Purification and properties of isopenicillin N epimerase from Streptomyces clavuligerus

78

Bioch/mica et BwphyaicaActa. 999 (1989) 78-85

Elsevier BBAPRO33492

Purification and properties of isopenicillin N epime;:ase from Streptomyces clavuligerus Shigeyuki U s u i a n d C h a n g - A n Y u Departmentof Biochemistry, OAFS. Oklahoma State University,Stillwater, OK (U.S.A.)

(Received5 June 1989,~ Key words: lsopenicillinN Epimerase:Substratespecificity

Isopenicillin N epimerase, which catalyzes conversion of isopenicillin N to penicillin N, has beel, purified to electrol~retic homogeneity from the cell.free extract of Strcptomyces clavuligems by a procedure involving ammonium sulfat~ fractionstion and chromstographles with DE-g2, DEAE Affi-gel blue, Sephedex G-200, calcium phosphate-cellulose, and Memo Q. The purified ephnerase is monomerie with a molecular weight of 47000 or 50000 as estimated by SDS.polyacrylmnide gel electropheresis or gel filtration, respectively. The enzyme contains I tool of pyridoxal S'-phesphate per mol of protein, and shows absorption maxima at 280 and 420 urn. The epimerase catalyzes the complete 'racemization' on both the L-a-uminomlipyl side-chain of isopenieillin N and the D-a-uminoedipyl side-chain of penicillin N, so that Jut approxima|ely equimolar mixture of the two penicillins is produced. The mixture is not truly racemie, since these peniciHins are diasteremne,s rather than optical isomers. The chemical modification of primary

amino groups of the epimerase by flnoreseamine results in a great loss of the enzyme activity. The activity of purified enzyme is partially stimulated by the addition of sulfhydryl compounds. The activity is strongly inhibited by sulfhythryl group modifiers such as p-ehioromercuribenzoate and N-ethylmeleimide.

Introduction Understanding of the pathway for the biosynthesis of penicillins and cephaiosporins has been extended using cell-free systems from the cephaiosporin C-producing fungus, Cephalosporium acremonium (syn. Acremonium chy.~ogenum) [1-5]. it has been demonstrated that cellfree extracts of C. acremonium protoplasts catalyze conversion of isopeniciUin N (IPN) to penicillin N [6,7]. The enzyme catalyzing this reaction has been named isopen;cillin epimerase [IPNE] [6,9]. IPN is produced from 8-(L-aminoadipyl)-L-cysteinyl-D-valine (ACV) in a reaction catalyzed by isopenicillin N synthetase (IPNS) (8,9). Penicillin N is ring-expanded to deacetoxycephaIosporin C (DAOC), and DAOC is hydroxylateA to dcacetylcephalosporin C (DAC). In C. acremonium, these oxygen-dependent reactions are catalyzed by a

Al:hreviations: ACV, bi.~-O~-a-aminoadipy0-L-cysteinyI-D-valine; DAC, deacetylcephalospotinC; DAOC. deacetoxycephalosporinC, isoDAOC, isodeacetoxysephalospotinC; IPN. isopenicillin N, IPNE, isopentcillin N epimerase; IPNS, isopenicilfin N synthetase; SDS, sodium dodecylsulfate. Correspondence: C.-A. Yu, Department of Biochemistry, OAES, OklahomaState University,Stillwater.OK 74078, U.S.A.

bifunctional enzyme, penicillin N expandase/DAOC hydroxylase [10]. Thus, IPNS, IPNE and penicillin N expandase/DAOC hydroxylase function sequentially in cephalosporin C biosynthesis. In C acremonium, only a low level of the epimerase activity has been detected, demonstration of activity has only occurred in cell-free extracts freshly prepared by osmotic lysis of protoplasts [5,6]. This epimerase activity has been reported to be extremely unstable [4-6,8]. These features have discouraged further irvestigation of epimerase activity in C. acremonium. Streptomyces clavuligerus is a prokaryote which pro-

duces a variety of fl-lactam antibiotics including cephamycin C, a 7-methoxylated cephalosporin. Biosynthesis of cephamycin C in S. clavuligerus has been shown to involve IPNS, IPNE, and separable penicillin N expandase and DAOC hydroxylase cnzyraes [11,12]. A partial purification of isopenicillin N epimerase in S. clavuligerus has been reported [13], and the epimerase activity in the partially purified preparation was found to be unstable, similar to IPNE from C. acremonium. A microbiological procedure which makes use of the differential sensitivity of certain indicator organisms to isopenicillin N and penicillin N [6] has been employed to assay for the epimerase activity in both C. acremonium and S. clavuligerus. The complexity and slow-

0167-4838/89/$03.50 ~:~19E9Elsevier~k:ien~ PublishersB.V.(BiomedicalDivision)

