CARBOXYPEPTIDASE, A ZINC METALLOENZYME

CARBOXYPEPTIDASE, A ZINC METALLOENZYME*
BY BERT L. VALLEE AND HANS NEURATH
(From the Biophysics Research Laboratory of the Department of Medicine, Harvard
Medical School and the Peter Bent Brigham Hospital, Boston, Massachusetts, and
the Department of Biochemistry, University of Washington, Seattle, Washington)
(Received for publication, April 22, 1955)
The physical, chemical, a.nd enzymatic properties of pancreatic carboxypeptidase
have been the subject of a recent, comprehensive review (24).
The present study was undertaken to reexamine several features of the
mechanism of action of this peptidase and to extend previous investigations
on the chemical composition and the inhibition of the enzyme. The
data presented herein demonstrate that carboxypeptidase is a zinc metalloenzyme.
The metal is firmly bound to the protein and is apparently
indispensable for enzymatic activity.
EXPERIMENTAL
Methods and Materials
Protein concentration was determined by precipitation with trichloroacetic
acid (5), by micro-Kjeldahl nitrogen determinations, assuming 15.1
per cent nitrogen (6), or by absorption at 278 rnp, assuming1 a molar extinction
coefficient of 8.6 X 104.
Zinc was determined by microchemical and spectrochemical techniques.
Two microchemical methods, in which dithizone was employed but differing
essentially in the mode of sample preparation, were found to yield
equivalent results, as described previously (7, 8).
Copper was measured microchemically by diethyldithiocarbamate (9),
with modifications of the spectrophotometry as subsequently reported (10).
Emission SpectrographySamples were dry ashed at 450” in a thermostatically
controlled electric muffle furnace. Internal standards were
added to the ash. Aliquots were sparked in porous cup electrodes. Jarrell-
Ash “varisources” were employed to generate and control the spark. The
two spectrographs were 21 foot Wadsworth mountings but differed in the
characteristics of their gratings. One instrument was provided with a
15,000 line per inch grating reflected maximally in the first order violet.
The reciprocal linear dispersions were 5.18 and 5.25 A per mm., respectively.
Eastman No. 103-O photographic plates were employed throughout. Lines
and backgrounds were measured with a densitometer. Working curves
* A preliminary account has been published (1).
1 Unpublished determinations by Dr. G. S. Albrecht.
253
Downloaded from http://www.jbc.org/ by guest on April 13, 2017
254 CARBOXYPEPTIDASE
were prepared with spectroscopically pure chemicals. Zinc was determined
both chemically and spectrographically to insure accuracy.
All analytical procedures were carried out with precautions against metal
contamination. Water wasobtained by slow passage through mixedIR-120
and IRA-420 ion exchange resins (Rohm and Haas). The effluent had a
specific resistance of at least 1.5 X lo6 ohms. Reagents were freed of metals
when necessary and stored in acid-cleaned polyethylene bottles throughout.
Enzymatic activity was determined by methods previously described (11,
12), with carbobenzoxyglycyl-L-phenylalanine as substrate. As a matter
of routine, enzymatic assays were performed at 25” with an initial substrate
concentration of 0.02 M in a 0.02 M Verona1 buffer, pH 7.5, containing
0.1 M NaCl. Activities were expressed as apparent proteolytic coefficients
(cf. (2)), calculated from the strictly linear portions of first order
reaction plots observed when hydrolysis did not exceed about 15 per cent.
Crystalline curboxypeptiduse was prepared from the exudate of freshly
collected, frozen beef pancreas glands2 (2,3, 13). Commercial preparations,
obtained from the Worthington Biochemical Sales Corporation, Freehold,
New Jersey, and from Armour and Company, Chicago, were analyzed without
further purification.
Substrate and Inhz’bitors-Carbobenzoxyglycyl-L-phenylalanine was prepared
in part by methods previously published (13) ; another lot was synthesized
by Dr. Murray Goodman, Department of Chemistry, Massachusetts
Institute of Technology, by a method to be described elsewhere.
