Studies of enzymes from two protease families: Tissue Kallikreins, ADAMs and MMPs.

Abstract

The human kallikrein family is a family of proteolytic enzymes, classified as serine proteases, that derive from chromosome 19, locus 13.3-13.4. These enzymes are widespread in pathophysiological processes such as cancer and neurodegenerative diseases; hence studies of catalytic sites and inhibitors are important in relation to the longer term of design of therapeutic drugs. One member of the family, human kallikrein 4 (hK4) which is thought to carry out crucial functions in the prostate, was expressed in this study as a secreted protein in a baculovirus expression system, bearing a His-tag and V5-epitope that were used for purification and detection respectively. Its mass was estimated to be 35kDa, ~2kDa less than the equivalent product expressed in monkey kidney cells. The protein was purified to 50-90% purity with a yield of 0.93mg/L-4.8mg/L based on methods derived from computational prediction of its properties, such as pI. Computational analysis was extended by applying high-performing computing techniques, such as molecular dynamics, and flexible ligand docking, to predict antigenic regions, the likely substrate specificity and putative inhibitors. These results show that hK4 has a loop, between Leu83-Ser94 that shows promise as a specific segment that can be exploited for generation of antibodies. Preferred substrates were also predicted to bear hydrophobic residues at the P'-region of the scissile bond and amphiphilic residues at the P-region. At the S-region, hK4 potentially involves its unique PLYH-motif in recognizing the P4/P5 position from the substrate. Flexible ligand-docking studies indicate that hK4 can be inhibited by inhibitors that carry a modified bulky hydrophobic sidechain with a guanidinium group at the P1-position and its own putative autoactivation region residues at the P2, P1' and P2' position. The computational study was extended to other members of the kallikrein family, predicting distinctions between these that could be used for future studies. These results show that 8 of the fifteen kallikrein members are very homologous in terms of specificity bearing typical trypsinlike activity and specificity, except for hK2, hK3, hK4, hK5, hK7, hK9, hK15 that retain certain distinct signatures in the binding pocket in terms of secondary specificity. The principles of substrate-specificity analysis that were developed were further applied on three metzincins, MMP-3, ADAM-9 and ADAM-10. These three enzymes are metalloproteases, which are involved in tissue remodeling, intracellular signalling and cell-to-cell mediation. The substrate-specificity analysis was carried out on all three metzincins using the structure of a crystallized complex of the MMP-3 enzyme with the TIMP-1 natural inhibitor as template. In this specific enzyme-substrate complex, the challenge was to model and suggest a possible orientation of the P-region, which is not known. The interactions on the P/S-region are therefore unclear and need to be clarified. In order to suggest the arrangement of the enzyme-substrate complex and the undefined S-subsites, four new residues were added in an extended beta-sheet conformation to the P1' residue (derived originally from the TIMP-1 inhibitor) to create a full-length modeled substrate spanning P4'-P4. This new modeled region, in particular, was bound through backbone H-bonds with the enzyme at position 169 (MMP nomenclature) suggesting a new crucial residue for substrate binding, and satisfied steric and chemical restraints in the S'-region of the enzyme. This modeling approach also indicated a putative presence of an S2/S3-pocket on these metzincins which is composed of different residues for MMP-3, ADAM-9 and ADAM-10, and which could prove useful for future drug design projects. Furthermore, the data argue against the involvement of a polarizable water molecule in catalysis, a mechanism that has been postulated by various groups. A new catalytic mechanism is suggested to involve an oxyanion anhydride transition state. This study is a demonstration of the power of combining bioinformatics with wet-lab biochemistry.

Full text

Page 1

Studies of enzymes from two protease families:
Tissue Kallikreins, ADAMs and MMPs.












A thesis submitted to the Queensland University of Technology for the degree of Master of
Applied Science by Research.




Sergio Manzetti, B. Sc. (Chemistry)
Cluster for Molecular Biotechnology
Science Research Centre










School of Life Sciences, Queensland University of Technology, Brisbane, Australia.
June 2005

Page 2

2

KEYWORDS
Human kallikrein family, prostate-cancer, kallikrein 4, baculovirus system, purification, metal-
affinity chromatography, immuno-detection, computer modelling, enzyme-specificity prediction,
inhibitor design, docking, metalloproteases, enzyme-substrate complex.

