Peptide microarrays for the determination of protease substrate specificity

Article abstract of DOI:10.1021/ja027477q

A method is described for the preparation of substrate microarrays that allow for the rapid determination of protease substrate specificity. Peptidyl coumarin substrates, synthesized on solid support using standard techniques, are printed onto glass slide

Full text of DOI:10.1021/ja027477q

Published on Web 11/22/2002
Peptide Microarrays for the Determination of Protease Substrate Specificity
Cleo M. Salisbury, Dustin J. Maly, and Jonathan A. Ellman*
Center for New Directions in Organic Synthesis, Department of Chemistry, UniVersity of California,
Berkeley, California 94720
Received June 26, 2002
Proteases, which hydrolyze amide bonds in peptides and proteins,
are abundant in nature and essential for cellular function and
viability. Many important biological pathways, such as hormone
activation and apoptosis, are dependent on the action of proteases.
Furthermore, proteases are involved in diverse disease states,
including cardiovascular disease, cancer, AIDS, and Alzheimer’s
disease, making these enzymes important therapeutic targets.1
Significantly, many proteases have recently been identified through
genome sequencing efforts. Only a fraction of the estimated 2% of
genes that encode for proteases in organisms from E. coli to humans
2
have been studied to date. Establishing the functional roles of these
proteases will greatly enhance our understanding of biological
systems and will likely provide a number of important new
Figure 1. Fluorogenic substrate microarrays.
therapeutic targets.
The ability of proteases to selectively cleave specific substrates
in the presence of many other potential substrates is essential for
their function. Knowledge of the substrate specificity of a protease
can therefore be used to facilitate the identification of its physi-
ological substrates, which is essential for defining its biological
function. Moreover, determination of substrate specificity greatly
aids in the design of potent and selective substrates and inhibitors.
Although a number of combinatorial methods have been described
aldehyde-derivatized surface and alkoxylamine-functionalized sub-
strates (Figure 1). These alkoxylamine-derivatized peptide substrates
show kinetic constants comparable to those of the underivatized
ACC substrates in solution (see Supporting Information).
The alkoxylamine-substituted fluorogenic substrates are prepared
by first loading the latent alkoxylamine functionality onto solid
support by reductive amination of monoprotected alkoxylamine 1
with the backbone amide linker (BAL) aldehyde resin 2 (Scheme
3
7
for the rapid determination of substrate specificity, they do not
1). The resultant secondary amine is then acylated with Fmoc-
provide a full specificity fingerprint because they do not allow for
kinetic evaluation of all of the prepared substrate sequences. Herein
we report the synthesis and evaluation of microarrays of fluorogenic
substrates that enable the rapid and complete characterization of
each substrate in the library.
Gly-OH activated with HATU. After Fmoc deprotection to provide
3, Fmoc-protected ACC is then added using standard coupling
conditions. The Fmoc protecting group is removed to yield
intermediate 4, and the resulting poorly nucleophilic free aniline is
acylated using the symmetric anhydride of the desired P amino
1
We have recently reported the development of the fluorogenic
peptidyl coumarin substrates, 7-amino-4-carbamoylmethyl coumarin
acid residue to yield intermediate 5. Any remaining unreacted
aniline is then capped with the nitrotriazole ester of acetic acid.
These conditions were found to give efficient loading of the P1
residue with low background fluorescence due to the efficient
capping of the unreacted coumarin. After the loading of the P1
residue, the desired peptide is constructed using standard Fmoc-
based solid-phase synthesis techniques. The desired alkoxylamine
fluorogenic peptide substrate 6 is cleaved from solid support using
trifluoroacetic acid.
(
ACC) peptides, which are effective tools for determining the
4
P-side (N-terminal) substrate specificity of serine and cysteine
proteases. Proteolytic cleavage at the peptide anilide bond liberates
5
the highly fluorescent coumarin leaving group, allowing for the
determination of cleavage rates through the increase in fluorescence.
