Synthesis and physical characterization of a P1 arginine combinatorial library, and its application to the determination of the substrate specificity of serine peptidases

Article abstract of DOI:10.1016/S0968-0896(02)00174-8

Serine peptidases are a large, well-studied, and medically important class of peptidases. Despite the attention these enzymes have received, details concerning the substrate specificity of even some of the best known enzymes in this class are lacking. One approach to rapidly characterizing substrate specificity for peptidases is the use of positional scanning combinatorial substrate libraries. We recently synthesized such a library for enzymes with a preference for arginine at P₁ and demonstrated the use of this library with thrombin (Edwards et al. Bioorg. Med. Chem. Lett. 2000, 10, 2291). In the present work, we extend these studies by demonstrating good agreement between the theroretical and measured content of portions of this library and by showing that the library permits rapid characterization of the substrate specificity of additional SA clan serine peptidases including factor Xa, tryptase, and trypsin. These results were consistent both with cleavage sites in natural substrates and cleavage of commercially available synthetic substrates. We also demonstrate that pH or salt concentration have a quantitative effect on the rate of cleavage of the pooled library substrates but that correct prediction of optimal substrates for the enzymes studied appeared to be independent of these parameters. These studies provide new substrate specificity data on an important class of peptidases and are the first to provide physical characterization of a peptidase substrate library.

Full text of DOI:10.1016/S0968-0896(02)00174-8

Bioorganic & Medicinal Chemistry 10 (2002) 3637–3647
Synthesis and Physical Characterization of a P1 Arginine
Combinatorial Library, and Its Application to the Determination
of the Substrate Specificity of Serine Peptidases
Stephen T. Furlong,a,* Russell C. Mauger,b Anne M. Strimpler,a Yi-Ping Liu,b
Frank X. Morrisb and Philip D. Edwardsb
aDepartment of Molecular Science, AstraZeneca, Wilmington, DE 19850, USA
bDepartment of Chemistry, AstraZeneca, Wilmington, DE 19850, USA
Received23 January 2002; accepted24 April 2002
Abstract—Serine peptidases are a large, well-studied, and medically important class of peptidases. Despite the attention these
enzymes have received, details concerning the substrate specificity of even some of the best known enzymes in this class are lacking.
One approach to rapidly characterizing substrate specificity for peptidases is the use of positional scanning combinatorial substrate
libraries. We recently synthesizedsuch a library for enzymes with a preference for arginine at P anddemonstratedthe use of this
1
library with thrombin (Edwards et al. Bioorg. Med. Chem. Lett. 2000, 10, 2291). In the present work, we extendthese studies by
demonstrating good agreement between the theroretical and measured content of portions of this library and by showing that the
library permits rapid characterization of the substrate specificity of additional SA clan serine peptidases including factor Xa, tryp-
tase, andtrypsin. These results were consistent both with cleavage sites in natural substrates andcleavage of commercially available
synthetic substrates. We also demonstrate that pH or salt concentration have a quantitative effect on the rate of cleavage of the
pooled library substrates but that correct prediction of optimal substrates for the enzymes studied appeared to be independent of
these parameters. These studies provide new substrate specificity data on an important class of peptidases and are the first to pro-
vide physical characterization of a peptidase substrate library.
# 2002 Elsevier Science Ltd. All rights reserved.
Introduction
provide the opportunity for rapid study of substrate
specificity for many enzymes in parallel.1 In principle,
substrate library approaches can result in detailed
knowledge of enzyme specificity for a peptidase without
prior knowledge of the native substrate. However, sur-
prising new results can be obtainedeven when char-
acterizing well-studied enzymes.
A key to understanding the activity of any peptidase is
knowledge of the substrate specificity of the enzyme.
Such understanding provides a route to the design of
enzyme assays, and, potentially, a starting place for
structure-based design of peptidic and/or non-peptidic
inhibitors. Until recently, determining substrate specifi-
city for peptidases required considerable effort, typically
involving studies with native or synthetic substrates,
sequence andstructural data, andnaturally occurring or
synthetic inhibitors. Traditionally, most such investi-
gations of substrate specificity have been conducted with a
single or small number of enzymes, andthis work can be
complex andtime consuming. Recently, however, combi-
natorial substrate libraries have been devised for reveal-
ing peptidase substrate specificity and these libraries
Recently, we synthesizedan aminocoumarin positional
scanning library for determining substrate specificity for
serine peptidases.2 This library is suitable for use with
serine peptidases that prefer Arg at the P1 position and
is similar to other recent work with positional scanning
libraries for peptidases with a preference for Asp at P1.3
Since many serine peptidases are known to have a pref-
erence for Arg at P1, such a library is potentially useful
in the comprehensive characterization of a large subset
of all the known enzymes in this class or for revealing
substrate specificity of novel serine peptidases. In the
present study, we have tested 12 serine peptidases with
this combinatorial library. These studies have revealed
*Corresponding author. Tel.: +1-302-886-8588; fax: +1-302-886-
2766; e-mail: stephen.furlong@astrazeneca.com
0968-0896/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(02)00174-8

3638
S. T. Furlong et al. / Bioorg. Med. Chem. 10 (2002) 3637–3647
detailed information on the substrate specificity of four
of these enzymes. In addition, we describe here the
complete synthetic details for construction of the
library, anda methodfor physical characterization of
the library to ensure synthetic fidelity.
Chemistry
The methodology we developed to determine the sub-
strate specificity of SA clan serine peptidases2 centered
aroundthe construction of a positional scanning com-
binatorial substrate library (PS-CSL) via solidphase
techniques. This library was usedto evaluate the speci-
ficity of the P4–P2 positions of tetrapeptides containing
a P1 arginine aminocoumarin (AMC). This is accom-
plishedby constructing separate P 2, P3 andP sub-
4
libraries. In order to evaluate 20 different amino acids at
each position, 20 reaction vessels or ‘pots’ are prepared
in each sub-library, with a known residue. Using the P2
library for illustration, the resin boundP -Arg-AMC is
1
placedin 20 pots andcoupledwith a single, different
amino acidresidue in each pot. Each pot is then coupled
with the same isokinetic mixture of all 20 amino acids4
to generate an equimolar mixture of all possible 20 tri-
peptides with a known P2 residue (e.g., O-V-Arg-AMC).
Next, each pot is again coupledwith the same isokinetic
mixture to generate all 400 tetrapeptides (e.g., Ac-O-O-
V-Arg-AMC). Following cleavage from the resin, each
pot is assayedfor its relative rate of hydrolysis by the
enzyme under evaluation. Since each pot contains all
possible tetrapeptides with a known, but different P2
residue, the pots displaying the greatest level of hydro-
Scheme 1. (i) Ethyl bromoacetate, K2CO3, DMF, rt, 18 h, 75%; (ii)
HSO3Cl, CH2Cl2, 4 ꢀC, 1 h, 85%; (iii) Boc-Arg(Z)2-OH, Boc2O, Py,
THF, rt, 18 h, 76%; (iv) Pd(OH)2, EtOH, 1 N HCl, H2/48 psi, 4 h,
93%; (v) 1 N NaOH (2 equiv), acetone, rt, 0.5 h, 32%; (vi) 1 N NaOH
(1 equiv), THF/EtOH (1:1), rt, 3 h, 82%.
sulfonamide forming reaction immediately upon solvent
removal. We have also foundthat the best overall yields
are obtainedwhen sulfonamied formation andester
hydrolysis are carried out in two independent, sequential
reactions. A significant amount of effort was requiredto
identify appropriate conditions for attaching 7 to a solid
support (Scheme 2). We initially investigatedtwo different
PEG polystyrene resins. Since the amino group attach-
ment point in these resins is distal to the cross-linked resin
backbone, we felt that these resins wouldafforda more
efficient amide coupling with the relatively bulky acid 7.
