award number: w81xwh-05-1-0300 chain antibody inhibitors of membrane-type serine ... · 2011. 5....

AD_________________ Award Number: W81XWH-05-1-0300 TITLE: Structural Characterization and Determinants of Specificity of Single- Chain Antibody Inhibitors of Membrane-Type Serine Protease 1 PRINCIPAL INVESTIGATOR: Christopher J. Farady CONTRACTING ORGANIZATION: University of California

San Francisco, CA 94143-2280

REPORT DATE: March 2007 TYPE OF REPORT: Annual Summary PREPARED FOR: U.S. Army Medical Research and Materiel Command Fort Detrick, Maryland 21702-5012 DISTRIBUTION STATEMENT: Approved for Public Release; Distribution Unlimited The views, opinions and/or findings contained in this report are those of the author(s) and should not be construed as an official Department of the Army position, policy or decision unless so designated by other documentation.

REPORT DOCUMENTATION PAGE Form Approved

OMB No. 0704-0188 Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing this collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden to Department of Defense, Washington Headquarters Services, Directorate for Information Operations and Reports (0704-0188), 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to any penalty for failing to comply with a collection of information if it does not display a currently valid OMB control number. PLEASE DO NOT RETURN YOUR FORM TO THE ABOVE ADDRESS. 1. REPORT DATE (DD-MM-YYYY)01/03/07

2. REPORT TYPEAnnual Summary

3. DATES COVERED (From - To)21 Feb 2006 – 20 Feb 2007

4. TITLE AND SUBTITLE

5a. CONTRACT NUMBER

Structural Characterization and Determinants of Specificity of Single- Chain Antibody Inhibitors of Membrane-Type Serine Protease 1

5b. GRANT NUMBER W81XWH-05-1-0300

5c. PROGRAM ELEMENT NUMBER

6. AUTHOR(S) Christopher J. Farady

5d. PROJECT NUMBER

5e. TASK NUMBER

E-Mail: [email protected] 5f. WORK UNIT NUMBER

7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES)

8. PERFORMING ORGANIZATION REPORT NUMBER

University of California San Francisco, CA 94143-2280

9. SPONSORING / MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSOR/MONITOR’S ACRONYM(S)U.S. Army Medical Research and Materiel Command Fort Detrick, Maryland 21702-5012 11. SPONSOR/MONITOR’S REPORT NUMBER(S) 12. DISTRIBUTION / AVAILABILITY STATEMENT Approved for Public Release; Distribution Unlimited

13. SUPPLEMENTARY NOTES

14. ABSTRACT: Membrane-type serine protease 1 (MT-SP1) is a cancer-associated serine protease implicated in the tumorogenesis and metastasis of breast cancer. Inhibition of MT-SP1 activity has been shown to decrease metastatic potential. We have developed a number of potent and specific single-chain (scFv) antibody inhibitors to MT-SP1, and have begun to characterize their mechanism of inhibition. Through kinetic characterization and site-directed mutagenesis experiments, it has been determined that three potent inhibitors have separate and novel mechanisms of inhibition which do not mimic either biologically or pharmaceutically relevant protease inhibitors. These novel modes of binding and inhibition are the basis for their specificity, and suggest these inhibitors will have less cross-reactivity and toxicity problems when used in vivo to further dissect the role of MT-SP1 in breast cancer.

15. SUBJECT TERMSAntibody inhibitors, Proteases in cancer, Protease specificity

16. SECURITY CLASSIFICATION OF:

17. LIMITATION OF ABSTRACT

18. NUMBER OF PAGES

19a. NAME OF RESPONSIBLE PERSONUSAMRMC

a. REPORT U

b. ABSTRACTU

c. THIS PAGEU

UU

25

19b. TELEPHONE NUMBER (include area code)

Standard Form 298 (Rev. 8-98)Prescribed by ANSI Std. Z39.18

mailto:[email protected]

Table of Contents

Introduction…………………………………………………………….………… 4

Body……………………………………………………………………………… 5-6

Key Research Accomplishments………………………………………….……. 6

Reportable Outcomes…………………………………………………………… 6

References……………………………………………………………………….. 7

Appendix 1……………………………………………………….................... 8-18

Appendix 2……………………………………………………….................... 19-26

2

Introduction: My research has focused on the mechanism of inhibition of a set of single-chain inhibitors of Membrane-Type Serine Protease 1 (MT-SP1). MT-SP1 is a type II transmembrane serine protease (TTSP) expressed on the surface of epithelial cells. Research over the past 10 years have shown that MT-SP1 is involved in a number of biological processes, including tissue development, cell adhesion, and growth factor activation. Furthermore, a number of experiments have suggested dysregulated MT-SP1 activity may have a critical role in tumor progression and metastasis (Uhland 2006). Immunoblotting, immunohistochemical analysis, and expression level analysis have found MT-SP1 to be differentially overexpressed in breast, prostate, and ovarian cancers. MT-SP1 has been shown to play a role in ovarian (Suzuki et al. 2004) and prostate (Galkin et al. 2004) tumor invasion using experimental methods including inhibition of MT-SP1 by small molecules and anti-sense. In breast cancer, MT-SP1 expression levels, when correlated with substrate expression levels have been prognostic in disease progression. High levels of MT-SP1 expression has been correlated with the expression of hepatocyte growth factor (HGF) and the Met/HGF receptor (Kang et al 2003), and with the glycosylation enzyme β1,6-N-Acetylglucosaminyltransferase V (Siddiqui et al. 1999), and in both cases, these clusters showed prognostic value for disease-related survival. MT-SP1 expression levels have also been correlated with macrophage stimulating protein (MSP) (Bhatt et al, 2007), and co-expression of MT-SP1, MSP, and its receptor, RON, have been implicated in breast cancer metastasis to the bone (Welm et al. 2007). Finally, modest orthotopic overexpression of MT-SP1 in mouse epidermal tissue led to spontaneous squamous cell carcinomas (List et al. 2005), further cementing MT-SP1’s role in cancer, and suggesting the enzyme is causally involved in malignant transformation. In order to tease apart the role of MT-SP1 in tumor progression, the Craik Lab has used phage display to develop a series of potent and specific single-chain antibody inhibitors (scFv) of the catalytic domain of MT-SP1 (Sun et al. 2003). With Ki’s ranging from 10pM to 10nM, these inhibitors are extremely potent in vitro, and showed no appreciable inhibition of a panel of closely related serine proteases including factor Xa, thrombin, kallikrein, tPA, and uPA. The potential benefits of these inhibitors are two-fold: they can be used to probe complex biology of MT-SP1, both its role in normal and cancer biology, and they can be used to validate MT-SP1 as both an imaging and therapeutic target. From a more biophysical standpoint, these inhibitors are unique in that they are the only reported antibody inhibitors of serine proteases, a large class of homologous enzymes in which the development of specific inhibitors has been a monumental challenge. Most protease inhibitors take advantage of either the catalytic machinery or topological fold of the protease. These scFv inhibitors bind and recognize a specific three-dimensional epitope near the active site of the enzyme, which allows for specificity among proteases, and allows for a fundamentally different mechanism of inhibition from other biologically active protease inhibitors. A thorough understanding of the mechanism of inhibition of these inhibitors will help us validate their putative mode of action in vivo, and will suggest new strategies for inhibition of MT-SP1 and other serine proteases.

4

Results: Significant progress has been made in my aims of kinetically and structurally characterizing the interactions between two potent scFv inhibitors of MT-SP1, named E2, S4. Using a combination of mutagenesis experiments, steady state kinetics, and stopped flow kinetics, we have determined the mechanism of inhibition of both of these novel macromolecular inhibitors. These results have recently been published in the Journal of Molecular Biology and are attached in appendix 1. S4 is a competitive inhibitor of MT-SP1 with a KI of 140pM, and a number of kinetic competition experiments have shown it binds at or near the S1-site in the protease active site, which is critical to substrate and inhibitor binding. The scFv has a fast association rate, of 1.2x108 M-1s-1, nearly an order of magnitude faster than that of a typical protein-protein interaction. It has a one-step binding mechanism, and rapidly comes to binding equilibrium with the protease. The surface loops of closely related serine proteases surround the protease active site and show a high degree of sequential diversity. Alanine scanning these loops has helped define the binding epitope of the antibody, and thus the basis of its specificity. S4 makes a number of moderate interactions with the six loops surrounding the protease active site, effectively capping the active site and preventing substrate binding. E2 has a more complex mechanism of inhibition. It too is a competitive inhibitor of MT-SP1, has a KI of 8.0pM, and competes with substrate binding in the active site. Stopped-flow experiments revealed that there are at least two binding steps in the inhibition process, suggesting E2 is a mechanistic inhibitor of MT-SP1. Digest experiments show that E2 binds in the protease active site in a substrate-like manner, and can be cleaved like a substrate-like manner. The multiple binding steps and the substrate-like binding suggest that E2 is a standard mechanism serine protease inhibitor. Standard mechanism serine protease inhibitors are ubiquitous in nature, and this mechanism is used to modulate nearly all serine proteases. Despite their robustness, though, standard mechanism inhibitors, such as bovine pancreatic trypsin inhibitor (BPTI), are not specific, and can inhibit many serine proteases. E2 shows a high degree of specificity, though. Alanine scanning of the surface loops of the protease reveal that E2 gains its binding specificity through interactions with the 90’s loop of MT-SP1. Mutations of F97 and D96 to alanine nearly abolish E2 binding to the protease. Therefore, E2 is a standard mechanism inhibitor that gains its specificity from its interaction with a ‘hot-spot’ centered on the 90’s loop of MT-SP1. The significance of these results are two-fold. The mechanisms of inhibition provide a rationale for the effectiveness of these inhibitors, and suggest that the development of specific antibody-based inhibitors against individual members of closely related enzyme families is feasible, and an effective way to develop tools to tease apart complex biological processes. Furthermore, it suggests that these inhibitors might be effective in vivo tools, either as biological inhibitors of MT-SP1, or as imaging or detection tools. E2 has been used as a tool to validate the growth factor MSP as a substrate of MT-SP1. E2 inhibited MT-SP1 on the surface of mouse peritoneal macrophages, and prevented the processing of pro-MSP, and the resulting activation of the macrophages (appendix 2). This helps provide evidence for a signaling axis consisting of MT-SP1, MSP, and its receptor, RON, which appears to be important in directing breast cancer metastases to the bone.

