H penicillinase serine-44 acylation by diazotized 6-aminopenicillanic acid

292

Biochimica et Biophysica Acta, 745 (1983)292-300

Elsevier BBA31633

B A C I L L U S C E R E U S 5 6 9 / H PENICILLINASE SERINE-44 ACYLATION BY DIAZOTIZED 6-AMINOPENICILLANIC ACID

THOMASG. HECKLERand RICHARDA. DAY Department of Chemistry, Unicersity of Cincinnati, Cincinnati, OH 45221 (U.S.A.)

(ReceivedMarch 28th, 1983)

Key words: Penicillinase; Affinity label; Peptide mappm&" Active site," Serine residue; (B. cereus)

Penieillinase from Bac/~hts cereus 5 6 9 / H was purified to homogeneity. Its active site was probed by use of an affinity label generated in situ by the diazotization of 6-aminopenicillanic acid, a catalytically poor substrate for this enzyme. The loss of activity arising during the inactivation is dependent upon pH and the penicillin:sodium nitrite ratio used. Optimal inactivation was obtained at pH 4.7 and reactivation could be prevented if subsequent purification and manipulations were performed at low pH. Inactivation by diazotized 6-aminopenicillanic acid was characterized further by tryptic and chymotryptic digestion of the inactivated enzyme and peptide mapping of the resulting digests. Amino acid analysis of the chymotryptic labeled peptide yielded a composition which corresponds to residues 41-46 (Ala-Phe-Ala-Ser-Thr-Tyr) in the published partial sequence of the enzyme (Thatcher, D. (1975) Biochem. J. 147, 313-326). Further digestion of this chymotryptic peptide with carboxypeptidese A reveals that serine-44 is modified in this affinity labeling procedure. Mass spectral analysis of the modified serine residue and alkali-released label, and comparison with spectra of model compounds indicates that the inactivation occurs with rearrangement of the ~-lactamthlazolidine structure to a dthydrothiazine.

Introduction The development of site-specific inhibit0rs against peniciUinases began with the discovery of the inhibitory action of cephalosporin C [1] towards Bacillus cereus penillinase. This substrate analog and, later, others, e.g. methiciUin [2] and nafcillin [3], while generally restricted in bactericidal action, are good reversible inhibitors. They are hydrolyzed very slowly, if at all, by penicillinases, and hence are somewhat effective clinically when used in combination with other penicillin antibiotics. However, there is still a great need for the further development of more potent penicillinase inhibitors to expand the l=actericidal activity of p-lactam antibiotics. Towards this goal some progress has been made in recent years in understanding the mechanism of action of penicillinases 0167-4838/83/$03.00 © 1983ElsevierSciencePublishersB.V.

by the use of modified substrates and substrate analogs to probe the active site of these enzymes. Patil and Day [4] accomplished irreversible inactivation of B. c e r e u s peni¢illinase by the diazotization of 6-aminopeniciUanic acid and ampicillin in the presence of the enzyme. In that work, and in later work by Durkin et at. [5], inactivation was suggested to be due to the alkylation of a carboxyl function in the active site via an ester linkage on the basis of alkaline pH reversal of inactivation and an appearance of a new circular dichroic band that could be reversed with hydroxylamine treatment. Similarly, the structurally related 6-fl-bromopenicillanic acid, prepared from 6-aminopenicillanic acid, has been found effective as an active site-directed inhibitor which inactivates B. cereus penicillinase [6,7], as well as the penicillinases of

293

Bacillus licheniformis, Staphylococcus aureus and Escherichia coli [7]. Inactivation by this inhibitor was also found to be sensitive to alkali treatment, indicating a possible similarity in inactivation mechanisms. In this report we describe work directed towards the characterization of diazotized 6-aminopenicillanic acid inactivation of B. cereus penicillinase. Materials and Methods