"79 ness of this assay procedure has somewhat hindered purification of IPNE. Taking advantage of HPLC to ~eparate the derivatives of isopenicillin N and penicillin N formed with o-phthalaldehyde and N-acetyl-L-cysteine, we have developed a routine assay procedure for IPN epimerase activity. Aided by this simple assay method, we have devised a procedure for purifying the epimerase from S. clavuligen,s to homogeneity. Purified epimerase catalyzed not only the conversion of isopenicillin N to penicillin N, a previously reported reaction, but also catalyzed conversion of penicillin N to isopenicillin N so that, starting wath either penicillin, an approximately equimolar mixture of the two diastereomers was produced. The mixture is not racemic, since IPN and penicillin N are not optical isomers.

Experimental procedure

Materials lsopenicillin N, penicillin N, deacetoxycephalosporin C (DAOC), isodeacetoxycephalosporin C (isoDAOC), bis-8-(L-a-aminoadipyl)-L-cystein$1-D-valine (ACV), deaeetylcephalosporin C (DAC), and cephalosporin C were gifts from Dr. W.-K. Yek and G. Huffman of Eli Lilly and Co., Indianapolis, IN. Pyridoxal 5'-phosphate, o-phthalaldehyde, N-aeetyl-L-cysteine, h)droxylamine, sodium borohydride and fluoreseamine were obtained from Sigma Chemical Co., St Louis, MO. Dithiothreitol, the molecular weight standard proteins for SDS-polyacrylamide gel electrophoresis, DEAE Affi-gel blue were purchased from Bio-Rad Laboratories, Richmond, CA. DEAE-cellulose (DE-52) was obtained from Whatman Inc., Clifton, NJ. Sephadex G-200, Sephadex-G-100 superfine, Mono Q, and the molecular weight standard proteins for gel filtration were purchased from Pharmacia Inc., Piscataway, NJ. Calcium phosphate was prepared by the method of Jenner [14]. HPLC-grade methanol was obtained from J.T. Backer Chemical Co., Philipsburg, NJ. Other chemicals were obtained commerically at the highest available grade. Culture conditions An improved cephamycin C-producing strain of Streptomyces clavuligerus ATCC 27064 was grown in 500 ml Erlenmeyer flasks under conditions describzd by Whitney et. al. [15]. After cultivated for 48 h, cells were harvested by centrifugatio,, washed with 15 ml~ TrisHCI (pH 7.5), containing 1.0 M KCI and 20% [w/w) ethanol and then with 15 mM Tris-l-:Cl ~pH 7.5). Washed cclls were stored at - 8 0 °C until used. Enzyme assay and other methods The epimerase activity was determined 5y product formation, using isopenicillin N and pen:,cillin N as substrates. Formation of isopenicillin N and penicillin

N was estimated by tfigh-performance liquid chromatography (HPLC). The standard assay mixture for measuring the net interconversion of isopenicillin N and penicillin N, unless otherwise noted, contained 1.4 mM iscpenicillin N or 1.4 mM penicillin N, 0.2 mM dithiothreitol, 0.1 mM pyridoxal 5'-phosphate, 50 mM pyrophosphate-HCl (pH 8.3), and enzyme in a final volume of 0.5 ml. Tile reaction mixture was incubated at 37°C for 20 ~nin. The enzymatic reaction was terminated by placing the mixture into boiling water for 10 rain. The mixture was immediately cooled in the ice and then filtered through a 0.45 #m filter. Since it was difficult to separate directly isopenicillin N and penicillin N by HPLC, both penicillins were derivatized with o-phthalaldehyde according to the method of Usher et al. [16]. 20/xl of the filtrates of the reaction mixture containing isopenicillin N and penicillin N were mixed with 5 #1 of the derivatizing reagent. The reagent was prepared by mixing 4 mg o-phthalaldehyde dissolved in 300 #1 methanol, 250 #1 of 0.4 M sodium borate (pH 9.4), 390 ~1 of water and 60/~1 of 1 M N-acetyl-L-cysteine, and then adjusting to pH 5-6 with NaOH. After 2 rain reaction, 200 #1 of 50 mM sodium acetate (pH 5.0), was added to the mixture, and 50 #1 of aliquots were taken and analyzed in tile HPLC system. A Varian HPLC system including a Vista 5560 unit with Fluorichrom fluorescence detector and model 4270 integrator (Varian Associates, Sugar Land, TX), and model 231-401 auto-sampling injector (Gilson Medical Electronics, Middleton, Wl) were used. The derivatized isopenicillin N and penicillin N were separated on a Cyclobond l column (0.46 × 25 cm) (Astec Inc, Whippany, N J) with a mixture of 59% (v/v) 50 mM sodium phosphate, 1% (v/v) 1 M sodium acetate, 40% (v/v) methanol (pH 6.0), as a mobile phase. A flow rate of 1 ml/min was used, and total fluorescence (Ex 360 n) was determined. Quantitation was achieved by comparison to external standards. The protein was determined by the method of Lowry et al. [17] witi~ bovine serum albumin as a standard Pyridoxal 5'-phosphate was measured by phenylhydrazine method [18].

