Citation
Site-directed mutagenesis of the aspartic proteinase rhizopuspepsin : an analysis of unique specificity

Material Information

Title:
Site-directed mutagenesis of the aspartic proteinase rhizopuspepsin : an analysis of unique specificity
Alternate title:
Analysis of unique specificity
Creator:
Lowther, William Todd, 1967-
Publication Date:
Language:
English
Physical Description:
xv, 138 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Active sites ( jstor )
Amino acids ( jstor )
Biochemistry ( jstor )
Enzyme substrates ( jstor )
Enzymes ( jstor )
Hydrogen bonds ( jstor )
Kinetics ( jstor )
pH ( jstor )
Proteins ( jstor )
Substrate specificity ( jstor )
Aspartic Endopeptidases -- chemistry ( mesh )
Aspartic Endopeptidases -- genetics ( mesh )
Aspartic Endopeptidases -- physiology ( mesh )
Department of Biochemistry and Molecular Biology thesis Ph.D ( mesh )
Dissertations, Academic -- College of Medicine -- Department of Biochemistry and Molecular Biology -- UF ( mesh )
Mutagenesis, Site-Directed ( mesh )
Rhizopus ( mesh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1994.
Bibliography:
Includes bibliographical references (leaves 129-137).
Additional Physical Form:
Also available online.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by William Todd Lowther.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright William Todd Lowther. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
50543947 ( OCLC )
028825855 ( ALEPH )

Downloads

This item has the following downloads:


Full Text













SITE-DIRECTED MUTAGENESIS OF THE ASPARTIC PROTEINASE RHIZOPUSPEPSIN: AN ANALYSIS OF UNIQUE SPECIFICITY














By


WILLIAM TODD LOWTHER
















A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY



UNIVERSITY OF FLORIDA 1994


































This work is dedicated to all those who have given freely of
their love, support, guidance and patience; my Lord and
Savior Jesus Christ, family, friends and colleagues.















ACKNOWLEDGMENTS



I have always been encouraged to be my best and to be true to myself. This support has extended from family and friends. In particular, I want to thank my mother and father, Ruth and Mike, for their love, generosity and willingness to do whatever it took to make sure that my brothers, Jason and Pat, and I had everything we could have ever needed or wanted. I also want to thank Bill, Dee, Brandon and Kara for their encouragement and love. My grandparents have also been a continual source of wisdom, support and love. Encouragement from Andy and Cheryl has also been very reassuring.

I am indebted to Dr. Ben Dunn for providing a fantastic environment for learning and scientific opportunity. I am particularly thankful for his insights and willingness to allow me to pursue my research interests, within reason.

I would like to thank my committee members for their suggestions and encouragement during my graduate work; Drs. Daniel Purich, Charles Allen, Sheldon Schuster, and Nigel Richards.

The support of the Dunn laboratory has made my stay at the University of Florida one of the best times of my life. In particular, I want to thank Chetana, Paula, Wieslaw, Bill,



iii










Wichet, Brian, Jenny, and the whole rest of gang for their friendships. The efforts and friendships of the Protein Chemistry Core Facility have also made my stay in Gainesville enjoyable. In particular, I want to thank Ruth, Nancy, Hung, Benne, and the whole rest of the gang.













































iv
















TABLE OF CONTENTS


ACKNOWLEDGMENTS ........................................... iii

LIST OF TABLES ............................................ iii

LIST OF FIGURES .. ...........................................


KEY TO SYMBOLS .............................................xii

ABSTRACT ................................... .... .......... xiv

CHAPTERS

1 HYDROGEN BONDING AND BIOLOGICAL SPECIFICITY .............. 1
Introduction ............................................. 1
Tyrosyl tRNA Synthetase ................................. 3
Trypsin ....................................................5
Aspartic Proteinases ...................................7
General Characteristics .............................7
Initial Kinetic Studies ..............................12
Rhizopuspepsin .....................................................14
Historical Background ................................15
Specificity Differences Studied by Site-Directed
Mutagenesis ...................................... 16


2 EXPERIMENTAL PROCEDURES ............................... 23
Introduction ..........................................23
Materials ...............................................23
Methods .............................................. 25
Cloning and Mutagenesis ............................. 25
Expression ......................................... 30
Refolding ........................................30
Size-Exclusion Chromatography ...................... 31
Activation and Ion-Exchange Chromatography .........32 Structural Characterization ........................ 32
Kinetic Analysis ...................................37
Analysis of Transition State Effects ............... 40
Molecular Graphics .................................41


3 EXPRESSION, REFOLDING, PURIFICATION, AND ACTIVATION
OF RECOMBINANT RHIZOPUSPEPSINS ........................42
Introduction ............................................42



v










Results ...... ............................ 43
Mutagenesis ....................................... 43
Expression and Refolding ..........................43
Activation and Purification ....................... 50
Discussion ............................................59


4 KINETIC AND STRUCTURAL AUTHENTICITY OF RECOMBINANT
RHIZOPUSPEPSINS ............................................. 65
Introduction ................................................ 65
Results ..... ........................ ......... 66
WT-REC and Native Isozymes of Rhizopuspepsin ....... 66 Structural Comparisons ..............................68
Discussion ................................................75


5 ANALYSIS OF THE SPECIFICITY OF RHIZOPUSPEPSIN
THROUGH THE USE OF INHIBITORS CONTAINING P1 AND P1'
SUBSTITUTIONS AND PEPTIDE BOND MIMETICS ...............80
Introduction .......................................................80
Results and Discussion ................................. 81


6 THE BROAD SUBSTRATE SPECIFICITY OF RHIZOPUSPEPSIN:
ANALYSIS WITH SYSTEMATIC SUBSTITUTIONS IN P5-P1 AND
P2'-P3' OF THE SUBSTRATE LYS-PRO-ALA-LYS-PHE*NPHARG-LEU ............... .................................. 91
Introduction ........................................................91
Results ................................................92
Lys-Pro-Ala-Lys-Phe*Nph-Arg-Leu based substrates ...92 Lys-Pro-Ile-P2-Phe*Nph-Arg-Leu based substrates .....96
Discussion .............................................. 96
Comparison to the Mammalian Enzymes ................98
Comparison to the Candida Aspartic Proteinases ..... 99


7 ENGINEERING THE SUBSTRATE SPECIFICITY OF
RHIZOPUSPEPSIN: THE ROLE OF ASP30 AND ASP77 OF A FUNGAL ASPARTIC PROTEINASE TO CLEAVE SUBSTRATES
WITH LYSINE IN P1 ....................................... 101
Introduction .........................................101
Results .............................................. 104
Kinetic Analysis of the Recombinant
Rhizopuspepsins .................................104
Kinetic Analysis of Porcine Pepsin ................ 109
Discussion .........................................110
Substrate Design .......................................111
Kinetic Analysis ......................................112


8 CONCLUSIONS AND FUTURE DIRECTIONS ....................123
Conclusions ..........................................123
Future Directions .........................................125



vi













LIST OF REFERENCES ... ........ ........ ................. ........129

BIOGRAPHICAL SKETCH ....................................... 138


























































vii















LIST OF TABLES

Table joage

1-1. Crystal structures of native and inhibitorcomplexed rhizopuspepsin ..............................10

1-2. Primary sequence comparison of rhizopuspepsin to several aspartic proteinases ....................... 18

1-3. Partial sequence alignment of several aspartic proteinases ...................... ................. 20

3-1. Representative yields during the purification of the recombinant rhizopuspepsins ..................... 63

4-1. Kinetic comparison between the naturally occurring isozymes and WT-REC rhizopuspepsin
using the substrate Lys-Pro-P3-Lys-Phe*Nph-ArgLeu ................................................ 69

4-2. Guanidinium hydrochloride denaturation parameters of native and mutant forms of rhizopuspepsin ....... 74

5-1. Inhibition constants for XaaY[CH2NH]Yaa modified derivatives ........................................ 83

5-2. Inhibition constants for Leu'I[CH(OH)CH2]Val and statine modified derivatives ........................ 84

6-1. Kinetic parameters for WT-REC rhizopuspepsin with the substrate Lys-Pro-Ala-Lys-Phe*Nph-Arg-Leu
containing systematic substitutions in P5-P and
P2'-P3' at pH 3.5 .................................. 94

6-2. Kinetic analysis of WT-REC rhizopuspepsin with substrates of the form Lys-Pro-Ile-P2-Phe*NphArg-Leu at pH 3.5 .................................. 97

7-1. Kinetic analysis of wild-type and mutant rhizopuspepsins: systematic substitution of
lysine into P3-P1 .................................. 106

7-2. Kinetic analysis of porcine pepsin: systematic substitution of lysine into P3-P1 .................. 108



viii









7-3. Transition state stabilization energy changes
seen with variation in pH from 3.5 to 5.0 for the
recombinant rhizopuspepsins and porcine pepsin ..... 113

7-4. Double mutant cycle analysis of the recombinant
rhizopuspepsins: substitution of lysine into P3P1 at pH 3.5 and 5.0 ............................... 117















































ix















LIST OF FIGURES

Figure page

1-1. Ribbon representation of the aspartic proteinase rhizopuspepsin complexed with a reduced peptide
bond inhibitor ...................................... 9

1-2. Closeup view of the reduced peptide bond inhibitor bound to the active site of
rhizopuspepsin ................... ..... ............. 11

1-3. Closeup view of the active site of rhizopuspepsin highlighting the catalytic aspartic acid
residues, Asp32 and 215, and Asp30 and Asp77 ....... 21

2-1. PCR mutagenesis procedure ........................... 26

2-2. pET3aE expression vector containing
rhizopuspepsinogen ................................. 29

3-1. SDS-PAGE analysis of the expression of wild-type
rhizopuspepsinogen (RPGN) in E. coli upon the
addition of IPTG from 0 to 3 hours ................. 46

3-2. SDS-PAGE analysis of wild-type rhizopuspepsinogen
at different stages of purification ................ 47

3-3. Gel filtration elution profile of refolded
Asp30Ile rhizopuspepsinogen ........................ 49

3-4. SDS-PAGE analysis of the fractions from the
purification of Asp30Ile by gel filtration in
Figure 3-3 ......................................... 52

3-5. SDS-PAGE analysis of the time course of
activation of wild-type rhizopuspepsinogen upon
lowering the pH at room temperature ................ 54

3-6. SDS-PAGE analysis of the activation of wild-type
recombinant rhizopuspepsinogen at pH 4.0 ........... 56

3-7. IEF analysis of the wild-type and mutant
rhizopuspepsinogens activated at pH 2.0 ........... 58

3-8. Elution profile of Asp77Thr from the Mono S
column after activation at pH 3.0 ................... 61



x









3-9. IEF comparison of the activated, purified
recombinant rhizopuspepsins ........................ 62

4-1. IEF comparison of wild-type rhizopuspepsinogen,
activated, purified WT-REC and the two naturally
occurring isozymes ................................. 67

4-2. CD spectra of the recombinant wild-type and
mutant forms of rhizopuspepsin ...................... 71

4-3. Fluorescence emission spectra for folded (0 M
GdnHC1) and unfolded (6 M GdnHC1) wild-type
recombinant rhizopuspepsin ......................... 72

4-4. Guanidinium hydrochloride induced unfolding of
the naturally occurring isozyme pI 6 and the
recombinant forms of rhizopuspepsin monitored by
the change in intrinsic fluorescence at 350 nm ..... 73

7-1. Hydrogen bonding interactions in penicillopepsin
between Asp77, Ser79 and the pepstatin derivative
containing lysine in the P1 position ............... 119

7-2. Ca carbon backbone superposition of
penicillopepsin and rhizopuspepsin complexed with
inhibitors ......................................... 120

7-3. Proposed hydrogen bonding in WT-REC
rhizopuspepsin and Asp77Thr mutants with
substrates containing lysine in P1 ................. 121

























xi














KEY TO SYMBOLS


amp ampicillin AU absorption units C carboxyl Ca alpha carbon CAPS 3-(cyclohexylamino)-l-propanesulfonic acid CD circular dichroism d deoxy DMSO dimethylsulfoxide E. coli Escherichia coli
EDTA ethylenediaminetetraacetic acid FPLC fast protein liquid chromatography h Planck's constant HC1 hydrochloric acid HIV human immunodeficiency virus HPLC high performance liquid chromatography IEF isoelectric focusing IPTG isopropylthio-0-D-galactopyranoside kcal kilocalorie kb Boltzmann's constant
kcat turnover number KC1 potassium chloride Ki inhibition constant Km Michaelis-Menten constant LB Luria Broth MES 2-(4-morpholino)-ethane sulfonic acid mg milligram MgC12 magnesium chloride min minutes






xii

















ml(s) milliliter(s) mM millimolar MOPS 3-(N-morpholino) propane-sulfonic acid MWCO molecular weight cut off N amino NaCl sodium chloride ng nanogram nM nanomolar nm nanometers Nph p-nitrophenylalanine NTP nucleotide triphospahtes OD optical density ori origin PAGE polyacrylamide gel electrophoresis PI isoelectric point pmol picomoles PVDF polyvinylidene difluoride rec recombinant s second SDS sodium dodecyl sulfate sec seconds tet tetracycline TFA trifluoroacetic acid Tricine N-[Tris-(hydroxymethyl) Methyl] glycine Tris tris (hydroxymethyl) aminomethane Vmax maximum velocity WT-REC wild-type recombinant 9g microgram gM micromolar





xiii















Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy SITE-DIRECTED MUTAGENESIS OF THE ASPARTIC PROTEINASE RHIZOPUSPEPSIN: AN ANALYSIS OF UNIQUE SPECIFICITY

By

William Todd Lowther

August, 1994




Chairman: Dr. Ben M. Dunn
Major Department: Biochemistry and Molecular Biology


Rhizopuspepsin is a secreted aspartic proteinase from the fungus Rhizopus chinensis. Despite the high degree of structural homology among the aspartic proteinases, the amino acid residues that create the individual binding pockets have been shown to profoundly influence substrate specificities and inhibitor preferences. Rhizopuspepsin and other fungal aspartic proteinases are distinct from the mammalian enzymes in that they are able to cleave substrates with lysine in the P1 position. Sequence and structural comparisons suggest that two aspartic acid residues, Asp77 and Asp30 (pig pepsin numbering), may be responsible for generating this unique specificity in rhizopuspepsin.

In order to analyze their contributions to specificity, Asp30 and Asp77 were changed to the corresponding residues in


xiv










porcine pepsin, Ile30 and Thr77, to create single and double mutants. The zymogen forms of the wild-type and mutant forms of the enzymes were overexpressed in E. coli as inclusion bodies. Following denaturation, refolding, activation and purification to homogeneity, structural and kinetic comparisons were made. These comparisons have shown that the wild-type recombinant enzyme is kinetically and structurally indistinguishable from the naturally occurring isozymes. The mutant proteins were also shown to exhibit a high degree of similarity.

Characterization of the wild-type and mutant proteins with chromogenic substrates with systematic substitution of lysine into the P1, P2 and P3 positions has shown that Asp77 is the predominant residue responsible for enabling the catalysis of substrates with lysine in P1. Mutation of Asp77 resulted in a loss of 7 kcal mol-I of transition state stabilization energy. The Asp30Ile mutant was still able to cleave the PI-Lys peptide with near wild-type efficiency. These observations suggest that it may be possible to exploit the Asp77 interaction to design compounds that preferentially inhibit a variety of related, secreted Candida proteinases in immunocompromised patients.











xv















CHAPTER 1
HYDROGEN BONDING AND BIOLOGICAL SPECIFICITY


Introduction


Biological processes are controlled by a variety of forces that influence intramolecular and intermolecular interactions. A balance of the positive and negative aspects of these forces, for example, determines the structure of DNA and proteins, establishes the specificity of binding interactions needed for information transfer and substrate recognition, and creates the environmental requirements for enzymatic catalysis. These forces have been classically divided into two categories: (1) forces which lead to binding energy, dispersion or van der Waals forces and the hydrophobic effect, and (2) forces which relate to specificity or the discrimination of one molecule over another, electrostatic interactions and hydrogen bonds (Jencks, 1969; Fersht, 1985; Fersht et al., 1985; Dill, 1990).

The contribution of van der Waals forces and the hydrophobic effect to protein folding and binding interactions is thought to extend primarily from an increase in entropy upon release of water molecules to the bulk solvent. Even though there is, for example, a large decrease




1






2



in entropy in the organizational process of protein folding, the release of water molecules and the establishment of an intricate network of intramolecular hydrogen bonds compensates for the entropy loss making the entire process slightly favorable overall. The observation that almost all possible hydrogen bonds are established upon the folding of a protein points to the difficulty, in separating and quantifying the forces involved. Despite this difficulty trends are still observable. A survey of many x-ray crystal structures of ligands bound to proteins has shown that there is a strong correlation between the binding strength of ligands and the degree of buried surface area (Horton & Lewis, 1992; Young et al., 1994). This correlation is dependent on the geometric or steric requirements for the optimization of packing and hydrophobic interactions, but is also influenced by the chemical complementarity between the two species.

Chemical complementarity and biological specificity in protein-ligand interactions are primarily governed by hydrogen bonds and electrostatic interactions (Fersht et al., 1985). A hydrogen bond occurs when two electronegative atoms share a hydrogen atom. The optimal arrangement for this interaction is linear with a distance between the electronegative atoms ranging from 2.85 to 3 angstroms. The study of the contribution of hydrogen bonds to specificity has been facilitated by the use of site-directed mutagenesis. The influence of hydrogen bonding in recognition and






3



catalysis has been studied by deleting a hydrogen bond donor or acceptor on the enzyme or by making similar modifications to the substrate. Two systems which have been extensively studied are tyrosyl-tRNA synthetase and the pancreatic proteinase trypsin.


Tvrosvl-tRNA Synthetase


One of the most thoroughly studied model systems for

studying the role of hydrogen bonding in specificity has been the tyrosyl-tRNA synthetase from Bacillus stearothermophilus. The tyrosyl-tRNA synthetase ensures the fidelity of information transfer from the genetic code to the final protein product by optimizing interactions with the structural components of the amino acid tyrosine that make it different from the other amino acids and, in particular, phenylalanine. The examination of the crystal structure of the enzyme bound aminoacyl adenylate has shown that there are eight hydrogen bonds between the enzyme and the substrate which can be studied by mutation of the enzyme. Fersht and his coworkers have analyzed the effects of mutations by comparing the kcat/Km values of the wild-type enzyme to the mutant enzyme for the activation of tyrosine and phenylalanine (Carter et al., 1984; Fersht, 1985, 1988; Fersht, et al., 1985; Leatherbarrow et al., 1985; Lowe et al., 1987). This comparison gives information about the overall apparent change in transition state stabilization






4



energy or the binding of the substrate in the transition state. A true comparison of the binding energies is not possible because of the different interactions of the wildtype and mutant proteins with water. From the systematic analysis of the mutant proteins, however, a pattern of transition state stabilization free energy changes emerged. The magnitude of the change was dependent on the type of mutation and whether or not the hydrogen bond donor/acceptor was charged or neutral.

The deletion of an enzyme side chain or substrate

hydrogen bond to an uncharged hydrogen bond donor/acceptor resulted in a loss of transition state stabilization energy of 0.5 to 1.8 kcal mol-1. This decrease represents a factor of 2 to 15 toward specificity. The deletion of hydrogen bond to a charged hydrogen donor/acceptor resulted in the loss of

3 to 6 kcal mol-I of stabilization energy representing a factor of 1000 or more in specificity. The same loss of specificity was seen when both the donor and acceptor were charged.

The results from these studies suggest that hydrogen bonding between uncharged donor/acceptor do contribute to specificity to some extent. Hydrogen bonds containing at least one charged donor/acceptor, however, contribute significantly to specificity. Further evidence to support the critical role of electrostatic interactions in the creation of specificity comes from the study of the interaction of proteinases with substrates and inhibitors.






5



Mutagenesis studies have been undertaken with trypsin in order to understand the source of its unique specificity among serine proteinases.


Trvsin


Trypsin provides an excellent example of how an enzyme has optimized electrostatic interactions to recognize and preferentially cleave a particular class of substrates. The primary specificity of trypsin is to cleave substrates containing Arg or Lys in the P1 position (nomenclature of Schechter & Berger, 1967). The x-ray crystal structures of inhibitor complexes of trypsin have given insight into this preference for basic residues (Ruhlmann et al., 1973; Krieger et al., 1974; Perona et al., 1993, 1994). The bottom of the S1 binding pocket contains an Asp at position 189. When lysine is present in the substrate, hydrogen bonding occurs directly to Serl90 and through a water molecule to Asp189. The presence of Arg in the substrate, however, expels the water molecule and hydrogen bonding occurs directly to Asp189 and Serl90.

Extensive site-directed mutagenesis studies have been

performed to understand the contribution of Asp189 and Serl90 to the specificity of trypsin. Mutagenesis has also been carried out to see if changing Asp189 converts trypsin into a chymotrypsin-like enzyme. The S1 binding pockets of chymotrypsin and trypsin show a high degree of structural






6



homology. Despite this similarity, chymotrypsin has a primary specificity for large aromatic residues. The most notable substitution that may be responsible for the specificity differences between trypsin and chymotrypsin occurs at position 189, where Asp has been replaced by Ser.

The mutation of Asp189 to Ser in trypsin resulted in a dramatic decrease in the kcat/Km values for P1 Arg and Lys containing substrates (Graf et al., 1988). The average loss of transition state stabilization energy as a consequence of this mutation was 6.7 kcal mol-1. This value is in the same range as those seen by Fersht for the deletion of a charged hydrogen bond donor/acceptor discussed above. Experiments were also performed to attempt to rescue the basic residue specificity. The substitution of Asp at position 190 did restore activity (Evnin et al., 1990). Experiments have also shown that activity can be restored when acetate is present in the buffer at very high levels (Perona, et al., 1994).

Interestingly, the Aspl89Ser mutant did show some

improvement toward cleaving large hydrophobic substrates. Further mutagenesis experiments tried to complete the conversion by mutating the remaining residues in the S1 pocket of trypsin to those in chymotrypsin. This effort still did not result in a complete metamorphosis to chymotrypsin. Only upon changing surface loops around P1 subsite of trypsin to those of chymotrypsin in conjunction with the P1 substitutions resulted in the desired enzyme specificity (Hedstrom et al., 1992).






7



The studies with tyrosyl-tRNA synthetase and trypsin are only two examples of the extensive literature establishing the importance of hydrogen bonding and electrostatic interactions in the creation of specificity in biological reactions. Mutational and kinetic analysis have also been used to dissect the structural components that generate the unique specificities of cysteine and metallo proteinases. Studies on the aspartic proteinases have the potential to give further insight into the role of electrostatic interactions in specificity. In particular, rhizopuspepsin, the aspartic proteinase from the fungus Rhizopus chinensis, contains several aspartic acid residues in the active site that may influence the selection of substrates for hydrolysis.


Aspartic Proteinases



General Characteristics


All members of the aspartic proteinase family show marked inhibition by pepstatin (Davies, 1990) and a high degree of amino acid sequence and three-dimensional structure homology, especially in the region of the active site, as shown by the studies of Blundell et al. (Pearl & Blundell, 1984; Blundell et al., 1987; Sali et al., 1989). The bilobal structure of an aspartic proteinase is created by the nearly symmetrical N-terminal and C-terminal domains of the protein.






8



The extended cleft of the active site is formed by the interaction of the two domains (Figure 1-1). Each domain also contributes one catalytic aspartic acid at the bottom of the active site. An elaborate network of hydrogen bonds maintains these aspartic acid residues (Asp32 and Asp215, porcine pepsin numbering) in a juxtaposed or opposing orientation. A centrally located water molecule, hydrogenbonded to each aspartic acid residue, is thought to act as the nucleophile in a base-catalyzed attack of the scissile bond carbonyl of the substrate (Suguna et al., 1987; Fraser et al., 1992; James et al., 1992).

From the examination of inhibitor complexes of

rhizopuspepsin (Table 1-1), porcine pepsin (Abad-Zapatero et al., 1991), endothiapepsin (Veerapandian et al., 1990; Lunney et al., 1993), cathepsin D (Baldwin et al., 1993; Metcalf & Fusek, 1993) and the HIV proteinase (Swain et al., 1990, 1991), it is evident that there is a consistent binding mode for ligands (Figure 1-2). Ligands seven to eight residues long completely fill the active site in an extended B-strand conformation with the amino acid side chains alternating in a regular fashion. This uniform binding or anchoring of ligands to the active site is attributed to a highly conserved hydrogen bonding network, between the enzyme and the a-carbon backbone of ligands, and the preference for large hydrophobic or aromatic substituents on either side of the scissile bond (P1-P1') or the site of cleavage.






9










































Figure 1-1. Ribbon represention of the aspartic proteinase rhizopuspepsin complexed with a reduced peptide bond inhibitor. The catalytic aspartic acid residues, Asp32 and Asp215, are shown in red. The flap which extends over the active site is shown in orange. The inhibitor is shown in yellow. Coordinates were obtained from the Brookhaven Protein Data Bank file 3APR (Suguna et al., 1987).






10


















Table 1-1. Crystal structures of native and inhibitorcomplexed rhizopuspepsin

Structure Resolution R value Reference
(A)



Native 1.8 0.143 Suguna et al., 1987a Pepstatin 2.5 0.145 Bott et al., 1982 U70531E 1.8 0.147 Suguna et al., 1987b CP-69,799 2.0 0.171 Parris et al., 1992 CP-88,218 1.9 0.175 Parris et al., 1992 U85548E 2.0 0.170 unpublished results






11








































Figure 1-2. Closeup view of a reduced peptide bond inhibitor bound to the active site of rhizopuspepsin (Suguna et al., 1987b). This figure illustrates the binding mode of ligands to the active site of aspartic proteinases and illustrates the Schechter and Berger nomenclature for describing active site interactions. For example, the side chain of the P3 residue of the ligand interacts with the S3 subsite of the enzyme. Bond cleavage occurs between P1 and P11'.






12


Initial Kinetic Studies


The initial studies characterizing the hydrolytic properties of mammalian and fungal aspartic proteinases utilized small tri- and tetrapeptides (Fruton, 1970, 1976). Even with these small, poorly binding substrates, differences in the "secondary interactions," those interactions not at the scissile bond, were seen for pepsin, cathepsin D and rhizopuspepsin (Voynick & Fruton, 1971; Fruton, 1976). Subsequent studies (Sampath-Kumar & Fruton, 1974; Hofmann et al., 1988; Balbaa et al., 1993) showed an increase in kcat as the substrate length was extended to eight residues, particularly when the S3, S2 and S2' subsites were occupied. A further enhancement of substrates for the study of this family of enzymes came when Hofmann and Hodges (1982) showed that the change in absorbance of a p-nitrophenylalanine residue in the P1' position of the substrate would be greater upon substrate hydrolysis than when present in the P1 position (Hofmann et al., 1984). A slight modification of Hofmann's substrates which contained lysine in P1 was made by Dunn and coworkers (Dunn et al., 1984) using the information from a large survey by Powers et al. (1977) on all known cleavage junctions of the pepsin at that time. This work yielded a chromogenic substrate, Pro-Thr-Glu-Phe*Nph-Arg-Leu (Nph = p-nitrophenylalanine), that could be cleaved by porcine pepsin in a continuous assay allowing the quantitation of initial rates. The binding mode of this






13



peptide was confirmed by the crystal structure of the inhibitor H-256, Pro-Thr-Glu-Phe'F[CH2NH] Phe-Arg-Glu, bound to the active site of endothiapepsin (Cooper et al., 1987).

Studies soon followed delving deeper into the variations seen in the specificity at different subsites. A systematic series of peptides with substitutions in the P3 position was made by Dunn et al. (1986), in the peptide Lys-Pro-Xaa-GluPhe*Nph-Arg-Leu, and screened against a large variety of mammalian and fungal enzymes. The results from these experiments again showed a large diversity of specificity in the S3 subsite. In another study, the pH dependence of the kinetic parameters was studied with pepsin, chymosin, and endothiapepsin cleaving substrates of the form Lys-Pro-XaaYaa-Phe*Nph-Arg-Leu where Xaa-Yaa were Ala-Glu, His-Ala and Thr-Val (Dunn et al., 1987). From the examination of the endothiapepsin structure complexed with H-256, and the trends seen in the data, it was proposed that specific electrostatic interactions with Glul3 (pepsin numbering) of the S3 pocket of porcine pepsin were occurring. These studies also suggested that electrostatic interactions may also be important in the S2 subsite of porcine pepsin. Additional studies by Pohl & Dunn (1988) provided further kinetic evidence for electrostatic interactions in the S3 and S2 subsites of porcine pepsin from the study of the substrates derived from Lys-Pro-Ala-Lys-Phe*Nph-Arg-Leu. Now that the genes have been cloned for a variety of aspartic proteinases, including rhizopuspepsin, the structure-activity






14



relationships proposed in these studies can be tested by site-directed mutagenesis and the heterologous expression of these proteins.


RhizopusDeDsin


Rhizopuspepsin, a fungal aspartic proteinase from

Rhizopus chinensis, has been the subject of many substrate, inhibitor and crystallographic studies in an effort to better understand factors contributing to catalysis and specificity differences between members of this enzyme family. Initial interest in the fungal enzymes arose from the need to find a suitable replacement for calf chymosin in the cheese-making process (Davies, 1990). This focus changed with the determination of the native structure (Suguna et al., 1987) and many different inhibitor complexes of rhizopuspepsin (Suguna, et al., 1987; Parris et al., 1992; Suguna et al., 1992) and other fungal enzymes (James et al., 1982; Cooper, et al., 1987; Lunney, et al., 1993). The insights gained from these complexes have led to proposals for the reaction mechanism (James et al., 1977, 1992; Suguna, et al., 1987 Fraser et al., 1992;) and the rationalization of subsite preferences of the aspartic proteinases (Rao et al., 1993; Scarborough et al., 1993). The goal has been to use this information to design specific or targeted therapeutics for renin and the HIV proteinase. Further clinical interest in rhizopuspepsin comes from its similarity to the secreted






15



aspartic proteinases of several opportunistic Candida species (Fusek et al., 1993; Morrison et al., 1993). The knowledge gleaned from studying rhizopuspepsin has the potential to foster the design of anti-fungal agents for use in treating vaginal infections and immunocompromised AIDS, organ transplant and cancer patients (Samaranayake & Holmstrup, 1989; Saral, 1991; Paya, 1993). Historical Backaround


Rhizopuspepsin was first isolated by Fukumoto et al. (1967). Purification of rhizopuspepsin by isoelectric focusing identified two major isozymes (Graham et al., 1973). These two isozymes have been shown to be very similar in molecular weight, amino acid composition, specific activity, as well as in three-dimensional structure as shown by the crystal structure of the isozyme mixture (Subramanian et al., 1977). N- and C-terminal sequencing showed that the two isozymes were identical (Sepulveda et al., 1975). Grippon et al. (1977) showed the first structural difference at residue 12 (pepsin numbering). Isozyme pI 5 has a Ile residue at position 12 while pI 6 has a Val. Delaney et al. (1987) solved the complete sequence of isozyme pI 6 by a combination of amino acid sequencing of HPLC-purified CNBr cleavage fragments by Edman degradation (154 residues/325 directly sequenced) and the DNA sequencing of a positive cDNA clone, 33E2. Subsequent work by Takahashi's group established the






16



complete sequence of both pI 5 and 6 by the 100% chemical sequencing of HPLC-purified trypsin and Staphylococcus aureus V8 protease generated peptide fragments (Takahashi, 1987, 1988). Rhizopuspepsin pI 5 and pI 6 differ only at eight positions in the entire 325 amino acid sequence. All substitutions are semi-conservative and only the Val/Ile 12 residue is in the active site.

The work by Chen et al. (1991) has lead to the production of the zymogen of rhizopuspepsin, rhizopuspepsinogen, in several different expression systems. The zymogen form is expressed to facilitate correct folding. Activation at low pH converts the zymogen to the active form in combination of inter and intramolecular processes yielding enzyme for kinetic studies. Kinetic studies comparing various mutant rhizopuspepsin enzymes to the native enzyme will help resolve the role hydrogen-bonding, electrostatic, and hydrophobic interactions play in the creation of specificity in the active site subsites SI, S2 and 53 of rhizopuspepsin as well as give clues for understanding these interactions in the other aspartic proteinases.


Specificity Differences Studied by Site-directed Mutaaenesis


Despite the high degree of primary sequence and

structural homology within the aspartic proteinases in and around the active site (Table 1-2), differences are observed in substrate specificities. The primary specificity of






17



aspartic proteinases is for cleavage between large hydrophobic-hydrophobic junctions, such as, Phe-Phe (Fruton, 1970, 1976). Secondary interactions have been shown to cause large increases in catalytic efficiency, kcat, without corresponding changes in Km (Balbaa et al., 1993). These interactions, though sometimes far from the site of cleavage, as previously discussed have been implicated in dictating highly specific preferences for substrates. Cathepsin D, for example, is unable to efficiently cleave substrates with basic residues in the P2 position (Scarborough et al., 1993). This effect is thought to be generated by the presence of a methionine residue at position 287 (pepsin numbering, Scarborough et al., 1994). Porcine pepsin exhibits a similar aversion for positive residues, but in the S3 subsite. This effect was demonstrated, through the use of mutagenesis and pH dependence kinetic studies, to be mediated by Glul3 (Pohl & Dunn, 1988; Rao-Naik, unpublished results).

Rhizopuspepsin and other fungal enzymes are different

from these examples and other mammalian enzymes in that they are able to cleave a wide range of substrates with similar efficacy and 2-10 fold higher kcat/Km values (Dunn et al., 1986: Lowther et al., 1991). Even though rhizopuspepsin has broad specificity, it does possess an additional characteristic that distinguishes itself from all mammalian aspartic proteinases.






18



















Table 1-2. Primary sequence comparison of rhizopuspepsin to several aspartic proteinases

Enzyme % identity % similarity


porcine pepsin 39.2 62.7 human pepsin 40.5 62.3 human cathepsin E 37.2 59.7 human cathepsin D 33.1 58.8

human renin 27.5 51.6 Candida albicans 28.1 49.4 aspartic proteinase






19



Rhizopuspepsin, as well as the majority of all other fungal enzymes, has the ability to cleave substrates (Hofmann et al., 1984; Balbaa et al., 1993) and to bind inhibitors with lysine in the P1 position (Salituro et al., 1987).

Structural and sequence comparisons suggest two possible residues in the active site that may account for this unique capacity. Table 1-3 shows an alignment of sequences from several aspartic proteinases in the active site and flap regions. Besides the two catalytic aspartic acids, Asp32 and Asp215, rhizopuspepsin has additional aspartic acid residues at positions 30, 37, and 77. Asp30 is situated at the boundary between the S3 and S1 subsites and has the potential to interact with positively charged residues in the P3 and the P1 positions of the substrate (Figure 1-3). Asp37 is located in the S2' subsite and probably does not directly affect primary specificity. Asp77, which is also present in the Candida enzymes, is positioned at the end of the flap and points down into the active site cavity making potential interactions with the P3, P2 and P1 positions of the substrate.

A precedent exists for the importance of Asp77 in the primary specificity of fungal enzymes from a complex of penicillopepsin with a pepstatin derivative containing lysine in P1 (James et al., 1984). In this structure the Asp77 side chain and the enzyme backbone NH hydrogen bond in a highly conserved manner to the P2 NH and the P2 carbonyl of the inhibitor backbone, respectively.






20













Table 1-3. Partial sequence alignment of several aspartic proteinasesa

HPEP 28TVVFDTGSSN37 74TYGTG78
PPEP TVIFDTGSSN TYGTG CATE TVIFDTGSSN QYGTG CATD TVVFDTGSSN HYGSG HREN KVVFDTGSSN RYSTG RCAP NLDFDTGSSD SYGDG CAAP NVIVDTGSSD GYGDG CTAP TVVIDTGSSD EYGDL CPAP TVIIDTGSSD RYGDG

aThis alignment was obtained using the PILEUP program, a module in the GCG Sequence Analysis Software Package (Devereux et al., 1984). The sequences are HPEP = human pepsin (Sogawa et al., 1983), PPEP = porcine pepsin (Chen et al., 1975), CATE = human cathepsin E (Azuma et al., 1992), CATD = human cathepsin D (Faust et al., 1985), HREN = human renin (Hobart et al., 1984), RCAP = Rhizopus chinesis aspartic proteinase (Chen et al., 1991), CAAP = Candida albicans aspartic proteinase (Hube et al., 1991), CTAP = Candida tropicalis asapartic proteinase (Togni et al., 1991), CPAP = Candida parapsilosis (de Viragh et al., 1993).






21














































Figure 1-3. Closeup view of the active site of rhizopuspepsin highlighting the catalytic aspartic acid residues, Asp32 and 215, and Asp30 and Asp77.






22



In contrast to the mammalian and fungal enzymes that contain either serine or threonine at position 77, the Asp77 residue was shown to be able to make an additional contact through its side chain by hydrogen bonding to the E-amino nitrogen of the lysine residue. Hydrogen bonds are also seen between the lysine residue and Ser79.

This study has focused on the contributions of Asp30 and Asp77 to fungal specificity through the use of site-directed mutagenesis. These residues were changed in the rhizopuspepsinogen gene to those present in porcine pepsin, Ile30 and Thr77. The proteins were overexpressed in E. coli, refolded from inclusion bodies, activated and purified for structural and kinetic comparisons. A series of systematically substituted substrates with lysine in either PI, P2 or P3 was assayed and analyzed by double mutant cycles (Carter et al., 1984; Wells, 1990) in order to ascertain and confirm the predominant interactions enabling substrate catalysis.















CHAPTER 2
EXPERIMENTAL PROCEDURES


Introduction


This chapter outlines the materials and methods used to characterize the unique specificity of rhizopuspepsin toward substrates and inhibitors. This study has used a combination of systematically substituted substrates and inhibitors and site-directed mutagenesis to accomplish this task.


Materials

Restriction and modifying enzymes were purchased from Promega, Life Technologies, Inc., New England Biolabs or United States Biochemical Corp. Deoxyadenosine-5'-[a-35S] thiotriphosphate, as its triethylammonium salt (Sp isomer, 1000 Ci/mmol), was purchased from Amersham Corp. The pet3a expression vector (Studier et al., 1990), containing the wild-type rhizopuspepsinogen gene (Chen et al., 1991), was kindly provided by Jordan Tang at the Oklahoma Medical Research Foundation, Oklahoma City, Oklahoma. The aminomethylene, '[CH2NH] (Spatola nomenclature, 1983), and the hydroxyethylene, '[CH(OH)CH2], containing inhibitors were a gift from Tomi Sawyer. Pepstatin was purchased from Sigma. Acetyl-pepstatin was a gift from Kohei Oda, Kyoto Institute




23






24



of Technology, Japan. The native isozymes, pI 5 and pI 6, were a gift from Kevin Parris and David Davies at the Laboratory of Molecular Biology, National Institutes of Health. The porcine pepsin used for comparison to the rhizopuspepsins was from Sigma. The synthetic oligonucleotides were synthesized by the University of Florida, Interdisciplinary Center for Biotechnology Research (ICBR) DNA Synthesis Core Facility using an Applied Biosystems 394 DNA synthesizer. Peptide substrates were synthesized by the ICBR Protein Chemistry Core Facility using an Applied Biosystems 430A peptide synthesizer. The oligonucleotides were used directly for mutagenesis and sequencing reactions. All peptides were shown to be >95% pure by reverse phase HPLC and capillary electrophoresis. Stock solutions of the peptides and the inhibitor U85548E were quantified by amino acid analysis on a Beckman System 6300 high performance amino acid analyzer following acid hydrolysis. The N-terminal sequence analyses of the rhizopuspepsins were performed on Applied Biosystems 470A and 473A protein sequencers. The activated enzymes were analyzed by matrix-assisted laser desorption ionization time of flight (MALDI-TOF) mass spectroscopic analysis on a Vestec (Houston, TX) LaserTec Bench Top System. All other materials were of the highest commercial grade.






25


Methods


All routine DNA manipulation procedures were performed as outlined by Sambrook et al. (1989). Competent cells were prepared by the calcium chloride method. All plasmid and PCR products were isolated by using Magic Minipreps or Magic PCR preps kits from Promega. These kits use a proprietary anion exchange column to efficiently purify plasmid DNA. Clonina and Mutagenesis


Mutations in the rhizopuspepsinogen gene were made by using a modified version of the overlap extension method of site-directed mutagenesis by the polymerase chain reaction (PCR) (Ho et al., 1989) with the use of TAQ Polymerase (United States Biochemical) or Vent polymerase (New England Biolabs). This modification, using only one mutant primer, has been discussed in detail (Scarborough & Dunn, 1994) and is shown in Figure 2-1. This procedure uses four oligonucleotides, one of which contains nucleotide changes corresponding to the desired amino acid change. In the first round of amplification two reactions were performed using 100 pmol of each primer, 10 ng plasmid template and 5 units of polymerase. The first reaction generated the 5'-end of the gene (A) by using a sense primer (1) containing an engineered NdeI site (5'CAT ATG GCA GTT AAC GCT GCC CC3') and an antisense primer (2) containing the mutations for residues 30 or 77 (Asp30 Ile; 5'GA GGA ACC GGT ATC AAA GAT AAG GTT GAA






26







1 3
8_1 pET3A-RPGN Rhizopuspepsinogen

2 primers 1 and 2 or
primers 3 and 4

A

B





Denature and Anneal


C D
5 5 3 3'




primers 1 and 4 IM BamH I Nde I
E 0 Mutation




Figure 2-1. PCR mutagenesis procedure. A, 5'-end of rhizopuspepsinogen; B, 3'-end; C, hybrid capable of extension; D, hybrid not capable of extension; E, full length rhizpuspepsinogen gene containing engineered mutations and Nde I and BamH I restriction sites.






27



C3'; Asp77 -4Thr; 5'GAT ACC GCT AGC AGA AGA GCC AGE ACC ATA AG3'). The underlined bases indicate the engineered restriction sites or the differences from wild-type. The second reaction generated the 3'-end of the gene (B) by using a sense primer (3, Asp30 -Ile; 5'TTT GAT ACC GGT TCC TCC GAT TTA TG3'; Asp77 -4Thr; 5'TCT TCT GCT AGC GGT ATC TTG GC3') capable of annealing to the 3' end of the PCR product above

(A) and an antisense primer (4) containing an engineered BamHI site (5'GGA TTC TTA TTG AGC GAC AGG AGC G3'). The cycling conditions for the first round of PCR were as follows: (1) 3 cycles; 96C for 40 sec, 500C for 40 sec, 720C for 2 minutes, (2) 25 cycles; 940C for 40 sec, 500C for 40 sec, 720C for 2 minutes, (3) 720C for 7 minutes. These PCR products (A and B) were purified on Seaplaque GTG or NuSieve GTG low melting agarose gels (FMC Bioproducts). In order to generate the full length gene, small amounts of each band were mixed together and heated to 1000C for 5 minutes, and then placed on ice. This step is crucial for insuring that the fragments are completely dissociated so they can form hybrid templates in the next PCR reaction (C and D). After the addition of polymerase and more of the outer primers (1 and 4, 80 pmol each), the second round of PCR was performed as follows: (1) 25 cycles; 940C for 40 sec, 550C for 1 minute, 720C for 2 minutes, (2) 720C for 7 minutes. The double mutant was generated by repeating this procedure using the Asp30Ile mutant rhizopuspepsinogen gene as the starting template. The resulting products (E) were ligated






28



either directly into the TA cloning kit vector pCR (Invitrogen) or after restriction digest with NdeI and BamHI and gel purification into the pGEM vector (Promega) for DNA sequence analysis. In order to confirm the presence of the desired mutation and to ensure that no spurious mutations occurred during the polymerization process, the entire 1136 bp coding region was dideoxy-sequenced according to the Sequenase 2.0 Kit protocol using 5 gg of plasmid template (United States Biochemical) and deoxyadenosine-5'-[a35S]thiotriphosphate with the insertion of one extra step. Before the addition of the stop solution, more reaction buffer, dNTPs, and terminal deoxynucleotidyl transferase (TdT) were added in order to extend prematurely terminated products resulting from high GC content and secondary structure (Kho & Zarbl, 1992). The mutant genes were transferred to a modified version of the pET3a expression system vector. The pET3a vector was modified by removing a 375 bp fragment between the BamHI restriction site and ampicillin resistance gene in order to generate a vector which does not contain an EcoRV site. This was performed by digesting the vector with EcoRI and EcoRV and by subsequently treating with Klenow polymerase and blunt-end ligation. The resulting vector is shown in Figure 2-2. The use of this new vector allowed the efficient screening of recombinant clones because of the unique EcoRV restriction site within the rhizopuspepsinogen gene.






29
















BamH Rhizopuspepsinogen


EcoR

Nde
pET3aE-RPGN AMP
SD

5364 bp






ori


Figure 2-2. pET3aE expression vector containing rhizopuspepsinogen. AMP, ampicillin resistance gene; ori, origin or replication; SD, ShineDalgarno sequence.






30


Expression


The native and mutant enzymes were expressed and

purified from BL21(DE3) E. coli cells as reported with minor changes (Chen, Koelsch et al., 1991). A 1:50 dilution of an overnight culture grown in M9 media (10pg/mL thiamine, 0.5% casamino acids, 0.2% glucose) containing 50 mg/L ampicillin was made into LB media containing the same amount of ampicillin and grown to an OD600 of 0.5. At that time IPTG was added to give a final concentration of 0.5 mM. The cells were pelleted at 3,500 x g for 10 minutes and resuspended in

4.2 mls of 50 mM Tris-HCl pH 7.4, 150 mM NaCl and 1 mM MgC12 (buffer A) per gram of cells. Following the addition of 80 Kunitz units of DNase (Sigma) per ml of suspension, the cells were lysed by two passes through a French Press cell. The resulting slurry was carefully layered over a 27% sucrose cushion (density = 1.1) and centrifuged at 12,000 x g in order to isolate the inclusion bodies which sediment through the sucrose solution (Taylor et al., 1986). The inclusion bodies were washed by resuspension in buffer A containing 1% Triton X-100 and pelleted through sucrose a second time. The resulting pellet was stored at -200C until refolding. Refoldina


In order to regain enzymatic activity, the wild-type and mutant recombinant proteins were refolded by a modified procedure for the refolding of prochymosin involving






31



denaturation, reduction and dialysis (Suzuki et al., 1989). The purified inclusion bodies were dissolved in freshly deionized 8 M urea, 50 mM CAPS pH 10.5, 1 mM EDTA, 1 mM glycine, 500 mM NaCl and 300 mM P-mercaptoethanol to a final concentration of approximately 1 mg(wet)/ml. After stirring at room temperature for one hour, the solution was centrifuged at 24,000 x g for 30 minutes to remove undissolved material. The supernatant was dialyzed for one hour at room temperature against five times the original volume in SpectraPor 1 (MWCO 6-8 kDa) membranes and 50 mM Tris-HC1 pH 11.0 buffer. Following a buffer change and dialysis at room temperature for another hour, the dialysis buffer was changed to 50 mM Tris-HCl pH 7.5 and dialysis continued overnight at 40C. The next morning the buffer was changed to 50 mM MOPS pH 7.0 and dialyzed for at least 6 more hours at 40C. The resulting solution was centrifuged at 24,000 x g for 30 minutes to remove precipitates and concentrated using a Minitan Ultrafiltration system outfitted with low protein binding, PLTK, 10,000 MWCO membrane plates (Millipore) and an Amicon pressurized cell with YM10 membranes (10,000 MWCO).


Size-exclusion ChromatoaraDhv


The zymogen was further processed by centrifugation at 45,000 x g for thirty minutes before loading onto a 2.5 cm x 90 cm Sephacryl S300 gel filtration column equilibrated with






32



50 mM MOPS pH 7.0 containing 300 mM NaCI. The zymogen was eluted at a flow rate of 25 ml/hr and the fractions showing the highest purity were concentrated and buffer exchanged with 10 mM MOPS pH 7.0.


Activation and Ion-Exchanae Chromatoaraohy


In the activation of the native and Asp30Ile mutant

proteins (0.5 mg/ml) for kinetic analysis, citric acid pH 2.0 was added to give a final concentration 0.1 M. The resulting solution was held for fifteen minutes at room temperature. The Asp77Thr and the double mutant Asp30Ile/Asp77Thr zymogens were activated for twenty-four hours in 0.1 M sodium formate, 370C at pH 3.0 and 3.5, respectively. After filtering through a 0.2 pm Millipore microcentrifuge unit, each enzyme was directly injected onto a Pharmacia Mono S column equilibrated with 50 mM sodium formate pH 3.0. The enzyme was eluted by running a 25 minute gradient to 25% 50 mM sodium formate pH 3.0 containing 1 M NaCI at a flow rate of 1 ml/min. Enzyme aliquots were quickly frozen and stored at

-200C.


Structural Characterization


N-terminal seauencina. N-terminal sequence analysis of the activated rhizopuspepsins was performed to determine the extent of processing during self-activation. The proteins were electroblotted at 90 volts for 2 hours or 20 volts






33



overnight from 12% Tris-Tricine SDS-PAGE gels (Schagger & von Jagow, 1987) to PVDF Immobilon P transfer membranes (Millipore) in 10 mM MES pH 6.0 containing 20% methanol. The excised bands were analyzed by the Protein Chemistry Core Facility.

Mass sDectrometry. The activated enzymes were analyzed by matrix-assisted laser desorption ionization time of flight (MALDI-TOF) mass spectroscopic analysis on a Vestec (Houston, TX) LaserTec Bench Top System. One to ten pmol of each sample or standard was mixed 1:1 with fresh 0.05% TFA, 40% acetonitrile, saturated sinnapinic acid. One ~1 of this mixture was applied to a stainless-steel sample pin and allowed to air dry. The mass spectrum was obtained from the average of at least 50 laser shots (337 nm nitrogen laser, 3 ns pulse width). Time to mass/charge calibration was performed from a calibration curve using bovine carbonic anhydrase II (Sigma, 28,980 daltons) immediately prior to the rhizopuspepsin samples.

Isoelectric focusing. The isoelectric points of the

proteins were determined by running precast PhastGel IEF gels ranging from pH 3 to 9 on the Pharmacia PhastSystem. The samples, including standards, were diluted in sample buffer (0.01% pyronin Y, 10% glycerol, 62.5 mM TRIS pH 6.8) prior to loading. The gels were run according to the PhastSystem profile No. 100 and subsequently stained by the Coomassie blue profile No. 200.






34



Circular dichroism. Circular dichroism (CD) spectra were determined on one day at room temperature on a Jasco J-500C spectropolarimeter equipped with a IF-500 II computer interface with a 0.1 cm pathlengh cell (Hellma). The polarimeter was standardized with D(+)-camphorsulfonic acid (Chen & Yang, 1977). The samples were diluted into buffer to a final concentration around 0.5 mg/ml in 0.1 M sodium formate pH 3.0. Just prior to loading into the CD cell, the samples were filtered through a 0.2 pim microcentrifuge filter (Rainin) and quantitated by reading their absorbance at 280 nm. The ellipticity values were converted to the molar ellipticity, [01, using the conversion factor E9 = 12.6 (Fukumoto et al., 1967) and a molecular weight of 35 kDa. The data points were fit with the smoothing algorithm of the KaleidaGraph program (version 3.0.2 Synergy software, PCS Inc.).

Fluorescence spectroscopy. The guanidium hydrochloride denaturation curves of the rhizopuspepsins were determined using excitation and emission wavelengths of 280 and 350 nm, respectively, on an SLM Aminco 4800C spectrofluorometer. Ultrapure 8 M guanidinium hydrochloride in water was purchased from Pierce. The proteins were diluted in duplicate into denaturant concentrations ranging from 0 to

6 M with a final buffer concentration of 0.1 M sodium formate pH 3.0 and an enzyme concentration around 100 nM. The protein/denaturant solutions were equilibrated at 250C for at






35



least one hour prior to spectroscopic measurements in 1 cm cuvettes.
Denaturation curve analysis. The denaturation curves were analyzed assuming a two-state model where only the native and the denatured states are populated. The fluorescence values for the native and unfolded states,FN and Fu can be used to determine the equilibrium constant for unfolding, Ku_,, and the free energy of unfolding, AGu_F, at different denaturant concentrations by using equation 1.


KuF = (FN F) / (F Fu) = exp(-AGu-F / RT) (1)



F is the observed fluorescence, R is the gas constant (1.987 cal mol-1 K-1) and T is the absolute temperature. The free energy of unfolding of proteins has been shown to be linearly dependent on the denaturant concentration as expressed by equation 2 (Pace, 1986).



AG F = AG& m[denaturant] (2)



AGDF and AGHf are the free energy of unfolding at denaturant concentration, D, and in water, respectively. m is the slope of the transition and is thought to related to the difference in the degree of accessible surface area between the native and unfolded states (Schellman, 1978). Two methods have been used to calculate the transition point,






36


[GdnHC],5, AGf2O and m. The method of Pace (1986) employs equation 2 by plotting AGg_F within the transition region ( 1.5 kcal mol-1) versus [denaturant] and linearly extrapolating back, usually quite a long distance, to zero denaturant to obtain AG,f. Because small errors in m can lead to large errors in the calculation of AGHF and [GdnHCI], Fersht and coworkers have used a method, represented by equation 3, which uses all the observed fluorescence data, F, to directly determine [GdnHC1m typically within 0.02 M guanidinium hydrochloride (Jackson et al., 1993).


F = (aN + P[GdnHC1]) + (au + l [GdnHC1]) x

exp[m([GdnHC1] [GdnHCl1],) / RT]} / (3)

{1 + exp[m(GdnHC1] [GdnHC1]5,) / RT]}



This equation combines equations 1 and 2 and assumes that FN and Fu are linearly dependent on the denaturant concentration. aN and au are the intercepts and $N and f, are the slopes of the baselines at denaturant concentrations before and after the transition region. These parameters, as well as, [GdnHC1],, and m were allowed to be variables in the KaleidaGraph non-linear regression analysis program. The values of m and [GdnHCI],. were obtained for the rhizopuspepsins with their standard errors. AGR values






37



were calculated from equation 2 where at [GdnHC1]5, the transition point, AGH ==m[GdnHC1],4.