Culture medium for growth of vegetative cells. Growth of B. cereus 569/H was performed in a casein hydrolysate medium by the method of Kogut et al. [8]. lodometric assay. An iodometric assay was used to quantitate penicillinase activity [9-11 ]. Sporulation 'S' broth. Sporulation of B. cereus was achieved in a 'S' broth medium. The medium was prepared by dissolving 10.0 g bacteriological peptone (Difco), 3.0 g bovine heart extract (Difco) and 2.0 g NaCI in 1 1 of distilled deionized water and the pH adjusted to 7.0. Spore seed culture for fermentor inoculation. Aliquots (0.1 ml) of high penicillinase titer casein cultures (9000 units/ml) were transferred to three 2-1 Erlenmeyer flasks containing 200 ml of 'S' broth each. The flasks were agitated on a wrist action shaker at 42°C for 6-8 h. These turbid spore cultures were then used to inoculate 30 1 of casein media. Large-scale fermentation of B. cereus 569/H. 30-1 batch cultures were grown in a 40 liter carboy system designed by A. Kornel [12]. When the culture was ripe, as determined by periodic assay, the fermentor was connected to a Cepa-Schnetl flow-thru centrifuge (Carl Padburg Co., F.R.G.) t o remove the bacterial cells. The crude penicillinase supernate was harvested at a rate of 5-10 l/h. Concentration of penicillinase by celite adsorption. The crude penicillinase supernate was adjusted to pH 4.5 by the addition of cone. HC1. Celite 545 (John.~Mansville, N J), at 1 g / m g expected penicillinase, was added and the suspension mechanically stirred overnight at 4°C. The penicillinase adsorbs to the celite and the penicillinase/celite f'dtered off on a large (38 cm diameter) Buchner f , nnel. The celite was then washed

with 3-4 1 of acetone at - 2 0 ° C to remove lipid material, followed by two 2-1 rinses of 0.02 M potassium citrate, pH 7.0. PeniciUinase was eluted off by slurrying with 1-1 aliquots of 50~ satd. (NH4)2SO4, containing 0.1 M phosphate buffer at pH 8.5, and filtering. The l-I eluate aliquots were assayed for peniciUinase activity and the protein material was precipitated by addition of solid (NH4)2SO 4 to achieve 100~ saturation. This was stored at 4°C. After sitting a few days, the supernate was aspirated off and the remaining suspension centrifuged at 12000 rpm (20000 x g) for 30 rain to collect the protein. Sephadex G-IO0 chromatography. The crude precipitated penicillinase was dissolved to an approximate concentration of 5 mg penicillinase/ml in a 0.05 M sodium phosphate buffer, pH 7.0, containing 0.2 M NaCI and 0.02~ (w/v) NaN 3. The solution was then chromatographed on a Sephadex G-100 (Pharmacia) column (10 x 120 cm) equilibrated with the same buffer. The flow rate through the column was maintained at 1.0 ml/min, and the effluent was monitored at 280 nm. The penicillinase activity was determined by removing aliquots from each 22-ml fraction and assaying. Penicillinase-containing fractions were pooled, exhaustively dialyzed against distilled deionized water and lyophili7ed. Gel electrophoresis. Polyacrylamide gel electrophoresis in the presence of SDS was carried out by the method of Weber and Osborn (13] on 10~ gels. Amino acid analysis. Hydrolyses were carried out in evacuated, sealed 8-ram internal diameter pyrex tubes containing 6 N HCI and maintained at 110°C for 24 h. The acid hydrolysates were evaporated to dryness, dissolved in 0.01 N HCI and analyzed on a Phoenix Biolyzer 9000 equipped with Durrum DC-1A resin. Large-scale inactivation. Batch inactivation was carried out as follows: 20 mg quantities of purified, lyophilized penicillinase was dissolved in 20 ml of 0.3 M sodium acetate, pH 4.7, and 400 mg of NaNO2 added. After dissolution of the NaNO2, 300 mg of 6-aminopenicillanic acid was added as a solid and the solution gently stirred for 15 min at 25°C. The penicillinase activity was assayed before and after inactivation by the iodometric assay. Sephadex G-25 desalting. Inactivated penicil-