Purification of epimerase Al I purification steps were carried out at 0-4 ° C. The frozen cells (250 g) were thawed and homogenized in 1 liter of buffer: 10 mM pyrophosphate-HCi (pH 8.0), containing 20% glycerol and 0.1 mM dithiothreitol. The cell suspension was divided into 250 ml portions and subjected to sonication four times for 20 s each with 30 s intervals. The cell-free extract was obtained by centrifugation at 30000 × g for 30 min and then subjected to ammonium s,:lfate fractionation. The precipitates formed between 35 and 70% ammonium sulfate satucation were collected by centrifugation and redissolved in 50 ml of 10 mM pyrophosphate-HCi (pH 8.0) contain-

80 ing 0.1 mM dithiothreitol and 0.15 M NaCI, and then dialyzed overnight against the same buffer. The dialyzed crude epimerase was applied onto a DEo52 column (3.7 x 25 cm) equilibrated with 10 mM pyrophosphate-HCi (pH 8.0) coataiaing 0.1 mM dithiothreitol and 0.15 M NaCI. After washing the column with the same buffer, the epimerase was eluted with a linear gradient formed from 0.15 to 0.3 M NaCI in 10 mM pyrophosphate-HCI (pH 8.0) containing 0.1 mM dithiothreitol. Fractions with the enzyme activity were pooled and concentrated by ultrafiltration with an Amizon PM-10 membrane, and then dialyzed overnight against 10 mM potassium phosphate (pH 7.0) containing 0.1 mM dithiothreitol and 0.1 M NaCl. The dialyzed solution was loaded onto a DEAE Affi-Gel-blue column (1.6 × 15 cm) equilibrated with the dialyzing buffer. The enzyme was retained in the column and was eluted with a linear gradient formed from 0.1 to 0.3 M NaCl in 10 mM potassium phosphate (pH 7.0) containing 0.1 mM dithiothreitol. The fractions exhibiting enzyme activity were combined and concentrated by ultrafiltration, and loaded onto a Sephadex G-200 column (2.5 × 42 cm) equilibrated with 10 mM potassium phosphate (pH 6.0) containing 0.1 mM dithiothreitol and 10 #M pyridoxal 5'-phosphate. The fractions with epimerase activity were combined, concentrated by ultrafiltration, and loaded onto a calcium phosphate-cellulose column (1.6 × 10 cm) equilibrated with 10 mM potassium phosphate (pH 6.0) containing 0.1 mM dithiothreitol and 10 #M pyridoxal 5'-pbosphate. The column was washed with the same buffer and the enzyme was eluted with a linear gradient formed from 10 to 100 mM potassium phosphate (pH 6.0) containing 0.1 mM dithiothreitoi and 10 pM pyridoxal 5'-phosphate. The fractions containing enzyme were combined and concentrated by ultrafiltration, and then dialyzed overnight against 10 mM pyrophosphate-HCI (pH 8.0) containing 0.1 mM dithiothreitol and 10 btM pyridoxal 5'-phosphate. The enzyme solution was applied onto a FPLC MonoQ column (0.5 × 5 cm) equilibrated with 10 mM pyrophosphateHCI (pH 8.0) containing 0.1 mM dithiothreitol and 10 /~M pyridoxal 5'-phosphate. The enzyme eluted with a linear gradient formed from 0 to 0.4 M NaCI in 10 mM pyrophosphate-HC! (pH 8.0) contailfing 0.1 mM dithiothreilol and 10 pM pyridoxal 5'-phosphate.

Molecular weight determination The molecular weight of the native epimerase was estimated by gel filtration using Sephadex G-100 column (1.6 × 68 cm) equilibrated with 10 mM potassium phosphate (pH 7.0) containing 0.1 mM dithiothreitol, 10 ~tM pyridoxal 5'-phosphate and 0.15 M NaCI. Yeast, alcohol dehydrogenase (M, 80000), bovine serum albumin (67000), ovalbumin (43000), chymotrypsinogen A (25000) and ribonuclease A (13700) were used as molecular weight standards. The molecular weight of

detergent-denatured epimerase was determined by SDS (12.55~)-polyacrylamide gel electrophoresis according to the method of Weber and Osborn [19]. The protein standards used were phosphorylase B (94000), bovine serum albumin, ovalbumin, carbonic anhydrase (31000), soybean trypsin inhibitor (21500) and lysozyme (14400). The sample and standards were treated with 2~ SDS and 5% /]-mercaptoethanol at 100*C for 5 rain before they were subjected to electrophoresis.

