Reaction of Lys-Tyr-Lys triad mimics with benzylpenicillin: Insight into the role of Tyr150 in Class C β-lactamase

Article abstract of DOI:10.1016/S0960-894X(01)00168-8

Small and simple molecules mimicking a Lys-Tyr-Lys triad and some "mutant" derivatives were designed and synthesized. These compounds react with benzylpenicillin in water (75 mM phosphate buffer, pH 7), apparently through general base assistance by the ph

Full text of DOI:10.1016/S0960-894X(01)00168-8

Bioorganic & Medicinal Chemistry Letters 11 (2001) 1161–1164
Reaction of Lys-Tyr-Lys Triad Mimics with Benzylpenicillin:
Insight into the Role of Tyr150 in Class C ꢀ-Lactamase
Yoko Kato-Toma and Masaji Ishiguro*
Suntory Institute for Bioorganic Research, 1-1-1 Wakayamadai, Shimamoto, Osaka 618-8503, Japan
Received 20 December 2000; accepted 3March 2001
Abstract—Small and simple molecules mimicking a Lys-Tyr-Lys triad and some ‘mutant’ derivatives were designed and synthe-
sized. These compounds react with benzylpenicillin in water (75 mM phosphate buffer, pH 7), apparently through general base
assistance by the phenolic moiety. Class C b-lactamase has a Lys-Tyr-Lys triad in its active site, and our finding gives some insight
into the role of this triad in the enzymatic b-lactam hydrolysis mechanism. # 2001 Elsevier Science Ltd. All rights reserved.
Class C b-lactamases efficiently hydrolyze b-lactam
antibiotics, including cephalosporines.1 Studies using
transition-state analogue inhibitors,2 X-ray crystal
structures of the native enzyme,3 inhibitor bound
enzymes,4 and natural mutants5 have identified the
active site. Particularly, the roles of Tyr150, Lys67, and
Lys315 in the active site are of interest since they appear
to form a hydrogen-bond network with Ser64 at which
the acyl-enzyme intermediate forms. Kinetic parameters
obtained with Lys67, Lys315, and Tyr150 mutants by
Frere et al.6 and Page et al.7 suggested that the basic
residue activating Ser64 for acylation is Lys67. How-
ever, later studies by Frere et al.8 suggested the
combination of Tyr150 and Lys315 as the base. Crys-
tallography2 and pH dependent kinetic solvent isotope
effect studies9 suggested that the phenol of Tyr150 has a
reduced pKa, and should therefore exist as the phenolate
anion capable of functioning as a base. Thus, the base
activating Ser64 remains enigmatic, yet these findings at
least narrow down the choice of basic residues to Lys67,
Lys315, or Tyr150. This means that either lysine or tyr-
osine is likely to have a largely reduced, near neutral pKa
rather than their typical value close to 10.
active site of class C b-lactamase is expected to share
some similarity to these water-soluble model com-
pounds. The findings with these model compounds
agreed well with the proposed negative charge at
Tyr150, and positive charges at Lys67 and Lys315 resi-
dues under physiological pH. We now describe the use
of these compounds to explore their behavior in the
presence of a b-lactam.
The study involves the use of compounds 1–5 shown in
Scheme 1. Compound 311 was synthesized by reductive
amination of 6 followed by hydrogenolysis of the benzyl
We have already reported a model compound which
mimics the positioning of Lys67-Tyr150-Lys315 side-
chain functional groups of class C b-lactamases, and
have used this and its derivatives to measure their func-
tional group acidities.10 The shallow, solvent-exposed
Scheme 1. (a) BnBr, K2CO3, acetone, reflux 12 h (90%); (b) (COCl)2,
DMSO, Et3N, CH2Cl2, ꢀ78 ꢁC to rt (87%); (c) ethanolamine, HOAc,
NaBH3CN, MeOH, overnight (54%); (d) H2 2.5 atm, 10% Pd–C,
MeOH, rt 4 h (83%);.
*Corresponding author. Tel.:+81-75-962-3742; fax:+81-75-962-2115;
e-mail: ishiguro@sunbor.or.jp
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00168-8

