MIAMI UNIVERSITY THE GRADUATE SCHOOL
CERTIFICATE FOR APPROVING THE DISSERTATION
We hereby approve the Dissertation
of
Frank Carl Golich
Candidate for the Degree:
Doctor of Philosophy
______Dr. Michael W. Crowder, Director
______Dr. Hongcai Zhou, Reader
______Dr. Neil Danielson, Reader
______Dr. Christopher A. Makaroff, Reader
______Dr. Michael Robinson, Graduate School Representative
ABSTRACT
STRUCTURAL AND FUNCTIONAL CHARACTERIZATION OF AMINOPEPTIDASE N (PEPN) FROM Escherichia coli
by Frank C. Golich
Aminopeptidases are ubiquitous hydrolytic enzymes that hydrolyze the N-terminal amino acids from peptides/proteins. Aminopeptidase N (PepN) primarily cleaves basic and hydrophobic residues from peptides. PepN in mammals is membrane-associated and has been implicated in viral infections and in tumor cell invasion. Bacterial PepNs are cytosolic, and E. coli PepN is the sole alanyl aminopeptidase in this organism. Bacterial proteinases have been suggested to be potential targets for the generation of novel antibiotics. Before inhibitor design efforts can be made, the structural and mechanistic characterization of the target is required. Escherichia coli PepN was cloned into pET26b, over-expressed in E. coli, and purified using Q-Sepharose chromatography. This protocol yielded over 17 mg of purified, recombinant PepN per liter of culture. Gel filtration chromatography revealed that PepN exists as a monomer. MALDI-TOF mass spectra showed that the enzyme has a molecular mass ca. 99 kDa. Metal analyses demonstrated that as-isolated, recombinant PepN binds 0.5 and <0.1 equivalents of iron and zinc, respectively. The addition of Zn(II) to recombinant PepN inhibits catalytic activity, while the addition of iron causes little change in activity. Further metal binding studies revealed that recombinant PepN tightly binds 5 equivalents of iron and < 0.1 equivalents of Zn(II). CD and fluorescence studies revealed that iron is not involved in the structure of PepN. EPR, NMR, and UV-Vis studies revealed that one of the irons binds at a site containing 2 histidines and at least 1 cysteine as ligands. Chemical modification studies suggest that cysteine is involved in the catalytic mechanism. pH Dependence and solvent isotope studies indicate several rate-limiting proton transfers, which can be tentatively assigned to cysteine and histidine residues. Stopped- flow kinetic studies, along with kinetic simulations, demonstrate that PepN utilizes a mechanism with ordered release of products and a rate-determining chemistry step. The first reaction mechanism for E. coli PepN is proposed. The results of this work can be used to guide future inhibitor design efforts, and these inhibitors may prove to be novel antibiotics.
STRUCTURAL AND FUNCTIONAL CHARACTERIZATION OF AMINOPEPTIDASE N (PEPN) FROM ESCHERICHIA COLI
A DISSERTATION
Submitted to the Faculty of
Miami University in partial
fulfillment of the requirements
for the degree of
Doctor of Philosophy
Department of Chemistry and Biochemistry
by
Frank Carl Golich
Miami University
Oxford, Ohio
2005
Dissertation Director: Dr. Michael Crowder
TABLE OF CONTENTS
Chapter 1: Introduction 1
Peptidases 1
Aminopeptidases 1
Occurrence 1
Biological Roles 2
General Classification 2
Classification Schemes 6
Aminopeptidase Inhibitors 8
Quaternary structures of aminopeptidases 11
Metallo-aminopeptidases 12
Mechanistic studies on aminopeptidases 13
VanX 13
Aminopeptidase N 15
Sections of this dissertation 16
References 17
Chapter 2: L-Alanine-p-nitroanilide is not a Substrate for VanX 28
Introduction 29
Materials and Methods 30
Preparation of VanX 30
Identification of contaminating enzyme 31
Coupled assay for VanX 31
ii Assays with protease inhibitors 32
In-gel trypsin digestions/MALDI-TOF mass spectrum 32
Results 33
L-ala-p-nitroanilide hydrolase activity due to a contaminating 33 enzyme
Aminopeptidase N is the contaminating enzyme 34
Discussion 35
References 37
Chapter 3: Over-expression, Purification, and Characterization of 43 Aminopeptidase N (PepN) from Escherichia coli