Amino acid composition and sequence analysis Amino acids were analyzed by the methods of Heinrikson and Meredith [20] and Bidlingmeyer et at. [21] with a HPLC reversed-phase column after derivatization with phenylisothiocyanate (PITC) to phenyithiocarbamoyl amino acids (FrC-amino acids). The purified epimerase was extensively dialyzed against water, iyophilized in acid-washed tubes, and hydrolyzed with 6 M HCI at l l 0 * C for 24 h. The hydrolysate was reacted with PITC, and PTC-amino acids were quantitated by HPLC procedure with external amino acid standards, using Ultrasphere-ODS column (0.46 × 25 cm) (Beckman, San Ramon, CA). The PTC-amino acids were eluted with the gradient of the solvents (A) 50 mM ammonium acetate (pH 6.0), and (B) mixture of acetonitrile/methanol/0.22 M ammonium acetate (pH 6.0) (44:10 : 46). The amino acid sequence analysis was performed by automated Edman degradation, using model 470A gasph&~e protein sequencer with released amino acid phe~ ylthiohydantoin derivatives (PTH-amino acids) detected on-line by model 120A PTH-amino acid analyzer (Applied Biosystems, Foster City, CA) [22]. The purified epimcrase was extensively dialyzed against water, lyophilized, dissolved in 5% acetonitrile containing 0.1~ TFA, and then absorbed into the polybrene-coated glass microfiber filter. Results and Discussion

Purification of epimerase from S. clavuligerus The IPN epimerase was purified to electrophoretic homogeneity by sequential application of ammonium sulfate prec~ipitation and five serial chromatographies. About 1 mg of the purified epimerase was obtained from 250 g wet cells of Streptomyces elavuligerus. The overall yield was about 18~. Using isopenicillin N as a substrate, a 650-fold purification was obtained (Table I). The activity ratio between penicilfin N and isopenicillin N as substrates was about 2, and the ratio remained constant during the enzyme purification. These results suggested that one epimerase in S. clavuligerus catalyzed both reactions: conversion of IPN to penicillin N and conversion of penicillin N to IPN. The molecular weights of denatured and native epimerase were estimated as 47000 and 50000 by SDS-

81

TABLE I

Purification of isopenicillin N epimerasefrom. S. clm,uligergs The enzyme activity was assayed for isopenicillin N and penicillin N as substrates in 0.5 ml of the reaction mixture consisted of 50 mM pyrophosphate-HCI (pH 8.3), 0.1 mM pyridoxal 5'-phosphate, 0.2 mM dithiothreitol, 1.4 mM substrate, and the enzyme solution. ,After the mixture was incubated at 37°C for 20 rain, the amounts of product were measured by HPLC. Step

Protein (mg)

C¢lbfree extract 35--70~

+~lff'flSO 4 f r a c t i o n

DE-52 eluate DEAE-Affi-gel blue eluate Sephadex (3-200 eluate Calcium phosphate eluate Mono Q eluate

4430 2210 180 40.4 9.75 1.89 1.15

Substrate isopenicillinN

Activity ratio" penicillin N

activity (/~mol/min)

spec.act, (pmol/min per rag)

activity (~amol/min)

spec. act. (t~moi/min per rag)

23.9 23.6 15.1 7.97 6.42 5.07 4.42

0.0054 0.0107 0.0837 0.197 0.659 2.68 3.85

53.5 50.7 27.2 16.1 13.1 10.3 8.93

0.012 0.023 0.151 0.399 1.34 5.46 7.77

2.24 2.15 1.80 2.03 2.04 2.04 2.02

" This presents the ratio of the enzyme activity for penicillin N to that for isopen!eillin N.

polyacrylamide gel electrophoresis and Sephadex G-100 gel filtration, respectively (Fig. 1). Thus, the native epimerase appeared to be monomeric. The molecular weight of partially purified I P N epimerase in S. clavuiigerus estimated by Scphadex G-200 gel filtration m e t h o d had been reported to be 60000 [13].