5

Future Directions My proposal aimed to crystallize the MT-SP1-inhibitor complex of the inhibitors. To date, despite setting up more than 1,500 crystallization conditions, I have not been able to get crystals that diffract to more than 8 angstroms. In an effort to improve the chances of crystallization, I plan on converting the antibodies to an Fab scaffold. There are also a number of surface residue mutations - such as removing cysteines - that have been designed that might improve the behavior of MT-SP1 in solution and make the complex more amenable to crystallization. Key Research and Training Accomplishments:

• Determined the modes and mechanisms of inhibition of the two most potent scFv inhibitors of the breast-cancer associated serine protease MT-SP1. These results have been summarized in the manuscript “The Mechanism of Inhibition of Antibody-Based Inhibitors of Membrane-Type Serine Protease 1 (MT-SP1)”, which has been accepted by the Journal of Molecular Biology (see Reportable Outcomes and Appendix 1).

• The scFv inhibitor EB-9 has been validated as a useful biological tool in cell-culture

assays; it inhibited MT-SP1 activity on the surface of mouse peritoneal macrophages. These results were published in “Coordinate expression and functional profiling identify an extracellular proteolytic signaling pathway” in the Proceedings of the National Academy of Sciences (see Reportable Outcomes and Appendix 2).

• Attended and presented a peer-reviewed poster of this work at the Gordon Conference on

“Proteases and their Inhibitors”, in July 2006. • Registered for and will present a poster at the 2007 UCSF Breast Oncology Program

annual conference. This two-day retreat this year is focused on molecular diagnostics, the UCSF Early Detection Research Program; the Integrative Cancer Biology Program.

Reportable Outcomes: Two papers have been accepted for publication in the past year.

• Bhatt, AS, Welm, A, Farady, CJ, Vasquez, M, Wilson, K, and Craik, CS. (2007). Coordinate expression and functional profiling identify an extracellular proteolytic signaling pathway. Proc Natl Acad Sci U S A 104, 5771-5776.

• Farady, CJ, Sun, J, Darragh, MR, Miller, SM, and Craik, CS. (2007). The mechanism of inhibition of antibody-based inhibitors of Membrane-Type Serine Protease 1 (MT-SP1). J Mol Biol in press.

Future results outlined above will be reportable, and will be published when experiments are completed.

6

References: Galkin AV, Mullen L, Fox WD, Brown J, Duncan D, Moreno O, Madison EL. CVS-3983, a selective

matriptase inhibitor, suppresses the growth of androgen independent prostate tumor xenografts. The Prostate 2004, 61, 228-35.

Kang JY, Dolled-Filhart M, Ocal IT, Singh B, Lin CY, Dickson RB, Lim DL, Camp RL. Tissue

microarray analysis of hepatocyte growth factor/met pathway components reveals a role for met, matriptase, and HGF-HAI1 in the progression of node-negative breast cancer. Cancer Research 2003, 63, 1101-5.

Lin CY, Anders J, Johnson M, Dickson RB. Purification and characterization of a complex containing matriptase and a Kunitz-type serine protease inhibitor from human milk. Journal of Biological Chemistry 1999, 274, 18237-42.

List K, Szabo R, Molinolo A, Sriuranpong V, Redeye V, Murdock T, Burke B, Nielsen BS, Gutkind JS, Bugge TH. Deregulated matriptase causes ras-independent multistage carcinogenesis and promotes ras-mediated malignant transformation. Genes Dev 2005, 16, 1934-50.

Sun J, Pons J, Craik CS. Potent and selective inhibition of membrane-type serine protease 1 by human single-chain antibodies. Biochemistry 2003, 42, 892-900.

Suzuki M, Kobayashi H, Kanayama N, Saga Y, Suzuki M, Lin CY, Dickson RB, Terao T. Inhibition of tumor invasion by genomic down-regulation of matriptase through suppression of activation of receptor-pound pro-urokinase. Journal of Biological Chemistry 2004, 279, 14899-908.

Takeuchi T, Shuman MA, Craik CS. Reverse biochemistry: use of macromolecular protease

inhibitors to dissect complex biological processes and identify a membrane-type serine protease in epithelial cancer and normal tissue. Proc Natl Acad Sci U S A 1999, 96, 11054-61.

7

1

2

3

4

56789101112131415

16171819202122232425262728293031

323334

3536

37

383940414243444546474849

ARTICLE IN PRESSYJMBI-59301; No. of pages: 11; 4C: 3, 4, 5, 7

doi:10.1016/j.jmb.2007.03.078 J. Mol. Biol. (2007) xx, xxx–xxx

The Mechanism of Inhibition of Antibody-basedInhibitors of Membrane-type Serine Protease 1 (MT-SP1)

Christopher J. Farady1, Jeonghoon Sun2, Molly R. Darragh1

Susan M. Miller1 and Charles S. Craik1,2⁎

1Graduate Group in Biophysics,University of California,San Francisco, 600 16th St.Genentech Hall, San Francisco,CA. 94143, USA2Department of PharmaceuticalChemistry, University ofCalifornia, San Francisco,600 16th St. Genentech Hall,San Francisco, CA. 94143, USA

U

Present address: J. Sun, Amgen InCA. 91320, USA.Abbreviations used: BPTI, bovine

inhibitor; uPA, urokinase-type plasmscFv, single-chain variable fragmentmembrane-type serine protease 1; Hcombinatorial antibody library; pNApAB, p-aminobenzamidine; ESI, elecE-mail address of the correspondi

[email protected]

0022-2836/$ - see front matter © 2007 P

Please cite this article as: Farady, C. J.Protease 1 (MT-SP1), J. Mol. Biol. (2007

D PRO

OF

The mechanisms of inhibition of two novel scFv antibody inhibitors of theserine protease MT-SP1/matriptase reveal the basis of their potency andspecificity. Kinetic experiments characterize the inhibitors as extremelypotent inhibitors with KI values in the low picomolar range that competewith substrate binding in the S1 site. Alanine scanning of the loopssurrounding the protease active site provides a rationale for inhibitorspecificity. Each antibody binds to a number of residues flanking the activesite, forming a unique three-dimensional binding epitope. Interestingly, oneinhibitor binds in the active site cleft in a substrate-like manner, can beprocessed by MT-SP1 at low pH, and is a standard mechanism inhibitor ofthe protease. The mechanisms of inhibition provide a rationale for theeffectiveness of these inhibitors, and suggest that the development ofspecific antibody-based inhibitors against individual members of closelyrelated enzyme families is feasible, and an effective way to develop tools totease apart complex biological processes.

© 2007 Published by Elsevier Ltd.
EKeywords: antibody; standard mechanism protease inhibitor; specificity;serine protease; HuCAL*Corresponding author TC
5051525354555657585960616263
NC
ORREIntroductionOf the 22 families of naturally occurring, protein-based protease inhibitors known to inhibit the

S1 clan of serine proteases, 18 use an identicalmechanism of inhibition.1 Standard mechanism(also known as canonical, or Laskowski mechan-ism) inhibitors all insert a reactive loop into theactive site of the protease, which binds in an ex-tended β-sheet in a substrate-like manner.2 Whilesome of these inhibitors have developed secondarymechanisms,3 the primary mechanism of inhibitionis extremely well conserved; so much so thatcrystal structures of unrelated inhibitors overlay

646566676869707172737475

c., Thousand Oaks,

pancreatic trypsininogen activator;; MT-SP1,uCAL, human, para-nitroanilide;tron spray ionization.ng author:

ublished by Elsevier Ltd.

et al., The Mechanism of Inh), doi:10.1016/j.jmb.2007.03.0

8

perfectly in the protease active site.4 As evidencedby this remarkable example of convergent evolu-tion, the standard mechanism is an efficient, robustway to inhibit serine proteases. However, thisrobustness often comes at the expense of specifi-city. With the exception of a small number ofparasitic anti-thrombin inhibitors that also bindto protease exosites,5 the majority of standardmechanism protease inhibitors have a relativelybroad specificity. Bovine pancreatic trypsin inhibi-tor (BPTI) inhibits efficiently almost all trypsin-foldserine proteases with P1-Arg specificity, but caninhibit chymotrypsin (P1-Phe specificity) with a KIof 10 nM.6

Much effort has been expended on the develop-ment of specific protease inhibitors for use both asbiological tools and as potential therapeutic agents.As attempts to make specific small molecules havebeset by difficulties,7 researchers have often at-tempted to gain specificity using peptide or protein-based scaffolds. Constrained peptide phage displaylibraries have yielded extremely potent exositeinhibitors of factor VIIa,8,9 and standard mechanisminhibitors of chymotrypsin,10 and urokinase-typeplasminogen activator (uPA),11 with moderatepotency and specificity. An alternate approach has

ibition of Antibody-based Inhibitors of Membrane-type Serine78

mailto:[email protected]://dx.doi.org/10.1016/j.jmb.2007.03.078http://dx.doi.org/10.1016/j.jmb.2007.03.078

C

767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128

129130131132133134135136137

138

139140141142143144145146147148149150

151

152153154155156157158159160161162163164165166167168169170171172173174175176

t1:1t1:2

t1:3

t1:4t1:5

t1:6t1:7

2

ARTICLE IN PRESS

UNCO

RRE

been to improve the specificity of naturally occurringprotease inhibitors.12–14 For example, maturation ofAlzheimer's amyloid β-protein precursor inhibitorKunitz domain, a canonical serine protease inhibitor,via competitive phage display improved its specifi-city for factor VIIa by increasing itsKI against a panelof some related serine proteases by two to five ordersofmagnitude.13 A third approach has been tomaturespecific protease inhibitors on other natural proteinscaffolds, such as ankyrin repeats or antibodies.15

Until now, the characterized protease antibodyinhibitors have been monoclonal antibodies raisedfrom hybridomas, and have tended towards twotypes of inhibitors; those that interfere with multi-merization (and thus activation) of the protease,16–18

and those that bind to loops and protein–proteininteraction sites19–22 and occlude substrate binding,instead of interfering with the catalytic machinery ofthe enzyme, and ensuring complete inhibition.23

Earlier, we reported the development of single-chain variable fragment (scFv) antibody inhibitorsof the serine protease membrane-type serine pro-tease 1 (MT-SP1).24 MT-SP1 (also called matriptase)was discovered and cloned in a search for serineproteases expressed in the PC-3 prostate cancer cellline,25 and was determined independently to be ahighly expressed protease in breast cancer tissue.26

