapproximately 20,000 clones in equal proportion and further con-firmation by activity analysis using normalized cell pellets led toidentification of point mutation isolates, such as P72L, T90A,V150A, P186L, V221A, V221T, M229V, T273A, A311V, Y38CT90A, and E16G T90A T304A. A model of the wild-type DACSstructure was produced by MODELLER using the structure ofDAOCS (Fig. 2A) as obtained through the ModBase database(24). The structure as such available in ModBase server is an apostructure. The Fe(II) atom was transferred into the DACS modelat its corresponding coordinate position consistent with the tem-plate DAOCS crystal structure (PDB ID 1UOB), with the help ofSwissPDBViewer v 4.0.2 software (25). The model structures ofDACS were also produced using SwissPDBViewer. The spatialorganization and binding sites for substrate and cosubstrate havebeen mapped and appear to be highly conserved between DAOCS

Modified Deacetylcephalosporin C Synthase

July 2016 Volume 82 Number 13 aem.asm.org 3715Applied and Environmental Microbiology

on July 1, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

and DACS (Fig. 2B). T90 is located in a highly conserved region atpositions 91 to 96 of DAOCS, DACS, and DAOCS/DACS, tenta-tively designated the N-terminal region (11) (see Fig. S3 in thesupplemental material) and was identified repeatedly under mul-tiple PCR conditions. P72, located in close proximity to theArg75/76 residues (suggested to be involved in substrate recogni-tion) and A311, is located in the C-terminal region implicated inenhanced catalysis as deduced from functional analysis of a struc-tural analogue of DAOCS (12). As a result, these three sites werecombined by site-directed mutagenesis for further evolution.When the combination isolates were expressed and assayed, oneof the mutant isolates, denoted the CefFTM mutant, was found toshow 2.3-fold-enhanced activity for biotransformation of CephGin relation to the wild type. Table 1 lists representative mutantswith the lower and upper limits of relative activity observed duringeach round of evolution, while the detailed list is included in TableS1 in the supplemental material.

(ii) Random mutagenesis by varying the template concentra-tions. cefF is GC rich, and the mutational bias exhibited by TaqDNA polymerase could skew the representation of random mu-tant libraries and possibly reduce the size of the mutant collection.In order to disclose further mutations causing enhanced activity,the strategy was modified to gradually accumulate single muta-tions by varying the template concentrations, which was followedby sequential in vitro recombination of these single mutations (26,27). Random mutagenesis with the GeneMorph II kit as recom-mended by the supplier by using the cefFTM template withamounts of 100, 200, and 400 ng each in three separate reactionswas performed as described in Materials and Methods. Screeningof the resulting mutant library of nearly 15,000 clones led to thedisclosure of 13 additional mutations (P7L, A40V, T51M, F195L,V206I, A210V, V226I, M229I, M233I, A237V, R250L, E258K, andA273A) and generation of variants containing 1 or 2 mutagenicsubstitutions. Many of these mutations appeared to be locatedin and around the active site (Fig. 2B), and hence additional

combinations containing 1 or 2 additional mutations wereconstructed by site-directed mutagenesis. In addition, in a par-allel investigation, the T90 and V221 residues were subjected tosite-directed saturation mutagenesis using cefFTM as the tem-plate and the 5=CCCAGGTGNNSAGAACCGGTTCCTACACGGACTACTC3= and 5=GGCGTCCTCCACSNNCGGCAAGCTTACCAGTTC3= oligonucleotides, respectively, to identifysubstitutions with improved activity. Activity analysis of thismutagenic enzyme library revealed 2.6- to 4.4-fold increases withreference to the wild-type enzyme (Table 1; see Table S1 in thesupplemental material). The mutant variant with a 4.4-fold im-provement in activity was denoted the CefFM mutant, and its cor-responding gene served as the parent for the third round of evo-lution. As variation of the template concentration was found todisclose a significant number of beneficial mutations, this wasmaintained as the primary vehicle for random mutagenesis usingthe GeneMorph kit for further generation of a diverse mutageniclibrary in conjunction with site saturation and quantitative struc-ture-activity relationship (QSAR) prediction-based site-directedmutagenesis. Random mutagenesis using 100 ng of cefFM templateand screening of approximately 20,000 clones led to the disclosureof 13 novel mutations (E82D, V171L, V171M, A177V, R182S,M184I, I193V, E209Q, L236V, V249I, S260G, S251F, and F267L)in a collection of 11 mutant isolates (Table 1; see Table S1 in thesupplemental material), and many of these residues were found tomap in the active-site region. In addition, R182W, identified byQSAR-like predictions for enhanced activity, was subjected to sat-uration site-directed mutagenesis using cefFM as the template andthe 5=GGC CAT CCG SNN CGG CTC GTG CTC CGC GGA CCGGTG C3= oligonucleotide, while V221P and V221H was intro-duced by site-directed mutagenesis. The V171L, I193V, R182W,and F267L mutations, located in the active-site region, were se-quentially introduced by site-directed mutagenesis to generate ad-ditional variants. Activity measurements of these isolates showedan improvement ranging from 5.3- to 10-fold with respect to the