Kinetic Analysis


Kinetic assays usina chromoaenic substrates. Substrate hydrolysis, where cleavage occurs between Phe*Nph, Nle*Nph or Lys*Nph (Nph = p-nitrophenylalanine, Nle = norleucine and = site of cleavage), was monitored by the decrease in the average absorbance from 284-324 nm using a Hewlett Packard 8452A diode array spectrophotometer (Scarborough et al., 1993). The Km and Vmax values were determined from the initial rates of at least six different peptide substrate concentrations using Marquardt analysis (Marquardt, 1963) and the equation v = Vmax[S]/(Km + [S]). The observed rates in AU s-1 were converted to M s-1 by dividing by the total change in absorbance for complete hydrolysis of a known concentration of each substrate used. The amount of active enzyme was determined by fitting the curve generated by the competitive titration at one substrate concentration and 2% DMSO with the inhibitor Val-Ser-Gln-Asn-LeuP[CH(OH) CH2]ValIle-Val (U85548E; Sawyer et al., 1992) with the Henderson equation for tight binding inhibitors using the Enzfitter program (Henderson, 1972; Leatherbarrow, 1987). The standard deviations of the kcat and kcat/Km values were propagated using equations derived by standard procedures for non-






38


independent or correlated errors as outlined by Meyer (1975). In those cases where the Km values were >>1 mM, the kcat/Km values were determined by fitting the initial rates of at least six substrate concentrations ranging from 25-250 pM to the equation v = (kcat/Km) [E]0oS]o with the Enzfitter program and the assumption that [S] << Km. The kcat/Km values for the cleavage of the P1 lysine-substituted peptides by the Asp77Thr and the Asp30Ile/Asp77Thr mutants were calculated with the same equation on Enzfitter as above with the initial rates determined by capillary electrophoresis.

Kinetic assays usina competitive inhibitors. The

inhibition constant, Ki, was determined by monitoring the competitive inhibition of the hydrolysis of the peptide LysAla-Ala-Lys-Phe*Nph-Arg-Leu (Km = 20 gM) where cleavage occurs strictly between Phe and Nph. All reactions were performed at 370C in 0.1 M sodium formate buffer, pH 3.5 and a final concentration of 4% DMSO. The initial rates of six different substrate concentrations were measured following preincubation of the enzyme without inhibitor for five minutes. Additional curves were obtained, after preincubation with two or more inhibitor concentrations, from the initial rates of at least three different substrate concentrations. The Ki value was determined from the family of curves by the equation, v = Vmax[S]/[Km(l + [I]/Ki) + [S]]. If the Ki value determined by this method was one nanomolar or lower, a competitive titration was performed as described for the enzyme titration above.






39



Product analysis. The fidelity of the cleavage sites

was verified by HPLC and capillary electrophoresis (CE). All substrates were incubated with enzyme at 370C overnight. The cleavage products of the substrates based on the parent peptide Lys-Pro-Ala-Lys-Phe*Nph-Arg-Leu, discussed in Chapter 6, were analyzed by reverse phase HPLC using a Waters C-18 Radial-pack column with a gradient from 10 to 80% acetonitrile in water containing 0.1% TFA in 45 minutes. The peaks were collected and hydrolyzed by the addition of 6 N HC1. The composition was determined by amino acid analysis by the Protein Chemistry Core Facility.

The cleavage products of the peptides used for the analysis of mutants of rhizopuspepsin (Chapter 7) were determined on a BioRad BioFocus 3000 capillary electrophoresis system. The samples at 40C were pressure injected for 20 psi*sec onto a BioRad 24 cm X 25 pm cartridge maintained at 150C. This type of injection procedure ensures reproducible loading of the sample onto the capillary. The capillary was equilibrated with 0.5 M sodium phosphate pH

2.5, electrophoresed at a constant 8 kV in the +-*direction and monitored at 200 nm. All the peptides studied, if cleaved properly, will have the same C-terminal product, Nph-Arg-Ala. This product was purified on a Rainin HPLC system using a 4.6 mm x 25 cm Dynamax-300A C-18 column with a gradient from 0 to 10% B acetonitrile in water containing 0.1% TFA in 32 minutes at a flow rate of 1.1 ml/min. The composition of the fraction was confirmed by amino acid






40



analysis. This fragment was used as a retention time standard for the CE runs to validate the correct site of cleavage for each substrate and all forms of rhizopuspepsin at pH 3.5 and 5.0.

CaDillarv electrophoresis kinetic assay. The initial rates of cleavage of the peptide containing lysine in P1 by Asp77Thr and Asp30Ile/Asp77Thr were determined by incubating the enzyme (900 nM) with either 50, 100 or 150 JM substrate for a period of twenty-four hours at 370C. A 20 t1 aliquot was taken at 0, 1, 3, 5, 7, 12, 15, and 24 hours and mixed with 1.5 pl of U85548E to give a final concentration of inhibitor of 5 pM (five-fold molar excess) and stored at

-200C until the electropherograms were run. The initial linear slopes of the change in intact substrate peak area versus time were converted to M s-1 by dividing by the slope (peak area/[S]) of a standard curve generated from 30 to 1000 M of the substrate. The same injection and electropherogram run parameters were used in the product analysis.


Analysis of Transition State Effects


In transition state theory, the energy difference

between the free enzyme and substrate and the transition state,AG, is related to the binding energy released upon binding of the substrate, AGs, and the activation energy, AG*, of the chemical steps responsible for bringing the






41



enzyme-substrate complex from the ground state to the transition state (Fersht, 1985) :


AG = AGs + AG* = RTln(kT / h)- RTln(kt /K) (4)



With the assumption that the energies associated with the bond breaking and making steps, AG*, are not significantly affected upon mutation of the enzyme or changes in the assay pH, the discrimination of the wild type and mutant enzymes for different substrates can be evaluated by their relative binding to the transition state:



AAGT = -RTnk / K.(mutant, mutant 2 or pH 5.0) (5)
k / K. (wild- type, mutant 1 or pH 3.5)



Molecular Graphics


Molecular graphic representations of the x-ray crystal structures were generated using the Insight II (version 2.3) from Biosym Technologies, Inc. (San Diego) on a Silicon Graphics Indigo system at the University of Florida Center for Structural Biology. Root mean square (RMS) superposition of the Ca (alpha carbon) backbones of rhizopuspepsin (3APR, Suguna et al., 1987) and penicillopepsin (1APT, James et al., 1985) inhibitor complexes were performed by selecting the active site residues 27-37 and 210-220 for the SUPERIMPOSE command in the TRANSFORM menu.















CHAPTER 3
EXPRESSION, REFOLDING, PURIFICATION, AND ACTIVATION OF
RECOMBINANT RHIZOPUSPEPSINS


Introduction


Overexpression of proteins in heterologous systems has become an indispensable method in the generation of large quantities of wild-type and mutant proteins for structural and biochemical analysis. Many different systems have been used to obtain the protein of interest; for example, E. coli, yeast and SF9 insect cells. The decision of which system to use primarily depends on the yields required and whether or not glycosylation of the protein is desired. Expression in E. coli usually gives the highest yields but the protein is frequently deposited in an insoluble form known as inclusion bodies. These deposits are readily purified and are usually greater than 95% pure protein. In order to regain biological activity the protein must be denatured and refolded. Upon optimization of the refolding conditions, peculiar to each protein and its mutants, a sufficient quantity of protein for analysis may be obtained.

This chapter outlines the used of E. coli to produce

sufficient quantities of wild-type (WT-REC) and mutant forms of rhizopuspepsin for structural and kinetic analysis. Expression of the zymogen form of rhizopuspepsin,



42






43



rhizopuspepsinogen, resulted in the formation of inclusion bodies. The inclusion bodies were isolated, denatured, and refolded by dialysis. The inactive, zymogen form of rhizopuspepsin were converted to the active, mature form by lowering the pH of the solution. The resulting proteins were purified by ion exchange chromatography and analyzed by isoelectric focusing and N-terminal sequencing.


Results



Mutaaenesis


Mutants of rhizopuspepsin were generated by overlap extension PCR. All reactions yielded the desired size fragments. The entire coding region of each protein was DNA sequenced to determine the presence of the desired base changes and if other spurious mutations had occurred. In order to obtain clear sequencing ladders, an extra step was added to the Sequenase procedure. Prior to adding stop solution, the DNA was incubated with terminal deoxynucleotidyl transferase in order to extend premature stops due to high GC content and secondary structure. No additional mutations were seen. Expression and Refoldina


In order to obtain sufficient quantities of enzyme for

structural and kinetic analyses, the rhizopuspepsinogens were






44



overexpressed in E. coli using the pET expression system (Studier et al., 1990). Upon addition of IPTG to the bacterial cultures, the proteins were expressed at high levels in the form of inclusion bodies. Figure 3-1 shows an expression time course experiment for wild-type rhizopuspepsinogen. By the end of three hours, rhizopuspepsinogen was one of the predominant proteins. The inclusion bodies were purified from the cell lysate by centrifugation through a 27% sucrose cushion. The resulting pellet was washed with buffer containing triton X-100 to remove any remaining cellular debris. Typical yields ranged from 0.5 to 1 g (wet weight, 5 to 18 % of the total cell pellet) for a 4 L expression. The purified inclusion bodies were judged to be >95 % pure zymogen with a molecular weight of 43kDa by SDS-PAGE (Figure 3-2).

The inactive zymogen inclusion bodies, containing

approximately 100 mg of protein, were refolded from 8 M urea. Two refolding methods were tried in order to maximize yields for subsequent activation to the mature enzyme. Refolding by the rapid dilution procedure of Chen et al. (1991) resulted in primarily polymeric material that was difficult to completely activate into an active, monomeric state (data not shown). Even though there was some precipitation, the largest yields of monomeric zymogen capable of activation were obtained when the proteins were refolded by dialysis. Figure 3-3 shows the elution profile from gel filtration chromatography of the refolded rhizopuspepsinogen.

































Figure 3-1. SDS-PAGE analysis of the expression of wild-type rhizopuspepsinogen (RPGN) in E. coli upon the additon of IPTG from 0 to 3 hours. The RPGN migrates at 43 kDa in comparison to the molecular weight markers, M.





46









M 0 30' 1 2 3 kDa
946743- A W 30

2014-






47












1 2 3 4 5




kDa



43



















Figure 3-2. SDS-PAGE analysis of wild-type rhizopuspepsinogen at different stages of purification. Lane 1, E. coli whole cell lysate; lane 2, 27% sucrose pellet; lane 3, Triton X-100 wash 1 supernatant; lane 4, Triton X-100 wash 2 supernatant; lane 5, concentrated, refolded zymogen prior to loading onto a S-300 gel filtration column.

































Figure 3-3. Gel filtration elution profile of refolded Asp30Ile rhizopuspepsinogen. Peak 1, polymeric material at the void volume; peak 2, rhizopuspepsinogen; peak 3, low molecular weight contaminants.






49




















2




ABS 280 rn

3




11 16 hours






50



Peak 2 was shown to be rhizopuspepsinogen by SDS-PAGE analysis (Figure 3-4). Peaks 1, which elutes at the void volume, was shown, upon silver staining of the gel, to be primarily polymeric rhizopuspepsinogen. Peak 3 was shown not to contain rhizopuspepsinogen or active rhizopuspepsin but to consist of low molecular weight proteins by the same method. Yields at this stage of purification ranged from 15 to 35 mg for a 4 L preparation.


Activation and Purification


Activation of zymogens was accomplished by lowering the pH of the solution. Activation has been shown to occur by intermolecular and intramolecular interactions (Chen et al., 1991). The optimal conditions for activating the zymogens varied. Figure 3-5 shows the time course of activation of Asp30Ile at a protein concentration of 0.5 mg/ml. The protein was efficiently converted at both pHs and room temperature to a molecular weight of 35 kDa. Sequence analysis of the WT-REC and Asp30Ile proteins, activated at pH

2.0, confirmed the N-terminus of these proteins to be ThrSer-Thr-Gly-Gly-Ile-Val-Pro-Asp-. This sequence represents an extension of naturally occurring rhizopuspepsin by 9 amino acids. Figure 3-6 shows the results of an activation experiment of WT-REC at pH 4.0 from 1 to 6 hours in an attempt to remove this extension by intermolecular processing.


































Figure 3-4. SDS-PAGE analysis of the fractions from the purification of Asp30Ile by gel filtration in Figure 3-3. Lane 1, peak 1; lanes 2-10, peak 2; lane 11, peak 3





52













M 1 2 3 4 5 6 7 8 9 10 11 M kDa
94- l 67- g



30

20

14-S



































Figure 3-5. SDS-PAGE analysis of the time course of activation of wild-type rhizopuspepsinogen upon lowering the pH at room temperature. A, pH 2.0; B, pH 3.5.







54













M 0 15' 30' 1 2 3 kDa 67-


43.


30 67


43-


30-

































Figure 3-6. SDS-PAGE analysis of the activation of wild-type recombinant rhizopuspepsinogen at pH 4.0. Incubations were performed from 0 to 6 hours with protein concentrations of
0.1, 0.5, 1.0, and 1.5 mg/ml. C, crude isozyme pI 6.





56










0.1 0.5 1.0 1.5

kDaM C 1 2 3 6 1 2 3 6 1 2 3 6 1 2 36 M 43



30



20- .



14- S g






57



Tang and coworkers have shown that intermolecular activation of rhizopuspepsinogen and porcine pepsinogen occurs principally at pHs above 3.0 (Al-Janabi et al., 1972; Lin et al., 1989; Chen, et al., 1991). Protein concentrations were varied from 0.1 to 1.5 mg/ml. Activation to the intermediate form of rhizopuspepsin was not observed until 6 hours of incubation at 0.1 mg/ml. An increase in the protein concentration above this level resulted in degradation of the protein. The 9 amino acid extended form of the proteins were used for all kinetic and structural comparisons. A discussion on the potential effect of this extension is given in Chapter 4.

The activation of the D77T and the D30I/D77T proteins at pH 2, however, produced a mixed population of N-termini: the

9 amino acid extension and a 15 amino acid extension, AsnLys-His-Lys-Ile-Asn-Thr-Ser-Thr-Gly-Gly-Ile-Val-Pro-Asp-.

These populations can readily be seen with IEF gel analysis. Figure 3-7 shows the IEF gel of all the recombinant rhizopuspepsins following activation at pH 2 prior to ion exchange chromatography. The WT-REC and the Asp30Ile proteins have a pI of 5.8. The Asp77Thr and Asp30Ile/Asp77Thr mutants show bands at pHs 6.1 and 6.9. The yield ratios from the sequencing analysis suggest that the pH

6.9 band corresponds to the 15 amino acid extension activation intermediate. This is not suprising since this extension contains three additional positively charged residues making the protein more basic.






58










M 1 2 3 4

8.7

8.5
8.2
7.4


6.9


pI
6.6


5.9



5.2




3.5-





Figure 3-7. IEF analysis of the wild-type and mutant rhizopuspepsinogens activated at pH 2.0. Samples analyzed prior to ion-exchange chromatography. Lane 1, wild-type; lane 2, Asp30Ile; lane 3, Asp77Thr; lane 4, Asp30Ile/Asp77Thr.






59



Interestingly, only upon mutation of Asp77 to Thr is there a notable effect on the overall pI value of the 9 amino acid extended form.

Increased yields of the 9 amino acid extension were obtained for the D77T and the D30I/D77T proteins when the activation was performed at 370C and pH 3.0 and 3.5, respectively. The three additional positively charged residues proved to be fortuitous in clearly separating the two activation intermediates by ion-exchange (Figure 3-8). The recombinant rhizopuspepsins at this final step in the protocol were shown by IEF to be highly pure (Figure 3-9). All structural and kinetic comparisons were made using the Thr-Ser-Thr-form of the rhizopuspepsins. Yields at this final step in the purification ranged from 1 to 5 mg for a

4 L expression. A summary of the yields during the purification of the recombinant rhizopuspepsins is shown in Table 3-1.


Discussion


The largest losses of rhizopuspepsinogen occurred during the solubilization of the inclusion bodies and refolding. Precipitation usually occurs during the refolding protocol. The solubilization and refolding of rhizopuspepsinogen is complicated by the presence of two disulfide bonds in its tertiary structure. The inclusion bodies of prochymosin, which contains three disulfide bonds, have been shown to


































Figure 3-8. Elution profile of Asp77Thr from the Mono S column after activation at pH 3.0. The desired 9 amino acid extended form elutes near 12% B.







61





















100






ABS 280 nm % B








0
0 minutes 38






62










M 1 2 3 4

8.7
8.5
8.2

7.4


6.9


pI 6.65.9


5.2

4.6


3.5







Figure 3-9. IEF comparison of the activated, purified recombinant rhizopuspepsins. Lane 1, WT-REC, lane 2, Asp30Ile; lane 3, Asp77Thr; lane 4, Asp30Ile/Asp77Thr.


















Table 3-1. Representative yields during the purification of the recombinant rhizopuspepsins

WT-REC Asp30Ile Asp77Thr Asp30Ile/Asp79Thr


Inclusion bodies (wet) 1000 mg 429 mg 805 mg 813 mg 8M urea 97.2 mg 99.4 mg 33.3 mg 98.4 mg Refold supernatant 25.8 mg 37.3 mg 19.5 mg 15.3 mg Gel filtration 7.6 mg 10.0 mg 6.4 mg 6.3 mg Ion exchange (Mono S) 2.0 mg 3.8 mg 1.7 mg 1.2 mg






64



contain mainly intermolecularly cross-linked protein (Schoemaker et al., 1985). In order to reduce the cysteine residues in the protein before refolding, high levels (10 to 1000 fold molar excess based on cysteines) of reducing agent, BME or DTT, are required. Even if all the cysteines are reduced, losses still occur for several different reasons: pH of the refolding solution, speed at which the denaturant and reducing agents are removed and the protein concentration. These factors must be optimized in order to obtain biologically active protein. Each protein has its own characteristic conditions for refolding. Different conditions may also have to used to refold mutant proteins.

The refolding and activation protocols for the WT-REC and mutant forms of rhizopuspepsin have been optimized. These procedures produced sufficient quantities of active enzyme for structural and kinetic analysis. All enzymes used exhibit homogeneous N-termini and show similar electrophoretic properties after purification. Structural analysis and kinetic comparisons of the recombinant rhizopuspepsins to the naturally occurring isozymes are discussed in Chapter 4. The inhibitor binding and substrate specificity characteristics of WT-REC are discussed in Chapters 5 and 6. Chapter 7 presents the analysis of the WTREC and mutant rhizopuspepsins toward substrates that contain lysine in P1.















CHAPTER 4
KINETIC AND STRUCTURAL AUTHENTICY OF RECOMBINANT RHIZOPUSPEPSINS


Introduction


Molecular biology techniques have enabled the production of large quantities of purified proteins that would other wise be difficult to study due to their low abundance or to the impractical nature of the source. These techniques also aid in the dissection and understanding of biological phenomena through site-directed mutagenesis. The resulting proteins, however, must be shown to be analogous to the naturally occurring enzymes, with respect to structure and biological property being examined, if the results from recombinant proteins are to extrapolated to what occurs in the physiological environment.

The recombinant rhizopuspepsinogen gene used in this study was constructed from the two naturally occuring isozymes, pI 5 and pI 6, by the fusion of the pro region through residue 12 of the pI 5 isozyme gene to the pI 6 gene at residue 12. The expression of this chimeric gene results in an enzyme after activation that is identical to the pI 6 isozyme except at position 12 where Val is replaced by a Ile. Ile12 is located in the S3 subsite of the protein. All the other residue differences seen between the two naturally


65






66



occurring isozymes are on the surface of the protein far from the active site. In order to investigate a possible kinetic difference of the WT-REC enzyme from the pI 5 and 6 isozymes, a series of substrates with systematic substitutions in P3 were examined. This comparison was also performed to rule out the possible effects of the 9 residue N-terminal extension of WT-REC on catalysis. IEF gel analysis was also performed.

The overall tertiary structure of the native isozyme

pI 6 was compared to the WT-REC and mutant rhizopuspepsins by guanidinium hydrochloride denaturation experiments. The recombinant enzyme structures were also examined by circular dichroism. These studies were undertaken to investigate whether or not the mutations introduced into rhizopuspepsin caused large conformational changes in the enzymes that may compromise the interpretation of kinetic experiments.


Results



WT-REC and Native Isozvmes of RhizoDusDeDsin


IE. A comparison of the pI values of the naturally occurring isozymes of rhizopuspepsin to WT-REC was made using IEF gels and protein standards (Figure 4-1). The two native isozymes, pI 5 and pI 6, exhibit pI values of 5.1 and 6.2, respectively. The WT-REC enzyme has a pI of 5.7. This gel also illustrates the large difference in pI values seen upon






67











M 1 2 3 4

8.2
8.0
7.8
7.5

7.1


pI 6.5



6.0

5.1


4.7




Figure 4-1. IEF comparison of wild-type rhizopuspepsinogen, activated, purified WT-REC and the two naturally occurring isozymes. Lane 1, rhizopuspepsinogen, lane 2, isozyme pI 5; lane 3, isozyme pI 6; lane 4, WT-REC.






68



activation and removal of the pro region of the enzyme. The wild-type recombinant rhizopuspepsinogen has a pI value of

7.4.

Kinetic analysis. A comparison of the substrate specificity of the naturally occurring isozymes of rhizopuspepsin to WT-REC was made using peptides with substitutions in P3 (Table 4-1). Even though the enzymes possess different eletrophoretic properties and N-termini, the kinetic parameters determined for each substrate, within experimental error, are directly comparable. The three different forms of rhizopuspepsin are kinetically indistinguishable from each other and exhibit similar substrate specificity showing a preference for Arg and Leu substitutions in P3.


Structural Comparisons


Mass spectrometry. The recombinant forms and the isozyme pI 6 of rhizopuspepsin were analyzed by mass spectrometry. The mass for each protein was determined to be as follows: pI 6, 34,173; WT-REC, 34,627; D30I, 34,638; D77T, 34,914; and D30I/D77T, 34,856. The sizes of the recombinant proteins are consistent with the addition of 9 amino acids to the N-terminus of the native isozyme. The values for the WT-REC and D30I proteins are slightly lower than expected. This difference may be the result of C-terminal processing.






69














Table 4-1. Kinetic comparison between the naturally occurring isozymes and WT-REC rhizopuspepsin using the substrate Lys-Pro-P3-Lys-Phe*Nph-Arg-Leu

Enzyme P3 kcat Km kcat/Km
(s-l) (JM) (M-1s-1)
x 10-6


pI 5 12 2 20 2 0.63 0.14 pI 6 Asp 15 2 21 1 0.69 0.09
WT-REC 13 1 20 2 0.65 0.09

pI 5 16 3 9 1 1.71 0.38 pI 6 Arg 17 2 16 2 1.07 + 0.17
WT-REC 13 2 9 1 1.39 0.25

pI 5 9 2 7 1 1.31 0.30 pI 6 Leu 12 1 8 1 1.53 0.22
WT-REC 10 1 7 1 1.37 0.21

pI 5 13 2 21 2 0.62 0.12 pI 6 Ser 12 1 19 1 0.67 0.08
WT-REC 9 1 16 2 0.60 0.08

Nph = p-nitrophenylalanine; WT-REC = wild-type recombinant rhizopuspepsin.






70



Circular dichroism. Figure 4-2 shows the CD spectra of the recombinant rhizopuspepsins in the far-UV region (200-250 nm). Even though there are slight wavelength shifts in the spectra, the mutants as a whole are, within experimental error, structurally similar to each other and the wild-type enzyme. These shifts may be due to slight changes in the ahelix/0-sheet ratios. The differences seen between the proteins from 200 to 205 nm cannot be considered to be significant because of the high degree of signal fluctuation in this region on the instrument used.

Fluorescence sDectroscovY. The denaturation of the rhizopuspepsins was followed by the change in intrinsic fluorescence at 350 nm using an excitation wavelength of 280 nm (Figure 4-3). Upon the addition of sufficient guanidinium hydrochloride to cause unfolding, the fluorescence signal shifted to longer wavelengths with a 75% decrease in intensity. Figure 4-4 shows the normalized denaturation curves for the recombinant rhizopuspepsins and the isozyme pI 6. Analysis of the transition curves by the method of Jackson (1993) is presented in Table 4-2. The pI 6 isozyme and wild-type recombinant proteins exhibit unfolding parameters which are indistinguishable from each other. The decrease in [Gdn-HCO], of the two single mutants suggests a slight decrease in stability from the wild-type enzymes, WTREC and isozyme pI 6. Further loss of stability is seen in the double mutant.







71














1 106


s\ WT-REC
-.......-- D30I
5 10s -----D77T
- - D30I/D77T
0







-5 10




-1 106




-1.5 106
200 210 220 230 240 250
Wavelength (NM)



Figure 4-2. CD spectra of the recombinant wild-type and mutant forms of rhizopuspepsin.







72

















10




8
oFolded



6
0
r4


4
4 4 (d


2
Unfolded





300 320 340 360 380 400 NM


Figure 4-3. Fluorescence emission spectra for
folded (0 M GdnHC1) and unfolded (6 M GdnHCl)
wild-type recombinant rhizopuspepsin.







73



















1 ---- p16
------- WT-REC
-a--D30I
---V--- D77T
-- -D30I/D77T '44


0.5
0











0 1 2 3 4 5 6 [GdnHCl] (M)


Figure 4-4. Guanidinium hydrochloride induced unfolding of the naturally occurring isozyme pI 6 and the recombinant forms of rhizopuspepsin monitored by the change in intrinsic fluorescence at 350 nm.






74















Table 4-2. Guanidinium hydrochloride denaturation parameters of native and mutant forms of rhizopuspepsina

Enzyme [Gdn-HC11] 50 m AGH2
(M) (kcal mol-1 M-1) (kcal mol-1)


pI 6 3.54 0.02 4.8 0.5 17.0 1.8 WT-REC 3.50 0.02 4.6 0.5 16.0 1.7 Asp30Ile 3.31 0.02 3.3 0.2 10.9 0.8 Asp77Thr 3.33 0.02 4.5 0.4 15.1 1.4 Asp30Ile/Asp77Thr 3.00 0.01 3.7 0.2 11.0 0.7 aParameters derived from the denaturation curves presented in Figure 4-4. Each enzyme was studied as outlined in Chapter 2 from 0 to 6 M guanidinium hydrochloride in 0.1 M sodium formate pH 3.0. All denaturant concentrations were performed in duplicate. Values determined by the method of Jackson et al. (1993); curve fit of observed fluorescence, F, versus [denaturant]. The errors represented are the standard errors from the KaleidaGraph program.






75



Discussion



The zymogen forms of the native and mutant proteins were efficiently expressed in E.coli and refolded from inclusion bodies. Several lines of evidence exist to support the conclusion that the recombinant rhizopuspepsins are structurally and enzymatically similar to the native isozymes. Maturation of rhizopuspepsinogen requires catalytic activity. Activation has been to shown to occur upon lowering of the environmental pH by intermolecular and intramolecular processes similar to that of porcine pepsinogen (Chen et al., 1991). Kinetic comparisons between the two naturally occurring isozymes and the wild-type recombinant enzyme have shown that the 9 amino acid Nterminal extension and differences in pI values do not result in significant deviations in catalytic activity. These observations suggest that the slight differences seen in the pI values of the mutants will not adversely affect their kinetic analysis (Chapter 2, Figure 3-9). The degree of similarity seen in the circular dichroism and denaturation studies lends additional support to the conclusion that the recombinant wild-type and mutant proteins are correctly folded overall. Denaturation studies with the aspartic proteinase zymogen prochymosin have also shown that the recombinant protein is directly comparable to the native enzymes (Sugrue et al., 1990).






76



Mutagenesis at positions 30 and 77 did cause slight

destabilization of the proteins in response to guanidinium hydrochloride. Evidence supporting the idea that these differences may be the product of local side-chain reorganization comes from the x-ray crystallographic analyses of mutants of chymosin, lysozyme, trypsin and many other proteins (Strop et al., 1990; reviewed by Shortle, 1992). The Ca backbones of these structures exhibited little or no deviation from the wild-type structures. The side chains, however, usually did show some small positional movements.

Since denaturation studies have not been carried out on mutants of the aspartic proteinases, trends seen in the change of stability of the large library of mutants and their crystal structures, reviewed by Shortle (1992), can be used to rationalize the change in denaturation parameters seen for the mutants of rhizopuspepsin. Mutations that have been shown to cause significant changes in stability in comparison to the wild-type protein can be grouped into several categories: (1) insertion or deletion of an amino acid, (2) addition or deletion of disulfide crosslinks, (3) changes made near the ends of loops and a-helices, and (4) the alteration of hydrophobic packing in the core of the protein by the deletion of methylene equivalents. The sites targeted for mutagenesis in this study do not fall within any of the categories mentioned above. Position 30 and 77 were replaced with the corresponding residues of porcine pepsin. These residues are highly conserved among aspartic proteinases and






77



their substitution into the rhizopuspepsin structure is not expected to change the structure of the mutants significantly. Added support for structural similarity between the different rhizopuspepsins extends from the observation that the mutation of surface residues of a protein generally does not affect its overall fold.

Asp30 points into the active site cavity between the S3 and Si binding pockets. The base of the active site cleft is made of a large P-sheet composed of strands from both the N- and C-terminal domains of the protein. Studies by Katz and Kossiakoff (1990) have shown that P-sheets undergo less distortion than a-helices and loops upon mutation. This is thought to be the result of an increased ability of the P-sheet to dissipate strain energy through slight changes in 4-Y angles. These observations suggest that mutation at position 30 will not causes significant changes in stability or structure of the protein.

Asp77 is located in the f-hairpin turn of the flap which extends over the active site. The flap region is thought to be quite flexible. Upon inhibitor binding, particularly in the HIV proteinase (Swain et al., 1990), the flap undergoes movements to optimize hydrogen bonding and van der Waals contacts. The decrease in the crystallographic thermal B factors seen in the flap region upon inhibitor binding are thought to be representative of this effect (Suguna et al., 1987). Studies by Hurley et al. (1992) have shown that there is a strong correlation between the B factor of the residue






78


in the wild-type protein and the change seen in stability of the mutant. If the residue was originally not very mobile, low B value, the stability of a mutant at this position would decrease. This observation suggests that mutations at position 77 of the flexible flap region should not greatly affect overall stability of the protein.

All of the information discussed above suggests that the structures of the mutants of rhizopuspepsin have the same overall folding pattern with small positional changes in the side chains. The discussion, however, still does not explain why differences, particularly between the m and AG' values of the Asp30 Ile and Asp30Ile/Asp77Thr mutant proteins, are observed when comparisons are made to the WT-REC and pI 6 proteins. One possible reason may be the differences between the folded and denatured states of the proteins. Several studies have shown that the m value is related to the solvent accessible surface area of the denatured state (Schellman, 1978; Shortle & Meeker, 1986). Shortle and Meeker (1989) have shown that there is a correlation between the m value and the solvent accessibility of a mutant. Mutants that exhibit m values less than the wild-type protein show more compact structures and residual structural components when compared to the wild-type proteins by size-exclusion chromatography and circular dichroism. These observations suggest that rhizopuspepsin mutants with Ile at position 30 may be able to optimize hydrophobic interactions in the denatured state when compared to enzymes with Asp in this






79


position. These new interactions may make the Asp30 mutants less stable with the equilibrium shifting slightly in favor of the denatured state.

This chapter has presented kinetic and structural

evidence that the recombinant wild-type enzyme accurately represents the naturally occurring isozymes of rhizopuspepsin. Structural studies also suggest that the mutant enzymes exhibit similar structures to the wild-type proteins. The kinetic analysis of inhibitors (Chapter 5) and substrates (Chapters 6 and 7) with the WT-REC and mutant enzymes should not be significantly affected by structural deviations. Kinetic comparisons of the WT-REC to the mutant forms of rhizopuspepsin also support the idea that the structures have not been significantly altered. These results are discussed in Chapter 7.















CHAPTER 5
ANALYSIS OF THE SPECIFICITY OF RHIZOPUSPEPSIN THROUGH THE USE OF INHIBITORS CONTAINING P1 AND Pl' SUBSTITUTIONS AND SCISSILE BOND MIMETICS


INTRODUCTION


The inhibitors used in this study are based on the

cleavage sites in angiotensinogen, His-Pro-Phe-His-Leu*ValIle-His-Asn, and the p17/p24 HIV polyprotein junction ValSer-Gln-Asn-Tyr*Pro-Ile-Val. Many potent renin and HIV proteinase (HIV-PR) inhibitors have been generated using these cleavage sites and the natural product pepstatin as leads (Wiley & Rich, 1993). Studies varying side chain functionalities, in order to probe possible "secondary interactions" (Fruton, 1970; Medzihradszky et al., 1970) in the enzyme binding subsites, inhibitor length, and nonhydrolyzable peptide bond analogs have been undertaken to optimize potency and bioavailability (Rosenberg et al., 1990; Sawyer et al., 1991; Wiley and Rich, 1993). Other approaches to inhibitor design have involved molecular modeling based on the results from crystal structures of enzyme-inhibitor complexes (Kempf et al., 1990; Thompson et al., 1992). The results from crystal structures also aid in the analysis and rationalization of differences seen for inhibitor interactions and substrate specificities between family



80






81



members (Parris et al., 1992; Rao et al., 1993; Scarborough et al., 1993).

This chapter reports the structure-activity

relationships of wild-type recombinant rhizopuspepsin with inhibitors containing various scissile bond isosteres and PIP1' substitutions. The results of inhibitors with different lengths are also discussed. The primary goal of this study is to identify inhibitors for use as active site titrants.




Results and Discussion


Renin inhibitors have been synthesized using information from many studies where variations have been made to create potential interactions with the active site binding cleft (Wiley and Rich, 1993). These extensive surveys have not only led to many novel and potent inhibitors for renin, but also decreased the time to find potent HIV-PR inhibitors. Three major directions have been taken to increase potency, bioavailability and stability of inhibitors: (1) substitution of the scissile peptide bond of naturally occurring cleavage junctions with mimetics, (2) elaboration of the Pl' side chain, and (3) the determination of inhibitor length requirements. The crystal structures of many aspartic proteinase complexes with a variety of inhibitors aids in the rationalization of binding interactions in the active site which lead to a preference for one inhibitor over another.






82



The importance of the PI-P1' isostere interactions in the structure-activity relationships of the methyleneamino (CH2NH) and hydroxyethylene (CH(OH)CH2) containing inhibitors is shown in Tables 5-1 and 5-2. A ten-fold increase in potency (lower Ki) was observed when cyclohexylalanine (Cha, also referred to as cyclohexylmethyl) is replaced by Phe in the P1 position when comparing compounds 1 to 5, and 2 to 4. This decrease in binding may be due to the loss of van der Waals contacts and stabilizing aromatic/hydrophobic interactions with Tyr77, Phell4, and Leu122 in the hydrophobic S1 subsite. These observations are consistent with Cha side chain interactions seen in several complexes with rhizopuspepsin and endothiapepsin (Sali et al., 1989; Cooper et al., 1992; Parris et al., 1992). The incorporation of Cha into the P1 position was shown by Sali et al. (Sali et al., 1989) to influence the bound conformation of the P3 residue. The Cha substitution forces the P3 residue to reside in a less favorable rotamer conformation possibly resulting in lower inhibition. Other factors were stressed, however, which must also be considered. The ability of a compound to readily inhibit is determined by a combination of the intermolecular interactions between the inhibitor and the enzyme and the thermodynamic affects associated with establishment or loss of van der Waals interactions, hydrogen bonds and the change in solvent accessible surface of the free enzyme and substrate.












Table 5-1. Inhibition constants for XaaY[CH2NH]Yaa modified derivatives

CMPD P5-P4-P3-P2 P1-P1' P2' -P3' Ki (IM)


1 U79465E Ac-Pro-Phe-His ChaY[ X]Phe NH2 > 200 2 U79211E Ac-Pro-Phe-His Cha' [X]Val NH2 > 200 3 U79464E Ac-Pro-Phe-His ChaW[X]Cha NH2 > 200
4 U79339E Ac-Pro-Phe-His Phe'Y[X]Val NH2 8.7 1.7
5 U71909E Ac-Pro-Phe-His Phe'Y[X]Phe NH2 21 4 6 U80011E Ac-Pro-Phe-His PheV [X]pClPhe NH2 13 2
7 U80445E Ac-Pro-Phe-His Phe?[X]Tyr NH2 105 18
8 U81330E Ac-Pro-Phe-His Phe'P[X]pNO2Phe NH2 40 4
9 U70531E dHis-Pro-Phe-His PheY[X]Phe Val-Tyr 5.2 0.7 10 U91990E Ac-Pro-Hph-NMeHis PheF[X]Phe NH2 1.6 0.2

Hph = homophenylalanine; pNO2Phe = p-nitrophenylalanine; X = CH2NH pClPhe = p-chlorophenylalanine; Cha = cyclohexylalanine













Table 5-2. Inhibition constants for Leu~[CH(OH)CH2]Val and Statine modified derivatives

Inhibitor P5-P4-P3-P2 PI-P1' P2 -P3' Ki (nM)



11 U85548E Val-Ser-Gln-Asn Leu [X]Val Ile-Val < 0.1 12 U92522E Ac-Ser-Gln-Asn Leu?[X]Val Ile-NH2 1.4 0.3 13 U92517E Ac-Gln-Asn Leu?[X]Val Ile-NH2 1.4 0.3 14 U92516E Ac-Asn Leu'[X]Val Ile-NH2 510 92 15 U84728E Ac Leu?[X]Val Ile-NH2 20 200 3 700 16 U85964E Ac-Val-Val LeuYt[X)Val Ile-Amp < 0.1 17 Pepstatin Iva-Val-Val Sta Ala-Sta 0.7 0.2 18 Ac-Pepstatin Ac-Val-Val Sta Ala-Sta 10 2 19 U77647E Ac-Pro-Phe-His Leu[X]Val Ile-NH2 54 7


X = CH(OH)CH2; Amp = aminomethylpyridine; Sta = statine = LeuY [CH(OH)]Gly Iva = isovaleryl; Ac = acetyl






85



These additional factors may be able to explain the increased potency seen for compounds containing Phe and Leu in the P1 position for rhizopuspepsin. From these observations it was clear that rhizopuspepsin has different requirements in the S1 pocket since PI-Cha-containing analogs were found early on to confer increased potency for renin (Boger et al., 1985; Sawyer et al., 1990; Wiley and Rich, 1993).

Systematic substitutions of the P1' position (compounds

4 to 8), where P1 was Phe, resulted in a large range of potency. The Val, Phe and p-CiPhe substitutions exhibited similar inhibitory capabilities while p-NO2Phe and Tyr in Pl' showed progressively higher Ki values. The later two substitutions may be unfavorable due to the interruption of hydrophobic/aromatic interactions with Ile216 and 298, Trp194 Trp294, and Phe296 of the S1' binding pocket or the introduction of a slightly altered hydrogen bonding arrangements. These observations point to further differences between rhizopuspepsin and renin. Substitution of the P1' phenyl ring of PheY[CH(OH)CH2]Phe containing compounds with halogens reduced potency for renin, while the nitro substitution had little effect on binding (Young et al., 1992), but did increase bioavailability (Thompson et al., 1992). Possible reasons for the small differences seen for these substitutions against renin extend from the much larger S1' pocket and the use of the hydroxyethylene isostere which resulted in subnanomolar Ki values (Szelke et al., 1980). The hydroxyethylene isostere shows significantly




Full Text
76
Mutagenesis at positions 30 and 77 did cause slight
destabilization of the proteins in response to guanidinium
hydrochloride. Evidence supporting the idea that these
differences may be the product of local side-chain
reorganization comes from the x-ray crystallographic analyses
of mutants of chymosin, lysozyme, trypsin and many other
proteins (Strop et al.f 1990; reviewed by Shortle, 1992).
The Ca backbones of these structures exhibited little or no
deviation from the wild-type structures. The side chains,
however, usually did show some small positional movements.
Since denaturation studies have not been carried out on
mutants of the aspartic proteinases, trends seen in the
change of stability of the large library of mutants and their
crystal structures, reviewed by Shortle (1992), can be used
to rationalize the change in denaturation parameters seen for
the mutants of rhizopuspepsin. Mutations that have been
shown to cause significant changes in stability in comparison
to the wild-type protein can be grouped into several
categories: (1) insertion or deletion of an amino acid, (2)
addition or deletion of disulfide crosslinks, (3) changes
made near the ends of loops and a-helices, and (4) the
alteration of hydrophobic packing in the core of the protein
by the deletion of methylene equivalents. The sites targeted
for mutagenesis in this study do not fall within any of the
categories mentioned above. Position 30 and 77 were replaced
with the corresponding residues of porcine pepsin. These
residues are highly conserved among aspartic proteinases and


92
Hofmann et al., 1988; Pohl & Dunn, 1988). Recent studies
with porcine pepsin, cathepsin E (Rao-Naik, unpublished data)
and cathepsin D (Scarborough et al., 1993) have confirmed
this information and extended the search for other enzyme
interactions with the P5, P4, P2' and P3' positions of
substrates.
This chapter describes the results of the kinetic
analysis of wild-type recombinant rhizopuspepsin.
Systematically substituted octapeptide substrates were
analyzed with rhizopuspepsin in an effort to identify
specificity differences from the mammalian enzymes.
Substitutions were made in the P5-P1 and the P2,-P3' positions
of the substrate Lys-Pro-Ala-Lys-Phe*Nph-Arg-Leu to explore
potential hydrophobic, electrostatic, and hydrogen bonding
interactions. An additional series of peptides, containing
lie in the P3 position of the parent above and substitutions
in P2, were studied to complement the analysis of S2
interactions. This information may prove to be useful in the
design of targeted anti-fungal agents for the treatment of
Candida infections.
Results
L.Y.£7PrD.-Ala-Lys-Phe*Nph-Arg-Leu Based Substrates
Table 6-1 lists the kinetic parameters at pH 3.5 for the
library of substituted peptides based on the parent substrate


122
loss of these interaction yields a free energy change of 2.3
kcal mol-1. When the Pi-lysine substrate (3) is bound to
Asp77Thr or the double mutant, two additional hydrogen bonds
are lost to the lysine side-chain. These combined losses
appear to be worth up to 7 kcal mol-1. If these effects are
additive, the two hydrogen bonds to the lysine side chain are
worth approximately 4.7 kcal mol-1. The free energy changes
determined by these experiments are in the same range,
between 3 and 6 kcal mol-1, as those seen for other mutants of
charged hydrogen bond donors and acceptors (Fersht et al.,
1985; Fersht, 1988).
This study has shown that Asp30 may be partially
responsible for the general broad specificity of
rhizopuspepsin when compared to the mammalian enzymes. Asp77
is the critical residue required for enabling rhizopuspepsin
to cleave substrates containing lysine in Pi. The presence
of an aspartic acid residue at position 77 in the flap may
allow the formation of an extensive hydrogen bonding network
not only to the Pi lysine side chain but between the enzyme
itself. These interactions somehow enable the anomalous
cleavage of Pi-lysine containing peptides by lowering the
energy barrier to the transition state in the fungal aspartic
proteinases.


86
greater inhibitory capacity, approximately 100-fold, to the
methyleneamino derivatives. This difference can be
illustrated by the comparison of compounds 4 and 19. This
increased affinity may mask subtle interactions only
discernible with the use of poorer inhibitors.
X-ray crystallographic analyses of inhibitor complexes,
kinetic analysis of substrates of different lengths, and
studies attempting to potentiate bioavailability by
decreasing the inhibitor size, have led to the understanding
that the extended active site of aspartic proteinases can
bind from 7-8 residues. Ligands bind in an extended P-strand
conformation with the amino acid side chains alternating from
side to side with the a-carbon backbone making a series of
highly conserved hydrogen bonds to the enzyme (Davies, 1990;
Suguna et al., 1992). Because of this arrangement, for
example, the S3/S1 and the S2/Si' subsites are adjacent to
each other and the residues at the corresponding positions in
ligand are also able to interact and influence each other.
The sum of these interactions generate the forces needed for
substrate binding and catalysis. Table 5-2 shows a deletion
series of U85548E (11). The P3, P2 and the P2' positions
contribute from 7 to 15 kcal/mole to binding. Removal of the
P5, P4 and the P3' residues of resulted in a 10-fold decrease
in potency. When the P3 and P2 residues were deleted,
however, a substantial increase in the Ki values of 5,000 and
200,000-fold were seen, respectively. These observations
mirror those seen in the analysis of HIV-PR with same set of


14
relationships proposed in these studies can be tested by
site-directed mutagenesis and the heterologous expression of
these proteins.
Rhizopuspepsin
Rhizopuspepsin, a fungal aspartic proteinase from
Rhizopus chinensis, has been the subject of many substrate,
inhibitor and crystallographic studies in an effort to better
understand factors contributing to catalysis and specificity
differences between members of this enzyme family. Initial
interest in the fungal enzymes arose from the need to find a
suitable replacement for calf chymosin in the cheese-making
process (Davies, 1990). This focus changed with the
determination of the native structure (Suguna et al., 1987)
and many different inhibitor complexes of rhizopuspepsin
(Suguna, et al., 1987; Parris et al., 1992; Suguna et al.,
1992) and other fungal enzymes (James et al., 1982; Cooper,
et al., 1987; Lunney, et al., 1993). The insights gained
from these complexes have led to proposals for the reaction
mechanism (James et al., 1977, 1992; Suguna, et al., 1987
Fraser et al., 1992;) and the rationalization of subsite
preferences of the aspartic proteinases (Rao et al., 1993;
Scarborough et al., 1993). The goal has been to use this
information to design specific or targeted therapeutics for
renin and the HIV proteinase. Further clinical interest in
rhizopuspepsin comes from its similarity to the secreted


102
irreversible step in the catalytic mechanism. By comparing
the specificity constants of the wild-type and mutant enzymes
for the same substrate, the change in the free energy barrier
to reach the transition state complex (ES*) from the free
enzyme and substrate (E+S) can be calculated. This
difference reflects the stabilization or destabilization of
the transition state.
From the examination of the high resolution crystal
structures of the inhibitor complexes of penicillopepsin
(James & Sielecki, 1985), endothiapepsin (Pearl & Blundell,
1984), rhizopuspepsin (Suguna et al., 1987) and the HIV
proteinase (Swain et al., 1990) and the analysis of kinetic
isotope effects, particularly in the HIV proteinase system
(Hyland et al., 1991a,b; Rodriquez et al., 1993), the
transition state in the general acid-base catalytic mechanism
of the aspartic proteinases is thought to resemble a
tetrahedral intermediate. In the cleavage of substrates
containing Tyr-Pro junctions by the HIV proteinase, a water
molecule located between the two catalytic aspartic acids is
considered to be made more nucleophilic by the transfer of a
proton to the negative Asp25', Asp215 in porcine pepsin,
after the formation of an intermediate where the nitrogen of
the scissile bond is partially positively charged. This
intermediate is thought to be a result of the establishment
of a strong low-barrier hydrogen bond (Cleland, 1992) from
the protonated Asp25 (the equivalent function is performed by
Asp32 in porcine pepsin) to the scissile bond carbonyl group.


109
Aso77Thr mutant. The mutation of Asp77 to Thr resulted
in an enzyme which displays reduced kcat values toward this
series of substrates. The most significant result of this
mutation appears to be the 100-fold increase in the Km
parameter and the parallel decrease in kcat/Km. This mutant
enzyme is no longer competent to cleave the Pi-lysine
peptide. This result is illustrated by the 105-fold decreases
in kcat/Km from 3 60,000 to 5 M-1s-1 at pH 3.5 and 1,660,000 to
13.7 M-1s-1 at pH 5.0.
Aso30lle/Aso77Thr mutant. The double mutant enzyme is
able to cleave all the substrates but with decreased
specificity (kcat/Km) This enzyme is similar to the Asp77Thr
mutant in that it is characterized by the inability to
readily cleave the Pi-lysine substrate.
Kinetic Analysis of Porcine Pepsin
Table 7-2 lists the kinetic parameters determined for
peptides 1-5 with porcine pepsin. Peptides 2, 4 and 5 are
readily cleaved with kcat values similar to that of wild-type
rhizopuspepsin. The kcat/Km values, however, are 2 to 5-fold
lower. Peptide 1 is poorly cleaved at pH 3.5, consistent
with previous observations (Rao-Naik & Dunn, unpublished
results; Pohl & Dunn, 1988). Improvement is seen upon
raising the pH, although not to levels seen for the other
peptides. Pepsin behaves similarly to rhizopuspepsin with
respect to peptide 2 in that the specificity (kcat/Km)


22
In contrast to the mammalian and fungal enzymes that contain
either serine or threonine at position 77, the Asp77 residue
was shown to be able to make an additional contact through
its side chain by hydrogen bonding to the e-amino nitrogen of
the lysine residue. Hydrogen bonds are also seen between the
lysine residue and Ser79.
This study has focused on the contributions of Asp30 and
Asp77 to fungal specificity through the use of site-directed
mutagenesis. These residues were changed in the
rhizopuspepsinogen gene to those present in porcine pepsin,
Ile30 and Thr77. The proteins were overexpressed in E. coli,
refolded from inclusion bodies, activated and purified for
structural and kinetic comparisons. A series of
systematically substituted substrates with lysine in either
Pi, P2 or P3 was assayed and analyzed by double mutant cycles
(Carter et al., 1984; Wells, 1990) in order to ascertain and
confirm the predominant interactions enabling substrate
catalysis.


19
Rhizopuspepsin, as well as the majority of all other fungal
enzymes, has the ability to cleave substrates (Hofmann et
al., 1984; Balbaa et al., 1993) and to bind inhibitors with
lysine in the Pi position (Salituro et al., 1987).
Structural and sequence comparisons suggest two possible
residues in the active site that may account for this unique
capacity. Table 1-3 shows an alignment of sequences from
several aspartic proteinases in the active site and flap
regions. Besides the two catalytic aspartic acids, Asp32 and
Asp215, rhizopuspepsin has additional aspartic acid residues
at positions 30, 37, and 77. Asp30 is situated at the
boundary between the S3 and Si subsites and has the potential
to interact with positively charged residues in the P3 and
the Pi positions of the substrate (Figure 1-3). Asp37 is
located in the S2' subsite and probably does not directly
affect primary specificity. Asp77, which is also present in
the Candida enzymes, is positioned at the end of the flap and
points down into the active site cavity making potential
interactions with the P3, P2 and Pi positions of the
substrate.
A precedent exists for the importance of Asp77 in the
primary specificity of fungal enzymes from a complex of
penicillopepsin with a pepstatin derivative containing lysine
in Pi (James et al., 1984). In this structure the Asp77 side
chain and the enzyme backbone NH hydrogen bond in a highly
conserved manner to the P2 NH and the P2 carbonyl of the
inhibitor backbone, respectively.


Table 6-2. Kinetic analysis of WT-REC rhizopuspepsin with substrates of the form
Lys-Pro-Ile-P2~Phe*Nph-Arg-Leu at pH 3.5
P5
P4
P3
P2
Pl*Pl*
P2'
P3 1
^cat
(sec-1)
Km
(JXM)
kCat/Km
(M-1S-1)
X 10-6
Lys
Pro
He
Glu
Phe*Nph
Arg
Leu
22
2
6

2
3.84
1.16
VO
Ala
8
1
5

1
1.56
0.42
Ser
8
1
11

2
0.76
0.16
Arg
3
0.4
9

1
0.36
0.06
Asp
2
0.3
6

1
0.33
0.06
Leu
1
0.1
5

1
0.21
0.05
His
0.03
0.003
5

1
0.005
0.001


105
all the substrates studied upon increase of the assay pH:
kcat values were independent of pH while two to five-fold
increases in kcat/Km were observed, mainly as the result of a
corresponding decrease in Km. Importantly, all the enzymes
have similar kcat values at pH 5.0 for peptide 4. This
observation confirms the results of the structural analyses
in Chapter 4 that supported the overall similarity of the
protein folds and that the enzymes can be compared with
confidence.
Wild-tvne recombinant. The wild-type enzyme exhibits
the characteristic broad specificity seen for similar
peptides discussed in Chapter 6. All substrates, with (1-3)
or without (4,5) the lysine residue placed in several
alternative positions, were readily cleaved. Even though
there is still a preference for the hydrophobic junction
(1 and 4), the enzyme is able to cleave the Pi-lysine peptide
(3) with equivalent efficiency (kcat) .
Asp30lle mutant. This mutant cleaves the substrates,
including the Pi-lysine derivative, with nearly identical kcat
values to the wild-type enzyme. A 2-3 fold increase in Km
is seen with peptides 1-4. This increase results in kcat/Km
values that are in the range seen for porcine pepsin with the
same substrates (Table 7-2). Larger changes are seen in all
the kinetic parameters for peptide 5. These differences are
probably due to the loss of hydrophobic interactions when the
aromatic phenylalanine is replaced by the aliphatic
norleucine.


7
The studies with tyrosyl-tRNA synthetase and trypsin are
only two examples of the extensive literature establishing
the importance of hydrogen bonding and electrostatic
interactions in the creation of specificity in biological
reactions. Mutational and kinetic analysis have also been
used to dissect the structural components that generate the
unique specificities of cysteine and metallo proteinases.
Studies on the aspartic proteinases have the potential to
give further insight into the role of electrostatic
interactions in specificity. In particular, rhizopuspepsin,
the aspartic proteinase from the fungus Rhizopus chinensis,
contains several aspartic acid residues in the active site
that may influence the selection of substrates for
hydrolysis.
Asoartic Proteinases
General Characteristics
All members of the aspartic proteinase family show
marked inhibition by pepstatin (Davies, 1990) and a high
degree of amino acid sequence and three-dimensional structure
homology, especially in the region of the active site, as
shown by the studies of Blundell et al. (Pearl & Blundell,
1984; Blundell et al., 1987; Sali et al., 1989). The bilobal
structure of an aspartic proteinase is created by the nearly
symmetrical N-terminal and C-terminal domains of the protein.