Work by a number of groups has since shown thatMT-SP1 may be a key upstream factor involved inthe ECM remodeling, and in signal transductioncascades involved in cell transformation.27 Abla-tion of MT-SP1 activity has been shown to decreasethe invasiveness of both ovarian and prostatetumor cells, and modest orthotopic over-expressionof MT-SP1 in mouse epidermal tissue led tospontaneous squamous cell carcinomas,28 furthercementing the role of MT-SP1s in cancer, and sug-gesting the enzyme is causally involved inmalignanttransformation.Here, we have characterized the mechanism of

inhibition of the two most potent scFv inhibitors ofMT-SP1, E2 and S4. The inhibitors were selectedfrom a fully synthetic human combinatorial anti-body library in the scFv format (HuCAL, Mor-phoSys AG). HuCAL-scFv contains consensusframeworks with diversified light and heavy chainCDR3 regions reflecting the natural human aminoacid composition.29 A combination of mutagenesisexperiments, steady-state kinetics, and stopped-flow kinetics reveal that, while the inhibitors gainspecificity by making a number of critical interac-tions with surface loops on the protease, they can bestandard mechanism inhibitors, which insert a re-

Table 1. Kinetic parameters of scFv inhibitors

kona (106M−1s−1) k off

a (10−3s−1) Kda (nM) Mode of inhibiti

E2 2.1 0.38 0.16 CompetitiveS4 11.5 5.8 0.51 CompetitiveMOI, multiplicity of infection.

a Values determined by SPR.24

Please cite this article as: Farady, C. J. et al., The Mechanism of InhProtease 1 (MT-SP1), J. Mol. Biol. (2007), doi:10.1016/j.jmb.2007.03.0

9

active loop in a substrate like manner into the activesite of the protease. This work suggests that theantibody scaffold can be used to create extremelyspecific standard mechanism protease inhibitors.Furthermore, the design of inhibitors that utilizemacromolecular recognition factors (variable loops,protein–protein interaction sites) can help to differ-entiate highly homologous proteases, and can thusimpart specificity upon the inhibitors.

TED P

ROOF

Results

Earlier, we described the maturation and initialcharacterization of a number of scFv inhibitors ofMT-SP1.24 The scFvs bound tightly to the catalyticdomain of MT-SP1, and showed a high degree ofspecificity, as they showed no appreciable inhibi-tion of a panel of closely related serine proteases,including factor Xa, thrombin, kallikrein, tissueplasminogen activator (tPA), and uPA at inhibitorconcentrations of 1 μM. Here, we characterize themechanism of inhibition of E2 and S4, the two mostpotent members of this novel class of serine proteaseinhibitors.

Steady-state kinetics

Previous experiments showed that E2 and S4 hadKD values of 160 pM and 500 pM (as determined bysurface plasmon resonance), and were potentinhibitors of MT-SP1. In the current study, a numberof steady-state kinetic experiments were performedin an attempt to understand the mechanism ofinhibition of these inhibitors. The results of theseexperiments are summarized in Table 1. Doublereciprocal plots revealed that both E2 and S4 arecompetitive inhibitors of MT-SP1 with respect toSpectrazyme-tPA, a small molecule para-nitroanilide(pNA) substrate of P1 arginine serine proteases. Tofurther characterize the tight-binding nature of theseinhibitors, accurate KI values were determined; E2and S4 are extremely tight-binding competitiveinhibitors of MT-SP1, with KI values of 8.0(±1.3)pM and 140(±6.0) pM, respectively.To verify that the mode of inhibition is similar in

the context of a macromolecular substrate, a discon-tinuous assay was developed to measure the activa-tion of uPA. The KM of uPA as a substrate forMT-SP1was determined to be 1.7(±0.2) μM, and the kcat ofMT-SP1 activation of sc-uPA was 0.89(±0.09) s−1.Double reciprocal plots showed that the inhibitorswere indeed competitive with respect to macromo-

on Ki (pM) Macromolecular MOI Macromolecular Ki (pM)

8.0±1.3 Competitive 12140±6 Competitive 160


http://dx.doi.org/10.1016/j.jmb.2007.03.078

C

177178179180181182183184

185

186187188189190191192193194195196197198199200201202203204205206207208209210211212

213214215216217218

219220221222223224225226227228229230231232233234235236237238239

Scheme 1.

3

ARTICLE IN PRESS

lecular substrates. From these data, approximate KIvalues of 12 pM for E2 and 160 pM for S4 could beextrapolated (Table 1). Since the substrate (uPA)concentration could not be increased above KM, theerrors associated with KI values are large; none-theless, they confirm that the inhibitors inhibitMT-SP1 equally well, regardless of the size of theprotease substrate.

Pre-steady-state kinetics

A closer examination of the progress curves ofthe steady-state reactions when enzyme was addedto a mixture of substrate and inhibitor revealeddifferent binding mechanisms for E2 and S4 (Figure1). The progress curves for S4 are linear, suggestingthe binding of scFv to enzyme comes to equili-brium rapidly. Conversely, the progress curves forE2 inhibition are curved, suggesting slow bindinginhibition.30 To define the binding mechanisms ofthese scFvs, stopped-flow experiments were per-formed to evaluate the onset of inhibition duringturnover at higher concentrations of enzyme.Stopped-flow experimentsmeasured the appearanceof pNA, and were carried out as described inMaterials and Methods.The stopped-flow traces from the S4 inhibitor

experiments were fit by nonlinear regression to therate equations for reversible, tight-binding inhibi-tion to obtain observed rate constants (kobs) for theonset of inhibition (equation (5)).31 Plots of kobsversus [S4] are linear with positive y-intercepts(Figure 2(a)), consistent with a one-step reversiblemechanism for binding the inhibitor. The y-inter-cepts of the plots give an average off-rate of k−1=1.7×10−2 s−1, and a secondary plot of the slopesversus substrate concentration (Figure 2(a), inset)defined the on rate as k1=1.2×10

8 M−1s−1. The KI

UNCO

RRE

Figure 1. Progress curves of MT-SP1 inhibition by scFvinhibitors reveal multiple mechanisms of inhibition. Theaddition of 0.2 nM enzyme to a mixture of substrate(300 μM Spec-tPA) and inhibitor results in a decrease inproteolytic activity. S4 inhibition results in a linearprogress curve, suggesting rapid-equilibrium inhibition,while the curved nature of the E2 progress curve suggestsslow-binding inhibition.


10

calculated from k−1/k1=147 pM, which is in verygood agreement with the steady-state KI of 140 pMfrom experiment. From these data, it can beconcluded that S4 binds and inhibits MT-SP1 withan extremely fast on-rate, and has a one-stepbinding mechanism as shown in Scheme 1.

240241242

243

244245246247248249250251252253254255256257258259260261262

TED P

ROOFThe stopped-flow traces from the E2 inhibitorexperiments (Figure 2(b)) revealed a more com-plicated binding mechanism. In this case, the prog-

ress curves fit well to a sum of two exponentials(equation (7)), indicating the presence of at least twosteps in the binding process, which leads to the onsetof inhibition. At minimum, a double-exponentialdecay is consistent with a two-step binding mechan-ism. This occurs when the first step in the bindingprocess is more rapid than the second and, asa result, the first observed rate constant (kobs1)shows a linear dependence on the concentration ofinhibitor.32 If kobs1 shows a hyperbolic dependenceon inhibitor concentration, the mechanism of inhibi-tion involves more than two steps. Unfortunately,due to the extremely tight nature of the enzyme–inhibitor interaction, the concentration of inhibitorcould not be increased sufficiently to distinguishbetween a linear or hyperbolic dependence of kobs1on the concentration of inhibitor. But, due to thepresence of two exponential decays, an absoluteminimal mechanism of E2 inhibition has two steps,and E2 can be classified as a slow, tight-bindinginhibitor.30

p-Aminobenzamidine competition assay

p-Aminobenzamidine (pAB) has been reported asa weak competitive inhibitor of P1-arginine-specificserine proteases,33 and can be used as a fluorescentprobe to monitor substrate or inhibitor binding inthe S1 site. The hydrophobic nature of the S1 sitecauses pAB to fluoresce with a maximum emissionaround 360 nm when bound to the enzyme, whilepAB in aqueous solution has both a lower intensityand longer wavelength of emission at 376 nm. pABhas been used as a probe to monitor binding ofinhibitors in the S1 site of inhibitors; competitiveinhibitors displace pAB from the protease active siteand reduce emission at 360 nm,33,34 while non-competitive inhibitors do not.35 pAB has a KI of28.8 μM forMT-SP1 (data not shown), and 1 μMMT-SP1 incubated with 270 μM PAB (to saturate theenzyme) shows a characteristic emission peak at361 nmwhen excited at 325 nm (Figure 3). When oneequivalent of either E2 or S4 is added to the pre-



RREC

TED P

ROOF

263264265266267

268

269270271272273274275276277

278279280281282283284285286287288289290291292293294

Figure 2. Stopped-flow experi-ments confirm disparate mechan-isms of inhibitor binding to MT-SP1. (a) Linear plots of kobs versus S4concentration confirm that S4 has aone-step binding mechanism, asillustrated by Schem 1. Individualtraces are for different concentra-tions substrate: black (×), 200 μMSpec-tPA; green (⋄), 300 μM Spec-tPA; blue (□), 500 μM Spec-tPA;and red (○),800 μM Spec-tPA. (a)The y-intercepts of the observedrate constant plots gave an averageoff rate of k−1=1.7×10

−2 s− 1, and asecondary plot of the slopes versusconcentration of substrate (inset)defined the on rate as k1=1.2×10

8

M− 1s− 1. (b) The raw stopped-flowtrace monitoring E2 inhibition ofMT-SP1 by measuring the appear-ance of pNA at 405 nm fits well tothe double exponential equation (7)with two observed rate constants.The inset shows the residuals of thenon-linear regression fit. Final con-centrations for this trace were240 nM E2, 10 nM MT-SP1, and500 μM Spec-tPA.

4

ARTICLE IN PRESS

UNCOincubated MT-SP1/pAB, the fluorescence isdecreased sharply (Figure 3). This suggests thatboth inhibitors bind at or near the S1 site, and most

likely insert an arginine or lysine side-chain into thepocket.

MT-SP1/inhibitor digest

The reactive site of many standard mechanismserine protease inhibitors has been determined byincubating protease and inhibitor at low pH, wherethe inhibitor can be cleaved in a substrate-likemanner, causing a processing of the inhibitor intotwo fragments, with the cleavage occurring betweenthe P1 and P1′ residues.36,37 When MT-SP1 and E2are incubated at pH 6.0 for an extended period oftime (>120 h), E2 is processed into two bands


11

(Figure 4). This processing is not seen at pH 8.0, orwithout MT-SP1 at pH 6.0. MT-SP1 shows no pro-teolytic activity below pH 6.0, making this thelowest pH at which processing can occur. Electronspray ionization (ESI) mass spectrometry verifiesthat the processing event takes place between R131and R132 in E2. The N-terminal fragment has amass of 12,013 Da (expected 12,014 Da) and theC-terminal fragment has a mass of 15,624 Da(expected 15,627 Da). This places the reactive loopin the CDR3 of the heavy chain of E2, which wouldbe expected from the HuCAL library from whichthese scFvs were matured, as the scaffold had large,diverse CDR3s.29 No S4 processing was observedupon incubation with MT-SP1 for extended periodsof time at low pH, suggesting a different, non-canonical mechanism of inhibition for the S4 scFv.