FIG 2 Homology modeling of DACS structure and mapping of mutant positions. (A) Crystal structure of DAOCS as obtained from PDB (PDB ID 1UOB). (B)Modeled structure and mapping of mutations in DACS. The N-terminal region (blue circle), C-terminal arm (green), and positions of the Arg75 and -6 residuesare highlighted. The substrate CephG and the cosubstrate (�KG) are represented in stick models in magenta and pink, respectively. The Fe(II) is represented assphere in red, and residues altered by mutations are denoted in stick models.

Balakrishnan et al.

3716 aem.asm.org July 2016 Volume 82 Number 13Applied and Environmental Microbiology

on July 1, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

wild type (Table 1; see Table S1 in the supplemental material).With a view to improving expression activity to a high level, thegene for a mutant with 10-fold-improved activity, named cefFS,was subjected to codon optimization, resulting in cefFOS. BothcefFS and cefFOS were used as templates for evolution of additionalvariants through random mutagenesis by using 200 ng of templatein the next round. Screening of clones led to the development of 11novel mutations (M53L, R91G, T96S, G108D, I149T, V226I,A241V, G255D, V307A, A311M, and N313D). Additional com-binations with selected residues using site-directed mutagene-sis led to the creation of variants with activity enhancementranging from 11- to 19-fold (Table 1; see Table S1 in the sup-plemental material). The mutant isolate with the highest activ-ity was designated CefFGOS.

Expression profiling and kinetic characterization. In order todeduce the functional role of these mutations, cefFTM, cefFM, cefFS,cefFOS, and cefFGOS clones, along with cefFW (wild type), wereexpressed, and their crude extracts were prepared and analyzed bySDS-PAGE (see Fig. S2 in the supplemental material). SDS-PAGEanalysis showed the presence of detectable amount of wild-typedeacetylcephalosporin C synthase in the soluble fraction, while amajority of the expressed protein was found to be in inclusionbodies. The mutant variants showed increasing levels of solubili-zation with a corresponding reduction in inclusion body forma-

tion, suggesting that several mutations possibly enhanced foldingof protein. It was also observed, as expected, that the total cellularexpression of deacetylcephalosporin C synthase was much higherfor cefFOS and cefFGOS. Determination of the total protein, activity,and relative activity (Table 2) of selected mutants further rein-forced the SDS-PAGE observation. In these mutants, the relativeactivity had increased many fold but the enhancement in proteinconcentration was marginal. This may be because several mutationshad altered the kinetic properties of the enzyme as well. Hence, theenzymes from these select variants were further purified to near ho-mogeneity, and their kinetic parameters were determined. The resultsare presented in Table 3. It is observed that in addition to enhance-ment in the levels of expression and solubilization, the Km increasedby 1.7-fold, while kcat increased by about 35-fold, for CefFGOS com-pared to CefFW (wild type). The increase in Km along with the in-crease in kcat can be attributed to the active-site enlargement, whichcauses the enzyme to release the product efficiently, coupled withweak substrate binding leading to high kcat and Km (28). The kineticparameters Km, kcat, and kcat/Km reported in this study for deacetyl-cephalosporin C synthase correlate well with the reported values fordeacetoxycephalosporin C synthase and its substrate analogue peni-cillin G (20).

