Degradation of β-lactam antibiotics in the presence of Zn2+ and 2-mino-2-hydroxymethylpropane-1,3-diol (Tris). A hypothetical non-enzymic model of β-lactamases

Degradation of β-lactam antibiotics in the presence of Zn2+ and 2-mino-2-hydroxymethylpropane-1,3-diol (Tris). A hypothetical non-enzymic model of β-lactamases

Degradation of /I-lactam antibiotics in the presence of Zn*+ and 2-amino-2hydroxymethylpropane-1,3-diol (Tris). A hypothetical non-enzymic model of P-...

756KB Sizes 0 Downloads 5 Views

Degradation of /I-lactam antibiotics in the presence of Zn*+ and 2-amino-2hydroxymethylpropane-1,3-diol (Tris). A hypothetical non-enzymic model of P-lactamases Mar Company, Maria J. Benitez and Juan S. JimCnez* Departamento de Quimica Fisica Aplicada, Universidad Autonoma de Madrid, 28049 Madrid, Spain (Received 15 November 1990; revised 1 March 1991) The system composed of 2-amino-2-hydroxymethylpropane-1,3-diol (Tris) and Zn2+ catalyses the degradation of cephalosporins. The fl-lactam opening fits to a first-order process, with a constant directly proportional to the zinc ion concentration. The pH and Tris concentration dependency displayed by the first-order constant, as well as the nature of the degradation products point to a mechanism that can be considered as an extension of that proposed for the benzylpenicillin degradation. The mechanism proposed here, ana’the values of the kinetic constants calculated, as compared with those of f?-lactamases, lead to the conclusion that the Tris-Zn2+ system simulates the catalytic action of the serine fNactamases rather than the action of the Zn’+-dependent type of enzymes. Keywords: p-Lactam antibiotics; Tris; kinetic mechanism; jI-Lactamases

Introduction One of the main mechanisms by which bacteria resist the action of the /I-lactam antibiotics is the production of b-lactamases (EC 3.5.2.6), which catalyse the hydrolysis of the b-lactam to innocuous products’ -4. There are various classes of /I-lactamases depending on the criterion used to classify them5. From a mechanistic point of view, however, these enzymes can be considered either as serine /I-lactamases or as Zn2+ b-lactamases. The serine /I-lactamases catalyse the hydrolysis of the b-lactam by way of an acyl-enzyme intermediate which is the ester between penicilloic acid and a serine residue of the enzyme6-9. Although displaying important differences, the /I-lactamase II from Bacillus cereusl’ and the fi-lactamase Ll from Pseudomonas maltophilia” are, to date, the two enzymes found to be specifically dependent on zinc ion for activity. The non-enzymic hydrolysis of p-lactam antibiotics catalysed by metal ions is well documented4~‘2~‘3. The divalent ions of Zn, Cu, Ni and Co have been found to catalyse efficiently the hydrolysis of penicillins (I) and cephalosporins (II). A complex between the metal ion and the /I-lactam is formed, and the more relevant hypotheses point to a mechanism in which the ion stabilizes the tetrahedral intermediate formed when the hydroxide ion attacks the fl-lactam. A different mechanism has been proposed for the system composed by zinc ion and Tris. This system has been proved to catalyse very efficiently the degradation of penicillins’4*‘s. An intramolecular nucleophilic attack of one of the hydroxyl groups of Tris on the /&lactam has been proposed as the *To whom correspondence should be addressed. 0141-8130/91/040225-06 0 1991 Butterworth-Heinemann

Limited

CEPHALOSFURIN HOOC, R, : H2N,CH-(CH2)3

C -

R2: -0-CO-CH3

CEPHALOTIN R2:-0-CO-CH3

COOH (II)

