Voet D., Voet Ju.G. Biochemistry

Подождите немного. Документ загружается.

b. The Lysozyme Catalytic Mechanism Proceeds

via a Covalent Intermediate

Lysozyme’s catalytic mechanism was largely formulated

by Phillips based on the foregoing information. However,

as we shall see, further investigations have required that

important changes be made to the original version of this

mechanism. The presently accepted mechanism occurs as

follows (Fig. 15-14):

Section 15-2. Lysozyme 521

Figure 15-14 The lysozyme reaction mechanism. Glu 35 acts

as an acid catalyst, and Asp 52 acts as a covalent catalyst. Only

the substrate D- and E-rings are shown. R represents the

N-acetyl group at C2, and R¿ represents the CH

CHCOO



group

at C3.The resonance-stabilized oxonium ion transition state

–

HOCH

⬘

Covalent

intermediate

Glu35

HOCH

General

acid

catalysis

General

base

catalysis

Substrate

polysaccharide

Binding

Covalent

catalysis

Water binding

⬘

–

Glu35

Asp52

–

HOCH

⬘

HOCH

⬘

HOCH

⬘

Oxonium ion

(transition state)

–

HOCH

⬘

Glu35

Asp52

requires that C1, C2, C5, and O5 be coplanar (orange shading)

creating a half-chair conformation. Step 5 includes the

participation of an oxonium ion transition state that is not

shown.

See Kinemage Exercise 9

JWCL281_c15_506-556.qxd 2/19/10 9:27 PM Page 521

1. Lysozyme attaches to a bacterial cell wall by binding

to a hexasaccharide unit. In the process, the D-ring is dis-

torted toward the half-chair conformation in response to

the unfavorable contacts that its ¬C6H

OH group would

otherwise make with the protein.

2. Glu 35 transfers its proton to the O1 atom linking the

D- and E-rings, the only polar group in its vicinity, thereby

cleaving the C1¬O1 bond (general acid catalysis).This step

converts the D-ring to a planar resonance-stabilized oxo-

nium ion transition state, whose formation is facilitated by

the strain distorting it to the half-chair conformation (catal-

ysis by the preferential binding of the transition state). The

positively charged oxonium ion is stabilized by the presence

of the nearby negatively charged Asp 52 carboxylate group

(electrostatic catalysis).The E-ring product is released.

3. The Asp 52 carboxylate group nucleophilically at-

tacks the now electron-poor C1 of the D ring to form a co-

valent glycosyl–enzyme intermediate (covalent catalysis).

4. Water replaces the E-ring product in the active site.

5. Hydrolysis of the covalent bond with the assistance

of Glu 35 (general base catalysis), which involves another

oxonium ion transition state, regenerates the active site

groups.The enzyme then releases the D-ring product, com-

pleting the catalytic cycle.

The double-displacement mechanism diagrammed in

Fig. 15-14 (in which the Asp 52 carboxylate group displaces

O1 from C1 of the D-ring and is, in turn,displaced by water)

allows the incoming water molecule to attach to the same

face of the D-ring as the E-ring it replaces. Consequently,

the configuration of the D-ring is retained. A single-dis-

placement reaction, in which water directly displaces O1,

would invert the configuration at C1 of the D-ring between

the substrate and product, a result that is not observed.

C. Testing the Catalytic Mechanism

The forgoing mechanism is the product of over 40 years of

enzymatic research. It differs in important ways from the

mechanism Phillips originally proposed based on his struc-

tural studies of lysozyme and a knowledge of the mecha-

nism of nonenzymatic acetal hydrolysis. In the remainder

of this section, we discuss the highlights of these enzymatic

studies to illustrate how scientific models evolve.

a. Confirming the Identities of Lysozyme’s

Catalytic Residues

The identities of lysozyme’s catalytically important

groups have been experimentally verified through site-

directed mutagenesis (Section 5-5Gc) and the use of

group-specific reagents:

Glu 35.The mutagenesis of Glu 35 to Gln yields a protein with

no detectable catalytic activity (0.1% of wild type),although

it has only a ⬃1.5-fold decrease in substrate affinity. Glu 35

must therefore be essential for lysozyme’s catalytic activity.

Asp 52. The mutagenesis of Asp 52 to Asn, which has a po-

larity comparable to that of Asp but lacks its negative

charge, yields an enzyme with no more than 5% of wild-

type lysozyme’s catalytic activity even though this mutation

causes an ⬃2-fold increase in the enzyme’s affinity for sub-

strate. Asp 52 is therefore important for enzymatic activity.