Mapping of mutations and their mechanistic implications.Mapping analysis of mutations in deacetylcephalosporin C syn-

TABLE 1 Representative list of DACS mutants and their relative activities

Round of evolution(template) Mutant

No. ofmutations Mutation(s)

Relative activity(fold)a

1 (CefFW/cefFW) W5 1 V221T 1.2W1 1 P72L 1.3W3 1 P186L 1.5W10 3 E16G T90A T304A 2.1W11 2 T90A�P72Lb 1.9W12 2 T90A�A311Vb 2.1W13 3 T90A�P72L�A311V (CefFTM)b 2.3

2 (cefFTM) TM2 5 T90A�P72L�A311V�V206I�A210V 2.6TM3 5 T90A�P72L�A311V�P7L�A237V 2.8TM10 4 T90A�P72L�A311V�R250L 3.8TM9 4 T90A�P72L�A311V�F195L 4.1TM15 5 T90A�P72L�A311V�P7L�T273Ab 3.4TM13 6 T90A�P72L�A311V�V206I�A210V�T273Ab 3.6TM23 6 T90A�P72L�A311V�A40V�M229I�T273A (CefFM)b 4.4

3 (cefFM) M7c 7 T90A�P72L�A311V�A40V�M229I�T273A�V221P 5.3M9d 7 T90A�P72L�A311V�A40V�M229I�T273A�R182S 5.8M12 7 T90A�P72L�A311V�A40V�M229I�T273A�S260G 6.6M13 7 T90A�P72L�A311V�A40V�M229I�T273A�S251F 7.5M24 9 T90A�P72L�A311V�A40V�M229I�T273A�M184I�I193V�F267Lb 5.0M26 8 T90A�P72L�A311V�A40V�M229I�T273A�V171L�F267Lb 8.0M30 9 T90A�P72L�A311V�A40V�M229I�T273A�V171L�R182W�F267L (CefFS)b 10.0

4 (cefFS) S5 10 T90A�P72L�A311V�A40V�M229I�T273A�V171L�R182W�F267L�V226I 11.0S1 10 T90A�P72L�A311V�A40V�M229I�T273A�V171L�R182W�F267L�N313D 12.0S2 10 T90A�P72L�A311V�A40V�M229I�T273A�V171L�R182W�F267L�R91G 13.0S6 11 T90A�P72L�A311V�A40V�M229I�T273A�V171L�R182W�F267L�A241V�V307A 15.0S17 12 T90A�P72L�A311V�A40V�M229I�T273A�V171L�R182W�F267L�T96S�G255D�A280Sb 13.0S18 12 T90A�P72L�A311V�A40V�M229I�T273A�V171L�R182W�F267L�T96S�A241V�V307Ab 15.0S15 12 T90A�P72L�A311V�A40V�M229I�T273A�V171L�R182W�F267L�R91G�A241V�V307Ab 17.0

5 (cefFOS) OS1 10 T90A�P72L�A311V�A40V�M229I�T273A�V171L�R182W�F267L�G108D (CefFGOS) 19.0

a Fold enhancement in activity of the mutant variant DACS enzyme relative to the activity of the wild-type DACS enzyme.b Combination generated from the indicated round by site-directed mutagenesis.c Site saturation mutagenesis at position 221.d Site saturation mutagenesis at position 182.

Modified Deacetylcephalosporin C Synthase

July 2016 Volume 82 Number 13 aem.asm.org 3717Applied and Environmental Microbiology

on July 1, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

thase indicates that majority of mutations (65%) are located in aconformationally flexible nonstructural region (12, 13, 14) (Fig.2B). Alignment of amino acid mutations identified in this study byClustal W (29) with DAOCS and DAOCS/DACS revealed that 11(Y38C, R91N, T96S, G108D, I193V, F195L, M233I, A237V,G255D, T273A, and T304G) of the 44 residues in DACS have beenmutated to residues found in native DAOCS (see Fig. S3B in thesupplemental material), while 5 (V150A, F195L, V221T, G255D,and T304G) of the 44 residues have been mutated into nativeDAOCS/DACS (see Fig. S3C in the supplemental material) andthree of the mutations (F195L, G255D, and T304G) remain com-mon between the DAOCS, DACS, and DAOCS/DACS enzymes.When cefE, cefF, and cefEF were expressed under identical condi-tions in E. coli and compared, it was found that the level of solubleexpression was highest for cefEF, while cefE had the lowest level ofexpression (unpublished data). This suggests that 11 of the 44mutations are likely to enhance the activity of the deacetylcepha-losporin C synthase enzyme. This again correlates with their loca-tion within the vicinity of the putative active site, except for G255and V150, which could affect folding and hence enhance solubleexpression. The V307A, A311M, A311V, and N313D mutationslie in the C-terminal arm of deacetylcephalosporin C synthase,and the C-terminal region has been implicated in various roles,such as enclosure of the active site, a shelter for reactive interme-diates, forming a “lid” over the active site, and binding and/ororienting the penicillin substrate during catalysis in case of deace-

