вход по аккаунту


Catalytic Activity of Enzymes Altered at Their Active Sites.

код для вставкиСкачать
Catalytic Activity of Enzymes Altered at Their Active Sites**
By Emil Thomas Kaiser*
Protein engineering has as its goals the design and construction of new peptides and proteins with novel binding and catalytic properties. In one approach to protein engineering,
new active sites have been introduced into naturally occurring proteins either by sitedirected mutagenesis or by chemical modification. Providing that important changes in the
tertiary structures d o not result from such alterations, at least a portion of the binding site
of the original protein should be available for the formation of complexes between the
altered enzyme and its substrates. Many examples of active-site mutations have been described, including the generation by us of a cysteine mutant of alkaline phosphatase. A
fundamental limitation of the site-directed mutagenesis methodology is that replacements
of residues are restricted to the twenty naturally occurring amino acids. The alternative,
chemical modification, is difficult to carry out for the specific replacement of one amino
acid by another. However, we have shown that through such modification coenzyme analogues can be introduced covalently into appropriate positions in proteins, allowing us to
produce semisynthetic enzymes with catalytic activities radically altered from those of their
precursor proteins. In another approach to protein engineering efforts have focused on the
construction of systems where, as a first approximation, folding can be neglected and the
preparation of secondary structural units is the target. Examples of the successful design of
biologically active peptides and proteins along such lines, taken from our own work, include molecules mimicking apolipoproteins, toxins, and many hormones. In recent studies
we have progressed to the stage where we are starting to combine the two general approaches to protein engineering we have described and are able to construct small enzymes
like ribonuclease T, and its structural analogues.
1. Introduction
An important goal of our research in the field of protein
engineering is the design of new enzymatic catalysts starting from the constituent amino acids. This is a daunting
challenge, particularly since there are, at the present time,
no reliable procedures for predicting tertiary o r folded
structures from primary amino acid sequences. Accordingly, a number of years ago when we began our protein engineering efforts we decided to proceed along two different
routes. I n one route, we elected to start with naturally occurring folded structures and to introduce new active sites
by either chemical
or site-directed mutagenesis.[41As long as major changes in the tertiary structure
d o not result from such alterations, this route should allow
us to utilize at least part of the binding site of the original
protein for the formation of complexes between the modified enzymes and substrates.
Our second approach to the construction of new proteins focuses on the design of structural regions, holding
active sites constant where they occur. In contrast to tertiary structures, it appears quite feasible to predict amino
Prof. E. T. Kaiser
Laboratory of Bioorganic Chemistry and Biochemistry
The Rockefeller University
1230 York Avenue, New York, NY 10021-6399 (USA)
For other relevant reviews see: D. L. Oxender, C. F. Fox (Eds.): Protein
Engineering, Alan R. Liss, New York 1987; A. R. Fersht, J.-P. Shi, A. J.
Wilkinson, D. M. Blow, P. Carter, M. M. Y. Waye, G. P. Winter, Angew.
Chem. 96 (1984) 455; Angew. Chem. Int. Ed. Engl. 23 (1984) 505; J. R.
Knowles, Science (Washington. D.C.) 236 (1987) 1252.
Angew Chem. In!. Ed. Engl. 27 (1988) 913-922
acid sequences which form well-defined secondary structures, particularly in cases where peptides or proteins bind
at biological interfaces. Thus, for surface-active peptides
and proteins ranging in their biological activity from apolipoproteins through chemotactic agents and peptide toxins to peptide hormones, we believe that amphiphilic secondary structures (that is, secondary structures which are
hydrophilic on one face and hydrophobic on the other) are
readily induced when binding occurs to amphiphilic surfaces such as those found at biological interface~.[~-’’
have developed design principles permitting the construction of model peptides and polypeptides which simulate
the biological and physical properties of the naturally occurring amphiphilic systems. For example, careful examination of a space-filling model of salmon calcitonin, a peptide hormone with high hypocalcemic activity, led us to
propose that the biologically active structure of this hormone consists of three
(Fig. 1A). First, there is the
“active site” comprising residues 1-7 which are contained
in a disulfide loop. The second structural region consists of
residues 8-22 which have the potential to form an amphiphilic a-helix (Fig. 1B). When one examines this helix, it is
striking that only one residue, the glutamic acid at position
15, violates the hydrophobic-hydrophilic segregation. The
final region consists of residues 23-32, a region which
serves as a kind of hydrophilic spacer. The proline amide
residue at the C-terminus is crucial for the biological activity of the hormone, and perhaps it should be considered to
be part of the active site of the hormone in addition to the
N-terminal region (residues 1-7). In our model building,
we have focused our attention on testing the importance of
0 VCH Verlagsgesellschaft mbH, 0-6940 Weinheim. 1988
0570-0833/88/0707-0913 $ 02.50/0
91 3
Fig. 1. A) Sequences of salmon calcitonin I (SCT-I)
and model calcitonins MCT-I1 and MCT-111.
MCT-I1 differs from MCT-I11 in having a leucine
residue at position 15 and has hypocalcemic activity roughly comparable to that of SCT-I, while
MCT-111 is somewhat more active. B) a-Helical
projections of the 8-22 sequences of SCT-I, MCT11, and MCT-111. The amino acid residues in the
shaded areas are hydrophobic and the unshaded
residues are hydrophilic. The segregation of the
faces of the helices into hydrophobic and hydrophilic regions can be clearly seen.
the amphiphilic secondary structure (residues 8-22)
through the construction of analogues. The N-terminal active-site region has been maintained constant, but the putative helical region has been replaced by sequences which
have an accentuated potential for forming amphiphilic ahelical structures but differ very significantly from the natural sequence in the region 8-22. The hydrophobic face of
salmon calcitonin’s helical region consists of a considerable number of leucine residues which are very good aliphatic a-helix formers. In designing our replacement helices, we have found that the most active analogue structures preserve, in large part, the sequence and distribution
of residues o n the hydrophobic face of the salmon calcitonin helix but have a radically different sequence with increased a-helical potential on the hydrophilic face (Fig.
1B). Our most recent analogue, MCT-111, has a hydrophilic face quite different from that of natural salmon calcitonin but preserves, except for a very minor change, the
residues present in the natural hormone’s hydrophobic
face (Fig. 1). This latest analogue shows about 2 % to 3
times the hypocalcemic activity of the naturally occurring
hormone.“’’ Studies in which we have designed and constructed models along similar lines have shown that for
many hormones, like B-endorphin, corticotropin-releasing
hormone, glucagon, growth hormone releasing factor, neuropeptide Y, and vasoactive intestinal peptide, arnphiphilic secondary structures play an important role in the
biological and physical properties of the hormone.[”]
In this article we are going to concentrate on the approach in which new active sites are introduced into existing tertiary structures. However, as will be seen later, we
are now proceeding from the design of secondary structural systems in molecules like calcitonin, where, to a first
approximation, tertiary structure has been neglected, to
cases in which tertiary structure is crucial. As we learn a n
increasing amount about the redesign of secondary structural regions in molecules where tertiary structure is all im914
portant, we are moving closer to the time when it will be
possible to put together what we have learned about building molecules in regions outside the active site with the
work which we have done on the active sites so that we can
consider the construction of whole enzymatic systems with
new structural features and catalytic functions.
