Investigations of an antibody ligase

Article abstract of DOI:10.1021/ja963167u

Further investigation of the monoclonal antibody 16G3 has revealed that it not only couples activated amino acids to form dipeptides with high turnover rates but also couples an activated amino acid with a dipeptide to form a tripeptide, as well as an act

Full text of DOI:10.1021/ja963167u

278
J. Am. Chem. Soc. 1997, 119, 278-282
Investigations of an Antibody Ligase
†
†
,†
David B. Smithrud, Patricia A. Benkovic, Stephen J. Benkovic,*
‡
‡
‡
‡
Carol M. Taylor, Kraig M. Yager, Jason Witherington, Barton W. Philips,
‡
,‡
,‡
Paul A. Sprengeler, Amos B. Smith, III,* and Ralph Hirschmann*
Contribution from the Department of Chemistry, PennsylVania State UniVersity,
UniVersity Park, PennsylVania 16802, and Department of Chemistry, UniVersity of PennsylVania,
Philadelphia, PennsylVania 19104
X
ReceiVed September 9, 1996
Abstract: Further investigation of the monoclonal antibody 16G3 has revealed that it not only couples activated
amino acids to form dipeptides with high turnover rates but also couples an activated amino acid with a dipeptide
to form a tripeptide, as well as an activated dipeptide with another dipeptide to give a tetrapeptide. The catalytic
rates for these reactions greatly exceed the background rate of ester hydrolysis providing average yields of 80%
within the assay time of 20 min. Importantly, the amount of product inhibition is low, allowing for high yields of
products using multiple addition of substrates to the same antibody reaction mixture. A sequential mechanism is
employed by 16G3 for dipeptide coupling, and this mechanism appears to hold for the formation of the other peptides.
High catalytic selectivity is observed for the nucleophilic R-amino group of an R,â-diamino nucleophile and for the
para substituent on the activated ester, traits that are consistent with hapten design. The former chemoselectivity is
crucial for the condensation of fragments which are unprotected at the ꢀ-amino group of lysine.
Introduction
Other methods for coupling are based on combining native
or nonnative peptide fragments using enzymes or antibodies.
Subtilisin has been modified by Wells et al. by a double
In the past few years significant strides have been made in
the pursuit of synthesizing unique proteins or peptides in Vitro.
At this time, peptides containing approximately 50 amino acids
can be synthesized efficiently in a stepwise manner using solid-
support methodology or solution chemistry-based fragment
condensation. Larger peptides have also been synthesized both
5
mutation to give an effective ligase without extensive protease
or esterase activity The resulting subtiligase binds seven amino
acids in its active site with the first, fourth, and fifth amino
acid from the N-terminus providing most of the binding energy.
A broad range of specificity was observed with some amino
acids being well recognized while others are recognized only
weakly, thereby limiting the choice of amino acid for chemical
reaction. So far, unnatural amino acids have not been shown
to be recognized by subtiligase, but in the synthesis of
in solution1 or on solid support. Because many proteins are
larger than these, other methods are needed to couple fragments.
One approach that has been employed in the construction of
a
1b
2a
the transcription factor-related protein cMyc-Max and a HIV-1
protease analog2b entails the stepwise synthesis of peptide
fragments of about 30-50 amino acids in length and then their
chemoselective ligation. Solid-phase synthesis was employed
to incorporate the unique active groups that provided for the
thioester and oxime linkage in the former study and the thioester
and disulfide bond in the latter. These syntheses did not require
protecting groups, and the proteins were active. Thus, coupling
occurred in positions that reconstructed the proteins with the
correct orientation and did not interfere with their function. An
6
ribonuclease A, Wells successfully incorporated unnatural
amino acids within peptide fragments that had been prepared
by solid-phase synthesis.
Catalytic antibodies are an attractive choice for peptide
coupling considering that peptide fragments are a natural epitope
and antibodies have been constructed to catalyze a wide range
7
8
of chemical reactions. X-ray crystallographic analyses reveal
that antibodies bind peptides of various lengths in elongated
grooves using van der Waal, hydrogen bonding, and ionic
contacts for recognition. Therefore, antibodies can either be
used for highly specific coupling or be designed to recognize a
variety of natural and unnatural amino acids. As a first step
toward generating antibodies capable of coupling unprotected
3
oxime linkage has also been used to form a cyclic peptide,
and disulfide bonds were used in the partial construction of the
protein basic pancreatic trypsin inhibitor.