Amino acid composition of N-terminal amino acid sequence of epimerase An experimentally determined amino acid composition for the epimerase is presented in Table II. The predominant residues of the enzyme were leucine, alanine and arginine. The following N-terminal amino acid sequence from the epimerase was obtained: AlaVal-Ala-Asp-Trp-Glu-Glu-Ala-Arg-Gly-Arg-Met-LeuLeu-Asp-Pro-Thr-Val-Val-Asn-Leu-Asn-Thr-

Catalytic properties of enimerase Since the enzyme catalyzed the conversion not only from IPN to peifcilfin N, but also from penicillin N to IPN, the kinetics of the interconversion was analyzed starting with either penicillin as substrate. The timecourse of the enzyme reaction was examined by following the amounts of both preduct formed and substrate remaining. The ratio between the product and the substrate was 1 when the reaction reached an equilibrium, regardless of whether I P N or penicillin N was used as substrate (Fig. 2). This result confirmed that the enzyme catalyzed 'racemization' of the L- and the V-a-anunoadipyl side-chains of IPN and penicillin N, respectively. The epimerase had a Vm=, of 3.93 p m o l / m i n per mg and a K m of 0.30 m M for IPN, whereas it had a Vma~ of 9.47 / t m o l / m i n per mg and a K m of 0.78 m M for penicillin N. When these valtles were used, the calculated K=q for the interconversion of IPN and penicillin N was 1.08, which agreed well with the value of 1.0 for a ~omplete

A O Tv

Q

a

_o

~O

2

1

0

I

0.2

I

I

0.4 0,6 ~elatlvt~ mobility

J

0.8

J

1.0

Fig. 1. Deten~fination of molecular weight of isopenicillin N epimerase by SDS-polyacrylamide gel electrophoresis. The purified epimerase (10 pg), treated with 2~o SDS and 5~ fl-mercaptoethanol, was appl;ed onto 12.5~ gel. The SDS-polyacrylamidc gel electrophoresis was performed according to Weber and Osborn [19]. The protein standards used were: (1) phosphorylase b (Mr 92500); (2) bovine serum albumin (66 200); (3) ovalbumin (45 000); (4) carbonic anhydrase (31000); (5) soybean trypsin inhibitor (21500); and (6) iysozyme (144130). Isopenicillin N epimerase was shown as E. i'he molecular weight was ,~stimated from a semilogarithmicplot of molecular weight against mobility.

82 TABLE l!

A

800

Amino aod composition of isopenicillin N epimerase. The enzMn¢ was hydrolyzed at I 1 0 ° C for 24 h in 6 N HCI as described under I~xpcrimental Procedure. After amino acids were derivatized with PITC, lYFC-amino acids were anaiyzed by HPLC. Cystcine was measured as cvsteic acid after the enzyme w ~ oxidized by perform/c acid. The integral numbers of amino acids based on the molecular weight of 47000 were pre~nted. Amino acid

No. of residue~ per M, 47000

Asx Thr ser

Olx

19.0 (t9) 25.'~ (25) t9.1 09) 36.0 (36)

Pro

34.3 (34)

Gly Ala Val Cys Met lie Leu Tyr Pbe Lys His Arg Trp

31.0 (31) 51.1 (51) 10.7 (!1) 3.9 (4) 3.1 (3) 12.6 (13) 49.7 (50) 2.9 (3,~ 16.8 (17) 2.8 (3) 13.1 (13) 42.0 (42) n.d."

"

n.d., not determined.

' racemization' of the L- and D-a-aminoadipyl side-chains of IPN and penicdlin N, (see also Scheme 1): is~penicillin N (L-side-chain) ** Penicillin N (D-side-chair.):

X.q - [ v.,., eL) K.(D)]/[ V,,~ (D) K,.(L)]

it should be noted that the 6-aminopenicillanic acid nucleus of IPN and penicillin N contains additional chiral centers. Thus IPN and penicillin N are diastereomers and not optical isomers; and even a 1:1 mixture of the p,micillias is not truly racemic. Since only one and not all of the optical centers is reversibly epimerized in the enzyme reaction, the name epimerase and not racemase is used. When the epimerase activity was assayed under various buffer systems, the optical pH of the epimerase activity was found at 7.8 :o 8.3 when 50 mM pyrophosphate-HCI was used a.~ a ouffer. The epimeras¢ was stable at 4 ° C for 24 h ~rom pH 5.3 to 10.0 when 50 mM sodium acetate, potessmm phosphate, Tris-HCI, pyrophosphate and carbottate/hicarbonate were used as buffers. When the epim~ase was incubated at the various temperatures for 10 rain in 50 mM pyrophosphateHCI (pH 8.0), no IGss of the enz)lrne activity was observed up to 4 5 ° C At higher temperatures the enzyme rapidly became ,~en~,tu:ed. Complete inactivation

600

A

400

200

¢J "0

0

|

I

I

t

I

t

8 0 0 ~ i CO0

_,oo[

0

:

20

40

60

80

:

100

120

Incubation time (mln)

Fig. 2. Interconversion of isopenicillin N and penicilfin N by isopenicillin N epimerase. The enzyme activity was assayed with isopenicillin N (a) and penicillin N (h) and substrates in 4 ml of the reaction mixture consisting of 50 mM pyrophosphate-HCl (pH 8.3), 0.t mM pyridoxal 5'-phosphate, 0.2 mM dithiothreitol, 1.4 mM substrate, and 3.4 /~g (e, o) or 10 ~g (a, z0 of the purified enzyme. After the mixture was incubated at 3 7 ° C for the indicated time, 0.5 ml portion of the mixture was withdrawn and the amounts of substrate remaining (e, a) and product formed (o, 4) were quantitated Ly HPLC).