1 , lO-Phenanthroline, 2,2’-dipyridyl, 8-hydroxyquinoline-5-sulfonic acid,
2-carboxy-2’-hydroxy-5’-sulformazylbenzene (Zincon), and 2-acetylamino-
1,3,4-thiodiazole-5-sulfonamide (Diamox) were obtained from commercial
sources and used without further purification.
Fractionation of Pancreatic Exudate-In order to follow the metal content
and the specific enzymatic activity through the process of isolation
and crystallization of carboxypeptidase, aliquots of the following fractions
were analyzed for zinc and other metals, nitrogen, and enzymatic activity:
(1) the crude pancreatic juice; (2) the euglobulin precipitate obtained after
lo-fold dilution with water of the activated juice, at pH 4.6; (3) the su,pernatant
solution of the euglobulin precipitate; (4) the enzymatically inactive
solution obtained by extraction of the euglobulin precipitate with barium
hydroxide at pH 6.0 (referred to in Table II as Ba(OH)2 extract); and (5)
the enzyme obtained after the first through the fifth crystallization. For
metal analysis these protein fractions were dried to constant weight first
in a vacuum desiccator and then in an oven at 108”. Protein concentra-
2 These glands were shipped to this laboratory, frozen, by Armour and Company,
Chicago, Illinois, and Spokane, Washington.
Downloaded from http://www.jbc.org/ by guest on April 13, 2017
B. L. VALLEE AND H. NEURATH 255
tion of fractions containing non-protein nitrogen was determined by the
micro-Kjeldahl method on the material precipitable by trichloroacetic acid
added to a final concentration of 5 per cent. Enzymatic activities were
determined on the freshly prepared fractions.
Results
Metal Analyses
Table I presents spectrographic analyses for all metals and microchemical
analyses for zinc and copper performed on four different preparations of
the crystalline enzyme. 10 to 20 mg. of enzyme were employed fol analy-
TABLE I
Metal Analysis of Crystalline Carboxypeptidaae+
Data expressed as micrograms of metal per gm. of carboxypeptidase.
Elementt
Zinc
“
Copper
Iron
Aluminum
Magnesium
Calcium
Barium
Strontium
Manganese
Method
Dithizone
Spectrography
‘I
‘I
“
I‘
‘I
“
“
“
Prepiration
1820
1820
33
40
6
6
3
18
Not found
“ “
Preparation
2
1770
Not done
“ ‘I
54
4
6
38
9
Not found
I‘ I‘
Preparation
3
1800
1700
Not done
36
11
17
16
48
1
Not found
-
1
.-
-
Preparation
4
1980
1665
64
75
73
16
6
114
1
1
* For a description of preparations, see the text.
t Not found: beryllium, boron, cadmium, chromium, cobalt, lead, lithium, molybdenum,
nickel, phosphorus, potassium, silver, tin.
sis. Preparations 1 and ‘2 were prepared in our laboratory and recrystallized
five and six times, respectively. Preparations 3 and 4 were obtained from
two different commercial sources and had been recrystallized five and three
times, respectively. All four preparations have comparable enzymatic
activities and show comparable and consistently high zinc contents, both
by spectrographic and chemical analyses. The observed differences are entirely
within the error of the techniques; since zinc is easily lost on dry
ashing, the microchemical data obtained by means of the trichloroacetic acid
precipitation technique (5) are considered more reliable. Based on a molecular
weight of carboxypeptidase of 34,400 (2, 6), the ratio of moles of zinc
per mole of carboxypeptidase for these preparations is 0.96, 0.93, 1.04, and
0.95, respectively. All other elements are present in stoichiometrically
and absolutely insignificant amounts. The two preparations from our
Downloaded from http://www.jbc.org/ by guest on April 13, 2017
256 CARBOXYPEPTIDABE
own laboratory, Preparations 1 and 2, contain 5.7 and 6.0 7 of Mg per gm.
of protein, and the two commercial preparations, (Nos. 3 and 4) 17 and
16 y of Mg per gm., respectively. Lithium, employed in recrystallization,
is not detected; barium, added in large quantities during purification, is
apparently removed during recrystallization. Preparation 4, material recrystallized
three times, contains larger amounts of extraneous metals than
do those recrystallized five or six times.