294 linase was purified from other dia:zotization products and buffer salts by passage through a Sephadex G-25 column (2.8 × 25 cm) previously equilibrated with 10% (v/v) acetic acid. The outlet of this column was connected to an Isco flow-through ultraviolet monitor and the void volume protein fraction absorbing at 280 nm was collected and lyopbilized. Peptide mapping. Proteolytic digests of inactivated peniciUinase were subjected to two-dimensional paper chromatography/electrophoresis for isolation of peptides. The digests, dissolved in 50% acetic acid, were appfied to Whatman 3 MM paper (46 × 57 cm) in amounts of 6 mg dry weight per map. Descending chromatography was carried out in the first dimension with the organic phase of freshly prepared 1-butanol/acetic acid/water (4:1:5) for 14-16 h. The chromatogram was then air-dried for several hours, moistened with pyridine/acetic acid/water (10:1 : 184, v/v) and electrophoresis carried out in the second dimension at 54 V/cm for 75 min. The paper was immersed in an inert hydrocarbon coolant. Starch /azide-I 2 detection of penicillin peptide. Chymotryptic and tryptic peptide maps of inactivated penicillinase were treated with starch/azide-I 2 to visualize penicillin-bound peptide(s) as described by Helberg [14]. The detection of a penicillin peptide is noted by the appearance of a white spot in a blue field. Detection with dimethylaminobenzaldehyde. General detection of all peptide spots on a peptide map was performed with p-dimethylaminobenzaldehyde [15]. Maps were dipped in a trough, of freshly prepared 0.1% (w/v) dimethylaminobenzaldehyde in ethyl ether containing 1% (v/v) acetic acid. The papers were air-dried and heated in a 100°C oven for 10 rain to facilitate Schiff base formation. The peptides show as pale yellow spots and can be best visualized by illumination with a long ultraviolet mineral lamp, having the paper submerged in a tray of liquid N 2. Elutlon peptides. Peptide spots of interest were cut out, rinsed in ethyl ether to remove excess dimethylaminobenzaldehyde, and eluted with approx. 2 ml of 10% aqueous acetic acid. Amino acid mapping. The lyophilized carboxypeptidase digest of the chymotryptic penicillin

peptide, dissolved in 50% acetic acid, was applied in a 1 × 2 cm spot in the top corner of Whatman No. 1 paper and descending Chromatography carried out with 1-butanol/acetic acid/water (4:1:5). The paper was air-dried and electrophoresis carried out in the second dimension in acetic acid/formic acid/water (30:3: 142) at 60 V/cm for 110 min. The paper was then air-dried and developed with p-dimethylaminobenzaldehyde [15]. A standard mix containing alanine, phenylalanine, serine, threonine and tyros.ine mapped identically under these same conditions. Elution of the dimethylaminobenzylidene amino acids. The same method as described for elution of peptides was used except that methanol was used as the eluant due to the decreased polarity of dimethylaminobenzylidene-derivatized amino acids. Preparation of serine-labei moiety for mass spectrometry. The serine-label moiety, which has been separated by two-dimensional mapping (spot X, Fig. 5) and derivatized to its p-dimethylaminobenzylidene derivative with p-dimethylaminobenzaldehyde, was eluted with 2 ml of anhydrous methanol and esterified by refluxing for 0.5 h after the addition of 2-3 drops of thionyl chloride. The solution was then evaporated to dryness and stored at - 1 0 ° C . Prior to mass spectral analysis, the sample was subjected to high-pressure liquid chromatography on an analytical scale silica gel column (Whatman Partisil PXS 10/25 PAC), eluting with chloroform. The sample peak, which was detected by fluorescence monitoring, was concentrated to a small volume and transferred to a solid inlet sample holder for mass spectrometry. The released affinity label moiety. The free 'label' itself was isolated by treatment of 16 mg of purified, inactivated penicillinase with 1 N NaOH for 3 h at 21°C and subsequent chromatography on a calibrated 9 ml Sephadex G-25 column equilibrated with 0.1 N NaOH. Fractions corresponding to the label were neutralized with 1 N HCI and the solvent was removed. The label was then picked up by extraction with anhydrous methanol and esterified by refluxing after adding a few drops of thionyl chloride.