1162
Y. Kato-Toma, M. Ishiguro / Bioorg. Med. Chem. Lett. 11 (2001) 1161–1164
group (Scheme 1). Compound 512 was prepared
according to known procedures. Compounds 1, 2 and 4
were prepared as described previously.10
corresponding to the synthetic compound plus benzyl-
penicillin. The carbonyl stretches in the IR spectra cor-
respond to amides. The absence of a stretch around
1740 cmꢀ1 suggested that the reaction yielded an amide,
and not an ester.
Compounds 1 and 2 both bring two primary amines
close to a phenol. This is similar to the arrangement of
the active-site side chains, Lys67, Tyr150, and Lys315
residues (Fig. 1). Compound 3 has functional groups
similar to that of 1 but the hydroxyethyl group is
attached to the amine, thereby yielding a secondary
amine. The methyl ether in ‘mutant’ 4 precludes depro-
tonation of the phenol. ‘Mutant’ 5 addresses the differ-
ence between having a phenol between two amines and
having a phenol next to a single amine.
Considering the reaction mechanism, there are two ways
in which the phenolic compounds, 1–3 and 5, can form
mono-amides. The reaction may proceed (1) via a phe-
nol ester intermediate, followed by subsequent acyl-
transfer, or (2) via activation of the amino group by the
adjacent phenolate resulting in a direct nucleophilic
attack by the amino group on the b-lactam carbonyl
carbon.
Since 1 and 2 mimic the active site, we were curious
whether these and related synthetic compounds enhance
b-lactam ring opening. Thus, we reacted compounds 1–5
with benzylpenicillin. Reactions on compounds 1–5
with benzylpenicillin were carried out at room tempera-
ture, with 5 mM of both reactants, in a 75 mM
phosphate buffer solution at pH 7.0. The decrease in
concentration of benzylpenicillin was monitored by
HPLC. Similarly, the decrease in benzylpenicillin
(5 mM) due to hydrolysis was monitored in the absence
of compounds 1–5. The time taken for 5 mM benzylpe-
nicillin to decrease to 4.5 mM (10% decrease) is shown
in Table 1.
If the first mechanism is operating, the phenolic ester
formation step should be rate limiting because the sec-
ond step would be a facile intramolecular (6-exo-trig)
reaction. Keeping in mind the pKa values of 1–3 (Table 1),
1
if the first mechanism holds, the time taken for a 10%
decrease in benzylpenicillin in the presence of 1 or 3
should be similar and should also be shorter compared
to that for 2. This is because 1 and 3, which have slightly
acidic pKa values (6.3and 6.4, respectively), should
have larger populations of deprotonated species at pH 7
compared to 2 which has a pKa of 7.0. Also, steric
1
1
hindrance around the phenol appears similar in all of
these compounds. Nevertheless, the fastest decrease in
benzylpenicillin concentration was observed with 2 (the
most basic) followed by 3, followed by 1 (the most
acidic). Thus, these findings did not strongly support the
phenol ester mechanism.
Compounds 1–5 were mono-acylated by benzylpeni-
cillin at the amine (Fig. 2). MS and IR spectra charac-
terized the b-lactam acylated products, 10–50.13 MS of
the product revealed a molecular ion peak with mass
The experimental results lend stronger support for a
mechanism where the amino group directly reacts with
the b-lactam, that is the ammonium activation mechan-
ism (Scheme 2). As Scheme 2 shows, compounds 1–3 are
all in an equilibrium at pH 7 between the overall +2
charge state and the overall +1 state. In the overall +1
state, one of the amines is capable of reacting directly
with the b-lactam at pH 7. The difference in 1 and 3
compared to 2 is in the hydroxyalkyl groups. These
groups appear to reduce the reactivity of 1 and 3 by
steric crowding around the amine, thereby making 2 the
most reactive. Between 1 and 3, although the hydroxyl
Figure 1. Class C b-lactamase active site and synthetic mimics 1 and 2.
Table 1. Time taken for 10% decrease in benzylpenicillin concentra-
tion and pKa of 1–5
Compound reacted
Time taken for 10% decrease in pKa pKa
1 2
with benzylpenicillin benzylpenicillin concentrationa (h)
Figure 2. Reaction between benzylpenicillin and 2. Analogous
reactions were carried out with 1 and 3–5.
2
3
1
5
1.38
2.45
3.60
8.25
7.0
6.4
6.3
8.7
9.7
9.7
9.0
n.d.b
4
none
44.9
50.8
8.5 10.2
n.a.c n.a.
aAll determinations were carried out at least twice. The values
obtained from separate runs were within 22% of each other.
bn.d., not determined.
cn.a., not applicable.
Scheme 2.

