2970.full, A Mechanism of Action For Carboxypeptidase A
2970.full, A Mechanism of Action For Carboxypeptidase A
2970.full, A Mechanism of Action For Carboxypeptidase A
USA
Vol. 69, No. 10, pp. 2970-2974, October 1972
ABSTRACT In an attempt to gain a better under- Before electron-density maps of carboxypeptidase A were
standing of the mechanism of action of carboxypeptidase available, chemical studies were undertaken in search of
A (EC 3.4.2.1), many kinetic studies have been undertaken
using numerous substrates-both esters and peptides functionally active amino-acid residues at the active site.
-that have exhibited substrate linearity, inhibition, acti- Among the numerous studies undertaken, acetylation of
vation, and sigmoid-shaped rate plots. Numerous interpre- tyrosine residues (6, 7) was shown to increase esterase activity
tations of the kinetic data have been proposed, none of and decrease peptidase activity. This, coupled with the strik-
which are fully in accord both with kinetic data and x-ray ing differences in pH-activity profiles between esters and
crystallographic studies. Much of the kinetic data has been
interpreted using multisite binding while the x-ray infor- peptides (7), led to the hypothesis that peptides and esters
mation seems to severely restrict these possibilities. bind at different, but perhaps overlapping, sites on the
We have examined the feasibility of a simple model enzyme. There are exceptions, however, to the effect of acetyla-
with a single active site, without modifier sites, that al- tion (3) and pH behavior (8).
lows only one substrate molecule to bind the enzyme at a
time. A random-pathway model was identified that simul- Kinetic studies-peptides
taneously accounts for the nonlinear kinetic data and
meets the restrictions imposed by the x-ray crystallo- Kinetic data from the study by Lumry et al. (9) show that
graphic studies. both carbobenzoxyglycyl-L-tryptophan and carbobenzoxy-
glycyl-L-phenylalanine exhibit substrate inhibition. It was
INTRODUCTION noted that the extent of inhibition by these substrates is
Carboxypeptidase A (EC 3.4.2.1) is an exopeptidase that proportional to the fourth power of the substrate concentra-
hydrolyzes peptide bonds in peptides and proteins in biological tion. From this relationship, it was concluded that four
systems, and also esters under experimental conditions, at molecules of substrate are required to inhibit an enzymic
the carboxy terminal end. It cleaves only L-amino acids with center. The model used for interpretation included both a
free carboxyl groups adjacent to the peptide or ester bond catalytically active and inactive binding site.
and is specific for amino acids that have aromatic or hydro- Upon investigation of the kinetics of benzoylglycylglycyl-L-
phobic side chains, such as phenylalanine, tryptophan, or phenylalanine and benzoylglycyl-L-phenylalanine, Auld and
leucine (4). Vallee (10) found that the latter exhibits substrate activation,
R3 H R2 H Ri while the former is linear. A model proposing the binding of
substrates at multiple loci was found to be in accord with the
etc.-CH-C-N-CH-C-N-CH-COO-(PP) + H20 data.
1 1 Kinetic studies esters
O 0
Carboxypeptidase A Data obtained by Awazu et al. (11) indicate that the hydroly-
sis of O-(N-benzoylglycyl)-L-mandelate exhibits substrate
R3 H R2 H R inhibition. The following theoretical equation was obtained
I I I I by fitting the experimental points.
etc.-CH-C-N-CH-C-OH (Ac) + HN-CH-C00-
V0 _
sO 0
(Am) [Eo = a[PPJ/{1 + b[PPJ + C[PP]2}
If we use the symbols indicated A model with two sites to which the substrate can bind-
Carboxypeptidase A one catalytic and the other inhibitory-was proposed to
PP + H20 Ac + Am account for this equation. The data thus interpreted supported
Carboxypeptidase A isolated from the pancreas has a the hypothesis that there are multiple sites to which substrates
zinc to carboxypeptidase A ratio of 0.96, indicating that the and modifiers can bind on the carboxypeptidase A molecule.