295

Synthesis of model compounds I. 3-S-carboxy-2,3-dihydro-2,2-dimethyl-6-methoxy-carbonyl-l, 4-thiazine (I). 6-Aminopenicillanic acid (500 rag, 2.3 retool) was dissolved in 20 ml of 1 N HCI cooled to 0°C. Sodium nitrite (250 mg, 3.6 mmol) was added as a solid and the solution stirred for 0.5 h, after which the solution was extracted with 5 ml of chloroform. The chloroform extract was then washed three times with 10 ml of water; dried over MgSO4 and evaporated to a yellow oil. Thin-layer chromatography on silica gel with chloroform revealed a single major component (6-chloropenicillanic acid). The oil was dissolved in 30 ml of methanol, and 4.6 ml of 1 N sodium methoxide (two equivalents) added. The reaction mixture was stirred at 25°C overnight and the precipitate that formed collected by centrifugation and washed with methanol. This salt was then converted to the free acid (I) by acidification with 2 N HCI. (melting point, 174°C; literature 177°C [16]).

II. Esterification of 3-S-carboxy-2,3-dihydro2,2-dimethyl-6-methoxycarbonly-l, 4-thiazine (I) to yield the diester (II). The acid (I) (100 mg, 0.4 mmol) was refluxed in methanol with a few drops of thionyl chloride for 1 h. The reaction mixture was then vacuum evaporated to yield the diester (II).

III. Treatment of free acid (I) with acid to yield methyb 2,3-dihydro-3-hydroxy-2,2-dimethyl-l, 4thiazine-6-carboxylate (III). [17]. 100 mg of the free acid (I) was stirred with 20% acetic acid for 3 h and vacuum-evaporated to yield (III). Mass spectrometry. Mass spectra were obtained on a Hitachi-Perkin-Elmer RMU-7 mass spectrometer. Samples were introduced through the solid inlet and the mass spectra recorded while slowly increasing the temperature of the sample probe. The ionization energy was varied from 30 to 70 eV. R ~

Penicillinase isolated from B. cereus 569/H by the procedure described, yielded pure, homogeneous enzyme by three criteria: (1) activity profile is symmetrical with A2s0 (Fig. 1), (2) migration as single band on SDS-polyacrylamide gel electrophoresis, and (3) amino acid composition corn-

O8

2OO

tZ5 ~ ,<

~oZ

a4

:).7'5 >

60 70 80 Froction$ (22 ml )

90

100

Fig. 1. Purification of penicillinase by gel filtration on Sephadex G-100. • • , penicillinase activity; - - , absorbance at 280 nm.

pares well with reported values [18]. About 220 mg of pure penicillinase was obtained per fermentation batch for an overall yield of 40% (Table I). The specific activity of the final purified penicillinase was 2.4 x 105 units/mg. Prior to the preparation of amounts of inactivated penicillinase suitable for structural analysis, the reaction conditions were characterized in order to optimize the extent of inactivation that could be obtained by the diazotization of 6aminopenicillanic acid in the presence of the enzyme. Maximum inactivation was achieved at pH 4.7, a result essentially identical to that previously reported by Patil and Day [4]. The 6-aminopenicil-

TABLE 1 PURIFICATION OF PENICILLINASE FROM B. CEREUS 569/H Purification stage

1. Culture supernatant 2. Eluate from celite in 50% satd. (NH4)2SO 4 3. Precipitate from 100% satd. (NH4) 2SO 4 dissolved in 0.05 M phosphate/0:2 M NaCI, pH 6.8, buffer 4. Sephadex G-100 eluate, dialyzed and lyophylized

Volume (ml)

Total activity

Total yield

(units)

(%)

28 500

1.7.108

100

4100

1.1.10 s

65

80

8.2. l0 T

48

-

6.8. l0 T

40

296 TABLE II

~-

DEPENDENCE OF INACTIVATION UPON 6-AMINOPENICILLANIC ACID AND NaNO 2 CONCENTRATION Penicillinase solutions (1.0 rag/m1) in 0.3 M sodium acetate buffer, pH 4.7, were incubated with the indicated concentrations of 6-aminopenicillanic acid and NaNO 2 for 15 min. Penicillinase activity was measured before and after by the iodometric assay and the percent inactivationcalculated.