4
†
Pennsylvania State University.
University of Pennsylvania.
Abstract published in AdVance ACS Abstracts, January 1, 1997.
‡
X
(4) Fotouhi, N.; Galakatos, N. G.; Kemp, D. S. J. Org. Chem. 1989, 54,
2803-2817.
(1) See, for example: (a) Hirschmann, R.; Nutt, R. F.; Veber, D. F.;
Vitali, R. A.; Varga, S. L.; Jacob, T. A.; Holly, F. W.; Denkewalter, R. G.
J. Am. Chem. Soc. 1969, 91, 507-508. Yajime, H.; Fujii, N. J. Am. Chem.
Soc. 1981, 103, 5867-5871. (b) Gutte, B.; Merrifield, R. B. J. Am. Chem.
Soc. 1969, 91, 501-502. Nutt, R. F.; Brady, S. F.; Darke, P. L.; Cicarone,
T. M.; Colton, C. D.; Nutt, E. M.; Rodkey, J. A.; Bennett, C. D.; Waxman,
L. H.; Sigal, I. S.; Anderson, P. S.; Veber, D. F. Proc. Natl. Acad. Sci.
U.S.A. 1988, 85, 7129-7133.
(5) (a) Abrahms e´ n, L.; Tom, J.; Burnier, J.; Butcher, K. A.; Kossiakoff,
A.; Wells, J. A. Biochemistry 1991, 30, 4151-4159. (b) Ballinger, M. D.;
Tom, J.; Wells, J. A. Biochemistry 1995, 34, 13312-13319. (c) Chang, T.
K.; Jackson, D. Y.; Burnier, J. P.; Wells, J. A. Proc. Natl. Acad. Sci. U.S.A.
1994, 91, 12544-12548.
(6) Jackson, D. Y.; Burnier, J.; Quan, C.; Stanley, M.; Tom, J.; Wells,
J. A. Science 1994, 266, 243-247.
(
2) (a) Canne, L. E.; Ferr e´ -D’Amar e´ , A. R.; Burley, S. K.; Kent, S. B.
(7) (a) Stewart, J. D.; Benkovic, S. J. Chem. Soc. ReV. 1993, 22, 213-
219. (b) Lerner, R. A.; Benkovic, S. J.; Schultz, P. G. Science 1991, 252,
659-667.
(8) (a) Rini, J. M.; Stanfield, R. L.; Stura, E. A.; Salinas, P. A.; Profy,
A. T.; Wilson, I. A. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 6325-6330.
(b) Wilson, I. A.; Stanfield, J. M.; Rini, J. M.; Arevalo, J. H.; Schulze-
Gahmen, U.; Fremont, D. H.; Stura, E. A. Ciba Found. Sym. 159, 13-28.
H. J. Am. Chem. Soc. 1995, 117, 2998-3007. (b) Baca, M.; Muir, T. W.;
Schn o¨ lzer, M.; Kent, S. B. H. J. Am. Chem. Soc. 1995, 117, 1881-1887.
(
3) Pallin, T. D.; Tam, J. P. J. Chem. Soc., Chem. Commun. 1995, 2021-
2
022. See also: Tam, J. P.; Lu, Y.-A.; Liu, C.-F.; Shao, J. Proc. Natl. Acad.
Sci. U.S.A. 1995, 92, 12485-12489. Liu, C.-F.; Rao, C.; Tam, J. P.
Tetrahedron Lett. 1996, 37, 933-936.
S0002-7863(96)03167-8 CCC: $14.00 © 1997 American Chemical Society

InVestigations of an Antibody Ligase
J. Am. Chem. Soc., Vol. 119, No. 2, 1997 279
Figure 1. Hapten (4) used to elicit antibody 16G3.