was observed when epimerase was incubated at 6 0 ° C for 10 rain. The epimerase appeared to have very narrow substrate specificity. Among the intermediates of the biosynthetic pathway to cephamycin C that were tested, only isopenicillin N and penicillin N served as efficiently utilized substrates for the enzyme. The reversible nature of epimerase was first described by Bowers et al. [23] IPN (3.6 pmmol, rain. mg -I, 100%), penicillin N (204%), and DAOC, (<1%) showed activity as substrates. Other related compounds tested (isoDAOC, DAC, cephalosporin C, ACV, L-amino adipate and

-.- 3 s,.NH2 I t S CH t ~IC-(CH2)3-CONH'~. "'~ CH3 HOOC 04r"m"-~ CO0 H I'~openl¢.llnN S T E,Imerase H

IdL.I

.ooc.,.Z:;c. H

--

.....

•

l'CH3

O O ' - ' v - ' ~ COOH Penicillin N

Scheme I. Reactio;,of isopenicillinN epimerase.

83 0.6,

0.12

D-amino adipate) were not substrates for the epimerase at 1.4 mM concentration. Since the Km for IPN (0.30 mM) is lower than that for penicillin N (0.78 mM), penicillin N produced in vivo might be consumed in the next biosynthetic step, ring-expansion to DAOC, before significant amounts of penicillin N could be epimcnzed back to IPN. In other words, at low concenttation~ of IPN and peaicillin N in the presence of expandase or expandase/hydroxylase in vivo, the epimerase is well suited for converting IPN to penicillin N, i.e., epimerization in direction towards cephalosporin biosynthesis.

[18], and an average value of 1 mol of pyridoxal 5"phosphate per 47000 g of enzyme was obtained. Although the purified enzyme contained a stoiehiometric amount of pyridoxal 5'-phosphate, a 1.5-fold increase Jn activity was observed when more than 10/xM pyridoxal 5'-phosphate was present in the reaction mixture [7]. The reason for activity enhancement is currently unknown. One possible explanation is the dissociation of the enzyme-cofactor complex at the low concentration used for enzyme assays. If this occurs, addition of excess cofactor would be expected to shift the dissociation equilibrium in favor of active holoenzyme. This also explains why the presence of 10 #M pyridoxal 5'-phosphate was required during the enzyme purification [13]. s ! The pyridoxal 5 -phosnhate in the holoenzyme can be removed by the incubation with iO mM potassium phosphate (pH 7.0) containing 10 mM hydroxylamine at 0 ° C for 60 rain followed by gel filtration. The apoenzyme produced by hydroxylamine treatment had no absorption maximum at 420 nm and was inactive in the absence of exogenous pyridoxal 5'-phosphate (Fig. 3, curve 2). The epimerase activity was restored by the addition of pyridoxal 5'-phosphate to the apoenzyme. The apparent Michelis constant for pyridoxal 5'-phosphate was determined to be 2.4 gM by Lineweaver-Burk plot (data not shown). Since the apoenzyme is not very stable, the age of the apoenzyme and the incubation times after addition of cofactor at various concentrations were kept constant in order to avoid possible comphcations in the km determination du~ to the denaturation of the apoenzyme. Reduction of the epimerase with sodium borohydride affected both the absorption spectrum (Fig. 3, curve 3) and the activity. The reduced enzyme was catalytically inactive and the addition of pyridoxal 5'-phosphate did not restore the enzyme activity. These results suggested that the aldimi.|e linkage was reduced to an aldamine bond when borohyddde was used.

Characterization of cofactor

Effect of fluorescamine on epimerase

Fig. 3 shows the absorption spectra of the epimerase. The purified epimerase had absorption maxima at 280 and 420 nm (Fig. 3, curve 1). The absorption ratio of A2so/A42o was calculated to be 7.57. Absorption coefficients for A2s0 and A420 were 8.96 and 1.18, respectively, in 10 mM potassium phosphate (pH 7.0). The appearance of the absorption peak at 420 nm suggested that the enzyme contained pyridoxal 5'-phosphate as cofactor with its formyl group in an aldimine linkage with an amino group of the protein. The presence of pyridoxal 5'-phosphate in the epimerase was similar to the presence of that cofactor in amino acid racemases involved in the cell ~.~," ~ymhesis [24,25]. The amount of pyridoxal 5'-phosphate bound to the purified epimerase was measured by the phenylhyd, azine method