naturally occurring enzyme contains one atom of zinc per Using hippuryl-DL-/-phenyllactate, McClure et al. (12)
protein molecule (5). The zinc atom plays an integral role in noted marked inhibition with excess substrate. It was sug-
the cleavage of the substrate; it polarizes the carbonyl group gested that these results could be accounted for by the forma-
of the-substrate, Zn+. ..0O- Cs+, in order to render the
- tion of several different inactive complexes or by the co-
Downloaded by guest on December 15, 2021
carbon atom of this carbonyl group more susceptible to existence of several forms of the enzyme. An interesting par-
nucleophilic attack (3). allel between hippuryl-DL-fl-phenyllactate and its peptide
2970
Proo. Nat. Acad. Sci. USA 69 (1972) Substrate Kinetics of Carboxypeptidase A 2971
analog carbobenzoxyglycyl-L,-phenylalanine was obtained ing sites, and in most cases from three to five binding sites
(12) by a generalization of the scheme proposed by Lumry for the substrates. This conclusion was reached from the
et al. (9). Assuming that the enzyme molecules exist in two study of a general mechanism involving a monomeric form of
species, deriving steady-state equations, and fitting them to the enzyme with various binding sites. This model can also
experimental data, they found that the velocity is inversely account for the sigmoid curve obtained by Kaiser et al. (13)
proportional to the fifth power of substrate concentration. with O-hippurylglycolate. This finding gave further credence
From this relationship, it was deduced that five molecules of to the proposed mechanism involving multiple binding sites
substrate are required to inhibit one enzymic center. for substrates and similar modifiers on a single macromolecule
With the investigation of the hydrolysis of O-hippuryl- with only one catalytic site.
glycolate by Kaiser et al. (13), another type of curve-sig- A second model endeavoring to encompass the observed
moid-was observed that is typically interpreted as being kinetic properties as a unified whole was proposed by Vallee
allosteric. These authors considered the possibility that et al. (2). This model, based on multiple modes of substrate
the hydrolysis of O-hippurylglucolate by carboxypepti- binding, assumes discrete, but overlapping, productive bind-
dase A does exhibit allosteric character and this idea ing sites for esters and peptides. Inhibition is conceived of
was strengthened by the observation that the sigmoid char- resulting from (a) "incorrect" binding of the peptide at the
acter observed with the hydrolysis of O-hippurylglycolate ester site (or of the ester at the peptide site), or (b) inversion
is no longer apparent after the addition of an activating of the peptide or ester within its respective binding site to
modifier, 0.05 M N-carbobenzyloxyglycine. yield an "unproductive" complex. Activation results when
The hydrolysis of acetyl-L-mandelate studied by Kaiser the mode of substrate inhibition is prevented. Thus, this
and Carson (14) and of cinnamoyl-Dx-p-phenyllactate in- model is based upon the assumptions that, (i) carboxypep-
vestigated by Awazu et al. (11) give linear Lineweaver-Burk tidase A can distinguish between peptides and esters, (ii)
plots, which can be interpreted using the Michaelis-Menten peptides and esters bind at different sites, and (iii) there are
model. multiple modes of substrate binding that can account for
General models for interpretation activation, sigmoidness, and inhibition.
Two more general models have been propounded in an at- Structural studies
tempt to account for all of the kinetic data available. Re- Electron density maps to 0.20-nm resolution of carboxy-
garding the existence of substrate inhibition for peptidase, peptidase A and its substrate (Gly-Tyr) complex were pre-
and especially esterase activity, Quiocho and Richards (1) pared by Lipscomb et al. (3). Upon making a Gly-Tyr differ-
stated that there appears to be no mechanism for this phe- ence map, they found that the native enzyme undergoes
nomenon that requires the assumption of less than two bind- several conformational changes when Gly-Tyr is bound, re-
TABLE 1. Empirical equations for 10 substrates for carboxypeptidase A
% Standard Source
Substrate Equation error of data
Peptides
1. Carbobenzoxyglycyl-Ltryptophan =a + b[PP] + [PP 4.24 9
2. Carbobenzoxyglycyl-L-phenylalanine /= a + b[PP] + d+[PP] 3.40 9
3. Benzoylglycyl-iphenylalanine C+
l =a+ +I [PP]
b 4.83 10
4. Benzoylglycylglycyl-L-phenylalanine 4.10 10
/=a + [PP]
5. Carbobenzoxyglycyl-ileucine =a + b 6.4 9
EPPI
Esters
6. 0-(N-Benzoylglycyl)-L-mandelate iisfr a + b[PP] + [PP 6.4 lit
7. Hippuryl-DL-,-phenyllacetate l/V=a+ [P + d 5.03 12
C + [PP] [PPJ
8. 0-Hippuryl-glycolate 1,f=a + b 2.02 13
c + d [PPJ + [PPI 2
9. Cinnamoyl-Drn6-phenyllactate i/r=a + b lit
[PPJ
10. Acetyl-L-mandelate i/r=a + b 2.98 14
[PP]
* PP = polypeptide.