R, -CO-NH_

CEPHALORIOINE

H

O-NH-R3

probable mechanism (scheme 1). In the same reports the authors pointed out the possibility that the Tris-Zn system may be considered as a model of Zn2+-dependent /I-lactamases. The Zn2+-dependent /I-lactamase II from Bacillus cereus catalyses the degradation of penicillins and cephalosporins with a similar efficiencyi6, whereas the, also Zn2+ -dependent, /I-lactamase Ll from Pseudomonas maltophilia possesses more activity against penicillins”. /I-Lactamase I from Bacillus cereus16, and many serine j?-lactamases are also more active against penicillins5. The action of the Tris-Zn system on cephalosporins, however, does not seem to have been reported. Therefore, in order to get a deeper insight into the properties of the Tris-Zn system, and to compare it with the mode of action of fi-lactamases on different types of substrates, we have studied the interaction of most

Int. J. Biol. Macromol., 1991, Vol. 13, August

225

/SLactam

degradation

by Zn2+ and Tris: M. Company et al.

0

H

/\

CH,OH

H

Results and discussion

CH,OH

Scheme 1

cephalosporins with this buffer-metal system from the points of view of rate of degradation, mechanism, and plausible degradation products. Experimental Penicillin G, cephalosporin C, cephaloridine and cephalothin were purchased from Sigma. 2-Amino-2-(hydroxymethyl)-1,3-propane diol (Tris), and ZnSO, were from Merck. Solutions of ZnCl, were also prepared from zinc metal dissolved in an excess of dilute HCl, although no significant difference was found with respect to the use of ZnSO,. B-Lactamases I and II from Bacillus cereus were purchased as a lyophilized powder from Sigma. The separation of both forms of the enzyme was carried out as described by Davis and Abraham”. The cephalosporin degradation rates were obtained by measuring the 260 nm’ 9 absorbance diminution with respect to the total change in absorbance produced when all the /?-lactam is degraded. The quotient between the degradation product absorbance and that of native cephaloridine was determined for each pH value and several Tris-Zn system concentrations, and was found to be practically constant. The spectra registered after reaction of cephaloridine either with /?-lactamase or with the Tris-Zn system, immediately after the 260 nm level, were practically the same. Absorbance measurements were carried out in a Beckman or Perkin-Elmer spectrophotometer equipped with a thermostatic cell holder at 35°C. After a few minutes for stabilizing the temperature of the appropriate Tris-Zn (II) mixture, a small volume of a cephalosporin solution was added to initiate the reaction. The first order rate constant observed was always corrected by subtracting the much lower first order constant of aminolysis obtained in separated runs in which Zn2+ was not included. All the pH measurements were made in a thermostatic cell with a Radiometer pH-meter. A pK = 8 was used to calculate the concentration of the free base of Tris at 35°C and the ionic strength 0.5, which was adjusted by adding NaCl. The rate of penicillin degradation was measured by following the absorbance at 232nm2’, in the same way described for cephaloridine. Non-linear regression analyses were carried out by using a BMDP computer program21 on an IBM 4381. The k,,,/K, values for benzylpenicillin were obtained from initial rates at different substrate concentrations. In the case of cephaloridine, the continuous curves of absorbance diminution at 260 nm were used to obtain k,,J K,, applying the integrated form of the Michaelis-Menten equation.

226

Int. J. Biol. Macromol., 1991, Vol. 13, August

The mechanism proposed to explain the action of the Tris-Zn system on penicillins is based on the formation of a ternary complex between Tris, Zn2 + and penicillin 14,’5. The metal ion would bind both penicillin and Tris in a conformation able to facilitate the nucleophilic attack by a bound ionized hydroxyl group of Tris on the B-lactam carbonyl of penicillin. According to this mechanism, and following the report of Schwartz14, the observed firstorder rate constant should be a linear function of the total Zn2+ concentration and should increase on increasing pH. The free base of Tris concentration dependency, however, is more complex, and can be illustrated by an example of our own experiments which are practically in agreement with those already reported (Figure 1). As can be observed, the first-order constant increases on increasing the concentration of the free base of Tris, but after reaching a maximum, an inhibitory effect seems to be exerted by Tris. When several cephalosporins were used as substrates for the Tris-Zn system, the results were, however, quite different. As can be seen in Figure 2, the apparent first-order rate constant for the degradation of cephalosporins is about two orders of magnitude lower for these fi-lactams than for penicillins. Beside that, the Tris concentration dependency is different. Figure 3 shows a more comprehensive set of experiments with cephaloridine, displaying the pH and Tris concentration dependency of k,,. As can be observed, the rate constant increases until reaching an apparent ‘plateau’, without showing the maximum observed for the degradation of penicillins (see Figure