Introduction 44
Materials and Methods 45
Cloning of pepN gene 45
Over-expression and purification of recombinant PepN 45
Metals analyses 46
Steady-state kinetics 46
Gel-Filtration chromatography 46
Amino Acid Analysis 47
MALDI-TOF MS 47
Results 47
Discussion 49
References 51
iii
Chapter 4: Spectroscopic studies on recombinant E. coli PepN 57
Introduction 57
Materials and Methods 58
Preparation of PepN 58
Preparation of Apo-PepN and Fe-containing PepN 59
Steady-state kinetics 59
Circular dichroism spectroscopy 60
Fluorescence emission spectroscopy 60
Electronic absorption spectroscopy 60
NMR spectroscopy 60
EPR spectroscopy 60
Results and Discussion 61
Effect of Fe binding 61
UV-Vis spectrophotometry 63
EPR spectroscopy 64
1H-NMR Spectroscopy 67
Conclusions 68
References 70
Chapter 5: Kinetic and mechanistic studies on recombinant PepN 81
Introduction 81
Materials and Methods 82
Preparation of PepN 82
iv Steady-State Kinetics 82
Chemical modification of PepN with iodoacetate 83
pH Dependence studies 83
Proton inventory studies 83
Stopped-flow kinetics 83
Kinetic simulations 84
Inhibition studies 84
Results and Discussion 84
Chemical modification 84
pH Dependence studies 85
Solvent isotope studies 86
Inhibition of PepN by L-alanine and p-nitroaniline 87
Stopped-flow kinetic studies 88
Kinetic simulations 88
Inhibition studies on PepN 89
Conclusions 90
References 92
Chapter 6: Summary and Conclusions 103
References 105
v
LIST OF TABLES
Chapter 1: Introduction
1. Selected Aminopeptidases in Nature 4
Chapter 2: L-Alanine-p-nitroanilide is not a Substrate for VanX
1. Protease inhibitors tested 42
Chapter 3: Over-expression, Purification, and Characterization of Aminopeptidase N (PepN) from Escherichia coli
1. Purification of recombinant E. coli PepN 55
Chapter 4: Spectroscopic studies on recombinant E. coli PepN
Chapter 5: Kinetic and mechanistic studies on recombinant PepN
1. Inhibitors of PepN 102
vi
LIST OF FIGURES
Chapter 1: Introduction
1. Peptidase classification 1
2. Selected inhibitors of aminopeptidases 10
3. Crystal structure of bleomycin hydrolase 11
Chapter 2: L-Alanine-p-nitroanilide is not a Substrate for VanX
1. SDS-PAGE gel of samples characterized in this work 41
Chapter 3: Over-expression, Purification, and Characterization of Aminopeptidase N (PepN) from Escherichia coli
1. SDS PAGE gel (5.2% acrylamide) of recombinant PepN from 56 pET26b-PepN over-expression system
Chapter 4: Spectroscopic studies on recombinant E. coli PepN 1. CD spectra of as-isolated PepN and Fe-added PepN 74
2. Fluorescence emission spectra of as-isolated PepN and after 75 addition of 1, 2, 3, 4, 5, 6, and 10 equivalents of Fe(II)
3. UV-Vis difference spectra of as-isolated PepN titrated with Fe(II) 76
4. EPR spectra of as-isolated PepN 77
5. EPR spectra of as-isolated PepN containing 5 equivalents of Fe 78
6. 1H NMR spectra of PepN 79
7. Proposed iron centers in recombinant E. coli PepN 80
vii Chapter 5: Kinetic and mechanistic studies on recombinant PepN
1. Matrix Science Mascot search results for trypsin digestion 96 of aminopeptidase N (PepN) from E. coli
2. pH dependence plots of as-isolated PepN with 97 L-alanine-p-nitroanilide
3. Proton inventory of recombinant PepN at pH(D) 7.0 98
4. Kinetic simulations of stopped-flow data using KINSIM 99
5. Proposed kinetic mechanism of PepN 100
6. Proposed reaction mechanism of E. coli PepN 101
viii
Acknowledgements
I am all too aware that no one can become the person they are without the influences of many people. This is definitely true in my case. To be more specific, research is impossible without the efforts of many people. It is, therefore, my desire to thank not only those individuals who helped directly in the completion of this research, but also those people; friends, acquaintances, family, etc., who helped me become the person I am. In so doing, they have contributed both tangibly and intangibly to the success of this research. It goes without saying that I could not have done it without them.