© oO

6-Aminopenicillanic acid (mg/ml)

% inactivation

30 30 20 20 20 15 10 0

0 10 10 20 15 20 30 30

0 45 56 74 92 76 36 0

1OO

1

t

I

1

I

I

I

I

-

4-

Oo (]

oQOo C:D cs o

lanic acid/NaNO 2 concentration dependence upon i n a c t i v a t i o n ( T a b l e II) s h o w e d t h a t g r e a t e r t h a n 90% i n a c t i v a t i o n c o u l d b e a c h i e v e d u s i n g the fol-

-

o o

9 O O0

N a N O 2 (mg/ml)

Elect~sis

, oo

Fig. 3. Peptide map of chymotryptic digest of inactivated penicillinase. Inactivated penicillinase (16 mg) was suspended in 1.6 ml of 2% ammonium bicarbonate buffer, pH 6.8. Chymotrypsin (I.5 rag) dissolved in 0.1 ml of 0.001 N HCI was added and the digestion allowed to proceed for 2.5 h at 25°C with mild stirring. Any precipitate that remained was removed by centrifigation and the digest lyophilized. Mapping was as described in Materials and Methods. The Helberg-positive peptide containing the modified serine is shaded.

I

l o w i n g c o n d i t i o n s : p e n i c i l l i n a s e (1 m g / m l ) , N a N O 2 (20 m g / m l ) a n d 6 - a m i n o p e n i c i l l a n i c a c i d (15 m g / m l ) at p H 4.7, T = 2 1 - 2 5 ° C . Inactivation of 20-mg quantities of penicil-

go g.o 80 70

PHe~o

8~)

Electro~sis

7.0

30

<3

-

O@

10 2

4

6

8

10

Hours Fig. 2. T i m e - d e p e n d e n t

reactivation of inactivated penicillinase

at various p H values. Penicillinase(I.0 mg/ml) in 0.05 M sodium acetate buffer, p H 4.7, was 92% inactivated by treatment with N a N O 2 (20 mg/ml) and 6-aminopenicillamc acid (15 mg/ml). Aliquots (0.5 ml) were removed and admixed with 0.5 nd of either0.3 M sodium acetate(pH 4.5,5.5 or 6.5) buffer or 0.3 M sodium phosphate (pH 7.0,7.5, 8.0,8.5 or 9.0) buffer. Aliquots of these admixed solutions were then assayed iodometrically for recovery of penicillinase activity at various times. Percent reactivation was calculated by comparison with an untreated penicillinase control.

OO°o ? 0

D

0

Fig. 4. Peptide map of tryptic digest of reactivated penicillinase. Inactivated penicillinase (approx. 16 nag) was suspended in 1.6 ml of 2% NH4HCO 3, pH 6.8. Trypsin (2.0 rag), dissolved in 0.I ml of 0.001 N NHCI was added. The digestion was allowed to proceed for 4 h at 25 ° with gentle stirrin8. The remaining precipitate was removed and the digest was lyophilized.