However, the physical properties of the coupledproducts
were such that severe mechanical losses were encountered.
These losses continuedto plague us throughout the solid-
phase synthetic sequence. To overcome this problem and
still maintain an extended attachment point to the resin,
we attacheda b-alanine linker to an AgroPoreTM-NH2
resin. When coupledwith 7, the resin boundinter-
mediate 9 hadexcellent handling properties andno sig-
nificant mechanical losses were observed. The couplings
of both the b-alanine to the resin andof 7 to 8 were
monitoredby following the growth of the amide carbonyl
absorption peak of a beadusing solid-phase IR since a
negative Kaiser test couldnot be obtained. Maximum
loading was considered the point where there was no fur-
ther increase in the carbonyl absorption, andwas achieved
using a two-step sequence involving basic coupling con-
ditions (HATU and DIPEA) followed by a more acidic
system (DIPCI andHOBt). Unreactedamino groups on
the resin were then cappedwith acetic anhyrdied.
Recently, a procedure has been reported for the solid-
phase synthesis of peptide substrates with a P1 argininne-
aminocoumarin using a safety-catch linker approach.8
lysis possess the preferredP
libraries are preparedfor the P andP positions. Fol-
residues. Similar sub-
2
3
4
lowing assay of these libraries, the combineddata is
usedto determine the preferredsubstrate specificity at
each position. While there have been several reports of
the solid-phase synthesis of amidines and guanidines,5
little has been publishedon the solid-phase synthesis of
peptides containing arginine linked to the resin through
the side chain guanidine group. As we began consider-
ing a strategy for attaching an arginine to a solid-phase
resin via its side chain guanidine group, two reports
appearedin the literature edscribing such a proce-
dure.6,7 Zhong et al. employedan aromatic sulfonyl
linker, which is conceptually similar to the procedure we
ultimately employed. However, attachment of the argi-
nine guanidine to the resin-bound sulfonyl group
required4 N KOH at 75 ꢀC for 2 days, conditions
incompatible with the presence of an aminocoumarin
amide required in our synthesis. Most relevant to our
needs was a poster presentation by Urban and Dattilo
describing the solid phase preparation of a P1 arginine
tripeptide p-nitroanilide substrate using a novel CMtr
linker. The CMtr is a modified Mtr containing a car-
boxyl group for attachment to the resin.7 We useda
modified version of this procedure for preparing the
intermediate Boc-Arg(CMtr)-Amc 7 (Scheme 1). The
chemistry is straightforwardwith a few notable excep-
tions. The sulfonyl chloride 3 is relatively unstable, and
the reaction time for its formation shouldbe limitedto
1 h. Furthermore, 3 shouldbe usedin the subsequent

S. T. Furlong et al. / Bioorg. Med. Chem. 10 (2002) 3637–3647
3639
Figure 1. MALDI-TOF mass spectra of P3 sublibrary pots: (a) Ac-O-
Val-O-Arg-Amc; (b) Ac-O-Asp-O-Arg-Amc.
Scheme 2. (i) Boc-b Ala-OH, HATU, DIPEA, DMF, 18 h, repeat for
3 h; (ii) HOBT, DIPCI, DMAP (cat), DMF, 18 h, repeat for 3 h; (iii)
Ac2O, DIPEA, DMF, 2 h; (iv) 40% TFA/CH2Cl2, 0.5 h; (v) HATU,
DIPEA, DMAP (cat), DMF, 18 h; (vi) HOBt, DIPCI, DMAP (cat),
DMF, 18 h; (vii) 40% TFA/CH2Cl2, 1 h; (viii) isokinetic mixture or
known amino acid, HATU, DIPEA, DMF, 2 h, repeat procedure
once; (ix) 20% piperidine, DMF, 15 min, repeat procedure once; (x)
thioanisole/TMSBr/ triisopropyl silane/phenol/TFA.
Results
Physical characterization of the Arg P1 positional
scanning library
Although there have been several reports on the pre-
paration anduse of positional scanning combinatorial
substrate libraries, these libraries have only been char-
acterizedbiochemically (functionally). The assumption
is made that since isokinetic mixtures of amino acids are
usedin the construction of the libraries, all possible
peptides will be present in equimolar mixtures. How-
ever, since the efficiency of solid-phase amide couplings
is dependent upon the exact sequence of the resin-bound
substrate, the experimentally derived isokinetic mixtures
may not affordequimolar products for a particular sub-
library. In order to validate the synthetic fidelity of our
libraries, we have developed a procedure for physical
analysis of the libraries. Using the P3 sub-libraries Ac-
O-Val-O-Arg-Amc andAc- O-Asp-O-Arg-Amc as
examples, the pots were analyzedby means of MALDI-
TOF mass spectrometry (Fig. 1). The mass spectral data
was assessedfor the presence of three features: (1) that
the smallest (G-V-G-; G-D-G-) andlargest (W-V-W-;
W-D-W) molecular weight peptides were present; (2)
that the center of the distribution for each of the two
pots was shiftedfrom each other by 16, the difference in
molecular weight between valine andaspartic acid; and
(3) that the molecular weight distribution was centered
Figure 2. SimulatedMALDI-TOF mass spectra of P 3 sublibrary pots:
(a) Ac-O-Val-O-Arg-AMC; (b) Ac-O-Asp-O-Arg-AMC.
aroundthe average molecular weight. Figure 1 is a
comparison of the mass spectra for the-Val- and-Asp-
pots. As anticipated, the high and low molecular weight
compounds were present and the center of distribution
between each library was shiftedby 16. The average
molecular weights for the Ac-O-Val-O-Arg andAc- O-
Asp-O-Arg pots are 705 and721, respectively. The cen-
ters of distribution in Figure 1 are in good agreement
with these calculatedvalues. In aiddtion, we have
developed a program for computationally simulating
the predicted MALDI-TOF spectrum of an individual
sublibrary (MALDIsym, see Experimental). The simu-
latedmass spectra for the Ac- O-Val-O-Arg-Amc and
Ac-O-Asp-O-Arg-Amc are depicted in Figure 2, and
agree favorably with the experimental mass spectra in

3640
S. T. Furlong et al. / Bioorg. Med. Chem. 10 (2002) 3637–3647
Figure 1. The mass spectral analysis we have presented
here provides an easy and valuable method for evaluat-
ing the fidelity of a combinatorial substrate library.
Since interpretation of the results from the biological
evaluation of a substrate libraryis dependent upon the
synthetic fidelity of the library, we believe that an analysis
of the librarysuch as the MALDI-TOF analysis presented
here is a prerequisite for utilization of combinatorial sub-
strate libraries for screening against target enzymes.
consistent with the known properties of this enzyme.9
Specifically, thrombin has a strong P2 preference for Pro
andis less specific at P 3 andP 4. In the present study, we
expandon this initial result by comparing the results of
thrombin with this library to results using other serine
peptidases: Factor Xa, trypsin, and tryptase. As depic-
ted in Figure 3, in addition to substrates with proline at
P2, thrombin also cleavedsubstrates with aliphatic
amino acids at this position such as alanine, isoleucine,
valine, norleucine, valine or leucine. However, this
cleavage occurredat markedly lower rates than for sub-
strates with proline at P2. Generally, substrates with
Screening of the Arg P1 positional scanning library
In our previous report using the present library with
thrombin,2 we demonstrated cleavage of library substrates
chargedamino acids at P were not cleavedat a sig-
2
nificant rate by thrombin. By contrast, tryptase generally
Figure 3. Substrate specificities for serine peptidases. Assays were conducted for four serine peptidases with a known preference for arginine at P1.