Fluorescamine reacts with primary amino groups such as the a-amino group of the N-terminal amino acid of a p ~ , ~ e t , ~ - ' _,r-~ e-amino group of iysines in a protein. The reaction products have a fluorescence with the maximum excitation and emission wavelength of 390 and 475 nm, respectively [261. When the apoenzyme, prepared by the treatment with hydroxylamine, was titrated with fluorescamine, about 60% of the amino groups in the epimerase reacted readily with fluorescamine, indicating that these ~n,inn groups were probably located on the surface of the molecule. When 3.5 mol of amino groups per mol of enzyme were reacted ;,i;h fiu~.rcscarr, inc, the enzyme lost only 20% of its activity, The enzyme, however, was dramatically inactivated b2~ further modification with fluorescamine

L 1,2, 0.5 ; 0.4 ~
"

\3

0.08

0.3 ,

0.1 0 250

300

1

I

I

350

400

450

0 SO0

Wavelength (nm)

Fig. 3. Absorption spectra of isopenicillin N epimerase. (l) Holoenzyme. (2) Apomzyme (epimerase after treatment with l0 mM hydroxylamine). (3) Epimerase reduced by 1 mM sodium borohydride. Absorption spectra were taken in 10 mM potassium phosphate (pH 7.0), at the enzymeconcentration of 0.57 mg/ml by using a Aminco DW-2000 spectrophotometer.Hydroxylamineor sodium borohydride was removedby gel filtration prior to spectrophotometry.

84 SO

1oo

80

~

40

i

8

~

4o

2o x

2o

lo

i

p___AL

2

4

i iS

8

o 10

Fluomll~mlne/l[plmorlloO (molYmol)

Fig. 4. Titration of isopeniciilin N epimerase by fluorescamine. The epimerase (0.5 mg/ml) was incubated in 10 mM potassium phosphate (pH 7.0) containing l0 mM hydroxylamine at 0 o C for 60 nun. After removal of hydroxylamine by gel filtration, the treated enzyme ,0.16 mg/ml) in 10 mM potassium phosphate (pH 7.0) was titrated with the indicated amounts of fluorescamine (50 mM) dissolved in dry acetone at 25eC. The sample was excited in 390 tun and the emission f l u ~ was followed at 475 nm (o). Simultaneously, aliquots of enzyme solution were withdrawn and the activity was assayed for epimera~e activity with isopenicillin N as substrate (@).

(Fig. 4). These results suggested that the lysine residue, capable of forming an aldimine linkage with pyridoxal 5'-phosphate and involved in the catalytic activity, was probably located inside the molecule. This deduction is supported bv the observation that the holoenzyme as prepared, or the apoenzyme in the presence of pyridoxa! 5'-phosphate, was not inactivated by the treatment of fluorescamine. TABLE I11

Effects of ~,erious compounds on isopenicillin N ~pimerase activity The enzyme (1.5 Pg) was incubated with the indicated amount of the compound at 370C for 5 rain in 0.25 ml of 100 mM pyrophosphateHCI (pH 8.3). The enzyme activity was subsequently measured by the addition of 0.25 ml of the mixture consisted o" 0.2 mM pyridoxal 5'-phospha*,e, 0.4 rruM ditiotheritol and 2,8 mM isopenicillin N. The activity w~hout compound was 3.8 p m o l / m m per nag. Compound

Conc. (raM)

Inhibition (%)

p-CMoromercuribenboate

0.1

N-Ethylmaleimide

0.1

70 88 26 45 28 80 95 23 23 25 80 33

1.0 1.0

lodoactmnide Hydroxylamine

1.0 0.1 1.0

D-Cyclosenne lsoniazid iproniazid FAD Phenylglyoxal

1.0 1.0

1.0 0.5 1.0

Effec, o[ activators and inhibitors on epimerase activity Table III lists various compounds which affected the epimerase acti'Aty. The enzyme activity was partially stimulated (1.3-1.8-fo!d) by 0.5 mM of sulfhydryl reducing reagents such as ~-mercaptoethanol, dithiothreitol and N-acetyl-t-cysteine in the reaction mixture. Reagents which react with sulfhydryi group such as p-chloromercuribenzoate, N-ethylmaleimide and iodoacetamide inactivated the enzyme. The sulfilydryl group appeared to be involved in the epimerase reaction. There are four sulfhydryl groups per molecule of epimerase. One of them has been shown to be near the pyridoxal 5'-phosphate binding site. Although a more detailed and systematic investigation on these sulfhydryl groups is needed before the role of each sulthydryl group can be assigned, it is tempting to speculate that the sulfhydryl group near the cofactor binding site may play an important role in catalysis. This of coarse does not exclude the possibility that modification of sulfhydryl group remote from the active site could also indirectly cause the enzyme to assume an inactive conformation and thus diminish the activity. Phenylglyoxal and pyridoxal 5'-phospliate antagonists such as hydroxylamine, D-cycloserine, iproniazid and isoniazid also inhibited the enzyme activity. Although the enzyme activity was strongly inhibited by 0.5 mM FAD, the inhibition mechanism is not dear. Chelating reagents such as EDTA or EGTA did not affect the enzyme activity, suggesting that no metal ion is involved in catalysis by epimerase.