297 TABLE 1II

Electrophore~s-

AMINO ACID ANALYSIS OF LABELED PEPTIDES RELEASED BY CHYMOTRYPTIC AND TRYPTIC DIGESTION OF INACTIVATED PENICILLINASE

Asp Thr Ser

Glu Gly Ala Val

Met lie Leu

Tyr Phe Lys His Arg

Chymotryptic

Tryptic

0.00 0.98 0.86 0.00 0.00 2.10 0.00 0.00 0.01 0.00 0.96 0.94 0.00 0.00 0.00

0.83 I. 14 0.93 2.56 0.97 3.74 1.00 0.00 0.00 2.68 0.89 2.08 1.72 0.00 0.65

linase, using the reaction conditions described above, consistently gave greater than 90~ inactivation. De.salting by gel filtration through Sephadex G-25, eluting with 10~ acetic acid, yielded upon lyophilization a very light fluffy protein that could be readily digested with proteolytic enzymes without further denaturation. Proteolytic digestions of the inactivated penicillinase were carried out separately with trypsin, chymotrypsin and pepsin. Mapping and comparison of these digests showed that chymotryptic digestion was the most efficient of the three, giving a faster, more complete digest as well as better separation of its corresponding penicillin-bound peptide upon two-dimensional mapping and staining with starch/azide-I 2. Reproduction of the peptide maps obtained from tryptic and

(~)

COOH

~-

Origin •-. S e r ',--,'

~Thr

O

~

Tyr

Phe

Fig. 5. Map of the carboxypeptidase A digest of the modified, Helberg-positive chymotryptide. The chymotryptic penicillin peptide, eluted and lyophilized, was dissolved in 0.25 ml of 2~ ammonium bicarbonate buffer, pH 6.8, and digested with carboxypeptidase A (25 units) dissolved in 0.1 N NaOH (10 1). After 4 h at 25°C the digest was centrifuged, lyophiliTed and mapped as described in Materials and Methods. The modified serine is indicated by 'X'.

chymotryptic digestions showing the locations of each respective penicillin derivatized peptide are given in Figs. 3 and 4 (see Ref. 19). The penicillin-peptides from the chymotryptic and tryptic maps were eluted and the amino acid compositions determined from their 6 N HCI hydrolysates. Comparison of the results given in Table III with the published amino acid sequence of B. cereus penicillinase [18] identified these two respective peptides as: chymotryptic: -41Ala-PheAla-Ser-Thr-Ty~46-; tryptic: -36pro-Asp-Gln-ArgPhe-Ala-Phe-Ala-Ser-Thr-Tyr-Lys-Ala-Leu-AlaAla-Gly-Val-Leu-Leu-Gln-Gln-LysS8-; the tryptic peptide overlapping in sequence over the chymotryptic peptide.

COOH

Scheme I. Direct alkylation mechanlsm. Possible mechanisms of inactivation. ~ ) represents the penicilfinase moiety attached through serine-44.

298 So

COOH

+

O

OH

COOH

®_CH /

/

(~)--CH 2 O

O s

o

ca2

a

H Scheme II. Dihydrothiazine mechanism. Possible mechanism of inactivation. (~) represents the penicillinase moiety attached through serine-44.

Further degradation of the chymotryptic peptide with carboxypeptidase A and two-dimensional mapping of this digest is given in Fig. 5. Comparison of this map with identically mapped amino acid standards, and amino acid analysis, shows that serine is missing and instead a single new spot appears, spot X in Fig. 5, which yields serine upon amino acid analysis of its 6 N HCI hydrolysate. As indicated by the reversal of inactivation at higher pH (Fig. 2), it was doubtful that covalent attachment at this serine residue was due to a direct alkylation mechanism (Scheme I), since labeling would therefore be through an ether linkage, a bond which is not compatible with the observed alkaline pH lability. Alternatively, based upon analogous chemistry described by Stoodley [20] and McMillan and Stoodley [16], it was postulated that inactivation may be occurring via a dihydrothiazine mechanism (Scheme II). To investigate this possibility, the labeled serine moiety was examined by mass spectrometry in an attempt to elucidate its structure. Derivatization of the serine NH 2 with p-dimethylaminobenzaldehyde and esterification of the free COOH(s) facilitated volatilization of the sample into the ion source. Although the molecular ion was not observed, due to the intrinsic lability of the molecule to thermal-electron impact ionization, a reasonable

TABLE IV MASS SPECTRAL FRAGMENTATION OF MODIFIED SERINE MOIETY DERIVATIZED WITH p-DIMETHYLAMINOBENZALDEHYDE Possible cleavages are shown around the structural formula and the observed significant ones are tabulated. The mass spectra is of the HPLC-purified derivative of the carboxypeptidase Areleased product ('X' in Fig. 5). Typical of dimethylaminobenzaldehyde derivatives, the 'expected' m / z 147 is replaced with m / z 148 by H-transfer (see Ref. 24).