Noninvolvement of Other Amino Acid Residues. Lysozyme’s

other carboxyl groups besides Glu 35 and Asp 52 do not

participate in the catalytic process, as was demonstrated by

reacting lysozyme with carboxyl-specific reagents in the

presence of substrate. This treatment yields an almost fully

active enzyme in which all carboxyl groups but Glu 35 and

Asp 52 are derivatized. Other group-specific reagents that

modify, for instance, His, Lys, Met, or Tyr residues but in-

duce no major protein structure disruptions, cause little

change in lysozyme’s catalytic efficiency.

b. Role of Strain

Many of the mechanistic investigations of lysozyme

have had the elusive goal of establishing the catalytic role

of strain. Not all of these studies, as we shall see, supported

the Phillips mechanism, thereby stimulating a series of in-

vestigations that eventually settled this issue.

Measurements of the binding equilibria of various

oligosaccharides to lysozyme indicate that all saccharide

residues except that binding to the D subsite contribute ener-

getically toward the binding of substrate to lysozyme; bind-

ing NAM in the D subsite requires a free energy input of

12.1 kJ ⴢ mol

1

(Table 15-3). The Phillips mechanism ex-

plains this observation as being indicative of the energy

penalty of straining the D-ring from its preferred chair con-

formation toward the half-chair form.

As we discussed in Section 15-1Fa, an enzyme that cat-

alyzes a reaction by the preferential binding of its transi-

tion state has a greater binding affinity for an inhibitor that

has the transition state geometry (a transition state analog)

than it does for its substrate. The -lactone analog of

(NAG)

(Fig. 15-15) is presumably a transition state analog

of lysozyme since this compound’s lactone ring has the half-

chair conformation that geometrically resembles the pro-

posed oxonium ion transition state of the substrate’s D-ring.

X-ray studies indicate, in accordance with prediction, that

this inhibitor binds to lysozyme’s subsites

such that the lactone ring occupies the D subsite in a half-

chairlike conformation.

A¬B¬C¬D

522 Chapter 15. Enzymatic Catalysis

Table 15-3 Binding Free Energies of HEW

Lysozyme Subsites

Binding

Bound Free Energy

Site Saccharide (kJ ⴢ mol

1

)

ANAG 7.5

BNAM 12.3

CNAG 23.8

DNAM ⴙ12.1

ENAG 7.1

FNAM 7.1

Source: Chipman, D.M. and Sharon, N., Science 165, 459 (1969).

JWCL281_c15_506-556.qxd 2/19/10 9:28 PM Page 522

Despite the foregoing, the role of substrate distortion in

lysozyme catalysis had been questioned. Theoretical studies

by Michael Levitt and Arieh Warshel on substrate binding by

lysozyme suggested that the protein is too flexible to me-

chanically distort the D-ring of a bound substrate. Rather,

these calculations implied that transition state stabilization

occurs through the displacement by substrate of several

tightly bound water molecules from the D subsite.The result-

ing desolvation of the Asp 52 carboxylate group would signif-

icantly enhance its capacity to electrostatically stabilize the

transition state oxonium ion.This study therefore concluded

that “electrostatic strain” rather than steric strain is the more

important factor in stabilizing lysozyme’s transition state.

In an effort to obtain further experimental information

bearing on the Phillips strain mechanism, Nathan Sharon

and David Chipman determined the D subsite–binding

affinities of several saccharides by comparing the

lysozyme-binding affinities of various substrate analogs.

The NAG lactone inhibitor binds to the D subsite with 9.2

kJ ⴢ mol

1

greater affinity than does NAG. This quantity

corresponds, according to Eq. [14.15], to no more than an

⬃40-fold rate enhancement of the lysozyme reaction as a

result of strain (recall that the difference in binding energy

between a transition state analog and a substrate is indica-

tive of the enzyme’s rate enhancement arising from the

preferential binding of the transition state complex). Such

an enhancement is hardly a major portion of lysozyme’s

⬃10

-fold rate enhancement (accounting for only ⬃20% of

the reaction’s

; Section 14-1Cd). Moreover, an N-

acetylxylosamine (XylNAc) residue,

HOH

-Acetylxylosamine residue

NHCOCH

¢¢G

‡

cat

which lacks the sterically hindered ¬C6H

OH group of

NAM and NAG, has only marginally greater binding affin-

ity for the D subsite (3.8 kJ ⴢ mol

1

) than does NAG

(2.5 kJ ⴢ mol

1

). Yet recall that the Phillips mechanism

postulates that it is the unfavorable contacts made by this

¬C6H

OH group that promotes D-ring distortion. Never-

theless, lysozyme does not hydrolyze saccharides with

XylNAc in the D subsite.