Acknowledgments We would like to thank Drs. S.W. Queener and W.-K. Yeh for critical review of this manuscript. We also thank A. Kreuzman and J.E. Dotzlaf (Eli Lilly and Company) for preparation of Streptomyces clavuligerus cells. This work was supported in part by a research grant from Eli Lilly and Company, and we thank Dr. David W. Dennen for arranging this support. References l Fawcett, P.A., Usher, J.J., Huddleston, J.A., Bleaney, R.C., Nisbett, J.J. and Abraham, E.P. (1976) Biochem. J. 157, 651-660. 2 Bost, P.E. and Demain, A.L. (1977) Biochem. J. 162, 681-687. 3 Yoshida, M., Konomi, T., Kohsaka, H., Baldwin, J.E., Herchen, S., Singh, P.D., Hunt, N.A. and Demain, A.L. (1978) Proc. Natl. Acad. Sci. USA 75, 6253-6357. 4 Konomi, T., Herchen, S., Baldwin, J.E., Yoshida, M., Hunt, N.A. and Demain, A.L. (1979) Biochem. J. 184, 427-430. 5 Baldwin, J.E., Keeping, J.W., Singh, P.D. and Vallejo, C.A. (1981) Biochem. J. 194, 649-651. 6 Jayatilake, G S., Huddleston, J.A. and Abraham, E.P. (1981) Biochem. J. 194, 645-647. 7 Lvbbe, C, Wolfe, S. and Demain, A.L (1986) AppL Microbiol. Biotechnol. 23, 367-368.

85 8 Sawada, Y., Baldwin, J.E., Singh, P.D., Solomon, N.A. and Demain, A.L. (1980) Antimicrob. Agents Chemother. 18, 465-470. 9 Abraham, E.P., Huddleston, J.A., Jayatilake, G.S., O'Sullivan, J. and White, R.L (1981) in Recent Advances in the Chemistry of ~-Lactam Antibiotics (Gregory, G.E., ed.), pp. 125-134, The Royal Society of Cbe:~tist:'y, Burlington House, London. 10 DotzlaL J.E. and Yeh, W.-K. (1987) J. Bacteriol. 169, 1611-1618. 11 Jensen, S.E., Westlake, D.W.S. and Wolfe, S. (1982) J. Antibiot. 35, 483-490. 12 Jensen, S.E., Westlake, D.W.S., Bowers, R.J. and Wolfe, S. (1982) J. Antibiot. 35, 1351-1360. 13 Jensen, S.E., Westlake, D.W.S. and Wolfe. S. (1983) Can. J. Microbiol. 29, 1526-1531. 14 Jenner, E.L. (June 5, 1773) U.S. Patent 3737516. 15 Whitney, J.G., Brannon, D.R., Mabe, J.A. and Wicker, K..;. (1972) Antimicrob. Agents Chemother. 1, 247-251. 16 Usher, J., Lewis, M. aild H~lghes, D.W. (1985) Anal~ Biochem 149, 105-110.

17 Lowry. O.H., Rosebrough, N.J., F,~trr, A.L. and Randall, R.J. (1951) J. Biol. Chem. 193, 265-275. 18 Wada, H. and Snell, E.E. (1961) J. Biol. Chem. 236, 2089-2095. 19 Weber, K. and Osborn, M. (1969) J. Biol. Chem. 244, 4406-4412. 20 Heinrikson, R.L. and Meredith, S.C. (1984) Anal. Bincbem. 136, 65-74. 21 Bidlingmeyer, B.A., Cohen, S.A. and Tarvin, T.L. (1984) J. Chromatogr. 336, 93-104. 22 Applied Biosystems Protein Sequencer User Bulletin No. 12 (1985). 23 Bowers, R.J., Jensen, S.E., Lyubechansky, L, Wertlake, D.W.S. and Wolfe, S. (1984) Biochem. Biophys. Res. Co'mr~un. 120, 607-613. 24 Adams, E. (1976) Adv. EnzymoL Relat. Areas Mol. Biol. 44, 69-138. 25 Esaki, N. and Walsh, C.T. (1986) Biochemistry 25, 3261-3267. 26 Weigle, M., Debernarde, S.L, Tensi, J.P. and Leingruber, W. (1972) J, Am. Chem. SOC.94. 5927-5928.