127
The changes in energetics could be compared to those seen for
the substrate series. Analysis of the triple mutant
Asp30lle/Asp37Asn/Asp77Thr of rhizopuspepsin may also show
the importance of Asp37 in the generation of a specific
electrostatic interaction with substrates containing an
arginine in the P2' position and the general overall
preference of rhizopuspepsin for positively charged
substrates.
The results of this study have implications for the
design of anti-fungal agents. The Candida aspartic
proteinases also contain an aspartic acid at position 77.
Kinetic studies have shown that these enzymes have substrate
preferences similar to rhizopuspepsin (Fusek et al., 1994).
Preliminary evidence that the Candida enzymes possess the
Pl-lysine specificity comes from their ability to cleave
collagen and keratin (Lin et al., 1993). Future mutational
and kinetic analysis of the Candida enzymes with Pi-lysine
containing substrates and the solution of crystal structures
already in progress (Cutfield et al., 1993) are needed to
confirm the importance of Asp77 and other potential
interactions in pathogenicity that will aid in the design of
targeted therapeutics. These experiments may include: (1)
mutation of the Candida enzymes at position 77 with the
subsequent analysis of tissue invasion and substrate
specificity, (2) analysis of the ability of the wild-type and
mutant forms of rhizopuspepsin to cleave keratin and
collagen.


2
in entropy in the organizational process of protein folding,
the release of water molecules and the establishment of an
intricate network of intramolecular hydrogen bonds
compensates for the entropy loss making the entire process
slightly favorable overall. The observation that almost all
possible hydrogen bonds are established upon the folding of a
protein points to the difficulty, in separating and
quantifying the forces involved. Despite this difficulty
trends are still observable. A survey of many x-ray crystal
structures of ligands bound to proteins has shown that there
is a strong correlation between the binding strength of
ligands and the degree of buried surface area (Horton &
Lewis, 1992; Young et al., 1994). This correlation is
dependent on the geometric or steric requirements for the
optimization of packing and hydrophobic interactions, but is
also influenced by the chemical complementarity between the
two species.
Chemical complementarity and biological specificity in
protein-ligand interactions are primarily governed by
hydrogen bonds and electrostatic interactions (Fersht et al.,
1985). A hydrogen bond occurs when two electronegative atoms
share a hydrogen atom. The optimal arrangement for this
interaction is linear with a distance between the
electronegative atoms ranging from 2.85 to 3 angstroms. The
study of the contribution of hydrogen bonds to specificity
has been facilitated by the use of site-directed mutagenesis.
The influence of hydrogen bonding in recognition and


103
Following the attack of the water molecule to form the
tetrahedral amide hydrate intermediate, the transition state
is proposed to collapse to form products upon protonation of
the proline nitrogen by Asp25' and the deprotonation of the
gem-diol hydroxyl group.
Kinetic analysis has shown that the rate of exchange of
180 from H2180 into the starting substrate exceeded the rate
of label incorporation for the reverse peptidolytic reaction.
This evidence suggests that the rate determining or first
irreversible step in the reaction mechanism is the breakdown
of the tetrahedral intermediate to form products.
Experiments analyzing solvent deuterium isotope effects
suggest that this breakdown or collapse of the transition
state occurs by the simultaneous transfer of two protons as
described above (Hyland et al., 1991a,b; Rodriquez et al.,
1993). Site-directed mutagenesis and double mutant cycle
analysis can give insight into how particular amino acid
residues of the aspartic proteinases contribute to the
stabilization of the transition state that results in the
overall preference for one substrate over another.
The fungal aspartic proteinases exhibit a unique
specificity toward substrates in comparison to the mammalian
enzymes. The fungal enzymes are able to cleave substrates
containing lysine in the Pi position. Sequence and
structural comparisons have suggested Asp30 and Asp77 in
establishing this characteristic ability.


ACKNOWLEDGMENTS
I have always been encouraged to be my best and to be
true to myself. This support has extended from family and
friends. In particular, I want to thank my mother and
father, Ruth and Mike, for their love, generosity and
willingness to do whatever it took to make sure that my
brothers, Jason and Pat, and I had everything we could have
ever needed or wanted. I also want to thank Bill, Dee,
Brandon and Kara for their encouragement and love. My
grandparents have also been a continual source of wisdom,
support and love. Encouragement from Andy and Cheryl has
also been very reassuring.
I am indebted to Dr. Ben Dunn for providing a fantastic
environment for learning and scientific opportunity. I am
particularly thankful for his insights and willingness to
allow me to pursue my research interests, within reason.
I would like to thank my committee members for their
suggestions and encouragement during my graduate work; Drs.
Daniel Purich, Charles Allen, Sheldon Schuster, and Nigel
Richards.
The support of the Dunn laboratory has made my stay at
the University of Florida one of the best times of my life.
In particular, I want to thank Chetana, Paula, Wieslaw, Bill,
iii


82
The importance of the Pi-Pi' isostere interactions in
the structure-activity relationships of the methyleneamino
(CH2NH) and hydroxyethylene (CH(0H)CH2) containing inhibitors
is shown in Tables 5-1 and 5-2. A ten-fold increase in
potency (lower Ki) was observed when cyclohexylalanine (Cha,
also referred to as cyclohexylmethyl) is replaced by Phe in
the Pi position when comparing compounds 1 to 5, and 2 to 4.
This decrease in binding may be due to tfie loss of van der
Waals contacts and stabilizing aromatic/hydrophobic
interactions with Tyr77, Phell4, and Leul22 in the
hydrophobic Si subsite. These observations are consistent
with Cha side chain interactions seen in several complexes
with rhizopuspepsin and endothiapepsin (Sali et al., 1989;
Cooper et al., 1992; Parris et al., 1992). The incorporation
of Cha into the Pi position was shown by Sali et al. (Sali et
al., 1989) to influence the bound conformation of the P3
residue. The Cha substitution forces the P3 residue to
reside in a less favorable rotamer conformation possibly
resulting in lower inhibition. Other factors were stressed,
however, which must also be considered. The ability of a
compound to readily inhibit is determined by a combination of
the intermolecular interactions between the inhibitor and the
enzyme and the thermodynamic affects associated with
establishment or loss of van der Waals interactions, hydrogen
bonds and the change in solvent accessible surface of the
free enzyme and substrate.


136
Newman, M., Frazao, C., Shearer, A., Tickle, I. J. &
Blundell, T. L. (1990) Biochemistry 29, 9863-9871.
Studier, F. W., Rosenberg, A. H., Dunn, J. J. & Dubendorf, J.
W. (1990) Meth. Enzymol. 185, 60-89.
Subramanian, E., Swan, I., Liu, M., Davies, D., Jenkins, J.,
Tickle, I. & Blundell, T. (1977) Proc. Natl. Acad. Sci.
USA 77, 556-559.
Sugrue, R., Marston, F., Lowe, P. & Freedman, R. (1990)
Biochein. J. 271, 541-547.
Suguna, K., Bott, R. R., Padlan, E. A., Subramanian, E.,
Sheriff, S., Cohen, G. H. & Davies, D. R. (1987a) J. Mol.
Biol. 196, 877-900.
Suguna, K., Padlan, E. A., Smith, C., Carlson, W. & Davies,
D. R. (1987b) Proc. Natl. Acad. Sci. U.S.A. 84, 7009-
7013.
Suguna, K., Padlan, E. A., Bott, R., Boger, J. & Davies, D.
R. (1992) Proteins: Struct. Func. Genet. 13, 195-205.
Suzuki, J., Sasaki, K., Sasao, Y., Hamu, A., Kawasaki, H.,
Nishiyama, M., Horinouchi, S. & Beppu, T. (1989) Protein
Eng. 2, 563-569.
Swain, A. L., Gustchina, A. & Wlodawer, A. (1991) in
Structure and Function of the Aspartic Proteinases:
Genetics, Structures and Mechanisms (Dunn, B. M., Ed.) pp
433-441, Plenum Press, New York.
Swain, A. L., Miller, M. M., Green, J., Rich, D. H.,
Schneider, J., Kent, S. B. & Wlodawer, A. (1990) Proc.
Natl. Acad. Sci. U.S.A. 87, 8805-8809.
Szelke, M., Jones, D., Atrash, B., Hallet, A. & Leckie, B.
(1980) in Peptides: Structure and Function, Proceedings
of the Eighth American Peptide Symposium (Hruby, V. J. &
Rich, D. H., Eds.) pp 579-582, Pierce Chemical,
Rockford,IL.
Takahashi, K. (1987) J. Biol. Chem. 262, 1468-1478.
Takahashi, K. (1988) J. Biochem. 103, 162-167.
Taylor, G., Hoare, M., Gray, D. & Marston, F. (1986)
Biotechnology 4, 553-557.
Thompson, W. J., Fitzgerald, P. M. D., Holloway, M. K.,
Emini, E. A., Darke, P. L., McKeever, B. M., Schleif, W.
A., Quintero, J. C., Zugay, J. A., Tucker, T. J.,


Figure 3-8. Elution profile of Asp77Thr from the Mono S
column after activation at pH 3.0. The desired 9 amino acid
extended form elutes near 12% B.


87
compounds (Sawyer et al., 1992) and in studies by many
investigators with truncation series of inhibitors (Rosenberg
et al., 1990; Wiley & Rich, 1993) and substrates (Fruton,
1976; Hofmann et al., 1988; Balbaa et al., 1993). The
results from substrate analyses have shown that the addition
of residues at the P3, P2 and P2' positions significantly
increases catalysis, kcat, without a corresponding change in
Km (Hofmann et al., 1988; Balbaa et al., 1993). It is thought
that the extra interactions and hydrogen bonds formed to
these residues produces a conformational change in the enzyme
or the substrate that lowers the activation barrier for the
formation of the tetrahedral intermediate (Pearl, 1985). The
loss of binding energy with N- and C-terminal truncations can
be recovered to yield nanomolar level inhibitors by modifying
the para position of phenylalanine residues in the P1-P1'
positions of pseudo peptides with large, extended hydrophobic
substituents, particularly for HIV-PR and renin (Roberts et
al., 1990; Thompson et al., 1992; Young et al., 1992).
The importance of the P3 and P2 residues in binding and
catalysis suggests that the bound conformation of inhibitors
in these regions is highly conserved. This is true for the
P3/S3 interactions, but a duality of binding has been seen in
the P2/S2 interactions in many endothiapepsin and
rhizopuspepsin inhibitor complexes (Foundling et al., 1987;
Sali, et al., 1989; Suguna et al., 1992). These differences
are thought to be partially dependent upon the residue
present in the Pi' position of the inhibitor. This


I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy^
^ '~yy\
Ben M. Dunn, Chairman
Professor of Biochemistry
and Molecular Biology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
a dissertation for the degree of Docto]
as
.ori Philosophy.
Charles Allen
Professor of Biochemistry
and Molecular Biology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and jqCiiiXity, as
a dissertation for the degree of Doctdr^of Phil<
JJwA<
DanielPurich
Professor of Biochemistry
and Molecular Biology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
a dissertation for the degree of Doctor off* Philosophy.
as
Sheldon Schuster
Professor of Biochemistry
and Molecular Biology


LIST OF TABLES
Table Page
1-1. Crystal structures of native and inhibitor-
complexed rhizopuspepsin 10
1-2. Primary sequence comparison of rhizopuspepsin to
several aspartic proteinases 18
1-3. Partial sequence alignment of several aspartic
proteinases 20
3-1. Representative yields during the purification of
the recombinant rhizopuspepsins 63
4-1. Kinetic comparison between the naturally
occurring isozymes and WT-REC rhizopuspepsin
using the substrate Lys-Pro-P3~Lys-Phe*Nph-Arg-
Leu 69
4-2. Guanidinium hydrochloride denaturation parameters
of native and mutant forms of rhizopuspepsin 74
5-1. Inhibition constants for Xaa'F[CH2NH]Yaa modified
derivatives 83
5-2. Inhibition constants for LeuvP[CH(OH) CH2]Val and
statine modified derivatives 84
6-1. Kinetic parameters for WT-REC rhizopuspepsin with
the substrate Lys-Pro-Ala-Lys-Phe*Nph-Arg-Leu
containing systematic substitutions in P5-P1 and
P2'-P3' at pH 3.5 94
6-2. Kinetic analysis of WT-REC rhizopuspepsin with
substrates of the form Lys-Pro-Ile-P2-Phe*Nph-
Arg-Leu at pH 3.5 97
7-1. Kinetic analysis of wild-type and mutant
rhizopuspepsins: systematic substitution of
lysine into P3-P1 106
7-2. Kinetic analysis of porcine pepsin: systematic
substitution of lysine into P3-P1 108
viii


37
were calculated from equation 2 where at [GdnHCl]^, the
transition point, AG.2 = ^[GdnHCl]^ .
Kinetic Analysis
Kinetic assays using chromoaenic substrates. Substrate
hydrolysis, where cleavage occurs between Phe*Nph, Nle*Nph or
Lys*Nph (Nph = p-nitrophenylalanine, Nle = norleucine and =
site of cleavage), was monitored by the decrease in the
average absorbance from 284-324 nm using a Hewlett Packard
8452A diode array spectrophotometer (Scarborough et al.,
1993). The Km and Vmax values were determined from the
initial rates of at least six different peptide substrate
concentrations using Marquardt analysis (Marquardt, 1963) and
the equation v = VmajjSl/fKm + [S] ) The observed rates in
AU s_1 were converted to M s-1 by dividing by the total change
in absorbance for complete hydrolysis of a known
concentration of each substrate used. The amount of active
enzyme was determined by fitting the curve generated by the
competitive titration at one substrate concentration and 2%
DMSO with the inhibitor Val-Ser-Gln-Asn-LeuvP[CH(0H)CH2]Val-
Ile-Val (U85548E; Sawyer et al., 1992) with the Henderson
equation for tight binding inhibitors using the Enzfitter
program (Henderson, 1972; Leatherbarrow, 1987) The standard
deviations of the kcat and kcat/Km values were propagated
using equations derived by standard procedures for non-


36
[GdnHCl]50%, AGu'p and m. The method of Pace (1986) employs
equation 2 by plotting AG_F within the transition region
( 1.5 kcal mol-1) versus [denaturant] and linearly
extrapolating back, usually quite a long distance, to zero
denaturant to obtain AG^p Because small errors in m can
lead to large errors in the calculation of Aand
[GdnHCl]50% Fersht and coworkers have used a method,
represented by equation 3, which uses all the observed
fluorescence data, F, to directly determine [GdnHCl]50%
typically within 0.02 M guanidinium hydrochloride (Jackson
et al., 1993).
F = {(aN + /JN[GdnHCl]) + {av + Pu [GdnHCl]) x
exp[j([GdnHCl] [GdnHCl]50J / RT]) / (3)
{1 + exp[m([GdnHCl] [GdnHCl]^) / /?!]}
This equation combines equations 1 and 2 and assumes
that Fn and Fv are linearly dependent on the denaturant
concentration. orN and are the intercepts and /JN and
are the slopes of the baselines at denaturant concentrations
before and after the transition region. These parameters, as
well as, [GdnHCl]^ and m were allowed to be variables in the
KaleidaGraph non-linear regression analysis program. The
values of m and [GdnHCl]^ were obtained for the
rhizopuspepsins with their standard errors. AG.2 values


UJ -U Os VO fr
O W vj ^
i i i p
2
OJ
o
o\
K>
U)


128
Information from this study may also prove to be useful
in the study of the paired basic residue-specific aspartic
proteinases from yeast and the pituitary (Loh et al., 1985;
Azaryan et al., 1993). Yeast aspartic proteinase 3 (YAP3)
and pro-opiomelanocortin converting enzyme have been
suggested to play critical roles in prohormone processing.
These enzymes are characterized by the ability to cleave
between or after Lys-Arg junctions.


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EZLFRK0GN_0LWAFU INGEST_TIME 2015-03-25T19:01:52Z PACKAGE AA00029769_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


114
porcine pepsin. A negative value indicates that substrate
interactions are improved at pH 5.0. The similarity in the
values, between -0.5 and -1.0 kcal mol-1, for the recombinant
enzymes over the range of substrates tested implies that the
interactions seen between rhizopuspepsin and a substrate may
have a general electrostatic component. Asp30 and Asp77 do
not play a role in this phenomenon.
Interactions between porcine pepsin and this series of
substrates also show a dependence on pH. The observations
with lysine in P3 and P2 are probably related to interactions
with Glul3 and Glu287, respectively (Pohl and Dunn, 1988;
Rao-Naik, unpublished data). The dependence seen for peptide
5 is probably related to slight alterations in active site
interactions in order to optimize interactions with the Nle
residue. The lack of pH dependence for peptide 4 with pepsin
suggests that there may be a specific difference between
pepsin and rhizopuspepsin that is responsible for the general
electrostatic effect seen with rhizopuspepsin. All the
substrates in this series contain an arginine in the P2'
position. There is a highly conserved difference between the
mammalian and fungal enzymes at residue 37 in the S2' pocket
(Table 1-3). Rhizopuspepsin contains an Asp at this position
while porcine pepsin has an Asn. The suggestion that Asp37
may form a specific electrostatic interaction with the P2'
arginine awaits further analysis by mutagenesis studies.
The contrasting pH dependence in the kinetic parameters
for rhizopuspepsin and porcine pepsin for peptide 4 may also


52
kDa
M
94 -
j
67 -

43 -

30 -

20 -

14 -

2 34 5 6789 10 11 M


17
aspartic proteinases is for cleavage between large
hydrophobic-hydrophobic junctions, such as, Phe-Phe (Fruton,
1970, 1976). Secondary interactions have been shown to cause
large increases in catalytic efficiency, kcat without
corresponding changes in Km (Balbaa et al., 1993). These
interactions, though sometimes far from the site of cleavage,
as previously discussed have been implicated in dictating
highly specific preferences for substrates. Cathepsin D, for
example, is unable to efficiently cleave substrates with
basic residues in the P2 position (Scarborough et al., 1993).
This effect is thought to be generated by the presence of a
methionine residue at position 287 (pepsin numbering,
Scarborough et al., 1994). Porcine pepsin exhibits a similar
aversion for positive residues, but in the S3 subsite. This
effect was demonstrated, through the use of mutagenesis and
pH dependence kinetic studies, to be mediated by Glul3 (Pohl
& Dunn, 1988; Rao-Naik, unpublished results).
Rhizopuspepsin and other fungal enzymes are different
from these examples and other mammalian enzymes in that they
are able to cleave a wide range of substrates with similar
efficacy and 2-10 fold higher kcat/Km values (Dunn et al.,
1986: Lowther et al., 1991). Even though rhizopuspepsin has
broad specificity, it does possess an additional
characteristic that distinguishes itself from all mammalian
aspartic proteinases.


26
1
N.
PET3A-RPGN
Rhizopuspepsinogen
2
\ jm
4
primers 1 and 2
or
primers 3 and 4
Denature and Anneal
c
D
3
c; "4
EvvW-3 Ml "\0
primers 1 and 4
mm
BamH I

Nde I
E

Mutation
Figure 2-1. PCR mutagenesis procedure. A, 5'-end of
rhizopuspepsinogen; B, 3'-end; C, hybrid capable of
extension; D, hybrid not capable of extension; E, full length
rhizpuspepsinogen gene containing engineered mutations and
Nde I and BamH I restriction sites.


38
independent or correlated errors as outlined by Meyer (1975).
In those cases where the Km values were 1 mM, the kcat/Km
values were determined by fitting the initial rates of at
least six substrate concentrations ranging from 25-250 |iM to
the equation v = (kcat/Km) [E] o [S] o with the Enzfitter program
and the assumption that [S] Km. The kcat/Km values for the
cleavage of the Pi lysine-substituted peptides by the
Asp77Thr and the Asp30lle/Asp77Thr mutants were calculated
with the same equation on Enzfitter as above with the initial
rates determined by capillary electrophoresis.
Kinetic assays using competitive inhibitors. The
inhibition constant, Ki, was determined by monitoring the
competitive inhibition of the hydrolysis of the peptide Lys-
Ala-Ala-Lys-Phe*Nph-Arg-Leu (Km = 20 |JM) where cleavage
occurs strictly between Phe and Nph. All reactions were
performed at 37C in 0.1 M sodium formate buffer, pH 3.5 and
a final concentration of 4% DMSO. The initial rates of six
different substrate concentrations were measured following
preincubation of the enzyme without inhibitor for five
minutes. Additional curves were obtained, after
preincubation with two or more inhibitor concentrations, from
the initial rates of at least three different substrate
concentrations. The Ki value was determined from the family
of curves by the equation, v = Vmax[S]/[Km(1 + [I]/Ki) +
[S]]. If the Ki value determined by this method was one
nanomolar or lower, a competitive titration was performed as
described for the enzyme titration above.


47
Figure 3-2. SDS-PAGE analysis of wild-type
rhizopuspepsinogen at different stages of purification. Lane
1, E. coli whole cell lysate; lane 2, 27% sucrose pellet;
lane 3, Triton X-100 wash 1 supernatant; lane 4, Triton X-100
wash 2 supernatant; lane 5, concentrated, refolded zymogen
prior to loading onto a S-300 gel filtration column.


4
energy or the binding of the substrate in the transition
state. A true comparison of the binding energies is not
possible because of the different interactions of the wild-
type and mutant proteins with water. From the systematic
analysis of the mutant proteins, however, a pattern of
transition state stabilization free energy changes emerged.
The magnitude of the change was dependent on the type of
mutation and whether or not the hydrogen bond donor/acceptor
was charged or neutral.
The deletion of an enzyme side chain or substrate
hydrogen bond to an uncharged hydrogen bond donor/acceptor
resulted in a loss of transition state stabilization energy
of 0.5 to 1.8 kcal mol-1. This decrease represents a factor
of 2 to 15 toward specificity. The deletion of hydrogen bond
to a charged hydrogen donor/acceptor resulted in the loss of
3 to 6 kcal mol-1 of stabilization energy representing a
factor of 1000 or more in specificity. The same loss of
specificity was seen when both the donor and acceptor were
charged.
The results from these studies suggest that hydrogen
bonding between uncharged donor/acceptor do contribute to
specificity to some extent. Hydrogen bonds containing at
least one charged donor/acceptor, however, contribute
significantly to specificity. Further evidence to support
the critical role of electrostatic interactions in the
creation of specificity comes from the study of the
interaction of proteinases with substrates and inhibitors.


59
Interestingly, only upon mutation of Asp77 to Thr is there a
notable effect on the overall pi value of the 9 amino acid
extended form.
Increased yields of the 9 amino acid extension were
obtained for the D77T and the D30I/D77T proteins when the
activation was performed at 37C and pH 3.0 and 3.5,
respectively. The three additional positively charged
residues proved to be fortuitous in clearly separating the
two activation intermediates by ion-exchange (Figure 3-8).
The recombinant rhizopuspepsins at this final step in the
protocol were shown by IEF to be highly pure (Figure 3-9).
All structural and kinetic comparisons were made using the
Thr-Ser-Thr-form of the rhizopuspepsins. Yields at this
final step in the purification ranged from 1 to 5 mg for a
4 L expression. A summary of the yields during the
purification of the recombinant rhizopuspepsins is shown in
Table 3-1.
Discussion
The largest losses of rhizopuspepsinogen occurred during
the solubilization of the inclusion bodies and refolding.
Precipitation usually occurs during the refolding protocol.
The solubilization and refolding of rhizopuspepsinogen is
complicated by the presence of two disulfide bonds in its
tertiary structure. The inclusion bodies of prochymosin,
which contains three disulfide bonds, have been shown to


81
members (Parris et al., 1992; Rao et al., 1993; Scarborough
et al., 1993).
This chapter reports the structure-activity
relationships of wild-type recombinant rhizopuspepsin with
inhibitors containing various scissile bond isosteres and Pi-
Pl' substitutions. The results of inhibitors with different
lengths are also discussed. The primary goal of this study
is to identify inhibitors for use as active site titrants.
Results and Discussion
Renin inhibitors have been synthesized using information
from many studies where variations have been made to create
potential interactions with the active site binding cleft
(Wiley and Rich, 1993). These extensive surveys have not
only led to many novel and potent inhibitors for renin, but
also decreased the time to find potent HIV-PR inhibitors.
Three major directions have been taken to increase potency,
bioavailability and stability of inhibitors: (1) substitution
of the scissile peptide bond of naturally occurring cleavage
junctions with mimetics, (2) elaboration of the Pi* side
chain, and (3) the determination of inhibitor length
requirements. The crystal structures of many aspartic
proteinase complexes with a variety of inhibitors aids in the
rationalization of binding interactions in the active site
which lead to a preference for one inhibitor over another.


Figure 3-1. SDS-PAGE analysis of the expression of wild-type
rhizopuspepsinogen (RPGN) in E. coli upon the additon of IPTG
from 0 to 3 hours. The RPGN migrates at 43 kDa in comparison
to the molecular weight markers, M.


5
Mutagenesis studies have been undertaken with trypsin in
order to understand the source of its unique specificity
among serine proteinases.
Trypsin
Trypsin provides an excellent example of how an enzyme
has optimized electrostatic interactions to recognize and
preferentially cleave a particular class of substrates. The
primary specificity of trypsin is to cleave substrates
containing Arg or Lys in the Pi position (nomenclature of
Schechter & Berger, 1967). The x-ray crystal structures of
inhibitor complexes of trypsin have given insight into this
preference for basic residues (Ruhlmann et al., 1973; Krieger
et al., 1974; Perona et al., 1993, 1994). The bottom of the
Si binding pocket contains an Asp at position 189. When
lysine is present in the substrate, hydrogen bonding occurs
directly to Serl90 and through a water molecule to Aspl89.
The presence of Arg in the substrate, however, expels the
water molecule and hydrogen bonding occurs directly to Aspl89
and Serl90.
Extensive site-directed mutagenesis studies have been
performed to understand the contribution of Aspl89 and Serl90
to the specificity of trypsin. Mutagenesis has also been
carried out to see if changing Aspl89 converts trypsin into a
chymotrypsin-like enzyme. The Si binding pockets of
chymotrypsin and trypsin show a high degree of structural


130
Cooper, J., Foundling, S., Hemmings, A., Blundell, T., Jones,
D. M., Hallett, A. & Szelke, M. (1987) FEBS Lett. 169,
215-221.
Cooper, J., Quail, W., Frazao, C., Foundling, S. I.,
Blundell, T. L., Humblet, C., Lunney, E. A., Lowther, W.
T. & Dunn, B. M. (1992) Biochemistry 31, 8142-8150.
Cutfield, S., Marshall, C., Moody, P., Sullivan, P. &
Cutfield, J. (1993) J. Mol. Biol. 234, 1266-1269.
Davies, D. R. (1990) Annu. Rev. Biophys. Biophys. Chem. 19,
189-215.
Delaney, R., Wong, R., Meng, G., Wu, N. & Tang, J. (1987) J.
Biol. Chem. 262, 1461-1467.
Devereux, J., Haeberli, P. & Smithies, O. (1984) Nucleic
Acids Res. 12, 387-395.
Dill, K. (1990) Biochemistry 29, 7133-7155.
Dunn, B. M., Kammerman, B. & McKurry, K. (1984) Anal.
Biochem. 138, 68-73.
Dunn, B. M., Jimenez, M., Parten, B. F., Valler, M. J.,
Rolph, C. E. & Kay, J. (1986) Biochem. J. 237, 899-906.
Dunn, B. M., Valler, M. J., Rolph, C. E., Foundling, S. I.,
Jimenez, M. & Kay, J. (1987) Biochem. Biophys. Acta 913,
122-130.
Evnin, L., Vasquez, J. & Craik, C. (1990) Proc. Natl. Acad.
Sci. U.S.A 87, 6659-6663.
Faust, P. F., Kornfeld, S. & Chirgwin, J. M. (1985) Proc.
Natl. Acad. Sci. U.S.A. 82, 4910-4914.
Fersht, A. (1985) in Enzyme Structure and Mechanism, 2nd ed.,
W. H. Freeman and Co., New York.
Fersht, A. (1988) Biochemistry 27, 1577-1580.
Fersht, A., Shi, J., Knill-Jones, J., Lowe, D., Wilkinson,
A., Blow, D., Brick, P., Carter, P., Waye, M. & Winter,
G. (1985) Nature 314, 235-238.
Foundling, S. I., Cooper, J., Watson, F. E., Cleasby, A.,
Pearl, L., Sibanda, B. L., Hemmings, A., Wood, S. P.,
Blundell, T. L., Valler, M., Norey, C. G., Kay, J.,
Boger, J., Dunn, B., Leckie, B. J., Jones, D. M., Atrash,
B., Hallett, A. & Szelke, M. (1987) Nature 327, 349-352.


10
Table 1-1. Crystal structures of native and inhibitor-
complexed rhizopuspepsin
Structure Resolution
(A)
R value
Reference
Native
1.8
0.143
Suguna et al., 1987a
Pepstatin
2.5
0.145
Bott et al., 1982
U70531E
00
rH
0.147
Suguna et al., 1987b
CP-69,799
o
CN
0.171
Parris et al., 1992
CP-88,218
1.9
0.175
Parris et al., 1992
U85548E
2.0
0.170
unpublished results


SITE-DIRECTED MUTAGENESIS OF THE ASPARTIC PROTEINASE
RHIZOPUSPEPSIN: AN ANALYSIS OF UNIQUE SPECIFICITY
By
WILLIAM TODD LOWTHER
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1994


49
6 11 16
hours


66
occurring isozymes are on the surface of the protein far from
the active site. In order to investigate a possible kinetic
difference of the WT-REC enzyme from the pi 5 and 6 isozymes,
a series of substrates with systematic substitutions in P3
were examined. This comparison was also performed to rule
out the possible effects of the 9 residue N-terminal
extension of WT-REC on catalysis. IEF gel analysis was also
performed.
The overall tertiary structure of the native isozyme
pi 6 was compared to the WT-REC and mutant rhizopuspepsins by
guanidinium hydrochloride denaturation experiments. The
recombinant enzyme structures were also examined by circular
dichroism. These studies were undertaken to investigate
whether or not the mutations introduced into rhizopuspepsin
caused large conformational changes in the enzymes that may
compromise the interpretation of kinetic experiments.
Results
WT-REC and Native Isozvmes of Rhizopuspepsin
IEF. A comparison of the pi values of the naturally
occurring isozymes of rhizopuspepsin to WT-REC was made using
IEF gels and protein standards (Figure 4-1). The two native
isozymes, pi 5 and pi 6, exhibit pi values of 5.1 and 6.2,
respectively. The WT-REC enzyme has a pi of 5.7. This gel
also illustrates the large difference in pi values seen upon


13
peptide was confirmed by the crystal structure of the
inhibitor H-256, Pro-Thr-Glu-Phe'F[CH2NH] Phe-Arg-Glu, bound to
the active site of endothiapepsin (Cooper et al., 1987).
Studies soon followed delving deeper into the variations
seen in the specificity at different subsites. A systematic
series of peptides with substitutions in the P3 position was
made by Dunn et al. (1986), in the peptide Lys-Pro-Xaa-Glu-
Phe*Nph-Arg-Leu, and screened against a large variety of
mammalian and fungal enzymes. The results from these
experiments again showed a large diversity of specificity in
the S3 subsite. In another study, the pH dependence of the
kinetic parameters was studied with pepsin, chymosin, and
endothiapepsin cleaving substrates of the form Lys-Pro-Xaa-
Yaa-Phe*Nph-Arg-Leu where Xaa-Yaa were Ala-Glu, His-Ala and
Thr-Val (Dunn et al., 1987). From the examination of the
endothiapepsin structure complexed with H-256, and the trends
seen in the data, it was proposed that specific electrostatic
interactions with Glul3 (pepsin numbering) of the S3 pocket
of porcine pepsin were occurring. These studies also
suggested that electrostatic interactions may also be
important in the S2 subsite of porcine pepsin. Additional
studies by Pohl & Dunn (1988) provided further kinetic
evidence for electrostatic interactions in the S3 and S2
subsites of porcine pepsin from the study of the substrates
derived from Lys-Pro-Ala-Lys-Phe*Nph-Arg-Leu. Now that the
genes have been cloned for a variety of aspartic proteinases,
including rhizopuspepsin, the structure-activity


112
active site. Asp30 is located between the S3 and the Si
subsites and has the capacity to interact with the P3 and Pi
backbones and side chains. Besides the Pi interactions seen
in the lysine pepstatin derivative complex, crystallographic
analyses of penicillopepsin with renin inhibitors have shown
that Asp77 can also interact with a histidine residue in the
P2 position (Blundell et al., 1987). Lysine was substituted
into P3-P1 of the substrate in order to definitively show the
positional requirement for interaction with Asp77. These
peptides are also designed to give information about the
contribution of Asp30, if any. In this study, two peptides
(4 and 5) were used as controls representing substrates that
did not contain charges in positions P3-P1. The Nle peptide
was used to mimic the lysine side chain.
Kinetic Analysis
The values and pH dependence of the kinetic parameters
seen in Table 7-1 for peptide 3 are directly comparable to
those historically seen for rhizopuspepsin (Hofmann et al.,
1984; Balbaa et al., 1993). Hofmann suggested that the pH
dependence was indicative of a specific interaction of a
carboxyl group with the lysine residue in Pi. Interestingly,
the same pH dependence of kcat/Km is seen for all the peptides
and forms of rhizopuspepsin in this study. Table 7-3 lists
the changes in transition state stabilization energies upon
changing the assay pH for the recombinant rhizopuspepsins and


61
O
minutes
38


90
used to titrate the native isozymes and recombinant forms of
rhizopuspepsin.


134
Pohl, J. & Dunn, B. M. (1988) Biochem. 27, 4827-34.
Powers, J. C., Harley, A. D. & Myers, D. V. (1977) Adv. Exp.
Med. Biol. 95, 141-157.
Rao, C., Scarborough, P. E., Kay, J., Batley, B., Rapundalo,
S., Klutchko, S., Taylor, M. D., Lunney, E. A., Humblet,
C. C. & Dunn, B. M. (1993) J. Med. Chem. 36, 2614-2620.
Rheinnecker, M., Baker, G., Eder, J. Sc Fersht, A. R. (1993)
Biochemistry 32, 1199-1203.
Roberts, N., Martin, J., Kinchington, D., Broadhurst, A.,
Craig, J., Duncan, I., Galpin, S., Handa, B., Kay, J.,
Krohn, A., Lambert, R., Merrett, J., Mills, J., Parkes,
K., Redshaw, S., Ritchie, A., Taylor, D., Thomas, G. &
Machin, P. (1990) Science 248, 358-361.
Rodriquez, E. J., Angeles, T. S. & Meek, T. D. (1993)
Biochemistry 32, 12380-12385.
Rosenberg, S., Woods, K., Sham, H., Kleinert, H., Martin, D.
Stein, H., Cohen, J., Egan, D., Boop, B., Merits, I.,
Garren, K., Hoffman, D. & Plattner, J. (1990) J. Med.
Chem. 33, 1962-1969.
Ruhlmann, A., Kukla, D., Schwager, P., Bartels, K. & Huber,
R. (1973) J-. Mol. Biol. 77, 417-436.
Sali, A., Veerapandian, B., Cooper, J. B., Foundling, S. I.,
Hoover, D. J. Sc Blundell, T. L. (1989) EMBO J. 8, 2179-
88.
Salituro, F. C., Agarwal, N., Hofmann, T. Sc Rich, D. H.
(1987) J. Med. Chem. 30, 286-295.
Samaranayake, L. & Holmstrup, P. (1989) J. Oral Pathol. Med.
18, 554-64.
Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) in
Molecular Cloning, Cold Spring Harbor Laboratory Press,,
New York.
Sampath-Kumar, P. S. & Fruton, J. S. (1974) Proc. Natl. Acad
Sci. U.S.A. 71:4, 1070-1072.
Saral, R. (1991) Rev. Infect. Dis. 13, 487-492.
Sawyer, T. K., Hester, J., Schostarez, H., Thaisrivongs, S.,
Bundy, G., Liu, L., Bradford, V., DeVaux, A., Staples, D
J., Maggiora, L. L., TenBrink, R., Kinner, J., Smith, C.
W., Pals, D. T., Couch, S., Hinzmann, J., Poorman, R. A.
Einspahr, H., Finzel, B., Watenbaugh, K., Mao, B., Epps,


133
Loh, Y., Parish, D. & Tuteja, R. (1985) J. Biol. Chem. 260,
7194-7205.
Lowe, D., Winter, G. & Fersht, A. (1987) Biochemistry 26,
6038-6043.
Lowther, W. T., Chen, Z., Lin, X., Tang, J. and Dunn, B. M.
(1991) in Structure and Function of the Aspartic
Proteinases: Genetics, Structure and Mechanisms. (Dunn,
B. M., Ed.) pp 275-279, Plenum Press, New York.
Lunney, E., Hamilton, H., Hodges, J., Kaltenbronn, J.,
Repine, J., Badasso, M., Cooper, J., Dealwis, C.,
Wallace, B., Lowther, W. T., Dunn, B. M. & Humblet, C.
(1993) J. Med. Chem. 36, 3809-3820.
Marquardt, D. (1963) J. Soc. Indust. Appl. Math. 11, 431-441.
Medzihradszky, K., Voynick, I., Medzihradszky-Schweiger, H. &
Fruton, J. (1970) Biochemistry 9, 1154-1162.
Metcalf, P. & Fusek, M. (1993) EMBO J. 12, 1293-1302.
Meyer, S. L. (1975) in Data Analysis for Scientists and
Engineers, John Wiley and Sons, Inc., New York.
Mildvan, A. S., Weber, D. J. & Kuliopulos, A. (1992) Arch.
Biochem. Biophys. 294, 327-340.
Morihara, K. & Oka, T. (1973) Arch. Biochem. Biophys. 157,
561-572.
Morrison, C., Hurst, S., Bragg, S., Kuyendall, R., Diaz, H.,
McLaughlin, D. & Reiss, E. (1993) J. Gen. Microbiol. 139,
1177-1186.
Pace, C. (1986) Methods Enzymol. 131, 266-280.
Parris, K. D., Hoover, D. J., Damon, D. B. & Davies, D. R.
(1992) Biochemistry 31, 8125-8141.
Paya, C. (1993) Clin. Infect. Dis. 16, 677-688.
Pearl, L. (1985) in Aspartic Proteinases and Their Inhibitors
(Kostka, V., Ed.) pp 189-195, Walter de Gruyter, Berlin.
Pearl, L. & Blundell, T. (1984) FEBS Lett. 174, 96-101.
Perona, J. J., Hedstrom, L., Wagner, R., Rutter, W., Craik,
C. S. Sc Fletterick, R. J. (1994) Biochemistry 33, 3252-
3259.
Perona, J. J., Tsu, C. A., Craik, C. S. & Fletterick, R. J.
(1993) J. Mol. Biol. 230, 919-933.


SITE-DIRECTED MUTAGENESIS OF THE ASPARTIC PROTEINASE
RHIZOPUSPEPSIN: AN ANALYSIS OF UNIQUE SPECIFICITY
By
WILLIAM TODD LOWTHER
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1994

This work is dedicated to all those who have given freely of
their love, support, guidance and patience; my Lord and
Savior Jesus Christ, family, friends and colleagues.

ACKNOWLEDGMENTS
I have always been encouraged to be my best and to be
true to myself. This support has extended from family and
friends. In particular, I want to thank my mother and
father, Ruth and Mike, for their love, generosity and
willingness to do whatever it took to make sure that my
brothers, Jason and Pat, and I had everything we could have
ever needed or wanted. I also want to thank Bill, Dee,
Brandon and Kara for their encouragement and love. My
grandparents have also been a continual source of wisdom,
support and love. Encouragement from Andy and Cheryl has
also been very reassuring.
I am indebted to Dr. Ben Dunn for providing a fantastic
environment for learning and scientific opportunity. I am
particularly thankful for his insights and willingness to
allow me to pursue my research interests, within reason.
I would like to thank my committee members for their
suggestions and encouragement during my graduate work; Drs.
Daniel Purich, Charles Allen, Sheldon Schuster, and Nigel
Richards.
The support of the Dunn laboratory has made my stay at
the University of Florida one of the best times of my life.
In particular, I want to thank Chetana, Paula, Wieslaw, Bill,
iii

Wichet, Brian, Jenny, and the whole rest of gang for their
friendships. The efforts and friendships of the Protein
Chemistry Core Facility have also made my stay in Gainesville
enjoyable. In particular, I want to thank Ruth, Nancy, Hung,
Benne, and the whole rest of the gang.
IV

TABLE OF CONTENTS
page
ACKNOWLEDGMENTS iii
LIST OF TABLES viii
LIST OF FIGURES X
KEY TO SYMBOLS xii
ABSTRACT xiv
CHAPTERS
1 HYDROGEN BONDING AND BIOLOGICAL SPECIFICITY 1
Introduction 1
Tyrosyl tRNA Synthetase 3
Trypsin 5
Aspartic Proteinases 7
General Characteristics 7
Initial Kinetic Studies 12
Rhizopuspepsin 14
Historical Background 15
Specificity Differences Studied by Site-Directed
Mutagenesis 16
2 EXPERIMENTAL PROCEDURES 23
Introduction 23
Materials 23
Methods 25
Cloning and Mutagenesis 25
Expression 30
Refolding 30
Size-Exclusion Chromatography 31
Activation and Ion-Exchange Chromatography 32
Structural Characterization 32
Kinetic Analysis 37
Analysis of Transition State Effects 40
Molecular Graphics 41
3 EXPRESSION, REFOLDING, PURIFICATION, AND ACTIVATION
OF RECOMBINANT RHIZOPUSPEPSINS 42
Introduction 42
v

Results 43
Mutagenesis 43
Expression and Refolding 43
Activation and Purification 50
Discussion 59
4 KINETIC AND STRUCTURAL AUTHENTICITY OF RECOMBINANT
RHIZOPUSPEPSINS 65
Introduction 65
Results 66
WT-REC and Native Isozymes of Rhizopuspepsin 66
Structural Comparisons 68
Discussion 75
5 ANALYSIS OF THE SPECIFICITY OF RHIZOPUSPEPSIN
THROUGH THE USE OF INHIBITORS CONTAINING Pi AND Pi'
SUBSTITUTIONS AND PEPTIDE BOND MIMETICS 80
Introduction 80
Results and Discussion 81
6 THE BROAD SUBSTRATE SPECIFICITY OF RHIZOPUSPEPSIN:
ANALYSIS WITH SYSTEMATIC SUBSTITUTIONS IN P5-P1 AND
P2 1-P31 OF THE SUBSTRATE LYS-PRO-ALA-LYS-PHE*NPH-
ARG-LEU 91
Introduction 91
Results 92
Lys-Pro-Ala-Lys-Phe*Nph-Arg-Leu based substrates ... 92
Lys-Pro-Ile-P2-Phe*Nph-Arg-Leu based substrates 96
Discussion 96
Comparison to the Mammalian Enzymes 98
Comparison to the Candida Aspartic Proteinases 99
7 ENGINEERING THE SUBSTRATE SPECIFICITY OF
RHIZOPUSPEPSIN: THE ROLE OF ASP30 AND ASP77 OF A
FUNGAL ASPARTIC PROTEINASE TO CLEAVE SUBSTRATES
WITH LYSINE IN Pi 101
Introduction 101
Results 104
Kinetic Analysis of the Recombinant
Rhizopuspepsins 104
Kinetic Analysis of Porcine Pepsin 109
Discussion 110
Substrate Design Ill
Kinetic Analysis 112
8 CONCLUSIONS AND FUTURE DIRECTIONS 123
Conclusions 123
Future Directions 125
vi

LIST OF REFERENCES 129
BIOGRAPHICAL SKETCH 138
Vll

LIST OF TABLES
Table Page
1-1. Crystal structures of native and inhibitor-
complexed rhizopuspepsin 10
1-2. Primary sequence comparison of rhizopuspepsin to
several aspartic proteinases 18
1-3. Partial sequence alignment of several aspartic
proteinases 20
3-1. Representative yields during the purification of
the recombinant rhizopuspepsins 63
4-1. Kinetic comparison between the naturally
occurring isozymes and WT-REC rhizopuspepsin
using the substrate Lys-Pro-P3~Lys-Phe*Nph-Arg-
Leu 69
4-2. Guanidinium hydrochloride denaturation parameters
of native and mutant forms of rhizopuspepsin 74
5-1. Inhibition constants for Xaa'F[CH2NH]Yaa modified
derivatives 83
5-2. Inhibition constants for LeuvP[CH(OH) CH2]Val and
statine modified derivatives 84
6-1. Kinetic parameters for WT-REC rhizopuspepsin with
the substrate Lys-Pro-Ala-Lys-Phe*Nph-Arg-Leu
containing systematic substitutions in P5-P1 and
P2'-P3' at pH 3.5 94
6-2. Kinetic analysis of WT-REC rhizopuspepsin with
substrates of the form Lys-Pro-Ile-P2-Phe*Nph-
Arg-Leu at pH 3.5 97
7-1. Kinetic analysis of wild-type and mutant
rhizopuspepsins: systematic substitution of
lysine into P3-P1 106
7-2. Kinetic analysis of porcine pepsin: systematic
substitution of lysine into P3-P1 108
viii

7-3. Transition state stabilization energy changes
seen with variation in pH from 3.5 to 5.0 for the
recombinant rhizopuspepsins and porcine pepsin 113
7-4. Double mutant cycle analysis of the recombinant
rhizopuspepsins: substitution of lysine into P3-
Pl at pH 3.5 and 5.0 117
IX

LIST OF FIGURES
Figure page
1-1. Ribbon representation of the aspartic proteinase
rhizopuspepsin complexed with a reduced peptide
bond inhibitor 9
1-2. Closeup view of the reduced peptide bond
inhibitor bound to the active site of
rhizopuspepsin 11
1-3. Closeup view of the active site of rhizopuspepsin
highlighting the catalytic aspartic acid
residues, Asp32 and 215, and Asp30 and Asp77 21
2-1. PCR mutagenesis procedure 26
2-2. pET3aE expression vector containing
rhizopuspepsinogen 29
3-1. SDS-PAGE analysis of the expression of wild-type
rhizopuspepsinogen (RPGN) in E. coli upon the
addition of IPTG from 0 to 3 hours 46
3-2. SDS-PAGE analysis of wild-type rhizopuspepsinogen
at different stages of purification 47
3-3. Gel filtration elution profile of refolded
Asp30lle rhizopuspepsinogen 49
3-4. SDS-PAGE analysis of the fractions from the
purification of Asp30lle by gel filtration in
Figure 3-3 52
3-5. SDS-PAGE analysis of the time course of
activation of wild-type rhizopuspepsinogen upon
lowering the pH at room temperature 54
3-6. SDS-PAGE analysis of the activation of wild-type
recombinant rhizopuspepsinogen at pH 4.0 56
3-7. IEF analysis of the wild-type and mutant
rhizopuspepsinogens activated at pH 2.0 58
3-8. Elution profile of Asp77Thr from the Mono S
column after activation at pH 3.0 61
x

3-9. IEF comparison of the activated, purified
recombinant rhizopuspepsins 62
4-1. IEF comparison of wild-type rhizopuspepsinogen,
activated, purified WT-REC and the two naturally
occurring isozymes 67
4-2. CD spectra of the recombinant wild-type and
mutant forms of rhizopuspepsin 71
4-3. Fluorescence emission spectra for folded (0 M
GdnHCl) and unfolded (6 M GdnHCl) wild-type
recombinant rhizopuspepsin 72
4-4. Guanidinium hydrochloride induced unfolding of
the naturally occurring isozyme pi 6 and the
recombinant forms of rhizopuspepsin monitored by
the change in intrinsic fluorescence at 350 nm 73
7-1. Hydrogen bonding interactions in penicillopepsin
between Asp77, Ser79 and the pepstatin derivative
containing lysine in the Pi position 119
7-2. Ca carbon backbone superposition of
penicillopepsin and rhizopuspepsin complexed with
inhibitors 120
7-3. Proposed hydrogen bonding in WT-REC
rhizopuspepsin and Asp77Thr mutants with
substrates containing lysine in Pi 121
xi

KEY TO SYMBOLS
amp
ampicillin
AU
absorption units
C
carboxyl
Ca
alpha carbon
CAPS
3-(cyclohexylamino)-1-propanesulfonic acid
CD
circular dichroism
d
deoxy
DMSO
dimethylsulfoxide
E. coli
Escherichia coli
EDTA
ethylenediaminetetraacetic acid
FPLC
fast protein liquid chromatography
h
Planck's constant
HCl
hydrochloric acid
HIV
human immunodeficiency virus
HPLC
high performance liquid chromatography
IEF
isoelectric focusing
IPTG
isopropylthio-P~D-galactopyranoside
kcal
kilocalorie
kb
Boltzmann's constant
kcat
turnover number
KCl
potassium chloride
Ki
inhibition constant
km
Michaelis-Menten constant
LB
Luria Broth
MES
2-(4-morpholino)-ethane sulfonic acid
mg
milligram
MgCl2
magnesium chloride
min
minutes
Xll

ml (s)
milliliter(s)
mM
millimolar
MOPS
3-(N-morpholino) propane-sulfonic acid
MWCO
molecular weight cut off
N
amino
NaCl
sodium chloride
ng
nanogram
nM
nanomolar
nm
nanometers
Nph
p-nitrophenylalanine
NTP
nucleotide triphospahtes
OD
optical density
ori
origin
PAGE
polyacrylamide gel electrophoresis
PI
isoelectric point
pmol
picomoles
PVDF
polyvinylidene difluoride
rec
recombinant
s
second
SDS
sodium dodecyl sulfate
sec
seconds
tet
tetracycline
TFA
trifluoroacetic acid
Tricine
N-[Tris-(hydroxymethyl) Methyl] glycine
Tris
tris (hydroxymethyl) aminomethane
Vmax
maximum velocity
WT-REC
wild-type recombinant
fig
microgram
|IM
micromolar
Xlll

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
SITE-DIRECTED MUTAGENESIS OF THE ASPARTIC PROTEINASE
RHIZOPUSPEPSIN: AN ANALYSIS OF UNIQUE SPECIFICITY
By
William Todd Lowther
August, 1994
Chairman: Dr. Ben M. Dunn
Major Department: Biochemistry and Molecular Biology
Rhizopuspepsin is a secreted aspartic proteinase from
the fungus Rhizopus chinensis. Despite the high degree of
structural homology among the aspartic proteinases, the amino
acid residues that create the individual binding pockets have
been shown to profoundly influence substrate specificities
and inhibitor preferences. Rhizopuspepsin and other fungal
aspartic proteinases are distinct from the mammalian enzymes
in that they are able to cleave substrates with lysine in the
Pi position. Sequence and structural comparisons suggest
that two aspartic acid residues, Asp77 and Asp30 (pig pepsin
numbering), may be responsible for generating this unique
specificity in rhizopuspepsin.
In order to analyze their contributions to specificity,
Asp30 and Asp77 were changed to the corresponding residues in
xiv

porcine pepsin, Ile30 and Thr77, to create single and double
mutants. The zymogen forms of the wild-type and mutant forms
of the enzymes were overexpressed in E. coli as inclusion
bodies. Following denaturation, refolding, activation and
purification to homogeneity, structural and kinetic
comparisons were made. These comparisons have shown that the
wild-type recombinant enzyme is kinetically and structurally
indistinguishable from the naturally occurring isozymes. The
mutant proteins were also shown to exhibit a high degree of
similarity.
Characterization of the wild-type and mutant proteins
with chromogenic substrates with systematic substitution of
lysine into the Pi, P2 and P3 positions has shown that Asp77
is the predominant residue responsible for enabling the
catalysis of substrates with lysine in Pi. Mutation of Asp77
resulted in a loss of 7 kcal mol-1 of transition state
stabilization energy. The Asp30lle mutant was still able to
cleave the Pi-Lys peptide with near wild-type efficiency.
These observations suggest that it may be possible to exploit
the Asp77 interaction to design compounds that preferentially
inhibit a variety of related, secreted Candida proteinases in
immunocompromised patients.
xv

CHAPTER 1
HYDROGEN BONDING AND BIOLOGICAL SPECIFICITY
Introduction
Biological processes are controlled by a variety of
forces that influence intramolecular and intermolecular
interactions. A balance of the positive and negative aspects
of these forces, for example, determines the structure of DNA
and proteins, establishes the specificity of binding
interactions needed for information transfer and substrate
recognition, and creates the environmental requirements for
enzymatic catalysis. These forces have been classically
divided into two categories: (1) forces which lead to binding
energy, dispersion or van der Waals forces and the
hydrophobic effect, and (2) forces which relate to
specificity or the discrimination of one molecule over
another, electrostatic interactions and hydrogen bonds
(Jencks, 1969; Fersht, 1985; Fersht et al., 1985; Dill,
1990) .
The contribution of van der Waals forces and the
hydrophobic effect to protein folding and binding
interactions is thought to extend primarily from an increase
in entropy upon release of water molecules to the bulk
solvent. Even though there is, for example, a large decrease
1

2
in entropy in the organizational process of protein folding,
the release of water molecules and the establishment of an
intricate network of intramolecular hydrogen bonds
compensates for the entropy loss making the entire process
slightly favorable overall. The observation that almost all
possible hydrogen bonds are established upon the folding of a
protein points to the difficulty, in separating and
quantifying the forces involved. Despite this difficulty
trends are still observable. A survey of many x-ray crystal
structures of ligands bound to proteins has shown that there
is a strong correlation between the binding strength of
ligands and the degree of buried surface area (Horton &
Lewis, 1992; Young et al., 1994). This correlation is
dependent on the geometric or steric requirements for the
optimization of packing and hydrophobic interactions, but is
also influenced by the chemical complementarity between the
two species.
Chemical complementarity and biological specificity in
protein-ligand interactions are primarily governed by
hydrogen bonds and electrostatic interactions (Fersht et al.,
1985). A hydrogen bond occurs when two electronegative atoms
share a hydrogen atom. The optimal arrangement for this
interaction is linear with a distance between the
electronegative atoms ranging from 2.85 to 3 angstroms. The
study of the contribution of hydrogen bonds to specificity
has been facilitated by the use of site-directed mutagenesis.
The influence of hydrogen bonding in recognition and

3
catalysis has been studied by deleting a hydrogen bond donor
or acceptor on the enzyme or by making similar modifications
to the substrate. Two systems which have been extensively
studied are tyrosyl-tRNA synthetase and the pancreatic
proteinase trypsin.
Tvrosvl-tRNA Synthetase
One of the most thoroughly studied model systems for
studying the role of hydrogen bonding in specificity has been
the tyrosyl-tRNA synthetase from Bacillus stearothermophilus.
The tyrosyl-tRNA synthetase ensures the fidelity of
information transfer from the genetic code to the final
protein product by optimizing interactions with the
structural components of the amino acid tyrosine that make it
different from the other amino acids and, in particular,
phenylalanine. The examination of the crystal structure of
the enzyme bound aminoacyl adenylate has shown that there are
eight hydrogen bonds between the enzyme and the substrate
which can be studied by mutation of the enzyme. Fersht and
his coworkers have analyzed the effects of mutations by
comparing the kcat/Km values of the wild-type enzyme to the
mutant enzyme for the activation of tyrosine and
phenylalanine (Carter et al., 1984; Fersht, 1985, 1988;
Fersht, et al., 1985; Leatherbarrow et al., 1985; Lowe et
al., 1987). This comparison gives information about the
overall apparent change in transition state stabilization

4
energy or the binding of the substrate in the transition
state. A true comparison of the binding energies is not
possible because of the different interactions of the wild-
type and mutant proteins with water. From the systematic
analysis of the mutant proteins, however, a pattern of
transition state stabilization free energy changes emerged.
The magnitude of the change was dependent on the type of
mutation and whether or not the hydrogen bond donor/acceptor
was charged or neutral.
The deletion of an enzyme side chain or substrate
hydrogen bond to an uncharged hydrogen bond donor/acceptor
resulted in a loss of transition state stabilization energy
of 0.5 to 1.8 kcal mol-1. This decrease represents a factor
of 2 to 15 toward specificity. The deletion of hydrogen bond
to a charged hydrogen donor/acceptor resulted in the loss of
3 to 6 kcal mol-1 of stabilization energy representing a
factor of 1000 or more in specificity. The same loss of
specificity was seen when both the donor and acceptor were
charged.
The results from these studies suggest that hydrogen
bonding between uncharged donor/acceptor do contribute to
specificity to some extent. Hydrogen bonds containing at
least one charged donor/acceptor, however, contribute
significantly to specificity. Further evidence to support
the critical role of electrostatic interactions in the
creation of specificity comes from the study of the
interaction of proteinases with substrates and inhibitors.