The y axis shows the rate of AMC production expressed as a percentage of the maximum rate observed in each experiment. The x axis represents the
amino acids (B=d-alanine; J=norleucine). The data indicate more selectivity at the P2- andP 3-positions for thrombin, factor Xa, andtryptase.
Porcine pancreatic trypsin shows less specificity than the other enzymes at all positions. Each bar represents meanÆstandard deviation for an n of
three assays carried out on different days.

S. T. Furlong et al. / Bioorg. Med. Chem. 10 (2002) 3637–3647
3641
preferredneutral-polar amino acids at P
andalso tol-
2
eratedhydrophobic aliphatic, but not aromatic amino
acids. Factor Xa preferred the aromatic residues phe-
nylalanine andtryptophan at P but, interestingly, did
2
not cleave substrates with tyrosine in that position.
Factor Xa also cleavedsubstrates with glycine or serine
at P2 at a significant rate but, like thrombin, generally
did not cleave substrates with charged P2 residues. As
might be expectedfor a idgestive enzyme, trypsin
showedlittle preference at P 2. Like the other three
enzymes illustrated, though, trypsin did not cleave sub-
strates with the non-physiological d-alanine at P2. For
all four enzymes a wide variety of amino acids were
Figure 4. Cleavage of a panel of commercially available aminocou-
marin-labeledsynthetic substrates by tryptase. Assays were carriedout
under identical pH and buffer conditions to those described for assays
with the combinatorial library. All substrates assayedwere AMC
labeled peptides (shown by single-letter amino acid designations),
most with R at P1. Results are expressedas rate of cleavage by com-
parison to that for the most efficiently cleavedsubstrate (DPR-AMC).
toleratedat P . With the exception of trypsin, which
3
showedlittle preference at P 3, andtryptase, which
demonstrated a clear preference for the basic amino
acids arginine and lysine at this position, general pat-
terns were less clear at P3 than at P2. At P4, again tryp-
sin showedlittle preference andall enzymes tolerateda
wide variety of amino acids. Both thrombin and Factor
Xa generally seemedto prefer hyrdophobic-aliphatic
residues at P4 andboth cleavedsubstrates containing
chargedamino acids at a lower rate (with the exception
of glutamine or arginine containing substrates by Xa).
Tryptase cleavedsubstrates containing glutamate, tryp-
tophan and d-alanine at P4 poorly, andthose contain-
ing lysine, asparagine, tyrosine andproline at P 4 best.
hadsome effect on the cleavage rate, especially since our
studies have revealed that tryptase prefers basic residues
at P3.
Effect of salt and pH on substrate cleavage
It is possible that assay conditions could affect the rela-
tive cleavage rate of different substrates and thus pro-
duce confusing or misleading results. This possibility
couldbe particularly important when evaluating a new
or poorly characterizedenzyme. Two specific assay
conditions that might influence the relative rate of clea-
vage of substrates (either within or between pots) using
the positional scanning approach are salt concentration
andpH. To our knowledge, these affects have not as yet
been investigated. As a comparison, the effect of salt
concentration on cleavage of one tripeptide tryptase
substrate, H-Gly-Pro-Arg-AMC, was comparedto
results obtainedusing substrate pots. The best cleavage
of H-Gly-Pro-Arg-AMC by tryptase was observedin
the absence of salt and a dose dependent inhibition by
sodium chloride was observed. Similar results were
observedusing selectedpots from the library. The pots
Cleavage of homogeneous synthetic substrates by
tryptase
Using a library such as that which we describe here, a
comparison of cleavage for substrates containing a spe-
cific amino acidat positions P 2–P4 can be obtained.
These results are obtainedby using mixtures of syn-
thetic substrates in each ‘pot’, however, so it wouldbe
of interest to compare library results to those obtained
using homogeneous substrates. Since a variety of
homogeneous aminocoumarin peptide substrates are
commercially available we chose to compare tryptase
cleavage of selectedhomogeneous substrates with the
results obtainedusing the combinatorial library. Tryp-
tase was chosen as the representative enzyme with which
to conduct these studies because it was shown to have
clear specificity at positions P2–P4 andtwo tripeptide
aminocoumarin substrates, H-Val-Pro-Arg-AMC and
H-Phe-Ser-Arg-AMC are known to be cleavedby this
enzyme. However, basedon our initial substrate library
results with tryptase, other substrates were also expec-
ted to be cleaved. This prediction proved correct. As
shown in Figure 4, H-Val-Pro-Arg-AMC andH-Phe-
Ser-Arg-AMC were efficiently cleavedby tryptase but,
consistent with the library results, three other sub-
strates, H-Gly-Pro-Arg-AMC, H-Gln-Ala-Arg-AMC
andH-Asp-Pro-Arg-AMC, were superior. Also con-
sistent with the library observations for tryptase, H-
Arg-Arg-Arg-AMC andH-Phe-Arg-Arg-AMC are not
goodsubstrates. Somewhat surprisingly, several of the
substrates that containedK, G or R at P 2 also showeda
reasonable rate of cleavage albeit at a lower rate than
predictedbetter substrates. It shouldbe notedthat the
homogeneous substrates usedin the present studies were
not terminatedwith an acetyl group andthis may have
selectedfor this experiment were those from the P
2
libraries in which tryptase hadbeen shown to produce
the most efficient cleavage. As shown in Figure 5, results
from the selected pots showed a dose dependent inhibi-
tion by salt comparable to that observedwith the tri-
peptide substrate. Significantly, however, the rank order
of substrate preference was unaffected. As a further
comparison, two P2 pots were chosen to compare salt
effect on thombin cleavage. Again these pots were cho-
sen as representative pots that contain substrates effi-
ciently cleavedby the enzyme. There was no evidence,
however, that the salt concentration at which the assays
were conducted would markedly alter the rank order
substrate preferences of the enzyme. Thrombin was
further chosen to test the effect of pH on library
screening assays andin particular to determine if the
markedpreference for Pro at P couldbe alteredby
2
changing this parameter. Figure 6 shows results from
assays carriedout at pH 9.0. For the P library relative
cleavage at pH 9.0 was essentially identical to that
2

3642
S. T. Furlong et al. / Bioorg. Med. Chem. 10 (2002) 3637–3647
each enzyme which is highly reproducible and consistent
with previously publisheddata. Furthermore, although
there were kinetic differences in the cleavage of these
mixedsubstrate pools depending on salt concentration
andpH, the specific assay conditions hadonly minor
impact on the ability to predict the substrate specificity
of a particular enzyme.