295

296297

298299300301302303304305306307308309310311312313314315316317

318

319320321322323324325326327328

Figure 3. Inhibitors displace pAB from the MT-SP1active site. pAB (270 μM) incubated with 1 μM MT-SP1emits a strong emission peak with a maximum at 361 nmwhen excited at 325 nm, due to hydrophobic interactionsbetween pAB and the P1 pocket of the protease. When oneequivalent of either S4 (blue trace) or E2 (green trace) isadded to 1 μM MT-SP1 saturated with pAB, the fluo-rescence decreases, suggesting pAB is released into theaqueous environment, where it is weakly fluorescent.Therefore, binding of both S4 and E2 are competitive withpAB binding, and both inhibitors bind in or near the P1pocket in a manner that precludes binding of pAB.

5

ARTICLE IN PRESS

Inhibitor point mutants

To verify the mechanism of inhibition of the scFvinhibitors, point mutants of the arginine residue in

UNCO

RREC

329330331332333334335336337338339340341342343344345346Figure 4. E2 is processed by MT-SP1 at pH 6.0. E2

(2 mM) was incubated at pH 6.0 with (lane 3) and without(lane 1) 0.1 μM MT-SP1 for 120 h. Samples were run on a12% (w/v) polyacrylamide gel and stained with Coomas-sie brilliant blue. At pH 6.0, E2 was processed into twoproducts, with molecular masses determined to be 15,624Da, and 12,013 Da by ESI mass spectrometry. Thesemasses, when added together, account for the mass of thefull-length inhibitor (27,219 Da) and the water moleculeadded to the products during the hydrolysis reaction. Thisprocessing does not take place when E2 and MT-SP1 areincubated at pH 8.0 (lane2). The diagram below shows thesite of the scissile bond in the middle of the heavy chain ofE2.


12

TED P

ROOF

the CDR3 loops of the inhibitors were constructed. Itwould be expected that mutations to residues thatbind in the S1 site of the protease would have thegreatest effect on binding. The mutational data aresummarized in Table 2. E2 R131A and R132A had KIvalues of 78 nM and 454 pM, respectively (KI=12.3pM for the wild-type E2). The mutation of R131 toalanine has a 6500-fold effect on protease inhibition,as would be expected from a residue that binds inthe S1 site. This is consistent with the data from theinhibitor digest at low pH. The mutation of R132caused a 38-fold increase in KI, suggesting that theP1′ arginine also makes significant contacts with theprotease. The CDR3 loop of S4 also has a doublearginine motif, R128 and R129. Both arginineresidues were mutated to alanine and had signifi-cant effects on protease inhibition: S4 R128A had aKI of 2.8 μM, while the R129 alanine mutant had a KIof 3.9 nM, a 4×104-fold and a 56-fold difference,respectively.

MT-SP1 point mutations

To footprint the binding site of the inhibitors, site-directed mutagenesis was used to alanine scan thesurface of the protease domain.38 On the basis of thecrystal structure of MT-SP1,39 30 point mutantswere identified as potential partners in macromo-lecular interactions (Table 3). The majority of theseresidues were located on the loops flanking theprotease active site. Proteolytic activity againstSpec-tPA was used to assure that the point muta-tions did not drastically affect MT-SP1 structure orfunction. The differences between the mutant andwild-type protease kcat/KM values were less thantwofold in most cases, suggesting that the muta-tions had minimal effect on protease structure. MT-SP1 T98A was a sixfold less efficient enzyme thanthe wild-type, which could be attributed primarilyto a lower kcat. MT-SP1 D217A had a threefolddecrease in protease specific activity, which wasdue to an increased KM of 210 μM. The F99A,Q192A, and W215A substitutions in MT-SP1 allresulted in inactive enzymes. The inactive variantseluted from a gel-filtration column at the same sizeas the zymogen protease, suggesting they areinactive because they could not autoactivate (datanot shown).The KI values for E2 and S4 were determined

against the MT-SP1 point mutants. As a positivecontrol, the fold-specific serine protease inhibitor

Table 2 t2:1. Inhibitor point mutant KI versus MT-SP1 t2:2

t2:3KI (nM) Fold difference

t2:4E2 0.01t2:5E2 R131A 78 6500t2:6E2 R132A 0.45 38t2:7S4 0.07t2:8S4 R128A 2800 4.0×104t2:9S4 R129A 3.9 56

All error values >6%. t2:10



CD P

ROOF

347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378

379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410

Table 3t3:1 . MT-SP1 point mutant/inhibitor KI valuest3:2

t3:3 BPTI E2 S4

t3:4 KI (pM) Fold difference KI (pM) Fold difference KI (pM) Fold difference

t3:5 MT-SP1 49.7 12.3 70.4t3:6 Q38A 20.7 0.42 6.1 0.5 73.6 1t3:7 I41A 12.4 0.25 (fourfold) 12.3 1 208 3t3:8 I60A 35.8 0.72 50.4 4.1 40.2 0.57t3:9 D60aA 37.2 0.75 25.1 2 125 1.8t3:10 D60bA 628 12.6 20.8 1.7 427 6.1t3:11 R60cA 134 2.7 11.4 0.93 11.7 0.17 (sixfold)t3:12 F60eA 24.1 0.48 11.4 0.93 102 1.4t3:13 R60fA 73.4 1.5 10.9 0.89 88.9 1.3t3:14 Y60gA 43.5 0.88 12.7 1 151 2.1t3:15 R87A 38.6 0.78 9.4 0.76 54.5 0.77t3:16 F94A 170 3.4 36.2 2.9 1036 15t3:17 N95A 83.6 1.7 45.4 3.7 108 1.5t3:18 D96A 150 3 >1uM >105 897 13t3:19 F97A 224 4.5 >1uM >105 154 2t3:20 T98A 76.4 1.5 83.2 6.7 239 3.4t3:21 H143A 48.5 1 14.2 1.2 1671 24t3:22 Q145A 83.4 1.7 15.4 1.3 116 1.6t3:23 Y146A 116 2.3 76.8 6.2 1405 20t3:24 T150A 57.8 1.2 20.1 1.6 94.6 1.3t3:25 L153A 116 2.3 21.7 1.8 116 1.6t3:26 E169A 163 3.3 23.1 1.9 199 2.8t3:27 Q174A 129 2.6 11.6 0.94 63.7 0.9t3:28 Q175A 39.7 0.8 851 69 246 3.5t3:29 D217A 2137 43 32 2.6 838 12t3:30 Q221aA 63.4 1.3 40.5 3.3 65.7 0.93t3:31 R222A 42.8 0.87 10.5 0.85 61.1 0.87t3:32 K224A 111 2.2 46.3 3.8 59.1 0.84

KI is calculated from the IC50 value; all errors >6%.t3:33

6

ARTICLE IN PRESS

UNCO

RRE

BPTI was screened against the protease point mu-tants, since the mechanism of inhibition is known40

and a the structure of a co-crystal of BPTI and MT-SP1 has been solved.39 As would be expected from afold-specific protease inhibitor, most point mutantshad little effect on BPTI inhibition. The I41Asubstitution moderately improved BPTI binding toMT-SP1, F94A, F97A, and E169A moderatelydecreased BPTI inhibition (corresponding to 1 μM, corresponding to adecrease in free energy of binding of >7.5 kcal/mol.The Q175A variant, on the loop adjacent to the 90sloop, also has a significant effect on E2 inhibition,increasing the KI of E2 by 69-fold (3.0 kcal/mol). The90s loop and 170s loop flank the extended bindingsites of MT-SP1,39,41 and F97 helps form the S4pocket, suggesting E2 binds in the extended bindingpockets of MT-SP1. Though E2 makes minorinteractions with Y146, Q221a, and K224, themajority of the binding energy of E2 for MT-SP1comes from interactions with the 90s loop, andminor interactions with residues flanking the 90s



CTED

PROO

F411412413414

415

416417418419420421422423424425426427428429430431

432433434435

Figure 5. MT-SP1 alanine point mutants and their effect on protease inhibition by (b) BPTI, (c) E2, and (d) S4. (a) The 6MT-SP1 surface loops surrounding the protease active site consisting of a binding cleft and the catalytic triad (sticks). Thespace-filling models shown in (b), (c), and (d) are oriented in the same manner, with the catalytic triad in yellow. Pointmutants that had minimal effect on protease inhibition are shaded in gray, mutations that had a three-to tenfold increasein inhibitor KI are shaded pink, and point mutants that increased inhibitor KI by more than tenfold are shaded in red.Point mutants that decreased inhibitor KI are shaded in green. The point-mutant/inhibitor KI values are given in Table 3.MT-SP1 point mutants have a minimal effect on BPTI inhibition, S4 interacts with moderate affinity to all six proteaseloops surrounding the active site, and E2 binds with high affinity to the 90s and 170s loop. This Figure was prepared usingPyMoL [http://www.pymol.sourceforge.net/].