2. Choice of Protein Templates for Introduction of
New Catalytic Functionalities
In choosing protein structures in which to introduce new
catalytic sites, we have felt that it was important to start
with systems where the tertiary structure was well defined
through X-ray crystallographic studies. Another important
consideration is that the proteins being altered be readily
available and easily purified. Since we wanted to make the
conceptual development of the design of active sites the
focal point of our work, we have tried to avoid working on
systems where technical difficulties in obtaining purified
and well-characterized modified enzymes will occur. The
four main proteins whose active sites we have altered are
E . coli alkaline phosphatase,”’] glyceraldehyde-3-phosphate dehydrogenase (rabbit skeletal
and B a d lus thermophil~s[’~]),
the case of alkaline phosphatase, the active site has been
altered by site-specific mutagenesis and for the other three
enzymes alterations have been made by chemical modification to produce what we have termed “semisynthetic”
enzymes. For each of the proteins where we have redesigned the active site through model building and the use
of computer graphics based on the X-ray coordinates, we
have examined whether the active-site modifications we
have proposed will permit the productive binding of potential substrates. We have found that such model building
has been enormously useful in helping us to predict what
sorts of substrates we might employ for the study of the
mutated enzymes.
Angew. Chem. Int. Ed. Engl 27 (1988) 913-922
3. Choice of Residues for Modification
When site-specific mutagenesis is used as the technique
for altering an enzyme’s active site, there is a great deal of
flexibility in picking a residue to be altered. Although
steric and electrostatic considerations are important, there
is no restriction o n which residues might be altered except
for problems which can be encountered in the expression
of the mutant gene. In altering enzyme active sites by
chemical modification, on the other hand, nucleophilic
groups in the enzyme are usually the most convenient targets since most of the reagents with which stoichiometric
modifications can be performed contain electrophilic centers. Thus, in our generation of semisynthetic enzymes by
the chemical modification of glyceraldehyde-3-phosphate
dehydrogenase, hemoglobin, and papain, we have utilized
uniquely reactive sulfhydryl groups in these proteins for
modification by coenzyme analogue^.['^-^^^ Of course, an
important factor in our choice of residues for modification
is the apparent accessibility of the modified active site to
the binding of appropriate substrates, as judged from
model building.
4. Approaches to the Introduction of New Active
Site Groups in Proteins
4.1. Chemical Methodology
Scheme 1
general, the mutations result in enzymes possessing low activity towards all but the more highly activated substrates,
just as had been observed with thiolsubtilisin. The catalytic
pathway for most acyl transfer reactions involves the formation and decomposition of tetrahedral intermediates
(except possibly in those cases where very good leaving
groups are involved) and requires a series of rapid proton
transfers. The replacement of a nucleophilic residue at the
active site of a proteolytic enzyme by a new nucleophile
may result in the retardation of some proton transfer steps
causing a bottleneck in the overall catalytic process. If this
hypothesis is correct, then it is possible that to obtain effective new active-site nucleophiles in acyl transfer enzymes there may have to be accompanying alterations of
the active-site environment through additional mutations.
The first example of the conversion of one active-site
nucleophilic residue to another was studied independently
by Bender1231
and K o ~ h l a n dand
[ ~ ~their
~ co-workers. They
found that the treatment of subtilisin with a-toluenesulfonyl fluoride, followed by displacement of the resulting sulfonate group using thiolacetate and subsequent hydrolysis
of the acetyl group, gave rise to thiolsubtilisin, a mutated
enzyme containing cysteine in place of serine at its active
site. Because thiolsubtilisin contains a histidine residue in
close proximity to the new cysteine, it seemed possible that
the enzyme would show proteolytic activity in much the
4.2. Site-Directed Mutagenesis
same way as the naturally occurring thiol enzyme papain
does. It was found, however, that thiolsubtilisin was an exThe possibility that alteration of active-site nucleophiles
tremely poor endopeptidase, although it retained the abilmay have less deleterious effects in enzymes catalyzing
ity to undergo acylation at the active-site thiol group by
other types of group transfer reactions where proton transactivated derivatives such as active esters and acyl imidafers may not be a requirement for effective catalysis has
zoles. Subsequent to the thiolsubtilisin work, we explored
led us to studies on E. coli alkaline p h o s p h a t a ~ e . ~ ’ Con~.~’’
the consequences of introducing new potential nucleoversion of the active-site serine to cysteine by site-directed
philic groups into enzyme active sites through modificamutagenesis in alkaline phosphatase results in a mutated
tion with highly reactive cyclic esters.125-281
We were able to
enzyme which shows high catalytic activity towards subshow, for example, that the phenolic hydroxyl group introstrates such as 4-nitrophenyl phosphate and 2,4-dinitroduced into papain by modification of the enzyme’s activephenyl phosphate. Unlike the wild-type enzyme, which
site sulfhydryl group (Cys-25) with the sultone 1 could act
shows for many substrates an invariant k,, with changes in
as an effective intramolecular nucleophile, resulting in the
the leaving-group tendency, the “thiol alkaline phosphadesulfonylation of the enzyme accompanied by re-formatase” exhibits a marked dependence of its k,,, values on
tion of the cyclic sulfonate ( k 2
of Scheme 1). Of course,
the leaving-group tendency. Furthermore, when experithis reaction did not involve generation of a new catalytic
ments were done using Tris as a trapping nucleophile, in
nucleophile, but it did show that a nucleophilic group difthe case of the wild-type enzyme the amount of trapping of
ferent from that originally in the enzyme active site could
the phosphoryl group by the nucleophile did not, in generbe introduced and could function effectively in the enzyme
al, depend upon the nature of the leaving group in the
phosphate monoester used. However, for the thiol enzyme
Active-site nucleophiles have been altered in a number
there was a strong dependence of the trapping reaction
of enzymes involved in acyl transfer r e a ~ t i o n s . [ ~ ,In
~ ~ - ~ ’ ~upon the nature of the leaving group. Attempts to titrate
Angew Chem Inr Ed Engl 27/1988) 913-922
the active-site thiol group in thiol alkaline phosphatase
with sulfhydryl reagents such as 5,5’-dithiobis(2-nitrobenzoic acid), DTNB, were unsuccessful. The lack of success
could conceivably be due either to steric problems in binding the DTNB in the active site or, alternatively, to the
binding of the thiol group by the active-site zinc ion in the
mutant enzyme. Only when the active-site metal ion had
been removed by treatment of the enzyme at low pH could
the sulfhydryl group be titrated. An attempt to detect the
formation of a phosphoryl enzyme in the action of the
thiol alkaline phosphatase by the use of stopped-flow
spectrophotometry employing 2,4-dinitrophenyl phosphate
under conditions where a “burst” could be observed in a
similar experiment with the wild-type enzyme[321revealed
no such burst for the mutant species.[331The interpretation
of these various results was ambiguous. In particular, one
possible explanation seemed to be that coordination of the
thiol group of thiol alkaline phosphatase to the active-site
zinc resulted in a mutant enzyme which reacted by a new
pathway, namely a carbonic anhydrase mechanism, involving the attack by either water o r hydroxide ion coordinated
to zinc on the phosphate monoester substrates (Scheme
2A). Alternatively, it seemed quite possible that no change
Michaeiis complex
Michaelis complex
+ H
phosphoryl group acceptor.[331Thus, it appears that the
mechanism of attack of the thiol alkaline phosphatase on
phosphate monoester substrates involves a normal pathway (Scheme 2B) with a change, however, in what is the
rate-limiting step. For the thiol enzyme under the conditions we have employed, it appears that rate-limiting phosphorylation occurs in contrast to the situation for the wildtype species where, depending upon the pH, dephosphorylation or release of product inorganic phosphate is ratecontrolling.