Table II contains metal analyses and activity measurements obtained
on pancreatic juice, the fractions attending the isolation of carboxypeptidase,
and the five recrystallizations of the enzyme. The zinc content
per gm. of protein of pancreatic juice is 310 y per gm. and rises to 1870 y
per gm. in the first crystals. There is virtually no change in zinc content
with recrystallization. The ratio of moles of zinc per mole of crystalline
carboxypeptidase remains close to 1 throughout five recrystallizations.
The specific activity in active fractions increases and is parallel to the zinc
content. The ratio of specific .activity to zinc rises as “extraneous protein”
and “extraneous zinc” are removed. The individual concentrations of
all other elements and their sum decrease with purification as zinc and specific
activity increase. Magnesium concentration falls from more than 1000
y per gm. of protein in the pancreatic juice to 29 y per gm. of protein in the
euglobulin precipitate, and to even lower levels in the crystals. Strontium,
molybdenum, and iron are apparently introduced with barium hydroxide
and are removed during crystallization. Chromium and lead appear as
spurious contaminants. A slight rise in calcium, magnesium, aluminum,
and iron, after the third crystallization, is attributed to impurities in LiCl
or water employed for recrystallization. The data in Tables I and II indicate
that recrystallization accomplishes the removal of extraneous metals.
There is virtually no change in protein composition of the crystals, as evidenced
by the constant Zn to protein ratio.
Dialysis of carboxypeptidase for 18 hours against water or ammonia does
not remove zinc from the protein. A preparation containing 2000 y of
Zn per gm. of carboxypeptidase before dialysis contained 1900 y of Zn per
gm. after dialysis against water, and 2400 y of Zn per gm. after dialysis
against 5 X 10m9 M ammonia.
Enzymatic Activity
The activity of carboxypeptidase recrystallized five times is inhibited
by metal-chelating agents, such as S-hydroxyquinoline-5-sulfonic acid, 1, lophenanthroline,
2,2’-dipyridyl, and, to a lesser extent, Versene. In these experiments,
the buffered enzyme solutions are incubated with the chelating
agent at pH 7.5, O”, for 1 hour prior to the addition of the substrate. Inhibition
does not occur when these chelating agents are first incubated with
Downloaded from http://www.jbc.org/ by guest on April 13, 2017
TABLE II
Metal Confent of Fractions Attending Zsolation of Carbozypeptidase from Pancreatic Juice
Unless ot(herwise indicated, metal content is given in micrograms per gm. of protein. All zinc determinations by the dithizone
method in duplicate, and in quadruplicate for the crystalline enzyme preparations. Copper determined by the sodium dithiocarbamate
method. Spectrographic analyses were made in duplicate, and separately on two different instruments.
Pancreatic juice.. .
6: Euglobulin ppt.. .
* Supernatant
Ba(OH)t extract..
1st crystals. . .
2nd “ . . . . . .
3rd “ . . . . . .
4th “ . . . .
5th “ . . . . . . .
310
590
450
410
1870
1860
0.16 0.18 0.58
0.31 1.43 2.43
0.24 0.05 0.11
0.21 0.05 0.12
0.98 13.2 7.05
0.98 # #
0.97 14.5 7.85
0.99 14.0 7.45
1.05 18.6 9.30
-
ACtk-
ACtiVity.
Ct
2
x?w
--
Sr
0.3
2.4
0.7
135
3.9
1.1
-
--
-
Ba ca Mg Al
1.6 520 1160 1.9
6.9 71.0 29.0 19.0
5.7 300 1250 20.0
>1006 212 1050 11.0
>lOOO 32.0 13.0 32.0
219 23.0 20.0 12.0
98.0 33.0 1.5 4.2
52.0 30.0 43.0 12.0
28.0 50.0 29.0 12.0
Fe
Other metals
65.0
460
66.0
760
130
36.0
51.0
50.0
-
--
-
* Calculated as moles of zinc versus per mole of carboxypeptidase.
t Expressed as the first order rate constant (in decimal logarithm) per mg. of enzyme N per ml.
t: Not determined.