7

ARTICLE IN PRESS

REloop (I60, Q175). This defines the 90s loop as a hot-spot for E2 binding; and as the 90s loop sequence isunique to MT-SP1, it helps explain E2s specificity forMT-SP1.
R 436
437438439440441442443444445446447448449450451452453454455

UNCODiscussionWe have described the mechanism by which twonovel scFv antibodies inhibit the cancer-associated

serine proteaseMT-SP1/matriptase. The S4 antibodyhas a fast association rate with MT-SP1 (1.2×108M−1s−1 as measured by stopped-flow kinetics) andbinds very tightly to the protease, making numerouscontacts with the loops surrounding the active site ofMT-SP1. The fast on-rate is likely influenced byelectrostatic steering, which can increase kon bymore than 104 over the basal diffusion-controlledassociation rate.42 Mutational data support thishypothesis, as nearly all the residues S4 makessignificant contacts with are polar or charged. Theinhibitor competes with pAB for the S1 site, and theR128A variant of S4 nearly abolishes proteaseinhibition. Despite these data, S4 cannot be consi-


14

dered a standard mechanism inhibitor of MT-SP1without further structural characterization. Stan-dard mechanism inhibitors have a characteristictwo-step binding mechanism; an initial bindingstep, followed by a tightening of the enzyme–inhibitor complex and, as such, have an associationrate approximately two orders of magnitude slowerthan S4. Furthermore, S4 is not processed byMT-SP1at low pH, meaning the substrate-like bindingcannot be assumed.E2, on the other hand, displays all the character-

istics of a standard mechanism serine proteaseinhibitor. While a crystal structure would help todetermine the mechanism of inhibition definitively,the data here are consistent with E2 being astandard-mechanism inhibitor. The enzyme–inhibi-tor complex reaches equilibrium slowly, E2 binds ina substrate-like manner, and inserts an arginineresidue into the S1 site of MT-SP1 (R131), which isimportant for, but not absolutely critical to, inhibi-tion. Furthermore, E2 shows a slight degree of foldspecificity; it inhibits the mouse homolog of MT-SP1,epithin, with a KI of 40 nM,

24 and can inhibit trypsin,a digestive protease with extremely broad specifi-


http://www.pymol.sourceforge.net/http://dx.doi.org/10.1016/j.jmb.2007.03.078

C

456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518

519520521522523524525526527528529530531532

533

534

535536537538539540541542543544

545

546547548549550551552553554555556557558559560561562563564565566567568

569570571572573

8

ARTICLE IN PRESS

UNCO

RRE

city, with an IC50 of 45 μM (data not shown). Incontrast to most standard mechanism serine pro-tease inhibitors, E2 is highly specific for a singleserine protease, MT-SP1. E2 gains much of itsspecificity through interactions with the 90s loopof MT-SP1, and makes significant interactions withresidues D96 and F97 of the protease. Perhaps notsurprisingly, known protease inhibitors that doexhibit a high degree of specificity, such as antic-oagulant protease inhibitors from ticks and leeches,often employ a similar mechanism of inhibition;they combine the robustness of competitive, activesite inhibition with protein extensions that bind torecognition sites on target enzymes.5,43

To our knowledge, these scFvs are the first docu-mented case of mechanistic protease inhibitors onan antibody scaffold that bind in the active site. Anumber of monoclonal antibody protease inhibitorshave been reported17–19,21,22,44 but, despite diversemechanisms, all have the same underlying mode ofaction; they bind to a small, linear peptide sequenceand prevent either a protein–protein or an enzyme–substrate interaction. While often sufficient forinhibition, these monoclonal antibodies can havecurious inhibitory profiles in which they cannotinhibit the hydrolysis of small-molecule substrates,or have different levels of inhibition against differentsubstrates.19,21 Because they are selected in vitroagainst the active form of the enzyme, antibodiesdeveloped by phage display have the inherentadvantage of recognizing three-dimensional epi-topes and the topography of the enzyme active site.With this comes the opportunity for tighter bindingdue to greater buried surface areas and minimalentropic penalties upon binding, and more completeinhibition through insertion of residues into theprotease active site. E2 and S4 have clearly usedthese advantages; they have fast on-rates, very lowKD values, bind in the active site groove, and makecontact with a number of loops flanking the activesite.The HuCAL-scFv library contains consensus

framework sequences for all frequently occurringVH and VL subfamilies with a germline sequence forthe CDR1 and CDR2 in each subfamily.29 Both theheavy and light chain CDR3 regions were diversi-fied according to the natural amino acid composi-tion and cover the natural length variation of the VHand VL CDR3 regions. In retrospect, this proves tobe an ideal scaffold for serine protease inhibition; itallows for a large, rigidified reactive loop to beinserted into the protease active site, while the rest ofthe antibody stabilizes the CDR3 of the heavy chainand makes additional contacts with the protease.While only the most potent scFv inhibitors of MT-SP1 were characterized, all inhibitors had heavychain CDR3 loops of at least 17 residues, suggestingthat large heavy chain CDR3s were critical to MT-SP1 inhibition.The explosion in antibody research over the past 15

years has revolutionized biotechnology. Antibodieshave been developed into extremely useful drugsand imaging devices, and have become critical tools


15

in many areas of biological research. Here, scFvfragments have shown the ability to inhibit speci-fically a single member of a family of closely relatedenzymes. While these molecules will be useful inhelping dissect the complex biology of MT-SP1,45

the mechanisms through which they work onceagain reveals the innate binding flexibility of anti-bodies, and the power of protein engineering. Thatthese inhibitors have developed the robust inhibi-tion mechanism of standard mechanism serineprotease inhibitors, suggests that we can developantibodies to mimic any protein–protein inter-action, and modulate nearly any biological processprecisely.

TED P

ROOFMaterials and MethodsProtein expression, purification, and mutagenesis

MT-SP1 and MT-SP1 mutants were expressed inEscherichia coli and purified from inclusion bodies asdescribed.25 Antibodies were selected from the HuCALscFv library (MorphoSys AG, Martinsreid, Germany).29

Expression and purification of inhibitory scFv antibodieswere as described.24 Point mutants were made using theStratagene Quickchange kit (Stratagene, La Jolla, CA). Oneor two base changes were sufficient to create the pointmutant in each case, and all sequences were verified byDNA sequencing.

Steady-state kinetics

All reaction volumes were 120 μl and were carried outin 50 mM Tris–HCl (pH 8.8), 50 mM NaCl, 0.01% (v/v)Tween-20 unless stated otherwise and all reactions werecarried out in triplicate. Reactions were run in 96-well,medium-binding, flat-bottomed plates (Corning), andcleavage of substrate was measured with a UVmaxMicroplate Reader (Molecular Devices Corporation, PaloAlto, CA.). MT-SP1 and mutant protease concentrationswere determined by 4-methylumbelliferyl p-guanidino-benzoate active-site titration with a Fluormax-2 spectro-fluorimeter.46 Kinetic parameters of MT-SP1 and mutantproteases were determined at 0.2 nM enzyme, withconcentrations of Spectrazyme-tPA (hexahydrotyrosyl-Gly-Arg-pNA, American Diagnostica, Greenwich, CT)varying from 1 μM to 400 μM. KM and kcat weredetermined using the Michaelis–Menten equation.Tight-binding inhibitors require that the effective

decrease in free enzyme be taken into account whendetermining KI values.

31 This is accomplished by incubat-ing enzyme and inhibitor so that the system can reachequilibrium, adding substrate, and then measuringsteady-state velocities at various concentrations of inhi-bitor and fitting the data to:

vi=vs¼ ½ET � IT � Ki*� þ ½ðIT þ Ki*� ETÞ2 þ ð4Ki*ETÞ�1=2

2ET

( )

ð1ÞKI* values are then plotted against substrate concentra-tion to extrapolate the KI at zero substrate concentration:

Ki* ¼ Kið1þ ½S�=KmÞ ð2Þ



574575576577578579580581582583584585586

587588589590

591592593594595596597598599

600

601602603604605606607608609610611612613614615616617618

619

620621622623624625626627628629630631

632633634

635636637638639640641642

643644645646

647648

649

650651

9

ARTICLE IN PRESS

When measuring the effect mutations had on thestrength of the interaction between the protease and in-hibitor, IC50 values were used instead of KI' as determinedabove. Though less accurate than KI,

47 IC50 is easier tocalculate when screening large numbers of inhibitor pointmutants, and is sufficient to monitor relative changes ininhibition versus the wild-type system. IC50 was deter-mined by incubating inhibitor and 0.2 nM enzyme for atleast 5 h at room temperature to assure steady-statebehavior of the system.3 There was no appreciabledecrease in protease activity during the incubation period.Steady-state velocities were then plotted against inhibitorconcentration and fit to:

v ¼ vmin þ ðvmax � vminÞð1þ 10ð½I��IC50ÞÞ ð3Þ

Relative KI' was calculated from IC50 values accordingto:

KI ¼ IC50ð1þ ½S�=KMÞ ð4Þ

Though nearly all protease mutants had a minimal (lessthan twofold) effect on substrate KM, and the substrateconcentration was held well above the KM, this correctionnormalizes the IC50 with respect to the strength of theprotease/substrate interaction. All graphs and equationswere fit using Kaliedagraph 3.6 (Synergy Software, Read-ing, PA).

C

652653654655656657

658

659660661662663664665666667668669
OR
RE

Macromolecular substrate assay

In an assay analogous to that used to monitor factorVIIa activation of FX,8 we developed a coupled assaythat monitors MT-SP1 activation of uPA in which 50 pMMT-SP1 was incubated with various concentrations (finalconcentrations, 12.5–400 nM) of single-chain uPA (Amer-ican Diagnostica). At various time-points (0–150 min),aliquots of the reaction were removed and quenched with10 nM E2. There was no residual MT-SP1 activity afterthe quench, and E2 showed no inhibition of uPA at aconcentration of 10 nM. The amount of active uPA wasmeasured by monitoring the activity for uPA against thepara-nitroanilide uPA substrate Spectrazyme-UK (Amer-ican Diagnostica). The mode of inhibition was determinedfrom double reciprocal plots, and kinetic parameters andinhibition constants were determined using theMichaelis–Menten equation. The KM of Spec-UK for uPA wasdetermined to be 42 μM, and the kcat of uPA turnoverwas 1.0 s−1.
C
670

671672673674675676677678679680

UNStopped-flow kineticsStopped-flow experiments were conducted using aHiTech SF-61DX2 instrument (TgK Scientific Ltd., Brad-ford on Avon, U.K.). Data were collected in dual beammode using photomultiplier detection of absorbance dataat 405 nm. MT-SP1 (10 nM for E2 experiments, 1 nM for S4experiments) was mixed rapidly with a solution ofsubstrate (Spec-tPA, 200–800 μM) and inhibitor (10–300 nM for S4, 100–340 nM for E2) and the appearanceof pNA was monitored for 20 s (for S4) or 150 s (for E2).Concentrations of enzyme and length of experiments werevaried between the two systems to ensure robust signaland equilibration of the system.