120~ C H 3 ~ _1

CH ..~.N

147

" / 1/ 6 0174

/

12g

oII

CH~N-.~I-CH~CH2~-O-- C

2.56

m/z 44

% relative intensity

120

15 39 IO0

128 147 (148) 160 174 256

7

II

2 4

S

H~

299 TABLE V MASS SPECTRAL FRAGMENTATION OF I, II, 11I AND RELEASED LABEL Values are [given as percent relative abundance.

m/z 245

231

203

186

126

59

44

100

-

100

59

17

32

-

I00

30

32

5

III

6

36

33

8

11

Label

5

59

27

19

12

I

II

80

-

Structure

" N r ~'CO 2 Me

fragmentation pattern was obtained. These data are presented in Table IV, along with the major fragmentations which strongly support the proposed dihydrothiazine mechanism. To substantiate further the dihydrothiazine inactivation scheme and to aid in better defining the structure of the label itself, mass spectra were obtained for the alkali-released label and compared with spectra of model dihydrothiazine compounds, I, II and III. The more intense ions , along with the major fragmentations corresponding to these ions are given in Table V. As shown in Table V, compound I gave an intense molecular ion at m / z 231; compound II (prepared by esterification of I) shows a strong molecular ion at m / z 245; and compound Ill (the product of a decarboxylation reaction of I occurring in mild acid) [21] has a molecular ion at m / z 203. Discussion

In this work, evidence is presented that the inactivation of penicillinase activity arising from the in situ diazotization of 6-aminopenicillanic acid is due to the labeling of a specific serine residue and that this inactivation occurs via a dihydrothiazine mechanism. The degree of inactivation obtained by the incubation of penicillinase with 6-aminopenicillanic acid and NaNO 2 is extremely dependent upon both pH [4] and the concentration ratio of NaNO2:6-aminopenicillanic acid (Table II). The

Me02C ~ . , ~ 0

H

concentrations of NaNO2 and 6-aminopenicillanic acid required to achieve complete inactivation is indeed quite high (about 10000-fold excess). However, this is not alarming when it is considered that the 6-diazopenicillanic acid, once formed, spontaneously hydrolyzes to the 6-hydroxy derivative

[22]. Peptide analysis of the chymotryptic and tryptic digests determined that the attachment site of the penicillin label occurred within the chymotryptic fragment encompassing residues 41-46. Further digestion of this chymotryptide with carboxypeptidase A enabled the identification of serine-44 as the labeled residue. Similar results have recently been reported by Knott-HunTiker et al. [21] utilizing 6-fl-bromopenicillanic acid, an affinity labeling reagent structurally similar to 6-aminopenicillanic acid. Additionally, it is interesting to note that this serine residue is conserved in the primary structure of all penicillinases sequenced to date. Mechanistically, covalent labeling of this serine residue via an ether linkage as described in Scheme I was not compatible with the demonstrated lability at mildly alkaline pH. As an alternative, we suspected that inactivation might be occurring via a dihydrothiazine mechanism (Scheme II). The basis for this hypothesis was founed upon the reasoning that: (1) because of the known spontaneous hydrolysis of 6-diazopenicillanic acid in tqueous solution [22], it was plausible to conclude that only those 6-aminopenicillanic acid molecules that are diazotized in (or very near) the active site would be subject to affinity labeling; (2) while in