Importantly, the bifunctional nature of ACC enables the fluorophore
to be linked to solid support through its C-terminus, and thus
libraries of fluorogenic substrates of any amino acid sequence can
be rapidly prepared using standard solid-phase peptide synthesis
methods. Positional scanning-substrate combinatorial libraries (PS-
SCLs) of peptidyl coumarins have proven to be very useful for
determining substrate specificity trends;3a-c,5 however, the mixture-
based nature of these libraries does not allow for cooperative
interactions between substrate residues to be determined.3n To fully
map the N-terminal specificity of a protease, we have applied
libraries of peptidyl coumarins to microarrays in a spatially
addressed fashion. The peptides are linked to glass slide microarray
A number of conditions were tested for oxime formation on
aldehyde-derivatized glass microscope slides. A concentration of
200-500 µM peptide in a pH 5.2 acetate buffer with 1-5%
dimethyl sulfoxide (DMSO) provided the best results. Higher
concentrations of peptide and DMSO were found to cause spot
spreading, and other buffer combinations did not lead to efficient
oxime formation. Fluorogenic ACC-peptides were first printed on
commercially available aldehyde-functionalized slides, but were
inefficiently cleaved by trypsin. A layer of bovine serum albumin
(BSA) has been shown to increase the accessibility of slide-bound
6
8
surfaces via a chemoselective oxime forming reaction, using an
peptides to enzymes. We therefore chose to derivatize BSA-covered
slides with 4-carboxybenzaldehyde, using standard carbodiimide
coupling conditions. Substrates bound to the BSA-aldehyde surface
*
To whom correspondence should be addressed. E-mail: jellman@
uclink.berkeley.edu.
1
4868 J. AM. CHEM. SOC. 2002, 124, 14868-14870
9
10.1021/ja027477q CCC: $22.00 © 2002 American Chemical Society

C O MMU N I C A T I O N S
Scheme 1. Substrate Synthesisa
a
(i) (a) NaBH3CN, DMF; (b) Fmoc-Gly-OH, DMF, HATU, collidine;
(
(
c) 20% piperidine in DMF. (ii) (a) Fmoc-ACC-OH, DIC, HOBt, DMF;
b) AcOH, DIC, HOBt, DMF; (c) 20% piperidine in DMF. (iii) (a)
Figure 2. Proteolytic cleavage of array-bound substrates. Panel A shows
a schematic of the array. Panels B-E show the fluorescence of the array
prior to cleavage (B), and after treatment with trypsin (C), granzyme B
Symmetric anhydride of P1 amino acid, DMF/dichloroethane (DCE); (b)
AcOH, DIC, DMF, nitrotriazole.
(D), and thrombin (E).
demonstrated an enhanced rate of cleavage by trypsin, reflecting
the greater accessibility of the substrates presented on this surface
and/or the increased stability of the enzyme.
To further demonstrate that array-bound fluorogenic substrates
are accessible for proteolytic cleavage, substrate arrays were
subjected to proteolysis by a variety of serine proteases. A split
pin arrayer was used to spot multiple copies of three different
protease substrates on the array surface at a spacing of 314 µm
These fluorescence differences are readily detected by standard
microarray scanning instrumentation. As little as 1% cleaved
substrate (unacylated ACC) can be detected, and the amount of
unacylated ACC is linearly proportional to the fluorescence (see
Supporting Information).
A library of substrates was next synthesized, printed, and assayed
to demonstrate the utility of peptide-ACC microarrays for the rapid
determination of protease substrate specificity. The 361-member
spatially separate library contained a fixed P -Lys and P -Ala with
(Figure 2). Specifically, two thrombin substrates with differing
1
4
specificity constants (kcat/K ), Ac-Leu-Gly-Pro-Lys-ACC-linker and
m
all combinations of proteinogenic amino acids (except cysteine) at
the P2 and P3 sites. To serve as standards for quantitation of
fluorescence, multiple copies of unacylated ACC and acetyl-capped
ACC were also printed on the array surface. Using a spacing of
440 µm, we printed the entire library and the control compounds
in duplicate in a 1.8 cm × 0.9 cm area, representing 800 individual
Ac-Nle-Thr-Pro-Lys-ACC-linker, and a known granzyme B sub-
strate, Ac-Ile-Glu-Pro-Asp-ACC-linker, were printed. To serve as
a fluorescence intensity standard, unacylated ACC-linker was also
printed. The general accessibility of the array to enzymatic reactions
was first demonstrated by exposing the array to trypsin (Figure
2
2
c). As expected, the two thrombin substrates with a P
showed a comparable rate of cleavage, whereas the granzyme B
substrate with a P -Asp residue showed no detectable cleavage.
1
-Lys residue
data points in an area of less than 1.7 cm . The array was then
used to determine the extended substrate specificity of the serine
protease thrombin by adding 15 µL of a 250 nM solution of the
enzyme to the array surface followed by incubation for 60 min.
The quantitative results of the assay are shown in Figure 3. The
1
Next, the array surface was subjected to the serine protease
granzyme B (Figure 2d). The array showed cleavage of the
granzyme B substrate with no detectable cleavage of either of the
thrombin substrates, demonstrating that enzymes maintain fidelity
for their desired substrate when affixed to the surface. To
demonstrate that subtle differences in extended substrate binding
by a protease could be detected, the array was subjected to cleavage
with the serine protease thrombin. The two thrombin substrates were
readily cleaved, whereas the granzyme B substrate showed no
detectable cleavage. As can be seen from Figure 2e, a significant
difference in the extent of cleavage of the two thrombin-susceptible
substrates could be detected by the greater increase in fluorescence
for the preferred substrate (Ac-Nle-Thr-Pro-Lys-ACC-linker) over
the less preferred substrate (Ac-Leu-Gly-Pro-Lys-ACC-linker).