Mu
4.7
5.2
3.7
MO
41.0
CU
19.0
%
43.0
51.0
82.0
:
Cr
1.7
10.0
Pb
33.0
Downloaded from http://www.jbc.org/ by guest on April 13, 2017
258 CARBOXYPEPTIDASE
an equimolar amount of zinc, cupric, or ferrous ions. Zincon, sodium diethyldithiocarbamate,
and Diamox (the latter employed because of its effect
on carbonic anhydrase) had little, if any, effect on carboxypeptidase activity.
The effects of increasing concentrations of 8-hydroxyquinoline-5-sulfonic
acid, 1, IO-phenanthroline, and 2 ,a’-dipyridyl on the activity of carboxypeptidase
at a constant substrate concentration of 0.02 M carbobenzoxyglycyl+
phenylalanine are shown in Fig. I. Activity of the inhibited
reaction is expressed as per cent of the proteolytic coeflicient observed at
IOO-
:: 80-
s
$ 80.
P 40.
h
5 20-
MOLAR CONE NTRATION OF INHIBITOR
FIN. 1. Semilogarithmic plot of the per cent activity of crystalline carboxypeptidase
versus the concentration of 1 ,lO-phenanthroline ( l ), &Lhydroxyquinoline-5-
sulfonic acid (A), and 2,2’-dipyridyl (0). The inhibitors were incubated with the
enzyme for 1 hour at 0” in a Verona1 buffer, pH 7.5, prior to the addition of the substrate.
Substrate concentration, 0.02 M throughout. The experimental points
were connected by plotting a curve tangent to them by the principle of least squares.
zero inhibitor concentration. The conditions of preincubation are indicated.
DISCUSSION
The present data establish carboxypeptidase as a zinc metalloenzyme
within the framework of previous definition (34, 15). It appears that 1
atom of zinc is firmly bound to the protein of carboxypeptidase.
The zinc content increases throughout fractionation in the very fractions
in which carboxypeptidase activity is increased. The zinc content becomes
constant with crystallization and is not altered materially by recrystallization.
At the same time, other metals decrease to absolutely and
stoichiometrically insignificant amounts. Zinc is firmly bound to the protein.
This is implied by the fact that the zinc to protein bond is maintained
through the changes in pH, ionic strength, and temperature and
against competition of other ions to which the enzyme is exposed in the
course of fractionation. The metal is not removed by dialysis against
Downloaded from http://www.jbc.org/ by guest on April 13, 2017
B. L. VALLEE AND H. NEURATH 259
water or by prolonged standing in water at pH values considered suitable
for the integrity of the protein. Zinc, not associated with the carboxypeptidase
of pancreatic juice, as indicated by the absence of enzymatic activity,
is removed during fractionation, as is ionic zinc, which may be introduced
from reagents, water, or glassware (Table II). No precautions
against contamination with zinc or other metals were taken during fractionation
of the enzyme in the commercial laboratories or in ours. The
levels of contamination introduced were probably different in all three
laboratories; yet the final ratio of moles of zinc to moles of enzyme is so
close in all instances as to be virtually identical. This indicates that the
firm Zn to carboxypeptidase bond exists in the “natural state” (16).
Zinc content and the activity of the enzyme are directly related. The
metal is aggregated in the fractions of highest activity (Table II).
Throughout recrystallization, both the enzymatic activities and the zinc
content remain at a constant level. The presence of all other metals is
apparently unrelated to activity through fractionation. Thus, constant
zinc to protein, zinc to activity, activity to protein ratios are achieved with
purification of carboxypeptidase. It is unlikely that the metal is a fortuitous
contaminant, since the molar zinc to protein ratio of the crystalline
enzyme is an integral number, i.e. 1.