toxycephalosporin C synthase (12). The T304G mutation createsthe highly conserved IGGNY C-terminal sequence (postitions 304to 307) and was previously identified to be part of the hinge sec-tion (positions 299 to 302 in DAOCS) and suggested to facilitatemovement of the C-terminal arm during catalysis (13). Sequencecomparison of DAOCS, DACS, and DAOCS/DACS reveals a con-served stretch of residues spanning the region from position 75 to105 (see Fig. S3A in the supplemental material), designated theN-terminal region, and its functional significance remains un-known (11). Site-directed mutagenesis studies of mutation ofArg74 (Arg75 in DACS) of DAOCS to hydrophobic residue Ile orhydrophilic Gln (i.e., R74I or R74Q) show uncoupling of 2-oxo-glutarate and penicillin N/penicillin G oxidation, although differ-entially, hence suggesting a role for substrate recognition and se-lection (12). The P72L, E82D, T90A, T90G, T90D, R91G, R91N,T96S, and G108D mutations are mapped as part of the N-terminalregion (Fig. 2B). The evolution of enhancement in catalytic effi-ciency was found to be incremental except in the case of singlepoint mutations such as R91G and G108D, where the enhance-ment was found to be substantial, and these are found to map inthe N-terminal region, thereby strengthening the possibility for arole in substrate recognition and selection. Closer scrutiny of themapping of the A177V, R182W/S, M184I, P186L, I193V, V221A,V226I, M229I, and V249I residues suggests that they are located inthe immediate vicinity of the putative active site as deduced fromour model. The F195 and F267 (�11) residues, located in � sheets6 and 11, respectively, are an integral part of the active site, andF267 is located within one of the most highly conserved stretchesof the deacetylcephalosporin C synthase sequence, containingR262 (see Fig. S3A in the supplemental material). Invariably, all ofthese residues carry either an isoleucine or leucine mutation,which possibly shifts the putative active site of deacetylcepha-losporin C synthase from a smaller and mildly polar environmentto an enlarged hydrophobic environment. M180/R258 in DAOCS(M184/R262 in DACS) is known to form a binding packet forhighly polar 2-oxoglutarate during the catalysis by DAOCS (14).Substitution of adjacent residues by hydrophobic residues mightretard the binding of 2-oxoglutarate while facilitating the bindingand departure of the hydrophobic and bulky substrate CephG andthus can rationalize the observed increase in Km and kcat. Cou-pling/decoupling studies of 2-oxoglutarate and substrate and theirkinetics can lead to further dissection of functional roles of theseresidues.

Biotransformation of CephG to DAG to HACA and indus-trial utility of CefFGOS. To assess the potential for application asan industrial biocatalyst, the CefFTM, CefFM, CefFS, CefFOS, andCefFGOS mutants along with CefFW were evaluated for biotrans-formation of 1.8% CephG, and the resulting data are presented inTable 4. It is observed that the CefFOS and CefFGOS variants could

TABLE 2 Expression and activity profile of wild-type and mutantDACSa

Enzyme MutationsActivity(Ub/ml)

Proteinconcn(mg/ml)

Relativeactivity(fold)c

CefFW None (wild type) 0.020 0.777 1CefFTM P72L�T90A�A311V 0.045 1.058 2.3CefFM CefFTM � A40V�M229I�

T273A0.096 1.090 4.8

CefFS CefFM � V171L�R182W�F267L

0.168 1.093 8.4

CefFOS CefFM � V171L�R182W�F267L

0.258 1.100 12.9

CefFGOS CefFOS � G108D 0.308 1.214 15.4a Mutant variants of DACS and the wild type were expressed, and their expressionprofile and activity were determined using crude extracts.b One unit of enzyme activity is defined as the amount of DACS required to cause theformation of 1 �mol of DAG per minute from CephG at 25°C and pH 7.c Fold enhancement in the activity of the mutant variant enzyme relative to the activityof the wild-type enzyme.