The apparent inconsistencies among the foregoing exper-

imental observations were largely rationalized by Michael

James’ highly accurate (1.5-Å resolution) X-ray crystal

structure determination of lysozyme in complex with

NAM–NAG–NAM. This trisaccharide binds, as expected,

to the B, C, and D subsites of lysozyme. The NAM in the

D subsite, in agreement with the Phillips mechanism, is dis-

torted to the half-chair conformation with its ¬C6H

group in a nearly axial position due to steric clashes that

would otherwise occur with the acetamido group of the

C subsite NAG (although, contrary to the original Phillips

mechanism, Glu 35 and Trp 108 are too far away from the

¬C6H

OH group to contribute to this distortion). This

strained conformation is stabilized by a strong hydrogen

bond between the D-ring O6 and the backbone NH of Val

109 (Fig. 15-12; transition state stabilization). Indeed, the

mutation of Val 109 to Pro, which lacks the NH group to

make such a hydrogen bond, inactivates the enzyme.

Lysozyme’s lack of hydrolytic activity when XylNAc occ-

upies its D subsite is likewise explained by the absence of

this hydrogen bond and the consequent lesser stability of

the XylNAc ring’s half-chair transition state.

The unexpectedly small free energy differences in bind-

ing NAG, NAG lactone, and XylNAc to the D subsite are

explained by the observation that undistorted NAG and

XylNAc can be modeled into the D subsite as it occurs in

the X-ray structure of the lysozyme ⴢ NAM–NAG–NAM

complex. NAM’s bulky lactyl side chain prevents it from

binding to the D subsite in this manner.

c. Evidence for a Covalent Intermediate

The original Phillips mechanism proposed that the

lysozyme reaction proceeds via the direct attack of a water

molecule on C1 of the half-chair-distorted D-ring so as to

directly displace the E-ring (a single-displacement reac-

tion) and therefore did not involve the intermediate for-

mation of a covalent bond (much like the mechanism dia-

grammed in Fig. 15-14 but skipping Step 3). This was based

on Phillip’s observation that the distance between C1 of

the D-ring and a carboxyl O of Asp 52 (which participates

in a network of hydrogen bonds that appear to hold this

side chain in place) are too long to form a covalent bond

(minimally 2.3 Å in the NAM–NAG–NAM complex with-

out significantly disrupting the protein structure vs ⬃1.4 Å

for a C¬O single bond). The fact that the lysozyme reac-

tion proceeds with retention of configuration was attrib-

uted to the shielding of the reaction intermediate by the

enzyme, thereby preventing a water molecule from ap-

proaching the bond being cleaved from its back side as nor-

mally occurs in a single-displacement reaction. This aspect

of the original Phillips mechanism was widely accepted for

Section 15-2. Lysozyme 523

Figure 15-15 Transition state analog inhibition of lysozyme.

The -lactone analog of (NAG)

(left) resembles the transition

state of the lysozyme reaction (right). Note that atoms C1, C2,

C5, and O5 in each structure are coplanar (as indicated by

orange shading), consistent with the half-chair conformation

of the hexose ring.

NHCOCH

(NAG)

␦-Lactone analog

of (NAG)

NHCOCH

...

R =

COO

Lysozyme transition state

JWCL281_c15_506-556.qxd 6/5/10 9:10 AM Page 523

over 35 years because no enzyme–substrate covalent bond

had been detected in any of the numerous experimental

studies of hen egg white (HEW) lysozyme.

Despite the foregoing,all other -glycosidases of known

structure that cleave glycosidic linkages with net retention

of configuration at the anomeric carbon (as does HEW

lysozyme) have been shown to do so via a covalent glycosyl–

enzyme intermediate. The active sites of these so-called

retaining ␤-glycosidases structurally resemble that of

HEW lysozyme. Moreover, there is no direct evidence in-

dicative of the existence of a long-lived oxonium ion at the

active site of any retaining -glycosidase, including HEW

lysozyme (the lifetime of a glucosyl oxonium ion in water is

⬃10

12

s, a time only slightly larger than that of a bond vi-

bration).Consequently, there had been a growing suspicion

that the HEW lysozyme reaction also proceeds via a cova-

lent intermediate, one between the D-ring’s anomeric car-

bon (C1) and the side chain carboxyl group of Asp 52 to

form an ester linkage (as generated by Step 3 in Fig. 15-14).

This intermediate presumably reacts with H

O in what is

essentially the reverse of the reaction leading to its forma-

tion, thereby yielding the reaction’s second product (Steps

4 and 5 of Fig. 15-14). In this double-displacement mecha-

nism, the oxonium ion is proposed to be the transition state

on the way to forming the covalent intermediate, rather

than being an intermediate itself as it was in the original

Phillips mechanism.

If, in fact, HEW lysozyme follows the mechanism in Fig.