The titles and ranks of some may have ultimately changed as their careers progressed, but they reflect those at the time they were most influential to me. In some cases, out of respect to their accomplishments and in the interest of being accurate, they have been changed to reflect their current status. Please, however, do not let any inaccuracies or omissions infer disrespect on my part. I am humbled by all who have taken their time to mold me, and my only wish is to repay them by this most modest gesture of appreciation and thanks.
Mr. Eugene Swetin, Mr. Norm Mills, Mr. Mark Barder, Mr. & Mrs. Jon Melhus, Ms. Björg Skaug, Mr. Frederick Wehrenberg Sr., Mrs. Nancey Lowenberg, Technical Sergeant Leo Williams, USAF, Staff Sergeant Donald Whitaker, USAF, Chief Master Sergeant William W. Wood, USAF (Retired), Norman Gevitz, Ph.D., Mark Zerwic, Ph.D., Mr. Glenn Shurney, Mr. Eric Hartman, August “Gus” Fiebig, Ph.D., Walter Eisenberg, Ph.D., Mr. Kevin Taylor, Ms. Louise Brusek, Chief Master Sergeant Fred Jacobson, USAFR, Mr. William “Ed” Pearson, Mr. George Michelini, Mr. Joseph “Joe” Giertych, Mrs. Cathy Herman, Theodore Roseman, Ph.D., Ms. Dorothy “Dotty” McMillin, Ms. Marina Cosgrove, Ed Gatliff, Ph.D., Mrs. Nina Fouts, Mrs. Susan Boyd, James Maas, M.D., Martin “Marty” Javors, Ph.D., Mrs. Jill Twardowski, Mrs. Jeanette Baker, Mr. David “Heavy D” Brown, Colonel Roger “Don” Stork, Ph.D., USAF, Wesley Baumgardner, Ph.D., Major George Kemper, Ph.D., USAF, Technical Sergeant Paul Valenzuela, USAF, Lieutenant Colonel
ix Christopher Phillips, M.D., USAF, Lieutenant Colonel Roberto Penne-Cassanova, M.D., USAF, Colonel John Bishop, M.D., USAF, Mr. Paul Labbe, Ms. Roberta “Bobbie” Dixon, Mr. David Chapman, Mrs. Shari Lewis-Hall, Mrs. Lisa Kobb, Beata Musial, Ph.D., Abdul- Qawi Numan, Ph.D.
In addition, I would like to express my sincere thanks to the past and present members of the Crowder Research Group: Patrick Crawford, Ph.D., Jeff Brandt, Ph.D., Ke- Wu Yang, Ph.D., Anne Carenbauer, Lisa Chatwood, Matt Breece, Lissa Herron, James Garrity, Ph.D., Nathan Wenzel, Rob Yates, Dan Sobieski, Sowmya Chandrasekar, Narayan Sharma, Allen Easton, Pattraranee Limphong, Patrick Hensley, Peter Wu, Priya Gursahaney, Renee Cilliers, Megan Matthews, James “P-Ditty” Pauff, and Maria Han. Their help and cooperation enabled me to learn and function in an area totally foreign to me prior to initiating this research. They have shone the light into some areas that were very dark. I would also like to express my love and sincere gratitude to my friends and colleagues Tara Sigdel, Ph.D. and Periyannan Gopal Raj, Ph.D. If people are fortunate, they may have an opportunity in their lifetime to meet one person destined for greatness. I believe I have met two such individuals in the persons of Tara and Gopal. I constantly marvel at how knowledgeable and perceptive and yet how humble they are.
For invaluable technical support I would like to thank Krishnan Damodharan, Ph.D., Ian Peat, Ph.D., Dr. Brian Bennett, Ph.D., Ms. Meghan Holdorf, John Hawes, Ph.D., Vilas Shukla, Ph.D., Alex Pisarenko, Christine Hajdin, and Ethan Carp.
I wish to thank my research committee: Dr. Neil Danielson, Dr. Christopher Makaroff, Dr. Hongcai “Joe” Zhou, Dr. Michael Robinson, and Dr. Michael Crowder as well as past committee members: Dr. Gilbert Gordon, Dr. David Pennock. You have my deepest respect and admiration, and I am honored to have had each of you on my committee. I also wish to extend a special thank you to Dr. Danielson for his considerate mentorship in the analytical teaching laboratory and in assisting me with my original research topic. Many thanks also go out to Dr. “Joe” for his tireless assistance with the advanced inorganic chemistry laboratory and his constant words of encouragement. I would like to offer my most
x sincere thanks to my research advisor, Professor Michael Crowder. In my opinion, the most valuable rewards from pursuing a course of graduate study come not only from the professional and academic exposure, although those do come. However, they also result from the interpersonal relationships which are developed. I was fortunate to have had the opportunity to experience both an intellectually and personally rewarding kinship with Dr. Crowder. It seems that I have had a doubly rewarding experience as a result of our numerous conversations which spanned the gambit from chemistry and current events to the Civil War, politics and religion. I could not have designed a better program of study, and I am not ashamed to admit that I feel fortunate, indeed blessed, to have had his mentorship.