TABLE 3 Kinetic parameters of DACS for the wild type and mutants

Enzyme Km (mM)a kcat (s�1)a

kcat/Km

(M�1 s�1)Relativekcat/Km (%)b

CefFW (wild type) 2.58 0.20 0.055 0.005 21.4 100CefFTM 2.58 0.11 0.122 0.007 47.3 222CefFM 4.64 0.28 0.326 0.002 70.33 330CefFS 7.24 0.68 1.805 0.08 249.3 1166CefFOS 9.20 0.71 2.302 0.049 250.2 1,170CefFGOS 4.41 0.14 1.899 0.01 430.6 2,022a Mean standard error from three independent experiments.b Percentage of the kcat/Km value of the mutant DACS enzyme relative to the kcat/Km

value of the wild-type DACS enzyme.

TABLE 4 Process-level bioconversion of CephG by wild-type DACS andmutants

Enzyme Reaction time (min) Conversion (%)

CefFW (wild type) 240 4.9CefFTM 240 20.4CefFM 240 34.0CefFS 240 87.6CefFOS 150 100CefFGOS 90 100

Balakrishnan et al.

3718 aem.asm.org July 2016 Volume 82 Number 13Applied and Environmental Microbiology

on July 1, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

convert 100% of the substrate in 150 and 90 min, respectively,which demonstrates the potential for large-scale transformations(Fig. 3A). When the concentration of CephG was doubled to3.6%, CefFGOS was found to transform equally well. Upon com-pletion of the transformation, the pH of the transformation mix-ture of CefFGOS was adjusted to 7.5 using ammonia, the phenyla-cetyl moiety was hydrolyzed by PenG amidase, and the resultswere analyzed by HPLC. HACA was found to elute with a reten-tion time of 0.3 min, and the reaction was nearing completion inabout 30 min, with a basal level of residual DAG seen (Fig. 3B). Incontrast to native enzyme, the CefFGOS enzyme variant was foundto be stable for 7 days at 4 to 8°C without loss of any catalyticactivity, thus removing an additional impediment towards devel-opment of an alternate process for the production of HACA and7ACA.

Future directions. In this study, numerous deacetylcepha-losporin C synthase mutants were identified, and one of the se-lected variants, CefFGOS, showed increased activity for conversionof CephG to deacetylcephalosporin G. This mutant is likely to playan important role in the commercialization of a novel route forHACA manufacture. Mapping of mutations led to identificationof potential residues and regions for further directional alterationthat can enhance catalytic activity significantly, thus simplifyingcommercial scalability. In an equally significant development, lo-calization of mutations and deduction of their possible role in

catalytic activity have led to the disclosure of critical regions in-volved in catalysis and the putative active site. Further character-ization using molecular, biochemical, and structural studies canfacilitate fundamental understanding of mechanism of catalysis,particularly pertaining to the choice of substrates, the process oftheir selection, and regulation of the catalytic activity, in the nearfuture.

ACKNOWLEDGMENTS

This work was supported by Orchid Chemicals & Pharmaceuticals Ltd.,Chennai, India.

We thank K. B. Ramachandran, IIT Chennai, for critical reading of themanuscript.

FUNDING INFORMATIONThe research work was done as part of in-house research activity sup-ported by the organization.

REFERENCES1. Elander RP. 2003. Industrial production of �-lactam antibiotics. Appl

Microbiol Biotechnol 61:385–392. http://dx.doi.org/10.1007/s00253-003-1274-y.

2. Brakhage AA. 1998. Molecular regulation of �-lactam biosynthesis infilamentous fungi. Microbiol Mol Biol Rev 62:547–585.