Scheme 1
Figure 2. Lineweaver-Burke plots for 16G3 catalysis with (a) 2 held
at constant concentrations of 200 µM (2), 750 µM (9), and 2000 µM
(b) reacting with 1b from 100 to 1000 µM (2), 100 to 1000 µM (9),
and 100 to 1270 µM (b) and (b) 1b held at constant concentrations of
100 µM (2), 266 µM (9), and 500 µM (b) reacting with 2 from 200
to 2000 µM (2, 9, and b).
peptide fragments, we generated antibody 16G3 which catalyzes
the formation of a di-, tri-, and tetrapeptide (Vide infra).
Results and Discussion
In a recent publication, we demonstrated that antibody 16G3
was able to catalyze the reaction between 2 and various activated
esters including 1a,c (Scheme 1) forming a series of dipeptides.
Product formation was accelerated by approximately 400-fold
over background for the coupling of activated N-acetyl-L-leucine
value for the background velocity of dipeptide formation is 0.04
M ‚min ; comparing kcat‚Km for the weaker binding (less
9
-1
-1
-1
2
-1
-1
favorable) substrate 2 (7.0 × 10 M ‚min ) gives a rate
4
enhancement of 2 × 10 . Access to a larger range of substrate
concentrations was limited by the poor solubility of 1b.
Bolstered by the successful coupling of amino acids, the
potential of 16G3 was tested further by the formation of larger
peptides. The activated esters of N-acetyl-L-phenylalanine (1c)
and N-acetyl-glycyl-L-phenylalanine (7) were allowed to react
with D-tryptophanyl-glycine (5) to produce the N-acetylated
tripeptide 6 and N-acetylated tetrapeptide 8, respectively
(1a) with D-tryptophan amide (2). 16G3 was obtained by using
10
hapten (4) (Figure 1) as an immunogen, and even though only
one diastereomer was used, collectively the antibodies react with
all stereomeric combinations of ester and amide. The smaller
size of the substrates compared to the hapten is one possible
explanation for the lack of chiral recognition observed in the
coupling reaction.9 A lack of preference is also observed in
hapten binding where 16G3 binds both diastereomers of 4 with
KD values less than 2 nM. It is noteworthy, however, that
antibody 16G3 does not catalyze the racemization of N-acylated
p-nitrophenyl esters such as 1c.
(
Schemes 2 and 3). Formation of 6 and 8 was extremely facile
with almost complete conversion (92% tripeptide formation and
0% tetrapeptide formation) within the assay time of 20 min.
Rigorous kinetic analyses were not performed, but a lower limit
7
-
1
-1
for product formation is 50 µM‚min for 6 and 30 µM‚min
-
1
for 8 with background velocities for both of 0.2 µM‚min .
These velocities for 16G3 are a lower estimate, considering that
the values are obtained at 30 s into the assay (the first data
point that could be accurately taken) and that the poor solubility
of substrates 1c and 7 prevented substrate saturation of the
antibody active site.
To determine the mechanism employed by 16G3, the reaction
of 1b with 2 was examined in detail under steady-state
conditions. The change in velocity of product formation was
measured with the concentration of substrate 1b held constant
while varying the concentration of 2. The converse experiment
was also performed; the resulting plots (Figure 2) reveal a
sequential mechanism with the binding of one substrate not
affecting the binding of the second substrate. Second-order plots
gave Michaelis constants of 100 µM for 1b and 900 µM for 2
The catalytic efficiency of 16G3 is a result of both its rate
enhancement for bond formation and the lack of substantial
product inhibition. The high turnover of 16G3 was first evident
with the production of 1.8 mM dipeptide 3a in approximately
-
1
and a kcat of 0.6 min /antibody binding site. An approximate
2
h using 20 µM antibody. A similar production curve was
obtained for 8 with 0.2 mM being synthesized in 2 h also using
0 µM 16G3 (Figure 3a). To obtain estimates for the bond
(
9) Hirschmann, R.; Smith, A. B., III; Taylor, C. M.; Benkovic, P. A.;
Taylor, S. D.; Yager, K. M.; Sprengeler, P. A.; Benkovic, S. J. Science
994, 265, 234-237.