16

EDPR

OOF

The stopped-flow traces from the S4 inhibitor experi-ments were fit by nonlinear regression to the rateequations for reversible, tight binding inhibition:31

P ¼ vstþ ðvi � vsÞð1� e�kobstÞ=kobs ð5Þ

The appearance of the product (P) is a function of theinitial (vi) and final (vs) velocities, and an apparent first-order rate constant, kobs for the onset of inhibition. Plots ofkobs versus inhibitor concentration were linear, and fit toequation (6), as would be expected when the inhibitorymechanism consists of one reversible binding step, as inScheme 1:

kobs ¼ k�1 þ k1½I�=ð1þ ½S�=KMÞ ð6Þ

E2 stopped-flow traces fit poorly to equation (5),but fit well to a mechanism with two observed rateconstants:32

P ¼ vstþ ðvi � vsÞð1� e�kobs1tÞ=kobs1þ ðvi � vsÞð1� e�kobs2tÞ=kobs2

ð7Þ

p-Aminobenzamidine fluorescence

Experiments were carried out in PBS with a Fluorolog 3(Instruments SA Inc. Edison, NJ) fluorimeter. Emissionspectra of MT-SP1/pAB were obtained by excitation at325 nm using a 4 nm excitation and 2 nm emissionbandpass, and were scanned from 335–430 nm. Spectrawere corrected for emission due to free pAB and protease.Data corrections were performed with Datamax 2.20software (Instruments SA).
T
Inhibitor digest

E2 (2 μM) or S4 (2 μM) was incubated with 0.1 nMMT-SP1 for 120 h at room temperature. Proteins wereincubated in 100 mM Mes (pH 6.0), 100 mM NaCl or in50 mM Tris–HCl (pH 8.0), 100 mM NaCl. Proteolysiswas monitored by gel mobility-shift on a 12% (w/v)polyacrylamide gel with a 4.5% stacking gel, andstained with Coomassie brilliant blue. ESI mass spectro-metry was carried out with an LCT Premier mass spec-trometer (Waters Corp. Milford, MA), and molecularmasses were determined using MassLynx (Waters) decon-volution software.

Acknowledgements

We thank Jill Winter (Chiron) and MorphoSys,AG for access to the HuCAL-scFv libraries, and DrAmi Bhatt, Dr Alan Marnett, and Dr Sami Mahrusfor many helpful discussions. This work wasfunded by a Program Project Grant for proteasesin cancer, NIH CA72006 (to C.S.C.), the Depart-ment of Defense Breast Cancer Research ProgramBC043431 (C.J.F.) and NIH training grant GM08284(M.R.D.). HuCAL is a registered trademark ofMorphoSys, AG.



C

681

682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748

749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817

10

ARTICLE IN PRESS

UNCO

RRE

References1. Rawlings, N. D., Morton, F. R. & Barrett, A. J. (2006).

MEROPS: the peptidase database. Nucl. Acids Res. 34,D270–D272.

2. Laskowski, M., Jr & Kato, I. (1980). Protein inhibitorsof proteinases. Annu. Rev. Biochem. 49, 593–626.

3. Eggers, C. T., Wang, S. X., Fletterick, R. J. & Craik, C. S.(2001). The role of ecotin dimerization in proteaseinhibition. J. Mol. Biol. 308, 975–991.

4. Tyndall, J. D., Nall, T. & Fairlie, D. P. (2005). Proteasesuniversally recognize beta strands in their active sites.Chem. Rev. 105, 973–999.

5. Rydel, T. J., Tulinsky, A., Bode, W. & Huber, R. (1991).Refined structure of the hirudin-thrombin complex.J. Mol. Biol. 221, 583–601.

6. Castro, M. J. & Anderson, S. (1996). Alanine point-mutations in the reactive region of bovine pancreatictrypsin inhibitor: effects on the kinetics and thermo-dynamics of binding to beta-trypsin and alpha-chymotrypsin. Biochemistry, 35, 11435–11446.

7. Coussens, L. M., Fingleton, B. & Matrisian, L. M.(2002). Matrix metalloproteinase inhibitors and can-cer: trials and tribulations. Science, 295, 2387–2392.

8. Dennis, M. S., Eigenbrot, C., Skelton, N. J., Ultsch,M. H., Santell, L., Dwyer, M. A. et al. (2000). Peptideexosite inhibitors of factor VIIa as anticoagulants.Nature, 404, 465–470.

9. Roberge, M., Santell, L., Dennis, M. S., Eigenbrot, C.,Dwyer, M. A. & Lazarus, R. A. (2001). A novel exositeon coagulation factor VIIa and its molecular interac-tions with a new class of peptide inhibitors. Biochem-istry, 40, 9522–9531.

10. Krook, M., Lindbladh, C., Eriksen, J. A. &Mosbach, K.(1997). Selection of a cyclic nonapeptide inhibitor toalpha-chymotrypsin using a phage display peptidelibrary. Mol. Divers. 3, 149–159.

11. Hansen, M., Wind, T., Blouse, G. E., Christensen, A.,Petersen, H. H., Kjelgaard, S. et al. (2005). A urokinase-type plasminogen activator-inhibiting cyclic peptidewith an unusual P2 residue and an extended proteasebinding surface demonstrates new modalities forenzyme inhibition. J. Biol. Chem. 280, 38424–38437.

12. Wang, C. I., Yang, Q. & Craik, C. S. (1995). Isolation ofa high affinity inhibitor of urokinase-type plasmino-gen activator by phage display of ecotin. J. Biol. Chem.270, 12250–12256.

13. Dennis, M. S. & Lazarus, R. A. (1994). Kunitz domaininhibitors of tissue factor-factor VIIa. II. Potent andspecific inhibitors by competitive phage selection.J. Biol. Chem. 269, 22137–22144.

14. Stoop, A. A. & Craik, C. S. (2003). Engineering of amacromolecular scaffold to develop specific proteaseinhibitors. Nature Biotechnol. 21, 1063–1068.

15. Binz, H. K., Amstutz, P., Kohl, A., Stumpp, M. T.,Briand, C., Forrer, P. et al. (2004). High-affinity bindersselected from designed ankyrin repeat proteinlibraries. Nature Biotechnol. 22, 575–582.

16. Rezacova, P., Lescar, J., Brynda, J., Fabry, M., Horejsi,M., Sedlacek, J. & Bentley, G. A. (2001). Structuralbasis of HIV-1 and HIV-2 protease inhibition by amonoclonal antibody. Structure, 9, 887–895.

17. Puchi, M., Quinones, K., Concha, C., Iribarren, C.,Bustos, P., Morin, V. et al. (2006). Microinjection of anantibody against the cysteine-protease involved inmale chromatin remodeling blocks the developmentof sea urchin embryos at the initial cell cycle. J. CellBiochem. 98, 335–342.

18. Fukuoka, Y. & Schwartz, L. B. (2006). The B12 anti-


17

TED P

ROOF

tryptase monoclonal antibody disrupts the tetramericstructure of heparin-stabilized beta-tryptase to formmonomers that are inactive at neutral pH and active atacidic pH. J. Immunol. 176, 3165–3172.

19. Petersen, H. H., Hansen, M., Schousboe, S. L. &Andreasen, P. A. (2001). Localization of epitopes formonoclonal antibodies to urokinase-type plasminogenactivator: relationship between epitope localizationand effects of antibodies on molecular interactions ofthe enzyme. Eur. J. Biochem. 268, 4430–4439.

20. Matias-Roman, S., Galvez, B. G., Genis, L., Yanez-Mo,M., de la Rosa, G., Sanchez-Mateos, P. et al. (2005).Membrane type 1-matrix metalloproteinase is in-volved in migration of human monocytes and is regu-lated through their interaction with fibronectin orendothelium. Blood, 105, 3956–3964.

21. Xuan, J. A., Schneider, D., Toy, P., Lin, R., Newton, A.,Zhu, Y. et al. (2006). Antibodies neutralizing hepsinprotease activity do not impact cell growth but inhibitinvasion of prostate and ovarian tumor cells inculture. Cancer Res. 66, 3611–3619.

22. Obermajer, N., Premzl, A., Zavasnik Bergant, T., Turk,B. & Kos, J. (2006). Carboxypeptidase cathepsin Xmediates beta2-integrin-dependent adhesion of differ-entiated U-937 cells. Expt. Cell Res. 312, 2515–2527.

23. Maun, H. R., Eigenbrot, C. & Lazarus, R. A. (2003).Engineering exosite peptides for complete inhibitionof factor VIIa using a protease switch with substratephage. J. Biol. Chem. 278, 21823–21830.

24. Sun, J., Pons, J. & Craik, C. S. (2003). Potent andselective inhibition of membrane-type serine protease1 by human single-chain antibodies. Biochemistry, 42,892–900.

25. Takeuchi, T., Shuman, M. A. & Craik, C. S. (1999).Reverse biochemistry: use of macromolecular pro-tease inhibitors to dissect complex biological pro-cesses and identify a membrane-type serine proteasein epithelial cancer and normal tissue. Proc. Natl Acad.Sci. USA, 96, 11054–110561.

26. Lin, C. Y., Anders, J., Johnson, M. & Dickson, R. B.(1999). Purification and characterization of a complexcontaining matriptase and a Kunitz-type serine pro-tease inhibitor from human milk. J. Biol. Chem. 274,18237–18242.

27. Uhland, K. (2006). Matriptase and its putative role incancer. Cell Mol. Life Sci. 63, 2968–2978.

28. List, K., Szabo, R., Molinolo, A., Sriuranpong, V.,Redeye, V., Murdock, T. et al. (2005). Deregulatedmatriptase causes ras-independent multistage carci-nogenesis and promotes ras-mediated malignanttransformation. Genes Dev. 19, 1934–1950.

29. Knappik, A., Ge, L., Honegger, A., Pack, P., Fischer,M., Wellnhofer, G. et al. (2000). Fully synthetic humancombinatorial antibody libraries (HuCAL) based onmodular consensus frameworks and CDRs rando-mized with trinucleotides. J. Mol. Biol. 296, 57–86.

30. Morrison, J. F. &Walsh, C. T. (1988). The behavior andsignificance of slow-binding enzyme inhibitors.Advan. Enzymol. Relat. Areas Mol. Biol. 61, 201–301.

31. Williams, J. W. & Morrison, J. F. (1979). The kinetics ofreversible tight-binding inhibition. Methods Enzymol.63, 437–467.

32. Hiromi, K. (1979). Kinetics of Fast Enzyme Reaction:Theory and Pratice. Kodansha Ltd., Wiley, Tokyo.

33. Evans, S. A., Olson, S. T. & Shore, J. D. (1982).p-Aminobenzamidine as a fluorescent probe for theactive site of serine proteases. J. Biol. Chem. 257,3014–3017.

34. Parry, M. A., Maraganore, J. M. & Stone, S. R. (1994).



818819820821822823824825826827828829830831832833834835836837838839840841842843844845846

847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876

877

878879

11

ARTICLE IN PRESS

Kinetic mechanism for the interaction of Hirulog withthrombin. Biochemistry, 33, 14807–14814.