15-14, the reason that its covalent intermediate had not

been observed is that its rate of breakdown must be much

faster than its rate of formation. Hence, if this intermedi-

ate is to be experimentally observed, its rate of formation

must be made significantly greater than its rate of break-

down.To do so, Stephen Withers capitalized on three phe-

nomena. First, if, as postulated, the reaction goes through

an oxonium ion transition state, all steps involving its for-

mation should be slowed by the electron withdrawing ef-

fects of substituting F (the most electronegative element)

at C2 of the D-ring. Second, mutating Glu 35 to Gln

(E35Q) removes the general acid–base that catalyzes the

reaction, further slowing all steps involving the oxonium

ion transition state. Third, substituting an additional F

atom at C1 of the D-ring accelerates the formation of the

intermediate because this F is a good leaving group. Mak-

ing all three of these changes should increase the rate of

formation of the proposed covalent intermediate relative

to its breakdown and hence should result in its accumula-

tion. Withers therefore incubated E35Q HEW lysozyme

with NAG-(1 S 4)-2-deoxy-2-fluoro--

D-glucopyranosyl

fluoride (NAG2FGlcF):

Electrospray ionization mass spectrometry (ESI-MS; Sec-

H F

HHO

H NHCOCH

H F

NAG2FGlcF

O O

tion 7-1I) of this reaction mixture revealed a sharp peak at

14,683 D, consistent with the formation of the proposed co-

valent intermediate, but no significant peak at or near the

14,314-D molecular mass of the mutant enzyme alone.

The X-ray structure of this covalent complex unam-

biguously reveals the expected ⬃1.4-Å-long covalent

bond between C1 of the D-ring NAG and a side chain

carboxyl O of Asp 52 (Fig. 15-16). This D-ring NAG

adopts an undistorted chair conformation, thus indicat-

ing that it is a reaction intermediate rather than an ap-

proximation of the transition state. The superposition of

this covalent complex with that of the above described

complex of NAM–NAG–NAM with wild-type HEW

lysozyme reveals how this covalent bond forms (Fig. 15-16).

The shortening of the 3.2-Å distance between the D-ring

NAG C1 and the Asp 52 O in the NAM–NAG–NAM

524 Chapter 15. Enzymatic Catalysis

Figure 15-16 The HEW lysozyme covalent intermediate. The

substrate C- and D-rings and Asp 52 are shown in the superposition

of the X-ray structures of the covalent complex formed by

reacting E35Q lysozyme with NAG2FGlcF (C green, N blue, O

red, and F magenta) and the noncovalent complex of wild-type

lysozyme with NAM–NAG–NAM (C yellow, N blue, and O red).

Note that the covalent bond between Asp 52 and C1 of the

D-ring forms when the D-ring in the noncovalent complex relaxes

from its distorted half-chair conformation to an undistorted chair

conformation and that the side chain of Asp 52 undergoes an

⬃45° rotation about its bond. [Based on X-ray structures

by David Vocadlo and Stephen Withers, University of British

Columbia,Vancouver, Canada; and Michael James, University of

Alberta, Edmonton, Canada. PDBids 1H6M and 9LYZ.]



¬C



JWCL281_c15_506-556.qxd 2/19/10 9:28 PM Page 524

complex to ⬃1.4 Å in the covalent complex is almost en-

tirely a consequence of the relaxation of the D-ring from

the half-chair to the chair conformation combined with an

⬃45°rotation of the Asp 52 side chain about its C



¬C



bond; the positions of the D-ring O4 and O6 atoms are es-

sentially unchanged. Hence, over 35 years after its formu-

lation, it was shown that the Phillips mechanism must be al-

tered to take into account the transient formation of this

covalent glycosyl–enzyme ester intermediate (covalent

catalysis). Keep in mind, however, that in order to form

this covalent linkage, the D-ring must pass through an ox-

onium-like transition state, which requires it to transiently

assume the half-chair conformation.

3 SERINE PROTEASES

Our next example of enzymatic mechanisms is a diverse

group of proteolytic enzymes known as the serine pro-

teases (Table 15-4). These enzymes are so named because

they have a common catalytic mechanism characterized by

the possession of a peculiarly reactive Ser residue that is

essential for their enzymatic activity. The serine proteases

are the most thoroughly understood family of enzymes as a

result of their extensive examination over a nearly 60-year

period by kinetic, chemical, physical, and genetic tech-

niques. In this section, we mainly study the best character-

ized serine proteases, chymotrypsin, trypsin, and elastase.

We also consider how these three enzymes, which are syn-

thesized in inactive forms, are physiologically activated.