10) Smith, A. B., III; Taylor, C. M.; Benkovic, S. J.; Hirschmann, R.
Tetrahedron Lett. 1994, 35, 6853-6856.
2
1
formation rates and product inhibition constants, the time courses
for multiple addition of substrates to a single antibody solution
(

280 J. Am. Chem. Soc., Vol. 119, No. 2, 1997
Smithrud et al.
Scheme 2
Scheme 3
Figure 3. (a) Amount of dipeptide 3a (0) (values are multiplied by
0
.5) and tetrapeptide 8 (4) formed with time using 20 µM 16G3. Also
given are the calculated amounts of 3a (×) (0.5×) and 8 (+) derived
from KINSIM analysis using the constants given in Table 1. (b) Amount
of p-nitrophenol produced in the reaction forming the dipeptide (0)
(
0.5×) and tetrapeptide (4) and their corresponding values derived from
KINSIM analysis (×) and (+), respectively.
for the reactions producing the di- and tetrapeptides (Figure 3a)
and p-nitrophenol (Figure 3b), along with their product versus
time curves, were simulated by the KINSIM computer program
using a simple kinetic mechanism (Scheme 4). Also simulated
was the time course for the tripeptide reaction (time course not
shown). It was assumed that the antibody establishes a rapid
equilibrium with the substrates prior to catalysis and that bond
formation is the rate-determining step, kcat. For simulations,
major variations were performed on the rates of bond formation
and product dissociation constants.
Scheme 4
k
k2
k
3
1
k –1
k –2
k
k –4
4
k
5
k –5
Table 1. Kinetic and Binding Constants Derived from KINSIM
Analysis for the Mechanism Displayed in Scheme 4 That Describes
the Reaction between the Various Substrates and 16G3
By examining the estimated rate and binding constants (Table
), it seems that the catalytic mechanism employed by 16G3 is
the same for all substrates because the corresponding individual
cat (s-1
reaction
k
)
product
K
D
(µM) reactant
K
m
(µM)
1
1a + 2
1c + 5
7 + 5
0.05
0.02
0.01
3a
6
8
170
140
100
330
1a
1c
2
200
150
870
900
120
steps have similar values. Relative to the Km value for N-acetyl-
9
L-valine ester (Km ) 4 mM) , the data also show that substrate
p-O NC H OH
5
2
6
5
binding is more favorable for ones having larger hydrophobic
substituents at the R-position (N-acetyl-L-leucine Km ) 200 µM,
N-acetyl-L-phenylalanine Km ) 150 µM). This is reasonable
considering that the hapten has a cyclohexyl ring at that position.
When the peptide is extended there is only a small effect on
binding, either as a substrate (Km for 1a,c and 7 and for 2 and
7
For catalytic antibodies to be useful in the condensation of
fragments unprotected at the ꢀ-amino groups of lysine, it is
critical that the antibodies catalyze acylation only at the R-amino
group of the nucleophile. This requirement was tested by
measuring the product distribution of a reaction between the
bis amino substrate 9 with 1c (Scheme 5). In the absence of
5
) or product (KD for 3a, 6, and 8). Thus, the smaller amount
of tetrapeptide produced over time compared to dipeptide is a
result of the poorer water solubility of 7 (its maximum
concentration is below its Km value) compared to 1a. For larger
peptides, the amount of specific binding and the rate of bond
formation should be roughly constant because only portions of
the substrates, roughly the size and shape of the hapten, interact
with the antibody pocket. Other binding energies can result
from nonspecific interactions between the polypeptide and the
antibody.
1
6G3, the reaction afforded 10 and 11 in approximately equal
-
1
amounts with a velocity of 0.1 µM‚min . But in the presence
of 16G3, the desired reaction proceeded with greater than 80-
fold preference for 1c reacting with the R- relative to the
â-amino group based on initial velocity measurements (i.e., 8
µM‚min for 10 and 0.1 µM‚min for 11). Importantly, the
production rate of 11 did not change in the presence of 16G3,
-
1
-1

InVestigations of an Antibody Ligase
J. Am. Chem. Soc., Vol. 119, No. 2, 1997 281
Scheme 5
by resonance through the aromatic ring onto the nitro group
which is also destabilizing.