4.3. Semisynthetic Enzymes Obtained by Covalent
Modification of Active Sites Using Coenzyme Analogues
Scheme 2.
in mechanism had occurred but that only a change in the
rate-limiting step of the reaction to give rate-limiting phosphorylation of the mutant enzyme had happened (Scheme
2B). To differentiate between these possibilities we performed an experiment with 4-nitrophenyl phosphate chirally labeled with different isotopes of oxygen in the phosphoryl group (2). If the mechanism had changed to a zinchydroxide or a zinc-water type (Scheme 2A), then we
would expect that the reaction would involve a single displacement and, if we transferred the phosphoryl group to a
suitable acceptor, we would expect to see stereochemical
inversion by such a route. On the other hand, if the thiol
group was acting by the usual mechanistic pathway involving the transient formation of a phosphoryl enzyme
(Scheme 2B), the stereochemical outcome should be one of
retention, due to the expected inversion in the phosphoryl
enzyme-forming step and another inversion in the decomposition of the phosphoryl
We found that retention occurred when we followed the course of reaction
with the thiol enzyme using (S)-propane-1,2-diol as the
0 2 N G O - P ( 0
In view of the results on proteolytic enzymes showing
that the replacement of nucleophiles at the active site was
not a strategy which readily produced highly active mutated enzymes, a number of years ago we decided to focus
on the construction of modified enzymes which we termed
semisynthetic enzymes where effective catalysis should not
require a series of rapid proton transfer steps.[’,3.‘3-221 In
our approach, we produced semisynthetic enzymes by
chemical modification of a n amino acid residue of an appropriate protein template with a reactive coenzyme analogue. We postulated that, if properly designed, the hybrid
species could exhibit a combination of the general binding
properties of the original protein template and the catalytic
activity of the covalently bound coenzyme. Flavin cofactors have proved to be particularly good prosthetic groups
in our work, although many other cofactors remain to be
examined. Such isoalloxazine derivatives are readily synthesized, act as good model compounds for the biological
cofactors FAD and FMN and catalyze in solution many of
the same reactions promoted by naturally occurring flavoenzymes, although with lower rates and selectivities. To
date, we have successfully used a series of isoalloxazine
derivatives to modify three templates : papain, glyceraldehyde-3-phosphate dehydrogenase, and hemoglobin. These
systems, all of which can function as efficient flavin-dependent oxidoreductases, will be discussed in turn.
Angew. Chem. lnt. Ed. Engl. 27 (1988) 913-922
E,, . NRNH
4.3. I . Flavopapain
Among the most successful targets for our approach to
the preparation of semisynthetic enzymes has been the hydrolytic enzyme papain. On the basis of both X-ray crystallographic investigation^[^^,^^] and solution studies, it is
known that there is a n extended groove approximately
25 A long in the vicinity of the active-site residue Cys-25.
Furthermore, this residue is highly reactive to modification
with reagents containing electrophilic centers, and the
course of modification can be readily monitored by a variety of techniques including sulfhydryl titration and the loss
of the enzyme’s proteolytic activity. Using the X-ray structural data it is possible to show with appropriate models
that modification of the Cys-25 sulfhydryl group with various flavin coenzyme analogues should permit the extended binding groove near the Cys residue to remain accessible to potential substrates which might interact with
the flavin moiety. Our choice of the flavins as the coenzyme analogues for the modification experiments was rnotivated by the knowledge that these compounds are fairly
effective even in the absence of enzyme in catalyzing various oxidation-reduction reaction^.[^" Therefore, we felt
that it might be sufficient to bind a potential substrate in a
geometrically favorable fashion in the vicinity of the enzyme-bound flavin in order to achieve a rapid rate of reaction. In other words, unlike hydrolysis reactions where it
seems likely that amino acid functional groups from the
protein have to be placed correctly to enhance proton
transfer reactions, for the enzyme-bound-flavin reactions
to occur it might be sufficient to utilize the binding cavity
of the enzyme without requiring specific participation of
any enzyme-bound functional groups.“, 13,16-20,36.371 Because our earlier work on flavopapain has been reviewed
rather e ~ t e n s i v e l y , ~ only
’ , ~ ~ ]the more recent studies will be
described here.
The most effective flavopapain prepared, 3, was obtained by alkylation with 8-bromoacety1-10-methy1isoalloxazine 4. This modified enzyme reacts rapidly with a va-
Scheme 3. E,, = the oxidized form of flavopapain 3, €HI = the reduced
form of flavopapain 3, E,,.NRNH = the Michaelis complex, NRNH =
N-alkyl- 1,4-dihydronicotinamide,N RN = N-alkylnicot~namide.
riety of hydrophobic N-alkyl-l,4-dihydronicotinamides.
The k,,,/K, value measured for the oxidation of the best
of the substrates examined, N-hexyl-1,4-dihydronicotinamide, is in the vicinity of lo6 M - ’ s-’ in air-saturated buffer
at pH 7.5 and 25°C.[20.2’1Comparison of this rate pararneter to the second-order rate constant for the corresponding
oxidation reaction catalyzed by the simple model compound 8-acetyl-10-methylisoalloxazine
shows that there is
more than a 103-fold rate acceleration for the enzymatic
system. Using anaerobic conditions and the stopped-flow
technique, we have investigated the individual steps in the
oxidation of N-alkyl-l,4-dihydronicotinamidescatalyzed
by 8-acetylflavopapain 3. The kinetic results obtained are
consistent with Scheme 3.