My heartfelt love and appreciation go out to my family beginning with my parents. They never accomplished much by this world’s standards, but they left me with gifts that are impossible to articulate. Included among these is an unquestionable sense of being loved and respected by them both. Two of the most important things I can remember being taught by them were the concepts of “looking it up” and “doing your best.” These simple concepts have instilled in me a sense of curiosity, tempered by the desire to not take things at face value. They also nurtured in me the belief that if something is worth having, it is worth working for. I know they would have been proud of my accomplishments.
My brothers and sisters in Christ, especially Marcie and Al Simon, Beth and Tim Massie, Pastor Doug Moss, Dr. Don and Patti Ruegsegger, Ms. Ann Ruegsegger, Mrs. Heather Loper, Mrs. Becky McCaskey, Ms. Hannah DeLange, Elvis Tiburu, Ph.D., Martin Waichigo, Ph.D., and Bill Gibbons. You were hands and feet, and were doers, not just hearers.
My most sincere thanks go to my wife Bonnie and my daughters Dana Marie and Emily Mae. They are the bedrock of my life. They have sacrificed in ways too numerous to count in order that I could pursue this goal. They have demonstrated to me the concept of unconditional love. It was the birth of our girls that has led me to believe that the miracle of each new life is God’s way of giving us a glimpse of His perfection.
xi And to my Lord, and Savior, Jesus Christ, who has preserved me throughout this season.
xii Chapter 1 Introduction
Peptidases Peptidases are enzymes that cleave proteins or peptides into smaller peptides (1-3). These enzymes have been broadly categorized as exopeptidases or endopeptidases depending upon whether the peptidases cleave at the termini of the substrate or cleave interior peptide bonds, respectively (1). Exopeptidases are often further classified as aminopeptidases (APs) and carboxypeptidases. Carboxypeptidases cleave at the C-terminus of polypeptides; analogously, APs are enzymes that selectively cleave amino acid residues from the N-terminus (1) (Figure 1).
(2) (1) (3)
H2N X1 X2 X3 X4 X5 X6 COOH
Figure 1. Peptidase classification. Endopeptidases cleave peptide bonds within the polypeptide chain (1). Exopeptidases cleave residues located at the N-terminal position (2) (Aminopeptidases) or C-terminal position (3) (Carboxypeptidases) of a polypeptide. Adapted from (1).
Aminopeptidases Occurrence. Aminopeptidases are ubiquitous in nature, and numerous varieties have been identified within the plant, animal, and bacterial kingdoms. Many are metalloenzymes, and zinc (Zn(II)) appears to be the most common metal ion cofactor in metallo-APs (4, 5). Most of the well-characterized enzymes have molecular masses between 20-30 kDa (6), although some lysosomal enzymes have masses greater than 1,000 kDa (6). According to Gonzales et al., almost half (47%) of the characterized APs are monomeric, while the remaining 53% take on a multimeric form (1). It has been estimated that irrespective of structure, the majority (97%) of APs are found in soluble fractions, with 65% being found in the cytoplasm, and 16% secreted to the external environment (1). Aminopeptidases occur in both soluble and membrane-associated forms and may be exported to the extracellular environment or maintained within various intracellular compartments depending upon their designated function (2). Roles for the various
1 aminopeptidases include post-translational modification, protein maturation, degradation of hormonal and signal peptides, and nutrient recycling (4, 5). Many intracellular proteins require some form of amino-terminal modification following translation. Removal of methionine from the N-terminus by aminopeptidases is required in eukaryotes (N-formyl methionine in prokaryotes), and this function is accomplished by a form of methionyl aminopeptidase (MAP or MetAP) that is found in most eukaryotes and prokaryotes (7). Research has also focused upon a possible role of aminopeptidases in a final step of ubiquitin-dependent protein degradation (8).