Preincubation of carboxypeptidase with various chelating agents produces
marked inhibition of enzymatic activity. Such inhibition does not
occur when chelating agents are first incubated with zinc, cupric, or ferrous
ions to form the respective metal chelate. This would seem to indicate
that the sites of chelation of these compounds are responsible for the
observed inhibition. Inhibition is therefore not caused by any structural
similarity between the inhibitors and the substrate.
The agents employed are known to form stable complexes with Zn++ in
solution, and their physical chemistry has been studied (17-19). These
agents apparently inhibit carboxypeptidase through their effects on its
zinc atom, possibly by formation of a complex. Calculations of the stoichiometry
for similar reversible enzyme-inhibitor complexes have been proposed
(20). Such calculations are based on several tacitly assumed conditions.
Inhibition is presumed to be fully and freely reversible under the
test conditions. The substrate is presumed not to compete or interact
with the inhibitor. Assuming that 2 moles of inhibitor bind to each zinc
atom, occupying four of the six possible covalent bonds of zinc, the data
are compatible with the calculations.
The inhibitory effects of chelating agents, while consistent with the stability
of their zinc complexes in solution, are no direct function of this parameter,
and their physicochemical interpretations present the difficulties
commonly encountered with data on mixed complexes (21). The geometric
arrangement of the zinc atoms with respect to protein and ligand mole-
Downloaded from http://www.jbc.org/ by guest on April 13, 2017
260 CARBOXYPEPTIDASE
cules, the steric and chemical factors contributed by the reactive polar
groups of protein and ligand and their respective charges would alter significantly
the constants arrived at on the basis of simpler systems (17-19).
A full elucidation of the rale of zinc in the functions of carboxypeptidase
has to await additional experiments, now in progress, on the reversal of
enzymatic inhibition, on the ease of exchange between bound zinc and free
zinc or other metal ions, and on the contribution of zinc to the stability
of the protein.
The studies of inhibition of carboxypeptidase presented here support the
conclusion that zinc is both a structural and functional component of the
enzyme and that it participates in its catalytic action.
It has been reported that a large fraction of Zna administered to dogs
is excreted in pancreatic juice (22). No explanation for this finding has
been made. The data presented here make it appear likely that at least
part of this zinc is associated with carboxypeptidase in pancreatic juice.
Carboxypeptidase was first crystallized by Anson in 1937 (23). Interest
in the mode of action of this enzyme was renewed by reports (24, 25) that
magnesium is concerned with carboxypeptidase activity. This conclusion
was based on the qualitative identification, by emission spectrography, of
significant amounts of magnesium in the ash of the enzyme. This qualitative
finding was given added weight, since only traces of copper and iron
could be found, while zinc, manganese, cobalt, barium, or lithium could
not be detected at all. The analytical data were thought to be supported
by studies indicating that cyanide, sulfide, phosphate, pyrophosphate, citrate,
oxalate, and cysteine inhibited the enzyme, though flouride alone or
in combination with 0.01 M orthophosphate did not. Subsequent work by
other investigators (12) failed to confirm some of these findings, while others
were shown to have resulted from inhibition by one of the products of the
reaction (n-phenylalanine). The present data give no support to the report
that magnesium is in any way associated with carboxypeptidase action;
the hypothetical considerations advanced on the basis of such an association
are therefore without foundation.
Our thanks are due to Mrs. Alice Abrahamian, Miss Elaine Cohen, Mr.
Thomas Coombs, Mr. Sumio Go, and Miss Flora Lerner for technical assistance.
This work has been supported by grants from the Rockefeller
Foundation, the United States Public Health Service, and by a contract
between the Office of Naval Research, Department of the Navy, and Harvard
University, contract NR 119277.
SUMMARY
Quantitative emission spectroscopy and chemical analyses have established
that crystalline pancreatic carboxypeptidase is a zinc metalloenzyme
Downloaded from http://www.jbc.org/ by guest on April 13, 2017
B. L. VALLEE AND H. NEURATH 261
containing 1 atom of zinc per molecule of enzyme protein. Metal analyses
of the fractions attending the isolation of the enzyme from pancreatic
exudate have shown that zinc is the only metal aggregated in the fractions
with increasing enzymatic activity, thus leading to constant activity to zinc
ratios.