3. Salehpour P, Reza RY, Hajmohammadi R. 2012. Determination ofoptimal operation conditions for production of cephalosporin G from

FIG 3 Process-level biotransformation of CephG to HACA by CefFGOS. (A) CephG (1.8%) was treated with CefFGOS and analyzed by HPLC. The elution profileillustrates complete bioconversion of CephG into DAG, with its retention time marked at its peak. (B) HPLC elution profile illustrating near-completebioconversion of DAG formed in the previous reaction into HACA by PenG amidase.

Modified Deacetylcephalosporin C Synthase

July 2016 Volume 82 Number 13 aem.asm.org 3719Applied and Environmental Microbiology

on July 1, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

penicillin G potassium. Org Process Res Dev 16:1507–1512. http://dx.doi.org/10.1021/op300076q.

4. Bianchi D, Bortolo R, Golini P, Cesti P. 1998. Enzymatic transformationof cephalosporin C to 7-ACA by simultaneous action of immobilized D-amino acid oxidase and glutaryl-7-ACA acylase. Appl Biochem Biotech-nol 73:257–268. http://dx.doi.org/10.1007/BF02785660.

5. Pollegioni L, Rosini E, Molla G. 2013. Cephalosporin C acylase: dreamand (/or) reality. Appl Microbiol Biotechnol 97:2341–2355. http://dx.doi.org/10.1007/s00253-013-4741-0.

6. Conti G, Pollegioni L, Molla G, Rosini E. 2014. Strategic manipulationof an industrial biocatalyst-evolution of a cephalosporin C acylase. FEBS J281:2443–2455. http://dx.doi.org/10.1111/febs.12798.

7. Aharonowitz Y, Cohen G, Martin JF. 1992. Penicillin and cephalosporinbiosynthetic genes: structure, organization, regulation and evolution.Annu Rev Microbiol 46:461– 495. http://dx.doi.org/10.1146/annurev.mi.46.100192.002333.

8. Hewitson KS, Granatino N, Welford RWD, Mcdonough MA, SchofieldCJ. 2005. Oxidation by 2-oxoglutarate oxygenases: non-haem iron sys-tems in catalysis and signaling. Philos Trans R Soc A 363:807– 828. http://dx.doi.org/10.1098/rsta.2004.1540.

9. Jensen SE, Westlake DWS, Wolf S. 1985. Deacetoxycephalosporin Csynthetase and deacetoxycephalosporin C hydroxylase are two separateenzymes in Streptomyces clavuligerus. J Antibiot 2:263–265.

10. Kovacevic S, Miller JR. 1991. Cloning and sequencing of the �-lactamhydroxylase gene (cefF) from Streptomyces clavuligerus: gene duplicationmay have led to separate hydroxylase and expandase activities in the Ac-tinomycetes. J Bacteriol 173:398 – 400.

11. Baker BJ, Dotzlaf JE, Yeh WK. 1991. Deacetoxycephalosporin C hydrox-ylase of Streptomyces clavuligerus. J Biol Chem 266:5087–5093.

12. Oster LM, van Scheltinga AC, Valegård K, Hose AM, Dubus A, HajduJ, Andersson I. 2004. Conformational flexibility of the C terminus withimplications for substrate binding and catalysis revealed in a new crystalform of deacetoxycephalosporin C synthase. J Mol Biol 343:157–171. http://dx.doi.org/10.1016/j.jmb.2004.07.049.

13. Lloyd MD, Lee HJ, Harlos K, Zhang ZH, Baldwin JE, Schofield CJ,Charnock JM, Garner CD, Hara T, van Scheltinga ACT, Valegard K,Viklund JAC, Hajdu J, Andersson I, Danielsson Å, Bhikhabhai R. 1999.Studies on the active site of deacetoxycephalosporin C synthase. J Mol Biol287:943–960. http://dx.doi.org/10.1006/jmbi.1999.2594.

14. Lee H-J, Lloyd MD, Harlos K, Clifton IJ, Baldwin JE, Schofield CJ. 2001.Kinetic and crystallographic studies on deacetoxycephalosporin C syn-thase (DAOCS). J Mol Biol 308:937–948. http://dx.doi.org/10.1006/jmbi.2001.4649.

15. Lee HJ, Schofield Lloyd MD. 2002. Active site mutations of recombinantdeacetoxycephalosporin C synthase. Biochem Biophys Res Commun 292:66 –70.