One advantage of using catalytic antibodies is that they readily
form a chiral binding pocket when challenged with a chiral
1
3
hapten. In a recent review, Kirby’s data accumulation on
catalytic antibodies showed they have a great potential for
asymmetric catalysis. Considering that hapten (4) contains three
stereogenic centers, it is surprising that the antibodies so far
isolated from the immune repertoire do not discriminate among
the different enantiomers of the substrates. On the basis of the
9
rates for different activated esters and the greater affinity for
esters containing larger side chains, we postulate that the loss
of chiral recognition for 16G3 could also be due in part to the
mismatch between the combining site induced by the p-
nitrobenzyl hapten and the p-nitrophenyl substrate. If one
assumes that the para nitro-substituted aromatic ring provides
the strongest binding epitope for the substrate, binding of the
substrate will result in a shift of the ester, by approximately
the one methylene unit away from the pocket designed for the
side chain, and thus a loss of chiral recognition.
Results from the study of antibody 9B5.1,14 which also
catalyzes peptide bond formation, corroborate our hypothesis
for the weaker chiral recognition observed with 16G3. 9B5.1
was raised against a hapten containing a phenyl ester and
designed to catalyze peptide bond formation between a phenyl-
alanyl amino group and a phenyl ester derivative of L-alanine.
Since the phenyl ester substrate was not reactive, the ester was
converted to an acyl azide giving a reactive substrate with a
and 12 was not formed with or without 16G3 (the reactions
were run with excess 9).
Specificity for the electrophilic species is also high. Within
the ester series, the p-NO2 substrates 1a-c are turned over by
the antibody, but the reactions involving the p-OCH3 (1d),
m-NO2 (1e), and p-Cl (1f) substrates are not catalyzed by 16G3.
Even though the spontaneous hydrolysis of 1e limited its reaction
time, it was determined that 1e does not bind strongly to the
antibody pocket because it was not an inhibitor of the reaction
-1
3
-1
-1
kcat‚Km value of 3.9 × 10 M ‚min and a rate enhancement
2
-1
-1
of 10 derived from the relationship kcat‚Km ‚kuncat . It should
be noted that calculations of rate enhancements using these ratios
for bimolecular reactions require the lower of the two kcat‚Km
-1
values found for the two substrates. The flux through the
catalyst is limited by the least favorably processed substrate,
2
and thus, the efficiency of antibody 9B5.1 is 10 rather than
1
1
between 1b and 2. It is evident that the p-NO2 moiety on the
substrate is crucial for both binding and activation. More
importantly, having the substrate contain a p-nitrophenyl ester,
whereas the antibody was induced by a hapten containing a
p-nitrobenzyl in the corresponding position, appears to provide
reduced product inhibition by p-nitrophenol (KD ) 330 µM,
Table 1) and subsequently higher yields of product. Evidence
for this hypothesis can be obtained by comparing the turnover
number for p-nitrophenyl ester hydrolysis by the antibody
4 14
the reported 10 . More importantly, changing the acylating
agent to a nonrecognizable epitope, an azide, did not substan-
tially alter the acyl donors orientation in the binding pocket
which is evident by its small Michaelis constant (Km ) 15 µM).
Thus, the methyl side chain of the acyl azide still had access to
the binding pocket, resulting in a 7.3-fold enhanced rate for
the L- versus D-alanine-derived azide. This level of chiral
recognition is on the same scale as observed for N-acetyl-L-
versus N-acetyl-D-phenylalanine with 16G3.
4
3C912 with 16G3 catalysis. 43C9 was raised against a hapten
Accordingly, greater chiral recognition should occur for
substrates containing larger side chains that can enter the
hydrophobic pocket. This is indeed true and evident by
examining the change in effective molarity of N-acetyl-L- versus
N-acetyl-D-valine (45 to 26 M) compared to N-acetyl-L- versus
containing a p-nitrophenyl substituent, and its catalysis was
limited by product inhibition of p-nitrophenol (KD ) 1 µM).