5
Mutagenesis studies have been undertaken with trypsin in
order to understand the source of its unique specificity
among serine proteinases.
Trypsin
Trypsin provides an excellent example of how an enzyme
has optimized electrostatic interactions to recognize and
preferentially cleave a particular class of substrates. The
primary specificity of trypsin is to cleave substrates
containing Arg or Lys in the Pi position (nomenclature of
Schechter & Berger, 1967). The x-ray crystal structures of
inhibitor complexes of trypsin have given insight into this
preference for basic residues (Ruhlmann et al., 1973; Krieger
et al., 1974; Perona et al., 1993, 1994). The bottom of the
Si binding pocket contains an Asp at position 189. When
lysine is present in the substrate, hydrogen bonding occurs
directly to Serl90 and through a water molecule to Aspl89.
The presence of Arg in the substrate, however, expels the
water molecule and hydrogen bonding occurs directly to Aspl89
and Serl90.
Extensive site-directed mutagenesis studies have been
performed to understand the contribution of Aspl89 and Serl90
to the specificity of trypsin. Mutagenesis has also been
carried out to see if changing Aspl89 converts trypsin into a
chymotrypsin-like enzyme. The Si binding pockets of
chymotrypsin and trypsin show a high degree of structural

6
homology. Despite this similarity, chymotrypsin has a
primary specificity for large aromatic residues. The most
notable substitution that may be responsible for the
specificity differences between trypsin and chymotrypsin
occurs at position 189, where Asp has been replaced by Ser.
The mutation of Aspl89 to Ser in trypsin resulted in a
dramatic decrease in the kcat/Km values for Pi Arg and Lys
containing substrates (Graf et al., 1988). The average loss
of transition state stabilization energy as a consequence of
this mutation was 6.7 kcal mol-1. This value is in the same
range as those seen by Fersht for the deletion of a charged
hydrogen bond donor/acceptor discussed above. Experiments
were also performed to attempt to rescue the basic residue
specificity. The substitution of Asp at position 190 did
restore activity (Evnin et al., 1990). Experiments have also
shown that activity can be restored when acetate is present
in the buffer at very high levels (Perona, et al., 1994).
Interestingly, the Aspl89Ser mutant did show some
improvement toward cleaving large hydrophobic substrates.
Further mutagenesis experiments tried to complete the
conversion by mutating the remaining residues in the Si
pocket of trypsin to those in chymotrypsin. This effort
still did not result in a complete metamorphosis to
chymotrypsin. Only upon changing surface loops around Pi
subsite of trypsin to those of chymotrypsin in conjunction
with the Pi substitutions resulted in the desired enzyme
specificity (Hedstrom et al., 1992).

7
The studies with tyrosyl-tRNA synthetase and trypsin are
only two examples of the extensive literature establishing
the importance of hydrogen bonding and electrostatic
interactions in the creation of specificity in biological
reactions. Mutational and kinetic analysis have also been
used to dissect the structural components that generate the
unique specificities of cysteine and metallo proteinases.
Studies on the aspartic proteinases have the potential to
give further insight into the role of electrostatic
interactions in specificity. In particular, rhizopuspepsin,
the aspartic proteinase from the fungus Rhizopus chinensis,
contains several aspartic acid residues in the active site
that may influence the selection of substrates for
hydrolysis.
Asoartic Proteinases
General Characteristics
All members of the aspartic proteinase family show
marked inhibition by pepstatin (Davies, 1990) and a high
degree of amino acid sequence and three-dimensional structure
homology, especially in the region of the active site, as
shown by the studies of Blundell et al. (Pearl & Blundell,
1984; Blundell et al., 1987; Sali et al., 1989). The bilobal
structure of an aspartic proteinase is created by the nearly
symmetrical N-terminal and C-terminal domains of the protein.

8
The extended cleft of the active site is formed by the
interaction of the two domains (Figure 1-1). Each domain
also contributes one catalytic aspartic acid at the bottom of
the active site. An elaborate network of hydrogen bonds
maintains these aspartic acid residues (Asp32 and Asp215,
porcine pepsin numbering) in a juxtaposed or opposing
orientation. A centrally located water molecule, hydrogen-
bonded to each aspartic acid residue, is thought to act as
the nucleophile in a base-catalyzed attack of the scissile
bond carbonyl of the substrate (Suguna et al., 1987; Fraser
et al., 1992; James et al., 1992).
From the examination of inhibitor complexes of
rhizopuspepsin (Table 1-1), porcine pepsin (Abad-Zapatero et
al., 1991), endothiapepsin (Veerapandian et al., 1990; Lunney
et al., 1993), cathepsin D (Baldwin et al., 1993; Metcalf &
Fusek, 1993) and the HIV proteinase (Swain et al., 1990,
1991), it is evident that there is a consistent binding mode
for ligands (Figure 1-2). Ligands seven to eight residues
long completely fill the active site in an extended |3-strand
conformation with the amino acid side chains alternating in a
regular fashion. This uniform binding or anchoring of
ligands to the active site is attributed to a highly
conserved hydrogen bonding network, between the enzyme and
the a-carbon backbone of ligands, and the preference for
large hydrophobic or aromatic substituents on either side of
the scissile bond (Pi-Pi1) or the site of cleavage.

9
Figure 1-1. Ribbon represention of the aspartic proteinase
rhizopuspepsin complexed with a reduced peptide bond
inhibitor. The catalytic aspartic acid residues, Asp32 and
Asp215, are shown in red. The flap which extends over the
active site is shown in orange. The inhibitor is shown in
yellow. Coordinates were obtained from the Brookhaven
Protein Data Bank file 3APR (Suguna et al., 1987).

10
Table 1-1. Crystal structures of native and inhibitor-
complexed rhizopuspepsin
Structure Resolution
(A)
R value
Reference
Native
1.8
0.143
Suguna et al., 1987a
Pepstatin
2.5
0.145
Bott et al., 1982
U70531E
00
rH
0.147
Suguna et al., 1987b
CP-69,799
o
CN
0.171
Parris et al., 1992
CP-88,218
1.9
0.175
Parris et al., 1992
U85548E
2.0
0.170
unpublished results

11
Figure 1-2. Closeup view of a reduced peptide bond inhibitor
bound to the active site of rhizopuspepsin (Suguna et al.,
1987b). This figure illustrates the binding mode of ligands
to the active site of aspartic proteinases and illustrates
the Schechter and Berger nomenclature for describing active
site interactions. For example, the side chain of the P3
residue of the ligand interacts with the S3 subsite of the
enzyme. Bond cleavage occurs between Pi and Pi'.

12
initial Kinetic Studies
The initial studies characterizing the hydrolytic
properties of mammalian and fungal aspartic proteinases
utilized small tri- and tetrapeptides (Fruton, 1970, 1976).
Even with these small, poorly binding substrates, differences
in the "secondary interactions," those interactions not at
the scissile bond, were seen for pepsin, cathepsin D and
rhizopuspepsin (Voynick & Fruton, 1971; Fruton, 1976).
Subsequent studies (Sampath-Kumar & Fruton, 1974; Hofmann et
al., 1988; Balbaa et al., 1993) showed an increase in kcat as
the substrate length was extended to eight residues,
particularly when the S3, S2 and S21 subsites were occupied.
A further enhancement of substrates for the study of this
family of enzymes came when Hofmann and Hodges (1982) showed
that the change in absorbance of a p-nitrophenylalanine
residue in the Pi' position of the substrate would be greater
upon substrate hydrolysis than when present in the Pi
position (Hofmann et al., 1984). A slight modification of
Hofmann's substrates which contained lysine in Pi was made by
Dunn and coworkers (Dunn et al., 1984) using the information
from a large survey by Powers et al. (1977) on all known
cleavage junctions of the pepsin at that time. This work
yielded a chromogenic substrate, Pro-Thr-Glu-Phe*Nph-Arg-Leu
(Nph = p-nitrophenylalanine), that could be cleaved by
porcine pepsin in a continuous assay allowing the
quantitation of initial rates. The binding mode of this

13
peptide was confirmed by the crystal structure of the
inhibitor H-256, Pro-Thr-Glu-Phe'F[CH2NH] Phe-Arg-Glu, bound to
the active site of endothiapepsin (Cooper et al., 1987).
Studies soon followed delving deeper into the variations
seen in the specificity at different subsites. A systematic
series of peptides with substitutions in the P3 position was
made by Dunn et al. (1986), in the peptide Lys-Pro-Xaa-Glu-
Phe*Nph-Arg-Leu, and screened against a large variety of
mammalian and fungal enzymes. The results from these
experiments again showed a large diversity of specificity in
the S3 subsite. In another study, the pH dependence of the
kinetic parameters was studied with pepsin, chymosin, and
endothiapepsin cleaving substrates of the form Lys-Pro-Xaa-
Yaa-Phe*Nph-Arg-Leu where Xaa-Yaa were Ala-Glu, His-Ala and
Thr-Val (Dunn et al., 1987). From the examination of the
endothiapepsin structure complexed with H-256, and the trends
seen in the data, it was proposed that specific electrostatic
interactions with Glul3 (pepsin numbering) of the S3 pocket
of porcine pepsin were occurring. These studies also
suggested that electrostatic interactions may also be
important in the S2 subsite of porcine pepsin. Additional
studies by Pohl & Dunn (1988) provided further kinetic
evidence for electrostatic interactions in the S3 and S2
subsites of porcine pepsin from the study of the substrates
derived from Lys-Pro-Ala-Lys-Phe*Nph-Arg-Leu. Now that the
genes have been cloned for a variety of aspartic proteinases,
including rhizopuspepsin, the structure-activity

14
relationships proposed in these studies can be tested by
site-directed mutagenesis and the heterologous expression of
these proteins.
Rhizopuspepsin
Rhizopuspepsin, a fungal aspartic proteinase from
Rhizopus chinensis, has been the subject of many substrate,
inhibitor and crystallographic studies in an effort to better
understand factors contributing to catalysis and specificity
differences between members of this enzyme family. Initial
interest in the fungal enzymes arose from the need to find a
suitable replacement for calf chymosin in the cheese-making
process (Davies, 1990). This focus changed with the
determination of the native structure (Suguna et al., 1987)
and many different inhibitor complexes of rhizopuspepsin
(Suguna, et al., 1987; Parris et al., 1992; Suguna et al.,
1992) and other fungal enzymes (James et al., 1982; Cooper,
et al., 1987; Lunney, et al., 1993). The insights gained
from these complexes have led to proposals for the reaction
mechanism (James et al., 1977, 1992; Suguna, et al., 1987
Fraser et al., 1992;) and the rationalization of subsite
preferences of the aspartic proteinases (Rao et al., 1993;
Scarborough et al., 1993). The goal has been to use this
information to design specific or targeted therapeutics for
renin and the HIV proteinase. Further clinical interest in
rhizopuspepsin comes from its similarity to the secreted

15
aspartic proteinases of several opportunistic Candida
species (Fusek et al., 1993; Morrison et al., 1993). The
knowledge gleaned from studying rhizopuspepsin has the
potential to foster the design of anti-fungal agents for use
in treating vaginal infections and immunocompromised AIDS,
organ transplant and cancer patients (Samaranayake &
Holmstrup, 1989; Saral, 1991; Paya, 1993).
Historical Background
Rhizopuspepsin was first isolated by Fukumoto et al.
(1967). Purification of rhizopuspepsin by isoelectric
focusing identified two major isozymes (Graham et al., 1973).
These two isozymes have been shown to be very similar in
molecular weight, amino acid composition, specific activity,
as well as in three-dimensional structure as shown by the
crystal structure of the isozyme mixture (Subramanian et al.,
1977). N- and C-terminal sequencing showed that the two
isozymes were identical (Sepulveda et al., 1975). Grippon et
al. (1977) showed the first structural difference at residue
12 (pepsin numbering). Isozyme pi 5 has a lie residue at
position 12 while pi 6 has a Val. Delaney et al. (1987)
solved the complete sequence of isozyme pi 6 by a combination
of amino acid sequencing of HPLC-purified CNBr cleavage
fragments by Edman degradation (154 residues/325 directly
sequenced) and the DNA sequencing of a positive cDNA clone,
33E2. Subsequent work by Takahashi's group established the

16
complete sequence of both pi 5 and 6 by the 100% chemical
sequencing of HPLC-purified trypsin and Staphylococcus aureus
V8 protease generated peptide fragments (Takahashi, 1987,
1988). Rhizopuspepsin pi 5 and pi 6 differ only at eight
positions in the entire 325 amino acid sequence. All
substitutions are semi-conservative and only the Val/Ile 12
residue is in the active site.
The work by Chen et al. (1991) has lead to the
production of the zymogen of rhizopuspepsin,
rhizopuspepsinogen, in several different expression systems.
The zymogen form is expressed to facilitate correct folding.
Activation at low pH converts the zymogen to the active form
in combination of inter and intramolecular processes yielding
enzyme for kinetic studies. Kinetic studies comparing
various mutant rhizopuspepsin enzymes to the native enzyme
will help resolve the role hydrogen-bonding, electrostatic,
and hydrophobic interactions play in the creation of
specificity in the active site subsites Si, S2 and S3 of
rhizopuspepsin as well as give clues for understanding these
interactions in the other aspartic proteinases.
Specificity Differences Studied bv Site-directed Mutagenesis
Despite the high degree of primary sequence and
structural homology within the aspartic proteinases in and
around the active site (Table 1-2), differences are observed
in substrate specificities. The primary specificity of

17
aspartic proteinases is for cleavage between large
hydrophobic-hydrophobic junctions, such as, Phe-Phe (Fruton,
1970, 1976). Secondary interactions have been shown to cause
large increases in catalytic efficiency, kcat without
corresponding changes in Km (Balbaa et al., 1993). These
interactions, though sometimes far from the site of cleavage,
as previously discussed have been implicated in dictating
highly specific preferences for substrates. Cathepsin D, for
example, is unable to efficiently cleave substrates with
basic residues in the P2 position (Scarborough et al., 1993).
This effect is thought to be generated by the presence of a
methionine residue at position 287 (pepsin numbering,
Scarborough et al., 1994). Porcine pepsin exhibits a similar
aversion for positive residues, but in the S3 subsite. This
effect was demonstrated, through the use of mutagenesis and
pH dependence kinetic studies, to be mediated by Glul3 (Pohl
& Dunn, 1988; Rao-Naik, unpublished results).
Rhizopuspepsin and other fungal enzymes are different
from these examples and other mammalian enzymes in that they
are able to cleave a wide range of substrates with similar
efficacy and 2-10 fold higher kcat/Km values (Dunn et al.,
1986: Lowther et al., 1991). Even though rhizopuspepsin has
broad specificity, it does possess an additional
characteristic that distinguishes itself from all mammalian
aspartic proteinases.

18
Table 1-2. Primary sequence comparison of
rhizopuspepsin to several aspartic proteinases
Enzyme
% identity
% similarity
porcine pepsin
39.2
62.7
human pepsin
40.5
62.3
human cathepsin E
37.2
59.7
human cathepsin D
33.1
58.8
human renin
27.5
51.6
Candida albicans
aspartic proteinase
28.1
49.4

19
Rhizopuspepsin, as well as the majority of all other fungal
enzymes, has the ability to cleave substrates (Hofmann et
al., 1984; Balbaa et al., 1993) and to bind inhibitors with
lysine in the Pi position (Salituro et al., 1987).
Structural and sequence comparisons suggest two possible
residues in the active site that may account for this unique
capacity. Table 1-3 shows an alignment of sequences from
several aspartic proteinases in the active site and flap
regions. Besides the two catalytic aspartic acids, Asp32 and
Asp215, rhizopuspepsin has additional aspartic acid residues
at positions 30, 37, and 77. Asp30 is situated at the
boundary between the S3 and Si subsites and has the potential
to interact with positively charged residues in the P3 and
the Pi positions of the substrate (Figure 1-3). Asp37 is
located in the S2' subsite and probably does not directly
affect primary specificity. Asp77, which is also present in
the Candida enzymes, is positioned at the end of the flap and
points down into the active site cavity making potential
interactions with the P3, P2 and Pi positions of the
substrate.
A precedent exists for the importance of Asp77 in the
primary specificity of fungal enzymes from a complex of
penicillopepsin with a pepstatin derivative containing lysine
in Pi (James et al., 1984). In this structure the Asp77 side
chain and the enzyme backbone NH hydrogen bond in a highly
conserved manner to the P2 NH and the P2 carbonyl of the
inhibitor backbone, respectively.

20
Table 1-3. Partial sequence alignment of
several aspartic proteinasesa
HPEP
28TWFDTGSSN37
7 4-pygtG7 8
PPEP
TVIFDTGSSN
TYGTG
CATE
TVIFDTGSSN
QYGTG
CATD
TWFDTGSSN
HYGSG
HREN
KWFDTGSSN
RYSTG
RCAP
NLDFDTGSSD
SYGDG
CAAP
NVIVDTGSSD
GYGDG
CTAP
TWIDTGSSD
EYGDL
CPAP
TVIIDTGSSD
RYGDG
aThis alignment was obtained using the PILEUP
program, a module in the GCG Sequence
Analysis Software Package (Devereux et al.,
1984) The sequences are HPEP = human pepsin
(Sogawa et al., 1983), PPEP = porcine pepsin
(Chen et al., 1975), CATE = human cathepsin E
(Azuma et al., 1992), CATD = human cathepsin
D (Faust et al., 1985), HREN = human renin
(Hobart et al., 1984), RCAP = Rhizopus
chinesis aspartic proteinase (Chen et al.,
1991) CAAP = Candida albicans aspartic
proteinase (Hube et al., 1991), CTAP =
Candida tropicalis asapartic proteinase
(Togni et al., 1991), CPAP = Candida
parapsilosis (de Viragh et al., 1993).

21
Figure 1-3. Closeup view of the active site of
rhizopuspepsin highlighting the catalytic aspartic acid
residues, Asp32 and 215, and Asp30 and Asp77.

22
In contrast to the mammalian and fungal enzymes that contain
either serine or threonine at position 77, the Asp77 residue
was shown to be able to make an additional contact through
its side chain by hydrogen bonding to the e-amino nitrogen of
the lysine residue. Hydrogen bonds are also seen between the
lysine residue and Ser79.
This study has focused on the contributions of Asp30 and
Asp77 to fungal specificity through the use of site-directed
mutagenesis. These residues were changed in the
rhizopuspepsinogen gene to those present in porcine pepsin,
Ile30 and Thr77. The proteins were overexpressed in E. coli,
refolded from inclusion bodies, activated and purified for
structural and kinetic comparisons. A series of
systematically substituted substrates with lysine in either
Pi, P2 or P3 was assayed and analyzed by double mutant cycles
(Carter et al., 1984; Wells, 1990) in order to ascertain and
confirm the predominant interactions enabling substrate
catalysis.

CHAPTER 2
EXPERIMENTAL PROCEDURES
Introduction
This chapter outlines the materials and methods used to
characterize the unique specificity of rhizopuspepsin toward
substrates and inhibitors. This study has used a combination
of systematically substituted substrates and inhibitors and
site-directed mutagenesis to accomplish this task.
Materials
Restriction and modifying enzymes were purchased from
Promega, Life Technologies, Inc., New England Biolabs or
United States Biochemical Corp. Deoxyadenosine-5'-[a-35S]
thiotriphosphate, as its triethylammonium salt (Sp isomer,
1000 Ci/mmol), was purchased from Amersham Corp. The pet3a
expression vector (Studier et al., 1990), containing the
wild-type rhizopuspepsinogen gene (Chen et al., 1991), was
kindly provided by Jordan Tang at the Oklahoma Medical
Research Foundation, Oklahoma City, Oklahoma. The
aminomethylene, VF[CH2NH] (Spatola nomenclature, 1983), and
the hydroxyethylene, 'P[CH(OH)CH2] containing inhibitors were
a gift from Tomi Sawyer. Pepstatin was purchased from Sigma.
Acetyl-pepstatin was a gift from Kohei Oda, Kyoto Institute
23

24
of Technology, Japan. The native isozymes, pi 5 and pi 6,
were a gift from Kevin Parris and David Davies at the
Laboratory of Molecular Biology, National Institutes of
Health. The porcine pepsin used for comparison to the
rhizopuspepsins was from Sigma. The synthetic
oligonucleotides were synthesized by the University of
Florida, Interdisciplinary Center for Biotechnology Research
(ICBR) DNA Synthesis Core Facility using an Applied
Biosystems 394 DNA synthesizer. Peptide substrates were
synthesized by the ICBR Protein Chemistry Core Facility using
an Applied Biosystems 430A peptide synthesizer. The
oligonucleotides were used directly for mutagenesis and
sequencing reactions. All peptides were shown to be >95%
pure by reverse phase HPLC and capillary electrophoresis.
Stock solutions of the peptides and the inhibitor U85548E
were quantified by amino acid analysis on a Beckman System
6300 high performance amino acid analyzer following acid
hydrolysis. The N-terminal sequence analyses of the
rhizopuspepsins were performed on Applied Biosystems 470A and
473A protein sequencers. The activated enzymes were analyzed
by matrix-assisted laser desorption ionization time of flight
(MALDI-TOF) mass spectroscopic analysis on a Vestec (Houston,
TX) LaserTec Bench Top System. All other materials were of
the highest commercial grade.

25
Methods
All routine DNA manipulation procedures were performed
as outlined by Sambrook et al. (1989). Competent cells were
prepared by the calcium chloride method. All plasmid and PCR
products were isolated by using Magic Minipreps or Magic PCR
preps kits from Promega. These kits use a proprietary anion
exchange column to efficiently purify plasmid DNA.
Cloning and Mutagenesis
Mutations in the rhizopuspepsinogen gene were made by
using a modified version of the overlap extension method of
site-directed mutagenesis by the polymerase chain reaction
(PCR) (Ho et al., 1989) with the use of TAQ Polymerase
(United States Biochemical) or Vent polymerase (New England
Biolabs). This modification, using only one mutant primer,
has been discussed in detail (Scarborough & Dunn, 1994) and
is shown in Figure 2-1. This procedure uses four
oligonucleotides, one of which contains nucleotide changes
corresponding to the desired amino acid change. In the first
round of amplification two reactions were performed using 100
pmol of each primer, 10 ng plasmid template and 5 units of
polymerase. The first reaction generated the 5'-end of the
gene (A) by using a sense primer (1) containing an engineered
Ndel site (5'CAT ATG GCA GTT AAC GCT GCC CC3') and an
antisense primer (2) containing the mutations for residues 30
or 77 (Asp30 -Ile; 5'GA GGA ACC GGT ATC AAA GAT AAG GTT GAA

26
1
N.
PET3A-RPGN
Rhizopuspepsinogen
2
\ jm
4
primers 1 and 2
or
primers 3 and 4
Denature and Anneal
c
D
3
c; "4
EvvW-3 Ml "\0
primers 1 and 4
mm
BamH I

Nde I
E

Mutation
Figure 2-1. PCR mutagenesis procedure. A, 5'-end of
rhizopuspepsinogen; B, 3'-end; C, hybrid capable of
extension; D, hybrid not capable of extension; E, full length
rhizpuspepsinogen gene containing engineered mutations and
Nde I and BamH I restriction sites.

27
C3'; Asp77 -Thr; 5'GAT ACC GCT AGC AGA AGA GCC AGT ACC ATA
AG3'). The underlined bases indicate the engineered
restriction sites or the differences from wild-type. The
second reaction generated the 3'-end of the gene (B) by using
a sense primer (3, Asp30 Ile; 5'TTT GAT ACC GGT TCC TCC GAT
TTA TG3' ; Asp77 -4Thr; 5'TCT TCT GCT AGC GGT ATC TTG GC3' )
capable of annealing to the 3' end of the PCR product above
(A) and an antisense primer (4) containing an engineered
BamHI site (5'GGA TTC TTA TTG AGC GAC AGG AGC G3'). The
cycling conditions for the first round of PCR were as
follows: (1) 3 cycles; 96C for 40 sec, 50C for 40 sec, 72C
for 2 minutes, (2) 25 cycles; 94C for 40 sec, 50C for
40 sec, 72C for 2 minutes, (3) 72C for 7 minutes. These
PCR products (A and B) were purified on Seaplaque GTG or
NuSieve GTG low melting agarose gels (FMC Bioproducts). In
order to generate the full length gene, small amounts of each
band were mixed together and heated to 100C for 5 minutes,
and then placed on ice. This step is crucial for insuring
that the fragments are completely dissociated so they can
form hybrid templates in the next PCR reaction (C and D).
After the addition of polymerase and more of the outer
primers (1 and 4, 80 pmol each), the second round of PCR was
performed as follows: (1) 25 cycles; 94C for 40 sec, 55C
for 1 minute, 72C for 2 minutes, (2) 72C for 7 minutes.
The double mutant was generated by repeating this procedure
using the Asp30lle mutant rhizopuspepsinogen gene as the
starting template. The resulting products (E) were ligated

28
either directly into the TA cloning kit vector pCR
(Invitrogen) or after restriction digest with Ndel and BamHI
and gel purification into the pGEM vector (Promega) for DNA
sequence analysis. In order to confirm the presence of the
desired mutation and to ensure that no spurious mutations
occurred during the polymerization process, the entire 1136
bp coding region was dideoxy-sequenced according to the
Sequenase 2.0 Kit protocol using 5 (Ig of plasmid template
(United States Biochemical) and deoxyadenosine-5' [cc-
35S]thiotriphosphate with the insertion of one extra step.
Before the addition of the stop solution, more reaction
buffer, dNTPs, and terminal deoxynucleotidyl transferase
(TdT) were added in order to extend prematurely terminated
products resulting from high GC content and secondary
structure (Kho & Zarbl, 1992). The mutant genes were
transferred to a modified version of the pET3a expression
system vector. The pET3a vector was modified by removing a
375 bp fragment between the BamHI restriction site and
ampicillin resistance gene in order to generate a vector
which does not contain an EcoRV site. This was performed by
digesting the vector with EcoRl and EcoRV and by subsequently
treating with Klenow polymerase and blunt-end ligation. The
resulting vector is shown in Figure 2-2. The use of this new
vector allowed the efficient screening of recombinant clones
because of the unique EcoRV restriction site within the
rhizopuspepsinogen gene.

29
Figure 2-2. pET3aE expression vector containing
rhizopuspepsinogen. AMP, ampicillin resistance
gene; ori, origin or replication; SD, Shine-
Dalgarno sequence.

30
Expression
The native and mutant enzymes were expressed and
purified from BL21(DE3) E. coli cells as reported with minor
changes (Chen, Koelsch et al., 1991). A 1:50 dilution of an
overnight culture grown in M9 media (lO^lg/mL thiamine, 0.5%
casamino acids, 0.2% glucose) containing 50 mg/L ampicillin
was made into LB media containing the same amount of
ampicillin and grown to an OD600 of 0.5. At that time IPTG
was added to give a final concentration of 0.5 mM. The cells
were pelleted at 3,500 x g for 10 minutes and resuspended in
4.2 mis of 50 mM Tris-HCl pH 7.4, 150 mM NaCl and 1 mM MgCl2
(buffer A) per gram of cells. Following the addition of 80
Kunitz units of DNase (Sigma) per ml of suspension, the cells
were lysed by two passes through a French Press cell. The
resulting slurry was carefully layered over a 27% sucrose
cushion (density = 1.1) and centrifuged at 12,000 x g in
order to isolate the inclusion bodies which sediment through
the sucrose solution (Taylor et al., 1986). The inclusion
bodies were washed by resuspension in buffer A containing 1%
Triton X-100 and pelleted through sucrose a second time. The
resulting pellet was stored at -20C until refolding.
Refolding
In order to regain enzymatic activity, the wild-type and
mutant recombinant proteins were refolded by a modified
procedure for the refolding of prochymosin involving

31
denaturation, reduction and dialysis (Suzuki et al., 1989).
The purified inclusion bodies were dissolved in freshly
deionized 8 M urea, 50 mM CAPS pH 10.5, 1 mM EDTA, 1 mM
glycine, 500 mM NaCl and 300 mM |3-mercaptoethanol to a final
concentration of approximately 1 mg(wet)/ml. After stirring
at room temperature for one hour, the solution was
centrifuged at 24,000 x g for 30 minutes to remove
undissolved material. The supernatant was dialyzed for one
hour at room temperature against five times the original
volume in SpectraPor 1 (MWCO 6-8 kDa) membranes and 50 mM
Tris-HCl pH 11.0 buffer. Following a buffer change and
dialysis at room temperature for another hour, the dialysis
buffer was changed to 50 mM Tris-HCl pH 7.5 and dialysis
continued overnight at 4C. The next morning the buffer was
changed to 50 mM MOPS pH 7.0 and dialyzed for at least 6 more
hours at 4C. The resulting solution was centrifuged at
24,000 x g for 30 minutes to remove precipitates and
concentrated using a Minitan Ultrafiltration system outfitted
with low protein binding, PLTK, 10,000 MWCO membrane plates
(Millipore) and an Amicon pressurized cell with YM10
membranes (10,000 MWCO).
Size-exclusion Chromatography
The zymogen was further processed by centrifugation at
45,000 x g for thirty minutes before loading onto a 2.5 cm x
90 cm Sephacryl S300 gel filtration column equilibrated with

32
50 mM MOPS pH 7.0 containing 300 mM NaCl. The zymogen was
eluted at a flow rate of 25 ml/hr and the fractions showing
the highest purity were concentrated and buffer exchanged
with 10 mM MOPS pH 7.0.
Activation and Ion-Exchange Chromatography
In the activation of the native and Asp30lle mutant
proteins (0.5 mg/ml) for kinetic analysis, citric acid pH 2.0
was added to give a final concentration 0.1 M. The resulting
solution was held for fifteen minutes at room temperature.
The Asp77Thr and the double mutant Asp30lle/Asp77Thr zymogens
were activated for twenty-four hours in 0.1 M sodium formate,
37C at pH 3.0 and 3.5, respectively. After filtering
through a 0.2 Jim Millipore microcentrifuge unit, each enzyme
was directly injected onto a Pharmacia Mono S column
equilibrated with 50 mM sodium formate pH 3.0. The enzyme
was eluted by running a 25 minute gradient to 25% 50 mM
sodium formate pH 3.0 containing 1 M NaCl at a flow rate of
1 ml/min. Enzyme aliquots were quickly frozen and stored at
-20C.
Structural characterisation
N-terminal sequencing. N-terminal sequence analysis of
the activated rhizopuspepsins was performed to determine the
extent of processing during self-activation. The proteins
were electroblotted at 90 volts for 2 hours or 20 volts

33
overnight from 12% Tris-Tricine SDS-PAGE gels (Schgger & von
Jagow, 1987) to PVDF Immobilon P transfer membranes
(Millipore) in 10 mM MES pH 6.0 containing 20% methanol. The
excised bands were analyzed by the Protein Chemistry Core
Facility.
Mass spectrometry. The activated enzymes were analyzed
by matrix-assisted laser desorption ionization time of flight
(MALDI-TOF) mass spectroscopic analysis on a Vestec (Houston,
TX) LaserTec Bench Top System. One to ten pmol of each
sample or standard was mixed 1:1 with fresh 0.05% TFA, 40%
acetonitrile, saturated sinnapinic acid. One [il of this
mixture was applied to a stainless-steel sample pin and
allowed to air dry. The mass spectrum was obtained from the
average of at least 50 laser shots (337 nm nitrogen laser, 3
ns pulse width). Time to mass/charge calibration was
performed from a calibration curve using bovine carbonic
anhydrase II (Sigma, 28,980 daltons) immediately prior to the
rhizopuspepsin samples.
Isoelectric focusing. The isoelectric points of the
proteins were determined by running precast PhastGel IEF gels
ranging from pH 3 to 9 on the Pharmacia PhastSystem. The
samples, including standards, were diluted in sample buffer
(0.01% pyronin Y, 10% glycerol, 62.5 mM TRIS pH 6.8) prior to
loading. The gels were run according to the PhastSystem
profile No. 100 and subsequently stained by the Coomassie
blue profile No. 200.

34
Circular dichroism. Circular dichroism (CD) spectra
were determined on one day at room temperature on a Jasco
J-500C spectropolarimeter equipped with a IF-500 II computer
interface with a 0.1 cm pathlengh cell (Hellma). The
polarimeter was standardized with D(+)-camphorsulfonic acid
(Chen & Yang, 1977). The samples were diluted into buffer to
a final concentration around 0.5 mg/ml in 0.1 M sodium
formate pH 3.0. Just prior to loading into the CD cell, the
samples were filtered through a 0.2 (im microcentrifuge filter
(Rainin) and quantitated by reading their absorbance at 280
nm. The ellipticity values were converted to the molar
ellipticity, [0], using the conversion factor E}^ = 12.6
(Fukumoto et al., 1967) and a molecular weight of 35 kDa.
The data points were fit with the smoothing algorithm of the
KaleidaGraph program (version 3.0.2 Synergy software, PCS
Inc.).
Fluorescence spectroscopy. The guanidium hydrochloride
denaturation curves of the rhizopuspepsins were determined
using excitation and emission wavelengths of 280 and 350 nm,
respectively, on an SLM Aminco 4800C spectrofluorometer.
Ultrapure 8 M guanidinium hydrochloride in water was
purchased from Pierce. The proteins were diluted in
duplicate into denaturant concentrations ranging from 0 to
6 M with a final buffer concentration of 0.1 M sodium formate
pH 3.0 and an enzyme concentration around 100 nM. The
protein/denaturant solutions were equilibrated at 25C for at

35
least one hour prior to spectroscopic measurements in 1 cm
cuvettes.
Denaturation curve analysis. The denaturation curves
were analyzed assuming a two-state model where only the
native and the denatured states are populated. The
fluorescence values for the native and unfolded states, FN and
Fy can be used to determine the equilibrium constant for
unfolding, KU_F, and the free energy of unfolding, AGV_F, at
different denaturant concentrations by using equation 1.
*u-f = (F* ~ F) / (F Fy) = exp(-AGv_v / RT) (1)
F is the observed fluorescence, R is the gas constant (1.987
cal mol-1 K_1) and T is the absolute temperature. The free
energy of unfolding of proteins has been shown to be linearly
dependent on the denaturant concentration as expressed by
equation 2 (Pace, 1986).
AGy_F = AGy.2 ^[denaturant] (2)
AG_f and AGy2 are the free energy of unfolding at
denaturant concentration, D, and in water, respectively. m
is the slope of the transition and is thought to related to
the difference in the degree of accessible surface area
between the native and unfolded states (Schellman, 1978).
Two methods have been used to calculate the transition point,

36
[GdnHCl]50%, AGu'p and m. The method of Pace (1986) employs
equation 2 by plotting AG_F within the transition region
( 1.5 kcal mol-1) versus [denaturant] and linearly
extrapolating back, usually quite a long distance, to zero
denaturant to obtain AG^p Because small errors in m can
lead to large errors in the calculation of Aand
[GdnHCl]50% Fersht and coworkers have used a method,
represented by equation 3, which uses all the observed
fluorescence data, F, to directly determine [GdnHCl]50%
typically within 0.02 M guanidinium hydrochloride (Jackson
et al., 1993).
F = {(aN + /JN[GdnHCl]) + {av + Pu [GdnHCl]) x
exp[j([GdnHCl] [GdnHCl]50J / RT]) / (3)
{1 + exp[m([GdnHCl] [GdnHCl]^) / /?!]}
This equation combines equations 1 and 2 and assumes
that Fn and Fv are linearly dependent on the denaturant
concentration. orN and are the intercepts and /JN and
are the slopes of the baselines at denaturant concentrations
before and after the transition region. These parameters, as
well as, [GdnHCl]^ and m were allowed to be variables in the
KaleidaGraph non-linear regression analysis program. The
values of m and [GdnHCl]^ were obtained for the
rhizopuspepsins with their standard errors. AG.2 values

37
were calculated from equation 2 where at [GdnHCl]^, the
transition point, AG.2 = ^[GdnHCl]^ .
Kinetic Analysis
Kinetic assays using chromoaenic substrates. Substrate
hydrolysis, where cleavage occurs between Phe*Nph, Nle*Nph or
Lys*Nph (Nph = p-nitrophenylalanine, Nle = norleucine and =
site of cleavage), was monitored by the decrease in the
average absorbance from 284-324 nm using a Hewlett Packard
8452A diode array spectrophotometer (Scarborough et al.,
1993). The Km and Vmax values were determined from the
initial rates of at least six different peptide substrate
concentrations using Marquardt analysis (Marquardt, 1963) and
the equation v = VmajjSl/fKm + [S] ) The observed rates in
AU s_1 were converted to M s-1 by dividing by the total change
in absorbance for complete hydrolysis of a known
concentration of each substrate used. The amount of active
enzyme was determined by fitting the curve generated by the
competitive titration at one substrate concentration and 2%
DMSO with the inhibitor Val-Ser-Gln-Asn-LeuvP[CH(0H)CH2]Val-
Ile-Val (U85548E; Sawyer et al., 1992) with the Henderson
equation for tight binding inhibitors using the Enzfitter
program (Henderson, 1972; Leatherbarrow, 1987) The standard
deviations of the kcat and kcat/Km values were propagated
using equations derived by standard procedures for non-

38
independent or correlated errors as outlined by Meyer (1975).
In those cases where the Km values were 1 mM, the kcat/Km
values were determined by fitting the initial rates of at
least six substrate concentrations ranging from 25-250 |iM to
the equation v = (kcat/Km) [E] o [S] o with the Enzfitter program
and the assumption that [S] Km. The kcat/Km values for the
cleavage of the Pi lysine-substituted peptides by the
Asp77Thr and the Asp30lle/Asp77Thr mutants were calculated
with the same equation on Enzfitter as above with the initial
rates determined by capillary electrophoresis.
Kinetic assays using competitive inhibitors. The
inhibition constant, Ki, was determined by monitoring the
competitive inhibition of the hydrolysis of the peptide Lys-
Ala-Ala-Lys-Phe*Nph-Arg-Leu (Km = 20 |JM) where cleavage
occurs strictly between Phe and Nph. All reactions were
performed at 37C in 0.1 M sodium formate buffer, pH 3.5 and
a final concentration of 4% DMSO. The initial rates of six
different substrate concentrations were measured following
preincubation of the enzyme without inhibitor for five
minutes. Additional curves were obtained, after
preincubation with two or more inhibitor concentrations, from
the initial rates of at least three different substrate
concentrations. The Ki value was determined from the family
of curves by the equation, v = Vmax[S]/[Km(1 + [I]/Ki) +
[S]]. If the Ki value determined by this method was one
nanomolar or lower, a competitive titration was performed as
described for the enzyme titration above.

39
Product, analysis. The fidelity of the cleavage sites
was verified by HPLC and capillary electrophoresis (CE). All
substrates were incubated with enzyme at 37C overnight. The
cleavage products of the substrates based on the parent
peptide Lys-Pro-Ala-Lys-Phe*Nph-Arg-Leu, discussed in Chapter
6, were analyzed by reverse phase HPLC using a Waters C-18
Radial-pack column with a gradient from 10 to 80%
acetonitrile in water containing 0.1% TFA in 45 minutes. The
peaks were collected and hydrolyzed by the addition of 6 N
HCl. The composition was determined by amino acid analysis
by the Protein Chemistry Core Facility.
The cleavage products of the peptides used for the
analysis of mutants of rhizopuspepsin (Chapter 7) were
determined on a BioRad BioFocus 3000 capillary
electrophoresis system. The samples at 4C were pressure
injected for 20 psi*sec onto a BioRad 24 cm x 25 Jim cartridge
maintained at 15C. This type of injection procedure ensures
reproducible loading of the sample onto the capillary. The
capillary was equilibrated with 0.5 M sodium phosphate pH
2.5, electrophoresed at a constant 8 kv in the +-
direction and monitored at 200 nm. All the peptides studied,
if cleaved properly, will have the same C-terminal product,
Nph-Arg-Ala. This product was purified on a Rainin HPLC
system using a 4.6 mm X 25 cm Dynamax-300 C-18 column with a
gradient from 0 to 10% B acetonitrile in water containing
0.1% TFA in 32 minutes at a flow rate of 1.1 ml/min. The
composition of the fraction was confirmed by amino acid

40
analysis. This fragment was used as a retention time
standard for the CE runs to validate the correct site of
cleavage for each substrate and all forms of rhizopuspepsin
at pH 3.5 and 5.0.
Capillary electrophoresis kinetic assay. The initial
rates of cleavage of the peptide containing lysine in Pi by
AspWThr and Asp30lle/Asp77Thr were determined by incubating
the enzyme (900 nM) with either 50, 100 or 150 (1M substrate
for a period of twenty-four hours at 37C. A 20 |i.l aliquot
was taken at 0, 1, 3, 5, 7, 12, 15, and 24 hours and mixed
with 1.5 H-l of U85548E to give a final concentration of
inhibitor of 5 |JM (five-fold molar excess) and stored at
-20C until the electropherograms were run. The initial
linear slopes of the change in intact substrate peak area
versus time were converted to M s-1 by dividing by the slope
(peak area/[S]) of a standard curve generated from 30 to 1000
|iM of the substrate. The same injection and electropherogram
run parameters were used in the product analysis.
Analysis of Transition State Effects
In transition state theory, the energy difference
between the free enzyme and substrate and the transition
state, AGf, is related to the binding energy released upon
binding of the substrate, AGs, and the activation energy,
AG*, of the chemical steps responsible for bringing the

41
enzyme-substrate complex from the ground state to the
transition state (Fersht, 1985) :
AG* = AGS + AG* = RTIn(kBTIK)-RTln^ /KJ (4)
With the assumption that the energies associated with the
bond breaking and making steps, AG*, are not significantly
affected upon mutation of the enzyme or changes in the assay
pH, the discrimination of the wild type and mutant enzymes
for different substrates can be evaluated by their relative
binding to the transition state:
. ^, k, / K (mutant, mutant 2 or pH 5.0)
AAG* = -RT\n (5)
kat / Km (wild type, mutant 1 or pH 3.5)
Molecular Graphics
Molecular graphic representations of the x-ray crystal
structures were generated using the Insight II (version 2.3)
from Biosym Technologies, Inc. (San Diego) on a Silicon
Graphics Indigo system at the University of Florida Center
for Structural Biology. Root mean square (RMS) superposition
of the Coe (alpha carbon) backbones of rhizopuspepsin (3APR,
Suguna et al., 1987) and penicillopepsin (1APT, James et al.,
1985) inhibitor complexes were performed by selecting the
active site residues 27-37 and 210-220 for the SUPERIMPOSE
command in the TRANSFORM menu.

CHAPTER 3
EXPRESSION, REFOLDING, PURIFICATION, AND ACTIVATION OF
RECOMBINANT RHIZOPUSPEPSINS
Introduction
Overexpression of proteins in heterologous systems has
become an indispensable method in the generation of large
quantities of wild-type and mutant proteins for structural
and biochemical analysis. Many different systems have been
used to obtain the protein of interest; for example, E. coli,
yeast and SF9 insect cells. The decision of which system to
use primarily depends on the yields required and whether or
not glycosylation of the protein is desired. Expression in
E. coli usually gives the highest yields but the protein is
frequently deposited in an insoluble form known as inclusion
bodies. These deposits are readily purified and are usually
greater than 95% pure protein. In order to regain biological
activity the protein must be denatured and refolded. Upon
optimization of the refolding conditions, peculiar to each
protein and its mutants, a sufficient quantity of protein for
analysis may be obtained.
This chapter outlines the used of E. coli to produce
sufficient quantities of wild-type (WT-REC) and mutant forms
of rhizopuspepsin for structural and kinetic analysis.
Expression of the zymogen form of rhizopuspepsin,
42

43
rhizopuspepsinogen, resulted in the formation of inclusion
bodies. The inclusion bodies were isolated, denatured, and
refolded by dialysis. The inactive, zymogen form of
rhizopuspepsin were converted to the active, mature form by
lowering the pH of the solution. The resulting proteins were
purified by ion exchange chromatography and analyzed by
isoelectric focusing and N-terminal sequencing.
Results
Mutagenesis
Mutants of rhizopuspepsin were generated by overlap
extension PCR. All reactions yielded the desired size
fragments. The entire coding region of each protein was DNA
sequenced to determine the presence of the desired base
changes and if other spurious mutations had occurred. In
order to obtain clear sequencing ladders, an extra step was
added to the Sequenase procedure. Prior to adding stop
solution, the DNA was incubated with terminal
deoxynucleotidyl transferase in order to extend premature
stops due to high GC content and secondary structure. No
additional mutations were seen.
Expression and Refolding
In order to obtain sufficient quantities of enzyme for
structural and kinetic analyses, the rhizopuspepsinogens were

44
overexpressed in E. coli using the pET expression system
(Studier et al., 1990). Upon addition of IPTG to the
bacterial cultures, the proteins were expressed at high
levels in the form of inclusion bodies. Figure 3-1 shows an
expression time course experiment for wild-type
rhizopuspepsinogen. By the end of three hours,
rhizopuspepsinogen was one of the predominant proteins. The
inclusion bodies were purified from the cell lysate by
centrifugation through a 27% sucrose cushion. The resulting
pellet was washed with buffer containing triton X-100 to
remove any remaining cellular debris. Typical yields ranged
from 0.5 to 1 g (wet weight, 5 to 18 % of the total cell
pellet) for a 4 L expression. The purified inclusion bodies
were judged to be >95 % pure zymogen with a molecular weight
of 43kDa by SDS-PAGE (Figure 3-2).
The inactive zymogen inclusion bodies, containing
approximately 100 mg of protein, were refolded from 8 M urea.
Two refolding methods were tried in order to maximize yields
for subsequent activation to the mature enzyme. Refolding by
the rapid dilution procedure of Chen et al. (1991) resulted
in primarily polymeric material that was difficult to
completely activate into an active, monomeric state (data not
shown). Even though there was some precipitation, the
largest yields of monomeric zymogen capable of activation
were obtained when the proteins were refolded by dialysis.
Figure 3-3 shows the elution profile from gel filtration
chromatography of the refolded rhizopuspepsinogen.

Figure 3-1. SDS-PAGE analysis of the expression of wild-type
rhizopuspepsinogen (RPGN) in E. coli upon the additon of IPTG
from 0 to 3 hours. The RPGN migrates at 43 kDa in comparison
to the molecular weight markers, M.

UJ -U Os VO fr
O W vj ^
i i i p
2
OJ
o
o\
K>
U)

47
Figure 3-2. SDS-PAGE analysis of wild-type
rhizopuspepsinogen at different stages of purification. Lane
1, E. coli whole cell lysate; lane 2, 27% sucrose pellet;
lane 3, Triton X-100 wash 1 supernatant; lane 4, Triton X-100
wash 2 supernatant; lane 5, concentrated, refolded zymogen
prior to loading onto a S-300 gel filtration column.

Figure 3-3. Gel filtration elution profile of refolded
Asp30lle rhizopuspepsinogen. Peak 1, polymeric material at
the void volume; peak 2, rhizopuspepsinogen; peak 3, low
molecular weight contaminants.

49
6 11 16
hours

50
Peak 2 was shown to be rhizopuspepsinogen by SDS-PAGE
analysis (Figure 3-4). Peaks 1, which elutes at the void
volume, was shown, upon silver staining of the gel, to be
primarily polymeric rhizopuspepsinogen. Peak 3 was shown not
to contain rhizopuspepsinogen or active rhizopuspepsin but to
consist of low molecular weight proteins by the same method.
Yields at this stage of purification ranged from 15 to 35 mg
for a 4 L preparation.
Activation and Purification
Activation of zymogens was accomplished by lowering the
pH of the solution. Activation has been shown to occur by
intermolecular and intramolecular interactions (Chen et al.,
1991). The optimal conditions for activating the zymogens
varied. Figure 3-5 shows the time course of activation of
Asp30lle at a protein concentration of 0.5 mg/ml. The
protein was efficiently converted at both pHs and room
temperature to a molecular weight of 35 kDa. Sequence
analysis of the WT-REC and Asp30lle proteins, activated at pH
2.0, confirmed the N-terminus of these proteins to be Thr-
Ser-Thr-Gly-Gly-Ile-Val-Pro-Asp-. This sequence represents
an extension of naturally occurring rhizopuspepsin by 9 amino
acids. Figure 3-6 shows the results of an activation
experiment of WT-REC at pH 4.0 from 1 to 6 hours in an
attempt to remove this extension by intermolecular
processing.

Figure 3-4. SDS-PAGE analysis of the fractions from the
purification of Asp30lle by gel filtration in Figure 3-3.
Lane 1, peak 1; lanes 2-10, peak 2; lane 11, peak 3 .

52
kDa
M
94 -
j
67 -

43 -

30 -

20 -

14 -

2 34 5 6789 10 11 M

Figure 3-5. SDS-PAGE analysis of the time course of
activation of wild-type rhizopuspepsinogen upon lowering the
pH at room temperature. A, pH 2.0; B, pH 3.5.

54
M O 15 30 1 2 3
kDa
67 _
A
43 -
30 -

Figure 3-6. SDS-PAGE analysis of the activation of wild-type
recombinant rhizopuspepsinogen at pH 4.0. Incubations were
performed from 0 to 6 hours with protein concentrations of
0.1, 0.5, 1.0, and 1.5 mg/ml. C, crude isozyme pi 6.

56
0.1 0.5 1.0 1.5
kDa M C 0 1 2 3 6 1 2 3 6 1 2 3 6 1 2 3 6 M

57
Tang and coworkers have shown that intermolecular activation
of rhizopuspepsinogen and porcine pepsinogen occurs
principally at pHs above 3.0 (Al-Janabi et al., 1972; Lin et
al., 1989; Chen, et al., 1991). Protein concentrations were
varied from 0.1 to 1.5 mg/ml. Activation to the intermediate
form of rhizopuspepsin was not observed until 6 hours of
incubation at 0.1 mg/ml. An increase in the protein
concentration above this level resulted in degradation of the
protein. The 9 amino acid extended form of the proteins were
used for all kinetic and structural comparisons. A
discussion on the potential effect of this extension is given
in Chapter 4.
The activation of the D77T and the D30I/D77T proteins at
pH 2, however, produced a mixed population of N-termini: the
9 amino acid extension and a 15 amino acid extension, Asn-
Lys-His-Lys-Ile-Asn-Thr-Ser-Thr-Gly-Gly-Ile-Val-Pro-Asp-.
These populations can readily be seen with IEF gel analysis.
Figure 3-7 shows the IEF gel of all the recombinant
rhizopuspepsins following activation at pH 2 prior to ion
exchange chromatography. The WT-REC and the Asp30lle
proteins have a pi of 5.8. The Asp77Thr and
Asp30lle/Asp77Thr mutants show bands at pHs 6.1 and 6.9. The
yield ratios from the sequencing analysis suggest that the pH
6.9 band corresponds to the 15 amino acid extension
activation intermediate. This is not suprising since this
extension contains three additional positively charged
residues making the protein more basic.

58
8.7-
8.5-
8.2-
7.4-
M
6.9-
2 3 4
3.5-
Figure 3-7. IEF analysis of the wild-type and mutant
rhizopuspepsinogens activated at pH 2.0. Samples analyzed
prior to ion-exchange chromatography. Lane 1, wild-type;
lane 2, Asp30lle; lane 3, Asp77Thr; lane 4,
Asp30lle/Asp77Thr.