2
data show a strong preference for proline at the P position and
broad specificity at the P position with basic, polar, and hydro-
3
phobic residues all being tolerated. This specificity profile is
consistent with prior studies using PS-SCLs and previously known
specificity determinants.9
To validate our results, we synthesized, purified, and assayed
four of these peptidyl coumarin substrates in solution (Table 1).
m
As the data show, the kcat/K for the substrates on the array is
10
predictive of the solution-phase results. Of note is that consider-
ably less enzyme was needed to determine this specificity profile
on the array relative to prior studies, due to the miniaturized nature
of the array and the increased sensitivity of fluorescence detection
J. AM. CHEM. SOC.
9
VOL. 124, NO. 50, 2002 14869

C O MMU N I C A T I O N S
Acknowledgment. We thank Dr. Adam Carroll for assistance
with the microarray printing and insightful discussions, Patty
Holman and Trent Su for the assistance with the microarray
scanning, and Sandra M. Waugh for the gift of recombinant
granzyme B. This work was supported by the NIH (grant #
GM54051). The Center for New Directions in Organic Synthesis
is supported by Bristol-Myers Squibb as a Sponsoring Member and
Novartis as a Supporting Member.
Supporting Information Available: Complete protocols for single
substrate and library syntheses, slide preparation, microarray printing,
enzyme assays, and kinetic analyses of microarrays and single
substrates, and microarray controls (PDF). This material is available
free of charge via the Internet at http://pubs.acs.org.
References
Figure 3. Characterization of the substrate specificity of thrombin using
substrates of the general structure Ac-Ala-P3-P2-Lys-linker. Each square is
colored in proportion to the quantitated fluorescence intensity of the
corresponding substrate after treatment with thrombin, corresponding to the
relative amount of cleavage (normalized data and 3D bar graph available
in Supporting Information).
(1) Leung, D.; Abbenante, G.; Fairlie, D. P. J. Med. Chem. 2000, 43, 305.
(
2) Rawlings, N. D.; O’Brien, E.; Barrett, A. J. Nucleic Acids Res. 2002, 30,
343.
(
3) (a) For a recent review, see: Maly, D. J.; Huang, L.; Ellman, J. A.
ChemBioChem 2002, 3, 16. (b) Thornberry, N. A.; Rano, T. A.; Peterson,
E. P.; Rasper, D. M.; Timkey, T.; Garcia-Calvo, M.; Houtzager, V. M.;
Nordstrom, P. A.; Roy, S.; Vaillancourt, J. P.; Chapman, K. T.; Nicholson,
D. W. J. Biol. Chem. 1997, 272, 17907. (c) Harris, J. L.; Niles, A.; Burdick,
K.; Maffitt, M.; Backes, B. J.; Ellman, J. A.; Kuntz, I.; Haak-Frendscho,
M.; Craik, C. S. J. Biol. Chem. 2001, 276, 34941. (d) Matthews, D. J.;
Wells, J. A. Science 1993, 260, 1113. (e) Smith, M. M.; Shi, L.; Navre,
M. J. Biol. Chem. 1995, 270, 6440. (f) Turk, B. E.; Huang, L. L.; Piro, E.
T.; Cantley, L. C. Nat. Biotechnol. 2001, 19, 661. (g) Petithory, J. R.;
Masiarz, F. R.; Kirsch, J. F.; Santi, D. V.; Malcolm, B. A. Proc. Natl.
Acad. Sci. U.S.A. 1991, 88, 11510. (h) Birkett, A. J.; Soler, D. F.; Wolz,
R. L.; Bond, J. S.; Wiseman, J.; Berman, J.; Harris, R. B. Anal. Biochem.
1991, 196, 137. (i) Meldal, M.; Svendsen, I.; Breddam, K.; Auzanneau,
F.-I. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 3314. (j) St. Hilaire, P. M.;
Willert, M.; Juliano, M. A.; Juliano, L.; Meldal, M. J. Comb. Chem. 1999,
1, 509. (k) Rosse, G.; Kueng, E.; Page, M. G.; Schauer-Vukasinovic, V.;
Giller, T.; Lahm, H.-W.; Hunziker, P.; Schlatter, D. J. Comb. Chem. 2000,
Table 1. Comparison of the Relative kcat/Km for Purified
Underivatized ACC-Peptides in Solution and the Correspondinga
Library Substrates on the Array, upon Treatment with Thrombin
solution phase
microarray
substrate
k
cat/K
m
kcat/K
m
Ac-ATPK-ACC
Ac-AGPK-ACC
Ac-ADAK-ACC
Ac-AFSK-ACC
1.00
0.27
0.02
0.02
1.00
0.26
0.09
0.00
a
In both cases, the data have been normalized to the Ac-ATPK-ACC
value.