The substrate specificity for each of the enzymes exam-
inedin the present study are summarizedin Table 1. In
this work we have shown that thrombin has a pref-
erence but not an absolute requirement for proline at
P2; other hydrophobic residues such as alanine and iso-
Figure 5. Effect of NaCl concentration on human lung tryptase clea-
vage of synthetic substrates. Each substrate has arginine at P1. The y
axis shows the rate of AMC production in arbitrary fluorescence units/
min. The x axis represents the amino acids at P2. The data shows a
decrease in the potency of the substrate with increasing concentrations
of salt. Although there is a decrease in the rate of cleavage with
increasing salt concentration, for all the amino acids shown there is
variation in the amount of the decrease that is dependent on the spe-
cific amino acid. Each bar represents the meanÆstandard deviation
for three assays.
leucine were also toleratedat P . By contrast, although
2
less specificity was demonstrated by thrombin at P3 than
P2, generally, there were some amino acids that were not
toleratedsuch as A, G, P, N andD. At P 4, the strongest
preference was shown for nor-leucine or other hydro-
phobic amino acids. Overall, these results for thrombin
agree with previous studies both in terms of known
cleavage sites in naturally occurring substrates andfrom
the results of other studies using positional scanning
libraries. For example, thrombin cleavage sites for sub-
strates in the clotting cascade, including multiple sites in
Factors V (709, 1018, 1545, 372) andVIII (373, 740,
1689), are consistent with the results shown here. Simi-
larly, known thrombin cleavage sites in other substrates,
including PAR1, Protein C, uPA, Factors XI, and Fac-
tor XIII, could also have been predicted. Interestingly,
while other studies yielded similar results to those
shown here, using a different approach to the synthesis
of the library, the results with our library yielded results
more consistent with the thrombin substrate specificity
for known substrates, overall. Specifically, while the
earlier studies suggested a nearly absolute requirement
for P at P2,10 we have shown here that thrombin has a
preference for P at P2 but will cleave substrates with
other hydrophobic residues at P2 nearly as well. Factor
Xa also yielded results with our library consistent with
but not identical to other studies.10 At P2 we foundthat
Xa prefers F>G, W>S> >A, P, N, H; hada pref-
erence for basic andaromatic hydrophobic residues at
P3; and preferred hydrophobic residues and R at P4.
shown at lower pH (Fig. 3). By contrast, thrombin
cleavage in the P3 pot showed a marked pH dependence.
For example at pH 9.0, the P3 ‘N’ pot containedthe best
substrates, while at lower pH (Fig. 3), this represented
one of the least efficiently cleavedsubstrates, andthe
‘R’, ‘L’, and‘Q’ pots were preferred. This is not unex-
pectedin that charge–charge electrostatic interactions
between enzyme andsubstrate wouldbe expectedto be
affectedby pH. In general, it appearedthat there was
less specificity at pH 9.0 than at pH 7.5. Generally, the
results showed that there was no pH dependence at the
P4 position.
Discussion and Summary
In the present study, we have used our Arg P1 positional
scanning library to characterize the substrate specificity
of a panel of serine peptidases, most of which were
already known to exhibit a preference for Arg at P1.
This work demonstrates that, with this library, these
serine peptidases show substrate specificity unique to
Figure 6. Cleavage of library substrates by human plasma thrombin at pH 9.0. Assays were conducted identically to those described in Figure 1
except that the pH was 9.0. Each bar represents meanÆstandard deviation for three separate assays.
Table 1. Summary of substrate preferences for enzymes tested(enzyme assays were carriedout as describedandreportedin Fig. 3)
Enzyme
Preference at P2
Preference at P3
Preference at P4
Factor Xa (human plasma)
Thrombin (human plasma)
Trypsin (human pancreatic)
Tryptase (human lung)
F, G, W, S
P then hydrophobic
Everything but d-Ala andR
Prefers N andother polar
neutral, andH
Basic, aromatic hydrophobic
Everything except A, G, P, N, D
Everything but P
Hydrophobic and R
Prefers J and other hydrophobic
Everything but G
Fairly broad, with significant activity
for all classes except acidic
Prefers K, R, then hydrophobic

S. T. Furlong et al. / Bioorg. Med. Chem. 10 (2002) 3637–3647
3643
Table 2. Substrate preferences for enzymes tested that produced no detectable cleavage using the substrate librarya
Enzyme
P1 Substrate preference
Enzyme
P1 Substrate preference
C1s
C1r
m-Calpain
Chymotrypsin
Arg
Arg
CVFBb/Factor B
Factor D
Plasmin
Arg
Arg
Lys/Arg
Tyr/Phe
Lys/Try, Arg/Lys
Tyr, Phe, Trp, Leu
Semenogelase
aAll enzymes shown were testedat concentrations equal to or higher than those reportedin other studies. Other than enzyme concentration, assay
conditions were identical to those used for enzymes that showed activity with this library.
Harris et al.,10 however, reportedthat Xa has a strong
preference for G at P2 andvery little preference at P
andP . In general, despite some discrepancies between
the studies reported here and those previous, there is an
overall remarkable degree of agreement.
PSA (preference for Phe at P1). The fact that no clea-
vage was observedwith these latter enzymes with the P
3
1
Arg library is consistent with expectations andprovides
further proof for the specificity of this library since none
of these enzymes prefer Arg at this position. A possible
explanation for why a particular enzyme did not work
with this library is that enzyme concentration was too
low. This explanation seems unlikely, however, since all
of the enzymes in this category were testedat con-
centrations at least ten-foldhigher than those reported
in other studies.
4
Two other serine peptidases tested included porcine
pancreatic trypsin andhuman lung tryptase. As might
be expectedfor a digestive enzyme, beyondthe require-
ment for R at P1, trypsin showedvery little preference
for a particular substrate. Trypsin did not efficiently
cleave substrates with d-Ala andR at P but the pre-
2
sence of any other amino acids at P2 made little differ-
ence. Similarly, trypsin cleavedsubstrates with P at P
In summary, in the present work we have synthesized
andcharacterizeda P Arginine substrate library. Phy-
3
1
andG at P 4 less efficiently, but otherwise little specificity
was observed. Tryptase, however, prefers N, other polar
neutral amino acids, and H at P2, a preference not pre-
viously described. Tryptase prefers K, R, and Q at P3
but also tolerates other hydrophobic residues. Little
sical analysis of this library yielded results in good
agreement with theoretical expectations. Biochemical
studies with four serine proteases demonstrated that, as
expected, experimental conditions, including salt and
pH, affect cleavage rates of library substrates. However,
determinations of substrate preferences for the enzymes
tested were largely independent of experimental condi-
tions confirming the usefulness of this andsimilar
libraries for predicting substrate preferences. Further
studies with additional enzymes may be required to
confirm these observations more broadly.
specificity was exhibitedat P with the exception of W,
4
E andD-Ala, which were not cleave.d Although a
complete analysis of tryptase specificity has not been
accomplishedprior to the present study, these results
were generally consistent both with synthetic substrates
known to be cleavedby this enzyme as well as a panel of
commercially available tripeptide synthetic substrates
testedfor the first time here (Fig. 2). For example, nat-
ural substrates include vasoactive intestinal peptide,
Experimental
11
fibrinogen, complement C3 andH-kininogen.
Analytical samples were homogeneous by TLC and
afforded spectroscopic results consistent with the
assignedstructures. Proton NMR spectra were obtained
using a Bruker AM-300 spectrometer. Chemical shifts
are reportedin parts per million ( d) relative to tetra-
methylsilane as internal standard. Mass spectra (MS)
were recorded on a Micromass VG electrospray instru-
ment (ES) or a Finnigan MAT-60 operating in the che-
mical ionization (APCI) mode (only peaks ꢁ10% of the
base peak are reported). Analytical thin-layer chroma-
tography (TLC) was conducted on pre-layered silica gel
GHLF plates (Analtech, Newark, DE, USA). Visuali-
zation of the plates was accomplishedusing UV light,
phosphomolybdic acid-ethanol and/or iodoplatinate
charring. Flash chromatography was conducted on
Kieselgel 60, 230–400 mesh (E. Merck, Darmstadt, West
Germany). Analytical HPLC was conducted on an HP
series 1050 instrument with a Vydac RP-C18 column.