59
Interestingly, only upon mutation of Asp77 to Thr is there a
notable effect on the overall pi value of the 9 amino acid
extended form.
Increased yields of the 9 amino acid extension were
obtained for the D77T and the D30I/D77T proteins when the
activation was performed at 37C and pH 3.0 and 3.5,
respectively. The three additional positively charged
residues proved to be fortuitous in clearly separating the
two activation intermediates by ion-exchange (Figure 3-8).
The recombinant rhizopuspepsins at this final step in the
protocol were shown by IEF to be highly pure (Figure 3-9).
All structural and kinetic comparisons were made using the
Thr-Ser-Thr-form of the rhizopuspepsins. Yields at this
final step in the purification ranged from 1 to 5 mg for a
4 L expression. A summary of the yields during the
purification of the recombinant rhizopuspepsins is shown in
Table 3-1.
Discussion
The largest losses of rhizopuspepsinogen occurred during
the solubilization of the inclusion bodies and refolding.
Precipitation usually occurs during the refolding protocol.
The solubilization and refolding of rhizopuspepsinogen is
complicated by the presence of two disulfide bonds in its
tertiary structure. The inclusion bodies of prochymosin,
which contains three disulfide bonds, have been shown to

Figure 3-8. Elution profile of Asp77Thr from the Mono S
column after activation at pH 3.0. The desired 9 amino acid
extended form elutes near 12% B.

61
O
minutes
38

62
Figure 3-9. IEF comparison of the activated, purified
recombinant rhizopuspepsins. Lane 1, WT-REC, lane 2,
Asp30lle; lane 3, AspWThr; lane 4, Asp30lle/Asp77Thr.

Table 3-1. Representative yields during the purification of the recombinant
rhizopuspepsins
WT-REC
Asp30lle
Asp77Thr
Asp30lle/Asp79Thr
Inclusion bodies (wet)
1000
mg
429 mg
805 mg
813 mg
8M urea
97.2
mg
99.4 mg
33.3 mg
98.4 mg
Refold supernatant
25.8
mg
37.3 mg
19.5 mg
15.3 mg
Gel filtration
7.6
mg
10.0 mg
6.4 mg
6.3 mg
Ion exchange (Mono S)
2.0
mg
3.8 mg
1.7 mg
1.2 mg

64
contain mainly intermolecularly cross-linked protein
(Schoemaker et al., 1985). In order to reduce the cysteine
residues in the protein before refolding, high levels (10 to
1000 fold molar excess based on cysteines) of reducing agent,
BME or DTT, are required. Even if all the cysteines are
reduced, losses still occur for several different reasons:
pH of the refolding solution, speed at which the denaturant
and reducing agents are removed and the protein
concentration. These factors must be optimized in order to
obtain biologically active protein. Each protein has its own
characteristic conditions for refolding. Different
conditions may also have to used to refold mutant proteins.
The refolding and activation protocols for the WT-REC
and mutant forms of rhizopuspepsin have been optimized.
These procedures produced sufficient quantities of active
enzyme for structural and kinetic analysis. All enzymes used
exhibit homogeneous N-termini and show similar
electrophoretic properties after purification. Structural
analysis and kinetic comparisons of the recombinant
rhizopuspepsins to the naturally occurring isozymes are
discussed in Chapter 4. The inhibitor binding and substrate
specificity characteristics of WT-REC are discussed in
Chapters 5 and 6. Chapter 7 presents the analysis of the WT-
REC and mutant rhizopuspepsins toward substrates that contain
lysine in Pi.

CHAPTER 4
KINETIC AND STRUCTURAL AUTHENTICY OF RECOMBINANT
RHIZOPUSPEPSINS
Introduction
Molecular biology techniques have enabled the production
of large quantities of purified proteins that would other
wise be difficult to study due to their low abundance or to
the impractical nature of the source. These techniques also
aid in the dissection and understanding of biological
phenomena through site-directed mutagenesis. The resulting
proteins, however, must be shown to be analogous to the
naturally occurring enzymes, with respect to structure and
biological property being examined, if the results from
recombinant proteins are to extrapolated to what occurs in
the physiological environment.
The recombinant rhizopuspepsinogen gene used in this
study was constructed from the two naturally occuring
isozymes, pi 5 and pi 6, by the fusion of the pro region
through residue 12 of the pi 5 isozyme gene to the pi 6 gene
at residue 12. The expression of this chimeric gene results
in an enzyme after activation that is identical to the pi 6
isozyme except at position 12 where Val is replaced by a lie.
Ilel2 is located in the S3 subsite of the protein. All the
other residue differences seen between the two naturally
65

66
occurring isozymes are on the surface of the protein far from
the active site. In order to investigate a possible kinetic
difference of the WT-REC enzyme from the pi 5 and 6 isozymes,
a series of substrates with systematic substitutions in P3
were examined. This comparison was also performed to rule
out the possible effects of the 9 residue N-terminal
extension of WT-REC on catalysis. IEF gel analysis was also
performed.
The overall tertiary structure of the native isozyme
pi 6 was compared to the WT-REC and mutant rhizopuspepsins by
guanidinium hydrochloride denaturation experiments. The
recombinant enzyme structures were also examined by circular
dichroism. These studies were undertaken to investigate
whether or not the mutations introduced into rhizopuspepsin
caused large conformational changes in the enzymes that may
compromise the interpretation of kinetic experiments.
Results
WT-REC and Native Isozvmes of Rhizopuspepsin
IEF. A comparison of the pi values of the naturally
occurring isozymes of rhizopuspepsin to WT-REC was made using
IEF gels and protein standards (Figure 4-1). The two native
isozymes, pi 5 and pi 6, exhibit pi values of 5.1 and 6.2,
respectively. The WT-REC enzyme has a pi of 5.7. This gel
also illustrates the large difference in pi values seen upon

67
M 1 2 3 4
Figure 4-1. IEF comparison of wild-type rhizopuspepsinogen,
activated, purified WT-REC and the two naturally occurring
isozymes. Lane 1, rhizopuspepsinogen, lane 2, isozyme pi 5;
lane 3, isozyme pi 6; lane 4, WT-REC.

68
activation and removal of the pro region of the enzyme. The
wild-type recombinant rhizopuspepsinogen has a pi value of
7.4.
Kinetic analysis. A comparison of the substrate
specificity of the naturally occurring isozymes of
rhizopuspepsin to WT-REC was made using peptides with
substitutions in P3 (Table 4-1). Even though the enzymes
possess different eletrophoretic properties and N-termini,
the kinetic parameters determined for each substrate, within
experimental error, are directly comparable. The three
different forms of rhizopuspepsin are kinetically
indistinguishable from each other and exhibit similar
substrate specificity showing a preference for Arg and Leu
substitutions in P3.
Structural Comparisons
Mass spectrometry. The recombinant forms and the
isozyme pi 6 of rhizopuspepsin were analyzed by mass
spectrometry. The mass for each protein was determined to be
as follows: pi 6, 34,173; WT-REC, 34,627; D30I, 34,638;
D77T, 34,914; and D30I/D77T, 34,856. The sizes of the
recombinant proteins are consistent with the addition of 9
amino acids to the N-terminus of the native isozyme. The
values for the WT-REC and D30I proteins are slightly lower
than expected. This difference may be the result of
C-terminal processing.

69
Table 4-1. Kinetic comparison between the naturally
occurring isozymes and WT-REC rhizopuspepsin using the
substrate Lys-Pro-P3~Lys-Phe*Nph-Arg-Leu
Enzyme
P3 kcat
(s'1)
Km kcat/Km
(\m) (m-1s_1 )
X 10"6
pi 5
12

2
20

2
0.63

0.14
pi 6
Asp
15

2
21

1
0.69

0.09
WT-REC
13

1
20

2
0.65

0.09
pi 5
16

3
9

1
1.71

0.38
pi 6
Arg
17

2
16

2
1.07

0.17
WT-REC
13

2
9

1
1.39

0.25
pi 5
9

2
7

1
1.31

0.30
pi 6
Leu
12

1
8

1
1.53

0.22
WT-REC
10

1
7

1
1.37

0.21
pi 5
13

2
21

2
0.62

0.12
pi 6
Ser
12

1
19

1
0.67

0.08
WT-REC
9

1
16

2
0.60

0.08
Nph = p-nitrophenylalanine; WT-REC = wild-type recombinant
rhizopuspepsin.

70
Circular dichroism. Figure 4-2 shows the CD spectra of
the recombinant rhizopuspepsins in the far-UV region (200-250
nm). Even though there are slight wavelength shifts in the
spectra, the mutants as a whole are, within experimental
error, structurally similar to each other and the wild-type
enzyme. These shifts may be due to slight changes in the (X-
helix/(3-sheet ratios. The differences seen between the
proteins from 200 to 205 nm cannot be considered to be
significant because of the high degree of signal fluctuation
in this region on the instrument used.
Fluorescence spectroscopy. The denaturation of the
rhizopuspepsins was followed by the change in intrinsic
fluorescence at 350 nm using an excitation wavelength of 280
nm (Figure 4-3). Upon the addition of sufficient guanidinium
hydrochloride to cause unfolding, the fluorescence signal
shifted to longer wavelengths with a 75% decrease in
intensity. Figure 4-4 shows the normalized denaturation
curves for the recombinant rhizopuspepsins and the isozyme
pi 6. Analysis of the transition curves by the method of
Jackson (1993) is presented in Table 4-2. The pi 6 isozyme
and wild-type recombinant proteins exhibit unfolding
parameters which are indistinguishable from each other. The
decrease in [Gdn-HCl]^ of the two single mutants suggests a
slight decrease in stability from the wild-type enzymes, WT-
REC and isozyme pi 6. Further loss of stability is seen in
the double mutant.

[ 0 ] deg cm2 dmol"
71
Wavelength (NM)
Figure 4-2. CD spectra of the recombinant wild-type
and mutant forms of rhizopuspepsin.

Relative Fluorescence
72
NM
Figure 4-3. Fluorescence emission spectra for
folded (0 M GdnHCl) and unfolded (6 M GdnHCl)
wild-type recombinant rhizopuspepsin.

Fraction Unfolded
73
Figure 4-4. Guanidinium hydrochloride induced
unfolding of the naturally occurring isozyme
pi 6 and the recombinant forms of rhizopuspepsin
monitored by the change in intrinsic fluorescence
at 350 nm.

74
Table 4-2. Guanidinium hydrochloride denaturation parameters
of native and mutant forms of rhizopuspepsina
Enzyme
[Gdn-
-HCl]50%
(M)
m
(kcal mol
_1 M-1)
AGu.V
(kcal mol-1)
pi 6
3.54

0.02
4.8

0.5
17.0
1.8
WT-REC
3.50

0.02
4.6

0.5
16.0
1.7
Asp30lle
3.31

0.02
3.3

0.2
10.9
0.8
Asp77Thr
3.33

0.02
4.5

0.4
15.1
1.4
Asp30lle/Asp77Thr
3.00

0.01
3.7

0.2
11.0
0.7
Parameters derived from the denaturation curves presented in
Figure 4-4. Each enzyme was studied as outlined in Chapter 2
from 0 to 6 M guanidinium hydrochloride in 0.1 M sodium
formate pH 3.0. All denaturant concentrations were performed
in duplicate. Values determined by the method of Jackson et
al. (1993); curve fit of observed fluorescence, F, versus
[denaturant]. The errors represented are the standard errors
from the KaleidaGraph program.

75
Discussion
The zymogen forms of the native and mutant proteins were
efficiently expressed in E.coli and refolded from inclusion
bodies. Several lines of evidence exist to support the
conclusion that the recombinant rhizopuspepsins are
structurally and enzymatically similar to the native
isozymes. Maturation of rhizopuspepsinogen requires
catalytic activity. Activation has been to shown to occur
upon lowering of the environmental pH by intermolecular and
intramolecular processes similar to that of porcine
pepsinogen (Chen et al., 1991). Kinetic comparisons between
the two naturally occurring isozymes and the wild-type
recombinant enzyme have shown that the 9 amino acid N-
terminal extension and differences in pi values do not result
in significant deviations in catalytic activity. These
observations suggest that the slight differences seen in the
pi values of the mutants will not adversely affect their
kinetic analysis (Chapter 2, Figure 3-9). The degree of
similarity seen in the circular dichroism and denaturation
studies lends additional support to the conclusion that the
recombinant wild-type and mutant proteins are correctly
folded overall. Denaturation studies with the aspartic
proteinase zymogen prochymosin have also shown that the
recombinant protein is directly comparable to the native
enzymes (Sugrue et al., 1990).

76
Mutagenesis at positions 30 and 77 did cause slight
destabilization of the proteins in response to guanidinium
hydrochloride. Evidence supporting the idea that these
differences may be the product of local side-chain
reorganization comes from the x-ray crystallographic analyses
of mutants of chymosin, lysozyme, trypsin and many other
proteins (Strop et al.f 1990; reviewed by Shortle, 1992).
The Ca backbones of these structures exhibited little or no
deviation from the wild-type structures. The side chains,
however, usually did show some small positional movements.
Since denaturation studies have not been carried out on
mutants of the aspartic proteinases, trends seen in the
change of stability of the large library of mutants and their
crystal structures, reviewed by Shortle (1992), can be used
to rationalize the change in denaturation parameters seen for
the mutants of rhizopuspepsin. Mutations that have been
shown to cause significant changes in stability in comparison
to the wild-type protein can be grouped into several
categories: (1) insertion or deletion of an amino acid, (2)
addition or deletion of disulfide crosslinks, (3) changes
made near the ends of loops and a-helices, and (4) the
alteration of hydrophobic packing in the core of the protein
by the deletion of methylene equivalents. The sites targeted
for mutagenesis in this study do not fall within any of the
categories mentioned above. Position 30 and 77 were replaced
with the corresponding residues of porcine pepsin. These
residues are highly conserved among aspartic proteinases and

77
their substitution into the rhizopuspepsin structure is not
expected to change the structure of the mutants
significantly. Added support for structural similarity
between the different rhizopuspepsins extends from the
observation that the mutation of surface residues of a
protein generally does not affect its overall fold.
Asp30 points into the active site cavity between the S3
and Si binding pockets. The base of the active site cleft is
made of a large P-sheet composed of strands from both the
N- and C-terminal domains of the protein. Studies by Katz
and Kossiakoff (1990) have shown that P~sheets undergo less
distortion than a-helices and loops upon mutation. This is
thought to be the result of an increased ability of the
P-sheet to dissipate strain energy through slight changes in
<})-Â¥ angles. These observations suggest that mutation at
position 30 will not causes significant changes in stability
or structure of the protein.
Asp77 is located in the P-hairpin turn of the flap which
extends over the active site. The flap region is thought to
be quite flexible. Upon inhibitor binding, particularly in
the HIV proteinase (Swain et al., 1990), the flap undergoes
movements to optimize hydrogen bonding and van der Waals
contacts. The decrease in the crystallographic thermal B
factors seen in the flap region upon inhibitor binding are
thought to be representative of this effect (Suguna et al.,
1987). Studies by Hurley et al. (1992) have shown that there
is a strong correlation between the B factor of the residue

78
in the wild-type protein and the change seen in stability of
the mutant. If the residue was originally not very mobile,
low B value, the stability of a mutant at this position would
decrease. This observation suggests that mutations at
position 77 of the flexible flap region should not greatly
affect overall stability of the protein.
All of the information discussed above suggests that the
structures of the mutants of rhizopuspepsin have the same
overall folding pattern with small positional changes in the
side chains. The discussion, however, still does not explain
why differences, particularly between the m and AG values
of the Asp30 lie and Asp30lle/Asp77Thr mutant proteins, are
observed when comparisons are made to the WT-REC and pi 6
proteins. One possible reason may be the differences between
the folded and denatured states of the proteins. Several
studies have shown that the m value is related to the solvent
accessible surface area of the denatured state (Schellman,
1978; Shortle & Meeker, 1986). Shortle and Meeker (1989)
have shown that there is a correlation between the m value
and the solvent accessibility of a mutant. Mutants that
exhibit m values less than the wild-type protein show more
compact structures and residual structural components when
compared to the wild-type proteins by size-exclusion
chromatography and circular dichroism. These observations
suggest that rhizopuspepsin mutants with lie at position 30
may be able to optimize hydrophobic interactions in the
denatured state when compared to enzymes with Asp in this

79
position. These new interactions may make the Asp30 mutants
less stable with the equilibrium shifting slightly in favor
of the denatured state.
This chapter has presented kinetic and structural
evidence that the recombinant wild-type enzyme accurately
represents the naturally occurring isozymes of
rhizopuspepsin. Structural studies also suggest that the
mutant enzymes exhibit similar structures to the wild-type
proteins. The kinetic analysis of inhibitors (Chapter 5) and
substrates (Chapters 6 and 7) with the WT-REC and mutant
enzymes should not be significantly affected by structural
deviations. Kinetic comparisons of the WT-REC to the mutant
forms of rhizopuspepsin also support the idea that the
structures have not been significantly altered. These
results are discussed in Chapter 7.

CHAPTER 5
ANALYSIS OF THE SPECIFICITY OF RHIZOPUSPEPSIN THROUGH THE USE
OF INHIBITORS CONTAINING Pi AND Pi' SUBSTITUTIONS AND SCISSILE
BOND MIMETICS
INTRODUCTION
The inhibitors used in this study are based on the
cleavage sites in angiotensinogen, His-Pro-Phe-His-Leu*Val-
Ile-His-Asn, and the pl7/p24 HIV polyprotein junction Val-
Ser-Gln-Asn-Tyr*Pro-Ile-Val. Many potent renin and HIV
proteinase (HIV-PR) inhibitors have been generated using
these cleavage sites and the natural product pepstatin as
leads (Wiley & Rich, 1993). Studies varying side chain
functionalities, in order to probe possible "secondary
interactions" (Fruton, 1970; Medzihradszky et al., 1970) in
the enzyme binding subsites, inhibitor length, and
nonhydrolyzable peptide bond analogs have been undertaken to
optimize potency and bioavailability (Rosenberg et al., 1990;
Sawyer et al., 1991; Wiley and Rich, 1993). Other approaches
to inhibitor design have involved molecular modeling based on
the results from crystal structures of enzyme-inhibitor
complexes (Kempf et al., 1990; Thompson et al., 1992). The
results from crystal structures also aid in the analysis and
rationalization of differences seen for inhibitor
interactions and substrate specificities between family
80

81
members (Parris et al., 1992; Rao et al., 1993; Scarborough
et al., 1993).
This chapter reports the structure-activity
relationships of wild-type recombinant rhizopuspepsin with
inhibitors containing various scissile bond isosteres and Pi-
Pl' substitutions. The results of inhibitors with different
lengths are also discussed. The primary goal of this study
is to identify inhibitors for use as active site titrants.
Results and Discussion
Renin inhibitors have been synthesized using information
from many studies where variations have been made to create
potential interactions with the active site binding cleft
(Wiley and Rich, 1993). These extensive surveys have not
only led to many novel and potent inhibitors for renin, but
also decreased the time to find potent HIV-PR inhibitors.
Three major directions have been taken to increase potency,
bioavailability and stability of inhibitors: (1) substitution
of the scissile peptide bond of naturally occurring cleavage
junctions with mimetics, (2) elaboration of the Pi* side
chain, and (3) the determination of inhibitor length
requirements. The crystal structures of many aspartic
proteinase complexes with a variety of inhibitors aids in the
rationalization of binding interactions in the active site
which lead to a preference for one inhibitor over another.

82
The importance of the Pi-Pi' isostere interactions in
the structure-activity relationships of the methyleneamino
(CH2NH) and hydroxyethylene (CH(0H)CH2) containing inhibitors
is shown in Tables 5-1 and 5-2. A ten-fold increase in
potency (lower Ki) was observed when cyclohexylalanine (Cha,
also referred to as cyclohexylmethyl) is replaced by Phe in
the Pi position when comparing compounds 1 to 5, and 2 to 4.
This decrease in binding may be due to tfie loss of van der
Waals contacts and stabilizing aromatic/hydrophobic
interactions with Tyr77, Phell4, and Leul22 in the
hydrophobic Si subsite. These observations are consistent
with Cha side chain interactions seen in several complexes
with rhizopuspepsin and endothiapepsin (Sali et al., 1989;
Cooper et al., 1992; Parris et al., 1992). The incorporation
of Cha into the Pi position was shown by Sali et al. (Sali et
al., 1989) to influence the bound conformation of the P3
residue. The Cha substitution forces the P3 residue to
reside in a less favorable rotamer conformation possibly
resulting in lower inhibition. Other factors were stressed,
however, which must also be considered. The ability of a
compound to readily inhibit is determined by a combination of
the intermolecular interactions between the inhibitor and the
enzyme and the thermodynamic affects associated with
establishment or loss of van der Waals interactions, hydrogen
bonds and the change in solvent accessible surface of the
free enzyme and substrate.

Table 5-1. Inhibition constants for Xaa*F[CH2NH]Yaa modified
derivatives
CMPD
P5-P4-P3-P2
Pl-Pl1
P2'-P3'
K (JXM)
1
U79465E
Ac-Pro-Phe-His
Cha^F [X] Phe
nh2
> 200
2
U79211E
Ac-Pro-Phe-His
Cha^F [X] Val
NH2
> 200
3
U79464E
Ac-Pro-Phe-His
Cha'F [X] Cha
nh2
> 200
4
U79339E
Ac-Pro-Phe-His
Phe*F [X] Val
nh2
8.7 1.7
5
U71909E
Ac-Pro-Phe-His
Phe'F [X] Phe
nh2
21 4
6
U80011E
Ac-Pro-Phe-His
Phe'F [X] pClPhe
nh2
13 2
7
U80445E
Ac-Pro-Phe-His
Phe^F [X] Tyr
nh2
105 18
8
U81330E
Ac-Pro-Phe-His
Phe4/[X]pN02Phe
nh2
40 4
9
U70531E
dHis-Pro-Phe-His
Phe'F [X] Phe
Val-Tyr
5.2 0.7
10
U91990E
Ac-Pro-Hph-NMeHis
Phe'F [X] Phe
nh2
1.6 0.2
Hph = homophenylalanine; pNC>2Phe = p-nitrophenylalanine; X = CH2NH
pClPhe = p-chlorophenylalanine; Cha = cyclohexylalanine

Table 5-2. Inhibition constants for Leu4,[CH(OH)CH2]Val and Statine
modified derivatives.
Inhibitor
P5-P4-P3-P2
Pl-Pl'
P2-P3'
Ki
(nM)
11
U85548E
Val-Ser-Gln-Asn
Leu1? [X] Val
Ile-Val
<
0.1
12
U92522E
Ac-Ser-Gln-Asn
Leu'F [X] Val
Ile-NH2
1.4
0.3
13
U92517E
Ac-Gln-Asn
Leu*? [X] Val
Ile-NH2
1.4
0.3
14
U92516E
Ac-Asn
Leu1? [X] Val
Ile-NH2
510
92
15
U84728E
Ac
Leu'F [X] Val
Ile-NH2
20 200
3 700
16
U85964E
Ac-Val-Val
Leu'? [X] Val
lie-Amp
<
0.1
17
Pepstatin
Iva-Val-Val
Sta
Ala-Sta
0.7
0.2
18
Ac-Pepstatin
Ac-Val-Val
Sta
Ala-Sta
10
2
19
U77647E
Ac-Pro-Phe-His
Leu1? [X] Val
lle-NH2
54
7
X = CH (OH) CH2; Amp = aminomethylpyridine; Sta = statine = Leu1? [CH (OH) ] Gly
Iva = isovaleryl; Ac = acetyl

85
These additional factors may be able to explain the increased
potency seen for compounds containing Phe and Leu in the Pi
position for rhizopuspepsin. From these observations it was
clear that rhizopuspepsin has different requirements in the
Si pocket since Pi-Cha-containing analogs were found early on
to confer increased potency for renin (Boger et al., 1985;
Sawyer et al., 1990; Wiley and Rich, 1993).
Systematic substitutions of the Pi' position (compounds
4 to 8), where Pi was Phe, resulted in a large range of
potency. The Val, Phe and p-ClPhe substitutions exhibited
similar inhibitory capabilities while p-NC>2Phe and Tyr in Pi'
showed progressively higher values. The later two
substitutions may be unfavorable due to the interruption of
hydrophobic/aromatic interactions with Ile216 and 298, Trpl94
Trp294, and Phe296 of the Si' binding pocket or the
introduction of a slightly altered hydrogen bonding
arrangements. These observations point to further
differences between rhizopuspepsin and renin. Substitution
of the Pi' phenyl ring of Phe'F[CH(OH)CH2] Phe containing
compounds with halogens reduced potency for renin, while the
nitro substitution had little effect on binding (Young et
al., 1992), but did increase bioavailability (Thompson et
al., 1992). Possible reasons for the small differences seen
for these substitutions against renin extend from the much
larger Si1 pocket and the use of the hydroxyethylene isostere
which resulted in subnanomolar Ki values (Szelke et al.,
1980). The hydroxyethylene isostere shows significantly

86
greater inhibitory capacity, approximately 100-fold, to the
methyleneamino derivatives. This difference can be
illustrated by the comparison of compounds 4 and 19. This
increased affinity may mask subtle interactions only
discernible with the use of poorer inhibitors.
X-ray crystallographic analyses of inhibitor complexes,
kinetic analysis of substrates of different lengths, and
studies attempting to potentiate bioavailability by
decreasing the inhibitor size, have led to the understanding
that the extended active site of aspartic proteinases can
bind from 7-8 residues. Ligands bind in an extended P-strand
conformation with the amino acid side chains alternating from
side to side with the a-carbon backbone making a series of
highly conserved hydrogen bonds to the enzyme (Davies, 1990;
Suguna et al., 1992). Because of this arrangement, for
example, the S3/S1 and the S2/Si' subsites are adjacent to
each other and the residues at the corresponding positions in
ligand are also able to interact and influence each other.
The sum of these interactions generate the forces needed for
substrate binding and catalysis. Table 5-2 shows a deletion
series of U85548E (11). The P3, P2 and the P2' positions
contribute from 7 to 15 kcal/mole to binding. Removal of the
P5, P4 and the P3' residues of resulted in a 10-fold decrease
in potency. When the P3 and P2 residues were deleted,
however, a substantial increase in the Ki values of 5,000 and
200,000-fold were seen, respectively. These observations
mirror those seen in the analysis of HIV-PR with same set of

87
compounds (Sawyer et al., 1992) and in studies by many
investigators with truncation series of inhibitors (Rosenberg
et al., 1990; Wiley & Rich, 1993) and substrates (Fruton,
1976; Hofmann et al., 1988; Balbaa et al., 1993). The
results from substrate analyses have shown that the addition
of residues at the P3, P2 and P2' positions significantly
increases catalysis, kcat, without a corresponding change in
Km (Hofmann et al., 1988; Balbaa et al., 1993). It is thought
that the extra interactions and hydrogen bonds formed to
these residues produces a conformational change in the enzyme
or the substrate that lowers the activation barrier for the
formation of the tetrahedral intermediate (Pearl, 1985). The
loss of binding energy with N- and C-terminal truncations can
be recovered to yield nanomolar level inhibitors by modifying
the para position of phenylalanine residues in the P1-P1'
positions of pseudo peptides with large, extended hydrophobic
substituents, particularly for HIV-PR and renin (Roberts et
al., 1990; Thompson et al., 1992; Young et al., 1992).
The importance of the P3 and P2 residues in binding and
catalysis suggests that the bound conformation of inhibitors
in these regions is highly conserved. This is true for the
P3/S3 interactions, but a duality of binding has been seen in
the P2/S2 interactions in many endothiapepsin and
rhizopuspepsin inhibitor complexes (Foundling et al., 1987;
Sali, et al., 1989; Suguna et al., 1992). These differences
are thought to be partially dependent upon the residue
present in the Pi' position of the inhibitor. This

88
bifurcation can be seen, for example, by comparing the
structure of the reduced inhibitor U70531E (9, Suguna et al.,
1987) with that of U85548E (11, unpublished data). U70531E
and the U85548E have Phe and Val in the Pi' positions,
respectively. When Phe is present in the Pi1 position, the P2
side chain is oriented toward the S4 pocket surrounded by
Thr221, Phe278 and Leu223. The presence of Val in Pi',
however, positions the P2 side chain toward the long,
extended Si' pocket. The P2 side chain in this position may
result in the lower Ki seen for U79339E (4) when compared to
5 since the P2 His of 4 could partially fill the Si' subsite.
This apparent flexibility in the binding of the P2 side chain
implies that interactions in the S2 pocket may not actually
be that specific, due to the smaller loops in this region and
the presence of glycine at position 287 in rhizopuspepsin and
renin, and that the major contribution of the P2 residue
comes from hydrogen bonds to backbone. Even though there is
flexibility in the orientation of the P2 side chain, studies
using inhibitors and substrates with systematic substitutions
have shown that the S2 subsite does contribute significantly
to the specificity differences seen for pepsin, cathepsin D,
cathepsin E and renin (Rao et al., 1993; Scarborough et al.,
1993). For example, cathepsin D cleaves substrates with
positively charged residues in the P2 position poorly
(Scarborough et al., 1993).

89
A comparison of 16 to the classical aspartic proteinase
inhibitors (17 and 18) containing the statine derivative
indicates the relative binding of isosteres to WT-REC in this
study.
Xaa'P [CH (OH) CH2] Yaa > statine >
Phe'F[CH2NH]Yaa Cha'F[CH2NH] Yaa
The hydroxyethylene and methyleneamino derivatives are better
models for substrate interactions because of the frame shift
seen in pepstatin complexes caused by the addition of two
extra main chain atoms and the loss of hydrophobic
interactions in the Si pocket (Szelke et al., 1980; Sawyer et
al., 1990).
In conclusion, the results from the x-ray crystallo
graphic analysis can be used to understand subtle
interactions in the active site which can be exploited for
drug design. It is important to remember, however, that
there will always be a balance between potency, selectivity
and bioavailability.
This analysis has yielded several potent inhibitors of
rhizopuspepsin which can be used to titrate the different
forms of rhizopuspepsin used to analyze and understand
substrate specificity (Chapter 6 and 7). An active site
titration of the enzyme stock solutions enables an accurate
calculation of the kinetic parameter kcat- U85548E (11) was

90
used to titrate the native isozymes and recombinant forms of
rhizopuspepsin.

CHAPTER 6
THE BROAD SUBSTRATE SPECIFICITY OF RHIZOPUSPEPSIN: ANALYSIS
WITH SYSTEMATIC SUBSTITUTIONS IN THE P5-P1 AND P2-P3'
POSITIONS OF THE SUBSTRATE LYS-PRO-ALA-LYS-PHE*NPH ARG-LEU
INTRODUCTION
The residues that line the active sites of enzymes play
a critical role not only in catalysis but also in creating
the environment for the discrimination between substrates
(Fersht, 1985). This selection process or specificity is
governed by a myriad of favorable and unfavorable
interactions. Favorable intermolecular interactions include
hydrogen bonds, solvent exclusion from hydrophobic surfaces
and van der Waals interactions. Energetically unfavorable
interactions include accommodating suboptimal substrates with
distorted or lost hydrogen bonds to the enzyme and cavity
formation.
The specificity of the aspartic proteinases has been
under intense study in the effort to understand interactions
that distinguish each enzyme in this family. The goal has
been to exploit these differences in order to create
exquisitely targeted therapeutics, especially for renin and
HIV proteinase. Many early studies have shown the importance
of the 'secondary interactions* of the S3 and S2 subsites in
specificity and catalysis (Fruton, 1976; Dunn et al., 1987;
91

92
Hofmann et al., 1988; Pohl & Dunn, 1988). Recent studies
with porcine pepsin, cathepsin E (Rao-Naik, unpublished data)
and cathepsin D (Scarborough et al., 1993) have confirmed
this information and extended the search for other enzyme
interactions with the P5, P4, P2' and P3' positions of
substrates.
This chapter describes the results of the kinetic
analysis of wild-type recombinant rhizopuspepsin.
Systematically substituted octapeptide substrates were
analyzed with rhizopuspepsin in an effort to identify
specificity differences from the mammalian enzymes.
Substitutions were made in the P5-P1 and the P2,-P3' positions
of the substrate Lys-Pro-Ala-Lys-Phe*Nph-Arg-Leu to explore
potential hydrophobic, electrostatic, and hydrogen bonding
interactions. An additional series of peptides, containing
lie in the P3 position of the parent above and substitutions
in P2, were studied to complement the analysis of S2
interactions. This information may prove to be useful in the
design of targeted anti-fungal agents for the treatment of
Candida infections.
Results
L.Y.£7PrD.-Ala-Lys-Phe*Nph-Arg-Leu Based Substrates
Table 6-1 lists the kinetic parameters at pH 3.5 for the
library of substituted peptides based on the parent substrate

93
Lys-Pro-Ala-Lys-Phe*Nph-Arg-Leu. Rhizopuspepsin is able to
cleave the majority of the substrates with nearly equivalent
efficacy. Some of the subtle differences in kinetic
parameters can be rationalized by examining the x-ray crystal
structures of rhizopuspepsin.
positions. Rhizopuspepsin appears to show no
preference for the amino acid residues placed in the P5
position of the substrate. The P4 substitutions, however,
lead to some differences. Rhizopuspepsin prefers leucine in
this position followed by proline. The Arg substitution is
clearly disfavored. This can be rationalized by the mode in
which ligands bind to the active sites of the aspartic
proteinases. Because ligands bind in a extended ^-strand
conformation, the P4 and P2 residues can potentially interact
with each other (Figure 1-2). Electrostatic repulsion
between the Arg and the Lys in P2 may be responsible for this
observed specificity difference.
and P? positions. All subtrates with P3 substitutions
are cleaved readily with a slight preference for hydrophobic
side chains. A similar preference is seen for the P2
position substitutions. Interestingly, rhizopuspepsin is
able to cleave the Arg substitutions in P3 and P2. This
observation is in direct contrast to the mammalian enzymes.
Pi position. The observations in this series confirms
the specificity of the aspartic proteinases for large
hydrophobic junctions. Phe and Leu are preferred in this
position.

Table 6-1. Kinetic parameters for WT-REC rhizopuspepsin with the substrate
Lys-Pro-Ala-Lys-Phe*Nph-Arg-Leu containing systematic substitutions in
Ps-Pl
and
P2'-P3
' at
pH 3.5a
P5
P4
P3
P2
Pi*Pi 1
P2
P3
kcat
(sec"
1)
Km
(|XM)
kCat/Km
(s-J-M-1) X
10"6
Lys
Pro
Ala
Lys
Phe*Nph
Arg
Leu
12

1
11

1
1.13

0.20
Ser
5

1
4

1
1.21

0.32
Asp
5

1
7

3
0.76

0.29
Arg
6

1
3

1
1.74

0.76
Ala
7

1
6

1
1.14

0.17
Leu
5

1
7

2
0.82

0.27
Ser
9

1
2

0.4
0.42

0.09
Asp
22

3
30

6
0.73

0.18
Arg
3

0.3
43

10
0.06

0.02
Ala
15

1
19

1
0.75

0.08
Leu
22

2
14

1
1.57

0.17
Ser
9

1
16

2
0.60

0.08
Asp
13

1
20

2
0.65

0.09
Arg
13

2
9

1
1.39

0.25
Leu
10

1
7

1
1.37

0.21
Ser
9

1
9

1
1.03

0.18
Asp
7

1
8

1
0.83

0.12
Arg
9

1
4

0.9
2.17

0.56
Ala
20

1
8

2
2.39

0.46
Leu
9

1
4

0.7
2.31

0.53
Ala*Nph
4

0.4
24

4
0.17

0.03
Val*Nph
b
b
b
Leu*Nph
16

2
19

5
0.86

0.27
Ser
1

0.1
2

0.5
0.55

0.17
Asp
1

0.1
9

2
0.09

0.03
Ala
7

1
6

1
1.15

0.31
Leu
3

0.3
4

0.9
0.73

0.18
Ser
39

5
63

10
0.61

0.13
Asp
26

3
36

5
0.72

0.14
Arg
30

4
79

10
0.38

0.07
Ala
36

5
45

7
0.81

0.16
aAll peptides are eight residues in length.
bCleavage seen only after an extended incubation period with excess enzyme.

95
The Ala substitution is cleaved but to a lesser extent. The
substitution of Val into this position results in a peptide
that is not cleaved readily. This can be rationalized
because of valine side chain proximity to the catalytic
aspartic acid Asp32. Analysis of the inhibitor complexes of
rhizopuspepsin shows that (^-branched amino acids cannot be
accommodated in this position because of steric restrictions
and potential collisions with Asp32.
PV position. There appears to be a preference for Ala
or Arg in this position. Because of the large difference in
size of these amino acids, this observation might appear to
be unusual. This trend can be rationalized with the aid of
the crystal structures and may be substantiated by the
results seen with the Asp substitution. The P2' side chain
of the substrate is bound into the S2' subsite of the enzyme
which is primarily created by residues that line the
underside of the flap. One strictly conserved residue in
this region that has been shown by site-directed mutagenesis
to have a large effect on specificity is Tyr75 (Suzuki et
al., 1989). The substitution of the small Ala side chain
should not be sterically restricted by Tyr75. The Arg is
probably tolerated in this position because of its high
degree of flexibility and length. An Arg side-chain may be
able to extend beyond Tyr75 and make possible interactions
with solvent. An Asp residue, however, is probably not able
to avoid steric constraints. Another possibility for the
poor cleavage of the Asp substituted peptide may be the

96
presence of a negative charge near the catalytic apparatus.
An explanation for the preference for the Arg substitution is
also discussed in Chapter 7.
Pd' position. The replacement of leucine in the parent
peptide resulted in superior substrates with respect to
catalytic efficiency, kcat- The kcat values increased 2 to
3-fold in comparison to the other substrates in this
substitution series. This observation was used in the design
of the substrates used in Chapter 7.
Lvs-Pro-Ile-P9-Phe*Noh-Arq-Leu Based Peptides
In contrast to the peptide series with Ala in P3, the
combination of lie in P3 and Leu in P2 resulted in a poorly
cleaved peptide (Table 6-2). The extension of the Asp side
chain to Glu results in a 10-fold increase in kcat-
Interestingly, rhizopuspepsin is able to cleave the Arg
substitution but is not able to readily cleave the substrate
with His in P2.
Discussion
Besides the few instances where the substitution of an
amino acid in the peptide resulted in an understandable
unfavorable interaction, rhizopuspepsin is a protease of
relatively broad specificity. Rhizopuspepsin is able to
efficiently cleave a wide variety of peptides. This ability
is in stark contrast to the mammalian enzymes with the same

Table 6-2. Kinetic analysis of WT-REC rhizopuspepsin with substrates of the form
Lys-Pro-Ile-P2~Phe*Nph-Arg-Leu at pH 3.5
P5
P4
P3
P2
Pl*Pl*
P2'
P3 1
^cat
(sec-1)
Km
(JXM)
kCat/Km
(M-1S-1)
X 10-6
Lys
Pro
He
Glu
Phe*Nph
Arg
Leu
22
2
6

2
3.84
1.16
VO
Ala
8
1
5

1
1.56
0.42
Ser
8
1
11

2
0.76
0.16
Arg
3
0.4
9

1
0.36
0.06
Asp
2
0.3
6

1
0.33
0.06
Leu
1
0.1
5

1
0.21
0.05
His
0.03
0.003
5

1
0.005
0.001

98
peptide series, but is comparable to the characteristics of
the Candida aspartic proteinases.
Comparison to the Mammalian Enzymes
Studies with the same peptide series described above
have shown that porcine pepsin, cathepsin E and cathepsin D
have different specificities toward substrates than those
seen for rhizopuspepsin. These differences center around
interactions in the S3 and S2 subsites.
Porcine pepsin is able to cleave substrates containing
basic residues in P2, but is not able to accommodate these
residues in P3 at pH 3.5 (Rao-Naik, unpublished results, Pohl
& Dunn, 1988). Upon raising the pH, however, a dramatic
increase in kcat is seen for the peptide containing Arg in P3.
This observation has been shown by site-directed mutagenesis
to be mediated by Glul3 (Rao-Naik, unpublished results).
Rhizopuspepsin also contains a glutamic acid at position 13
but does not show the same pH dependence.
Cathepsin D has been shown not to accommodate basic
residues in P2 and to have a preference for larger
hydrophobic residues in P3 (Scarborough et al., 1993). Site-
directed mutagenesis has shown that the P2 effect is mediated
by Met287. Upon changing Met287 to Glu, the residue present
in porcine pepsin, basic substitutions were tolerated and
improved cleavage was seen.

99
Cathepsin E exhibits characteristics which make it
similar to porcine pepsin, cathepsin D and rhizopuspepsin.
Cathepsin E and porcine pepsin are alike in that they are
able to cleave substrates with a positive charge in P3.
Cathepsin E is able to cleave substrates with Lys and Arg
residues in P2, but it prefers hydrophobic substitutions as
seen for cathepsin D. Interestingly, cathepsin E is not able
to cleave the His substitution in this position. This trend
is similar to that seen for rhizopuspepsin.
Comparison to the Candida Aspartic Proteinases
Three Candida aspartic proteinases have been analyzed
with the same family of peptide substrates: enzymes isolated
from Candida albicans (CAAP), Candida parapsilosis (CPAP) and
Candida tropicalis (CTAP) (Fusek et al., 1994). These
enzymes are similar to rhizopuspepsin in that they exhibit
broad specificity. The majority of the substrates are
cleaved efficiently. The Candida enzymes are able to cleave
substrates containing basic substitutions in P3 and P2. They
also show a similar aversion for Asp in P2'. The most
notable difference of the Candida enzymes from rhizopuspepsin
is their inability to cleave the substrate with leucine
substituted in P4. The CTAP enzyme also shows a unique
difference from rhizopuspepsin and the other Candida
proteinases in that is also unable to cleave the substrate
with Glu substituted in P2. An understanding of the possible

100
structural variations that lead to the differences between
the Candida enzymes and rhizopuspepsin awaits the solution of
inhibitor complexes of the Candida proteinases by x-ray
crystallography. Efforts in this direction are already
underway (Cutfield et al., 1993).
From the substrate studies of rhizopuspepsin and the
Candida enzymes, it is clear that the fungal enzymes have a
broad specificity toward substrates with substitutions in P5-
Pl and P2'-P3' Distinct interactions do not appear to be
present, in contrast to the mammalian enzymes, which can be
exploited in the design of targeted anti-fungal agents.
Despite this apparent lack of specificity, the fungal enzymes
have the unique ability to cleave substrates containing
lysine in Pi (Hofmann & Hodges, 1982). This particular
characteristic of the fungal enzymes has been studied by
site-directed mutagenesis and is described in Chapter 7.

CHAPTER 7
ENGINEERING THE SUBSTRATE SPECIFICITY OF RHIZOPUSPEPSIN: THE
ROLE OF ASP30 AND ASP77 IN THE ABILITY OF A FUNGAL ASPARTIC
PROTEINASE TO CLEAVE SUBSTRATES WITH LYSINE IN Pi
Introduction
Double mutant cycle analysis has been used in the effort
to interpret and quantitate the amount of function lost or
gained by a mutant protein by comparing the properties of two
single mutants to the corresponding double mutant (Carter et
al., 1984; Akers & Smith, 1985; Fersht, 1985; Wells, 1990;
Mildvan et al., 1992). These functions include catalysis,
binding and intrinsic stability. An assumption made in this
type of analysis is that the loss of function directly
measures the contribution of the mutated residue. The
evaluation of mutation induced effects are facilitated by
comparing the changes seen in kinetic, kcat and kcat/Km, and
thermodynamic, equilibrium constants, parameters.
The substrate specificity or the ability of a proteinase
to distinguish between two substrates that are highly similar
in structure has been routinely studied by comparing the
kcat/Km parameter. This parameter or 'specificity constant'
is a complex function of all the kinetic constants, involving
the binding and subsequent turnover of substrate to form the
products of the reaction, up to and including the first
101

102
irreversible step in the catalytic mechanism. By comparing
the specificity constants of the wild-type and mutant enzymes
for the same substrate, the change in the free energy barrier
to reach the transition state complex (ES*) from the free
enzyme and substrate (E+S) can be calculated. This
difference reflects the stabilization or destabilization of
the transition state.
From the examination of the high resolution crystal
structures of the inhibitor complexes of penicillopepsin
(James & Sielecki, 1985), endothiapepsin (Pearl & Blundell,
1984), rhizopuspepsin (Suguna et al., 1987) and the HIV
proteinase (Swain et al., 1990) and the analysis of kinetic
isotope effects, particularly in the HIV proteinase system
(Hyland et al., 1991a,b; Rodriquez et al., 1993), the
transition state in the general acid-base catalytic mechanism
of the aspartic proteinases is thought to resemble a
tetrahedral intermediate. In the cleavage of substrates
containing Tyr-Pro junctions by the HIV proteinase, a water
molecule located between the two catalytic aspartic acids is
considered to be made more nucleophilic by the transfer of a
proton to the negative Asp25', Asp215 in porcine pepsin,
after the formation of an intermediate where the nitrogen of
the scissile bond is partially positively charged. This
intermediate is thought to be a result of the establishment
of a strong low-barrier hydrogen bond (Cleland, 1992) from
the protonated Asp25 (the equivalent function is performed by
Asp32 in porcine pepsin) to the scissile bond carbonyl group.

103
Following the attack of the water molecule to form the
tetrahedral amide hydrate intermediate, the transition state
is proposed to collapse to form products upon protonation of
the proline nitrogen by Asp25' and the deprotonation of the
gem-diol hydroxyl group.
Kinetic analysis has shown that the rate of exchange of
180 from H2180 into the starting substrate exceeded the rate
of label incorporation for the reverse peptidolytic reaction.
This evidence suggests that the rate determining or first
irreversible step in the reaction mechanism is the breakdown
of the tetrahedral intermediate to form products.
Experiments analyzing solvent deuterium isotope effects
suggest that this breakdown or collapse of the transition
state occurs by the simultaneous transfer of two protons as
described above (Hyland et al., 1991a,b; Rodriquez et al.,
1993). Site-directed mutagenesis and double mutant cycle
analysis can give insight into how particular amino acid
residues of the aspartic proteinases contribute to the
stabilization of the transition state that results in the
overall preference for one substrate over another.
The fungal aspartic proteinases exhibit a unique
specificity toward substrates in comparison to the mammalian
enzymes. The fungal enzymes are able to cleave substrates
containing lysine in the Pi position. Sequence and
structural comparisons have suggested Asp30 and Asp77 in
establishing this characteristic ability.

104
This chapter outlines the results achieved by mutating
residues Asp30 and Asp77 to the corresponding residues in
porcine pepsin, Ile30 and Thr77. The resulting mutant
proteins were overexpressed in E. coli and shown, after
refolding and activation, to be structurally similar to the
wild-type recombinant enzyme and the native isozyme pi 6
(Chapter 3 and 4). The recombinant rhizopuspepsins were
tested with oligopeptide substrates containing systematic
substitutions of lysine into the Pi, P2 and P3 positions.
Double mutant cycle analysis was performed to definitively
identify the residue or residues responsible for the ability
to cleave carboxy-terminal to a lysine residue and to
estimate the energy contributions to this specificity through
hydrogen bonding interactions (Carter et al., 1984; Akers &
Smith, 1985; Fersht, 1985; Wells, 1990; Mildvan et al.,
1992).
Results
Kinetic Analysis of the Recombinant Rhizoousoeosins
The summary of the kinetic analysis of the recombinant
rhizopuspepsins with substrates with the systematic
substitution of lysine into the P3-P1 positions is shown in
Table 7-1. All peptides were cleaved at the expected
positions, that is, between Phe*Nph, Nle*Nph or Lys*Nph. The
enzymes show consistent trends in the kinetic parameters for

105
all the substrates studied upon increase of the assay pH:
kcat values were independent of pH while two to five-fold
increases in kcat/Km were observed, mainly as the result of a
corresponding decrease in Km. Importantly, all the enzymes
have similar kcat values at pH 5.0 for peptide 4. This
observation confirms the results of the structural analyses
in Chapter 4 that supported the overall similarity of the
protein folds and that the enzymes can be compared with
confidence.
Wild-tvne recombinant. The wild-type enzyme exhibits
the characteristic broad specificity seen for similar
peptides discussed in Chapter 6. All substrates, with (1-3)
or without (4,5) the lysine residue placed in several
alternative positions, were readily cleaved. Even though
there is still a preference for the hydrophobic junction
(1 and 4), the enzyme is able to cleave the Pi-lysine peptide
(3) with equivalent efficiency (kcat) .
Asp30lle mutant. This mutant cleaves the substrates,
including the Pi-lysine derivative, with nearly identical kcat
values to the wild-type enzyme. A 2-3 fold increase in Km
is seen with peptides 1-4. This increase results in kcat/Km
values that are in the range seen for porcine pepsin with the
same substrates (Table 7-2). Larger changes are seen in all
the kinetic parameters for peptide 5. These differences are
probably due to the loss of hydrophobic interactions when the
aromatic phenylalanine is replaced by the aliphatic
norleucine.

Table 7-1. Kinetic analysis of wild-type and mutant rhizopuspepsins: systematic
substitution of lysine into P3~Pia
PEPTIDEb
WT-REC
Asp30Ile
pH
heat
Km
kcat/Km
heat
Km
hcat/Km
(s1)
(HM)
(M-
1S
-1)
(s'1)
(HM)
(M-
ls-1)
x 10
-6
XlO"6
3.5
24
3
19

2
1.24

0.17
29
3
64

2
0.46
0.05
1
KPKAF*XRA
5.0
28
3
6

1
4.90

0.95
29
3
14

1
2.15
0.30
3.5
36
5
45

7
0.81

0.16
47
7
139

23
0.34
0.07
2
KPAKF*XRA
5.0
41
5
19

2
2.13

0.34
30
3
25

2
1.22
0.18
3.5
35
4
98

6
0.36

0.05
33
5
395

51
0.08
0.02
3
KPAAK*XRA
5.0
35
4
21

1
1.66

0.20
40
5
99

9
0.40
0.06
3.5
33
4
15

1
2.17

0.26
15
2
31

1
0.49
0.05
4
KPAAF*XRA
5.0
34
4
7

1
4.73

0.63
16
2
17

2
0.94
0.17
3.5
18
2
18

1
1.00

0.12
67
8
173

15
0.39
0.06
5
KPAAZ*XRA
5.0
22
2
10

1
2.26

0.41
56
6
60

5
0.94
0.13
aThe kinetic parameters ( standard deviations) were determined at 37C with either 0.1 M sodium
formate, pH 3.5, or 0.1 M sodium acetate, pH 5.0, containing NaCl to maintain constant ionic
strength as outlined in Chapter 2.
Abbreviations used in the substrates are: X = p -nitrophenylalanine; Z = norleucine; = site of
cleavage.
ckCat/Km values were determined spectrophotometrically with the assumption that [S] Km;
v = (kcat/Km)[E]0[S]0.
dkcat/Km values were determined by capillary electrophoresis; v = (kcat/Km) [E]o tsl0-
106

Table 7-1 continued
PEPTIDEb
PH
Asp77Thr
Asp30Ile/Asp77Thr
kCat
Km
kcat/Km
kcat
Km
kcat/Km
(s'1
(|iM)
(M-
is'1)
(s'1)
(pLM)
(M'i-s-1)
X10*4
XlO-4
3.5
6
1
561

115
1.00
0.30
_
_
C0.26 0.02
1
KPKAF*XRA
5.0
6
1
283

25
1.90
0.30
-
-
c0.71 0.04
3.5
6
1
487

83
1.12
0.30
-
-
c0.67 0.04
2
KPAKF*XRA
5.0
7
1
296

44
2.50
0.60
9
1 449
36
2.03 0.24
3.5
-
-
d5
.0
0.7 X lo*4
-
-
d0.9 0.1 x 10-4
3
KPAAK*XRA
5.0
-
-
d13
.7
2.1 X 10-4
-
-
d3.4 0.3 X 10-4
3.5
15
3
408

48
3.80
0.80
-
-
C1.90 0.10
4
KPAAF*XRA
5.0
17
2
228

17
7.30
1.20
13
1 253
25
5.30 0.70
3.5
2
0.4
297

41
0.80
0.20
-
-
c0.07 0.01
5
KPAAZ*XRA
5.0
5
1
289

39
1.60
0.40
1
0.1 462
62
0.21 0.04
aThe kinetic parameters ( standard deviations) were determined at 37'C with either 0.1 M sodium
formate, pH 3.5, or 0.1 M sodium acetate, pH 5.0, containing NaCl to maintain constant ionic
strength as outlined in Chapter 2.
bAbbreviations used in the substrates are: X = p -nitrophenylalanine; Z = norleucine; = site of
cleavage.
Ckcat/Km values were determined spectrophotometrically with the assumption that [S] Km
v = (kcat/Km)[E]o[S]o-
dkCat/Km values were determined by capillary electrophoresis; v = (kCat/Km)[E]o[S]q.
107

108
Table 7-2. Kinetic analysis of porcine pepsin: systematic
substitution of lysine into P3~Pia
PEPTIDE
PH
PPEP
kcat
Fm
kcat/
(s
-1)
(|IM)
(M-
ls-1)
X
10"6
3.5
_
_
b0.0030
0.0005
1
KPKAF*XRA
5.0
2.4
0.5
122
20
0.019
0.005
3.5
21
5
339
77
0.06
0.02
2
KPAKF*XRA
5.0
20
3
31
2
0.66
0.11
3.5
c
3
KPAAK*XRA
5.0
c
3.5
29
6
83
12
0.41
0.09
4
KPAAF*XRA
5.0
24
5
61
11
0.39
0.10
3.5
22
4
103
12
0.21
0.04
5
KPAAZ*XRA
5.0
16
3
34
3
0.46
0.09
aAssay conditions described in Chapter 2. Abbreviations are
the same as those used in Table 7-1.
bkCat/Km values were determined spectrophotometrically with
the assumption that [S] Km; v = (kCat/Km) [E] o [S] o.
cCleavage occurred between Nph-Arg (X-R) instead of Phe*Nph
(F*X).