A. Kinetics and Catalytic Groups

Chymotrypsin, trypsin, and elastase are digestive enzymes

that are synthesized by the pancreatic acinar cells (Fig. 1-10c)

and secreted, via the pancreatic duct, into the duodenum

(the small intestine’s upper loop). All of these enzymes

catalyze the hydrolysis of peptide (amide) bonds but with

different specificities for the side chains flanking the scissile

(to be cleaved) peptide bond (recall that chymotrypsin is

specific for a bulky hydrophobic residue preceding the scis-

sile peptide bond, trypsin is specific for a positively charged

residue, and elastase is specific for a small neutral residue;

Table 7-2).Together, they form a potent digestive team.

a. Ester Hydrolysis as a Kinetic Model

That chymotrypsin can act as an esterase as well as a

protease is not particularly surprising because the chemical

mechanisms of ester and amide hydrolysis are almost iden-

tical. The study of chymotrypsin’s esterase activity has led

to important insights concerning this enzyme’s catalytic

mechanism. Kinetic measurements by Brian Hartley of the

chymotrypsin-catalyzed hydrolysis of p-nitrophenylacetate

indicated that the reaction occurs in two phases (Fig. 15-17):

1. The “burst phase,” in which the highly colored p-

nitrophenolate ion is rapidly formed in amounts stoichio-

metric with the quantity of active enzyme present.

2. The “steady-state phase,” in which p-nitrophenolate

is generated at a reduced but constant rate that is inde-

pendent of substrate concentration.

chymotrypsin

–

p-Nitrophenylacetate

p-NitrophenolateAcetate

Section 15-3. Serine Proteases 525

Table 15-4 A Selection of Serine Proteases

Enzyme Source Function

Trypsin Pancreas Digestion of proteins

Chymotrypsin Pancreas Digestion of proteins

Elastase Pancreas Digestion of proteins

Thrombin Vertebrate serum Blood clotting

Plasmin Vertebrate serum Dissolution of blood clots

Kallikrein Blood and tissues Control of blood flow

Complement C1 Serum Cell lysis in the immune response

Acrosomal protease Sperm acrosome Penetration of ovum

Lysosomal protease Animal cells Cell protein turnover

Cocoonase Moth larvae Dissolution of cocoon after metamorphosis

-Lytic protease Lysobacter enzymogenes Possibly digestion

Proteases A and B Streptomyces griseus Possibly digestion

Subtilisin Bacillus subtilis Possibly digestion

Source: Stroud, R.M., Sci.Am. 231(1), 86 (1974).

JWCL281_c15_506-556.qxd 2/19/10 9:28 PM Page 525

These observations have been interpreted in terms of a two-

stage reaction sequence in which the enzyme (1) rapidly re-

acts with the p-nitrophenylacetate to release p-nitrophenolate

ion forming a covalent acyl–enzyme intermediate that (2) is

slowly hydrolyzed to release acetate:

Chymotrypsin evidently follows a Ping Pong Bi Bi mecha-

nism (Section 14-5A). Chymotrypsin-catalyzed amide hy-

drolysis has been shown to follow a reaction pathway similar

to that of ester hydrolysis but with the first step of the reac-

tion, enzyme acylation, being rate determining rather than

the deacylation step.

C O

C Enzyme

Enzyme

–

p-Nitrophenylacetate

p-Nitrophenolate

Acetate

Enzyme

Chymotrypsin

Acyl–enzyme intermediate

fast

slow

b. Identification of the Catalytic Residues

Chymotrypsin’s catalytically important groups were

identified by chemical labeling studies.These are described

below.

Ser 195. A diagnostic test for the presence of the active Ser

of serine proteases is its reaction with diisopropylphospho-

fluoridate (DIPF):

which irreversibly inactivates the enzyme. Other Ser

residues, including those on the same protein, do not react

with DIPF. DIPF reacts only with Ser 195 of chymotrypsin,

thereby demonstrating that this residue is the enzyme’s ac-

tive Ser.

Diisopropylphospho-

fluoridate (DIPF)

DIP–Enzyme

F O

CH(CH

)

CH(CH

)

(Active Ser)

CH(CH

)

CH(CH

)

(Active Ser)

526 Chapter 15. Enzymatic Catalysis

Figure 15-17 Time course of p-nitrophenylacetate hydrolysis

as catalyzed by two different concentrations of chymotrypsin.