Solvents were either reagent or HPLC grade. Reactions
were run at ambient temperature andunder a nitrogen
atmosphere unless otherwise noted. Solvent mixtures are
expressedas volume/volume ratios. Solutions were eva-
porated under reduced pressure on a rotary evaporator.
Other enzymes were testedwith this library that may
have been expected(basedon existing information) to
show activity with the Arg P1 library but did not (Table
2). In this category were enzymes of the complement
cascade including C1s, C1r, CVFBb, Factor B and
Factor D. For each of these enzymes, no cleavage was
observedeven at high concentrations of enzyme. How-
ever, these enzymes generally appear to be atypical ser-
ine peptidases for which inability to cleave small
synthetic substrates (or inefficient cleavage) has been
demonstrated.12 Thus, it is not surprising that cleavage
of library substrates couldnot be demonstratedhere.
Another enzyme testedthat might have been expected
to yieldsome cleavage but didnot was plasmin. This
enzyme has been reportedto have a preference for Lys
or Arg at P1 with an overall specificity similar to tryp-
sin. However, this plasmin is also known to be much
less efficient than trypsin andcleaves only some of these
bonds in proteins.13 Three other enzymes that were tes-
tedwith the library for which no cleavage was observed
included calpain (preference for Lys/Tyr at P1), chymo-
trypsin (preference for Tyr, Phe, Trp, or Leu at P1), and

3644
S. T. Furlong et al. / Bioorg. Med. Chem. 10 (2002) 3637–3647
All starting materials were commercially available
unless otherwise indicated. Enzymes were purchased
from Sigma, Calbiochem or Enzyme Systems Products.
The buffers andsalts were purchasedfrom Fisher Scien-
tific andthe plates usedwere Costar #3912 opaque white.
The following synthetic substrates were purchasedfrom
Sigma: Cbz-Arg-Arg-AMC, Boc-Glu-Lys-Lys-AMC,
Cbz-Phe-Arg-AMC hydrochloride, Boc-Nle-Gln-Leu-
Gly-Arg-AMC, Boc-Leu-Gly-Arg-AMC acetate, Boc-
Leu-Arg-Arg-AMC dihydrochloride, Boc-Ala-Gly-Pro-
Arg-AMC hydrochloride, Boc-Gln-Gly-Arg-AMC
hydrochloride, Boc-Gln-Arg-Arg-AMC acetate, Boc-
Leu-Lys-Arg-AMC hydrochloride, Boc-g-benzyl-Glu-
Gly-Arg, Boc-Val-Pro-Arg-AMC, Boc-Phe-Ser-Arg,
N-p-tosyl-Gly-Pro-Arg-AMC hydrochloride, Boc-Gln-
Ala-Arg-AMC hydrochloride, and Boc-b-benzyl-Asp-
Pro-Arg-AMC hydrochloride.
(150 mL). The solution was stirredat room temperature
for 16 h, diluted with EtOAc (300 mL), and washed with
saturatedNaHCO
(2 Â 200 mL), 1 N HCl (2 Â
3
200 mL), andbrine (1 Â 200 mL). The organic phase
was dried (Na2SO)4, filtered, and the solvent removed
under reduced pressure. The crude product was purified
by flash chromatography on silica gel eluting with ace-
tone/hexane (30:70) to afford 5 (8.21 g, 76%). HPLC
Rt=18.02 min. (RP C-18 at 210 nm, acetonitrile/water/
trifluoroacetic acid, (10:90:0.1); MS (ES) m/z=700
1
(M+1, base); H NMR (300 MHz, DMSO-d6/TFA) d
7.84 (s, 1H), 7.30 (d, 1H, J=8.7 Hz), 7.52 (d, 1H,
J=8.5 Hz), 7.33–7.44 (m, 10H), 6.27 (s, 1H), 5.28 (s,
2H), 5.13 (s, 2H), 3.90 (br s, 1H), 3.80 (br s, 2H), 2.42 (s,
3H), 1.70 (br s, 4H), 1.41 (s, 9H).
BOC-Arg-Aminomethylcoumarin (6). A solution of 5
(8.21 g, 11.73 mmol) in absolute ethanol (400 mL) was
treatedwith Pd(OH) 2 on carbon (1.23 g) followedby the
dropwise addition of 1 N HCl (12.9 mL, 12.9 mmol).
The reaction vessel was chargedwith hydrogen (50 psi)
andshook for 4 h on a Parr Shaker. The catalyst was
filteredandthe filter cake was washedwith absolute
ethanol (4 Â 50 mL). The solvent was removedunder
reduced pressure to afford 6 (5.10 g, 93%), which was
usedwithout further purification. HPLC Rt=8.44 min.
(RP C-18 at 210 nm, acetonitrile/water/trifluoroacetic,
2,3,5-Trimethylphenoxyethylacetate (2). A solution of
2,3,5-trimethylphenol (12 g, 88 mmol) in DMF (50 mL)
was treatedwith K 2CO3 (13.4 g, 97 mmol). After stirring
at room temperature for 15 min, ethylbromoacetate
(11.73 mL, 106 mmol) was added dropwise. The solution
was stirredfor 16 h at room temperature andthe solvent
removed under reduced pressure. The residue was dis-
solvedin EtOAc (500 mL), washedwith 1 N HCl (2
Â
250 mL), 1 N NaOH (2 Â 250 mL), andthen brine (1 Â
250 mL). The organic phase was dried (Na2SO4), fil-
tered, and the solvent removed under reduced pressure.
The crude product was purified by flash chromato-
graphy on silica gel eluting with ethyl acetate/hexane
(5:95) to afford 2 (13.4 g, 68%). HPLC Rt=17.3 min
(RP C-18 at 210 nm, acetonitrile/water/trifluoroacetic
acid; 10:90:0.1); MS (APCI) m/z=223 (M+1, base); 1H
NMR (300 MHz, DMSO-d6/TFA) d 6.77 (s, 1H), 6.51
(s, 1H), 4.79 (s, 2H), 4.21 (q, 2H, J=7.2 Hz), 2.21 (s,
3H), 2.18 (s, 3H), 2.08 (s, 3H), 1.25 (t, 3H, J=7.2 Hz).
1
10:90:0.1); MS (ES) m/z=432 (M+1, base); H NMR
(300 MHz, DMSO-d6/TFA) d 7.92 (s, 1H), 7.86 (d, 1H,
J=8.7 Hz), 7.55 (d, 1H, J=8.5 Hz), 6.27 (s, 1H), 4.16
(br s, 1H), 3.18 (br t, 2H), 2.42 (s, 3H), 1.20–1.80 (br m,
4H), 1.41 (s, 9H).
BOC-Arg(CMtr)-Aminomethylcoumarin (7). A suspen-
sion of 6 (8.35 g, 17.84 mmol) in anhydrous acetone
(200 mL) was treatedwith 1 N NaOH (35.7 mL,
35.68 mmol). DMF (25 mL) was added to increase
solubility. Sulfonyl chloride 3 (6.01 g) in anhydrous
acetone (160 mL) was added and the cloudy pale yellow
solution was stirredat room temperature for 30 min.