16. Schofield CJ, Zhang Z. 1999. Structural and mechanistic studies on2-oxoglutarate-dependent oxygenases and related enzymes. Curr OpinStruct Biol 9:722–731. http://dx.doi.org/10.1016/S0959-440X(99)00036-6.

17. Wei C-L, Yang Y-B, Wang W-C, Liu W-C, Hsu J-S, Tsai Y-C. 2003.

Engineering Streptomyces clavuligerus deacetoxycephalosproin C synthasefor optimal ring expansion activity toward penicillin G. Appl EnvironMicrobiol 69:2306 –2312. http://dx.doi.org/10.1128/AEM.69.4.2306-2312.2003.

18. Chin HS, Sim J, Sim TS. 2001. Mutation of N304 to leucine in Strepto-myces clavuligerus deacetoxycephalosporn C synthase creates an enzymewith increased penicillin analogue conversion. Biochem Biophys ResCommun 287:507–513. http://dx.doi.org/10.1006/bbrc.2001.5552.

19. Hsu JS, Yang YB, Deng CH, Wei CL, Liaw SH, TsaiYC. 2004. Familyshuffling of expandase genes to enhance substrate specificity for penicillinG. Appl Environ Microbiol 70:6257– 6263. http://dx.doi.org/10.1128/AEM.70.10.6257-6263.2004.

20. Wei C-L, Yang Y-B, Deng C-H, Liu W-C, Hsu J-S, Lin Y-C, Liaw S-H,Tsai Y-C. 2005. Directed evolution of Streptomyces clavuligerus deace-toxycephalosporin C synthase for enhancement of penicillin G expansion.Appl Environ Microbiol 71:8873– 8880. http://dx.doi.org/10.1128/AEM.71.12.8873-8880.2005.

21. Sambrook J, Fritsch EF, Maniatis T. 1989. Molecular cloning: a labora-tory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold SpringHarbor, NY.

22. Bradford MM. 1976. A rapid and sensitive method for the quantitation ofmicrogram quantities of protein utilizing the principle of protein-dyebinding. Anal Biochem 72:248 –254. http://dx.doi.org/10.1016/0003-2697(76)90527-3.

23. Barik S. 1996. Site-directed mutagenesis in vitro by megaprimer PCR.Methods Mol Biol 57:203–215.

24. Kazlauskas R, Lutz S. 2009. Engineering enzymes by ‘intelligent’ design.Curr Opin Chem Biol 13:1–2. http://dx.doi.org/10.1016/j.cbpa.2009.02.022.

25. Tracewell CA, Arnold FH. 2009. Directed enzyme evolution: climbingfitness peaks one amino acid at a time. Curr Opin Chem Biol 13:3–9.http://dx.doi.org/10.1016/j.cbpa.2009.01.017.

26. Pieper U, Webb BM, Barkan DT, Schneidman DD, Schlessinger A,Braberg H, Yang Z, Meng EC, Pettersen EF, Huang CC, Datta RS,Sampathkumar P, Madhusudhan MS, Sjolander K, Ferrin TE, BurleySK, Andrej Sali A. 2011. MODBASE, a database of annotated compara-tive protein structure models and associated resources. Nucleic Acids Res39:465– 474.

27. Guex N, Peitsch MC. 1997. SWISS-MODEL and the Swiss-PdbViewer:an environment for comparative protein modeling. Electrophoresis 18:2714 –2723. http://dx.doi.org/10.1002/elps.1150181505.

28. Chen J, Chen H, Shi Y, Hu F, Lao X, Gao X, Zheng H, Yao W. 2013.Probing the effects of the non-active-site mutation Y229W in New Delhimetallo-�-lactamase-1 by site-directed mutagenesis, kinetic studies andmolecular dynamic simulations. PLoS One http://dx.doi.org/10.1371/journal.pone.0082080.

29. Thompson JD, Higgins DG, Gibson TJ. 1994. CLUSTAL W: improvingthe sensitivity of progressive multiple sequence alignment through se-quence weighting, position-specific gap penalties and weight matrixchoice. Nucleic Acids Res 22:4673– 4680. http://dx.doi.org/10.1093/nar/22.22.4673.

Balakrishnan et al.

3720 aem.asm.org July 2016 Volume 82 Number 13Applied and Environmental Microbiology

on July 1, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from