1).

Several possibilities have been analysed in order to explain the results displayed in Figure 3. In the first place, the Tris concentration dependency at constant pH excludes the hydrolysis catalysed by Zn2 ’ as the single mechanism to account for. The contribution of this mechanism to the cephaloridine degradation may be obtained by extrapolating k,, at zero concentration of the free base of Tris. Figure 4 shows that extrapolation

25

50 Trls (free hose) (mM)

75

Figure 1 First-order rate constants for the benzylpenicillin degradation in the presence of Zn’+ and Tris. The rate constants

have been normalized with respect to the total concentration of Zn’+. All measurements were carried out at 35°C and pH 8. The concentrations of penicillin and Zn*’ were 0.5 mM and 0.01 mM respectively. The continuous line is theoretical and has been obtained by computer simulation of equation (6), using k,,K, = 9.6 x lo5 1mol-’ min-’ and k,,K, = 1.3 x 10’1mol-’ mini’

/I-Lactam degradation by Zn2+ and Tris: M. Company et al.

0

40 Tris (free baseI

120

80

Figure 2 First-order rate constants for the degradation of cephalosporins in the presence of Zn2+ and Tris. The rate constants have been normalized with respect to the total concentration of Zn”. All measurements were made at 35°C and pH 8. The concentration of Zn’+ and 0.1 mM. The /&la&am concentrations were within the range of 0.07 mM to 0.2 mM. 0, cephaloridine. V, cephalothin. 0, cephalosporin C

r------l

--1 64

0

30 TM

60 (freebase)

90

Figure 3 First-order rate constants for the cephaloridine degradation as a function of pH and the concentration of the free base of Tris. All the experiments were carried out at 35°C. The Zn’+ concentration was between 0.05 mM and 0.5 mM. The cephaloridine concentration was between 0.04 mM and 0.12 mM. The continuous lines are theoretical and have been obtained by a computer simulation of equation (6), using k,,K, = 916 and k,,K, = 2726. The figures at the end of the curves indicate the pH of the experiments

for pH 8 and 7.55. The results agree well with the mechanism and constants reported by Gensmantel et al.12T13, and correspond to about 10% of the kob values obtained at the higher Tris concentrations used. This small contribution becomes negligible in the presence of Tris due to the metal ion depletion exerted by the buffer. Nevertheless it was theoretically estimated and subtracted from k,,. On the other hand, the pH dependency observed at constant concentration of the free base of Tris rules out the Zn2+ catalysed aminolysis. According to the mechanism proposed for the aminolysis”, the first order constant would not be pH-dependent at pH values below

IO

20

Tris (free bose)(mM)

Figure 4 First-order rate constants for the cephaloridine degradation. 0, pH 7.55; 0, pH 8. The experiments were carried out as described under Figure 3

the pK corresponding to the hydrolysis of Zn’+. Upon increasing the pH above neutrality, the first order constants would decrease as a consequence of the Zn2 + hydrolysis, thus yielding a pH dependency opposite to the one observed in Figure 3. Although the results of penicillins and cephalosporins degradation are different, the pH and Tris concentration dependency displayed by both types of j?-lactams suggest the possibility of a common mechanism able to yield the same kinetic equation. The absence of any maximum in Figure 3 suggests a mechanism of cephalosporin degradation similar to that described by Schwartz14 for the degradation of penicillin, in which the inhibitory effect produced by high concentration of Tris would be absent. This inhibitory effect is attributed to the formation of the unproductive complex ZnT: + (Ref. 14). Taking into account that the formation of this complex does not depend on the j?-lactam used, and considering also the higher capability of cephalosporins to complex Zn2 + as compared with penicillin,13~22 we propose the following mechanism which would explain the results obtained for the degradation of penicillin as well as those reported here for cephalosporins: ZnA+ A-/?