Page 3

3

ABSTRACT
The human kallikrein family is a family of proteolytic enzymes, classified as serine proteases,
that derive from chromosome 19, locus 13.3-13.4. These enzymes are widespread in
pathophysiological processes such as cancer and neurodegenerative diseases; hence studies of
catalytic sites and inhibitors are important in relation to the longer term of design of therapeutic
drugs. One member of the family, human kallikrein 4 (hK4) which is thought to carry out crucial
functions in the prostate, was expressed in this study as a secreted protein in a baculovirus
expression system, bearing a His-tag and V5-epitope that were used for purification and detection
respectively. Its mass was estimated to be 35kDa, ~2kDa less than the equivalent product
expressed in monkey kidney cells. The protein was purified to 50-90% purity with a yield of
0.93mg/L-4.8mg/L based on methods derived from computational prediction of its properties,
such as pI. Computational analysis was extended by applying high-performing computing
techniques, such as molecular dynamics, and flexible ligand docking, to predict antigenic regions,
the likely substrate specificity and putative inhibitors. These results show that hK4 has a loop,
between Leu83-Ser94 that shows promise as a specific segment that can be exploited for
generation of antibodies. Preferred substrates were also predicted to bear hydrophobic residues at
the P’-region of the scissile bond and amphiphilic residues at the P-region. At the S-region, hK4
potentially involves its unique PLYH-motif in recognizing the P4/P5 position from the substrate.
Flexible ligand-docking studies indicate that hK4 can be inhibited by inhibitors that carry a
modified bulky hydrophobic sidechain with a guanidinium group at the P1-position and its own
putative autoactivation region residues at the P2, P1’ and P2’ position.
The computational study was extended to other members of the kallikrein family, predicting
distinctions between these that could be used for future studies. These results show that 8 of the
fifteen kallikrein members are very homologous in terms of specificity bearing typical trypsin-
like activity and specificity, except for hK2, hK3, hK4, hK5, hK7, hK9, hK15 that retain certain
distinct signatures in the binding pocket in terms of secondary specificity.
The principles of substrate-specificity analysis that were developed were further applied on three
metzincins, MMP-3, ADAM-9 and ADAM-10. These three enzymes are metalloproteases, which
are involved in tissue remodeling, intracellular signalling and cell-to-cell mediation. The
substrate-specificity analysis was carried out on all three metzincins using the structure of a
crystallized complex of the MMP-3 enzyme with the TIMP-1 natural inhibitor as template. In this
specific enzyme-substrate complex, the challenge was to model and suggest a possible orientation
of the P-region, which is not known. The interactions on the P/S-region are therefore unclear and

Page 4

4
need to be clarified. In order to suggest the arrangement of the enzyme-substrate complex and the
undefined S-subsites, four new residues were added in an extended beta-sheet conformation to the
P1’ residue (derived originally from the TIMP-1 inhibitor) to create a full-length modeled
substrate spanning P4’-P4. This new modeled region, in particular, was bound through backbone
H-bonds with the enzyme at position 169 (MMP nomenclature) suggesting a new crucial residue
for substrate binding, and satisfied steric and chemical restraints in the S’-region of the enzyme.
This modeling approach also indicated a putative presence of an S2/S3-pocket on these
metzincins which is composed of different residues for MMP-3, ADAM-9 and ADAM-10, and
which could prove useful for future drug design projects. Furthermore, the data argue against the
involvement of a polarizable water molecule in catalysis, a mechanism that has been postulated
by various groups. A new catalytic mechanism is suggested to involve an oxyanion anhydride
transition state.
This study is a demonstration of the power of combining bioinformatics with wet-lab
biochemistry.

Page 5

5
TABLE OF CONTENTS
CHAPTER 1.................................................................................................................................................14
1.1 Prekallikreins................................................................................................................................14
1.1.1 Introduction to the kallikreins..................................................................................................14
1.1.2
Nomenclature.........................................................................................................................14
1.1.3 Gene structure..........................................................................................................................17
1.1.4 Individual organization and expression patterns of the kallikreins..........................................18
1.1.5 Physiological functions of the kallikreins................................................................................23
1.1.6 The kallikreins in cancer..........................................................................................................25
1.2 The kallikrein enzymes (serine proteases)...................................................................................27
1.2.1
The catalytic mechanism of serine proteases.........................................................................27
1.2.2 Structural features of kallikreins..............................................................................................31
1.2.3 Conserved residues in serine proteases....................................................................................33
1.2.4 Natural inhibition mechanism of trypsins and chymotrypsins.................................................33
1.3 Cultivation, expression and purification of kallikreins..............................................................36
1.3.1 Expression systems used for kallikrein expression..................................................................36
1.3.2 Expression of kallikreins in insect cells...................................................................................36
1.3.3 Purification procedures............................................................................................................37
1.4 The metzincins (Zinc-endopeptidases).........................................................................................38
1.4.1 Introduction to the metzincins: functions and expression patterns..........................................38
1.4.2 Natural inhibition of metzincins..............................................................................................41
1.5 Computational analysis of proteins...............................................................................................43
1.5.1 Homology modelling...............................................................................................................43
1.5.2 Enzyme-substrate modelling ...................................................................................................44
1.5.3 Force Fields.............................................................................................................................45
1.5.4 Loop modelling........................................................................................................................46
1.5.5 Docking ...................................................................................................................................46
1.5.6 Molecular design .....................................................................................................................47
1.5.7 Molecular dynamics ................................................................................................................48
1.6 Aims.................................................................................................................................................51
CHAPTER 2. ...........................................................................................................................................52
2.1 Material and Methods: Biochemical experiments.......................................................................52
2.1.1. Expression system. ..................................................................................................................52
2.1.2 Exclusion of contaminating proteins .......................................................................................52
2.1.3 Equilibration and incubation with metal affinity resin. ...........................................................52
2.1.4 Elution of hK4 from the column..............................................................................................52
2.1.5 Analysis of fractions................................................................................................................53
2.1.5.1 SDS PAGE ...........................................................................................................................53
2.1.5.2 Western blot and stripping of membranes............................................................................53
2.1.5.3 Silver-staining.......................................................................................................................54
2.1.5.4 Protein estimation and densitometry.....................................................................................54
2.2 Material and Methods: Computational experiments. ................................................................54
2.2.1 Sequence analysis and alignments...........................................................................................54
2.2.2 Protein modelling ....................................................................................................................54
2.2.2.1 Modelling the Kallikreins......................................................................................................54
2.2.2.2 Modelling ADAM-9 and ADAM-10.....................................................................................56
2.1.3 Enzyme-substrate modelling and subsite analysis...................................................................56
2.1.3.1 Kallikreins. ............................................................................................................................56
2.1.3.2 Metzincins .............................................................................................................................56
2.1.4 Molecular design .....................................................................................................................57
2.1.5 Docking ...................................................................................................................................58