Angew. Chem. lnt. Ed. Engl. 27 (1988) 913-922
For three dihydronicotinamides, N-hexyl-, N-benzyl-,
and N-propyl-l,4-dihydronicotinamide,very high values
of k2, the rate constant for the conversion of the Michaelis
complex to the reduced dihydroflavopapain species, were
obtained (370k70 s-I, 29k5.5 s-’, and 31 k 11 s-’, respectively). Indeed, for the N-hexyl compound the kz value
( ~ 4 0 s-I)
0 was sufficiently large that even using a very efficient stopped-flow instrument only an estimate of this
rate constant could be calculated.
Because oxygen is a relatively poor oxidant for the dihydro form of flavopapain 3, under the turnover conditions
used for the aerobic reactions of the dihydronicotinamides, the reaction with dihydroflavopapain was found to be
largely rate-limiting over the substrate concentration range
employed.[”] Therefore, to achieve turnover conditions
where the reduction step rather than the oxidation of the
dihydroflavopapain is rate-limiting, it is necessary to employ an acceptor that reacts more rapidly than oxygen. We
found that 3-(4,5-dimethylthiazol-2-yl)-2,4-diphenyltetrazolium bromide (MTT), was an especially effective acceptor for achieving high turnover rates in the action of
flavopapain 3. With MTT at about 400 PM, we obtained
k,,, values for the oxidation of N-benzyl- and N-propyl1,4-dihydronicotinamides(31.5k 1.0 s - ’ and 62.549.5 s - ’ ,
respectively) which corresponded reasonably well to the
rate constants k , (Scheme 3) obtained under anaerobic
conditions for the reduction of 8-acetylflavopapain 3 by
these substrates. This work demonstrates that a flavoenzyme exhibiting high catalytic efficiency both in terms of
k,,,/K, and of the value of the turnover number can be
constructed by utilizing the active and binding sites of the
hydrolytic enzyme papain. The naturally occurring coenzyme NADH is a relatively poor substrate for flavopapain
3 ; this is understandable since the structure of papain indicates that the binding region of the enzyme is quite hydrophobic and should, therefore, interact much better with
the N-alkyl-l,4-dihydronicotinamidesthan with NADH.
We have also shown that flavopapains can mediate the
oxidation of thiols to d i s ~ l f i d e s . ~Once
~ ] 8-acetylflavopapain 3 was found to be an effective catalyst for
such reactions. Under anaerobic conditions the k , / K , values (cf. Scheme 3) for the reduction of the flavopapain by
DL-dihydrolipoamide and DL-dihydrolipoic acid were 4400
and 3400 M - I s-’, respectively. These rate parameters
were 126 and 200 times larger than the second-order rate
constants for the corresponding reactions of DL-dihydrolipoamide and DL-dihydrolipoic acid with the model compound 8-acetyl-10-methylisoa11oxazine.
With the dye MTT
as a n electron acceptor we obtained k,,, and K , values for
the oxidation of dihydrolipoamide by flavopapain 3, under turnover conditions, which were in approximate agreement with the k2 and K , values measured under anaerobic
conditions, demonstrating that the rate-limiting step of the
catalytic cycle is substrate oxidation rather than oxidation
of dihydroflavin.[22J
9 I7
4.3.2. Studies with Flauo-Glyceraldehyde-3-phosphate
Dehydrogenase (Flauo-GAPDH)
Recently, we have worked extensively with a new enzyme template, GAPDH, a readily available tetrameric enzyme for which both the primary and tertiary structures are
The active site contains a binding site for
NAD'/NADH adjacent to Cys-149 and the catalytic region where glyceraldehyde-3-phosphate binds. Alkylation,
even by sterically demanding groups, of this sulfhydryl
group, which is essential for catalysis, does not prevent nicotinamide binding.[42,431Computer graphics modeling
(Evans and Sutherland PS-300) using the X-ray coordinates for the thermophilic enzyme from Bacillus stearotherrnophilus[441suggested that GAPDH might be a good template for the rational design of flavin-dependent oxidoreductases selective for NADH rather than for the hydrophobic N-alkyl- 1,4-dihydronicotinamides which are oxidized preferentially by flavopapain 3. At first, we employed GAPDH from rabbit muscle tissue"31 but most of
our more recent work has been done on the thermophilic
enzyme from Bacillus ~tearotherrnophilus~'~~
which is much
more stable than the muscle protein. Modification of apoGAPDH from the respective species with 7-bromoacetyl10-methylisoalloxazine 5 resulted in incorporation of only
one flavin molecule per subunit of these tetrameric enzymes.
y 3
5, x
6, X
= Br
Our HPLC and gel filtration data indicate that the initially formed 7-acetylflavo-GAPDH 6 is tetrameric. However, in a process that is apparently irreversible the bacterial semisynthetic enzyme subsequently undergoes dissociation (approximately 1 day at 4°C) into dimers. For the
thermophilic enzyme we have principally studied the dimer. The rabbit muscle flavoenzyme is far less robust, losing 40-60% of its activity within about 12 h.[l3]Therefore,
in the latter case, only the tetramer was studied. In examining the reactivity of the dihydronicotinamides with the
flavo-GAPDH species, we found that, as expected, NADH
is the optimal substrate for both proteins. The data obtained with the tetrameric rabbit muscle enzyme are readily fit by standard Michaelis-Menten kinetics. However,
pronounced negative cooperativity is observed with the
bacterial dimer. We cannot conclusively rule out trivial explanations for the apparent cooperativity (a population of
heterogeneous active sites, for instance) at this time. Still,
since half-of-the-sites reactivity is well documented for
GAPDH itself, cooperative catalysis by bacterial 6 may
not be an unreasonable possibility.