tiT

H+j I&2

II ZllTi

This mechanism encloses the one proposed for the and allows for the possibility penicillin degradationI

Int. J. Biol. Macromol., 1991, Vol. 13, August

227

lJ-Lactam degradation by Zn2+ and Tris: M. Company et al.

that the quaternary complex ZnT,A were able to yield the same degradation of b-la&am that the ternary complex ZnTA. The formation of the complex between the /?-la&am and Zn2 + is also included. The rate of the reaction would be: r = k,, [ZnTA]

+ k,, [ZnT,A]

(1)

where ZnTA and ZnT,A represent the complexes between Zn2+, Tris and a B-la&am, in both of which one hydroxyl group of Tris is ionized. Assuming that the breakdown of both complexes is rate determining, the previous steps being in equilibrium, we have

The data for the degradation of cephaloridine fitted well to a first-order reaction with respect to the fl-lactam (Figure 5). Therefore it seems plausible to consider negligible the terms including the concentration of the antibiotic, appearing in the denominator of equation (5). On the other hand, the first-order constant was found to be a linear function of the stoichiometric concentration of Zn2+ (Figure 6). It appears then that the last term (z) can also be considered negligible when compared with the rest of the denominator of equation (5). Therefore the apparent first-order constant normalysed for the total

r = ~,,~~,IC~~I~~~~,C~IC~IC~~2~1 +

+ ~,2~~,IC~~I~~~~2~,C~12~~lC~~2~1 (2) being assumed that the dissociation constant of the hydroxyl group of Tris has the same value in both complexes, K,. Taking into account the formation constants of each of the species as well as the proton dissociation constants, the stoichiometric concentration of Zn2+ will be: [Zn2+],

= [Zn2+](1

+ (K,/[H+])

+ CWK,K,CAI(l

+ K,,[A]

+ W,ICH+I))

+ K,(l + (K,,ICH+I)))

0

+ CT12W~WLCAl(1 + W,ICH+I)) + K,K,( 1 + (KdH+

I))))

(3)

The whole concentration of the /I-lactam, [A] t, during the course of the reaction is given by:

[Al, = [AIt1 + CZn2+1(KII + K,K,CTI(l + W,ICH+lU + WWWU2U

+ W,ICH+l)H)

(4)

C~ltCZ~lo(aC~l + bCT12) 1 + t + (c + cl)[T]

+ (d + dl)[T12

+ z

0 (mln)

005

01

( Cephalorldme)(mM)

Figure 5 First-order degradation of cephaloridine. (Panel a) First-order plots of the reaction time course under different conditions. (a), 10 mM Tris, 0.1 mM ZnSO,, pH 7.75. (b), 86 mM Tris, 0.05 mM ZnSO,, pH 8.4. (c), 100 mM Tris, 0.1 mM ZnSO,, pH 7.75. (d), 10m~ Tris, 0.1 mM ZnSO,, pH 8.20. (e), 20Om~ Tris, O.lOrn~ ZnSO,, pH 8.0. (f), 150m~ Tris, 0.5 mM ZnSO,, pH 7.50. (Panel b) Initial rates of cephaloridine degradation as a function of the initial antibiotic concentration. 0, 16 mM Tris, 0.05 mM ZnSO,, pH 8.4. 0, 20Om~ Tris, 0.5 mM ZnSO,, pH 7.55