2, 461. (l) Singh, J.; Allen, M. P.; Ator, M. A.; Gainor, J. A.; Whipple,
D. A.; Solowiej, J. E.; Treasurywala, A. M.; Morgan, B. A.; Gordon, T.
D.; Upson, D. A. J. Med. Chem. 1995, 38, 217. (m) Dekker, N.; Cox, R.
C.; Kramer, R. A.; Egmond, M. R. Biochemistry 2001, 40, 1694. (n)
Takeuchi, T.; Harris, J. L.; Huang, W.; Yan, K. W.; Coughlin, S. R.;
Craik, C. S. J. Biol. Chem 2000, 275, 34, 26333.
that results from the high substrate density on the array surface.11
The minimal enzyme requirements allow for the determination of
substrate specificity of proteases that are available in only limited
quantities.
(
(
(
4) In the standard notation, substrate residues are referred to as P
, where amide bond hydrolysis occurs between P
and P1′. Schechter, I.; Berger, A. Biochem. Biophys. Res. Commun. 1967,
7, 157.
n 2
, Pn-1...P ,
1
P , P1′, P2′...Pm-1′, P
m
1
In conclusion, we have described an approach to obtain a
complete analysis of protease specificity using microarrays of
peptidyl coumarin substrates. The ACC substrates are uniquely
suited to this application because, unlike fluorescence resonance
energy transfer (FRET)-based substrates, ACC peptides do not
require further analysis to determine site of cleavage.3 The arrays
were constructed with standard DNA microarraying equipment,
using aldehyde-derivatized glass slides and alkoxylamine-function-
alized peptidyl coumarin substrates. The fidelity of proteolytic
cleavage of array-bound substrates was demonstrated using a variety
of proteases. The speed and efficiency of the methodology were
demonstrated by obtaining a proteolytic “fingerprint” of thrombin
in a single experiment. This method should enable fingerprints of
proteolytic activity to be rapidly obtained for numerous proteases,
because thousands of substrates can be printed on a single slide
and large numbers of slides can be readily printed using microarray
automation. Furthermore, due to the sensitivity and efficiency of
this assay, substrate cleavage profiles obtained by this methodology
could potentially be used to detect differential proteolytic profiles
in cell lysates and, thus, could have diagnostic applications.
2
5) (a) 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. (b) Maly, D.
J.; Leonetti, F.; Backes, B. J.; Dauber, D. S.; Harris, J. L.; Craik, C. S.;
Ellman, J. A. J. Org. Chem. 2002, 67, 910.
6) (a) Falsey, J. R.; Renil, M.; Park, S.; Li, S.; Lam, K. S. Bioconjugate
Chem. 2001, 12, 346. (b) Cornish, V. W.; Hahn, K. M.; Schultz, P. G. J.
Am. Chem. Soc. 1996, 118, 8150. (c) Shao, J.; Tam, J. P. J. Am. Chem.
Soc. 1995, 117, 3893. (d) Rodriguez, E. C.; Marcaurelle, L. A.; Bertozzi,
C. R. J. Org. Chem. 1998, 63, 7134. (e) Rose, K. J. Am. Chem. Soc.
a
1994, 116, 30.
(
7) Jensen, K. J.; Alsina, J.; Songster, M. F.; Vagner, J.; Albericio, F.; Barany,
G. J. Am. Chem. Soc. 1998, 120, 5441.
(
(
8) MacBeath, G.; Schreiber, S. L. Science 2000, 289, 1760.
9) (a) Edwards, P. D.; Mauger, R. C.; Cottrell, K.; Morris, F. X.; Pine, K.
K.; Sylvester, M. A.; Scott, C. W.; Furlong, S. T. Bioorg. Med. Chem.
Lett. 2000, 10, 2291. (b) Backes, B. J.; Harris, J. L.; Leonetti, F.; Craik,
C. S.; Ellman, J. A. Nat. Biotechnol. 2000, 18, 187.
(10) The assay is run under conditions where the enzyme and substrate
concentrations are well below the K , and, therefore, the relative rates of
hydrolysis reflect kcat/K
m
m
ratios. See: Fersht, A. R. Enzyme Structure and
Mechanism, 3rd ed.; Freeman: New York, 1985; pp 105-106.
(11) Ekins, R. P.; Berger, H.; Chu, F. W.; Finckh, P.; Krause, F. Nanobiology
1998, 4, 197.
JA027477Q
14870 J. AM. CHEM. SOC.
9
VOL. 124, NO. 50, 2002