Y. Kato-Toma, M. Ishiguro / Bioorg. Med. Chem. Lett. 11 (2001) 1161–1164
1163
group is three bonds away from the amino nitrogen, in
both cases, 1 has primary amines and 3 has secondary
amines. Secondary amines are generally more basic
the 10% decrease is solely due to the halved effective
amine concentration, 5 should take approximately twice
as long as 2 to decrease benzylpenicillin by 10%. Also,
the pKa values show that the second amine contributes
compared to primary amines, as revealed by their pKa
2
values, and its nitrogen is expected to be more nucleo-
philic. This agrees well with the faster reaction observed
with 3 compared to 1. In the presence of compound 5,
which is a mono-amine, a 10% decrease in benzylpeni-
cillin concentration takes nearly 6 times longer com-
pared to the case with 2. If the longer time required for
to reduce pKa in 2. Compound 4 lacking the phe-
1
nolic hydroxyl group was the least reactive. Its pre-
sence hardly affected the reduction of benzylpenicillin
concentration.
Briefly summarizing these results, the time taken for
10% decrease in benzylpenicillin concentration only
loosely corresponds to the pKa value, which according
1
to previous UV studies is the pKa of the phenolic
hydroxyl group for compounds 1–3 and 5.10 This means
that the precise amount of phenolate species is probably
not so important in determining the time required for a
10% decrease of benzylpenicillin.
Certain aspects of our findings with the model com-
pounds seemed to have some relevance to the actual
enzyme mechanism. Combining our findings with pre-
vious findings by others on class C b-lactamase, the
mechanism of Ser64 activation, and subsequent forma-
tion of the tetrahedral intermediate may be like that
shown in Scheme 3. In this hypothetical mechanism, the
free enzyme has a partially deprotonated Tyr150 at
physiological pH, based on the pKa’s of compounds 1–
3.14 The substrate then binds and forms the Michaelis
complex. According to molecular modeling, the nega-
tively charged carboxylate (or sulfonate) of the b-lactam
antibiotic should be at a hydrogen bonding distance to
Lys315 and Thr316.3,15ꢀ17 Crystal structures of enzyme–
inhibitor complexes3 reveal that there is space near
Lys315 to accommodate a hydrogen bond acceptor.
Next, as a result of the carboxylate–Lys315 salt bridge,
the effect of positive charge on Tyr150 and the hydrogen
bond between Lys315 and Tyr150 are expected to
weaken. This would create a situation similar to that of
model compound 5, and the phenolic pKa of Tyr150
should rise. At this point, Tyr150 is hydrogen bonded to
Ser64 and Lys67, therefore these hydrogen bonds
should be strengthened, and this might trigger the
nucleophilic attack of Ser64 on the b-lactam carbonyl
carbon.18 This step which could be either stepwise or
concerted has some similarity to amine activation seen
in compounds 1–3. The circuitous proton abstraction
via Lys67 should also be possible, but would involve
reorientation of the hydrogen bonds. The two routes to
Ser64 proton abstraction have been assessed by b-sec-
ondary and solvent deuterium kinetic isotope effects.19
Although a clear-cut conclusion could not be drawn, the
results seem to favor the direct proton acceptance
mechanism by Tyr150. Yet, perhaps this possibility of
alternative pathways to achieve the same goal brings
forth the versatility of these enzymes and makes
mechanism elucidation through mutation studies and
modified substrates difficult.
In summary, class C b-lactamase active-site model
compounds and their ‘mutants’ demonstrated the
variability of the phenolic pKa and to a lesser extent the
amino pKa. Acylation of these compounds by benzyl-
penicillin suggested that the phenol activates the
Scheme 3.