By substituting equation (3) into equation (4), and then both expressions for [Zn”] and [A] into equation (2) we obtain, after dividing both numerator and denominator by 1 + (K,/[H+]), and rearranging, r=

IO Time

1

E

(5)

_

E

where a, b, t, c, cl, d, d,, and z are given by: a=

k,,K,K&

CH’I + K,

b=

_

k,,K,K,K,K,

CH’I + K,

N-

12 “4 2

Figure 6 First-order constants for the cephaloridine degradation as a function of the Zn2+ concentration. The reactions were carried out as described under Figure 3. 0, pH 8.2. 0, pH 7.75. In both cases the total concentration of Tris was 20Om~

228

Int. J. Biol. Macromol., 1991, Vol. 13, August

&Lactam degradation by Zn2+ and Tris: M. Company et al.

Zn2 + concentration

would

be:

(6) The term including [ T12 in the numerator appears as a consequence of the possibility that the quaternary complex ZnT,A were also able to carry out the intramolecular reaction occurring in the ternary complex, thus making the difference between this equation and the degradation. one reported l4 for the benzylpenicillin We have assumed that the same values for K,, K,, and also K, can be used for both K al7 Ka23 K,,

degradating systems, penicillins and cephalosporins. Therefore, the difference in the degradating behaviour of both fi-lactams would be attributable to the terms k,,K, and k,,K,. Both terms correspond to the product of the formation constants for the ternary and quaternary complexes and the respective catalytic constants. The continuous line of Figure 2 represents the result of a non-linear regression fitting of the experimental points to equation (6). We have used as fixed values K, = 10w9, K, = 6.4 x lo-‘, and K,, = K,, = 1.7 x lo- ‘as reported by Schwartz14, and K, = 184, K, = 59 as obtained previously by a pH titration methodz3. These values together with the parameters obtained in the fitting yield k,,K, = 9.6 x lo5 1mol-’ min-‘. This value is obtained with a standard deviation of 0.29 x 104, and agrees well with that reported by Schwartz14. kc2K4 is, however, much more unattainable; a value of 1.3 x lo5 1mol-’ min-’ is obtained but with a standard deviation of 1.9 x 104. This implies that large variations in kc2K4 do not provoke significant changes in the theoretical curve, and in fact the data reported by Schwartz14 are fitted considering that this constant is zero. Therefore, although we cannot exclude the possibility that the ZnT, complex may be able to catalyse the intramolecular degradation of benzylpenicillin, its efficiency would be much lower than that of the ZnT complex. The results for the cephaloridine degradation are indeed different. The theoretical lines in Figure 3 have been obtained in the following way: a non-linear regression fitting of the experimental points at pH 8.2 to equation (6) yielded values of k,,K, = 916 ( f 179 s.d.) lmol-‘min-‘andk,,K4=2726(f87s.d.)lmol-’min-’. Once these constants were obtained, equation (6) was simulated using a computer program. The values of K,, K,, K,, K,, K,, and K,, were the same as those used for the theoretical adjustment of the penicillin degradation, and the [H’] concentration used was that of the experimental points. Therefore we would say that both penicillin and cephaloridine are degraded by the Tris-Zn system by the same mechanism, although displaying two main differences: the penicillin degradation seems to be mainly a consequence of k,,K,, while in the case of cephaloridine both constants kclK3 and k,,K, possess a similar magnitude, being k,,K, even somewhat higher than k,,K,. On the other hand, the penicillin degradation is about two orders of magnitude more efficient than that of cephaloridine, mainly as a consequence of the difference displayed in the respective values of k,,K,. The time evolution of the cephaloridine degradation