The enzyme rapidly binds substrate and releases the first

product, p-nitrophenolate ion, but the second product, acetate

ion, is released more slowly. Consequently, the rate of

p-nitrophenolate generation begins rapidly (burst phase) but

2 4 6 8 10 12

Time (min)

0.8 mg ⋅ mL

–1

0.4 mg ⋅ mL

–1

[p-Nitrophenolate] (mM)

Steady state

phase

Burst

phase

slows as acyl–enzyme complex accumulates until the rate of

p-nitrophenolate generation approaches that of acetate release

(steady state).The extrapolation of the steady state curve to zero

time (dashed lines) indicates the initial concentration of active

enzyme. [After Hartley, B.S. and Kilby, B.A., Biochem. J. 56, 294

(1954).]

JWCL281_c15_506-556.qxd 2/19/10 9:28 PM Page 526

The use of DIPF as an enzyme inactivating agent came

about through the discovery that organophosphorus com-

pounds such as DIPF are potent nerve poisons. The neu-

rotoxicity of DIPF arises from its ability to inactivate

acetylcholinesterase, a serine esterase that catalyzes the

hydrolysis of acetylcholine:

Acetylcholine is a neurotransmitter: It transmits nerve im-

pulses across the synapses (junctions) between certain

types of nerve cells (Sections 12-4Da and 20-5Cb). The in-

activation of acetylcholinesterase prevents the otherwise

rapid hydrolysis of the acetylcholine released by a nerve

impulse and thereby interferes with the regular sequence

of nerve impulses. DIPF is of such great toxicity to humans

that it has been used militarily as a nerve gas. Related com-

pounds, such as parathion and malathion,

are useful insecticides because they are far more toxic to

insects than to mammals.

His 57. A second catalytically important residue was dis-

covered through affinity labeling. In this technique, a sub-

strate analog bearing a reactive group specifically binds at

the enzyme’s active site, where it reacts to form a stable co-

valent bond with a nearby susceptible group (these reac-

tive substrate analogs have therefore been described as the

“Trojan horses” of biochemistry). The affinity labeled

groups can subsequently be identified by peptide mapping

(Section 7-1J). Chymotrypsin specifically binds tosyl-

L-

phenylalanine chloromethyl ketone (TPCK),

CH CCH

ClS

O O

Parathion

Malathion

(CH

)

NCH

–

acetylcholinesterase

C +

(CH

)

NCH

COH +

Acetylcholine

Choline

because of its resemblance to a Phe residue (one of chy-

motrypsin’s preferred residues;Table 7-2).Active site–bound

TPCK’s chloromethyl ketone group is a strong alkylating

agent; it reacts with His 57 (Fig. 15-18), thereby inactivat-

ing the enzyme. The TPCK reaction is inhibited by ␤-

phenylpropionate,

a competitive inhibitor of chymotrypsin that presumably

competes with TPCK for its enzymatic binding site. More-

over,the TPCK reaction does not occur in 8M urea, a dena-

turing reagent, or with DIP–chymotrypsin, in which the ac-

tive site is blocked.These observations establish that His 57

is an essential active site residue of chymotrypsin.

B. X-Ray Structures

Bovine chymotrypsin, bovine trypsin, and porcine elastase

are strikingly homologous: The primary structures of these

⬃240-residue monomeric enzymes are ⬃40% identical and

their internal sequences are even more alike (in compari-

son, the  and  chains of human hemoglobin have a 44%

sequence identity). Furthermore, all of these enzymes have

an active Ser and a catalytically essential His as well as simi-

lar kinetic mechanisms. It therefore came as no surprise

when their X-ray structures all proved to be closely related.

To most conveniently compare the structures of these

three digestive enzymes, they have been assigned the same

amino acid residue numbering scheme. Bovine chy-

motrypsin is synthesized as an inactive 245-residue precur-

sor named chymotrypsinogen that is proteolytically con-

verted to chymotrypsin (Section 15-3Ea). In what follows,

the numbering of the amino acid residues in chymotrypsin,

trypsin, and elastase will be that of the corresponding

residues in bovine chymotrypsinogen.

The X-ray structure of bovine chymotrypsin was eluci-

dated in 1967 by David Blow. This was followed by the

COO

–

␤-Phenylpropionate

Section 15-3. Serine Proteases 527

Chymotrypsin

HCl

His 57

TPCK

Figure 15-18 Reaction of TPCK with chymotrypsin to

alkylate His 57.

JWCL281_c15_506-556.qxd 2/19/10 9:28 PM Page 527

528 Chapter 15. Enzymatic Catalysis

Figure 15-19 X-ray structure of bovine trypsin in covalent

complex with its inhibitor leupeptin. The protein, viewed looking

into its upper  barrel, is represented by its transparent

molecular surface with its polypeptide chain in ribbon form

colored in rainbow order from its N-terminus (blue) to its

C-terminus (red), but with loop L1 (residues 185–188)

blue-green, loop L2 (residues 221–225) violet, and Tyr 172 and

Gly 216 purple. The side chains of the catalytic triad (Ser 195, His

57, and Asp 102) and Asp 189 are drawn in ball-and-stick form

colored according to atom type (C green, N blue, O red) with

hydrogen bonds represented by dashed black lines. Leupeptin

(acetyl-Leu-Leu-Arg in which the terminal carboxyl group is

replaced by ) is drawn in stick form (C cyan, N blue, O

red) with its Arg side chain occupying the enzyme’s specificity

pocket (magenta mesh). [Based on an X-ray structure by Daniel

Koshland, Jr., University of California at Berkeley. PDBid

2AGI.]