35. Fernandez, A. Z., Tablante, A., Beguin, S., Hemker,H. C. & Apitz-Castro, R. (1999). Draculin, the antic-oagulant factor in vampire bat saliva, is a tight-binding, noncompetitive inhibitor of activated factorX. Biochim. Biophys. Acta, 1434, 135–142.

36. Ozawa, K. & Laskowski, M., Jr (1966). The reactive siteof trypsin inhibitors. J. Biol. Chem. 241, 3955–3961.

37. McGrath, M. E., Hines, W. M., Sakanari, J. A.,Fletterick, R. J. & Craik, C. S. (1991). The sequenceand reactive site of ecotin. A general inhibitor of pan-creatic serine proteases from Escherichia coli. J. Biol.Chem. 266, 6620–6625.

38. Cunningham, B. C. & Wells, J. A. (1989). High-resolution epitope mapping of hGH-receptor interac-tions by alanine-scanning mutagenesis. Science, 244,1081–1085.

39. Friedrich, R., Fuentes-Prior, P., Ong, E., Coombs, G.,Hunter, M., Oehler, R. et al. (2002). Catalytic domainstructures of MT-SP1/matriptase, a matrix-degradingtransmembrane serine proteinase. J. Biol. Chem. 277,2160–2168.

40. Luthy, J. A., Praissman, M., Finkenstadt, W. R. &Laskowski, M., Jr (1973). Detailed mechanism of inter-action of bovine-trypsin with soybean trypsin inhibi-tor (Kunitz). I. Stopped flow measurements. J. Biol.Chem. 248, 1760–1771.

41. Sriprapundh, D., Craik, C. S. (2006).

UNCO

RREC


18

PROO

F

42. Schreiber, G. & Fersht, A. R. (1996). Rapid, electro-statically assisted association of proteins. NatureStruct. Biol. 3, 427–431.

43. Rezaie, A. R. (2004). Kinetics of factor Xa inhibition byrecombinant tick anticoagulant peptide: both activesite and exosite interactions are required for a slow-and tight-binding inhibition mechanism. Biochemistry,43, 3368–3375.

44. Martin, F., Volpari, C., Steinkuhler, C., Dimasi, N.,Brunetti, M., Biasiol, G. et al. (1997). Affinity selec-tion of a camelized V(H) domain antibody inhibitorof hepatitis C virus NS3 protease. Protein Eng. 10,607–614.

45. Bhatt, A. S., Welm, A., Farady, C. J., Vasquez, M.,Wilson, K. & Craik, C. S. (2007). Coordinate expres-sion and functional profiling identify and extracellularproteolytic signaling pathway. Proc. Natl Acad. Sci.USA.

46. Jameson, G. W., Roberts, D. V., Adams, R. W., Kyle,W. S. & Elmore, D. T. (1973). Determination of theoperational molarity of solutions of bovine alpha-chymotrypsin, trypsin, thrombin and factor Xa byspectrofluorimetric titration. Biochem. J. 131, 107–117.

47. Chou, T. (1974). Relationships between inhibitionconstants and fractional inhibition in enzyme-cata-lyzed reactions with different numbers of reactants,different reaction mechanisms, and different typesand mechanisms of inhibition. Mol. Pharmacol. 10,235–247.

Edited by I. Wilson
DE(Received 23 February 2007; accepted 20 March 2007)T


pathwayCoordinate expression and functional profiling identify an extracellular proteolytic signaling

Charles S. Craik Ami S. Bhatt, Alana Welm, Christopher J. Farady, Maximiliano Vásquez, Keith Wilson, and

doi:10.1073/pnas.0606514104 published online Mar 27, 2007; PNAS

This information is current as of March 2007.

Supplementary Material www.pnas.org/cgi/content/full/0606514104/DC1

Supplementary material can be found at:

www.pnas.org#otherarticlesThis article has been cited by other articles:

E-mail Alerts. click hereat the top right corner of the article or

Receive free email alerts when new articles cite this article - sign up in the box

Rights & Permissions www.pnas.org/misc/rightperm.shtml

To reproduce this article in part (figures, tables) or in entirety, see:

Reprints www.pnas.org/misc/reprints.shtml

To order reprints, see:

Notes:

19

http://www.pnas.org/cgi/content/full/0606514104/DC1http://www.pnas.org#otherarticleshttp://www.pnas.org/cgi/alerts/ctalert?alertType=citedby&addAlert=cited_by&saveAlert=no&cited_by_criteria_resid=pnas;0606514104v1&return_type=article&return_url=http%3A%2F%2Fwww.pnas.org%2Fcgi%2Freprint%2F0606514104v1.pdfhttp://www.pnas.org/misc/rightperm.shtmlhttp://www.pnas.org/misc/reprints.shtml

Coordinate expression and functional profilingidentify an extracellular proteolytic signaling pathwayAmi S. Bhatt*, Alana Welm†‡, Christopher J. Farady*, Maximiliano Vásquez§, Keith Wilson§, and Charles S. Craik*¶

*Department of Pharmaceutical Chemistry, University of California, 600 16th Street, San Francisco, CA 94158; †The G. W. Hooper Foundation, Universityof California, 513 Parnassus Avenue, San Francisco, CA 94153; and §PDL Biopharma, Inc., 34801 Campus Drive, Fremont, CA 94555

Edited by James A. Wells, University of California, San Francisco, CA, and approved February 6, 2007 (received for review July 30, 2006)

A multidisciplinary method combining transcriptional data, speci-ficity profiling, and biological characterization of an enzyme maybe used to predict novel substrates. By integrating protease sub-strate profiling with microarray gene coexpression data fromnearly 2,000 human normal and cancerous tissue samples, threefundamental components of a protease-activated signaling path-way were identified. We find that MT-SP1 mediates extracellularsignaling by regulating the local activation of the prometastaticgrowth factor MSP-1. We demonstrate MT-SP1 expression in peri-toneal macrophages, and biochemical methods confirm the abilityof MT-SP1 to cleave and activate pro-MSP-1 in vitro and in a cellularcontext. MT-SP1 induced the ability of MSP-1 to inhibit nitric oxideproduction in bone marrow macrophages. Addition of HAI-1 or anMT-SP1-specific antibody inhibitor blocked the proteolytic activa-tion of MSP-1 at the cell surface of peritoneal macrophages. Takentogether, our work indicates that MT-SP1 is sufficient for MSP-1activation and that MT-SP1, MSP-1, and the previously shownMSP-1 tyrosine kinase receptor RON are required for peritonealmacrophage activation. This work shows that this triad of growthfactor, growth factor activator protease, and growth factor recep-tor is a protease-activated signaling pathway. Individually, MT-SP1and RON overexpression have been implicated in cancer progres-sion and metastasis. Transcriptional coexpression of these genessuggests that this signaling pathway may be involved in severalhuman cancers.

cancer � macrophage activation � protease substrate specificity �proteomics

Despite the successful physiological and biochemical character-ization of many proteases, the vast majority of the �2% of thehuman genome that encodes proteases has yet to be functionallyclassified. Although many approaches demonstrate the sufficiencyof a protease to cleave a given substrate, very few are able to addressthe physiological relevance of such in vitro findings. Cell-surfaceproteolysis is suggested to play a major role in cancer progressionand metastasis through the processing of macromolecules impor-tant for regulating the extracellular environment. The cell-surfacelocalization, high activity, and exquisite specificity of type II trans-membrane serine proteases (TTSPs) suggest a role in outside-insignaling and interaction with the microenvironment. We elected toapply a multifaceted approach to identify physiologically relevantsubstrates of one prominent member of this family, membrane typeserine protease 1 (MT-SP1/matriptase).

Members of the TTSP family, such as hepsin and MT-SP1, arehighly expressed in many cancers, including those of the prostate,breast, colon, and ovary (1–9). Both overexpression and inhibi-tion studies have supported the role of MT-SP1 in tumorigenesisand tumor growth. Targeted overexpression of MT-SP1 insquamous epithelia in mice results in skin-limited nodules ofsquamous cell carcinoma that become metastatic in the presence ofthe chemical carcinogen DMBA (10). Small molecule and macro-molecular inhibitors of MT-SP1 have been developed and appliedin a mouse model of cancer, resulting in growth suppression ofandrogen-independent prostate cancer xenografts (2, 11). Takentogether, these findings suggest a role for MT-SP1 in cancer.

In this study, we sought to explore the mechanism ofMT-SP1’s activity in more detail. Although transcriptionalcoregulation has been reported between known pathway com-ponents, it has not been applied for the prediction of novelenzyme substrates (12, 13). Transcriptional profiling of nearly2,000 human samples, including those from normal tissues,cancer cell lines, and 17 types of cancer tissue was performedto determine expression levels of MT-SP1, its candidatesubstrates, and its proposed endogenous inhibitor, the hepa-tocyte growth factor activator 1 (HAI-1) (14). Candidatesubstrates whose expression correlated with that of MT-SP1 ina statistically significant fashion were chosen for subsequentbiochemical validation. These substrates were tested and val-idated in primary cells. Using this approach, we identified thecancer-associated growth factor macrophage-stimulating pro-tein 1 (MSP-1) as a substrate of MT-SP1.

ResultsUse of PS-SCL and Other Specificity Data to Guide Candidate SubstrateSelection. Limited information on MT-SP1 substrate specificity wascollected by using a complete diverse positionally scanned syntheticcombinatorial library (PS-SCL) of synthetic substrates. The methodcan be used to identify consensus, nonprime side cleavage motifs forproteases (15). Our functional characterization of the bindingspecificity of MT-SP1 at the substrate-binding cleft is in accord withthe information obtained from structural studies revealing trypsin-like specificity at the S1 position, a shallow pocket for small,hydrophobic residues at the S2 position, and an open negativelycharged cavity at the S4 position, allowing for binding of a basicresidue at P3 or P4 (16). Given the relative degeneracy of thespecificity information obtained through biochemical profiling andstructural studies, additional information was considered in thedevelopment of a consensus cleavage sequence. By using thespecificity determinants obtained from the PS-SCL data and analignment of known macromolecular substrates, a set of consensussequences for MT-SP1 cleavage was deduced. The specificity ofMT-SP1 was in fairly good agreement with the described cleavagesequence of the HGF-homolog, MSP-1 (Table 1).

Author contributions: A.S.B. and C.S.C. designed research; A.S.B. and A.W. performedresearch; A.S.B., A.W., C.J.F., M.V., and K.W. contributed new reagents/analytic tools;A.S.B., A.W., and C.S.C. analyzed data; and A.S.B. and C.S.C. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.