products is known to depend on the initial compound formed upon opening of the p-la&am ringlg. When an amide is formed, an absorbance at 280nm appears in a very slow process taking a few days. This absorbance seems to be due to fragmentation of the molecule, and formation of penamaldates (III)lg. If the initial product is a cephalosporanic acid, however, that kind of compound does not seem to be formed. Figure 7 shows the time evolution of the UV spectrum registered after having degraded the cephaloridine by Tris and Zn2+. As can be observed, a band centred at 280 nm is generated, rendering the maximal absorbance a few days after the b-lactam degradation had come to an end. This result clearly contrasts with that obtained when the cephaloridine was hydrolysed by NaOH. No band centred at 280 nm was generated when the antibiotic was reacted with NaOH, neutralized and then transferred to Tris-Zn buffers of different Tris concentrations (Figure 7). According to the mechanism proposed here to explain the cephaloridine degradation by the Tris-Zn system, the formation of an amide is what one would have expected. The nucleophilic attack of the ionized hydroxyl group of Tris would give rise to an ester able to yield

1.2

[email protected]

0.4

0

Wavelength

Figure 7

350

250 (nm)

Ultraviolet spectra of the cephaloridine

degradation

products. The reaction mixture was composed by 1 M Tris and 0.5 mM ZnSO,, and was carried out at 35°C and pH 8. The cephaloridine concentration was 8.75 mM. At the indicated times aliquots ( 10 ~1) were removed from the reaction mixture, diluted into 1 ml of 50m~ phosphate pH 6.5, and the spectrum registered. Figures indicate the time (h) at which the spectra were obtained. (- -. -), spectrum of the hydrolysis degradation product obtained 65 h after reacting cephaloridine with NaOH. The spectrum of native cephaloridine in phosphate 50m~, pH 6.5 (-- -) has been included

Int. J. Biol. Macromol., 1991, Vol. 13, August

229

j-Lactam degradation by Zn2+ and Eis: M. Company et al. Table 1 Kinetic constants at 35°C for benzylpenicillin and cephaloridine degradation by fl-lactamases I and II, and by the Tris-Zn system

(ImolI’ min-‘)

Benzylpenicillin

Cephaloridine

L,lL L,lK,

1.9 x 2.9 x 9.6 x 1.3 x

0.78 x 2.3 x 0.92 x 2.73 x

k,,K, k,,K,

(P-II) (B-I)” (Tris-Zn) (Tris-Zn)

10’ 10’ lo5 105

107 10’ lo3 lo3

a In 30 mM Pipes, 0.5 M NaCl, 0.3 mM ZnSO,, pH 7 b In 50 ItIM phosphate, 0.5 M KCI, 0.1 mM EDTA, pH 7

same action exerted by the serine residue of a ,&lactamase. The Zn2+-dependent /I-lactamase II from Bacillus cereus, however, seems to act by way of a non-covalent mechanism. Zn2 ’ bound to one cysteine and three histidine residues of the enzyme would facilitate the attack by a water molecule on the /Llactam25S26. Therefore, we would conclude that the Tris-Zn system offers in its catalytic action more resemblance to the serine fllactamases than to the Zn2+-dependent enzyme. Acknowledgements This work was supported de Ciencia y Technologia

either cephalosporanic acid or the corresponding amide by interaction with the free base of Tris. A similar situation is described for the penicillin degradation by the Tris-Zn system14. Tris concentrations higher than 80m~ give rise to more than 95% of amide formation. Therefore, in the range of Tris concentration within which the contributions of aminolysis and Zn2+ catalysed hydrolysis can be considered negligible, one would expect a maximal formation of penamaldate derivative. From 50 mM to 0.5 M of the free base of Tris, this was indeed the case, as it could be deduced by comparing with the penamaldate obtained after aminolysis by Tris alone. The term k,,,/K, currently used to compare the efficiency with which different j?-lactams are hydrolysed by /3-lactamases possesses an analogous significance to the terms k,,K, and kc2K4, used here to describe the different behaviour of benzylpenicillin and cephaloridine. Both k,,K, and k,,K, are about two orders of magnitude higher for penicillin than for cephaloridine, being of most interest to note that that difference parallels the different values of k,,JK, measured for the action of the serine fi-lactamase I on benzylpenicillin and cephaloridine, but not so the k,,JK, values corresponding to the Zndependent /?-lactamase II (Table I). This enzyme is known to catalyse the degradation of penicillins and cephalosporins with a similar efficiencyi6, while fi-lactamase I, as many serine /I-lactamases, displays more activity against penicillins5. Penicillins and cephalosporins are distinguished, as good or poor substrates for the serine /I-lactamase I, by the values of the rate constants of acylation, being lower for cephalosporins than for penicillinsz4. On the other hand, the proposed nucleophilic attack by a CH,-OH group of Tris would simulate the