Biological Roles. Aminopeptidases have been implicated in a number of disease states including SARS (9), malaria (10, 11), coronavirus, HIV (12), and cancer (13). Aminopeptidases also appear to be involved in various immunological conditions such as systemic lupus erythematosus, collagenase and connective tissue disorders (14), Alzheimer’s disease (15), and bacterial and viral infections (16), and some aminopeptidases may play a role in various inflammatory mechanisms (17). Most prokaryotes utilize proteolytic enzymes in order to access nutrients for growth and proliferation, but it is precisely their ability to degrade tissues and interfere with host defenses that increases their invasiveness (18). As a result, bacterial proteinases and metallopeptidases have shown promise as potential drug targets for combating bacterial infections. The emergence of numerous antibiotic-resistant, pathogenic bacteria has necessitated the search for second-generation antibiotics as well as additional target sites within bacteria (19). The study of a number of mammalian counterparts of bacterial metallopeptidases and their inhibition is currently ongoing as some of these enzymes have been implicated in human pathological conditions (19, 20).
General Classification. Aminopeptidases catalyze the cleavage of amino acids from the N-terminus of proteins and peptide substrates (5). Their expression may be induced by high temperature, oxygen limitation or anaerobiosis (21), or carbon, nitrogen, and phosphate starvation (1, 2, 22). The expression levels of several APs in plants have been shown to be enhanced during development and growth, whereas the levels of several aminopeptidases in Lactobacillus were diminished during stationary growth phases (23). These enzymes are clearly essential, as mutational analyses have often proven lethal to the host organisms (24).
2 Aminopeptidases have been categorized as processive or nonprocessive depending upon the specificity with which they hydrolyze substrates. Processive aminopeptidases will continue to cleave amino acid residues from a polypeptide until the enzymes reach an unfavorable residue or combination of residues, resulting in substrate release (2). Nonprocessive aminopeptidases, such as methionine aminopeptidase (2, 4), are more highly-specific for a specific residue and will not hydrolyze a substrate unless that specific amino acid is present (Table 1). Likewise, nonprocessive APs typically will not continue to degrade a substrate after the preferred residue has been removed (4). Many aminopeptidases show a marked preference for certain types of amino acids in the N-terminal position of the peptide substrate. This substrate preference is often related to the biochemical properties of the N-terminal residue (charged, neutral, or hydrophobic, etc.). In some cases the substrate specificity of a given aminopeptidase is dependent upon both the terminal and the penultimate amino acids present, and peptide cleavage is facilitated by both (24). Yeast methionine aminopeptidase, for example, was shown to cleave the terminal methionine from a Met-Ala-Ser tripeptide. It also cleaved Met from other tripeptides whose penultimate residues were small and / or uncharged (e.g., Pro, Gly, Val, Thr, or Ser) but not bulky and / or charged (e.g., Arg, His, Leu, Met, or Tyr) (7). Peptide cleavage in many instances depends upon the amino acid residues that flank the cleavage site (25). Structural features in and around the cleavage site may also determine the specificity with which the aminopeptidase reaction occurs. Bovine lens aminopeptidase (bLueAP) and Aeromonas proteolytica aminopeptidase (ApAP) have both been classified as
broad range aminopeptidases and cleave substrate without regard for the penultimate (P1’) residue (25). Some aminopeptidases, referred to as “pita bread” enzymes because of their unique structural folds, are more restricted in terms of the N-terminal and P1’ residues with which they will react (25). Methionyl aminopeptidase and aminopeptidase P (APP) fall into this category (25).