The reaction mixture was adjusted to pH 4 with 1 N
HCl andthe proudct was extractedwith EtOAc
(500 mL). The organic extract was washedwith 1 N HCl
(2 Â 300 mL) andbrine (1 Â 100 mL). The organic
phase was dried (MgSO4), filtered, and the solvent
removed under reduced pressure. The crude material
was purifiedby flash chromatography on silica gel
eluting with EtOAc/CH2Cl2 (1:1) to affordthe inter-
mediate ester (1.67 g, 20%). HPLC Rt=14.98 min. (RP
C-18 at 210 nm, acetonitrile/water/trifluoroacetic acid,
10:90:0.1); MS (ES) m/z=716 (M+1), 660 (M-55,
base), 616; 1H NMR (300 MHz, DMSO-d6/TFA) d 7.97
(s, 1H), 7.80 (d, 1H, J=8.7 Hz), 7.52 (d, 1H, J=8.5 Hz),
6.60 (br s, 1H), 6.27 (s, 1H), 4.80 (s, 2H), 4.13 (q, 2H,
J=6.5 Hz), 4.07 (br t, 1H), 3.12 (br t, 2H), 2.59 (s, 3H),
2.55 (s, 3H), 2.42 (s, 3H), 2.13 (s, 3H) 1.20–1.80 (br m,
4H), 1.41 (s, 9H), 1.26 (t, 3H, J=6.3 Hz). The inter-
mediate ester (2.35 g, 3.28 mmol) in absolute ethanol/
THF (66 mL, 1:1) was treatedwith 1 N NaOH (3.45 mL,
3.45 mmol). The solution was stirredat room tempera-
ture for 3 h and the solvent was removed under reduced
pressure. The gummy residue was redissolved in water
(10 mL) with a minimum amount of MeOH (0.5 mL) to
2,3,5-Trimethylphenoxyethylacetate-4-sulfonyl chloride
(3). A solution of 2 (500 mg, 2.25 mmol) in dichloro-
methane (10 mL) was cooledto 4 ꢀC followedby the
addition of chlorosulphonic acid (0.37 mL, 5.63 mmol).
The cloudy pink solution was stirred for 1 h at 4 ꢀC.
Coldwater (50 mL) was addedandthe mixture extrac-
tedwith dichloromethane (3 Â 50 mL). The combined
organic extracts were washedwith water (2 Â 50 mL)
andbrine (1 Â 50 mL). The organic phase was dried
(MgSO4), filtere,d andthe solvent removedunedr
reduced pressure affording 3 (648 mg, 90%). HPLC
Rt=5.5 min (RP C-18 at 210 nm, acetonitrile/water/tri-
fluoroacetic acid; 10:90:0.1). 1H NMR (300 MHz,
DMSO-d6/TFA) d 6.77 (s, 1H), 4.79 (s, 2H), 4.15 (q,
2H, J=7.2 Hz), 2.55 (s, 6H), 2.12 (s, 3H), 1.22 (t, 3H,
J=7.2 Hz).
BOC-Arg(di-CBZ)-Aminomethylcoumarin (5). A solu-
tion of BOC-Arg(di-CBZ)-OH (25 g, 46.07 mmol) in
THF (150 mL) was treated dropwise with anhydrous
pyridine (3.72 mL, 46.07 mmol) followed by the addition
of di-tert-butyl dicarbonate (10.05 g, 46.07 mmol). The
solution was stirredfor 1 h at room temperature to form
the activated ester, followed by the addition of 7-amino-
4-methylcoumarin
4 (2.70 g, 15.36 mmol) in THF

S. T. Furlong et al. / Bioorg. Med. Chem. 10 (2002) 3637–3647
3645
insure homogeneity. The solution was acidified to pH 4
with 1 N HCl. The resulting white powder was filtered and
driedin a heated(50 ꢀC) vacuum oven for 16h to yield
pure 7 (1.86 g, 82%). HPLC Rt=11.43min. (RP C-18 at
210 nm, acetonitrile/water/trifluoroacetic acid, 10:90:0.1);
present on the resin. To cap these free amino groups, a
5 mL aliquot of a reaction cocktail consisting of acetic
anhydride (5.30 mL, 55.4 mmol) and diisopropylethyla-
mine (9.7 mL, 55.4 mmol) in DMF (10 mL) was added
to the resin andagitatedwith positive nitrogen gas
pressure for 5 h. The reaction mixture was drained and
the resin washedwith DMF (3 Â 20 mL). Another 5 mL
of the acetic anhydride reaction cocktail was added to
the resin followedby agitation for 16 h with positive
nitrogen gas pressure. The resin was drained and washed
with DMF (8 Â 10 mL), CH2Cl2 (5 Â 10 mL), ether (3 Â
20 mL), anddriedwith positive nitrogen gas to afford
Boc-Arg(CMtr)-AMC resin 9. The Kaiser test was
performedon a small aliquot of resin beads andgave a
pale orange color indicating no free amine was present.
The Boc-Arg(CMtr)-AMC resin was swelledin 20mL
CH2Cl2 for 15min andthe solvent rdaine.d Tri-
fluoroacetic acid/CH2Cl2 (20mL, 2:3) was added to the
resin followedby agitation for 1 h with positive nitrogen
gas. The reaction mixture was drained and resin washed
with CH2Cl2 (8 Â 20mL), ether (5 Â 20mL) anddried
with nitrogen gas to affordH-Arg(CMtr)-AMC resin 9.
The Kaiser test was performedon a small aliquot of
resin beads giving a dark blue color indicating free
amine was present.
1
MS (ES) m/z=688 (M+1), 632 (M-55, base), 588; H
NMR (300 MHz, DMSO-d6/TFA) d 7.83 (s, 1H), 7.75
(d, 1H, J=8.7 Hz), 7.52 (d, 1H, J=8.5 Hz), 6.60 (br s,
1H), 6.32 (s, 1H), 4.80 (s, 2H), 4.16 (br s, 1H), 3.12 (br t,
2H), 2.59 (s, 3H), 2.55 (s, 3H), 2.42 (s, 3H), 2.13 (s, 3H)
1.20–1.80 (br m, 4H), 1.41 (s, 9H).
ꢀ-Ala-Aminomethyl-AgroPoreTM resin (8). Amino-
methyl AgroPoreTM resin (3.85 g, 2.77 mmol,
0.72 mmol/g resin loading) was swelled in 50 mL of
DMF for 15 min andthen the solvent was drained. BOC-
b-Alanine (2.67 g, 14.1 mmol), HATU (5.27 g, 13.8 mmol),
andidisopropylethylamine (4.83 mL, 27.7 mmol) were
dissolved in 20 mL of DMF and 10 mL of this reaction
cocktail was added to the resin and agitated with posi-
tive nitrogen gas pressure for 2 h. The reaction mixture
was drained and the resin was washed with DMF (2 Â
20 mL). Another 10 mL of the reaction cocktail was
added to the resin and the resin agitated for 16 h with
positive nitrogen gas pressure to complete the double
coupling. The reaction mixture was drained and the
resin washedwith DMF (8 Â 20 mL). The resin was
treated with acetic anhydride (15.7 mL, 166 mmol) and
diisopropylethylamine (28.9 mL, 166 mmol) in DMF
(30 mL) andagitatedwith nitrogen gas for 2 h to endcap
any residual free NH2 sites on the resin. The reaction
mixture was drained and the resin washed with DMF (8
Library synthesis example: Ac-O-X-O-Arg-AMC. The
resin 9 (6.00 g, 4.32 mmol) was divided equally into 20
reaction vessels, swelledwith 10 mL DMF for 15 min
and the solvent was drained. A reaction cocktail con-
sisting of the isokinetic mixture of 20 amino acids (var-
ious wt, 35.3 mmol of each amino acid), HATU (13.7 g,
33.5 mmol), anidisdopropylethylamine (12.3 mL,
70.5 mmol) in DMF (88 mL) was prepared. An aliquot
of the reaction cocktail (2.5 mL) was added to each of
the 20 reaction pots containing resin 9 andthe resin
agitatedwith positive nitrogen gas pressure for 4 h. The
reaction mixture was drained and the resin washed with
DMF (2 Â 5 mL). Another 2.5 mL of the isokinetic
amino acidreaction cocktail was addedto the resin and
the resin agitatedfor 16 h with positive nitrogen gas
pressure to complete the double coupling. The reaction
mixture was drained and the resin washed with DMF (5
 5 mL). The Kaiser test was performedon a small ali-
quot of resin beads giving a dark orange color. A solu-
tion of piperidine/DMF (3 mL, 1:4) was added to each
reaction vessel andthe resin agitatedwith positive
nitrogen pressure for 10 min. The reaction mixture was
drained and the resin washed with DMF (3 Â 5 mL). A
solution of piperidine/DMF (3 mL, 1:4) was then added
to resin andthe resin agitatedfor 10 min with nitrogen
gas. The reaction mixture was drained and the resin
washedwith DMF (8 Â 5 mL) affording H-O-Arg-
AMC resin. The Kaiser test was performedon a small
aliquot of resin beads giving a dark blue color indicat-
ing free amine was present andFmoc cleavage was
complete.