Binding of a nitrophenyl ester in a pocket sculpted for a
nitrobenzyl moiety was anticipated to affect the catalytic steps
depicted in Scheme 4. Binding should be weaker (especially
for substrates with small side chains) because optimal binding
of the nitro group exposes the phenolic oxygen atom to a
hydrophobic patch on the antibody surface reserved for the
methylene group. Binding of the transition state should also
be weaker due to improper alignment of the atoms, but this
loss in stabilization, which reduces the rate of bond cleavage,
could be countered by the less stable nitrophenyl ester. Finally,
once bond cleavage occurs, p-nitrophenolate will be driven from
the binding pocket as a result of its greater negative charge.
The increased electron density should interact more unfavorably
with the hydrophobic patch and thereby be partially transferred
1
5
N-acetyl-D-phenylalanine (550 to 71 M). As the side chain
substituent becomes larger, chiral recognition increases from
5
7% to 87% along with a 10-fold rate enhancement. Another
means for enhancing chiral recognition would be to use a
p-nitrobenzyl ester as a substrate, but this ester is less reactive
and not turned over by the antibody. In fact, 16G3 was
9
originally designed to catalyze a variety of reactions via the
cyclohexyl ring in 4. The weaker chiral recognition, albeit
fortuitous, is also of benefit as it expands the repertoire of
reaction partners. Having a single antibody capable of multiple
fragment couplings would save time and effort by eliminating
(
11) The inhibition study was run with equal concentrations of 1e,b (100
(13) Kirby, A. J. Acta Chem. Scand. 1996, 50, 203-210.
(14) Jacobsen, J. R.; Schultz, P. G. Proc. Natl. Acad. Sci. U.S.A. 1994,
91, 5888-5892.
(15) The effective molarities were derived from rates given in Table 1
of ref 9.
µM) and 2 mM 2. For the experimental procedure see HPLC analysis under
Steady-State Kinetics in the Experimental Section.
(12) Gibbs, R. A.; Benkovic, P. A.; Janda, K. D.; Lerner, R. A.; Benkovic,
S. J. J. Am. Chem. Soc. 1992, 114, 3528-3534.

282 J. Am. Chem. Soc., Vol. 119, No. 2, 1997
Smithrud et al.
the need for the production of unique antibodies for each set of
coupling partners.
antibodies capable of catalyzing the ligation of large, unprotected
peptides.
4
The rate enhancement provided by 16G3 of 2 × 10 is
Experimental Section
consistent with the values obtained from the limited number of
bimolecular reactions catalyzed by antibodies16 but approxi-
D
Determination of K by Fluorescence Quenching. The quenching
of the intrinsic antibody fluorescence caused by hapten addition was
monitored using a SLM 8000 fluorimeter. Inner filtering effects were
corrected for by subtracting the fluorescent change observed by the
addition of hapten to a tryptophan solution of known concentration.
5
a
mately 10-fold slower than observed for substiligase. The
reaction efficiencies for both ligases are generally around 70%
or greater, although subtiligase substrates are longer peptides
and cyclizations catalyzed by this enzyme are also constrained
to large peptide fragments. To date, 16G3 has not been
challenged with larger peptides; however, we do expect that
cyclization of small peptide fragments may be possible. We
also anticipate that the catalytic rate may be improved because,
as in the case of most antibodies, the conversion of binding
energy to catalysis is not complete. For example, incorporation
of additional amino acids into the hapten may enhance both
binding and selectivity.
17
The resulting plots were then fit, by a procedure described previously,
to give dissociation constants.