See Kinemage Exercise 10-1

¬CHO

JWCL281_c15_506-556.qxd 2/19/10 9:28 PM Page 528

determination of the structures of bovine trypsin (Fig.15-19)

by Robert Stroud and Richard Dickerson, and porcine

elastase by David Shotton and Herman Watson. Each of

these proteins is folded into two structurally similar do-

mains, each of which have extensive regions of antiparallel

␤ sheets arranged in a 6-stranded ␤ barrel, but contain lit-

tle helix. The catalytically essential His 57 and Ser 195 are

located in a cleft between the ␤ barrels, as is the invariant (in

all serine proteases) Asp 102, which is buried in a solvent-

inaccessible pocket. These three residues form a hydrogen

bonded constellation referred to as the catalytic triad (Figs.

15-19 and 15-20).

a. The Structural Basis of Substrate Specificity Can

Be Quite Complex

The X-ray structures of the above three enzymes

suggest the basis for their differing substrate specificities

(Fig. 15-21):

1. In chymotrypsin, the bulky aromatic side chain of the

preferred Phe, Trp, or Tyr residue (Table 7-2) that con-

tributes the carbonyl group of the scissile peptide fits

snugly into a slitlike hydrophobic pocket, the specificity

pocket, that is located near the catalytic groups (Fig. 15-19).

Section 15-3. Serine Proteases 529

Figure 15-20 The active site residues of trypsin. Residues are

drawn in ball-and-stick form with C green, N blue, and O red and

viewed similarly to Fig. 15-19. The catalytic triad consists of Ser

Figure 15-21 Specificity pockets of three serine proteases. The

side chains of key residues that determine the size and nature of

the specificity pocket are shown along with a representative

substrate for each enzyme. Chymotrypsin prefers to cleave

peptide bonds following large hydrophobic side chains; trypsin

prefers Lys or Arg; and elastase prefers Ala, Gly, or Val. [After a

drawing in Branden, C. and Tooze, J., Introduction to Protein

Structure (2nd ed.), Garland Publishing, p. 213 (1999).]

–

Phe

Lys

Ala

Gly

216

Gly

216

Thr 216

Gly

226

Gly

226

Val 226

Ser 189

Asp 189

Chymotrypsin

Trypsin

Elastase

Scissile bond

195, His 57, and Asp 102. [Based on an X-ray structure by Daniel

Koshland, Jr., University of California at Berkeley. PDBid

2AGI.]

JWCL281_c15_506-556.qxd 6/7/10 2:08 PM Page 529

2. In trypsin, the residue corresponding to chy-

motrypsin Ser 189, which lies at the back of the specificity

pocket, is the anionic residue Asp (Fig. 15-19).The cationic

side chains of trypsin’s preferred residues, Arg or Lys, can

therefore form ion pairs with this Asp residue. The rest of

chymotrypsin’s specificity pocket is preserved in trypsin so

that it can accommodate the bulky side chains of Arg and

Lys.

3. Elastase is so named because it rapidly hydrolyzes

the otherwise nearly indigestible Ala, Gly, and Val-rich

protein elastin (a connective tissue protein with rubber-

like elastic properties). Elastase’s specificity pocket is

largely occluded by the side chains of a Val and a Thr

residue that replace two Gly’s lining this pocket in both

chymotrypsin and trypsin. Consequently elastase, whose

specificity pocket is better described as a depression,

specifically cleaves peptide bonds after small neutral

residues, particularly Ala. In contrast, chymotrypsin and

trypsin hydrolyze such peptide bonds extremely slowly be-

cause these small substrates cannot be sufficiently immo-

bilized on the enzyme surface for efficient catalysis to

occur (Section 15-1E).

Thus, for example, trypsin catalyzes the hydrolysis of pep-

tidyl amide substrates with an Arg or Lys residue preceding

the scissile bond with an efficiency, as measured by k

cat

(Section 14-2Ba), that is 10

-fold greater than that for the

corresponding Phe-containing substrates. Conversely, chy-

motrypsin catalyzes the hydrolysis of substrates after Phe,

Trp, and Tyr residues 10

-fold more efficiently than after

the corresponding Lys-containing substrates.