300 the active site, the spatial alignment of the molecule in the binding site might be such that this reaction is favored over hydrolysis; and (3) this type of mechanism could account for the reversal of inactivation that had been observed at alkaline pH, that is, by hydrolysis of the ester linkage that would necessarily be formed. To investigate the possibility of this process taking place, mass spectral data was obtained for the labeled serine moiety and the alkali-released label itself, and a comparison made with the spectra obtained of model dihydrothiazine compounds I, II and III. As illustrated in Table IV, the fragmentation data obtained of the labeled serine moiety is entirely consistent with a dihydrothiazine-type styructure. Furthermore, as shown in Table V, the alkali-released label exhibited a fragmentation pattern identical to that observed for the model dihydrothiazine compounds, as well as an apparent molecular ion ( m / z 203) consistent with the identification of the released label as being structurally identical to compound III. Therefore, it can be concluded that: (1) the initially formed dihydrothiazine undergoes the same acid-mediated decarboxylation reaction as I (purification of inactivated enzyme was performed under low pH conditions); (2) inactivation proceeds through the dihydrothiazine mechanism by attack of the serine hydroxyl at the fl-lactam carbonyl (Scheme II). Similarly, a dihydrothiazine mechanism has also been indicated for the 6-fl-bromopenicillanic acid inactivation of B. cereus penicillinase [7,23]. In these reports, the initially formed dihydrothiazine was also found to undergo the acid-mediated transformation to the alcohol. Although the exact nature of this transformation is unknown, one suggestion has been that it is the result of an autoxidation, since it did not appear to take place under strictly anaerobic conditions [17].

Acknowledgements The initial B. cereus 569/H spores were kindly provided by D. Thatcher. The work was supported in part by NSF Grant GP8490.

References i Abraham, E.P. and Chain, E,B. (1940) Nature 146, 837 2 Rolinson, G.N., Stevens, S., Batchelor, F.R., CameronWood, J. and Chain, B. (1960) Lancet ii, 564-567 3 Sabath, L.D., Gerstein, D.A., Leaf, C.D. and Finland, M. (1970) Clin. Pharm. Ther. 11, 161-167 4 Patil, G.V. and Day, R.A. (1973) Biochim. Biophys. Acta 293, 490-496 5 Durkin, J.P., Dmitrienko, G.I. and Viswanatha, T. (1977) Can. J. Biochem, 55, 453-457 6 Knon-Hunziker, V., Oriel B.S., Sammes, P.G. and Waley, S.G. (1979) Biochem. J. 177, 361-365 7 Pratt, R.F. and Loosemore, M2. (1978) Proc. Natl. Acad. Sci. U.S.A. 75, 4145-4149 8 Kosut, M., Pollock, M.R. and Tridgell, E.J. (1956) Binchem. J. 62, 391-396 9 Alcino, J.F. (1946) Ind. Eng. Chem., Anal. Ed. 18, 619-620 10 Perret' C.J. (1954) Nature 174, 1012 II Pollock, R.M. (1960) in The Enzymes (Boyer, P.D., ed.), Vol. IV, p. 264, Academic Press, New York 12 Kornel, A. (1983) Ph.D. Dissertation, University of Cincinnati 13 Weber, K. and Osborn, M. (1969) J. Biol. Chem. 246, 4406-4412 14 Helberg, H. (1968) J. Assoc. Off. Anal. Chem. 51,552 15 Falter, H., Jayasimhulu, K. and Day, R.A. (1975) Anal. Biochem. 67, 359-371 16 McMillan, I. and Stoodley, R. (1968) J. Chem. Soc. (C), 2533-2537 17 Orlek, B., Sammes, P., Knott-HunTJker, V. and Waiey, S. (1979) J. Chem. Soc. Chem. Comm. 962-963 18 Thatcher, D.R. (1975) Biochem. J. 147, 313-326 19 Madaiah, M., and Day, R.A. (1971) Biochim. Biophys. Acta 236, 191-196 20 Stoodley, R. (1966) Tetrahedron. Len. i 1, 1205-1210 21 Knott-Hunziker, V., Waley, S.G., O r i e l B.S. and Sammes, P.G. (1979) FEBS Lett. 99, 59-60 22 Cignarella, G., Pifferi, G. and Testa, E. (1962) J. Or 8. Chem. 27, 2668 23 Cohen, S.A. and Pratt, R.F. (1980) Biochemistry 19, 3996-4003 24 Day, R.A., Falter, H., Lehman, J.P. and Hamilton, R.E. (1973) J. Orig. Chem. 38, 782-788