1164
Y. Kato-Toma, M. Ishiguro / Bioorg. Med. Chem. Lett. 11 (2001) 1161–1164
adjacent ammonium for acylation at physiological pH.
A hypothetical mechanism of Ser64 activation by
Tyr150 in the formation of the tetrahedral intermediate
in class C b-lactamase can be proposed by incorporating
the findings from compounds 1–5 to previous experi-
mental and computational findings by others. The likely
course of proton transfer is a direct transfer from Ser64
to Tyr150; however, a relay of proton transfers from
Lys67 to Tyr150, and then from Ser64 to Lys67 may
also be possible.
9. Page, M. I.; Vilanova, B.; Layland, N. J. J. Am. Chem. Soc.
1995, 117, 12092.
10. Kato-Toma, Y.; Imajo, S.; Ishiguro, M. Tetrahderon Lett.
1998, 39, 51.
11. Spectral data of the new compound are shown below: 3.
1H NMR (CD3OD) d 6.851 (s, 2H), 3.870 (s, 4H), 3.659 (t, 4H,
J=5.6 Hz), 2.731 (t, 4H, J=5.8 Hz), 2.201 (s, 3H); 13C NMR
(CD3OD)
d 154.206, 135.772, 133.340, 122.623, 58.633,
48.428, 21.161; HRMS (FAB in 3-nitrobenzyl alcohol and
glycerol) calcd for C13H23O3N2 (M+H), 255.1709; found,
255.1691.
12. Ludeman, S. M.; Zon, G. J. Med. Chem. 1975, 18, 1251.
13. Epimerization of the N,S-acetal and formation of diaster-
Our next goal in understanding the reaction mechanism
at the atomic level is to clarify the pKa values and pro-
tonation state of the active-site residues in the native
enzyme. Accordingly, construction of class C b-lacta-
mase variants that should allow clarification of the
basicity of Lys67 and Tyr150 is now under way.
eomers in compound 10 rendered H NMR peak assignment
1
difficult. IR and MS data of 10–50 are shown below: 10: IR
(film) 3298, 1651, 1394, 1208, 1132, 1046 cmꢀ1; MS (FAB)
calcd for C27H37O7N4S (M+H), 561.24; found, 561.24; 20: IR
(film) 3059, 1671, 1198, 1138, 723 cmꢀ1; MS (FAB) calcd for
C25H33O5N4S (M+H), 501.22; found, 501.22; 30: IR (film)
3300, 1615, 1357, 1070 cmꢀ1 MS (FAB) calcd for
C29H41O7N4S (M+H), 589.27; found, 589.1; 40: IR (film)
3289, 2974, 1674, 1203, 1137, 722 cmꢀ1 MS (FAB) calcd for
C26H35O5N4S (M+H), 515.23; found, 515.23; 50: IR (film)
3288, 2978, 1730, 1664, 1536, 1200, 1145, 722 cmꢀ1; MS (FAB)
calcd for C23H28O5N3S (M+H), 458.18; found, 458.18.
14. This is similar to the case reported in UDP-galactose
4-epimerase where a positive electrostatic field stabilizes the
phenolate form of tyrosine thereby giving it a low pKa: Liu,
Y.; Thoden, J. B.; Kim, J.; Berger, E.; Gulick, A. M.; Ruzicka,
F. J.; Holden, H. M.; Frey, P. A. Biochemistry 1997, 36,
10675.
15. Lamotte-Brasseur, J.; Dive, G.; Dideberg, O.; Charlier, P.;
Frere, J.-M.; Ghuysen, J.-M. Biochem. J. 1991, 279, 213.
16. Juteau, J.-M.; Billings, E.; Knox, J. R.; Levesque, R. C.
Protein Eng. 1992, 5, 693.
17. Although a modified substrate with an esterified carboxyl-
ate did not show severely impaired activity,7 an ester is still a
good hydrogen bond acceptor. Therefore, this result should
not rule out the carboxylate–Lys315 interaction.
References and Notes
1. For a review on b-lactamases, see: Page, M. I.; Laws, A. P.
Chem. Commun. 1998, 1609.
2. Li, N.; Pratt, R. F. J. Am. Chem. Soc. 1998, 120, 4264.
Curley, K.; Pratt, R. F. J. Am. Chem. Soc. 1997, 119, 1529.
3. Oefner, C.; D’Arcy, A.; Daly, J. J.; Gubernator, K.; Char-
nas, R. L.; Heinze, I.; Hubschwerlen, C.; Winkler, F. K. Nat-
ure 1990, 343, 284.
4. Usher, K. C.; Blszczak, L. C.; Weston, G. S.; Shoichet, B.
K.; Remington, S. J. Biochemistry 1998, 37, 16082. Lob-
kovsky, E.; Billings, E. M.; Moews, P. C.; Rahil, J.; Pratt, R.
F.; Knox, J. R. Biochemistry 1994, 33, 6762. Lobkovsky, E.;
Moews, P. C.; Liu, H.; Zhao, H.; Frere, J.-M. Proc. Natl.
Acad. Sci. U.S.A. 1993, 90, 11257.
5. Crichlow, G. V.; Kuzin, A. P.; Nukaga, M.; Mayama, K.;
Sawai, T.; Knox, J. R. Biochemistry 1999, 38, 10256.
6. Monnaie, D.; Dubus, A.; Cooke, D.; Marchand-Brynaert,
J.; Normark, S.; Frere, J.-M. Biochemistry 1994, 33, 5193.
Monnaie, D.; Dubus, A.; Frere, J.-M. Biochem. J. 1994, 302, 1.
7. Dubus, A.; Normark, S.; Kania, M.; Page, M. G. P. Bio-
chemistry 1994, 33, 8577.
18. Although tyrosine usually does not abstract protons, such
a role has been assigned to a tyrosine in the active site of thy-
midylate synthase: Liu, Y.; Barrett, J. E.; Schultz, P. G.; Santi,
D. V. Biochemistry 1999, 38, 848.
19. Adedrian, S. A.; Pratt, R. F. Biochemistry 1999, 38,
1469.
8. Dubus, A.; Ledent, P.; Lamotte-Brasseur, J.; Frere, J.-M.
Proteins: Struct., Funct., Genet. 1996, 25, 473.