Despite the foregoing, the mutagenic change in trypsin

of Asp 189 S Ser (D189S) by William Rutter did not

switch its specificity to that of chymotrypsin but instead

yielded a poor, nonspecific protease. Moreover, even re-

placing the other three residues in trypsin’s specificity

pocket that differ from those in chymotrypsin, with those

of chymotrypsin, fails to yield a significantly improved en-

zyme. However, trypsin is converted to a reasonably active

chymotrypsin-like enzyme when, in addition to the forego-

ing changes (collectively designated S1), both of its two

surface loops that connect the walls of the specificity

pocket (Fig. 15-19), L1 (residues 185–188) and L2 (residues

221–225), are replaced by those of chymotrypsin (termed

Tr S Ch[S1  L1  L2]). Although this mutant enzyme

still has a low substrate-binding affinity, K

, the additional

mutation Y172W in a third surface loop yields an enzyme

(Tr S Ch[S1  L1  L2  Y172W]) that has 15% of chy-

motrypsin’s catalytic efficiency. Curiously, these loops,

whose sequences are largely conserved in each enzyme,

are not structural components of either the specificity

pocket or the extended substrate binding site in chy-

motrypsin or in trypsin (Fig. 15-19).

Careful comparisons, by Charles Craik and Robert Flet-

terick, of the X-ray structures of chymotrypsin and trypsin

with those of the closely similar Tr S Ch[S1  L1  L2]

and Tr S Ch[S1  L1  L2  Y172W] in complex with a

Phe-containing chloromethyl ketone inhibitor reveal the

structural basis of substrate specificity in trypsin and chy-

motrypsin. Efficient catalysis in the serine proteases re-

quires that the enzyme’s active site be structurally intact

and that the substrate’s scissile bond be properly posi-

tioned relative to the catalytic triad and other components

of the active site (see below). The above mutagenic

changes do not affect the structure of the catalytic triad or

those portions of the active site that bind the substrate’s

leaving group (that segment on the C-terminal side of the

scissile bond). However, the main chain conformation of

the conserved Gly 216 (which forms two hydrogen bonds

to the backbone of the third residue before the substrate’s

scissile bond in an antiparallel  pleated sheet–like

arrangement) differs in trypsin and chymotrypsin and

adopts a chymotrypsin-like structure in both hybrid pro-

teins. Evidently, if Gly 216 adopts a trypsin-like conforma-

tion, the scissile bond in Phe-containing substrates is mis-

oriented for efficient catalysis. Thus, despite the fact that

Gly 216 is conserved in trypsin and chymotrypsin, the dif-

fering structures of loop L2 in the two enzymes maintain it

in distinct conformations.

Loop L1, which interacts with L2 in both trypsin and

chymotrypsin, is largely disordered in the X-ray structure

of Tr S Ch[S1  L1  L2]. Modeling a trypsin-like L1 into

Tr S Ch[S1  L1  L2] results in severe steric clashes

with the chymotrypsin-like L2. Thus, the requirement of a

chymotrypsin-like L1 for the efficient catalysis by

Tr S Ch[S1  L1  L2] appears to arise from the need to

permit L2 to adopt a chymotrypsin-like conformation.

The side chain of Tyr 172 extends toward the base of

the specificity pocket. The improvement in substrate

binding affinity of Tr S Ch[S1  L1  L2  Y172W]

over Tr S Ch[S1  L1  L2] arises from structural re-

arrangements in this region of the enzyme caused by the

increased bulk and different hydrogen bonding require-

ments of Trp versus Tyr. These changes appear to improve

both the structural stability of residues forming the speci-

ficity pocket and their specificity for chymotrypsin-like

substrates. These results therefore highlight an important

caveat for genetic engineers: Enzymes are so exquisitely tai-

lored to their functions that they often respond to mutagenic

tinkering in unexpected ways.

b. Evolutionary Relationships Among

Serine Proteases

We have seen that sequence and structural homologies

among proteins reveal their evolutionary relationships

(Sections 7-3 and 9-6). The great similarities among chy-

motrypsin, trypsin, and elastase indicate that these proteins

evolved through gene duplications of an ancestral serine

protease followed by the divergent evolution of the resulting

enzymes (Section 7-3C).

Several serine proteases from various sources provide

further insights into the evolutionary relationships among

the serine proteases. Streptomyces griseus protease A

(SGPA) is a bacterial serine protease of chymotryptic

specificity that exhibits extensive structural similarity, al-

though only ⬃20% sequence identity, with the pancreatic

530 Chapter 15. Enzymatic Catalysis

JWCL281_c15_506-556.qxd 2/19/10 9:28 PM Page 530