 10 mL) andCH Cl2 (5  10 mL) affording BOC-b-
2
Ala-aminomethyl-AgroPoreTM resin. The Kaiser test
was performedon a small aliquot of resin beads andgave
a pale orange color indicating no free amine was present.
This resin was then treatedwith trifluoroacetic acid/
CH2Cl2 (20 mL, 2:3) andthe resin agitatedfor 30 min
with positive nitrogen gas. The reaction mixture was
drained and the resin washed with CH2Cl2 (8 Â 20mL),
ether (3 Â 20mL) anddriedwith nitrogen gas to afford
b-Ala-aminomethyl-AgroPoreTM resin 8. The Kaiser test
was performedon a small aliquot of resin beads giving a
dark blue color indicating free amine was present.
H-Arg(CMtr)-AMC Resin (9). b-Ala-Aminomethyl-
AgroPoreTM resin 8 (3.85, 2.77 mmol) was swelledin
DMF (50 mL) for 15 min and the solvent drained. A
solution of acid 7 (2.00 g, 2.91 mmol), HATU (1.07 g,
2.83 mmol), anidisdopropylethylamine (1.00 mL,
5.65 mmol) in DMF (5 mL) was added to the resin and
agitatedwith positive nitrogen gas pressure for 16 h.
The reaction mixture was drained and the resin washed
with DMF (5 Â 20 mL). A secondcoupling was per-
formed by adding to the resin a solution of acid 7
(481 mg, 0.700 mmol), HOBT (86 mg, 0.64 mmol),
diisopropylcarbodiimide (160 mg, 1.27 mmol), and
DMAP (17 mg, 0.14 mmol) in DMF (5 mL) andthen
agitating the resin with positive nitrogen gas pressure
for 16 h. The reaction mixture was drained and the resin
washedwith DMF (5 Â 20 mL). The Kaiser test was
performedon a small aliquot of resin beads to give a
very pale blue color indicating some free amine is still
A reaction cocktail was preparedby placing the 20 dis-
tinct amino acids, X, (various wts, 1.73 mmol) in 20
distinct vessels, along with HATU (624 mg, 1.64 mmol),
diisopropylethylamine (0.60 mL, 3.46 mmol) and DMF
(6 mL). 3 mL of the 20 distinct reaction cocktails was

3646
S. T. Furlong et al. / Bioorg. Med. Chem. 10 (2002) 3637–3647
then added to each of the 20 reaction pots containing
the resin from the step above andthe resin agitatedwith
positive nitrogen gas pressure for 4 h. The reaction
mixture was drained and the resin washed with DMF (4
 5 mL). Another 3 mL of the distinct amino acid reac-
tion cocktail was added to the resin and the resin agi-
tatedfor 16 h with positive nitrogen gas pressure to
complete the double coupling. The reaction mixture was
drained and the resin washed with DMF (8 Â 5 mL).
The Kaiser test was performedon a small aliquot of
resin beads, giving a dark orange color. A solution of
piperidine/DMF (3 mL 1:4) was added to each reaction
vessel andthe resin agitatedwith positive nitrogen
pressure for 10 min, the reaction mixture was drained
andthe resin washedwith DMF (3 Â 5 mL). Another
solution of piperidine/DMF (3 mL, 1:4) was added to
each reaction vessel andthe resin agitatedwith positive
nitrogen pressure for 10 min, the reaction mixture
drained and the resin washed with DMF (3 Â 5 mL) to
give H-X-O-Arg-AMC resin. The Kaiser test was per-
formed on a small aliquot of resin beads giving a dark
blue color indicating free amine was present and the
Fmoc cleavage was complete. The isokinetic amino acid
coupling and the FMOC cleavage procedure described
in the first step of the library synthesis was repeatedto
afford H-O-X-O-Arg-AMC resin. A reaction cocktail of
acetic anhydride (16.4 mL, 172.8 mmol) and diisopro-
pylethylamine (30 mL, 172.8 mmol) in DMF (80 mL)
was prepared. 3.1 mL of this reaction cocktail was
added to each of the 20 reaction vessels and the resin
agitatedwith positive nitrogen gas pressure for 1.5 h.
The reaction mixture was drained and the resin washed
with DMF (3 Â 5 mL). Another 3.1 mL of the reaction
cocktail was added to the reaction vessels and the resin
agitatedfor 1 h with positive nitrogen gas pressure to
complete the resin-endcapping procedure. The resin was
drained and washed with DMF (8 Â 5 mL), CH2Cl2 (5
 5 mL), ether (3  5 mL), anddriedwith positive
nitrogen gas. The Kaiser test was performedon a small
aliquot of resin beads giving a pale orange color indi-
cating no free amine is present andcoupling is complete
to give Ac-O-X-O-Arg-AMC resin.
mL, mixed1:1 with matrix and1 mL depositedon the
target. The matrix solution was 10 mg/mL alpha-cyano-
3-hydroxycinnamic acid in 2:2:1 acetonitrile/ethanol/
water. The mass range was calibratedto 7000 Da. A
nominal number of scans were acquiredwith each
recorded spectrum consisting of an average of 10 laser
shots. Data were combined, smoothed and background
subtractedto produce the final spectrum.
For the simulatedmass spectrum, the molecular mass of
the peptides in a library resulting from all possible
combinations of 20 amino acids is calculated using the
averagedatomic mass. Accoridng to the isokinetic
assumption, each peptide in the library is given equal
weight, anda MALDI spectrum for a library is simply
the distribution of molecular mass of the peptides in the
library. The computation is carriedout with a computer
program MALDIsym, which is available upon request.