Enzymatic activity is inhibited by metal-chelating agents such as 8-
hydroxyquinoline-5-sulfonic acid, 1, lo-phenanthroline, or 2,2’-dipyridyl.
Zinc is both a structural and functional component of carboxypeptidase
and participates in the mechanisms of its catalytic action.
BIBLIOGRAPHY
1. Vallee, B. L., and Neurath, H., J. Am. Chem. Sot., 76, !NO6 (1954).
2. Green, N. M., and Neurath, H., in Neurath, H., and Bailey, K., The proteins,
New York, 2, chapter 25 (1954).
3. Neurath, H., and Schwert, G. W., Chem. Rev., 46, 69 (1950).
4. Smith, E. L., Advances in Enzymol., 12.191 (1951).
5. Hoch, F. L., and Vallee, B. L., Anal. Chem., 26,317 (1953).
6. Smith, E. L., and Stockell, A., J. Biol. Chem., 207, 501 (1954).
7. Vallee, B. L., and Gibson, J. G., 2nd, J. Biol. Chem., 176, 435 (1948).
8. Hoch, F. L., and Vallee, B. L., J. Biol. Chem., 181,295 (1949).
9. Gubler, C. J., Lahey,M.E., Ashenbrucker,H., Cartwright, G. E., and Wintrobe,
M. M., J. Biol. Chem., 196,299 (1952).
10. Vallee, B. L., Anal. Chem., 26,985 (1953).
11. Snoke, J. E., and Neurath, H., J. BioZ. Chem., 181,789 (1949).
12. Neurath, H., and De Maria, G., J. BioZ. Chem., 186, 653 (1950).
13. Neurath, H., in Colowick, S. P., and Kaplan, N. O., Methods in enzymology,
New York, 2, chapter 8, in press.
14. Vallee, B. L., Scientijic MO., 72,368 (1951).
15. Vallee, B. L., in Harrison, T. R., Principles of internal medicine, Philadelphia,
chapter 51 (1954).
16. Edsall, J. T., Enzymes and enzyme systems, Cambridge (1951).
17. Kolthoff, I. M., Leussing, D. L., and Lee, T. S., J. Am. Chem. Sot., 73,390 (1951).
18. Albert, A., Biochem. J., 64,646 (1953).
19. Liiuger, P. G., Fallab, S., and Erlenmeyer, H., HeZv. chim. acta, 28, 92 (1955).
20. Smith, E. L., Lumry, R., and Polglase, W. J., J. Phys. and CoZZoid. Chem., 66,
125 (1951).
21. Klotz, I. M., and Loh Ming, W.-C., J. Am. Chem. Sot., 76,805 (1954).
22. Montgomery, M. L., Sheline, G. E., and Chaikoff, I. L., J. Exp. Med., 78. 151
(1943).
23. Anson, M. L., J. Gen. Physiol., 20, 663 (1937).
24. Smith, E. L., and Hanson, H. T., J. BioZ. Chem., 176,997 (1948).
25. Smith, E. L., and Hanson, H. T., J. BioZ. Chem., 179,803 (1949).
Downloaded from http://www.jbc.org/ by guest on April 13, 2017
Bert L. Vallee and Hans Neurath
METALLOENZYME
CARBOXYPEPTIDASE, A ZINC
J. Biol. Chem. 1955, 217:253-262.
http://www.jbc.org/content/217/1/253.citation
Access the most updated version of this article at
Alerts:
• When a correction for this article is posted
• When this article is cited
alerts
Click here to choose from all of JBC's e-mail
tml#ref-list-1
http://www.jbc.org/content/217/1/253.citation.full.h
accessed free at
This article cites 0 references, 0 of which can be
Downloaded from http://www.jbc.org/ by guest on April 13, 2017