230

Int. J. Biol. Macromol., 1991, Vol. 13, August

in part by a Comision Interministerial Grant from the Spanish Government.

References 1 2 3 4 5 6 7

Bush, K. Antimicrob. Agents Chemother. 1989, 33, 264 Bush, K. Antimicrob. Agents Chemother. 1989, 33, 271 Ghuysen, J. M. in ‘Topics in Antibiotic Chemistry’, (Ed. P. G. Sammes), Ellis Horwood Ltd, Chichester, 1980, vol. 5, p. 9 Page, M. I. Adu. Phys. Org. Chem. (Ed. D. Bethell), Academic Press, Inc, New York, 1987, vol. 23, p. 165 Sykes, R. B. .I. Infect. Dis. 1982, 145, 765 Bicknell, R. and Waley, S. G. Biochem. J. 1985, 231, 83 Fisher, J., Belasco, J. G., Khosla, S. and Knowles, J. R. Biochemistry 1980, 19, 2895

8 9 10 11

Faraci, W. S. and Pratt, R. F. Biochem. J. 1987, 246, 651 Buckwell, S. C., Page, M. I., Waley, S. G. and Longridge, J. L. J. Chem. Sot. Perkin Trans. II, 1988, 1815 Sabath, L. D. and Abraham, E. P. Biochem. J. 1966, 98, llc Saino, Y., Kobayashi, F., Inoue, M. and Mitsuhashi, S. Antimicrob. Agents Chemother. 1982, 22, 564

12 13

Gensmantel, N. P., Gowling, E. W. and Page, M. I. .!. Chem. Sot. Perkin Trans. II, 1978, 335 Gensmantel, N. P., Proctor, P. and Page, M. I. J. Chem. Sot.

14 15 16 17

Schwartz, M. A. Bioorg. Chem. 1982, 11,4 Tomida, H. and Schwartz, M. A. J. Pharm. Sci. 1983,72, 331 Kuwabara, S. and Abraham, E. P. Biochem. J. 1967, 103, 27c Bicknell, R., Emanuel, E. L., Gagnon, J. and Waley, S. G. Biochem.

18 19

Davis, R. B. and Abraham, E. P. Biochem. J. 1974, 143, 115 Hamilton-Miller, J. M. T., Newton, G. G. F. and Abraham. E. P.

20 21

Waley, S. G. Biochem. J. 1974, 139, 789 Jennrich, R. in ‘BMDP Statistical Software’, (Ed. W. J. Dixon), Universitv of California Press. Berkelev. 1983. D. 290 Davis, R.-B. and Abraham, E.‘P. &o&m. J. 1974, 143, 129 Benitez, M. J., Company, M., Arevalillo, A. and Jimknez, J. S. Antimicrob. Agents Chemother. (in press) Martin. M. T. and Walev. S. G. Biochem. J. 1988. 254. 923 Bicknell, R. and Waley, S: G. Biochemistry 1985, ‘U, 6876 Murphy, B. P. and Pratt, R. F. Biochem. J. 1989, 258, 765

Perkin Trans. II 1980, 1725

.I. 1985, 229, 791

Biochem. J. 1970, 116, 371

22 23 24 25 26