Enzyme assays. All fluorescent measurements were con-
ducted on a Perkin–Elmer HTS 7000 with the following
parameters: Excitation filter 360 nm; Emission 465 nm;
Gain 40–80; Number of flashes 20; Integration time
1000 ms; intervals between measurements were typically
1 min. The substrates were preparedat a concentration
of 1.0 mM in DMSO. The enzymes were preparedin
buffer (100 mM HEPES, 20 mM NaCl, 10 mM MgCl2
pH 7.5). Enzyme concentrations were as follows: trypsin,
0.11 mg/well; thrombin, 10.1 U/well; chymotrypsin,
2.5 mg/well; human urine kallikrein, 2.0 mU/well; human
plasma kallikrein, 5.0 mU/well; porcine pancreatic kal-
likrein, 0.1 ng/well; PSA, 0.05–0.5 mg/well; calpain,
0.013–0.05 U/well; human lung tryptase, 40 ng/well;
coagulation factor Xa, 0.2 U/well. The standard buffer
defined above was used for most experiments. However,
for some experiments the effect of different pH or salt
conditions was also tested. 80 mL of buffer and10 mL of
substrate were added to appropriate wells in the plate
and mixed. The reaction was initiated by the addition of
10 mL of enzyme andthe change in fluorescent intensity
at ex/em 360/465 nm was measuredfor 30 min at 25 ꢀC.
The slope was determined by linear regression and the
results expressedas percent maximum observedrate for
each enzyme andsubstrate library.
A cleavage cocktail consisting of thioanisole (17.6 mL),
phenol (1.07 g), andtriisopropylsilane (5 mL) in tri-
fluoroacetic acid(100 mL) was preparedand4 mL of
the cleavage cocktail was added to each of the 20 reac-
Acknowledgements
tion pots andthe pots cooledto 0
ꢀC. Bromo-
trimethylsilane14 (0.66 mL) was added to the 20 reaction
pots at 0 ꢀC andshaken gently for 0.5 h, allowedto
warm to room temperature and shaken for an addi-
tional 3 h. The solvents were removed under reduced
pressure andthe residue trituratedwith ether (5 mL)
The authors gratefully acknowledge Kara Pine and Mark
Sylvester for technical assistance, Clay Scott for his helpful
discussions andsupport andEileen Furlong for editing.
andfilteredaffording the
Ac-O-X-O-AMC library as
References and Notes
yellow powders (ꢂ85 mg from each pot). The Ac-O-O-
X-AMC and Ac-X-O-O-AMC libraries were also syn-
thesizedusing the above procedures.
1. (a) Harris, J. L.; Peterson, E. P.; Hudig, D.; Thornberry,
N. A.; Craik, C. S. J. Biol. Chem. 1998, 273, 27364. (b) Harris,
J. L.; Backes, B. J.; Leonetti, F.; Mahrus, S.; Ellman, J. A.;
Craik, C. S. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 7754. (c)
Takeuchi, T.; Harris, J. L.; Huang, W.; Yan, K. W.; Coughlin,
S. R.; Craik, C. S. J. Biol. Chem. 2000, 275, 26333. (d) Yang,
C. F.; Porter, E. S.; Boths, J.; Kanyi, D.; Hsieh, M.; Cooper-
man, B. S. J. Pept. Res. 1999, 54, 444.
Mass spectral analysis of libraries. Analyses were per-
formedon a Micromass (VG) TofSpec SE operating
with a flat plate target, pulsedion extraction anda
reflectron analyzer. Samples were dissolved in 2:2:1
acetonitrile/ethanol/water to approximately 10 pmol/
2. Edwards, P. D.; Mauger, R. C.; Cottrell, K. M.; Morris,

S. T. Furlong et al. / Bioorg. Med. Chem. 10 (2002) 3637–3647
3647
F. X.; Pine, K. K.; Sylvester, M. A.; Scott, C. W.; Furlong,
S. T. Bioorg. Med. Chem. Lett. 2000, 10, 2291.
11. (a) Caughey, G. H.; Leidig, F.; Viro, N. F.; Nadel, J. A.
J. Pharmacol. Exp. Ther. 1988, 244, 133. (b) Maier, M.;
Spragg, J.; Schwartz, L. B. J. Immunol. 1983, 130, 2352. (c)
Schwartz, L. B.; Bradford, T. R.; Littman, B. H.; Wintroub,
B. U. J. Immunol. 1985, 135, 2762. (d) Schwartz, L. B.;
Kawahara, M. S.; Hugli, T. E.; Vik, D.; Fearon, D. T.; Aus-
ten, K. F. J. Immunol. 1983, 130, 1891.
12. (a) Furlong, S. T.; Dutta, A. S.; Coath, M. M.; Gormley,
J. J.; Hubbs, S. J.; Lloyd, D.; Mauger, R. C.; Strimpler, A. M.;
Sylvester, M. A.; Scott, C. W.; Edwards, P. D. Immuno-
pharmacology 2000, 48, 199. (b) Volanakis, J. E.; Narayana,
S. V. Protein Sci. 1996, 5, 553. (c) Xu, Y.; Circolo, A.; Jing,
H.; Wang, Y.; Narayana, S. V.; Volanakis, J. E. J. Biol. Chem.
2000, 275, 378. (d) Jing, H.; Babu, Y. S.; Moore, D.; Kilpa-
trick, J. M.; Liu, X. Y.; Volanakis, J. E.; Narayana, S. V.
J. Mol. Biol. 1998, 282, 1061.
3. (a) Rano, T. A.; Timkey, T.; Peterson, E. P.; Rotonda, J.;
Nicholson, D. W.; Becker, J. W.; Chapman, K. T.; Thorn-
berry, N. A. Chem. Biol. 1997, 4, 149. (b) Kang, S. J.; Wang,
S.; Hara, H.; Peterson, E. P.; Namura, S.; Amin-Hanjani, S.;
Huang, Z.; Srinivasan, A.; Tomaselli, K. J.; Thornberry, N. A.;
Moskowitz, M. A.; Yuan, J. J. Cell Biol. 2000, 149, 613.
4. Herman, L. W.; Tarr, G.; Kates, S. A. Mol. Divers. 1996, 2,
147.
5. (a) Josey, J. A.; Tarlton, C. A.; Payne, C. E. Tetrahedron
Lett. 1998, 39, 5899. (b) Roussell, P.; Bradley, M.; Matthews,
I.; Kane, P. Tetrahedron Lett. 1997, 38, 4861. (c) Bonnat, M.;
Bradley, M.; Kilburn, J. Tetrahedron Lett. 1996, 37, 5409.
6. Zhong, H. M.; Greco, M. N.; Maryanoff, B. E. J. Org.
Chem. 1997, 62, 9326.
7. Urban, J.; Dattilo, J. W. Abstracts of Papers; 15th Amer-
ican Peptide Symposium, Nashville, TN, 14–19 June 1997.
8. Backes, B. J.; Harris, J. L.; Leonetti, F.; Craik, C. S.; Ell-
man, J. A. Nat. Biotechnol. 2000, 18, 187.
9. McRae, B. J.; Kurachi, K.; Heimark, R. L.; Fujikawa, K.;
Davie, E. W.; Powers, J. C. Biochemistry 1981, 20, 7196.
10. Harris, J. L.; Backes, B. J.; Leonetti, F.; Mahrus, S.; Ellman,
J. A.; Craik, C. S. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 7754.
13. (a) Castellino, F. J. In Handbook of Proteolytic Enzymes;
Barrett, A. J., Rawlings, N. D., Woessner, J. F., Eds.; Aca-
demic: New York, 1998, pp 190–199. (b) Di Cera, E.; Guinto,
E. R.; Vindigni, A.; Dang, Q. D.; Ayala, Y. M.; Wuyi, M.;
Tulinsky, A. J. Biol. Chem. 1995, 270, 22089.
14. (a) Hughes, J. L.; Leopold, E. J. Tetrahedron Lett. 1993,
48, 7713. (b) Guo, L.; Funakoshi, S.; Fujii, N.; Yajima, H.
Chem. Pharm. Bull. 1988, 36, 4989.