The second-order rate constant, kZ, for the oxidation of
NADH by 7-acetyl-10-methylisoalloxazine
at 25°C in airsaturated buffer is only 12.9 M - ' s-'. Under these conditions, the rate enhancement observed in the case of the
rabbit muscle enzyme as judged from the rate constant
k , , , / K , is 83-f0ld['~~
and for the bacterial flavoprotein it
is nearly 6000-fold (under circumstances where negative
cooperativity is minimal).['41 Clearly, the magnitude of
k , , , / K , for bacterial 7-acetylflavo-GAPDH 6 at low substrate concentrations is in the range of k,,,/K, values observed for naturally occurring flavoenzymes that oxidize
NADH. As we have seen for 8-acetylflavopapain 3, turnover of 7-acetylflavo-GAPDH 6 under aerobic conditions
is limited by the rate of reoxidation of the dihydroflavin
generated during each catalytic cycle. For this reason, the
value of k,,, measured depends explicitly on the oxygen
concentration and does not reflect the true efficiency of
the hydride transfer step. Therefore, we used rapid mixing
techniques under anaerobic conditions to determine the
rates of reduction of the flavin in the enzyme-substrate
complex. In contrast to the situation with rabbit muscle 7acetylflavo-GAPDH 6, which is too fragile for stoppedflow experiments, the bleaching of the enzyme-bound
flavin cofactor in the bacterial 6, which followed biphasic
kinetics in the presence of excess NADH, could be readily
observed by stopped-flow methodology. Approximately
half of the isoalloxazines are reduced in a very rapid step
and the remainder at a much slower rate, behavior typical
of negatively cooperative enzymes that exhibit half-of-thesites
Saturation kinetics are observed for the
rate data in the fast phase step of the reduction of bacterial
6. A rate constant of k , = 1.14 s - ' for the reduction of the
rapidly reacting subunit proceeding from the corresponding Michaelis complex and an apparent dissociation constant for the Michaelis complex of 12.4 p~ were obtained
from kinetic measurements on the fast phase. Thus, the apparent second-order rate constant for the reduction of the
rapidly reacting subunit of the enzyme is 91 900 M - ' s-', a
value only slightly larger than the steady-state value of
k , , , / K , (75600 M - ' s-I) determined for NADH with this
enzyme and comparing very favorably with k,,,/K, obtained for naturally occurring flavoenzymes that oxidize
NADH. If a suitable electron acceptor reacting much more
rapidly than oxygen to reoxidize dihydroflavo-GAPDH
could be found, the turnover number for this bacterial flavoenzyme could approach 1 s - ' , a value comparable to
numbers seen for several natural systems. The work that
we have done to date amply demonstrates that GAPDH is
an excellent choice as a template for the construction of
artificial enzymes.
4.3.3. Studies on Flavohemoglobin
Recently, we have generated a new type of semisynthetic
enzyme, flavohemoglobin, in which the heme group is
maintained in the protein but a flavin molecule is added by
covalent m~dification."~'Heme proteins like cytochrome
P-450, which carry out redox reactions, function typically
by means of single electron transfers in each
they operate with the usual biological electron donors like
NADPH, which act by two electron transfers, there is a
requirement for an electron-transport system. In a typical
electron-transport system consisting of
NADPH, 02,and NADPH-cytochrome P-450 reductase, a
major flavoprotein in the microsomal electron-transport
system,[471hemoglobin has been found to catalyze a wide
variety of monooxygenase r e a ~ t i o n s . ~ We
~ * -sought
~ ~ ~ to deAngew. Chem. Int. Ed. Engl. 27 (1988) 913-922
termine whether hemoglobin could be modified in a fashion such that it might react directly with two electron donors without the intervention of the full electron transport
system. Accordingly, we attached a n isoalloxazine residue
covalently to hemoglobin in the vicinity of the heme with
the hope that the flavin moiety could mediate electron
transfer, obviating the need for the P-450 reductase (EC We have found that the flavohemoglobin species
did indeed serve as a hydroxylase for aniline in the absence of the reductase protein!”’ The chemical modification was carried out by reaction of a derivative of 7-cyanoisoalloxazine 7,containing a reactive disulfide group attached through a tether to the N-3 position, with the thiol
group of cysteine p-93 in carbon monoxide-hemoglobin
(COHb2@). Subsequently, ferric flavohemoglobin (FlHb’@) 8 was prepared by the oxidation of the covalently
modified species Fl-COHb2@with potassium ferricyanide.
Kinetic measurements on the aniline hydroxylase activity of Fl-Hb3@at p H 7.5 showed that the apparent K , values for aniline observed were essentially the same as those
for ferric hemoglobin, Hb3@,assayed in the presence and
absence of NADPH-cytochrome P-450reductase. The apparent k,,, for Hb3@was increased considerably when we
reconstituted that system with P-450 r e d ~ c t a s e . [ ~ Still,
the apparent k,,, for Hb3@in the reconstituted system was
smaller than that seen for F1-Hb3@by itself. Thus, it is
clear that the flavin covalently bound to hemoglobin in F1Hb3@ is able to substitute effectively for the reductase.
Electron transfer between the neighboring prosthetic
groups in FI-Hb3@proceeds even more efficiently than
that in the combined Hb3@-reductasesystem.
Since a free model flavin does not catalyze the p-hydroxylation of aniline in the absence of hemoglobin, it is clear
that the heme group is directly involved in the reaction of
flavohemoglobin with aniline. The predominance of p-hydroxylation over o-hydroxylation @ / o > 4/1) suggests also
that the active oxygen species is likely to be bound to heme
and is probably not freely diffusible as in the case of aniline hydroxylation mediated by Fenton’s reagent where oaminophenol is the major product. It thus may be that flavohemoglobin acts in a fashion analogous to the cytochrome P-450 system (microsomal mixed-function oxidase). However, there are ambiguities in the detailed
mechanism for oxygen activation by flavohemoglobin that
need to be resolved, and this is a problem to which we are
currently addressing ourselves.
5. Synthetic Applications of Enzymes Altered at
Their Active Sites
Recently, we initiated studies o n the first practical application of a semisynthetic enzyme to synthesis.[531In work
which will be described briefly here, we have been engaged
Angew. Chern. Inr. Ed. Engl. 27(1988) 913-922
in the total synthesis of small proteins such as ribonuclease
T, and its structural analogues. In principle, we could have
approached the synthesis of this enzyme through genetic
engineering utilizing either the naturally occurring gene or
a synthetic gene,[54]and we could consider making the various structural mutants that we would like to design
through either site-directed mutagenesis on the natural
gene o r through alterations in the synthesis of the gene itself. We have chosen, however, to carry out the preparation of the enzyme by a peptide segment synthesis-condensation approach.[551There are several reasons why we
have done this rather than using the genetic methodology.
First, of course, we can avoid the problem of expression
when we make mutant structures. It is already known that
in making some mutants of ribonuclease T, the expression
has been rather poor because the mutants have had a deleterious effect o n the bacteria in which they are being prod ~ c e d . [Another
point where the synthetic methodology
has an advantage is that we can introduce labeled or unnatural amino acids at specific locations where we wish to
incorporate them. This ability to introduce labels specifically has great advantages in terms of possible use in fluorescence and NMR studies. Further, we can not only introduce unnatural amino acids at various points in the protein structure by synthesis, but also we have the possibility
even of introducing nonpeptidic linkages. This means that
we may be able to engineer modified proteins containing
structural regions where the use of nonpeptidic segments
might prevent degradation by proteolysis, which might
normally occur in a biological milieu. A case in point
where we have already achieved such a replacement is in
our work on models for the peptide hormone p-endorphin1561where we have replaced the part of the sequence
from residue 6 to 13, which we believe to function as a
hydrophilic Iinker region, with a repeating unit of four yhydroxymethyl-y-aminobutyric acid residues hooked together through amide linkages. While this redesigned region still contains amide bonds, the linkages are not the
usual peptide linkages, and yet we have found that the Bendorphin model so generated is still very effective, for example, as a n analgesic agent.