Abbreviations: MT-SP1, membrane type serine protease 1; MSP-1, macrophage stimulatingprotein 1; HAI-1, hepatocyte growth factor activator inhibitor 1; TTSP, type two transmem-brane serine protease; PS-SCL, positional scanning synthetic combinatorial library.

‡Present address: Department of Oncological Sciences, Huntsman Cancer Institute, Univer-sity of Utah, Salt Lake City, UT 84112.

¶To whom correspondence should be addressed at: University of California, GenentechHall, 600 16th Street, San Francisco, CA 94158-2517. E-mail: [email protected].

This article contains supporting information online at www.pnas.org/cgi/content/full/0606514104/DC1.

© 2007 by The National Academy of Sciences of the USA

www.pnas.org�cgi�doi�10.1073�pnas.0606514104 PNAS � April 3, 2007 � vol. 104 � no. 14 � 5771–5776

BIO

CHEM

ISTR

Y

20

http://www.pnas.org/cgi/content/full/0606514104/DC1http://www.pnas.org/cgi/content/full/0606514104/DC1

Transcriptional Profiling of Candidate MT-SP1 Substrates Demon-strates Two Candidate Proteins That Are Coexpressed with MT-SP1.To measure candidate gene RNA levels in human tissues and celllines, tumor samples and cancer-derived cell lines were obtainedfrom multiple sources. Tumor samples were obtained from thefollowing tissue types: bladder, breast, cervical, colon, esopha-geal, head and neck, lung, ovarian, pancreatic, prostate, renal,stomach, testicular, and uterine cancers as well as Ewing’ssarcoma, glioblastoma, and melanoma. Fifty-nine cancer celllines, largely obtained from the American Type Culture Collec-tion (Manassas, VA), were cultured in vitro and as SCID mousexenografts. RNA was extracted from these samples, and RNAfrom 382 samples representing 86 types of nonpathogenic tissuewas obtained from commercial sources. A single, custom mi-croarray chip was designed to contain 400,000 perfect-matchprobes (�59,000 probe sets). These arrays were then probed withbiotinylated cDNA derived from the sample RNAs, and bindingwas quantitated by fluorescence. Genes of interest, such asdescribed and predicted MT-SP1 substrates, the hepatocytegrowth factor activator inhibitor 1 (HAI-1), and common cancermarkers were chosen for further analysis. Both Pearson’s prod-uct moment correlation coefficients and Spearman’s rank-ordercorrelation coefficients were calculated for all the given genespaired with MT-SP1. Bonferroni-corrected P values and falsediscovery rate P values are presented in supporting information(SI) Table 2. The data are summarized in Fig. 1, with statisticalsignificance of the correlation established as P � 0.05 afterBonferroni correction or the false discovery rate correction (seeMethods and SI Appendix). Significant correlations were notdetected in cervical cancer, esophageal cancer, fetal tissues, headand neck cancer, Ewing’s sarcoma, renal cancer, melanoma, ortesticular cancer.

Expression of HAI-1 and MT-SP1 was significantly correlatedin 13 of the tissue/cell line categories studied. The range ofunadjusted P values calculated for the Pearson’s product mo-ment correlation coefficient range from 1.83 � 10�45 (HAI-1;normal tissues) to 0.994. The range of unadjusted P valuescalculated for the Spearman’s rank-order correlation coefficientrange from 5.85 � 10�34 (HAI-1; normal tissues) to 1. The mosthighly significant correlation detected was between HAI-1 andMT-SP1 in the aggregated sample of individually characterizednormal body tissues.

This analysis showed that the expression of MT-SP1 and twoof its previously described substrates, Trask and PAR2, was alsosignificantly correlated in many tissue types. Interestingly, ex-pression of the described MT-SP1 substrate HGF did not

correlate well with MT-SP1 expression at the transcriptionallevel. Transcriptional expression of the HGF homolog, MSP-1,however, correlated well with expression of the protease innormal tissues and in certain cancers (Fig. 1). Furthermore,expression of the receptor for MSP-1, RON, very stronglycorrelated with MT-SP1 expression. Although RON has a fairlynarrow expression pattern in normal tissues, including terminallydifferentiated macrophages, keratinocytes, several types of co-lumnar epithelium, and osteoclasts (17), our data demonstratedaberrant receptor expression in certain cancer tissues. MT-SP1and RON transcript levels were correlated in several tissue types

Table 1. Alignment of MSP-1 activation sequence with the predicted MT-SP1 cleavagesequence consensus

P4 P3 P2 P1

Filaggrin R K R RHGF/SF K Q L RMT-SP1/matriptase R Q A RPAR2 S K G RTrask/CDCP1/SIMA135 K Q S RuPA/urokinase P R F K*P1-diverse PS-SCL K/R K/R S � P � G � L K/RPhage Display K/R vs. X X vs. K/R Small/hydrophobic K/RCrystal Structure K/R vs. X X vs. K/R Small/hydrophobic K/RConsensus K/R vs. S/P Q vs. K/R Small/hydrophobic R � K*MSP-1 S K L R

Per standard notation, P1 is designated as the amino acid N-terminal to the scissile bond with P2 being theamino acid N-terminal to P1, etc. The P4 through P1 amino acids for known substrates of MT-SP1, P1-diverse PS-SCLdata, phage display data, and crystallographic determinations are displayed. An MT-SP1 consensus cleavagesequence was derived from this information. The MSP-1 activation sequence is also presented. Asterisks indicatecompiled PS-SCL data or MT-SP1 cleavage sequences determined by N-terminal sequencing in this work.

selpmas fo reb

mun

Tissue Categories

FG

Fb

1-R

FG

F

1R

FG

EV

FG

H

golomoh 1-delzzir

F

FG

E-B

H

DR

PT

P

2R

FG

EV

FG

E

5AF

ED

RF

GE

PS

M

APu

FG

EV

(p

robe

set

1)

2R

AP

(pro

be s

et 2

) 2

RA

P

ksarT

NO

R

1-IA

H

74 Prostate cancer - 2153 Lung cancer212 Cell lines46 Stomach cancer33 Prostate cancer - 147 Glioblastoma

103 Bladder cancer 85 Breast cancer71 Ovarian cancer - 181 Colon cancer - Metastatic70 Ovarian cancer - 261 Pancreatic cancer

204 Colon cancer328 Normal tissues56 Uterine cancer

Bonferroni, p < 0.05FDR, p < 0.05

Fig. 1. Correlational cluster diagram of MT-SP1 and associated proteins.Transcriptional profiling of 19 genes was performed for nearly 2,000 samplesfrom cell lines, normal tissues, and cancer tissues. Pearson’s and Spearman’scorrelation coefficients were calculated for each gene paired with MT-SP1.Associated P values were also calculated and corrected by using the Bonferronicorrection method (highly specific) and the false discovery rate (FDR) method(highly sensitive). Significantly correlated gene pairs (adjusted P � 0.05) areindicated by shading. Tissue categories without any significant correlationsare not displayed. All correlations determined to be significant by using theBonferroni method (darker shading) are significant by using the FDR method(lighter shading). Transcript levels of HAI-1, the proposed endogenous inhib-itor of MT-SP1, are significantly correlated with transcript levels of MT-SP1 inthe largest proportion of tissues.

5772 � www.pnas.org�cgi�doi�10.1073�pnas.0606514104 Bhatt et al.21

http://www.pnas.org/cgi/content/full/0606514104/DC1http://www.pnas.org/cgi/content/full/0606514104/DC1http://www.pnas.org/cgi/content/full/0606514104/DC1

with coexpression strength on par with MT-SP1/Trask andMT-SP1/HAI-1. The MT-SP1/MSP-1/RON interaction was cho-sen for subsequent biochemical validation.

MT-SP1 Is Present on the Cell Surface of Peritoneal Macrophages.Quantitative RT-PCR (Fig. 2) and immunoblotting (data notshown) demonstrated the expression of both MT-SP1 and RONin mouse peritoneal macrophages at levels �10-fold higher thanin bone marrow-derived macrophages (Fig. 2). The endogenousinhibitor of MT-SP1, HAI-1, was not highly expressed in eithercell type. MSP-1 and RON were originally described as impor-tant components of signaling cascades at the surface of certainpopulations of mature and differentiated macrophages. Previousstudies have identified a cell-surface, trypsin-fold serine proteaseactivity on macrophages that activates MSP-1, potentiatingMSP-1 binding to RON and subsequent macrophage activation(18–20).

Cell Surface-Bound MT-SP1 Can Cleave the Predicted Substrate pro-MSP-1. We investigated whether MT-SP1 could activate MSP-1for several reasons: (i) MSP-1 requires proteolytic activation to

bind and activate its receptor RON (18–21) (ii) MSP-1 containsthe consensus sequence for cleavage by MT-SP1 (Table 1), and(iii) MT-SP1 and RON are coexpressed in both peritonealmacrophages and cancer tissues, consistent with previous ob-servations that MSP-1 is activated at the surface of cells thatrespond to MSP-1 (20). Activation of MSP-1 is the result ofproteolysis C-terminal to R483 (21). The specificity determinantN-terminal to R483 is the peptide sequence SKLR (P4-P1) (22),and is in agreement with the PS-SCL results for MT-SP1 (Table1). MT-SP1 cleaved pro-MSP-1 into two major fragments, whichcorresponded to the � and � chains of the mature MSP-1 (Fig.2). Cleavage at R483 was confirmed by N-terminal sequencing ofthe � chain (data not shown). The dose-dependence of thiscleavage event was established by analyzing pro-MSP-1 cleavageover a 1,000-fold concentration range. Cleavage of MSP into the� and � chains was detectable at concentrations of protease aslow as 1 nM, with nearly complete activation of pro-MSP-1 by100 nM MT-SP1 within 1 h at 37°C. Activation of RON at R309after the specificity determinant RRRR (P4-P1) was also dem-onstrated by the nearly complete processing of single chain RONinto the mature �/� complex by MT-SP1 (data not shown).

MSP-1 Activation by MT-SP1 Results in Macrophage MorphologyChanges and Inhibition of Nitric Oxide Production by Macrophages.The cleavage of MSP-1 by MT-SP1 was then tested in primarycells in culture. Macrophages respond to MSP-1 via RONactivation, leading to changes in cell shape and increased mi-gration, as well as an inhibition of nitric oxide production(23–25). We found that primary mouse peritoneal macrophages,but not primary mouse bone marrow macrophages, underwenta characteristic shape change upon activation by MSP-1 (Fig. 3)(25). The effects of the endo

award number: w81xwh-05-1-0300 chain antibody inhibitors of membrane-type serine ... · 2011. 5....

Documents