109
Aso77Thr mutant. The mutation of Asp77 to Thr resulted
in an enzyme which displays reduced kcat values toward this
series of substrates. The most significant result of this
mutation appears to be the 100-fold increase in the Km
parameter and the parallel decrease in kcat/Km. This mutant
enzyme is no longer competent to cleave the Pi-lysine
peptide. This result is illustrated by the 105-fold decreases
in kcat/Km from 3 60,000 to 5 M-1s-1 at pH 3.5 and 1,660,000 to
13.7 M-1s-1 at pH 5.0.
Aso30lle/Aso77Thr mutant. The double mutant enzyme is
able to cleave all the substrates but with decreased
specificity (kcat/Km) This enzyme is similar to the Asp77Thr
mutant in that it is characterized by the inability to
readily cleave the Pi-lysine substrate.
Kinetic Analysis of Porcine Pepsin
Table 7-2 lists the kinetic parameters determined for
peptides 1-5 with porcine pepsin. Peptides 2, 4 and 5 are
readily cleaved with kcat values similar to that of wild-type
rhizopuspepsin. The kcat/Km values, however, are 2 to 5-fold
lower. Peptide 1 is poorly cleaved at pH 3.5, consistent
with previous observations (Rao-Naik & Dunn, unpublished
results; Pohl & Dunn, 1988). Improvement is seen upon
raising the pH, although not to levels seen for the other
peptides. Pepsin behaves similarly to rhizopuspepsin with
respect to peptide 2 in that the specificity (kcat/Km)

110
increases without changes in kcat- Importantly, hydrolysis of
the Pi-lysine substrate (3) by porcine pepsin did not occur
between Lys*Nph. After an extended incubation period with
excess enzyme, cleavage was seen between Nph-Arg.
Discussion
Site-directed mutagenesis has become an invaluable
technique in dissecting the molecular interactions which give
rise to the binding, discrimination and catalytic functions
of proteins. The use of double mutant cycles and the
analysis of additivity between mutants can give information
regarding the contribution of an amino acid residue to a
particular function of a protein, as well as to its own
stability (reviewed by Wells, 1990; Shortle, 1992). In this
study, site-directed mutants were made to investigate the
contributions of Asp30 and Asp77 to the unique ability of
rhizopuspepsin to cleave substrates with lysine in the Pi
position. The use of rhizopuspepsin as a model system for
studying active site interactions of fungal aspartic
proteinases is facilitated by: (1) high resolution crystal
structures of the native and inhibited forms of the enzyme,
(2) ease of cloning and expression in E. coli, and (3)
sensitive kinetic methods to quantify the effects of the
mutations (Rheinnecker et al., 1993).

Ill
Substrate Pesian
The initial experimental evidence that the fungal
aspartic proteinases were distinct from the mammalian enzymes
came from the observation that they were able to activate
trypsinogen at a Lys-Ile bond (Graham et al.f 1973; Morihara
& Oka, 1973). Hofmann and his coworkers have confirmed this
specificity by using substrates of the form Ac-(Ala)m-
Lys*Nph-(Ala)n-amide to study the effects of secondary
substrate binding interactions (Hofmann & Hodges, 1982;
Hofmann et al., 1984; Balbaa et al., 1993). The substrates
used in this study are slightly different and were derived
from the highly soluble peptide Lys-Pro-Ala-Lys-Phe*Nph-Arg-
Leu. This substrate and its P5-P1 and P2'-P3' systematic
substitution derivatives have been extensively used in the
exploration of subsite specificities of rhizopuspepsin
(Chapter 6), pepsin (Pohl and Dunn, 1988; Rao-Naik,
unpublished data), cathepsin E (Rao-Naik, unpublished data)
and cathepsin D (Scarborough et al., 1993). Rhizopuspepsin
is able to cleave all the substrates in this series with
nearly equivalent specificities (kcat/Km). Changing the P3'
Leu to Asp, Ala, Arg or Ser, however, caused a two-fold
increase in kcat. This information suggested that the
substitution of Ala in P3' would be advantageous for the
study of mutants that may have decreased catalytic activity.
As illustrated in Figure 1-3, Asp77 has the potential to
interact with the P3-P1 residues of a ligand bound in the

112
active site. Asp30 is located between the S3 and the Si
subsites and has the capacity to interact with the P3 and Pi
backbones and side chains. Besides the Pi interactions seen
in the lysine pepstatin derivative complex, crystallographic
analyses of penicillopepsin with renin inhibitors have shown
that Asp77 can also interact with a histidine residue in the
P2 position (Blundell et al., 1987). Lysine was substituted
into P3-P1 of the substrate in order to definitively show the
positional requirement for interaction with Asp77. These
peptides are also designed to give information about the
contribution of Asp30, if any. In this study, two peptides
(4 and 5) were used as controls representing substrates that
did not contain charges in positions P3-P1. The Nle peptide
was used to mimic the lysine side chain.
Kinetic Analysis
The values and pH dependence of the kinetic parameters
seen in Table 7-1 for peptide 3 are directly comparable to
those historically seen for rhizopuspepsin (Hofmann et al.,
1984; Balbaa et al., 1993). Hofmann suggested that the pH
dependence was indicative of a specific interaction of a
carboxyl group with the lysine residue in Pi. Interestingly,
the same pH dependence of kcat/Km is seen for all the peptides
and forms of rhizopuspepsin in this study. Table 7-3 lists
the changes in transition state stabilization energies upon
changing the assay pH for the recombinant rhizopuspepsins and

Table 7-3. Transition state stabilization energy changes seen with
variation in pH from 3.5 to 5.0 for the recombinant rhizopuspepsins
and porcine pepsina
PEPTIDE
WT-REC
Asp30lle
Asp77Thr
Asp30lle/Asp77Thr
PPEP
AAG^ (kcal
mole-1)b
1
KPKAF*XRA
-0.9
-1.0
-0.4
-0.6
-1.1
2
KPAKF*XRA
-0.6
-0.8
-0.5
-0.7
-1.5
3
KPAAK*XRA
-0.9
-1.0
-0.6
-0.8
c
4
KPAAF*XRA
-0.5
-0.4
-0.4
-0.6
o
o
5
KPAAZ*XRA
-0.5
-0.5
-0.4
-0.7
-0.5
aAbbreviations and assay conditions are described in Table 7-1.
bAAGt = -RT/Km(pH 5.0)lIKm(pH 3.5)]. The standard deviations
range from 0.1 to 0.2 kcal mol-1.
cCleavage occured between Nph-Arg (X-R) instead of Phe*Nph (F*X).
113

114
porcine pepsin. A negative value indicates that substrate
interactions are improved at pH 5.0. The similarity in the
values, between -0.5 and -1.0 kcal mol-1, for the recombinant
enzymes over the range of substrates tested implies that the
interactions seen between rhizopuspepsin and a substrate may
have a general electrostatic component. Asp30 and Asp77 do
not play a role in this phenomenon.
Interactions between porcine pepsin and this series of
substrates also show a dependence on pH. The observations
with lysine in P3 and P2 are probably related to interactions
with Glul3 and Glu287, respectively (Pohl and Dunn, 1988;
Rao-Naik, unpublished data). The dependence seen for peptide
5 is probably related to slight alterations in active site
interactions in order to optimize interactions with the Nle
residue. The lack of pH dependence for peptide 4 with pepsin
suggests that there may be a specific difference between
pepsin and rhizopuspepsin that is responsible for the general
electrostatic effect seen with rhizopuspepsin. All the
substrates in this series contain an arginine in the P2'
position. There is a highly conserved difference between the
mammalian and fungal enzymes at residue 37 in the S2' pocket
(Table 1-3). Rhizopuspepsin contains an Asp at this position
while porcine pepsin has an Asn. The suggestion that Asp37
may form a specific electrostatic interaction with the P2'
arginine awaits further analysis by mutagenesis studies.
The contrasting pH dependence in the kinetic parameters
for rhizopuspepsin and porcine pepsin for peptide 4 may also

115
result from differences in the pi values of the proteins.
Pepsin has a pi value less than 2 while the different
rhizopuspepsins have pi values ranging from 5 to 6 (Chapter 3
and 4). At pH 3.5 and 5.0 porcine pepsin is highly
negatively charged. In contrast, the rhizopuspepsins are
positively charged at pH 3.5 and become more negatively
charged at pH 5.0 with a net charge near zero. In order for
wild-type rhizopuspepsin at pH 5, for example, to have an
overall net charge of zero, many acidic residues must be
titrated to balance the large number of positive residues on
the enzyme surface. This increase in negative charges may
attract and facilitate the catalysis of positively charged
substrates.
Rhizopuspepsin readily cleaves peptide 3 containing
lysine in Pi. The Asp77Thr mutants cleave this substrate
poorly and porcine pepsin cleaves the peptide in the wrong
place altogether. The degree to which Asp30 and Asp77
contribute to specificity can be analyzed by using double
mutant cycles (Carter et al., 1984; Ackers & Smith, 1985;
Fersht, 1985). The free energy of transition state
stabilization can be affected by the introduction of
mutations. The changes seen upon multiple mutations are
generally additive providing (1) the residues do not directly
contact each other, (2) large structural changes do not occur
or (3) no change in the reaction mechanism has occurred
(Wells, 1990).

116
The free energy change of the double mutant AAGD30ID77T is
related to that of the single mutants:
^^^D30I,D77T A^^D30I + AAGD77T + AGj .
The AGj term or the coupling energy reflects the extent to
which the single mutations affect each other. When the sites
are functioning independently from each other, the coupling
energy is zero.
Table 7-4 shows the double mutant cycles at pH 3.5 and
5.0 for peptides 1-5. The cycles for peptides 1-3 are
strictly additive suggesting that severe changes in structure
are not seen as suggested by the structural results
(Chapter 4). Slight deviations from additivity are seen for
peptides 4 and 5 but do exhibit the same trends. Mutation of
Asp30 to lie results in an average loss in transition state
stabilization free energy of 0.7 kcal mol-1. This loss may be
attributable to a change in the hydration shell of the
protein or the loss of a hydrogen bond to the inhibitor
mediated through a water molecule. For peptides 1, 2, 4 and
5 an average loss in transition state stabilization free
energy of 2.3 kcal mol-1 is seen when Asp77 is changed to Thr.
The largest effect on mutating position 77 is clearly seen
with the 7 kcal mol-1 decrease in transition state
stabilization for peptide 3. This observation indicates that

117
Table 7-4. Double mutant cycle analysis of the recombinant
rhizopuspepsins: substitution of lysine into P3-P1 at pH 3.5
and 5.0a
Peptide pH 3.5b pH 5.0
4 900 000
0.5
2 150 000
WT
D30I
3.4
3.5
D77T
DBL
19 000
0.6
7 100
1 240 000
0.6
460 000
WT
D30I
3.0
3.2
D77T
DBL
10 000
0.8
2 600
2 130 000
0.3
1 220 000
WT
D30I
2.7
2.5
D77T
DBL
25 000
0.1
20 300
810 000
0.5
340 000
WT
D30I
2.6
2.4
D77T
DBL
11 200
0.3
6 700
3
4
KPAAK*XRA
360 000
0.9
80 000
WT
D3 0I
6.9
7.0
D77T
DBL
5.0
1.0
0.9
KPAAF*XRA
2 170 000
0.9
490 000
WT
D30I
2.5
2.0
D77T
DBL
38 000
0.4
19 200
1 660 000
0.9
400 000
WT
D30I
7.2
7.2
D77T
DBL
13.7
0.9
3.4
4 730 000
1.0
940 000
WT
D30I
2.6
1.8
D77T
DBL
73 000
0.2
53 000
5
KPAAZ*XRA
1 000 000
0.6
390 000
WT
D30I
3.0
3.9
D77T
DBL
8 000
1.5
730
2 260 000
0.5
940 000
WT
D30I
3.0
3.8
D77T
DBL
16 000
1.3
2 100
aWT = wild-type recombinant, D30I = Asp30Ile, D77T = Asp77Thr, and DBL
= Asp30Ile/Asp77Thr. The numbers above and below the enzyme type are
the kcat/Km values in M*1s~1 from Table 7-1. The free energy changes of
transition state stablization, AAGj, are shown in italics with standard
deviations that range from 0.1 to 0.2 kcal mol'1.
bAAGj = -RTlnlk^ / ^(mutant, mutant 2 )//Tm(wild- type, mutant 1)]

118
the Asp77 is the crucial residue for enabling cleavage of
peptides with lysine in the Pi position.
Figure 7-1 shows the potential hydrogen bonding
arrangement in the active site of penicillopepsin complexed
with a pepstatin based inhibitor containing a lysine in Pi
(James et al., 1985). Asp77 forms two hydrogen bonds to the
inhibitor backbone and one to the lysine side chain. An
intra-residue hydrogen bond is also seen for Asp77. The
enzyme also forms a hydrogen bond to Asp77 and to the lysine
side chain through Ser79. The hydrogen bonds to the P2 NH
and carbonyl of the inhibitor are strictly maintained in
inhibitor complexes of the aspartic proteinases (Davies,
1990). Extending from this conserved binding mode and the
high degree of structural homology in the active site region
between rhizopuspepsin and penicillopepsin (Figure 7-2), a
possible explanation for the inability of the Asp77 mutants
and the mammalian enzymes to cleave Pi-lysine substrates can
be proposed.
Figure 7-3 shows the proposed interactions for a Pi~
lysine substrate bound to the active site of rhizopuspepsin
and the Asp77 mutants that were designed to resemble porcine
pepsin. When Asp77 is mutated, several potential hydrogen
bonding interactions are lost, depending on the particular
type of substrate bound. When substrates that do not contain
a lysine in Pi (1, 2, 4 and 5) are bound, the Asp77 intra
residue hydrogen bond and the hydrogen bond between Asp77 and
Ser79 are lost. The double mutant cycles suggest that the

119
Figure 7-1. Hydrogen bonding interactions in penicillopepsin
between Asp77, Ser79 and the pepstatin derivative containing
lysine in the Pi position. Yellow dotted lines indicate
hydrogen bond distances in angstroms.

120
Figure 7-2. Ca carbon backbone superposition of
penicillopepsin and rhizopuspepsin complexed with inhibitors.
The pepstatin derivative containing lysine in the Pi position
is shown in cyan. Penicillopepsin and rhizopuspepsin are
shown in yellow and blue, respectively.

121
A
B
FLAP
Figure 7-3. Proposed hydrogen bonding interactions in WT-REC
rhizopuspepsin and Asp77Thr mutants with substrates
containing lysine in Pi. A, WT-REC; B, Asp77Thr mutants.

122
loss of these interaction yields a free energy change of 2.3
kcal mol-1. When the Pi-lysine substrate (3) is bound to
Asp77Thr or the double mutant, two additional hydrogen bonds
are lost to the lysine side-chain. These combined losses
appear to be worth up to 7 kcal mol-1. If these effects are
additive, the two hydrogen bonds to the lysine side chain are
worth approximately 4.7 kcal mol-1. The free energy changes
determined by these experiments are in the same range,
between 3 and 6 kcal mol-1, as those seen for other mutants of
charged hydrogen bond donors and acceptors (Fersht et al.,
1985; Fersht, 1988).
This study has shown that Asp30 may be partially
responsible for the general broad specificity of
rhizopuspepsin when compared to the mammalian enzymes. Asp77
is the critical residue required for enabling rhizopuspepsin
to cleave substrates containing lysine in Pi. The presence
of an aspartic acid residue at position 77 in the flap may
allow the formation of an extensive hydrogen bonding network
not only to the Pi lysine side chain but between the enzyme
itself. These interactions somehow enable the anomalous
cleavage of Pi-lysine containing peptides by lowering the
energy barrier to the transition state in the fungal aspartic
proteinases.

CHAPTER 8
CONCLUSIONS AND FUTURE DIRECTIONS
Conclusions
A combination of site-directed mutagenesis and kinetic
analysis has clearly established the importance of Asp77 in
the unique lysine primary specificity of fungal aspartic
proteinases. Asp30, which is not conserved in the fungal
enzymes, was shown not to influence this specificity but may
play a role in general electrostatic interactions. The
presence of an aspartic acid residue at position 77 has the
potential to establish an extensive hydrogen bonding network
between the enzyme and the substrate that enables the fungal
enzymes, but not the mammalian enzymes because they contain a
Thr77, to cleave substrates containing lysine in the Pi
position.
Another way to look at this unique ability is that the
Asp77 residue in the fungal enzymes may only be providing
extra interactions that stabilize what would normally be a
very unfavorable residue interaction in the hydrophobic Si
subsite of the aspartic proteinases. The lysine side chain
in the penicillopepsin complex extends out of the Si subsite
in order to make interactions with Asp77 which is located on
the surface of the protein. This orientation of the lysine
123

124
side chain allows for hydrophobic interactions with the
aliphatic portions of the side chain in the Si subsite and
the stabilization of a positive charge in a hydrophobic
region.
Specific and general electrostatic interactions have
also been shown, through the use of site-directed mutagenesis
and kinetic analysis of substrates containing arginine and
lysine in the P3 and P2 positions, to play a key role in the
substrate specificity of porcine pepsin, particularly in the
S3 and S2 binding subsites (Rao-Naik, unpublished data). The
electrostatic interactions of positively charged side-chains
of the substrate in the S3 subsite were shown to be specific
by the loss of pH dependence of kcat when Glul3 was mutated to
an alanine. Interactions in the S2 subsite have been shown
to be a combination of general and specific effects. In
these studies the mutation of Glu287 to a methionine residue
present in cathepsin D resulted in a decrease in specificity
for these substrates but did not result in a complete loss in
the ability to tolerate positively charged residue in the P2
position.
Cathepsin D is different from pepsin and rhizopuspepsin
in that it has a preference for hydrophobic interactions. An
example of this general overall specificity is that it has a
strong aversion to substrates containing positive charges in
the P2 position (Scarborough et al., 1993). Site-directed
mutagenesis of cathepsin D at position 287 has shown that
this strict preference for hydrophobic residues in the P2

125
position can be altered by changing Met287 to Glu present in
porcine pepsin discussed above (Scarborough et al, 1994).
Besides the unique ability of the fungal enzymes to
cleave substrates containing lysine in the Pi position,
rhizopuspepsin and other fungal aspartic proteinases exhibit
broad specificity. This ability to digest or degrade a wide
variety of synthetic substrates may confer a selective
advantage to the organism in the cleavage of proteinaceous
food sources and the invasion of host tissues.
The study of the intermolecular interactions of
rhizopuspepsin between the enzyme and the substrate has given
further insight into how a proteinase discriminates between
substrates. In particular, this study has added to the
understanding of how charged residues and hydrogen bonding
play a role in creating biological specificity.
Future directions
The observations from this study have raised additional
questions related to the specificity differences between
fungal and mammalian aspartic proteinases. These questions
may be answered by additional mutagenesis of rhizopuspepsin
to the corresponding residues in the mammalian enzymes.
Rhizopuspepsin is able to readily cleave the peptide
containing Arg in the P3 position of the substrate series
studied in Chapter 6. Porcine pepsin, on the other hand,
requires an increase in the assay pH in order for efficient

126
cleavage to be seen. Both enzymes contain a Glul3 residue.
One possible residue in rhizopuspepsin that may account for
this difference is Asnll7. The majority of the mammalian
enzymes, including porcine pepsin, contain a phenylalanine
residue at this position. A comparison of the uncomplexed
structure (Suguna et al., 1987) to many of the inhibitor
complexes of rhizopuspepsin (Suguna et al., 1987, 1992) shows
that upon binding of the inhibitor a large movement of Asnll7
occurs. Upon binding of the inhibitor the Asn residue alters
its interactions with solvent to hydrogen bond to Glul3.
Site-directed mutagenesis of Asnll7 to Phe may lead to some
insight into this phenomenon.
Experiments which may prove to be very interesting might
be those where Thr77 of porcine pepsin has been replaced with
an aspartic acid residue. This experiment would test whether
or not Pi-lysine specificity could be engineered into a
mammalian aspartic proteinase. Other experiments along these
same lines would include mutation of Ser79. These studies
may give further insight into the contribution of the serine
side chain hydrogen bonds to the cleavage of substrates
containing lysine in Pi.
Future experiments with the Asp30 and Asp77 mutants of
rhizopuspepsin described in Chapter 7 should provide
additional support for the importance of Asp77 interactions
with substrates. A comparison of the interaction of an
inhibitor containing lysine in Pi with the different mutant
enzymes may also show the requirement of Asp77 for binding.

127
The changes in energetics could be compared to those seen for
the substrate series. Analysis of the triple mutant
Asp30lle/Asp37Asn/Asp77Thr of rhizopuspepsin may also show
the importance of Asp37 in the generation of a specific
electrostatic interaction with substrates containing an
arginine in the P2' position and the general overall
preference of rhizopuspepsin for positively charged
substrates.
The results of this study have implications for the
design of anti-fungal agents. The Candida aspartic
proteinases also contain an aspartic acid at position 77.
Kinetic studies have shown that these enzymes have substrate
preferences similar to rhizopuspepsin (Fusek et al., 1994).
Preliminary evidence that the Candida enzymes possess the
Pl-lysine specificity comes from their ability to cleave
collagen and keratin (Lin et al., 1993). Future mutational
and kinetic analysis of the Candida enzymes with Pi-lysine
containing substrates and the solution of crystal structures
already in progress (Cutfield et al., 1993) are needed to
confirm the importance of Asp77 and other potential
interactions in pathogenicity that will aid in the design of
targeted therapeutics. These experiments may include: (1)
mutation of the Candida enzymes at position 77 with the
subsequent analysis of tissue invasion and substrate
specificity, (2) analysis of the ability of the wild-type and
mutant forms of rhizopuspepsin to cleave keratin and
collagen.

128
Information from this study may also prove to be useful
in the study of the paired basic residue-specific aspartic
proteinases from yeast and the pituitary (Loh et al., 1985;
Azaryan et al., 1993). Yeast aspartic proteinase 3 (YAP3)
and pro-opiomelanocortin converting enzyme have been
suggested to play critical roles in prohormone processing.
These enzymes are characterized by the ability to cleave
between or after Lys-Arg junctions.

LIST OF REFERENCES
Abad-Zapatero, C., Rydel, T. J., Neidhart, D. J., Luly, J. &
Erickson, J. W. (1991) in Structure and Function of
Aspartic Proteinases: Genetics, Structures and Mechanisms
(Dunn, B. M., Ed.) pp 9-21, Plenum Press, New York.
Ackers, G. & Smith, F. (1985) Ann. Rev. Biochem. 54, 597-629.
Al-Janabi, A., Hartsuck, J. A. & Tang, J. (1972) J. Biol.
Chem. 247, 4628-4632.
Azaryan, A., Wong, M., Friedman, T., Cawley, N., Estivariz,
F., Chen, H. & Loh, Y. (1993) J. Biol. Chem. 268, 11968-
11975.
Azuma, T., Liu, W. G., Vanderlaan, D. J., Bowcock, A. M. &
Taggart, R. T. (1992) J. Biol. Chem. 267, 1609-1614.
Balbaa, M., Cunningham, A. & Hofmann, T. (1993) Arch.
Biochem. Biophys. 306, 297-303.
Baldwin, E. T., Bhat, T. N., Gulnick, S., Hosar, M. V.,
Sowder II, R. C., Cachau, R. E., Collins, J., Silva, A.
M. & Erickson, J. W. (1993) Proc. Natl. Acad. Sci. USA
90, 6796-6800.
Blundell, T., Cooper, J., Foundling, S., Jones, D., Atrash,
B. & Szelke, M. (1987) Biochemistry 26, 5585-5590.
Boger, J., Payne, L. S., Perlow, D. S., Lohr, N. S., Poe, M.,
Blaine, E. H., Ulm, E. H., Schorn, T. W., LaMont, B. I.,
Lin, T. Y., Kawai, M., Rich, D. & Veber, D. F. (1985) J.
Med. Chem. 28, 1779-1790.
Bott, R., Subramanian, E., & Davies, D. R. (1982)
Biochemistry 21, 6956-6962.
Carter, P., Winter, G., Wilkinson, A. & Fersht, A. (1984)
Cell 38, 835-840.
Chen, G. & Yang, J. (1977) Anal. Lett. 10, 1195-1207.
Chen, K., Tao, N. & Tang, J. (1975) J. Biol. Chem. 250, 5068-
5075.
Chen, Z., Koelsch, G., Han, H. P., Wang, X. J., Lin, X. L.,
Hartsuck, J. A. & Tang, J. (1991) J. Biol. Chem. 266,
11718-11725.
Cleland, W. W. (1992) Biochemistry 31, 317-319
129

130
Cooper, J., Foundling, S., Hemmings, A., Blundell, T., Jones,
D. M., Hallett, A. & Szelke, M. (1987) FEBS Lett. 169,
215-221.
Cooper, J., Quail, W., Frazao, C., Foundling, S. I.,
Blundell, T. L., Humblet, C., Lunney, E. A., Lowther, W.
T. & Dunn, B. M. (1992) Biochemistry 31, 8142-8150.
Cutfield, S., Marshall, C., Moody, P., Sullivan, P. &
Cutfield, J. (1993) J. Mol. Biol. 234, 1266-1269.
Davies, D. R. (1990) Annu. Rev. Biophys. Biophys. Chem. 19,
189-215.
Delaney, R., Wong, R., Meng, G., Wu, N. & Tang, J. (1987) J.
Biol. Chem. 262, 1461-1467.
Devereux, J., Haeberli, P. & Smithies, O. (1984) Nucleic
Acids Res. 12, 387-395.
Dill, K. (1990) Biochemistry 29, 7133-7155.
Dunn, B. M., Kammerman, B. & McKurry, K. (1984) Anal.
Biochem. 138, 68-73.
Dunn, B. M., Jimenez, M., Parten, B. F., Valler, M. J.,
Rolph, C. E. & Kay, J. (1986) Biochem. J. 237, 899-906.
Dunn, B. M., Valler, M. J., Rolph, C. E., Foundling, S. I.,
Jimenez, M. & Kay, J. (1987) Biochem. Biophys. Acta 913,
122-130.
Evnin, L., Vasquez, J. & Craik, C. (1990) Proc. Natl. Acad.
Sci. U.S.A 87, 6659-6663.
Faust, P. F., Kornfeld, S. & Chirgwin, J. M. (1985) Proc.
Natl. Acad. Sci. U.S.A. 82, 4910-4914.
Fersht, A. (1985) in Enzyme Structure and Mechanism, 2nd ed.,
W. H. Freeman and Co., New York.
Fersht, A. (1988) Biochemistry 27, 1577-1580.
Fersht, A., Shi, J., Knill-Jones, J., Lowe, D., Wilkinson,
A., Blow, D., Brick, P., Carter, P., Waye, M. & Winter,
G. (1985) Nature 314, 235-238.
Foundling, S. I., Cooper, J., Watson, F. E., Cleasby, A.,
Pearl, L., Sibanda, B. L., Hemmings, A., Wood, S. P.,
Blundell, T. L., Valler, M., Norey, C. G., Kay, J.,
Boger, J., Dunn, B., Leckie, B. J., Jones, D. M., Atrash,
B., Hallett, A. & Szelke, M. (1987) Nature 327, 349-352.

131
Fraser, M. E., Strynadka, N. C. J., Bartlett, P. A., Hanson,
J. E. Sc James, M. N. G. (1992) Biochemistry 31, 5201-
5214.
Fruton, J. S. (1970) Adv. Enzymol. 33, 401-443.
Fruton, J. S. (1976) Adv. Enzymol. 44, 1-36.
Fukumoto, J., Tsuru, D. & Yamamoto, T. (1967) Agr. Biol.
Chem. 31, 710-717.
Fusek, M., Smith, E., Monod, M. & Foundling, S. (1993) FEBS
Lett. 327, 108-112.
Fusek, M., Smith, E., Monod, M., Dunn, B. & Foundling, S.
(1994) Biochemistry in press.
Graf, L., Jansco, A., Szilagyi, L., Hegyi, G., Pinter, K.,
Naray-Szabo, G., Hepp, J., Medzihradszky, K. & Rutter, W.
(1988) Proc. Natl. Acad. Sci. U.S.A. 85, 4961-4965.
Graham, J., Sodek, J.
51, 789-796.
Sc Ho fmann, T.
(1973)
Can.
J.
Biochem.
Grippon, J., Phee, S.
55, 504-506.
& Hofmann, T.
(1977)
Can.
J.
Biochem.
Hedstrom, L., Szilagyi, L. & Rutter, W. J. (1992) Science
255, 1249-1253.
Henderson, P. (1972) Biochem. J. 127, 321-333.
Ho, S. N., Hunt, H. D., Horton, R. M., Pullen J. K. & Pease,
L. R. (1989) Gene 77, 51-59.
Hobart, P., Fogliano, M., O'Connor, B., Schaefer, I. &
Chirgwin, J. (1984) Proc. Natl. Acad. Sci. U.S.A. 81,
5026-5030.
Hofmann, T. & Hodges, R. (1982) Biochem. J. 203, 603-610.
Hofmann, T., Allen, B., Bendiner, M., Blum, M. & Cunningham,
A. (1988) Biochemistry 27, 1140-1146.
Hofmann, T., Hodges, R. S. & James, M. N. G. (1984)
Biochemistry 23, 635-643.
Horton, N. & Lewis, M. (1992) Protein Sci. 1, 169-181.
Hube, B., Turver, C., Odds, F., Eiffert, H., Boulnois, G.,
Kochel, H. & Ruchel, R. (1991) J. Med. Vet. Mycology 29,
129-132.

132
Hurley, J. H., Baase, W. A. & Matthews, B. w. (1992) J. Mol.
Biol. 224, 1143-1159.
Hyland, L. J., Tomasek, T. A., Jr. & Meek, T. D. (1991a)
Biochemistry 30, 8454-8463.
Hyland, L. J., Tomasek, T. A., Jr., Roberts, G. D., Carr, S.
A., Maggard, V. W., Bryan, H. L., Fakhoury, S. A., Moore
M. L., Minnich, M. D., Culp, J. F., DesJarlais, R. L. &
Meek, T. D. (1991b) Biochemistry 30, 8441-8456.
Jackson, S., Moracci, M., elMarsy, Johnson, C. & Fersht, A.
(1993) Biochemistry 32, 11259-11269.
James, M. N. G.& Sielecki, A. R. (1985) Biochemistry 24,
3701-37013.
James, M., Sielecki, A. & Hofmann, T. (1985) in Aspartic
Proteinases and Their Inhibitors (Kostka, V., Ed.) pp
163-177, Walter de Gruyter, Berlin.
James, M. N. G., Sielecki, A. R., Hayakawa, K. & Gelb, M. H.
(1992) Biochemistry 31, 3872-3886.
Jencks, W. (1969) in Catalysis in Chemistry and Biology,
McGraw-Hill, New York.
Katz, B. & Kossiakoff, A. (1990) Proteins: Struc. Func.
Genet. 7, 343-357.
Kempf, D. J., Norbeck.D.W., Codacovi, L., Wang, X. C.,
Kohlbrenner, W. E., Wideburg, N. E., Paul, D. A., Knigge
M. F., Vasavanonda, S., Craig-Kinnard, A., Saldivar, A.,
Rosenbrook, J., W., Clement, J. J., Plattner, J. J. &
Erickson, J. (1990) J. Med. Chem. 33, 2687-2689.
Kho, C. & Zarbl, H. (1992) BioTechniques 12, 228-230.
Krieger, M., Kay, L. M. & Stroud, R. M. (1974) J. Mol. Biol.
83, 209-230.
Leatherbarrow, R., Fersht, A. & Winter, G. (1985) Proc. Natl
Acad. Sci. U.S.A. 82, 7840-7844.
Leatherbarrow, R. (1987). Enzfitter, a non-linear regression
analysis program for the IBM PC, Elsevier-BIOSOFT,
Amsterdam.
Lin, X., Wong, R. & Tang, J. (1989) J. Biol. Chem. 264, 4482
4489.
Lin, X., Tang, J., Koelsch, G., Monod, M. & Foundling, S.
(1993) J. Biol. Chem. 268, 20143-20147.

133
Loh, Y., Parish, D. & Tuteja, R. (1985) J. Biol. Chem. 260,
7194-7205.
Lowe, D., Winter, G. & Fersht, A. (1987) Biochemistry 26,
6038-6043.
Lowther, W. T., Chen, Z., Lin, X., Tang, J. and Dunn, B. M.
(1991) in Structure and Function of the Aspartic
Proteinases: Genetics, Structure and Mechanisms. (Dunn,
B. M., Ed.) pp 275-279, Plenum Press, New York.
Lunney, E., Hamilton, H., Hodges, J., Kaltenbronn, J.,
Repine, J., Badasso, M., Cooper, J., Dealwis, C.,
Wallace, B., Lowther, W. T., Dunn, B. M. & Humblet, C.
(1993) J. Med. Chem. 36, 3809-3820.
Marquardt, D. (1963) J. Soc. Indust. Appl. Math. 11, 431-441.
Medzihradszky, K., Voynick, I., Medzihradszky-Schweiger, H. &
Fruton, J. (1970) Biochemistry 9, 1154-1162.
Metcalf, P. & Fusek, M. (1993) EMBO J. 12, 1293-1302.
Meyer, S. L. (1975) in Data Analysis for Scientists and
Engineers, John Wiley and Sons, Inc., New York.
Mildvan, A. S., Weber, D. J. & Kuliopulos, A. (1992) Arch.
Biochem. Biophys. 294, 327-340.
Morihara, K. & Oka, T. (1973) Arch. Biochem. Biophys. 157,
561-572.
Morrison, C., Hurst, S., Bragg, S., Kuyendall, R., Diaz, H.,
McLaughlin, D. & Reiss, E. (1993) J. Gen. Microbiol. 139,
1177-1186.
Pace, C. (1986) Methods Enzymol. 131, 266-280.
Parris, K. D., Hoover, D. J., Damon, D. B. & Davies, D. R.
(1992) Biochemistry 31, 8125-8141.
Paya, C. (1993) Clin. Infect. Dis. 16, 677-688.
Pearl, L. (1985) in Aspartic Proteinases and Their Inhibitors
(Kostka, V., Ed.) pp 189-195, Walter de Gruyter, Berlin.
Pearl, L. & Blundell, T. (1984) FEBS Lett. 174, 96-101.
Perona, J. J., Hedstrom, L., Wagner, R., Rutter, W., Craik,
C. S. Sc Fletterick, R. J. (1994) Biochemistry 33, 3252-
3259.
Perona, J. J., Tsu, C. A., Craik, C. S. & Fletterick, R. J.
(1993) J. Mol. Biol. 230, 919-933.

134
Pohl, J. & Dunn, B. M. (1988) Biochem. 27, 4827-34.
Powers, J. C., Harley, A. D. & Myers, D. V. (1977) Adv. Exp.
Med. Biol. 95, 141-157.
Rao, C., Scarborough, P. E., Kay, J., Batley, B., Rapundalo,
S., Klutchko, S., Taylor, M. D., Lunney, E. A., Humblet,
C. C. & Dunn, B. M. (1993) J. Med. Chem. 36, 2614-2620.
Rheinnecker, M., Baker, G., Eder, J. Sc Fersht, A. R. (1993)
Biochemistry 32, 1199-1203.
Roberts, N., Martin, J., Kinchington, D., Broadhurst, A.,
Craig, J., Duncan, I., Galpin, S., Handa, B., Kay, J.,
Krohn, A., Lambert, R., Merrett, J., Mills, J., Parkes,
K., Redshaw, S., Ritchie, A., Taylor, D., Thomas, G. &
Machin, P. (1990) Science 248, 358-361.
Rodriquez, E. J., Angeles, T. S. & Meek, T. D. (1993)
Biochemistry 32, 12380-12385.
Rosenberg, S., Woods, K., Sham, H., Kleinert, H., Martin, D.
Stein, H., Cohen, J., Egan, D., Boop, B., Merits, I.,
Garren, K., Hoffman, D. & Plattner, J. (1990) J. Med.
Chem. 33, 1962-1969.
Ruhlmann, A., Kukla, D., Schwager, P., Bartels, K. & Huber,
R. (1973) J-. Mol. Biol. 77, 417-436.
Sali, A., Veerapandian, B., Cooper, J. B., Foundling, S. I.,
Hoover, D. J. Sc Blundell, T. L. (1989) EMBO J. 8, 2179-
88.
Salituro, F. C., Agarwal, N., Hofmann, T. Sc Rich, D. H.
(1987) J. Med. Chem. 30, 286-295.
Samaranayake, L. & Holmstrup, P. (1989) J. Oral Pathol. Med.
18, 554-64.
Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) in
Molecular Cloning, Cold Spring Harbor Laboratory Press,,
New York.
Sampath-Kumar, P. S. & Fruton, J. S. (1974) Proc. Natl. Acad
Sci. U.S.A. 71:4, 1070-1072.
Saral, R. (1991) Rev. Infect. Dis. 13, 487-492.
Sawyer, T. K., Hester, J., Schostarez, H., Thaisrivongs, S.,
Bundy, G., Liu, L., Bradford, V., DeVaux, A., Staples, D
J., Maggiora, L. L., TenBrink, R., Kinner, J., Smith, C.
W., Pals, D. T., Couch, S., Hinzmann, J., Poorman, R. A.
Einspahr, H., Finzel, B., Watenbaugh, K., Mao, B., Epps,

135
D., Kezdy, F. & Heinrickson, R. L. (1991) in Structure
and Function of the Aspartic Proteinases (Dunn, B. M.,
Ed.) pp 307-323, Plenum Press, New York.
Sawyer, T. K., Maggiora, L. L., Liu, L., Staples, D. J.,
Bradford, V. S., Mao, B., Pals, D. T., Dunn, B. M.,
Poorman, R. A., Hinzman, J., deVaux, A. E., Affholter, J.
A.& Smith, C. W. (1990) in Peptides: Chemistry, Structure
and Biology, Proceedings of the Eleventh American Peptide
Symposium (Rivier, J. E. & Marshall, G. R., Eds.) pp 46-
48, ESCOM Science Publishers, Leiden.
Sawyer, T. K., Staples, D. J., Liu, L., Tomasselli, A. G.,
Hui, J. 0., O'Connell, K., Schostarez, H., Hester, J. B.
Moon, J., Howe, W. J., Smith, C., Decamp, D. L., Craik,
C. S., Dunn, B. M., Lowther, W. T., Harris, J., Poorman,
R. A., Wlodower, A., Jaskolski, M. & Heinrickson, R. L.
(1992) Int. J. Peptide Protein Res. 40, 274-281.
Scarborough, P. E., Guruprasad, K., Topham, C., Richo, G. R.,
Conner, G. E., Blundell, T. L. & Dunn, B. M. (1993)
Protein Sci. 2, 264-276.
Scarborough, P. E. & Dunn, B. M. (1994) Protein Eng. 7,
Schgger, H. & von Jagow, G. (1987) Anal. Biochem. 166, 368-
379.
Schechter, I. & Berger, A. (1967) Biochem. Biophys. Res.
Comm. 27, 157-162.
Schellman, J. A. (1978) Biopolymers 17, 1305-1322.
Schoemaker, J., Brasnett, A. & Marston, F. (1985) EMBO J. 4,
775-780.
Sepulveda, P., Jackson, K. & Tang, J. (1975) Biochem.
Biophys. Res. Comm. 63, 1106-1112.
Shortle, D. & Meeker, A. (1986) Proteins: Struct. Func.
Genet. 1, 81-89.
Shortle, D. & Meeker, A. (1989) Biochemistry 28, 936-944.
Shortle, D. (1992) Q. Rev. Biophys. 25, 205-250.
Sogawa, K., Fujii-Kuriyama, Y., Mizukami, Y., Ichihara, Y. &
Takahashi, K. (1983) J. Biol. Chem. 258, 5306-5311.
Spatola, A. (1983) in Chemistry and Biochemistry of Amino
Acids, Peptides and Proteins, Dekker, New York.
Strop, P., Sedlacek, J., Stys, J., Kaderabkova, Z., Blaha,
I., Pavlickova, L., Pohl, J., Fabry, M., Kostka, V.,

136
Newman, M., Frazao, C., Shearer, A., Tickle, I. J. &
Blundell, T. L. (1990) Biochemistry 29, 9863-9871.
Studier, F. W., Rosenberg, A. H., Dunn, J. J. & Dubendorf, J.
W. (1990) Meth. Enzymol. 185, 60-89.
Subramanian, E., Swan, I., Liu, M., Davies, D., Jenkins, J.,
Tickle, I. & Blundell, T. (1977) Proc. Natl. Acad. Sci.
USA 77, 556-559.
Sugrue, R., Marston, F., Lowe, P. & Freedman, R. (1990)
Biochein. J. 271, 541-547.
Suguna, K., Bott, R. R., Padlan, E. A., Subramanian, E.,
Sheriff, S., Cohen, G. H. & Davies, D. R. (1987a) J. Mol.
Biol. 196, 877-900.
Suguna, K., Padlan, E. A., Smith, C., Carlson, W. & Davies,
D. R. (1987b) Proc. Natl. Acad. Sci. U.S.A. 84, 7009-
7013.
Suguna, K., Padlan, E. A., Bott, R., Boger, J. & Davies, D.
R. (1992) Proteins: Struct. Func. Genet. 13, 195-205.
Suzuki, J., Sasaki, K., Sasao, Y., Hamu, A., Kawasaki, H.,
Nishiyama, M., Horinouchi, S. & Beppu, T. (1989) Protein
Eng. 2, 563-569.
Swain, A. L., Gustchina, A. & Wlodawer, A. (1991) in
Structure and Function of the Aspartic Proteinases:
Genetics, Structures and Mechanisms (Dunn, B. M., Ed.) pp
433-441, Plenum Press, New York.
Swain, A. L., Miller, M. M., Green, J., Rich, D. H.,
Schneider, J., Kent, S. B. & Wlodawer, A. (1990) Proc.
Natl. Acad. Sci. U.S.A. 87, 8805-8809.
Szelke, M., Jones, D., Atrash, B., Hallet, A. & Leckie, B.
(1980) in Peptides: Structure and Function, Proceedings
of the Eighth American Peptide Symposium (Hruby, V. J. &
Rich, D. H., Eds.) pp 579-582, Pierce Chemical,
Rockford,IL.
Takahashi, K. (1987) J. Biol. Chem. 262, 1468-1478.
Takahashi, K. (1988) J. Biochem. 103, 162-167.
Taylor, G., Hoare, M., Gray, D. & Marston, F. (1986)
Biotechnology 4, 553-557.
Thompson, W. J., Fitzgerald, P. M. D., Holloway, M. K.,
Emini, E. A., Darke, P. L., McKeever, B. M., Schleif, W.
A., Quintero, J. C., Zugay, J. A., Tucker, T. J.,

137
Schwering, J. E., Homnick, C. F., Nunberg, J., Springer,
J. P. & Huff, J. R. (1992) J". Med. Chem. 35, 1685-1701.
Togni, G., Sanglard, D., Falchetto, R. & Monod, M. (1991)
FEBS Lett. 286, 181-185.
Veerapandian, B., Cooper, J. B., Sali, A. & Blundell, T. L.
(1990) J. Mol. Biol. 216, 1017.
de Viragh, P. A., Sanglard, D., Togni, G., Falchetto, R. &
Monod, M. (1993) J. Gen. Microbiol. 139, 335-342.
Voynick, I. & Fruton, J. (1971) Proc. Nat. Acad. Sci. U.S.A.
68, 257-259.
Wells, J. A. (1990) Biochemistry. 29, 8509-8517.
Wiley, R. Sc Rich, D. (1993) Med. Res. Rev. 13, 327-384.
Young, L., Jernigan, R. & Coveil, D. (1994) Protein Sci. 3,
717-729.
Young, S. D., Payne, L. S., Thompson, W. J., Gaffin, N.,
Lyle, T. A., Britcher, S. F., Graham, S. L., Schultz, T.
H., Deana, A. A., Darke, P., Zugay, J., Schleif, W. A.,
Quintero, J. C., Emini, E. A., Anderson, P. S. & Huff, J.
R. (1992) J. Med. Chem. 35, 1702-1709.

BIOGRAPHICAL SKETCH
William Todd Lowther was born on February 18, 1967. His
interest in science began at an early age with his
fascination of animals and plants. He credits his desire for
learning to the encouragement of his parents. While in
junior and senior high school he also became enamored with
music and playing the clarinet. He graduated from Apopka
High School in 1985. He attended Stetson University,
majoring in chemistry while on a music scholarship. His
professors were a source of wisdom and friendship. Their
support and passion for teaching lead to an interest in
pursuing research. The summer before his senior year he
participated in a research project at the Inhalation
Toxicology Research Institute in Albuquerque, New Mexico.
This experience confirmed Todd's desire to pursue graduate
studies and research in biochemistry. He graduated from
Stetson in 1989. Todd began his graduate studies in the
laboratory of Professor Ben Dunn in 1989. Dr. Dunn provided
a stimulating environment for furthering Todd's interest in
protein-ligand interactions and protein engineering. Upon
graduation from the University of Florida, Todd plans to
continue his study of proteins and their specific
interactions with ligands.
138

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy^
^ '~yy\
Ben M. Dunn, Chairman
Professor of Biochemistry
and Molecular Biology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
a dissertation for the degree of Docto]
as
.ori Philosophy.
Charles Allen
Professor of Biochemistry
and Molecular Biology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and jqCiiiXity, as
a dissertation for the degree of Doctdr^of Phil<
JJwA<
DanielPurich
Professor of Biochemistry
and Molecular Biology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
a dissertation for the degree of Doctor off* Philosophy.
as
Sheldon Schuster
Professor of Biochemistry
and Molecular Biology

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor /of
Nigel Richards-"^
Assistant Professor of
Chemistry
This dissertation was submitted to the Graduate Faculty
of the College of Medicine and to the Graduate School and was
accepted as partial fulfillment of the requirements for the
degree of Doctor of Philosophy.
August, 1994
Dean, College of Medicine
Dean, Graduate School




This work is dedicated to all those who have given freely of
their love, support, guidance and patience; my Lord and
Savior Jesus Christ, family, friends and colleagues.


57
Tang and coworkers have shown that intermolecular activation
of rhizopuspepsinogen and porcine pepsinogen occurs
principally at pHs above 3.0 (Al-Janabi et al., 1972; Lin et
al., 1989; Chen, et al., 1991). Protein concentrations were
varied from 0.1 to 1.5 mg/ml. Activation to the intermediate
form of rhizopuspepsin was not observed until 6 hours of
incubation at 0.1 mg/ml. An increase in the protein
concentration above this level resulted in degradation of the
protein. The 9 amino acid extended form of the proteins were
used for all kinetic and structural comparisons. A
discussion on the potential effect of this extension is given
in Chapter 4.
The activation of the D77T and the D30I/D77T proteins at
pH 2, however, produced a mixed population of N-termini: the
9 amino acid extension and a 15 amino acid extension, Asn-
Lys-His-Lys-Ile-Asn-Thr-Ser-Thr-Gly-Gly-Ile-Val-Pro-Asp-.
These populations can readily be seen with IEF gel analysis.
Figure 3-7 shows the IEF gel of all the recombinant
rhizopuspepsins following activation at pH 2 prior to ion
exchange chromatography. The WT-REC and the Asp30lle
proteins have a pi of 5.8. The Asp77Thr and
Asp30lle/Asp77Thr mutants show bands at pHs 6.1 and 6.9. The
yield ratios from the sequencing analysis suggest that the pH
6.9 band corresponds to the 15 amino acid extension
activation intermediate. This is not suprising since this
extension contains three additional positively charged
residues making the protein more basic.


Figure 3-3. Gel filtration elution profile of refolded
Asp30lle rhizopuspepsinogen. Peak 1, polymeric material at
the void volume; peak 2, rhizopuspepsinogen; peak 3, low
molecular weight contaminants.


31
denaturation, reduction and dialysis (Suzuki et al., 1989).
The purified inclusion bodies were dissolved in freshly
deionized 8 M urea, 50 mM CAPS pH 10.5, 1 mM EDTA, 1 mM
glycine, 500 mM NaCl and 300 mM |3-mercaptoethanol to a final
concentration of approximately 1 mg(wet)/ml. After stirring
at room temperature for one hour, the solution was
centrifuged at 24,000 x g for 30 minutes to remove
undissolved material. The supernatant was dialyzed for one
hour at room temperature against five times the original
volume in SpectraPor 1 (MWCO 6-8 kDa) membranes and 50 mM
Tris-HCl pH 11.0 buffer. Following a buffer change and
dialysis at room temperature for another hour, the dialysis
buffer was changed to 50 mM Tris-HCl pH 7.5 and dialysis
continued overnight at 4C. The next morning the buffer was
changed to 50 mM MOPS pH 7.0 and dialyzed for at least 6 more
hours at 4C. The resulting solution was centrifuged at
24,000 x g for 30 minutes to remove precipitates and
concentrated using a Minitan Ultrafiltration system outfitted
with low protein binding, PLTK, 10,000 MWCO membrane plates
(Millipore) and an Amicon pressurized cell with YM10
membranes (10,000 MWCO).
Size-exclusion Chromatography
The zymogen was further processed by centrifugation at
45,000 x g for thirty minutes before loading onto a 2.5 cm x
90 cm Sephacryl S300 gel filtration column equilibrated with


39
Product, analysis. The fidelity of the cleavage sites
was verified by HPLC and capillary electrophoresis (CE). All
substrates were incubated with enzyme at 37C overnight. The
cleavage products of the substrates based on the parent
peptide Lys-Pro-Ala-Lys-Phe*Nph-Arg-Leu, discussed in Chapter
6, were analyzed by reverse phase HPLC using a Waters C-18
Radial-pack column with a gradient from 10 to 80%
acetonitrile in water containing 0.1% TFA in 45 minutes. The
peaks were collected and hydrolyzed by the addition of 6 N
HCl. The composition was determined by amino acid analysis
by the Protein Chemistry Core Facility.
The cleavage products of the peptides used for the
analysis of mutants of rhizopuspepsin (Chapter 7) were
determined on a BioRad BioFocus 3000 capillary
electrophoresis system. The samples at 4C were pressure
injected for 20 psi*sec onto a BioRad 24 cm x 25 Jim cartridge
maintained at 15C. This type of injection procedure ensures
reproducible loading of the sample onto the capillary. The
capillary was equilibrated with 0.5 M sodium phosphate pH
2.5, electrophoresed at a constant 8 kv in the +-
direction and monitored at 200 nm. All the peptides studied,
if cleaved properly, will have the same C-terminal product,
Nph-Arg-Ala. This product was purified on a Rainin HPLC
system using a 4.6 mm X 25 cm Dynamax-300 C-18 column with a
gradient from 0 to 10% B acetonitrile in water containing
0.1% TFA in 32 minutes at a flow rate of 1.1 ml/min. The
composition of the fraction was confirmed by amino acid


131
Fraser, M. E., Strynadka, N. C. J., Bartlett, P. A., Hanson,
J. E. Sc James, M. N. G. (1992) Biochemistry 31, 5201-
5214.
Fruton, J. S. (1970) Adv. Enzymol. 33, 401-443.
Fruton, J. S. (1976) Adv. Enzymol. 44, 1-36.
Fukumoto, J., Tsuru, D. & Yamamoto, T. (1967) Agr. Biol.
Chem. 31, 710-717.
Fusek, M., Smith, E., Monod, M. & Foundling, S. (1993) FEBS
Lett. 327, 108-112.
Fusek, M., Smith, E., Monod, M., Dunn, B. & Foundling, S.
(1994) Biochemistry in press.
Graf, L., Jansco, A., Szilagyi, L., Hegyi, G., Pinter, K.,
Naray-Szabo, G., Hepp, J., Medzihradszky, K. & Rutter, W.
(1988) Proc. Natl. Acad. Sci. U.S.A. 85, 4961-4965.
Graham, J., Sodek, J.
51, 789-796.
Sc Ho fmann, T.
(1973)
Can.
J.
Biochem.
Grippon, J., Phee, S.
55, 504-506.
& Hofmann, T.
(1977)
Can.
J.
Biochem.
Hedstrom, L., Szilagyi, L. & Rutter, W. J. (1992) Science
255, 1249-1253.
Henderson, P. (1972) Biochem. J. 127, 321-333.
Ho, S. N., Hunt, H. D., Horton, R. M., Pullen J. K. & Pease,
L. R. (1989) Gene 77, 51-59.
Hobart, P., Fogliano, M., O'Connor, B., Schaefer, I. &
Chirgwin, J. (1984) Proc. Natl. Acad. Sci. U.S.A. 81,
5026-5030.
Hofmann, T. & Hodges, R. (1982) Biochem. J. 203, 603-610.
Hofmann, T., Allen, B., Bendiner, M., Blum, M. & Cunningham,
A. (1988) Biochemistry 27, 1140-1146.
Hofmann, T., Hodges, R. S. & James, M. N. G. (1984)
Biochemistry 23, 635-643.
Horton, N. & Lewis, M. (1992) Protein Sci. 1, 169-181.
Hube, B., Turver, C., Odds, F., Eiffert, H., Boulnois, G.,
Kochel, H. & Ruchel, R. (1991) J. Med. Vet. Mycology 29,
129-132.


98
peptide series, but is comparable to the characteristics of
the Candida aspartic proteinases.
Comparison to the Mammalian Enzymes
Studies with the same peptide series described above
have shown that porcine pepsin, cathepsin E and cathepsin D
have different specificities toward substrates than those
seen for rhizopuspepsin. These differences center around
interactions in the S3 and S2 subsites.
Porcine pepsin is able to cleave substrates containing
basic residues in P2, but is not able to accommodate these
residues in P3 at pH 3.5 (Rao-Naik, unpublished results, Pohl
& Dunn, 1988). Upon raising the pH, however, a dramatic
increase in kcat is seen for the peptide containing Arg in P3.
This observation has been shown by site-directed mutagenesis
to be mediated by Glul3 (Rao-Naik, unpublished results).
Rhizopuspepsin also contains a glutamic acid at position 13
but does not show the same pH dependence.
Cathepsin D has been shown not to accommodate basic
residues in P2 and to have a preference for larger
hydrophobic residues in P3 (Scarborough et al., 1993). Site-
directed mutagenesis has shown that the P2 effect is mediated
by Met287. Upon changing Met287 to Glu, the residue present
in porcine pepsin, basic substitutions were tolerated and
improved cleavage was seen.