In carrying out the segment synthesis-condensation
methodology, for the preparation of small proteins, the
first problem that we had to solve was to develop an efficient synthesis of peptide segments with the protecting
groups intact so that the segments could be readily coupled further after purification. The method that we have
used for the peptide segment synthesis is based on our development of an oxime p ~ l y m e r ~ ~ and
’ - ~ ~is ] illustrated in
Scheme 4. Briefly, the generation of the polymer involves
treatment of a polystyrene-1% divinylbenzene copolymer
with p-nitrobenzoyl chloride. The resultant ketone species
is treated with hydroxylamine to generate the corresponding oxime. Subsequently, the first amino acid is attached to
the resin using dicyclohexylcarbodiimide as the coupling
agent and employing protection such as the Boc group on
the a-amino function and appropriate groups o n potentially reactive side chains. Once the first amino acid has
been attached, a series of deprotection and coupling reactions is performed leading to the formation of a protected
peptide derivative attached to the polymer through the ox919
Scheme 4. The three downward pointing arrows
represent repetitive deprotection and coupling.
X =protecting group.
@ -d
b) Zn, HOAc
ime ester linkage. Because the linkage is more labile than
the usual linkages such as the benzyl group used in the
stepwise Merrifield synthesis, peptides substantially longer
than ten amino acid residues are not usually prepared on
the oxime resin since losses from the resin start to mount
with increases in the length of the peptide. A great advantage of the oxime polymer is, however, that the growing
peptide chain can be removed in a variety of ways which
can afford a protected peptide segment containing either a
free C-terminal carboxyl group o r a free C-terminal ester
function. An exhaustive list of the methods for removing
the peptide from the polymer will not be given here. However, it should suffice to indicate that the peptide chain
can be removed by treatment with N-hydroxypiperidine,
an a-nucleophile, resulting in the formation of the protected peptide ester of N-hydroxypiperidine. This peptide
ester can be then converted into the protected peptide with
a free carboxyl group by the use of zinc and acetic acid, a
treatment which is quite general, except for peptides containing sulfur functions, as in the case of methionine. Alternatively, the protected peptide can be removed from the
resin by the use of the tetra-n-butylammonium salt@'] of
the amino acid o r by treatment of the resin-bound peptide
with an amino acid ester having a free amino group, several equivalents of acetic acid being employed to catalyze
the reaction. These procedures result in the insertion at the
C-terminus of either the amino acid o r the amino acid ester
used for the displacement reaction.
In our synthesis of ribonuclease T,'"' we have used a
strategy in which the molecule has been divided into three
major fragments chosen both for their synthetic accessibility and because they contain structural features of the enzyme on which we want to focus. For example, from the
known X-ray structure of the enzyme the segment corresponding to residues 1-34 can be seen to contain a helical
region from residue 12 to 29 which has amphiphilic char920
acter.I6'' One of the objectives of our current work is to
synthesize a helix replacement preserving the amphiphilic
characteristics of the natural system but utilizing a significantly different sequence of amino acids. The objective of
this work is to establish whether or not incorporation of
such a redesigned helix into the ribonuclease T, structure
will still allow the enzyme to fold properly and function
effectively in catalysis.
We have utilized segments prepared by the oxime polymer method to assemble each of the three major segments of ribonuclease TI. This assembly has been achieved
by taking the smaller segments, which have been purified
and fully characterized by standard chemical methodology
including NMR, mass spectroscopy, and amino acid analysis, and coupling the segments sequentially on a polymeric support. The coupling has typically been carried out
using dicyclohexylcarbodiimide and N-hydroxybenzotriazole to suppress racemization. We have tried as much as
possible to avoid segment couplings where racemization is
likely to be an important problem, but the possibility of
racemization is one which cannot be totally avoided by our
procedures. Of course, after we have assembled the segments we have taken the major segments off the resin and
purified them extensively. The final coupling of the three
major segments (Scheme 5 ) has been carried out in solution.
While the synthesis of small proteins such as ribonuclease T, is now accessible through fragment synthesiscoupling, it is clear that this approach suffers by comparison with D N A synthesis in an important way. While our
syntheses of protected peptide segments by the oxime ester
approach are quite convenient, there are still many difficulties in the subsequent condensation of the segments.
Yields are not always high, one must be on guard against
the possibility of racemization, and sometimes, because
protected peptides are used, solubility problems can be enAngew. Chem. Inr. E L . Engl. 27 11988) 913-922
Z- 1 -34-OH
is the reaction of an active ester with an amino component,
and once the peptide bond is formed, we d o not have to
concern ourselves with rapid hydrolysis because thiolsubtilisin is a damaged endopeptidase. We have carried out
numerous peptide syntheses using this approach, and the
syntheses of [Le~]~-enkephalin
amide (Scheme 6A) and the
segment of ribonuclease T, corresponding to residues 1223 (Scheme 6B) by the enzyme-catalyzed procedure have
been accomplished.
A) Z-Tyr-Gly,-Phe-OR
Scheme 5. Synthesis
or rihonuclease T ,
+ Leu-NH,
67% yield of pure product
B) F,,,-Ser3-Asp-Val-Ser-Thr- Ala-OR + Gln-Ala,-Gly
countered. A major difference between the synthesis of
small proteins by the methodology outlined here and the
47% yield of pure p r o d u c t
R = C6H4CI-p
synthesis of genes by D N A methodology is that in the case
of the D N A synthesis ligases are available which can conScheme 6. A) Synthesis of (Leu)'-enkephalin amide. Conditions: pH 8.0,
nect D N A segments. It is true that, in many instances, pep0.1 M phosphate huffer/DMF (S0/50, v/v), room temperature, 5 h. B) Syntides have been coupled utilizing enzymatic c a t a l y s i ~ . ~ ' ~ - ' ~ ~ thesis of segment 12-23 of ribonuclease T,. Conditions: p H 7.0,0.01 M phosphate buffer/DMF (50/50, v/v) room temperature, 30 min. Thiolsuhtilisin
However, these reactions, which generally involve the use
was used as catalyst in A and B.
of endopeptidases or exopeptidases, necessitate finding
conditions where the equilibrium between peptide-bond
synthesis and peptide-bond hydrolysis favors the former
Of course, thiolsubtilisin will be useful in coupling pepprocess. Sometimes, reactions can be driven by the detide segments which have specificities appropriate to the
creased solubility of the product obtained from coupling.
binding pocket of the enzyme. To broaden the scope of the
In other cases, solvent conditions, including solvent comenzyme-catalyzed ligation, it will be necessary to utilize
position and pH, can be varied to favor peptide-bond synother
enzymes with different specificities in a related fashthesis. The difficulty is that each case of a peptide-bond
although the cysteine mutants of other
synthesis by the use of a proteolytic enzyme involves deterserine
appropriate targets for further work
mining the optimal conditions for the coupling process. In
possibility that other types of muother words, peptide-bond synthesis using enzymes usually
reactivity with active esters
requires finding a new set of conditions for each reaction
must be considered.