Steady-State Kinetics. The initial velocity of product formation
was determined by measuring the change in absorbance of p-nitrophenol
-
1
-1
at 400 nM (ꢀ ) 9600 cm ‚M ) using a Cary (Olis-14) spectropho-
tometer. The appropriate amounts of substrates and antibody were
mixed at 25 °C in 0.1 M 3-(N-morpholino)propanesulfonic acid (MOPS)
at pH 7.0, µ ) 0.1 M NaCl, and containing 5% DMSO. After
correcting for the background rate of ester hydrolysis, the kcat and K
m
values were derived by fitting the data to the Michaelis-Menten
equation. The simulations of kinetic data were done with the program
KINSIM (Barshop et al.) as modified by Anderson.18
The amount of product formed during a reaction was also obtained
by mixing the reagents (same solvents and buffer as previously
discussed), removing aliquots, quenching with perchloric acid to pH
One can measure the catalytic efficiency by comparing the
difference in the binding energy of hapten to substrate with the
-
1
ratio of the catalyzed versus the uncatalyzed rate: (Km1‚Km2)‚Ki
≈
-
1 16
kcat‚kuncat
.
For haptens that are competitive inhibitors,
KD can substitute for Ki. According to this relationship, the
∼
2.5, and separating by HPLC (Water 600E) equipped with an
antibody is not a very efficient catalyst for the reaction between
analytical C18 reversed-phase column (VYDAC, Hesperia, CA) with
an acetonitrile-water:0.1% trifluoroacetic acid gradient. Products were
identified by matching retention times with authentic samples, and
concentrations were derived by integrating the peaks.
2
and the p-nitrophenyl esters of N-acetyl-L-valine and N-acetyl-
L-phenylalanine giving 5% and 17% of the maximum rate
enhancement, respectively. On the other hand, the differences
in the values are consistent with hapten design, where the
isopropyl side chain of valine binds weaker than the phenyl
side chain of phenylalanine. Even with improved technology
for developing catalytic antibodies, the efficiency for catalysis
may never approach 100% especially for stable acylating agents.
It is possible that antibodies may require active substrates, acting
only to guide more energetic species along a desired reaction
coordinate.
Antibody-Catalyzed Product Formation. The total amount of
product that can be formed by 16G3 catalysis was determined by mixing
20 µM 16G3 with 2 mM 5 and either 0.1 mM 1c or 7 at 25 °C in 0.25
mL of MOPS (pH 7.0, µ ) 0.1 M NaCl, 5% DMSO). At 20 min
intervals, 0.03 mL aliquots were removed and analyzed for the amount
of product formed. Because the substrates are almost completely
depleted at 20 min, appropriate amounts of the substrates were added
to the antibody and product mixture at this time to return their
concentrations back to the initial conditions. The substrate aliquots
were withdrawn from very concentrated stock solutions so as to
maintain the antibody concentration, giving a final concentration of
Preparative Experiments
The requisite substrates and all possible peptidyl products
required for HPLC analysis of the catalytic antibody reaction
mixtures and subsequent product identification are described
in detail in the Supporting Information. The synthesis of the
hapten was reported previously.1
1
6 µM for 16G3 after six additions. In other multiple-addition
experiments, the tripeptide and tetrapeptide fractions were collected,
and their authenticity was confirmed by mass spectrometry.
Acknowledgment. We are pleased to acknowledge financial
support provided by the National Institutes of Health (Institute
of General Medical Sciences) through Grant GM-45611,
Bachem, CA, and the Merck Research Laboratories.
0
Conclusion
We have demonstrated that antibody 16G3 is capable of
catalyzing the coupling of an amino acid with a dipeptide and
the coupling of two dipeptides to form tri- and tetrapeptides,
respectively, even though the hapten was originally designed
to create antibodies that only catalyze the production of
dipeptides. This result, along with the demonstrated rate
enhancements, modest product inhibition, and critical chemose-
lectivity, provides impetus to producing the next generation of
Supporting Information Available: Complete experimental
details for compounds 1c-f, 5-12, 14-17, 19, 20, and 22-28
(
18 pages). See any current masthead page for ordering
information and Internet access instructions.
JA963167U
(
17) Taira, K.; Benkovic, S. J. J. Med. Chem. 1988, 31, 129-137.
(18) (a) Barshop, B. A.; Wrenn, R. F.; Frieden, C. Anal. Biochem. 1983,
30, 134-135. (b) Anderson, K.; Silorski, J. A.; Johnson, K. A. Biochem-
istry 1988, 27, 7395-7406.
1
(16) Stewart, J. D.; Benkovic, S. J. Nature 1995, 375, 388-391.