Downloaded by guest on December 15, 2021
E1Es 7E6E
1
Ec
Steps 1 and c
Steps 2,
7, and d Steps 4 and b Step 5
PP w r p p
w p r p
w r r p, p
0 r w p
E d- E5 X
E a ) Eb w p, p
r p w p
r r w p, p
FIG. 1. A plausible random pathway model for the following r r r p, W, p
reaction catalyzed by carboxypeptidase A p w r p
polypeptide [PP] + H20 -- cleaved amino acid [Am] + p r w p
p r r W, p
Remaining (poly)peptide [Ac]
Steps 1 and c; 2, 7, and d; and b and 4 may be either rearrange- Symbols: p = product-release step, w = water-addition step,
ment, water addition, or product-release .steps. Step 5 may be r = rearrangement step.
product release; rearrangement and product release; water addi-
tion and product release; or water addition, rearrangement, and
product release, depending upon the assignments made for the active substrates, and since (d) most of the requisites of
other steps. substrate and enzyme are the same for peptide and analogous
ester substrates, it was hypothesized (3) that the most favor-
sulting in several complementary rearrangements in the able productive binding mode for ester substrates is, in its
enzyme structure. essential interactions, the same as that depicted for peptides.
Mechanisms for the cleavage of a peptide substrate were EMPIRICAL EQUATIONS
considered. The attack on the susceptible carbonyl group is
either (i) nucleophilic catalysis by Glu-270, or (ii) general Empirical equations are a useful way of summarizing experi-
base catalysis by water. It was impossible to choose between mental data and are valuable in choosing the kinds of models
these two mechanisms. that can account for the data. Most of the kinetic data avail-
To account for the inhibition and activation exhibited by able for carboxypeptidase A has not been presented in this
kinetic studies, Lipscomb et al. (3), through model-building form; therefore, it was necessary, in most cases, to obtain the
experiments, have elicited two abortive modes of binding for data points from published figures and tables and to develop a
tripeptide substrates-displaced binding and reversed bind- suitable empirical equation by use of the procedure described
ing-that may result in substrate inhibition, the binding of by Barton and Fisher (15) for original data. Since only average
two tripeptide substrate molecules to give inhibition that is values could be obtained, each point was arbitrarily assigned
not competitive, and the binding of a tripeptide substrate a weight of one, whereas with original data the standard devi-
and a dipeptide product molecule to yield an activation ation for each point was used in weighting the points.
Table 1 shows the results for 10 representative substrates
process. Crystallographic experiments devised to demonstrate analyzed by this approach. The standard errors are in the
the abortive modes of binding were unsuccessful. In dis-
agreement with some kinetic analyses, which suggest that range of 2.0-6.4%. It should be noted that the standard error
from three to five molecules are bound, it was found that the reported by Awazu et al. (11) for O-(N-benzoylglycyl)-L-
binding region is probably limited to two substrate molecules. mandelate is equal to the highest standard error recorded
Attempts to prepare crystals of carboxypeptidase with in this paper. Thus, the standard errors defining the experi-
ester substrates have been unsuccessful; however, since (a) mental data in this paper are within reasonable limits.
the very nature of the active site restricts and severely limits
the number of productive binding modes, (b) the only obvious 5.0[
difference between analogous peptides and esters is the steric
and electronic configuration about the ester oxygen, (c) pep- C 4.0
E
tide nitrogen can be replaced by oxygen in corresponding
o 3.0
PP Am x
1.01
EH20 . ® EPPH2O E-AcH20 I I
0 1.0 2.0 3.0 4.0
4
x 10-3(m~i)
H20<____
itA H20
EAC E E-AC FIG. 3. A Lineweaver-Burk plot of the peptide benzoyl-
Am glycylgylcyl-i.phenylalanine. The line was calculated from the
model proposed in Fig. 1. The data are actual experimental
Downloaded by guest on December 15, 2021
FIG. 2. A specific example of the model in Fig. 1. points abstracted from an article by Auld and Vallee (10).
Proc. Nat. Acad. Sci. USA 69
.E
0
-4
x
i>2
81
6
ll
1.0
_ ~ ~ ~ ~ ~ ~ ~ ~ in
EPP3
(1972)
2.0
X 10 3(m9)
l
30 4.0
is l~~~~~~~~~~~~
1.0
[PPj X 102 (m)
2973