which is run. Ideally, one would be able to find peptide
can be utilized
ligases where a general set of conditions could be develwhere
to pepoped for coupling peptide segments. One can ask the questide
tion, at this point, whether o r not it is realistic to consider
of peptide bonds.
that such ligases can be found or developed. In work with
the semisynthetic enzyme, thiolsubtilisin, we have obtained
data which suggest that we have found a beginning to the
6. Future Prospects
solution of the problem of developing peptide ligases. As
mentioned near the beginning of this article, when BenAt the beginning of this article we saw that the properder'231 and K o s h l ~ n d [ *described
thiolsubtilisin they
ties of the first semisynthetic enzyme, thiolsubtilisin, did
not appear to augur well for the development of the field.
found that the endopeptidase activity of the enzyme was
However, the low activity of this mutated enzyme may repoor but that the nucleophilic SH group at the active site
still did accept acyl groups from activated derivatives such
flect the importance of achieving rapid proton transfer in
as active esters and acyl imidazoles. We have recognized
several steps involved in the overall catalysis process, steps
that just these properties are suitable for the use of a n enwhich could be affected by the difference in the acidity of
zyme as a peptide ligase. Specifically, thiolsubtilisin can be
the original nucleophile, the hydroxyl group in the native
acylated at its active site thiol group by an N-protected
enzyme, from that of the sulfhydryl group in the mutated
peptide segment containing a C-terminal active ester
species. Through attachment of flavin analogues to suitagroup, the p-chlorophenyl ester function. Since the acyl
ble locations in several proteins, we have shown that we
derivative formed is a thiol ester, it is very susceptible to
can obtain highly efficient semisynthetic enzymes which
aminolysis by appropriate amino components. Thus, thiolcan perform reactions such as oxidation-reduction and hysubtilisin should be able to catalyze an amide-bond-formdroxylation. In these reactions it seems unlikely that the
ing reaction between an N-protected peptide active ester
precise positioning of enzyme-bound functional groups,
and a free N-terminal amino group of another peptide segwhich would be involved in proton transfer reactions, is
ment. The driving force for the peptide-bond-forming step
required. Rather, the effectiveness of our semisynthetic enAngenz. Chem. Inr. Ed. Engl. 27 (1988) 913-922
92 1
zymes is likely to be due to the ability of these species to
bind substrates in close proximity to the reactive coenzyme
functions. In future studies it will be important to increase
the range of the coenzymatic species utilized. Also, the catalytic versatilities of semisynthetic enzymes such as flavohemoglobin have not been explored fully, and this is an
important pursuit of our laboratory at the present time. To
increase the flexibility of the chemical modification approach, we contemplate in the future to combine it with
site-specific mutagenesis. In particular, we are initiating
research in which we will place, by site-directed mutagenesis, appropriate “handles” in locations where chemical
modification with the coenzyme analogue is likely to result
in a species with desirable binding properties for substrates. The semisynthetic enzyme approach will be far
more powerful if we will rely not only on the presence of
suitable sites for chemical modification which are available in the natural protein systems but also if we can utilize
sites which we can introduce by genetic engineering. In
closing, it should be noted that very interesting activities
have been obtained through the generation of catalytic ant i b o d i e ~ [ ’ ~and
~ ~ *that
~ a promising route for the application of the semisynthetic concept will be to apply reactions
to the modification of the antibodies similar to those
which we have employed with the semisynthetic enzymes.
The experimental efforts and conceptual contributions oj
my co-workers whose names are given in the various articles
cited here are gratefully acknowledged. This research was
supported in part by the National Science Foundation (CHE8 41 8 8 78).
Received: December 21, 1987 [A 677 IE]
German version: Angew. Chem. 100 (1988) 945
[I] E. T. Kaiser, D. S. Lawrence, Science (Washington, D.C.) 226 (1984)
[2] E. T. Kaiser, Ann. N. Y. Acad. Sci. (Enzyme Eng. 8) 501 (1987) 14.
[3] D. Hilvert, E. T. Kaiser, Biotechnol. Genet. Eng. Rev. 5 (1987) 297.
141 E. T. Kaiser, Nature (London) 313 (1985) 630.
[S] E. T. Kaiser, F. J. Kezdy, Proc. Natl. Acad. Sci. USA 80 (1983) 1137.
161 E. T. Kaiser, F. J. Kezdy, Science (Washington. D.C.) 223 (1964) 249.
171 E. T. Kaiser, Trends Biochem. Sci. (Pers. Ed.) 12 (1987) 305.
[S] G. R. Moe, R. J. Miller, E. T. Kaiser, J. Am. Chem. SOC. I05 (1983)
4 100.
191 G. R. Moe, E. T. Kaiser, Biochemistry 24 (1985) 1971.
[lo] F. R. Green 111, B. Lynch, E. T. Kaiser, Proc. Natl. Acad. Sci. USA 84
(1987) 8340.
[I 11 J. W. Taylor, E. T. Kaiser, Pharmacol. Rev. 38 (1986) 291.
[I21 S . S. Ghosh, S . C. Bock, S . E. Rokita, E. T. Kaiser, Science(Washmgton.
D.C.) 231 (1986) 145.
1131 D. Hilvert, E. T. Kaiser, J. Am. Chem. SOC.107(1985) 5805.
[I41 D. Hilvert, Y. Hatanaka, E. T. Kaiser, J. Am. Chem. Soc. I10 (1988)
[IS] T. Kokubo, S. Sassa, E. T. Kaiser, J. Am. Chem. SOC.109 (1987) 606.
[I61 H. L. Levine, Y . Nakagawa, E. T. Kaiser, Biochem. Biophys. Res. Commun. 76 (1977) 64.
[I71 H. L. Levine, E. T. Kaiser, J. Am. Chem. SOC.100 (1978) 7670.
[IS] E. T. Kaiser, H. L. Levine, T. Otsuki, H. E. Fried, R. M. Dupeyre, Adu.
Chem. Ser. 191 (1980) 35.
1191 J. T. Slama, S. R. Oruganti, E. T. Kaiser, J. A m . Chem. SOC. 103 (1981)
I201 J. T. Slama, C. Radziejewski, S. R. Oruganti, E. T. Kaiser, J. Am. Chem.
Soc. 106 (1984) 6778.
[21] C . Radziejewski, D. P. Ballou, E. T. Kaiser, J. A m . Chem. Soc. 107
(1985) 3352.
[22] K. D. Stewart, C. Radziejewski, E. T. Kaiser, J. Am. Chem. SOC. 108
(1986) 3480.
[23] L. Polgar, M. L. Bender, J. Am Chem. SOC.88 (1966) 3153.