I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor /of
Nigel Richards-"^
Assistant Professor of
Chemistry
This dissertation was submitted to the Graduate Faculty
of the College of Medicine and to the Graduate School and was
accepted as partial fulfillment of the requirements for the
degree of Doctor of Philosophy.
August, 1994
Dean, College of Medicine
Dean, Graduate School


89
A comparison of 16 to the classical aspartic proteinase
inhibitors (17 and 18) containing the statine derivative
indicates the relative binding of isosteres to WT-REC in this
study.
Xaa'P [CH (OH) CH2] Yaa > statine >
Phe'F[CH2NH]Yaa Cha'F[CH2NH] Yaa
The hydroxyethylene and methyleneamino derivatives are better
models for substrate interactions because of the frame shift
seen in pepstatin complexes caused by the addition of two
extra main chain atoms and the loss of hydrophobic
interactions in the Si pocket (Szelke et al., 1980; Sawyer et
al., 1990).
In conclusion, the results from the x-ray crystallo
graphic analysis can be used to understand subtle
interactions in the active site which can be exploited for
drug design. It is important to remember, however, that
there will always be a balance between potency, selectivity
and bioavailability.
This analysis has yielded several potent inhibitors of
rhizopuspepsin which can be used to titrate the different
forms of rhizopuspepsin used to analyze and understand
substrate specificity (Chapter 6 and 7). An active site
titration of the enzyme stock solutions enables an accurate
calculation of the kinetic parameter kcat- U85548E (11) was


35
least one hour prior to spectroscopic measurements in 1 cm
cuvettes.
Denaturation curve analysis. The denaturation curves
were analyzed assuming a two-state model where only the
native and the denatured states are populated. The
fluorescence values for the native and unfolded states, FN and
Fy can be used to determine the equilibrium constant for
unfolding, KU_F, and the free energy of unfolding, AGV_F, at
different denaturant concentrations by using equation 1.
*u-f = (F* ~ F) / (F Fy) = exp(-AGv_v / RT) (1)
F is the observed fluorescence, R is the gas constant (1.987
cal mol-1 K_1) and T is the absolute temperature. The free
energy of unfolding of proteins has been shown to be linearly
dependent on the denaturant concentration as expressed by
equation 2 (Pace, 1986).
AGy_F = AGy.2 ^[denaturant] (2)
AG_f and AGy2 are the free energy of unfolding at
denaturant concentration, D, and in water, respectively. m
is the slope of the transition and is thought to related to
the difference in the degree of accessible surface area
between the native and unfolded states (Schellman, 1978).
Two methods have been used to calculate the transition point,


119
Figure 7-1. Hydrogen bonding interactions in penicillopepsin
between Asp77, Ser79 and the pepstatin derivative containing
lysine in the Pi position. Yellow dotted lines indicate
hydrogen bond distances in angstroms.


24
of Technology, Japan. The native isozymes, pi 5 and pi 6,
were a gift from Kevin Parris and David Davies at the
Laboratory of Molecular Biology, National Institutes of
Health. The porcine pepsin used for comparison to the
rhizopuspepsins was from Sigma. The synthetic
oligonucleotides were synthesized by the University of
Florida, Interdisciplinary Center for Biotechnology Research
(ICBR) DNA Synthesis Core Facility using an Applied
Biosystems 394 DNA synthesizer. Peptide substrates were
synthesized by the ICBR Protein Chemistry Core Facility using
an Applied Biosystems 430A peptide synthesizer. The
oligonucleotides were used directly for mutagenesis and
sequencing reactions. All peptides were shown to be >95%
pure by reverse phase HPLC and capillary electrophoresis.
Stock solutions of the peptides and the inhibitor U85548E
were quantified by amino acid analysis on a Beckman System
6300 high performance amino acid analyzer following acid
hydrolysis. The N-terminal sequence analyses of the
rhizopuspepsins were performed on Applied Biosystems 470A and
473A protein sequencers. The activated enzymes were analyzed
by matrix-assisted laser desorption ionization time of flight
(MALDI-TOF) mass spectroscopic analysis on a Vestec (Houston,
TX) LaserTec Bench Top System. All other materials were of
the highest commercial grade.


70
Circular dichroism. Figure 4-2 shows the CD spectra of
the recombinant rhizopuspepsins in the far-UV region (200-250
nm). Even though there are slight wavelength shifts in the
spectra, the mutants as a whole are, within experimental
error, structurally similar to each other and the wild-type
enzyme. These shifts may be due to slight changes in the (X-
helix/(3-sheet ratios. The differences seen between the
proteins from 200 to 205 nm cannot be considered to be
significant because of the high degree of signal fluctuation
in this region on the instrument used.
Fluorescence spectroscopy. The denaturation of the
rhizopuspepsins was followed by the change in intrinsic
fluorescence at 350 nm using an excitation wavelength of 280
nm (Figure 4-3). Upon the addition of sufficient guanidinium
hydrochloride to cause unfolding, the fluorescence signal
shifted to longer wavelengths with a 75% decrease in
intensity. Figure 4-4 shows the normalized denaturation
curves for the recombinant rhizopuspepsins and the isozyme
pi 6. Analysis of the transition curves by the method of
Jackson (1993) is presented in Table 4-2. The pi 6 isozyme
and wild-type recombinant proteins exhibit unfolding
parameters which are indistinguishable from each other. The
decrease in [Gdn-HCl]^ of the two single mutants suggests a
slight decrease in stability from the wild-type enzymes, WT-
REC and isozyme pi 6. Further loss of stability is seen in
the double mutant.


62
Figure 3-9. IEF comparison of the activated, purified
recombinant rhizopuspepsins. Lane 1, WT-REC, lane 2,
Asp30lle; lane 3, AspWThr; lane 4, Asp30lle/Asp77Thr.


28
either directly into the TA cloning kit vector pCR
(Invitrogen) or after restriction digest with Ndel and BamHI
and gel purification into the pGEM vector (Promega) for DNA
sequence analysis. In order to confirm the presence of the
desired mutation and to ensure that no spurious mutations
occurred during the polymerization process, the entire 1136
bp coding region was dideoxy-sequenced according to the
Sequenase 2.0 Kit protocol using 5 (Ig of plasmid template
(United States Biochemical) and deoxyadenosine-5' [cc-
35S]thiotriphosphate with the insertion of one extra step.
Before the addition of the stop solution, more reaction
buffer, dNTPs, and terminal deoxynucleotidyl transferase
(TdT) were added in order to extend prematurely terminated
products resulting from high GC content and secondary
structure (Kho & Zarbl, 1992). The mutant genes were
transferred to a modified version of the pET3a expression
system vector. The pET3a vector was modified by removing a
375 bp fragment between the BamHI restriction site and
ampicillin resistance gene in order to generate a vector
which does not contain an EcoRV site. This was performed by
digesting the vector with EcoRl and EcoRV and by subsequently
treating with Klenow polymerase and blunt-end ligation. The
resulting vector is shown in Figure 2-2. The use of this new
vector allowed the efficient screening of recombinant clones
because of the unique EcoRV restriction site within the
rhizopuspepsinogen gene.


27
C3'; Asp77 -Thr; 5'GAT ACC GCT AGC AGA AGA GCC AGT ACC ATA
AG3'). The underlined bases indicate the engineered
restriction sites or the differences from wild-type. The
second reaction generated the 3'-end of the gene (B) by using
a sense primer (3, Asp30 Ile; 5'TTT GAT ACC GGT TCC TCC GAT
TTA TG3' ; Asp77 -4Thr; 5'TCT TCT GCT AGC GGT ATC TTG GC3' )
capable of annealing to the 3' end of the PCR product above
(A) and an antisense primer (4) containing an engineered
BamHI site (5'GGA TTC TTA TTG AGC GAC AGG AGC G3'). The
cycling conditions for the first round of PCR were as
follows: (1) 3 cycles; 96C for 40 sec, 50C for 40 sec, 72C
for 2 minutes, (2) 25 cycles; 94C for 40 sec, 50C for
40 sec, 72C for 2 minutes, (3) 72C for 7 minutes. These
PCR products (A and B) were purified on Seaplaque GTG or
NuSieve GTG low melting agarose gels (FMC Bioproducts). In
order to generate the full length gene, small amounts of each
band were mixed together and heated to 100C for 5 minutes,
and then placed on ice. This step is crucial for insuring
that the fragments are completely dissociated so they can
form hybrid templates in the next PCR reaction (C and D).
After the addition of polymerase and more of the outer
primers (1 and 4, 80 pmol each), the second round of PCR was
performed as follows: (1) 25 cycles; 94C for 40 sec, 55C
for 1 minute, 72C for 2 minutes, (2) 72C for 7 minutes.
The double mutant was generated by repeating this procedure
using the Asp30lle mutant rhizopuspepsinogen gene as the
starting template. The resulting products (E) were ligated


99
Cathepsin E exhibits characteristics which make it
similar to porcine pepsin, cathepsin D and rhizopuspepsin.
Cathepsin E and porcine pepsin are alike in that they are
able to cleave substrates with a positive charge in P3.
Cathepsin E is able to cleave substrates with Lys and Arg
residues in P2, but it prefers hydrophobic substitutions as
seen for cathepsin D. Interestingly, cathepsin E is not able
to cleave the His substitution in this position. This trend
is similar to that seen for rhizopuspepsin.
Comparison to the Candida Aspartic Proteinases
Three Candida aspartic proteinases have been analyzed
with the same family of peptide substrates: enzymes isolated
from Candida albicans (CAAP), Candida parapsilosis (CPAP) and
Candida tropicalis (CTAP) (Fusek et al., 1994). These
enzymes are similar to rhizopuspepsin in that they exhibit
broad specificity. The majority of the substrates are
cleaved efficiently. The Candida enzymes are able to cleave
substrates containing basic substitutions in P3 and P2. They
also show a similar aversion for Asp in P2'. The most
notable difference of the Candida enzymes from rhizopuspepsin
is their inability to cleave the substrate with leucine
substituted in P4. The CTAP enzyme also shows a unique
difference from rhizopuspepsin and the other Candida
proteinases in that is also unable to cleave the substrate
with Glu substituted in P2. An understanding of the possible


8
The extended cleft of the active site is formed by the
interaction of the two domains (Figure 1-1). Each domain
also contributes one catalytic aspartic acid at the bottom of
the active site. An elaborate network of hydrogen bonds
maintains these aspartic acid residues (Asp32 and Asp215,
porcine pepsin numbering) in a juxtaposed or opposing
orientation. A centrally located water molecule, hydrogen-
bonded to each aspartic acid residue, is thought to act as
the nucleophile in a base-catalyzed attack of the scissile
bond carbonyl of the substrate (Suguna et al., 1987; Fraser
et al., 1992; James et al., 1992).
From the examination of inhibitor complexes of
rhizopuspepsin (Table 1-1), porcine pepsin (Abad-Zapatero et
al., 1991), endothiapepsin (Veerapandian et al., 1990; Lunney
et al., 1993), cathepsin D (Baldwin et al., 1993; Metcalf &
Fusek, 1993) and the HIV proteinase (Swain et al., 1990,
1991), it is evident that there is a consistent binding mode
for ligands (Figure 1-2). Ligands seven to eight residues
long completely fill the active site in an extended |3-strand
conformation with the amino acid side chains alternating in a
regular fashion. This uniform binding or anchoring of
ligands to the active site is attributed to a highly
conserved hydrogen bonding network, between the enzyme and
the a-carbon backbone of ligands, and the preference for
large hydrophobic or aromatic substituents on either side of
the scissile bond (Pi-Pi1) or the site of cleavage.


32
50 mM MOPS pH 7.0 containing 300 mM NaCl. The zymogen was
eluted at a flow rate of 25 ml/hr and the fractions showing
the highest purity were concentrated and buffer exchanged
with 10 mM MOPS pH 7.0.
Activation and Ion-Exchange Chromatography
In the activation of the native and Asp30lle mutant
proteins (0.5 mg/ml) for kinetic analysis, citric acid pH 2.0
was added to give a final concentration 0.1 M. The resulting
solution was held for fifteen minutes at room temperature.
The Asp77Thr and the double mutant Asp30lle/Asp77Thr zymogens
were activated for twenty-four hours in 0.1 M sodium formate,
37C at pH 3.0 and 3.5, respectively. After filtering
through a 0.2 Jim Millipore microcentrifuge unit, each enzyme
was directly injected onto a Pharmacia Mono S column
equilibrated with 50 mM sodium formate pH 3.0. The enzyme
was eluted by running a 25 minute gradient to 25% 50 mM
sodium formate pH 3.0 containing 1 M NaCl at a flow rate of
1 ml/min. Enzyme aliquots were quickly frozen and stored at
-20C.
Structural characterisation
N-terminal sequencing. N-terminal sequence analysis of
the activated rhizopuspepsins was performed to determine the
extent of processing during self-activation. The proteins
were electroblotted at 90 volts for 2 hours or 20 volts


118
the Asp77 is the crucial residue for enabling cleavage of
peptides with lysine in the Pi position.
Figure 7-1 shows the potential hydrogen bonding
arrangement in the active site of penicillopepsin complexed
with a pepstatin based inhibitor containing a lysine in Pi
(James et al., 1985). Asp77 forms two hydrogen bonds to the
inhibitor backbone and one to the lysine side chain. An
intra-residue hydrogen bond is also seen for Asp77. The
enzyme also forms a hydrogen bond to Asp77 and to the lysine
side chain through Ser79. The hydrogen bonds to the P2 NH
and carbonyl of the inhibitor are strictly maintained in
inhibitor complexes of the aspartic proteinases (Davies,
1990). Extending from this conserved binding mode and the
high degree of structural homology in the active site region
between rhizopuspepsin and penicillopepsin (Figure 7-2), a
possible explanation for the inability of the Asp77 mutants
and the mammalian enzymes to cleave Pi-lysine substrates can
be proposed.
Figure 7-3 shows the proposed interactions for a Pi~
lysine substrate bound to the active site of rhizopuspepsin
and the Asp77 mutants that were designed to resemble porcine
pepsin. When Asp77 is mutated, several potential hydrogen
bonding interactions are lost, depending on the particular
type of substrate bound. When substrates that do not contain
a lysine in Pi (1, 2, 4 and 5) are bound, the Asp77 intra
residue hydrogen bond and the hydrogen bond between Asp77 and
Ser79 are lost. The double mutant cycles suggest that the


120
Figure 7-2. Ca carbon backbone superposition of
penicillopepsin and rhizopuspepsin complexed with inhibitors.
The pepstatin derivative containing lysine in the Pi position
is shown in cyan. Penicillopepsin and rhizopuspepsin are
shown in yellow and blue, respectively.


50
Peak 2 was shown to be rhizopuspepsinogen by SDS-PAGE
analysis (Figure 3-4). Peaks 1, which elutes at the void
volume, was shown, upon silver staining of the gel, to be
primarily polymeric rhizopuspepsinogen. Peak 3 was shown not
to contain rhizopuspepsinogen or active rhizopuspepsin but to
consist of low molecular weight proteins by the same method.
Yields at this stage of purification ranged from 15 to 35 mg
for a 4 L preparation.
Activation and Purification
Activation of zymogens was accomplished by lowering the
pH of the solution. Activation has been shown to occur by
intermolecular and intramolecular interactions (Chen et al.,
1991). The optimal conditions for activating the zymogens
varied. Figure 3-5 shows the time course of activation of
Asp30lle at a protein concentration of 0.5 mg/ml. The
protein was efficiently converted at both pHs and room
temperature to a molecular weight of 35 kDa. Sequence
analysis of the WT-REC and Asp30lle proteins, activated at pH
2.0, confirmed the N-terminus of these proteins to be Thr-
Ser-Thr-Gly-Gly-Ile-Val-Pro-Asp-. This sequence represents
an extension of naturally occurring rhizopuspepsin by 9 amino
acids. Figure 3-6 shows the results of an activation
experiment of WT-REC at pH 4.0 from 1 to 6 hours in an
attempt to remove this extension by intermolecular
processing.


12
initial Kinetic Studies
The initial studies characterizing the hydrolytic
properties of mammalian and fungal aspartic proteinases
utilized small tri- and tetrapeptides (Fruton, 1970, 1976).
Even with these small, poorly binding substrates, differences
in the "secondary interactions," those interactions not at
the scissile bond, were seen for pepsin, cathepsin D and
rhizopuspepsin (Voynick & Fruton, 1971; Fruton, 1976).
Subsequent studies (Sampath-Kumar & Fruton, 1974; Hofmann et
al., 1988; Balbaa et al., 1993) showed an increase in kcat as
the substrate length was extended to eight residues,
particularly when the S3, S2 and S21 subsites were occupied.
A further enhancement of substrates for the study of this
family of enzymes came when Hofmann and Hodges (1982) showed
that the change in absorbance of a p-nitrophenylalanine
residue in the Pi' position of the substrate would be greater
upon substrate hydrolysis than when present in the Pi
position (Hofmann et al., 1984). A slight modification of
Hofmann's substrates which contained lysine in Pi was made by
Dunn and coworkers (Dunn et al., 1984) using the information
from a large survey by Powers et al. (1977) on all known
cleavage junctions of the pepsin at that time. This work
yielded a chromogenic substrate, Pro-Thr-Glu-Phe*Nph-Arg-Leu
(Nph = p-nitrophenylalanine), that could be cleaved by
porcine pepsin in a continuous assay allowing the
quantitation of initial rates. The binding mode of this


Table 3-1. Representative yields during the purification of the recombinant
rhizopuspepsins
WT-REC
Asp30lle
Asp77Thr
Asp30lle/Asp79Thr
Inclusion bodies (wet)
1000
mg
429 mg
805 mg
813 mg
8M urea
97.2
mg
99.4 mg
33.3 mg
98.4 mg
Refold supernatant
25.8
mg
37.3 mg
19.5 mg
15.3 mg
Gel filtration
7.6
mg
10.0 mg
6.4 mg
6.3 mg
Ion exchange (Mono S)
2.0
mg
3.8 mg
1.7 mg
1.2 mg


Wichet, Brian, Jenny, and the whole rest of gang for their
friendships. The efforts and friendships of the Protein
Chemistry Core Facility have also made my stay in Gainesville
enjoyable. In particular, I want to thank Ruth, Nancy, Hung,
Benne, and the whole rest of the gang.
IV


Figure 3-6. SDS-PAGE analysis of the activation of wild-type
recombinant rhizopuspepsinogen at pH 4.0. Incubations were
performed from 0 to 6 hours with protein concentrations of
0.1, 0.5, 1.0, and 1.5 mg/ml. C, crude isozyme pi 6.


Fraction Unfolded
73
Figure 4-4. Guanidinium hydrochloride induced
unfolding of the naturally occurring isozyme
pi 6 and the recombinant forms of rhizopuspepsin
monitored by the change in intrinsic fluorescence
at 350 nm.


115
result from differences in the pi values of the proteins.
Pepsin has a pi value less than 2 while the different
rhizopuspepsins have pi values ranging from 5 to 6 (Chapter 3
and 4). At pH 3.5 and 5.0 porcine pepsin is highly
negatively charged. In contrast, the rhizopuspepsins are
positively charged at pH 3.5 and become more negatively
charged at pH 5.0 with a net charge near zero. In order for
wild-type rhizopuspepsin at pH 5, for example, to have an
overall net charge of zero, many acidic residues must be
titrated to balance the large number of positive residues on
the enzyme surface. This increase in negative charges may
attract and facilitate the catalysis of positively charged
substrates.
Rhizopuspepsin readily cleaves peptide 3 containing
lysine in Pi. The Asp77Thr mutants cleave this substrate
poorly and porcine pepsin cleaves the peptide in the wrong
place altogether. The degree to which Asp30 and Asp77
contribute to specificity can be analyzed by using double
mutant cycles (Carter et al., 1984; Ackers & Smith, 1985;
Fersht, 1985). The free energy of transition state
stabilization can be affected by the introduction of
mutations. The changes seen upon multiple mutations are
generally additive providing (1) the residues do not directly
contact each other, (2) large structural changes do not occur
or (3) no change in the reaction mechanism has occurred
(Wells, 1990).


54
M O 15 30 1 2 3
kDa
67 _
A
43 -
30 -


41
enzyme-substrate complex from the ground state to the
transition state (Fersht, 1985) :
AG* = AGS + AG* = RTIn(kBTIK)-RTln^ /KJ (4)
With the assumption that the energies associated with the
bond breaking and making steps, AG*, are not significantly
affected upon mutation of the enzyme or changes in the assay
pH, the discrimination of the wild type and mutant enzymes
for different substrates can be evaluated by their relative
binding to the transition state:
. ^, k, / K (mutant, mutant 2 or pH 5.0)
AAG* = -RT\n (5)
kat / Km (wild type, mutant 1 or pH 3.5)
Molecular Graphics
Molecular graphic representations of the x-ray crystal
structures were generated using the Insight II (version 2.3)
from Biosym Technologies, Inc. (San Diego) on a Silicon
Graphics Indigo system at the University of Florida Center
for Structural Biology. Root mean square (RMS) superposition
of the Coe (alpha carbon) backbones of rhizopuspepsin (3APR,
Suguna et al., 1987) and penicillopepsin (1APT, James et al.,
1985) inhibitor complexes were performed by selecting the
active site residues 27-37 and 210-220 for the SUPERIMPOSE
command in the TRANSFORM menu.




3
catalysis has been studied by deleting a hydrogen bond donor
or acceptor on the enzyme or by making similar modifications
to the substrate. Two systems which have been extensively
studied are tyrosyl-tRNA synthetase and the pancreatic
proteinase trypsin.
Tvrosvl-tRNA Synthetase
One of the most thoroughly studied model systems for
studying the role of hydrogen bonding in specificity has been
the tyrosyl-tRNA synthetase from Bacillus stearothermophilus.
The tyrosyl-tRNA synthetase ensures the fidelity of
information transfer from the genetic code to the final
protein product by optimizing interactions with the
structural components of the amino acid tyrosine that make it
different from the other amino acids and, in particular,
phenylalanine. The examination of the crystal structure of
the enzyme bound aminoacyl adenylate has shown that there are
eight hydrogen bonds between the enzyme and the substrate
which can be studied by mutation of the enzyme. Fersht and
his coworkers have analyzed the effects of mutations by
comparing the kcat/Km values of the wild-type enzyme to the
mutant enzyme for the activation of tyrosine and
phenylalanine (Carter et al., 1984; Fersht, 1985, 1988;
Fersht, et al., 1985; Leatherbarrow et al., 1985; Lowe et
al., 1987). This comparison gives information about the
overall apparent change in transition state stabilization


TABLE OF CONTENTS
page
ACKNOWLEDGMENTS iii
LIST OF TABLES viii
LIST OF FIGURES X
KEY TO SYMBOLS xii
ABSTRACT xiv
CHAPTERS
1 HYDROGEN BONDING AND BIOLOGICAL SPECIFICITY 1
Introduction 1
Tyrosyl tRNA Synthetase 3
Trypsin 5
Aspartic Proteinases 7
General Characteristics 7
Initial Kinetic Studies 12
Rhizopuspepsin 14
Historical Background 15
Specificity Differences Studied by Site-Directed
Mutagenesis 16
2 EXPERIMENTAL PROCEDURES 23
Introduction 23
Materials 23
Methods 25
Cloning and Mutagenesis 25
Expression 30
Refolding 30
Size-Exclusion Chromatography 31
Activation and Ion-Exchange Chromatography 32
Structural Characterization 32
Kinetic Analysis 37
Analysis of Transition State Effects 40
Molecular Graphics 41
3 EXPRESSION, REFOLDING, PURIFICATION, AND ACTIVATION
OF RECOMBINANT RHIZOPUSPEPSINS 42
Introduction 42
v


30
Expression
The native and mutant enzymes were expressed and
purified from BL21(DE3) E. coli cells as reported with minor
changes (Chen, Koelsch et al., 1991). A 1:50 dilution of an
overnight culture grown in M9 media (lO^lg/mL thiamine, 0.5%
casamino acids, 0.2% glucose) containing 50 mg/L ampicillin
was made into LB media containing the same amount of
ampicillin and grown to an OD600 of 0.5. At that time IPTG
was added to give a final concentration of 0.5 mM. The cells
were pelleted at 3,500 x g for 10 minutes and resuspended in
4.2 mis of 50 mM Tris-HCl pH 7.4, 150 mM NaCl and 1 mM MgCl2
(buffer A) per gram of cells. Following the addition of 80
Kunitz units of DNase (Sigma) per ml of suspension, the cells
were lysed by two passes through a French Press cell. The
resulting slurry was carefully layered over a 27% sucrose
cushion (density = 1.1) and centrifuged at 12,000 x g in
order to isolate the inclusion bodies which sediment through
the sucrose solution (Taylor et al., 1986). The inclusion
bodies were washed by resuspension in buffer A containing 1%
Triton X-100 and pelleted through sucrose a second time. The
resulting pellet was stored at -20C until refolding.
Refolding
In order to regain enzymatic activity, the wild-type and
mutant recombinant proteins were refolded by a modified
procedure for the refolding of prochymosin involving


Table 7-1. Kinetic analysis of wild-type and mutant rhizopuspepsins: systematic
substitution of lysine into P3~Pia
PEPTIDEb
WT-REC
Asp30Ile
pH
heat
Km
kcat/Km
heat
Km
hcat/Km
(s1)
(HM)
(M-
1S
-1)
(s'1)
(HM)
(M-
ls-1)
x 10
-6
XlO"6
3.5
24
3
19

2
1.24

0.17
29
3
64

2
0.46
0.05
1
KPKAF*XRA
5.0
28
3
6

1
4.90

0.95
29
3
14

1
2.15
0.30
3.5
36
5
45

7
0.81

0.16
47
7
139

23
0.34
0.07
2
KPAKF*XRA
5.0
41
5
19

2
2.13

0.34
30
3
25

2
1.22
0.18
3.5
35
4
98

6
0.36

0.05
33
5
395

51
0.08
0.02
3
KPAAK*XRA
5.0
35
4
21

1
1.66

0.20
40
5
99

9
0.40
0.06
3.5
33
4
15

1
2.17

0.26
15
2
31

1
0.49
0.05
4
KPAAF*XRA
5.0
34
4
7

1
4.73

0.63
16
2
17

2
0.94
0.17
3.5
18
2
18

1
1.00

0.12
67
8
173

15
0.39
0.06
5
KPAAZ*XRA
5.0
22
2
10

1
2.26

0.41
56
6
60

5
0.94
0.13
aThe kinetic parameters ( standard deviations) were determined at 37C with either 0.1 M sodium
formate, pH 3.5, or 0.1 M sodium acetate, pH 5.0, containing NaCl to maintain constant ionic
strength as outlined in Chapter 2.
Abbreviations used in the substrates are: X = p -nitrophenylalanine; Z = norleucine; = site of
cleavage.
ckCat/Km values were determined spectrophotometrically with the assumption that [S] Km;
v = (kcat/Km)[E]0[S]0.
dkcat/Km values were determined by capillary electrophoresis; v = (kcat/Km) [E]o tsl0-
106


67
M 1 2 3 4
Figure 4-1. IEF comparison of wild-type rhizopuspepsinogen,
activated, purified WT-REC and the two naturally occurring
isozymes. Lane 1, rhizopuspepsinogen, lane 2, isozyme pi 5;
lane 3, isozyme pi 6; lane 4, WT-REC.


ml (s)
milliliter(s)
mM
millimolar
MOPS
3-(N-morpholino) propane-sulfonic acid
MWCO
molecular weight cut off
N
amino
NaCl
sodium chloride
ng
nanogram
nM
nanomolar
nm
nanometers
Nph
p-nitrophenylalanine
NTP
nucleotide triphospahtes
OD
optical density
ori
origin
PAGE
polyacrylamide gel electrophoresis
PI
isoelectric point
pmol
picomoles
PVDF
polyvinylidene difluoride
rec
recombinant
s
second
SDS
sodium dodecyl sulfate
sec
seconds
tet
tetracycline
TFA
trifluoroacetic acid
Tricine
N-[Tris-(hydroxymethyl) Methyl] glycine
Tris
tris (hydroxymethyl) aminomethane
Vmax
maximum velocity
WT-REC
wild-type recombinant
fig
microgram
|IM
micromolar
Xlll


21
Figure 1-3. Closeup view of the active site of
rhizopuspepsin highlighting the catalytic aspartic acid
residues, Asp32 and 215, and Asp30 and Asp77.


Table 5-2. Inhibition constants for Leu4,[CH(OH)CH2]Val and Statine
modified derivatives.
Inhibitor
P5-P4-P3-P2
Pl-Pl'
P2-P3'
Ki
(nM)
11
U85548E
Val-Ser-Gln-Asn
Leu1? [X] Val
Ile-Val
<
0.1
12
U92522E
Ac-Ser-Gln-Asn
Leu'F [X] Val
Ile-NH2
1.4
0.3
13
U92517E
Ac-Gln-Asn
Leu*? [X] Val
Ile-NH2
1.4
0.3
14
U92516E
Ac-Asn
Leu1? [X] Val
Ile-NH2
510
92
15
U84728E
Ac
Leu'F [X] Val
Ile-NH2
20 200
3 700
16
U85964E
Ac-Val-Val
Leu'? [X] Val
lie-Amp
<
0.1
17
Pepstatin
Iva-Val-Val
Sta
Ala-Sta
0.7
0.2
18
Ac-Pepstatin
Ac-Val-Val
Sta
Ala-Sta
10
2
19
U77647E
Ac-Pro-Phe-His
Leu1? [X] Val
lle-NH2
54
7
X = CH (OH) CH2; Amp = aminomethylpyridine; Sta = statine = Leu1? [CH (OH) ] Gly
Iva = isovaleryl; Ac = acetyl


25
Methods
All routine DNA manipulation procedures were performed
as outlined by Sambrook et al. (1989). Competent cells were
prepared by the calcium chloride method. All plasmid and PCR
products were isolated by using Magic Minipreps or Magic PCR
preps kits from Promega. These kits use a proprietary anion
exchange column to efficiently purify plasmid DNA.
Cloning and Mutagenesis
Mutations in the rhizopuspepsinogen gene were made by
using a modified version of the overlap extension method of
site-directed mutagenesis by the polymerase chain reaction
(PCR) (Ho et al., 1989) with the use of TAQ Polymerase
(United States Biochemical) or Vent polymerase (New England
Biolabs). This modification, using only one mutant primer,
has been discussed in detail (Scarborough & Dunn, 1994) and
is shown in Figure 2-1. This procedure uses four
oligonucleotides, one of which contains nucleotide changes
corresponding to the desired amino acid change. In the first
round of amplification two reactions were performed using 100
pmol of each primer, 10 ng plasmid template and 5 units of
polymerase. The first reaction generated the 5'-end of the
gene (A) by using a sense primer (1) containing an engineered
Ndel site (5'CAT ATG GCA GTT AAC GCT GCC CC3') and an
antisense primer (2) containing the mutations for residues 30
or 77 (Asp30 -Ile; 5'GA GGA ACC GGT ATC AAA GAT AAG GTT GAA


75
Discussion
The zymogen forms of the native and mutant proteins were
efficiently expressed in E.coli and refolded from inclusion
bodies. Several lines of evidence exist to support the
conclusion that the recombinant rhizopuspepsins are
structurally and enzymatically similar to the native
isozymes. Maturation of rhizopuspepsinogen requires
catalytic activity. Activation has been to shown to occur
upon lowering of the environmental pH by intermolecular and
intramolecular processes similar to that of porcine
pepsinogen (Chen et al., 1991). Kinetic comparisons between
the two naturally occurring isozymes and the wild-type
recombinant enzyme have shown that the 9 amino acid N-
terminal extension and differences in pi values do not result
in significant deviations in catalytic activity. These
observations suggest that the slight differences seen in the
pi values of the mutants will not adversely affect their
kinetic analysis (Chapter 2, Figure 3-9). The degree of
similarity seen in the circular dichroism and denaturation
studies lends additional support to the conclusion that the
recombinant wild-type and mutant proteins are correctly
folded overall. Denaturation studies with the aspartic
proteinase zymogen prochymosin have also shown that the
recombinant protein is directly comparable to the native
enzymes (Sugrue et al., 1990).


132
Hurley, J. H., Baase, W. A. & Matthews, B. w. (1992) J. Mol.
Biol. 224, 1143-1159.
Hyland, L. J., Tomasek, T. A., Jr. & Meek, T. D. (1991a)
Biochemistry 30, 8454-8463.
Hyland, L. J., Tomasek, T. A., Jr., Roberts, G. D., Carr, S.
A., Maggard, V. W., Bryan, H. L., Fakhoury, S. A., Moore
M. L., Minnich, M. D., Culp, J. F., DesJarlais, R. L. &
Meek, T. D. (1991b) Biochemistry 30, 8441-8456.
Jackson, S., Moracci, M., elMarsy, Johnson, C. & Fersht, A.
(1993) Biochemistry 32, 11259-11269.
James, M. N. G.& Sielecki, A. R. (1985) Biochemistry 24,
3701-37013.
James, M., Sielecki, A. & Hofmann, T. (1985) in Aspartic
Proteinases and Their Inhibitors (Kostka, V., Ed.) pp
163-177, Walter de Gruyter, Berlin.
James, M. N. G., Sielecki, A. R., Hayakawa, K. & Gelb, M. H.
(1992) Biochemistry 31, 3872-3886.
Jencks, W. (1969) in Catalysis in Chemistry and Biology,
McGraw-Hill, New York.
Katz, B. & Kossiakoff, A. (1990) Proteins: Struc. Func.
Genet. 7, 343-357.
Kempf, D. J., Norbeck.D.W., Codacovi, L., Wang, X. C.,
Kohlbrenner, W. E., Wideburg, N. E., Paul, D. A., Knigge
M. F., Vasavanonda, S., Craig-Kinnard, A., Saldivar, A.,
Rosenbrook, J., W., Clement, J. J., Plattner, J. J. &
Erickson, J. (1990) J. Med. Chem. 33, 2687-2689.
Kho, C. & Zarbl, H. (1992) BioTechniques 12, 228-230.
Krieger, M., Kay, L. M. & Stroud, R. M. (1974) J. Mol. Biol.
83, 209-230.
Leatherbarrow, R., Fersht, A. & Winter, G. (1985) Proc. Natl
Acad. Sci. U.S.A. 82, 7840-7844.
Leatherbarrow, R. (1987). Enzfitter, a non-linear regression
analysis program for the IBM PC, Elsevier-BIOSOFT,
Amsterdam.
Lin, X., Wong, R. & Tang, J. (1989) J. Biol. Chem. 264, 4482
4489.
Lin, X., Tang, J., Koelsch, G., Monod, M. & Foundling, S.
(1993) J. Biol. Chem. 268, 20143-20147.


126
cleavage to be seen. Both enzymes contain a Glul3 residue.
One possible residue in rhizopuspepsin that may account for
this difference is Asnll7. The majority of the mammalian
enzymes, including porcine pepsin, contain a phenylalanine
residue at this position. A comparison of the uncomplexed
structure (Suguna et al., 1987) to many of the inhibitor
complexes of rhizopuspepsin (Suguna et al., 1987, 1992) shows
that upon binding of the inhibitor a large movement of Asnll7
occurs. Upon binding of the inhibitor the Asn residue alters
its interactions with solvent to hydrogen bond to Glul3.
Site-directed mutagenesis of Asnll7 to Phe may lead to some
insight into this phenomenon.
Experiments which may prove to be very interesting might
be those where Thr77 of porcine pepsin has been replaced with
an aspartic acid residue. This experiment would test whether
or not Pi-lysine specificity could be engineered into a
mammalian aspartic proteinase. Other experiments along these
same lines would include mutation of Ser79. These studies
may give further insight into the contribution of the serine
side chain hydrogen bonds to the cleavage of substrates
containing lysine in Pi.
Future experiments with the Asp30 and Asp77 mutants of
rhizopuspepsin described in Chapter 7 should provide
additional support for the importance of Asp77 interactions
with substrates. A comparison of the interaction of an
inhibitor containing lysine in Pi with the different mutant
enzymes may also show the requirement of Asp77 for binding.


CHAPTER 1
HYDROGEN BONDING AND BIOLOGICAL SPECIFICITY
Introduction
Biological processes are controlled by a variety of
forces that influence intramolecular and intermolecular
interactions. A balance of the positive and negative aspects
of these forces, for example, determines the structure of DNA
and proteins, establishes the specificity of binding
interactions needed for information transfer and substrate
recognition, and creates the environmental requirements for
enzymatic catalysis. These forces have been classically
divided into two categories: (1) forces which lead to binding
energy, dispersion or van der Waals forces and the
hydrophobic effect, and (2) forces which relate to
specificity or the discrimination of one molecule over
another, electrostatic interactions and hydrogen bonds
(Jencks, 1969; Fersht, 1985; Fersht et al., 1985; Dill,
1990) .
The contribution of van der Waals forces and the
hydrophobic effect to protein folding and binding
interactions is thought to extend primarily from an increase
in entropy upon release of water molecules to the bulk
solvent. Even though there is, for example, a large decrease
1


CHAPTER 7
ENGINEERING THE SUBSTRATE SPECIFICITY OF RHIZOPUSPEPSIN: THE
ROLE OF ASP30 AND ASP77 IN THE ABILITY OF A FUNGAL ASPARTIC
PROTEINASE TO CLEAVE SUBSTRATES WITH LYSINE IN Pi
Introduction
Double mutant cycle analysis has been used in the effort
to interpret and quantitate the amount of function lost or
gained by a mutant protein by comparing the properties of two
single mutants to the corresponding double mutant (Carter et
al., 1984; Akers & Smith, 1985; Fersht, 1985; Wells, 1990;
Mildvan et al., 1992). These functions include catalysis,
binding and intrinsic stability. An assumption made in this
type of analysis is that the loss of function directly
measures the contribution of the mutated residue. The
evaluation of mutation induced effects are facilitated by
comparing the changes seen in kinetic, kcat and kcat/Km, and
thermodynamic, equilibrium constants, parameters.
The substrate specificity or the ability of a proteinase
to distinguish between two substrates that are highly similar
in structure has been routinely studied by comparing the
kcat/Km parameter. This parameter or 'specificity constant'
is a complex function of all the kinetic constants, involving
the binding and subsequent turnover of substrate to form the
products of the reaction, up to and including the first
101


56
0.1 0.5 1.0 1.5
kDa M C 0 1 2 3 6 1 2 3 6 1 2 3 6 1 2 3 6 M


135
D., Kezdy, F. & Heinrickson, R. L. (1991) in Structure
and Function of the Aspartic Proteinases (Dunn, B. M.,
Ed.) pp 307-323, Plenum Press, New York.
Sawyer, T. K., Maggiora, L. L., Liu, L., Staples, D. J.,
Bradford, V. S., Mao, B., Pals, D. T., Dunn, B. M.,
Poorman, R. A., Hinzman, J., deVaux, A. E., Affholter, J.
A.& Smith, C. W. (1990) in Peptides: Chemistry, Structure
and Biology, Proceedings of the Eleventh American Peptide
Symposium (Rivier, J. E. & Marshall, G. R., Eds.) pp 46-
48, ESCOM Science Publishers, Leiden.
Sawyer, T. K., Staples, D. J., Liu, L., Tomasselli, A. G.,
Hui, J. 0., O'Connell, K., Schostarez, H., Hester, J. B.
Moon, J., Howe, W. J., Smith, C., Decamp, D. L., Craik,
C. S., Dunn, B. M., Lowther, W. T., Harris, J., Poorman,
R. A., Wlodower, A., Jaskolski, M. & Heinrickson, R. L.
(1992) Int. J. Peptide Protein Res. 40, 274-281.
Scarborough, P. E., Guruprasad, K., Topham, C., Richo, G. R.,
Conner, G. E., Blundell, T. L. & Dunn, B. M. (1993)
Protein Sci. 2, 264-276.
Scarborough, P. E. & Dunn, B. M. (1994) Protein Eng. 7,
Schgger, H. & von Jagow, G. (1987) Anal. Biochem. 166, 368-
379.
Schechter, I. & Berger, A. (1967) Biochem. Biophys. Res.
Comm. 27, 157-162.
Schellman, J. A. (1978) Biopolymers 17, 1305-1322.
Schoemaker, J., Brasnett, A. & Marston, F. (1985) EMBO J. 4,
775-780.
Sepulveda, P., Jackson, K. & Tang, J. (1975) Biochem.
Biophys. Res. Comm. 63, 1106-1112.
Shortle, D. & Meeker, A. (1986) Proteins: Struct. Func.
Genet. 1, 81-89.
Shortle, D. & Meeker, A. (1989) Biochemistry 28, 936-944.
Shortle, D. (1992) Q. Rev. Biophys. 25, 205-250.
Sogawa, K., Fujii-Kuriyama, Y., Mizukami, Y., Ichihara, Y. &
Takahashi, K. (1983) J. Biol. Chem. 258, 5306-5311.
Spatola, A. (1983) in Chemistry and Biochemistry of Amino
Acids, Peptides and Proteins, Dekker, New York.
Strop, P., Sedlacek, J., Stys, J., Kaderabkova, Z., Blaha,
I., Pavlickova, L., Pohl, J., Fabry, M., Kostka, V.,


11
Figure 1-2. Closeup view of a reduced peptide bond inhibitor
bound to the active site of rhizopuspepsin (Suguna et al.,
1987b). This figure illustrates the binding mode of ligands
to the active site of aspartic proteinases and illustrates
the Schechter and Berger nomenclature for describing active
site interactions. For example, the side chain of the P3
residue of the ligand interacts with the S3 subsite of the
enzyme. Bond cleavage occurs between Pi and Pi'.


124
side chain allows for hydrophobic interactions with the
aliphatic portions of the side chain in the Si subsite and
the stabilization of a positive charge in a hydrophobic
region.
Specific and general electrostatic interactions have
also been shown, through the use of site-directed mutagenesis
and kinetic analysis of substrates containing arginine and
lysine in the P3 and P2 positions, to play a key role in the
substrate specificity of porcine pepsin, particularly in the
S3 and S2 binding subsites (Rao-Naik, unpublished data). The
electrostatic interactions of positively charged side-chains
of the substrate in the S3 subsite were shown to be specific
by the loss of pH dependence of kcat when Glul3 was mutated to
an alanine. Interactions in the S2 subsite have been shown
to be a combination of general and specific effects. In
these studies the mutation of Glu287 to a methionine residue
present in cathepsin D resulted in a decrease in specificity
for these substrates but did not result in a complete loss in
the ability to tolerate positively charged residue in the P2
position.
Cathepsin D is different from pepsin and rhizopuspepsin
in that it has a preference for hydrophobic interactions. An
example of this general overall specificity is that it has a
strong aversion to substrates containing positive charges in
the P2 position (Scarborough et al., 1993). Site-directed
mutagenesis of cathepsin D at position 287 has shown that
this strict preference for hydrophobic residues in the P2


9
Figure 1-1. Ribbon represention of the aspartic proteinase
rhizopuspepsin complexed with a reduced peptide bond
inhibitor. The catalytic aspartic acid residues, Asp32 and
Asp215, are shown in red. The flap which extends over the
active site is shown in orange. The inhibitor is shown in
yellow. Coordinates were obtained from the Brookhaven
Protein Data Bank file 3APR (Suguna et al., 1987).


Table 5-1. Inhibition constants for Xaa*F[CH2NH]Yaa modified
derivatives
CMPD
P5-P4-P3-P2
Pl-Pl1
P2'-P3'
K (JXM)
1
U79465E
Ac-Pro-Phe-His
Cha^F [X] Phe
nh2
> 200
2
U79211E
Ac-Pro-Phe-His
Cha^F [X] Val
NH2
> 200
3
U79464E
Ac-Pro-Phe-His
Cha'F [X] Cha
nh2
> 200
4
U79339E
Ac-Pro-Phe-His
Phe*F [X] Val
nh2
8.7 1.7
5
U71909E
Ac-Pro-Phe-His
Phe'F [X] Phe
nh2
21 4
6
U80011E
Ac-Pro-Phe-His
Phe'F [X] pClPhe
nh2
13 2
7
U80445E
Ac-Pro-Phe-His
Phe^F [X] Tyr
nh2
105 18
8
U81330E
Ac-Pro-Phe-His
Phe4/[X]pN02Phe
nh2
40 4
9
U70531E
dHis-Pro-Phe-His
Phe'F [X] Phe
Val-Tyr
5.2 0.7
10
U91990E
Ac-Pro-Hph-NMeHis
Phe'F [X] Phe
nh2
1.6 0.2
Hph = homophenylalanine; pNC>2Phe = p-nitrophenylalanine; X = CH2NH
pClPhe = p-chlorophenylalanine; Cha = cyclohexylalanine


64
contain mainly intermolecularly cross-linked protein
(Schoemaker et al., 1985). In order to reduce the cysteine
residues in the protein before refolding, high levels (10 to
1000 fold molar excess based on cysteines) of reducing agent,
BME or DTT, are required. Even if all the cysteines are
reduced, losses still occur for several different reasons:
pH of the refolding solution, speed at which the denaturant
and reducing agents are removed and the protein
concentration. These factors must be optimized in order to
obtain biologically active protein. Each protein has its own
characteristic conditions for refolding. Different
conditions may also have to used to refold mutant proteins.
The refolding and activation protocols for the WT-REC
and mutant forms of rhizopuspepsin have been optimized.
These procedures produced sufficient quantities of active
enzyme for structural and kinetic analysis. All enzymes used
exhibit homogeneous N-termini and show similar
electrophoretic properties after purification. Structural
analysis and kinetic comparisons of the recombinant
rhizopuspepsins to the naturally occurring isozymes are
discussed in Chapter 4. The inhibitor binding and substrate
specificity characteristics of WT-REC are discussed in
Chapters 5 and 6. Chapter 7 presents the analysis of the WT-
REC and mutant rhizopuspepsins toward substrates that contain
lysine in Pi.


BIOGRAPHICAL SKETCH
William Todd Lowther was born on February 18, 1967. His
interest in science began at an early age with his
fascination of animals and plants. He credits his desire for
learning to the encouragement of his parents. While in
junior and senior high school he also became enamored with
music and playing the clarinet. He graduated from Apopka
High School in 1985. He attended Stetson University,
majoring in chemistry while on a music scholarship. His
professors were a source of wisdom and friendship. Their
support and passion for teaching lead to an interest in
pursuing research. The summer before his senior year he
participated in a research project at the Inhalation
Toxicology Research Institute in Albuquerque, New Mexico.
This experience confirmed Todd's desire to pursue graduate
studies and research in biochemistry. He graduated from
Stetson in 1989. Todd began his graduate studies in the
laboratory of Professor Ben Dunn in 1989. Dr. Dunn provided
a stimulating environment for furthering Todd's interest in
protein-ligand interactions and protein engineering. Upon
graduation from the University of Florida, Todd plans to
continue his study of proteins and their specific
interactions with ligands.
138


18
Table 1-2. Primary sequence comparison of
rhizopuspepsin to several aspartic proteinases
Enzyme
% identity
% similarity
porcine pepsin
39.2
62.7
human pepsin
40.5
62.3
human cathepsin E
37.2
59.7
human cathepsin D
33.1
58.8
human renin
27.5
51.6
Candida albicans
aspartic proteinase
28.1
49.4


CHAPTER 3
EXPRESSION, REFOLDING, PURIFICATION, AND ACTIVATION OF
RECOMBINANT RHIZOPUSPEPSINS
Introduction
Overexpression of proteins in heterologous systems has
become an indispensable method in the generation of large
quantities of wild-type and mutant proteins for structural
and biochemical analysis. Many different systems have been
used to obtain the protein of interest; for example, E. coli,
yeast and SF9 insect cells. The decision of which system to
use primarily depends on the yields required and whether or
not glycosylation of the protein is desired. Expression in
E. coli usually gives the highest yields but the protein is
frequently deposited in an insoluble form known as inclusion
bodies. These deposits are readily purified and are usually
greater than 95% pure protein. In order to regain biological
activity the protein must be denatured and refolded. Upon
optimization of the refolding conditions, peculiar to each
protein and its mutants, a sufficient quantity of protein for
analysis may be obtained.
This chapter outlines the used of E. coli to produce
sufficient quantities of wild-type (WT-REC) and mutant forms
of rhizopuspepsin for structural and kinetic analysis.
Expression of the zymogen form of rhizopuspepsin,
42


137
Schwering, J. E., Homnick, C. F., Nunberg, J., Springer,
J. P. & Huff, J. R. (1992) J". Med. Chem. 35, 1685-1701.
Togni, G., Sanglard, D., Falchetto, R. & Monod, M. (1991)
FEBS Lett. 286, 181-185.
Veerapandian, B., Cooper, J. B., Sali, A. & Blundell, T. L.
(1990) J. Mol. Biol. 216, 1017.
de Viragh, P. A., Sanglard, D., Togni, G., Falchetto, R. &
Monod, M. (1993) J. Gen. Microbiol. 139, 335-342.
Voynick, I. & Fruton, J. (1971) Proc. Nat. Acad. Sci. U.S.A.
68, 257-259.
Wells, J. A. (1990) Biochemistry. 29, 8509-8517.
Wiley, R. Sc Rich, D. (1993) Med. Res. Rev. 13, 327-384.
Young, L., Jernigan, R. & Coveil, D. (1994) Protein Sci. 3,
717-729.
Young, S. D., Payne, L. S., Thompson, W. J., Gaffin, N.,
Lyle, T. A., Britcher, S. F., Graham, S. L., Schultz, T.
H., Deana, A. A., Darke, P., Zugay, J., Schleif, W. A.,
Quintero, J. C., Emini, E. A., Anderson, P. S. & Huff, J.
R. (1992) J. Med. Chem. 35, 1702-1709.


34
Circular dichroism. Circular dichroism (CD) spectra
were determined on one day at room temperature on a Jasco
J-500C spectropolarimeter equipped with a IF-500 II computer
interface with a 0.1 cm pathlengh cell (Hellma). The
polarimeter was standardized with D(+)-camphorsulfonic acid
(Chen & Yang, 1977). The samples were diluted into buffer to
a final concentration around 0.5 mg/ml in 0.1 M sodium
formate pH 3.0. Just prior to loading into the CD cell, the
samples were filtered through a 0.2 (im microcentrifuge filter
(Rainin) and quantitated by reading their absorbance at 280
nm. The ellipticity values were converted to the molar
ellipticity, [0], using the conversion factor E}^ = 12.6
(Fukumoto et al., 1967) and a molecular weight of 35 kDa.
The data points were fit with the smoothing algorithm of the
KaleidaGraph program (version 3.0.2 Synergy software, PCS
Inc.).
Fluorescence spectroscopy. The guanidium hydrochloride
denaturation curves of the rhizopuspepsins were determined
using excitation and emission wavelengths of 280 and 350 nm,
respectively, on an SLM Aminco 4800C spectrofluorometer.
Ultrapure 8 M guanidinium hydrochloride in water was
purchased from Pierce. The proteins were diluted in
duplicate into denaturant concentrations ranging from 0 to
6 M with a final buffer concentration of 0.1 M sodium formate
pH 3.0 and an enzyme concentration around 100 nM. The
protein/denaturant solutions were equilibrated at 25C for at


CHAPTER 8
CONCLUSIONS AND FUTURE DIRECTIONS
Conclusions
A combination of site-directed mutagenesis and kinetic
analysis has clearly established the importance of Asp77 in
the unique lysine primary specificity of fungal aspartic
proteinases. Asp30, which is not conserved in the fungal
enzymes, was shown not to influence this specificity but may
play a role in general electrostatic interactions. The
presence of an aspartic acid residue at position 77 has the
potential to establish an extensive hydrogen bonding network
between the enzyme and the substrate that enables the fungal
enzymes, but not the mammalian enzymes because they contain a
Thr77, to cleave substrates containing lysine in the Pi
position.
Another way to look at this unique ability is that the
Asp77 residue in the fungal enzymes may only be providing
extra interactions that stabilize what would normally be a
very unfavorable residue interaction in the hydrophobic Si
subsite of the aspartic proteinases. The lysine side chain
in the penicillopepsin complex extends out of the Si subsite
in order to make interactions with Asp77 which is located on
the surface of the protein. This orientation of the lysine
123


LIST OF REFERENCES
Abad-Zapatero, C., Rydel, T. J., Neidhart, D. J., Luly, J. &
Erickson, J. W. (1991) in Structure and Function of
Aspartic Proteinases: Genetics, Structures and Mechanisms
(Dunn, B. M., Ed.) pp 9-21, Plenum Press, New York.
Ackers, G. & Smith, F. (1985) Ann. Rev. Biochem. 54, 597-629.
Al-Janabi, A., Hartsuck, J. A. & Tang, J. (1972) J. Biol.
Chem. 247, 4628-4632.
Azaryan, A., Wong, M., Friedman, T., Cawley, N., Estivariz,
F., Chen, H. & Loh, Y. (1993) J. Biol. Chem. 268, 11968-
11975.
Azuma, T., Liu, W. G., Vanderlaan, D. J., Bowcock, A. M. &
Taggart, R. T. (1992) J. Biol. Chem. 267, 1609-1614.
Balbaa, M., Cunningham, A. & Hofmann, T. (1993) Arch.
Biochem. Biophys. 306, 297-303.
Baldwin, E. T., Bhat, T. N., Gulnick, S., Hosar, M. V.,
Sowder II, R. C., Cachau, R. E., Collins, J., Silva, A.
M. & Erickson, J. W. (1993) Proc. Natl. Acad. Sci. USA
90, 6796-6800.
Blundell, T., Cooper, J., Foundling, S., Jones, D., Atrash,
B. & Szelke, M. (1987) Biochemistry 26, 5585-5590.
Boger, J., Payne, L. S., Perlow, D. S., Lohr, N. S., Poe, M.,
Blaine, E. H., Ulm, E. H., Schorn, T. W., LaMont, B. I.,
Lin, T. Y., Kawai, M., Rich, D. & Veber, D. F. (1985) J.
Med. Chem. 28, 1779-1790.
Bott, R., Subramanian, E., & Davies, D. R. (1982)
Biochemistry 21, 6956-6962.
Carter, P., Winter, G., Wilkinson, A. & Fersht, A. (1984)
Cell 38, 835-840.
Chen, G. & Yang, J. (1977) Anal. Lett. 10, 1195-1207.
Chen, K., Tao, N. & Tang, J. (1975) J. Biol. Chem. 250, 5068-
5075.
Chen, Z., Koelsch, G., Han, H. P., Wang, X. J., Lin, X. L.,
Hartsuck, J. A. & Tang, J. (1991) J. Biol. Chem. 266,
11718-11725.
Cleland, W. W. (1992) Biochemistry 31, 317-319
129


Figure 3-4. SDS-PAGE analysis of the fractions from the
purification of Asp30lle by gel filtration in Figure 3-3.
Lane 1, peak 1; lanes 2-10, peak 2; lane 11, peak 3 .