[24] K. E. Neet, D. E. Koshland, Proc. Natl. Acad. Sci. USA 56 (1966)
[25] E. T. Kaiser, Ace. Chem. Res. 3 (1970) 145.
1261 J. H. Heidema, E. T. Kaiser, J. Am. Chem. SOC.92 (1970) 6050.
[271 P. Campbell, E. T. Kaiser, Bioorg. Chem. 1 (1971) 432.
[28] P. Campbell, E. T. Kaiser, J. Am. Chem. SOC.95 (1973) 3735. The products in the last step of Scheme 1 are those which would result from the
normal first step in thiol sulfonate hydrolysis. Although the existence of
the enzyme sulfenic acid is speculative, there is some evidence that sulfenic acids may be stable at the active sites of enzymes. See: A. N. Glazer, Annu. Reu. Biochem. 39 (1970) 101.
[29] 1. S . Sigal, B. G. Harwood, R. Arentzen, Proc. Natl. Acad. Sci. USA 79
(1982) 7157.
[30] G. Dalbodie-McFarland, L. W. Cohen, A. D. Riggs, C. Morin, K. Itakura, J. H. Richards, Proc. Natl. Acad. Sci. USA 79 (1982) 6409.
[311 D. A. Kendall, E. T. Kaiser in A. Torriani-Gorini, F. G. Rothman, S .
Silver, A. Wright, E. Yagil (Eds.): Phosphate Metabolism and Cellular
Regulation in Microorganisms, American Society for Microbiology,
Washington, D.C. 1987, pp. 115-117.
I321 S. H. D. KO, F. J. Kezdy, J. Am. Chem. SOC.89 (1967) 7139.
[33] J. Butler-Ransohoff, personal communication.
[34] S. R. Jones, L. A. Kindman, J. R. Knowles, Nature (London) 257 (1978)
1351 J. R. Knowles, Annu. Rev. Biochem. 49 (1980) 877.
I361 J. Drenth, J. N. Jansonius, R. Koekoek, B. G. Wolthers, Adu. Protein
Chem. 25 (1971) 79.
[37] J. Drenth, K. H. Kalk, H. M. Swen, Biochemistry 15 (1976) 3731.
[38] T. C. Bruice, Prog. Bioorg. Chem. 4 (1976) I.
[39] H. L. Levine, E. T. Kaiser, J. A m . Chem. Soc. 102 (1980) 343.
1401 H. E. Fried, E. T. Kaiser, J. A m . Chem. Soc. 103 (1981) 182.
[41] J. L. Harris, M. Waters in P. Boyer (Ed.): The Enzymes, Yo/. 13. Academic Press, New York 1976, pp. 1-49.
[42] G. N. Rafter, S . P. Colowick, Arch. Biochem. Biophys. 66 (1957) 190.
1431 L. V. Benitez, W. S . Allison, Arch. Biochem. Biophys. 159 (1973) 89.
[44] G. Biesecher, J. L. Harris, J. C. Thierry, J. E. Walker, A. J. Wonacott,
Nature (London) 244 (1977) 328.
[45] A. Levitzski, J. Mol. Biol. 90 (1974) 451.
1461 R. Sato, T. Omura (Eds.): Cytochrome P-450. Kodansha/Academic
Press, Tokyo/New York 1978.
[47] R. W. Estabrook, Methods Enzymol. 52 (1978) 43.
[48] J. J. Mieyal, R. S. Ackerman, J. L. Blumer, L. S. Freeman, J. Biol. Chem.
25 (1976) 3436.
[49] 1. Golly, P. Hlavica, Blochim. Biophys. Acta 760 (1983) 69.
[SO] 0. Takikawa, R. Yoshida, 0. Hayaishi, J. Biol. Chem. 258 (1983) 6808.
[Sl] B. L. Ferraiolo, G. M. Onady, J. J. Mieyal, Biochemistry 23 (1984)
[52] D. W. Starke, K. S. Blisard, J. J. Mieyal, Mol. Pharmacol. 25 (1984)
1531 T. Nakatsuka, T. Sasaki, E. T. Kaiser, J . Am. Chem. SOC.109 (1987)
[54] M. Ikehara, E. Ohtsuka, T. Tokunaga, S. Nishikawa, S. Uesugi, T. Tanaka, Y. Aoyama, S. Kilyodani, K. Fujimoto, K. Yanase, K. Fuchimura, H.
Morioka, Proc. Natl. Acad. Sci. USA 83 (1986) 4695.
I551 T. Sasaki, M. Findeis, personal communication.
[56] B. Rajashekhar, E. T. Kaiser, J . Biol. Chem. 241 (1986) 13617.
[57] W. F. DeGrado, E. T. Kaiser, J. Org. Chem. 45 (1980) 1295.
1581 W. F. DeGrado, E. T. Kaiser, J. Org. Chem. 47 (1982) 3258.
[59] S. H. Nakagawa, H. S. H. Lau, F. J. Kezdy, E. T. Kaiser, J. Am. Chem.
SOC.107 (1985) 7087.
[60] P. Lansbury, personal communication.
[61] U. Heinemann, W. Saenger, Nature (London) 299 (1982) 27.
[62] M. Bergmann, J. S. Fruton, Adu. Enzymol. 1 (1941) 63.
[63] J. S . Fruton, Adu. Enzymol. 53 (1982) 239.
[64] Y. V. Mitin, N. P. Zapevalova, E. V. Gorbunova, lnt. J. P e p . Protein
Res. 23 (1984) 528.
[65] W. Kullmann, J. Org. Chem. 47 (1982) 5300; J. Biol. Chem. 255 (1980)
[66] P. Luthi, P. L. Luisi, J. A m . Chem. SOC.106 (1984) 7285.
[67] H.-D. Jakubke, P. Kuhl, A. Konnecke, Angew. Chem. 97 (1985) 79; Angew. Chem. l n f . Ed. Engl. 24 (1985) 8 5 .
[68] K. Nakanishi, R. Matsuno, Eur. J. Biochem. 161 (1986) 533.
1691 J. B. West, C.-H. Wong, J. Chem. SOC.Chem. Commun. 1986, 417; J.
Org. Chem. 51 (1986) 5728.
1701 A. D. Napper, S. J. Benkovic, A. Tramontano, R. A. Lerner, Science
(Washington, D.C.) 237 (1987) 1041.
[71] J . W. Jacobs, P. G. Schultz, R. Sugasawara, M. Powell, J. A m . Chem.
SOC. 109 (1987) 2187.
[72] R. A. Lerner, A. Tramontano, Trends Biochem. Sci. (Perr. Ed.) 12 (1987)
Angew. Chem. l n f . Ed. Engl. 27 (1988) 913-922
Без категории
Размер файла
1 116 Кб
site, thein, enzymes, activ, catalytic, activity, alteren
Пожаловаться на содержимое документа