6
homology. Despite this similarity, chymotrypsin has a
primary specificity for large aromatic residues. The most
notable substitution that may be responsible for the
specificity differences between trypsin and chymotrypsin
occurs at position 189, where Asp has been replaced by Ser.
The mutation of Aspl89 to Ser in trypsin resulted in a
dramatic decrease in the kcat/Km values for Pi Arg and Lys
containing substrates (Graf et al., 1988). The average loss
of transition state stabilization energy as a consequence of
this mutation was 6.7 kcal mol-1. This value is in the same
range as those seen by Fersht for the deletion of a charged
hydrogen bond donor/acceptor discussed above. Experiments
were also performed to attempt to rescue the basic residue
specificity. The substitution of Asp at position 190 did
restore activity (Evnin et al., 1990). Experiments have also
shown that activity can be restored when acetate is present
in the buffer at very high levels (Perona, et al., 1994).
Interestingly, the Aspl89Ser mutant did show some
improvement toward cleaving large hydrophobic substrates.
Further mutagenesis experiments tried to complete the
conversion by mutating the remaining residues in the Si
pocket of trypsin to those in chymotrypsin. This effort
still did not result in a complete metamorphosis to
chymotrypsin. Only upon changing surface loops around Pi
subsite of trypsin to those of chymotrypsin in conjunction
with the Pi substitutions resulted in the desired enzyme
specificity (Hedstrom et al., 1992).


88
bifurcation can be seen, for example, by comparing the
structure of the reduced inhibitor U70531E (9, Suguna et al.,
1987) with that of U85548E (11, unpublished data). U70531E
and the U85548E have Phe and Val in the Pi' positions,
respectively. When Phe is present in the Pi1 position, the P2
side chain is oriented toward the S4 pocket surrounded by
Thr221, Phe278 and Leu223. The presence of Val in Pi',
however, positions the P2 side chain toward the long,
extended Si' pocket. The P2 side chain in this position may
result in the lower Ki seen for U79339E (4) when compared to
5 since the P2 His of 4 could partially fill the Si' subsite.
This apparent flexibility in the binding of the P2 side chain
implies that interactions in the S2 pocket may not actually
be that specific, due to the smaller loops in this region and
the presence of glycine at position 287 in rhizopuspepsin and
renin, and that the major contribution of the P2 residue
comes from hydrogen bonds to backbone. Even though there is
flexibility in the orientation of the P2 side chain, studies
using inhibitors and substrates with systematic substitutions
have shown that the S2 subsite does contribute significantly
to the specificity differences seen for pepsin, cathepsin D,
cathepsin E and renin (Rao et al., 1993; Scarborough et al.,
1993). For example, cathepsin D cleaves substrates with
positively charged residues in the P2 position poorly
(Scarborough et al., 1993).


15
aspartic proteinases of several opportunistic Candida
species (Fusek et al., 1993; Morrison et al., 1993). The
knowledge gleaned from studying rhizopuspepsin has the
potential to foster the design of anti-fungal agents for use
in treating vaginal infections and immunocompromised AIDS,
organ transplant and cancer patients (Samaranayake &
Holmstrup, 1989; Saral, 1991; Paya, 1993).
Historical Background
Rhizopuspepsin was first isolated by Fukumoto et al.
(1967). Purification of rhizopuspepsin by isoelectric
focusing identified two major isozymes (Graham et al., 1973).
These two isozymes have been shown to be very similar in
molecular weight, amino acid composition, specific activity,
as well as in three-dimensional structure as shown by the
crystal structure of the isozyme mixture (Subramanian et al.,
1977). N- and C-terminal sequencing showed that the two
isozymes were identical (Sepulveda et al., 1975). Grippon et
al. (1977) showed the first structural difference at residue
12 (pepsin numbering). Isozyme pi 5 has a lie residue at
position 12 while pi 6 has a Val. Delaney et al. (1987)
solved the complete sequence of isozyme pi 6 by a combination
of amino acid sequencing of HPLC-purified CNBr cleavage
fragments by Edman degradation (154 residues/325 directly
sequenced) and the DNA sequencing of a positive cDNA clone,
33E2. Subsequent work by Takahashi's group established the


100
structural variations that lead to the differences between
the Candida enzymes and rhizopuspepsin awaits the solution of
inhibitor complexes of the Candida proteinases by x-ray
crystallography. Efforts in this direction are already
underway (Cutfield et al., 1993).
From the substrate studies of rhizopuspepsin and the
Candida enzymes, it is clear that the fungal enzymes have a
broad specificity toward substrates with substitutions in P5-
Pl and P2'-P3' Distinct interactions do not appear to be
present, in contrast to the mammalian enzymes, which can be
exploited in the design of targeted anti-fungal agents.
Despite this apparent lack of specificity, the fungal enzymes
have the unique ability to cleave substrates containing
lysine in Pi (Hofmann & Hodges, 1982). This particular
characteristic of the fungal enzymes has been studied by
site-directed mutagenesis and is described in Chapter 7.


3-9. IEF comparison of the activated, purified
recombinant rhizopuspepsins 62
4-1. IEF comparison of wild-type rhizopuspepsinogen,
activated, purified WT-REC and the two naturally
occurring isozymes 67
4-2. CD spectra of the recombinant wild-type and
mutant forms of rhizopuspepsin 71
4-3. Fluorescence emission spectra for folded (0 M
GdnHCl) and unfolded (6 M GdnHCl) wild-type
recombinant rhizopuspepsin 72
4-4. Guanidinium hydrochloride induced unfolding of
the naturally occurring isozyme pi 6 and the
recombinant forms of rhizopuspepsin monitored by
the change in intrinsic fluorescence at 350 nm 73
7-1. Hydrogen bonding interactions in penicillopepsin
between Asp77, Ser79 and the pepstatin derivative
containing lysine in the Pi position 119
7-2. Ca carbon backbone superposition of
penicillopepsin and rhizopuspepsin complexed with
inhibitors 120
7-3. Proposed hydrogen bonding in WT-REC
rhizopuspepsin and Asp77Thr mutants with
substrates containing lysine in Pi 121
xi


Results 43
Mutagenesis 43
Expression and Refolding 43
Activation and Purification 50
Discussion 59
4 KINETIC AND STRUCTURAL AUTHENTICITY OF RECOMBINANT
RHIZOPUSPEPSINS 65
Introduction 65
Results 66
WT-REC and Native Isozymes of Rhizopuspepsin 66
Structural Comparisons 68
Discussion 75
5 ANALYSIS OF THE SPECIFICITY OF RHIZOPUSPEPSIN
THROUGH THE USE OF INHIBITORS CONTAINING Pi AND Pi'
SUBSTITUTIONS AND PEPTIDE BOND MIMETICS 80
Introduction 80
Results and Discussion 81
6 THE BROAD SUBSTRATE SPECIFICITY OF RHIZOPUSPEPSIN:
ANALYSIS WITH SYSTEMATIC SUBSTITUTIONS IN P5-P1 AND
P2 1-P31 OF THE SUBSTRATE LYS-PRO-ALA-LYS-PHE*NPH-
ARG-LEU 91
Introduction 91
Results 92
Lys-Pro-Ala-Lys-Phe*Nph-Arg-Leu based substrates ... 92
Lys-Pro-Ile-P2-Phe*Nph-Arg-Leu based substrates 96
Discussion 96
Comparison to the Mammalian Enzymes 98
Comparison to the Candida Aspartic Proteinases 99
7 ENGINEERING THE SUBSTRATE SPECIFICITY OF
RHIZOPUSPEPSIN: THE ROLE OF ASP30 AND ASP77 OF A
FUNGAL ASPARTIC PROTEINASE TO CLEAVE SUBSTRATES
WITH LYSINE IN Pi 101
Introduction 101
Results 104
Kinetic Analysis of the Recombinant
Rhizopuspepsins 104
Kinetic Analysis of Porcine Pepsin 109
Discussion 110
Substrate Design Ill
Kinetic Analysis 112
8 CONCLUSIONS AND FUTURE DIRECTIONS 123
Conclusions 123
Future Directions 125
vi


Table 7-3. Transition state stabilization energy changes seen with
variation in pH from 3.5 to 5.0 for the recombinant rhizopuspepsins
and porcine pepsina
PEPTIDE
WT-REC
Asp30lle
Asp77Thr
Asp30lle/Asp77Thr
PPEP
AAG^ (kcal
mole-1)b
1
KPKAF*XRA
-0.9
-1.0
-0.4
-0.6
-1.1
2
KPAKF*XRA
-0.6
-0.8
-0.5
-0.7
-1.5
3
KPAAK*XRA
-0.9
-1.0
-0.6
-0.8
c
4
KPAAF*XRA
-0.5
-0.4
-0.4
-0.6
o
o
5
KPAAZ*XRA
-0.5
-0.5
-0.4
-0.7
-0.5
aAbbreviations and assay conditions are described in Table 7-1.
bAAGt = -RT/Km(pH 5.0)lIKm(pH 3.5)]. The standard deviations
range from 0.1 to 0.2 kcal mol-1.
cCleavage occured between Nph-Arg (X-R) instead of Phe*Nph (F*X).
113


108
Table 7-2. Kinetic analysis of porcine pepsin: systematic
substitution of lysine into P3~Pia
PEPTIDE
PH
PPEP
kcat
Fm
kcat/
(s
-1)
(|IM)
(M-
ls-1)
X
10"6
3.5
_
_
b0.0030
0.0005
1
KPKAF*XRA
5.0
2.4
0.5
122
20
0.019
0.005
3.5
21
5
339
77
0.06
0.02
2
KPAKF*XRA
5.0
20
3
31
2
0.66
0.11
3.5
c
3
KPAAK*XRA
5.0
c
3.5
29
6
83
12
0.41
0.09
4
KPAAF*XRA
5.0
24
5
61
11
0.39
0.10
3.5
22
4
103
12
0.21
0.04
5
KPAAZ*XRA
5.0
16
3
34
3
0.46
0.09
aAssay conditions described in Chapter 2. Abbreviations are
the same as those used in Table 7-1.
bkCat/Km values were determined spectrophotometrically with
the assumption that [S] Km; v = (kCat/Km) [E] o [S] o.
cCleavage occurred between Nph-Arg (X-R) instead of Phe*Nph
(F*X).


33
overnight from 12% Tris-Tricine SDS-PAGE gels (Schgger & von
Jagow, 1987) to PVDF Immobilon P transfer membranes
(Millipore) in 10 mM MES pH 6.0 containing 20% methanol. The
excised bands were analyzed by the Protein Chemistry Core
Facility.
Mass spectrometry. The activated enzymes were analyzed
by matrix-assisted laser desorption ionization time of flight
(MALDI-TOF) mass spectroscopic analysis on a Vestec (Houston,
TX) LaserTec Bench Top System. One to ten pmol of each
sample or standard was mixed 1:1 with fresh 0.05% TFA, 40%
acetonitrile, saturated sinnapinic acid. One [il of this
mixture was applied to a stainless-steel sample pin and
allowed to air dry. The mass spectrum was obtained from the
average of at least 50 laser shots (337 nm nitrogen laser, 3
ns pulse width). Time to mass/charge calibration was
performed from a calibration curve using bovine carbonic
anhydrase II (Sigma, 28,980 daltons) immediately prior to the
rhizopuspepsin samples.
Isoelectric focusing. The isoelectric points of the
proteins were determined by running precast PhastGel IEF gels
ranging from pH 3 to 9 on the Pharmacia PhastSystem. The
samples, including standards, were diluted in sample buffer
(0.01% pyronin Y, 10% glycerol, 62.5 mM TRIS pH 6.8) prior to
loading. The gels were run according to the PhastSystem
profile No. 100 and subsequently stained by the Coomassie
blue profile No. 200.


CHAPTER 4
KINETIC AND STRUCTURAL AUTHENTICY OF RECOMBINANT
RHIZOPUSPEPSINS
Introduction
Molecular biology techniques have enabled the production
of large quantities of purified proteins that would other
wise be difficult to study due to their low abundance or to
the impractical nature of the source. These techniques also
aid in the dissection and understanding of biological
phenomena through site-directed mutagenesis. The resulting
proteins, however, must be shown to be analogous to the
naturally occurring enzymes, with respect to structure and
biological property being examined, if the results from
recombinant proteins are to extrapolated to what occurs in
the physiological environment.
The recombinant rhizopuspepsinogen gene used in this
study was constructed from the two naturally occuring
isozymes, pi 5 and pi 6, by the fusion of the pro region
through residue 12 of the pi 5 isozyme gene to the pi 6 gene
at residue 12. The expression of this chimeric gene results
in an enzyme after activation that is identical to the pi 6
isozyme except at position 12 where Val is replaced by a lie.
Ilel2 is located in the S3 subsite of the protein. All the
other residue differences seen between the two naturally
65


16
complete sequence of both pi 5 and 6 by the 100% chemical
sequencing of HPLC-purified trypsin and Staphylococcus aureus
V8 protease generated peptide fragments (Takahashi, 1987,
1988). Rhizopuspepsin pi 5 and pi 6 differ only at eight
positions in the entire 325 amino acid sequence. All
substitutions are semi-conservative and only the Val/Ile 12
residue is in the active site.
The work by Chen et al. (1991) has lead to the
production of the zymogen of rhizopuspepsin,
rhizopuspepsinogen, in several different expression systems.
The zymogen form is expressed to facilitate correct folding.
Activation at low pH converts the zymogen to the active form
in combination of inter and intramolecular processes yielding
enzyme for kinetic studies. Kinetic studies comparing
various mutant rhizopuspepsin enzymes to the native enzyme
will help resolve the role hydrogen-bonding, electrostatic,
and hydrophobic interactions play in the creation of
specificity in the active site subsites Si, S2 and S3 of
rhizopuspepsin as well as give clues for understanding these
interactions in the other aspartic proteinases.
Specificity Differences Studied bv Site-directed Mutagenesis
Despite the high degree of primary sequence and
structural homology within the aspartic proteinases in and
around the active site (Table 1-2), differences are observed
in substrate specificities. The primary specificity of


69
Table 4-1. Kinetic comparison between the naturally
occurring isozymes and WT-REC rhizopuspepsin using the
substrate Lys-Pro-P3~Lys-Phe*Nph-Arg-Leu
Enzyme
P3 kcat
(s'1)
Km kcat/Km
(\m) (m-1s_1 )
X 10"6
pi 5
12

2
20

2
0.63

0.14
pi 6
Asp
15

2
21

1
0.69

0.09
WT-REC
13

1
20

2
0.65

0.09
pi 5
16

3
9

1
1.71

0.38
pi 6
Arg
17

2
16

2
1.07

0.17
WT-REC
13

2
9

1
1.39

0.25
pi 5
9

2
7

1
1.31

0.30
pi 6
Leu
12

1
8

1
1.53

0.22
WT-REC
10

1
7

1
1.37

0.21
pi 5
13

2
21

2
0.62

0.12
pi 6
Ser
12

1
19

1
0.67

0.08
WT-REC
9

1
16

2
0.60

0.08
Nph = p-nitrophenylalanine; WT-REC = wild-type recombinant
rhizopuspepsin.


85
These additional factors may be able to explain the increased
potency seen for compounds containing Phe and Leu in the Pi
position for rhizopuspepsin. From these observations it was
clear that rhizopuspepsin has different requirements in the
Si pocket since Pi-Cha-containing analogs were found early on
to confer increased potency for renin (Boger et al., 1985;
Sawyer et al., 1990; Wiley and Rich, 1993).
Systematic substitutions of the Pi' position (compounds
4 to 8), where Pi was Phe, resulted in a large range of
potency. The Val, Phe and p-ClPhe substitutions exhibited
similar inhibitory capabilities while p-NC>2Phe and Tyr in Pi'
showed progressively higher values. The later two
substitutions may be unfavorable due to the interruption of
hydrophobic/aromatic interactions with Ile216 and 298, Trpl94
Trp294, and Phe296 of the Si' binding pocket or the
introduction of a slightly altered hydrogen bonding
arrangements. These observations point to further
differences between rhizopuspepsin and renin. Substitution
of the Pi' phenyl ring of Phe'F[CH(OH)CH2] Phe containing
compounds with halogens reduced potency for renin, while the
nitro substitution had little effect on binding (Young et
al., 1992), but did increase bioavailability (Thompson et
al., 1992). Possible reasons for the small differences seen
for these substitutions against renin extend from the much
larger Si1 pocket and the use of the hydroxyethylene isostere
which resulted in subnanomolar Ki values (Szelke et al.,
1980). The hydroxyethylene isostere shows significantly


Relative Fluorescence
72
NM
Figure 4-3. Fluorescence emission spectra for
folded (0 M GdnHCl) and unfolded (6 M GdnHCl)
wild-type recombinant rhizopuspepsin.


43
rhizopuspepsinogen, resulted in the formation of inclusion
bodies. The inclusion bodies were isolated, denatured, and
refolded by dialysis. The inactive, zymogen form of
rhizopuspepsin were converted to the active, mature form by
lowering the pH of the solution. The resulting proteins were
purified by ion exchange chromatography and analyzed by
isoelectric focusing and N-terminal sequencing.
Results
Mutagenesis
Mutants of rhizopuspepsin were generated by overlap
extension PCR. All reactions yielded the desired size
fragments. The entire coding region of each protein was DNA
sequenced to determine the presence of the desired base
changes and if other spurious mutations had occurred. In
order to obtain clear sequencing ladders, an extra step was
added to the Sequenase procedure. Prior to adding stop
solution, the DNA was incubated with terminal
deoxynucleotidyl transferase in order to extend premature
stops due to high GC content and secondary structure. No
additional mutations were seen.
Expression and Refolding
In order to obtain sufficient quantities of enzyme for
structural and kinetic analyses, the rhizopuspepsinogens were


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
SITE-DIRECTED MUTAGENESIS OF THE ASPARTIC PROTEINASE
RHIZOPUSPEPSIN: AN ANALYSIS OF UNIQUE SPECIFICITY
By
William Todd Lowther
August, 1994
Chairman: Dr. Ben M. Dunn
Major Department: Biochemistry and Molecular Biology
Rhizopuspepsin is a secreted aspartic proteinase from
the fungus Rhizopus chinensis. Despite the high degree of
structural homology among the aspartic proteinases, the amino
acid residues that create the individual binding pockets have
been shown to profoundly influence substrate specificities
and inhibitor preferences. Rhizopuspepsin and other fungal
aspartic proteinases are distinct from the mammalian enzymes
in that they are able to cleave substrates with lysine in the
Pi position. Sequence and structural comparisons suggest
that two aspartic acid residues, Asp77 and Asp30 (pig pepsin
numbering), may be responsible for generating this unique
specificity in rhizopuspepsin.
In order to analyze their contributions to specificity,
Asp30 and Asp77 were changed to the corresponding residues in
xiv


93
Lys-Pro-Ala-Lys-Phe*Nph-Arg-Leu. Rhizopuspepsin is able to
cleave the majority of the substrates with nearly equivalent
efficacy. Some of the subtle differences in kinetic
parameters can be rationalized by examining the x-ray crystal
structures of rhizopuspepsin.
positions. Rhizopuspepsin appears to show no
preference for the amino acid residues placed in the P5
position of the substrate. The P4 substitutions, however,
lead to some differences. Rhizopuspepsin prefers leucine in
this position followed by proline. The Arg substitution is
clearly disfavored. This can be rationalized by the mode in
which ligands bind to the active sites of the aspartic
proteinases. Because ligands bind in a extended ^-strand
conformation, the P4 and P2 residues can potentially interact
with each other (Figure 1-2). Electrostatic repulsion
between the Arg and the Lys in P2 may be responsible for this
observed specificity difference.
and P? positions. All subtrates with P3 substitutions
are cleaved readily with a slight preference for hydrophobic
side chains. A similar preference is seen for the P2
position substitutions. Interestingly, rhizopuspepsin is
able to cleave the Arg substitutions in P3 and P2. This
observation is in direct contrast to the mammalian enzymes.
Pi position. The observations in this series confirms
the specificity of the aspartic proteinases for large
hydrophobic junctions. Phe and Leu are preferred in this
position.


116
The free energy change of the double mutant AAGD30ID77T is
related to that of the single mutants:
^^^D30I,D77T A^^D30I + AAGD77T + AGj .
The AGj term or the coupling energy reflects the extent to
which the single mutations affect each other. When the sites
are functioning independently from each other, the coupling
energy is zero.
Table 7-4 shows the double mutant cycles at pH 3.5 and
5.0 for peptides 1-5. The cycles for peptides 1-3 are
strictly additive suggesting that severe changes in structure
are not seen as suggested by the structural results
(Chapter 4). Slight deviations from additivity are seen for
peptides 4 and 5 but do exhibit the same trends. Mutation of
Asp30 to lie results in an average loss in transition state
stabilization free energy of 0.7 kcal mol-1. This loss may be
attributable to a change in the hydration shell of the
protein or the loss of a hydrogen bond to the inhibitor
mediated through a water molecule. For peptides 1, 2, 4 and
5 an average loss in transition state stabilization free
energy of 2.3 kcal mol-1 is seen when Asp77 is changed to Thr.
The largest effect on mutating position 77 is clearly seen
with the 7 kcal mol-1 decrease in transition state
stabilization for peptide 3. This observation indicates that


CHAPTER 2
EXPERIMENTAL PROCEDURES
Introduction
This chapter outlines the materials and methods used to
characterize the unique specificity of rhizopuspepsin toward
substrates and inhibitors. This study has used a combination
of systematically substituted substrates and inhibitors and
site-directed mutagenesis to accomplish this task.
Materials
Restriction and modifying enzymes were purchased from
Promega, Life Technologies, Inc., New England Biolabs or
United States Biochemical Corp. Deoxyadenosine-5'-[a-35S]
thiotriphosphate, as its triethylammonium salt (Sp isomer,
1000 Ci/mmol), was purchased from Amersham Corp. The pet3a
expression vector (Studier et al., 1990), containing the
wild-type rhizopuspepsinogen gene (Chen et al., 1991), was
kindly provided by Jordan Tang at the Oklahoma Medical
Research Foundation, Oklahoma City, Oklahoma. The
aminomethylene, VF[CH2NH] (Spatola nomenclature, 1983), and
the hydroxyethylene, 'P[CH(OH)CH2] containing inhibitors were
a gift from Tomi Sawyer. Pepstatin was purchased from Sigma.
Acetyl-pepstatin was a gift from Kohei Oda, Kyoto Institute
23


68
activation and removal of the pro region of the enzyme. The
wild-type recombinant rhizopuspepsinogen has a pi value of
7.4.
Kinetic analysis. A comparison of the substrate
specificity of the naturally occurring isozymes of
rhizopuspepsin to WT-REC was made using peptides with
substitutions in P3 (Table 4-1). Even though the enzymes
possess different eletrophoretic properties and N-termini,
the kinetic parameters determined for each substrate, within
experimental error, are directly comparable. The three
different forms of rhizopuspepsin are kinetically
indistinguishable from each other and exhibit similar
substrate specificity showing a preference for Arg and Leu
substitutions in P3.
Structural Comparisons
Mass spectrometry. The recombinant forms and the
isozyme pi 6 of rhizopuspepsin were analyzed by mass
spectrometry. The mass for each protein was determined to be
as follows: pi 6, 34,173; WT-REC, 34,627; D30I, 34,638;
D77T, 34,914; and D30I/D77T, 34,856. The sizes of the
recombinant proteins are consistent with the addition of 9
amino acids to the N-terminus of the native isozyme. The
values for the WT-REC and D30I proteins are slightly lower
than expected. This difference may be the result of
C-terminal processing.


29
Figure 2-2. pET3aE expression vector containing
rhizopuspepsinogen. AMP, ampicillin resistance
gene; ori, origin or replication; SD, Shine-
Dalgarno sequence.


96
presence of a negative charge near the catalytic apparatus.
An explanation for the preference for the Arg substitution is
also discussed in Chapter 7.
Pd' position. The replacement of leucine in the parent
peptide resulted in superior substrates with respect to
catalytic efficiency, kcat- The kcat values increased 2 to
3-fold in comparison to the other substrates in this
substitution series. This observation was used in the design
of the substrates used in Chapter 7.
Lvs-Pro-Ile-P9-Phe*Noh-Arq-Leu Based Peptides
In contrast to the peptide series with Ala in P3, the
combination of lie in P3 and Leu in P2 resulted in a poorly
cleaved peptide (Table 6-2). The extension of the Asp side
chain to Glu results in a 10-fold increase in kcat-
Interestingly, rhizopuspepsin is able to cleave the Arg
substitution but is not able to readily cleave the substrate
with His in P2.
Discussion
Besides the few instances where the substitution of an
amino acid in the peptide resulted in an understandable
unfavorable interaction, rhizopuspepsin is a protease of
relatively broad specificity. Rhizopuspepsin is able to
efficiently cleave a wide variety of peptides. This ability
is in stark contrast to the mammalian enzymes with the same


77
their substitution into the rhizopuspepsin structure is not
expected to change the structure of the mutants
significantly. Added support for structural similarity
between the different rhizopuspepsins extends from the
observation that the mutation of surface residues of a
protein generally does not affect its overall fold.
Asp30 points into the active site cavity between the S3
and Si binding pockets. The base of the active site cleft is
made of a large P-sheet composed of strands from both the
N- and C-terminal domains of the protein. Studies by Katz
and Kossiakoff (1990) have shown that P~sheets undergo less
distortion than a-helices and loops upon mutation. This is
thought to be the result of an increased ability of the
P-sheet to dissipate strain energy through slight changes in
<})-Â¥ angles. These observations suggest that mutation at
position 30 will not causes significant changes in stability
or structure of the protein.
Asp77 is located in the P-hairpin turn of the flap which
extends over the active site. The flap region is thought to
be quite flexible. Upon inhibitor binding, particularly in
the HIV proteinase (Swain et al., 1990), the flap undergoes
movements to optimize hydrogen bonding and van der Waals
contacts. The decrease in the crystallographic thermal B
factors seen in the flap region upon inhibitor binding are
thought to be representative of this effect (Suguna et al.,
1987). Studies by Hurley et al. (1992) have shown that there
is a strong correlation between the B factor of the residue


78
in the wild-type protein and the change seen in stability of
the mutant. If the residue was originally not very mobile,
low B value, the stability of a mutant at this position would
decrease. This observation suggests that mutations at
position 77 of the flexible flap region should not greatly
affect overall stability of the protein.
All of the information discussed above suggests that the
structures of the mutants of rhizopuspepsin have the same
overall folding pattern with small positional changes in the
side chains. The discussion, however, still does not explain
why differences, particularly between the m and AG values
of the Asp30 lie and Asp30lle/Asp77Thr mutant proteins, are
observed when comparisons are made to the WT-REC and pi 6
proteins. One possible reason may be the differences between
the folded and denatured states of the proteins. Several
studies have shown that the m value is related to the solvent
accessible surface area of the denatured state (Schellman,
1978; Shortle & Meeker, 1986). Shortle and Meeker (1989)
have shown that there is a correlation between the m value
and the solvent accessibility of a mutant. Mutants that
exhibit m values less than the wild-type protein show more
compact structures and residual structural components when
compared to the wild-type proteins by size-exclusion
chromatography and circular dichroism. These observations
suggest that rhizopuspepsin mutants with lie at position 30
may be able to optimize hydrophobic interactions in the
denatured state when compared to enzymes with Asp in this


117
Table 7-4. Double mutant cycle analysis of the recombinant
rhizopuspepsins: substitution of lysine into P3-P1 at pH 3.5
and 5.0a
Peptide pH 3.5b pH 5.0
4 900 000
0.5
2 150 000
WT
D30I
3.4
3.5
D77T
DBL
19 000
0.6
7 100
1 240 000
0.6
460 000
WT
D30I
3.0
3.2
D77T
DBL
10 000
0.8
2 600
2 130 000
0.3
1 220 000
WT
D30I
2.7
2.5
D77T
DBL
25 000
0.1
20 300
810 000
0.5
340 000
WT
D30I
2.6
2.4
D77T
DBL
11 200
0.3
6 700
3
4
KPAAK*XRA
360 000
0.9
80 000
WT
D3 0I
6.9
7.0
D77T
DBL
5.0
1.0
0.9
KPAAF*XRA
2 170 000
0.9
490 000
WT
D30I
2.5
2.0
D77T
DBL
38 000
0.4
19 200
1 660 000
0.9
400 000
WT
D30I
7.2
7.2
D77T
DBL
13.7
0.9
3.4
4 730 000
1.0
940 000
WT
D30I
2.6
1.8
D77T
DBL
73 000
0.2
53 000
5
KPAAZ*XRA
1 000 000
0.6
390 000
WT
D30I
3.0
3.9
D77T
DBL
8 000
1.5
730
2 260 000
0.5
940 000
WT
D30I
3.0
3.8
D77T
DBL
16 000
1.3
2 100
aWT = wild-type recombinant, D30I = Asp30Ile, D77T = Asp77Thr, and DBL
= Asp30Ile/Asp77Thr. The numbers above and below the enzyme type are
the kcat/Km values in M*1s~1 from Table 7-1. The free energy changes of
transition state stablization, AAGj, are shown in italics with standard
deviations that range from 0.1 to 0.2 kcal mol'1.
bAAGj = -RTlnlk^ / ^(mutant, mutant 2 )//Tm(wild- type, mutant 1)]


40
analysis. This fragment was used as a retention time
standard for the CE runs to validate the correct site of
cleavage for each substrate and all forms of rhizopuspepsin
at pH 3.5 and 5.0.
Capillary electrophoresis kinetic assay. The initial
rates of cleavage of the peptide containing lysine in Pi by
AspWThr and Asp30lle/Asp77Thr were determined by incubating
the enzyme (900 nM) with either 50, 100 or 150 (1M substrate
for a period of twenty-four hours at 37C. A 20 |i.l aliquot
was taken at 0, 1, 3, 5, 7, 12, 15, and 24 hours and mixed
with 1.5 H-l of U85548E to give a final concentration of
inhibitor of 5 |JM (five-fold molar excess) and stored at
-20C until the electropherograms were run. The initial
linear slopes of the change in intact substrate peak area
versus time were converted to M s-1 by dividing by the slope
(peak area/[S]) of a standard curve generated from 30 to 1000
|iM of the substrate. The same injection and electropherogram
run parameters were used in the product analysis.
Analysis of Transition State Effects
In transition state theory, the energy difference
between the free enzyme and substrate and the transition
state, AGf, is related to the binding energy released upon
binding of the substrate, AGs, and the activation energy,
AG*, of the chemical steps responsible for bringing the


125
position can be altered by changing Met287 to Glu present in
porcine pepsin discussed above (Scarborough et al, 1994).
Besides the unique ability of the fungal enzymes to
cleave substrates containing lysine in the Pi position,
rhizopuspepsin and other fungal aspartic proteinases exhibit
broad specificity. This ability to digest or degrade a wide
variety of synthetic substrates may confer a selective
advantage to the organism in the cleavage of proteinaceous
food sources and the invasion of host tissues.
The study of the intermolecular interactions of
rhizopuspepsin between the enzyme and the substrate has given
further insight into how a proteinase discriminates between
substrates. In particular, this study has added to the
understanding of how charged residues and hydrogen bonding
play a role in creating biological specificity.
Future directions
The observations from this study have raised additional
questions related to the specificity differences between
fungal and mammalian aspartic proteinases. These questions
may be answered by additional mutagenesis of rhizopuspepsin
to the corresponding residues in the mammalian enzymes.
Rhizopuspepsin is able to readily cleave the peptide
containing Arg in the P3 position of the substrate series
studied in Chapter 6. Porcine pepsin, on the other hand,
requires an increase in the assay pH in order for efficient


95
The Ala substitution is cleaved but to a lesser extent. The
substitution of Val into this position results in a peptide
that is not cleaved readily. This can be rationalized
because of valine side chain proximity to the catalytic
aspartic acid Asp32. Analysis of the inhibitor complexes of
rhizopuspepsin shows that (^-branched amino acids cannot be
accommodated in this position because of steric restrictions
and potential collisions with Asp32.
PV position. There appears to be a preference for Ala
or Arg in this position. Because of the large difference in
size of these amino acids, this observation might appear to
be unusual. This trend can be rationalized with the aid of
the crystal structures and may be substantiated by the
results seen with the Asp substitution. The P2' side chain
of the substrate is bound into the S2' subsite of the enzyme
which is primarily created by residues that line the
underside of the flap. One strictly conserved residue in
this region that has been shown by site-directed mutagenesis
to have a large effect on specificity is Tyr75 (Suzuki et
al., 1989). The substitution of the small Ala side chain
should not be sterically restricted by Tyr75. The Arg is
probably tolerated in this position because of its high
degree of flexibility and length. An Arg side-chain may be
able to extend beyond Tyr75 and make possible interactions
with solvent. An Asp residue, however, is probably not able
to avoid steric constraints. Another possibility for the
poor cleavage of the Asp substituted peptide may be the


Table 7-1 continued
PEPTIDEb
PH
Asp77Thr
Asp30Ile/Asp77Thr
kCat
Km
kcat/Km
kcat
Km
kcat/Km
(s'1
(|iM)
(M-
is'1)
(s'1)
(pLM)
(M'i-s-1)
X10*4
XlO-4
3.5
6
1
561

115
1.00
0.30
_
_
C0.26 0.02
1
KPKAF*XRA
5.0
6
1
283

25
1.90
0.30
-
-
c0.71 0.04
3.5
6
1
487

83
1.12
0.30
-
-
c0.67 0.04
2
KPAKF*XRA
5.0
7
1
296

44
2.50
0.60
9
1 449
36
2.03 0.24
3.5
-
-
d5
.0
0.7 X lo*4
-
-
d0.9 0.1 x 10-4
3
KPAAK*XRA
5.0
-
-
d13
.7
2.1 X 10-4
-
-
d3.4 0.3 X 10-4
3.5
15
3
408

48
3.80
0.80
-
-
C1.90 0.10
4
KPAAF*XRA
5.0
17
2
228

17
7.30
1.20
13
1 253
25
5.30 0.70
3.5
2
0.4
297

41
0.80
0.20
-
-
c0.07 0.01
5
KPAAZ*XRA
5.0
5
1
289

39
1.60
0.40
1
0.1 462
62
0.21 0.04
aThe kinetic parameters ( standard deviations) were determined at 37'C with either 0.1 M sodium
formate, pH 3.5, or 0.1 M sodium acetate, pH 5.0, containing NaCl to maintain constant ionic
strength as outlined in Chapter 2.
bAbbreviations used in the substrates are: X = p -nitrophenylalanine; Z = norleucine; = site of
cleavage.
Ckcat/Km values were determined spectrophotometrically with the assumption that [S] Km
v = (kcat/Km)[E]o[S]o-
dkCat/Km values were determined by capillary electrophoresis; v = (kCat/Km)[E]o[S]q.
107


Table 6-1. Kinetic parameters for WT-REC rhizopuspepsin with the substrate
Lys-Pro-Ala-Lys-Phe*Nph-Arg-Leu containing systematic substitutions in
Ps-Pl
and
P2'-P3
' at
pH 3.5a
P5
P4
P3
P2
Pi*Pi 1
P2
P3
kcat
(sec"
1)
Km
(|XM)
kCat/Km
(s-J-M-1) X
10"6
Lys
Pro
Ala
Lys
Phe*Nph
Arg
Leu
12

1
11

1
1.13

0.20
Ser
5

1
4

1
1.21

0.32
Asp
5

1
7

3
0.76

0.29
Arg
6

1
3

1
1.74

0.76
Ala
7

1
6

1
1.14

0.17
Leu
5

1
7

2
0.82

0.27
Ser
9

1
2

0.4
0.42

0.09
Asp
22

3
30

6
0.73

0.18
Arg
3

0.3
43

10
0.06

0.02
Ala
15

1
19

1
0.75

0.08
Leu
22

2
14

1
1.57

0.17
Ser
9

1
16

2
0.60

0.08
Asp
13

1
20

2
0.65

0.09
Arg
13

2
9

1
1.39

0.25
Leu
10

1
7

1
1.37

0.21
Ser
9

1
9

1
1.03

0.18
Asp
7

1
8

1
0.83

0.12
Arg
9

1
4

0.9
2.17

0.56
Ala
20

1
8

2
2.39

0.46
Leu
9

1
4

0.7
2.31

0.53
Ala*Nph
4

0.4
24

4
0.17

0.03
Val*Nph
b
b
b
Leu*Nph
16

2
19

5
0.86

0.27
Ser
1

0.1
2

0.5
0.55

0.17
Asp
1

0.1
9

2
0.09

0.03
Ala
7

1
6

1
1.15

0.31
Leu
3

0.3
4

0.9
0.73

0.18
Ser
39

5
63

10
0.61

0.13
Asp
26

3
36

5
0.72

0.14
Arg
30

4
79

10
0.38

0.07
Ala
36

5
45

7
0.81

0.16
aAll peptides are eight residues in length.
bCleavage seen only after an extended incubation period with excess enzyme.


Figure 3-5. SDS-PAGE analysis of the time course of
activation of wild-type rhizopuspepsinogen upon lowering the
pH at room temperature. A, pH 2.0; B, pH 3.5.


74
Table 4-2. Guanidinium hydrochloride denaturation parameters
of native and mutant forms of rhizopuspepsina
Enzyme
[Gdn-
-HCl]50%
(M)
m
(kcal mol
_1 M-1)
AGu.V
(kcal mol-1)
pi 6
3.54

0.02
4.8

0.5
17.0
1.8
WT-REC
3.50

0.02
4.6

0.5
16.0
1.7
Asp30lle
3.31

0.02
3.3

0.2
10.9
0.8
Asp77Thr
3.33

0.02
4.5

0.4
15.1
1.4
Asp30lle/Asp77Thr
3.00

0.01
3.7

0.2
11.0
0.7
Parameters derived from the denaturation curves presented in
Figure 4-4. Each enzyme was studied as outlined in Chapter 2
from 0 to 6 M guanidinium hydrochloride in 0.1 M sodium
formate pH 3.0. All denaturant concentrations were performed
in duplicate. Values determined by the method of Jackson et
al. (1993); curve fit of observed fluorescence, F, versus
[denaturant]. The errors represented are the standard errors
from the KaleidaGraph program.


7-3. Transition state stabilization energy changes
seen with variation in pH from 3.5 to 5.0 for the
recombinant rhizopuspepsins and porcine pepsin 113
7-4. Double mutant cycle analysis of the recombinant
rhizopuspepsins: substitution of lysine into P3-
Pl at pH 3.5 and 5.0 117
IX


[ 0 ] deg cm2 dmol"
71
Wavelength (NM)
Figure 4-2. CD spectra of the recombinant wild-type
and mutant forms of rhizopuspepsin.


79
position. These new interactions may make the Asp30 mutants
less stable with the equilibrium shifting slightly in favor
of the denatured state.
This chapter has presented kinetic and structural
evidence that the recombinant wild-type enzyme accurately
represents the naturally occurring isozymes of
rhizopuspepsin. Structural studies also suggest that the
mutant enzymes exhibit similar structures to the wild-type
proteins. The kinetic analysis of inhibitors (Chapter 5) and
substrates (Chapters 6 and 7) with the WT-REC and mutant
enzymes should not be significantly affected by structural
deviations. Kinetic comparisons of the WT-REC to the mutant
forms of rhizopuspepsin also support the idea that the
structures have not been significantly altered. These
results are discussed in Chapter 7.


20
Table 1-3. Partial sequence alignment of
several aspartic proteinasesa
HPEP
28TWFDTGSSN37
7 4-pygtG7 8
PPEP
TVIFDTGSSN
TYGTG
CATE
TVIFDTGSSN
QYGTG
CATD
TWFDTGSSN
HYGSG
HREN
KWFDTGSSN
RYSTG
RCAP
NLDFDTGSSD
SYGDG
CAAP
NVIVDTGSSD
GYGDG
CTAP
TWIDTGSSD
EYGDL
CPAP
TVIIDTGSSD
RYGDG
aThis alignment was obtained using the PILEUP
program, a module in the GCG Sequence
Analysis Software Package (Devereux et al.,
1984) The sequences are HPEP = human pepsin
(Sogawa et al., 1983), PPEP = porcine pepsin
(Chen et al., 1975), CATE = human cathepsin E
(Azuma et al., 1992), CATD = human cathepsin
D (Faust et al., 1985), HREN = human renin
(Hobart et al., 1984), RCAP = Rhizopus
chinesis aspartic proteinase (Chen et al.,
1991) CAAP = Candida albicans aspartic
proteinase (Hube et al., 1991), CTAP =
Candida tropicalis asapartic proteinase
(Togni et al., 1991), CPAP = Candida
parapsilosis (de Viragh et al., 1993).


121
A
B
FLAP
Figure 7-3. Proposed hydrogen bonding interactions in WT-REC
rhizopuspepsin and Asp77Thr mutants with substrates
containing lysine in Pi. A, WT-REC; B, Asp77Thr mutants.


104
This chapter outlines the results achieved by mutating
residues Asp30 and Asp77 to the corresponding residues in
porcine pepsin, Ile30 and Thr77. The resulting mutant
proteins were overexpressed in E. coli and shown, after
refolding and activation, to be structurally similar to the
wild-type recombinant enzyme and the native isozyme pi 6
(Chapter 3 and 4). The recombinant rhizopuspepsins were
tested with oligopeptide substrates containing systematic
substitutions of lysine into the Pi, P2 and P3 positions.
Double mutant cycle analysis was performed to definitively
identify the residue or residues responsible for the ability
to cleave carboxy-terminal to a lysine residue and to
estimate the energy contributions to this specificity through
hydrogen bonding interactions (Carter et al., 1984; Akers &
Smith, 1985; Fersht, 1985; Wells, 1990; Mildvan et al.,
1992).
Results
Kinetic Analysis of the Recombinant Rhizoousoeosins
The summary of the kinetic analysis of the recombinant
rhizopuspepsins with substrates with the systematic
substitution of lysine into the P3-P1 positions is shown in
Table 7-1. All peptides were cleaved at the expected
positions, that is, between Phe*Nph, Nle*Nph or Lys*Nph. The
enzymes show consistent trends in the kinetic parameters for


porcine pepsin, Ile30 and Thr77, to create single and double
mutants. The zymogen forms of the wild-type and mutant forms
of the enzymes were overexpressed in E. coli as inclusion
bodies. Following denaturation, refolding, activation and
purification to homogeneity, structural and kinetic
comparisons were made. These comparisons have shown that the
wild-type recombinant enzyme is kinetically and structurally
indistinguishable from the naturally occurring isozymes. The
mutant proteins were also shown to exhibit a high degree of
similarity.
Characterization of the wild-type and mutant proteins
with chromogenic substrates with systematic substitution of
lysine into the Pi, P2 and P3 positions has shown that Asp77
is the predominant residue responsible for enabling the
catalysis of substrates with lysine in Pi. Mutation of Asp77
resulted in a loss of 7 kcal mol-1 of transition state
stabilization energy. The Asp30lle mutant was still able to
cleave the Pi-Lys peptide with near wild-type efficiency.
These observations suggest that it may be possible to exploit
the Asp77 interaction to design compounds that preferentially
inhibit a variety of related, secreted Candida proteinases in
immunocompromised patients.
xv


58
8.7-
8.5-
8.2-
7.4-
M
6.9-
2 3 4
3.5-
Figure 3-7. IEF analysis of the wild-type and mutant
rhizopuspepsinogens activated at pH 2.0. Samples analyzed
prior to ion-exchange chromatography. Lane 1, wild-type;
lane 2, Asp30lle; lane 3, Asp77Thr; lane 4,
Asp30lle/Asp77Thr.


CHAPTER 5
ANALYSIS OF THE SPECIFICITY OF RHIZOPUSPEPSIN THROUGH THE USE
OF INHIBITORS CONTAINING Pi AND Pi' SUBSTITUTIONS AND SCISSILE
BOND MIMETICS
INTRODUCTION
The inhibitors used in this study are based on the
cleavage sites in angiotensinogen, His-Pro-Phe-His-Leu*Val-
Ile-His-Asn, and the pl7/p24 HIV polyprotein junction Val-
Ser-Gln-Asn-Tyr*Pro-Ile-Val. Many potent renin and HIV
proteinase (HIV-PR) inhibitors have been generated using
these cleavage sites and the natural product pepstatin as
leads (Wiley & Rich, 1993). Studies varying side chain
functionalities, in order to probe possible "secondary
interactions" (Fruton, 1970; Medzihradszky et al., 1970) in
the enzyme binding subsites, inhibitor length, and
nonhydrolyzable peptide bond analogs have been undertaken to
optimize potency and bioavailability (Rosenberg et al., 1990;
Sawyer et al., 1991; Wiley and Rich, 1993). Other approaches
to inhibitor design have involved molecular modeling based on
the results from crystal structures of enzyme-inhibitor
complexes (Kempf et al., 1990; Thompson et al., 1992). The
results from crystal structures also aid in the analysis and
rationalization of differences seen for inhibitor
interactions and substrate specificities between family
80


110
increases without changes in kcat- Importantly, hydrolysis of
the Pi-lysine substrate (3) by porcine pepsin did not occur
between Lys*Nph. After an extended incubation period with
excess enzyme, cleavage was seen between Nph-Arg.
Discussion
Site-directed mutagenesis has become an invaluable
technique in dissecting the molecular interactions which give
rise to the binding, discrimination and catalytic functions
of proteins. The use of double mutant cycles and the
analysis of additivity between mutants can give information
regarding the contribution of an amino acid residue to a
particular function of a protein, as well as to its own
stability (reviewed by Wells, 1990; Shortle, 1992). In this
study, site-directed mutants were made to investigate the
contributions of Asp30 and Asp77 to the unique ability of
rhizopuspepsin to cleave substrates with lysine in the Pi
position. The use of rhizopuspepsin as a model system for
studying active site interactions of fungal aspartic
proteinases is facilitated by: (1) high resolution crystal
structures of the native and inhibited forms of the enzyme,
(2) ease of cloning and expression in E. coli, and (3)
sensitive kinetic methods to quantify the effects of the
mutations (Rheinnecker et al., 1993).


Ill
Substrate Pesian
The initial experimental evidence that the fungal
aspartic proteinases were distinct from the mammalian enzymes
came from the observation that they were able to activate
trypsinogen at a Lys-Ile bond (Graham et al.f 1973; Morihara
& Oka, 1973). Hofmann and his coworkers have confirmed this
specificity by using substrates of the form Ac-(Ala)m-
Lys*Nph-(Ala)n-amide to study the effects of secondary
substrate binding interactions (Hofmann & Hodges, 1982;
Hofmann et al., 1984; Balbaa et al., 1993). The substrates
used in this study are slightly different and were derived
from the highly soluble peptide Lys-Pro-Ala-Lys-Phe*Nph-Arg-
Leu. This substrate and its P5-P1 and P2'-P3' systematic
substitution derivatives have been extensively used in the
exploration of subsite specificities of rhizopuspepsin
(Chapter 6), pepsin (Pohl and Dunn, 1988; Rao-Naik,
unpublished data), cathepsin E (Rao-Naik, unpublished data)
and cathepsin D (Scarborough et al., 1993). Rhizopuspepsin
is able to cleave all the substrates in this series with
nearly equivalent specificities (kcat/Km). Changing the P3'
Leu to Asp, Ala, Arg or Ser, however, caused a two-fold
increase in kcat. This information suggested that the
substitution of Ala in P3' would be advantageous for the
study of mutants that may have decreased catalytic activity.
As illustrated in Figure 1-3, Asp77 has the potential to
interact with the P3-P1 residues of a ligand bound in the


LIST OF REFERENCES 129
BIOGRAPHICAL SKETCH 138
Vll


CHAPTER 6
THE BROAD SUBSTRATE SPECIFICITY OF RHIZOPUSPEPSIN: ANALYSIS
WITH SYSTEMATIC SUBSTITUTIONS IN THE P5-P1 AND P2-P3'
POSITIONS OF THE SUBSTRATE LYS-PRO-ALA-LYS-PHE*NPH ARG-LEU
INTRODUCTION
The residues that line the active sites of enzymes play
a critical role not only in catalysis but also in creating
the environment for the discrimination between substrates
(Fersht, 1985). This selection process or specificity is
governed by a myriad of favorable and unfavorable
interactions. Favorable intermolecular interactions include
hydrogen bonds, solvent exclusion from hydrophobic surfaces
and van der Waals interactions. Energetically unfavorable
interactions include accommodating suboptimal substrates with
distorted or lost hydrogen bonds to the enzyme and cavity
formation.
The specificity of the aspartic proteinases has been
under intense study in the effort to understand interactions
that distinguish each enzyme in this family. The goal has
been to exploit these differences in order to create
exquisitely targeted therapeutics, especially for renin and
HIV proteinase. Many early studies have shown the importance
of the 'secondary interactions* of the S3 and S2 subsites in
specificity and catalysis (Fruton, 1976; Dunn et al., 1987;
91


44
overexpressed in E. coli using the pET expression system
(Studier et al., 1990). Upon addition of IPTG to the
bacterial cultures, the proteins were expressed at high
levels in the form of inclusion bodies. Figure 3-1 shows an
expression time course experiment for wild-type
rhizopuspepsinogen. By the end of three hours,
rhizopuspepsinogen was one of the predominant proteins. The
inclusion bodies were purified from the cell lysate by
centrifugation through a 27% sucrose cushion. The resulting
pellet was washed with buffer containing triton X-100 to
remove any remaining cellular debris. Typical yields ranged
from 0.5 to 1 g (wet weight, 5 to 18 % of the total cell
pellet) for a 4 L expression. The purified inclusion bodies
were judged to be >95 % pure zymogen with a molecular weight
of 43kDa by SDS-PAGE (Figure 3-2).
The inactive zymogen inclusion bodies, containing
approximately 100 mg of protein, were refolded from 8 M urea.
Two refolding methods were tried in order to maximize yields
for subsequent activation to the mature enzyme. Refolding by
the rapid dilution procedure of Chen et al. (1991) resulted
in primarily polymeric material that was difficult to
completely activate into an active, monomeric state (data not
shown). Even though there was some precipitation, the
largest yields of monomeric zymogen capable of activation
were obtained when the proteins were refolded by dialysis.
Figure 3-3 shows the elution profile from gel filtration
chromatography of the refolded rhizopuspepsinogen.


KEY TO SYMBOLS
amp
ampicillin
AU
absorption units
C
carboxyl
Ca
alpha carbon
CAPS
3-(cyclohexylamino)-1-propanesulfonic acid
CD
circular dichroism
d
deoxy
DMSO
dimethylsulfoxide
E. coli
Escherichia coli
EDTA
ethylenediaminetetraacetic acid
FPLC
fast protein liquid chromatography
h
Planck's constant
HCl
hydrochloric acid
HIV
human immunodeficiency virus
HPLC
high performance liquid chromatography
IEF
isoelectric focusing
IPTG
isopropylthio-P~D-galactopyranoside
kcal
kilocalorie
kb
Boltzmann's constant
kcat
turnover number
KCl
potassium chloride
Ki
inhibition constant
km
Michaelis-Menten constant
LB
Luria Broth
MES
2-(4-morpholino)-ethane sulfonic acid
mg
milligram
MgCl2
magnesium chloride
min
minutes
Xll


LIST OF FIGURES
Figure page
1-1. Ribbon representation of the aspartic proteinase
rhizopuspepsin complexed with a reduced peptide
bond inhibitor 9
1-2. Closeup view of the reduced peptide bond
inhibitor bound to the active site of
rhizopuspepsin 11
1-3. Closeup view of the active site of rhizopuspepsin
highlighting the catalytic aspartic acid
residues, Asp32 and 215, and Asp30 and Asp77 21
2-1. PCR mutagenesis procedure 26
2-2. pET3aE expression vector containing
rhizopuspepsinogen 29
3-1. SDS-PAGE analysis of the expression of wild-type
rhizopuspepsinogen (RPGN) in E. coli upon the
addition of IPTG from 0 to 3 hours 46
3-2. SDS-PAGE analysis of wild-type rhizopuspepsinogen
at different stages of purification 47
3-3. Gel filtration elution profile of refolded
Asp30lle rhizopuspepsinogen 49
3-4. SDS-PAGE analysis of the fractions from the
purification of Asp30lle by gel filtration in
Figure 3-3 52
3-5. SDS-PAGE analysis of the time course of
activation of wild-type rhizopuspepsinogen upon
lowering the pH at room temperature 54
3-6. SDS-PAGE analysis of the activation of wild-type
recombinant rhizopuspepsinogen at pH 4.0 56
3-7. IEF analysis of the wild-type and mutant
rhizopuspepsinogens activated at pH 2.0 58
3-8. Elution profile of Asp77Thr from the Mono S
column after activation at pH 3.0 61
x