3 Table 1. Selected Aminopeptidases in Nature.
Aminopeptidase Other Names Class a Reactive Substratesb M.W. Organism Known Refs. Groups (kDa)c Inhibitors
Leucine AP LAP M 2 Zn(II) Leucine, carbonyl, 32 A. protelytica L-Leucine (5, 26) thionyl-containing phosphonic substrates acid
Methionine AP(I), (II) MetAP-I (II), MAP M 2 Co(II) or N-terminal 43 (I) E. coli, Methionine (2, 27) 1 Co(II), Methionine S. cerevisiae phosphonate 1 Zn(II) (Met-X) where: (II) H.sapiens X=small, uncharged amino acids residues
Aminopeptidase A PepA, XerB M 2 Zn(II) ArgR protein, 55.3 E. coli (5, 28, Cysteinyl-glycine, S. typhimurium Amastatin 29) large peptides
Aminopeptidase B PepB C Thiol Cysteinyl-glycine E. coli Bestatin, (30) “B”= Basic AAs Zn(II)* Amastatin
Aminopeptidase M PepM, membrane- M E. coli (30) bound AP Rat Kidney
Aminopeptidase N PepN, “N”=Neutral ?(C) ? N, B, A 99 Lactococcus/ Bestatin, Lactobacilli (31-33) aminopeptidase, Zn(II)* Cbz-Leu-Leu-Gly- Amaststin “Naphthylamide- βNA E. coli cleaving” AP H. sapiens Lysyl AP Yeast Mammals
Bleomysin Hydrolase BLH1, pepC, amino- C Thiol Bleomycin, poly- 55.4-H L. lactis Leupeptin (34) peptidase C glycine peptides: 330kDa H. sapiens tetra>tri>dipeptides Rabbit Lung
X-Pro dipeptidyl PepX S Serine X-Pro-Y peptides 88 Lb. debrueckii PMSFd (1) aminopeptidase hydroxl Where: X=any AA, ssp. lactis (DPAP) Pro=Proline,
Y=any AA
Enterococcus (35-37) VanX D-Alanyl-D-Alanine M 1 Zn(II)- D-ala-D-ala 23 Phosphinate, dipeptidase OH dipeptides faecium Phosphonate
aM - Metalloaminopeptidase, S - Serine aminopeptidase, C - Cysteine Aminopeptidase bN - Neutral amino acids, B - Basic amino acids, A - acidic amino acids βNA - β-Naphthylamide derivative pNA - p-Nitroanilide derivative cSuffices: D - dimer, T - trimer, Te - tetramer, P - pentamer, H - hexamer dPMSF - Phenylmethylsulfonyl fluoride * Based upon sequence homology only
Classification Schemes. Numerous, often-overlapping classification schemes have been developed to categorize aminopeptidases (2). Aminopeptidases have been categorized according to: the identity of amino acid residues efficiently cleaved from the N-terminus (i.e., an enzyme that cleaves a leucine residue most efficiently would be designated a “leucine” aminopeptidase), their cellular location (i.e., whether the APs are cytosolic, membrane-bound, soluble, etc.), inhibitor susceptibility, metal ion content (38), pH of maximal activity (i.e., acidic, neutral, basic), size, thermostability, amino acid sequence homology, and the specific reaction mechanism and the number of reactions the enzymes catalyze (2). Gonzales and Robert-Baudouy developed a classification scheme for bacterial aminopeptidases based on physico-chemical properties (catalytic function, molecular weight, pH optimum, etc.) and peptide sequence that is probably the most comprehensive to date (1). Families 1-6 are aminopeptidases with broad substrate specificities, while enzymes in families 7- 14 exhibit much more narrow substrate specificities. Family 1 contains PepN, which is discussed in detail below. Family 2 contains all of the PepA-type aminopeptidases such as leucine aminopeptidase. Enzymes in family 2 are Zn(II)-requiring, and several of the members are activated by Ca(II). Family 2 enzymes exhibit basic pH optima, have monomeric molecular masses ranging from 20 to 60 kDa, and are typically extracellular. Family 3 contains the other metallo-aminopeptidases with broad substrate specificities. Family 4 contains the PepC-type cysteine aminopeptidases. These are broad-specificity aminopeptidases that utilize a cysteine sulfhydryl group for catalysis. These enzymes exhibit neutral pH optima and have monomeric molecular masses of 50 kDa. Representative members of this family include PepCs from Lactococcus and Lactobacillus, as well as PepC from Streptococcus thermophilus that shares sequence homology (and catalytic activity) with bleomycin hydrolase (1, 34, 39, 40)and references within). Family 5 includes additional examples of cysteine aminopeptidases that do not fit neatly in with the Family 4 cysteine aminopeptidases. Family 6 is comprised of the serine aminopeptidases. These enzymes utilize the oxygen atom of a hydroxyl group (or alkoxide) as a nucleophile during catalysis. Members of this family are by far the least-frequently encountered aminopeptidases. Members of families 7-14 exhibit narrow substrate specificities. Family 7 contains the methionine aminopeptidases (MAP). These have been more thoroughly studied because of their role in N-terminal protein modification and their implicated roles in ubiquitination and disease
6 (2, 25, 41). The methionine aminopeptidases are monomeric metalloenzymes that often contain two metal ions per monomer, purportedly Co(II), but many reports exist claiming the enzymes may contain other metals and remain active (1, 41). Family 8 is comprised of the aspartate aminopetidases, the most representative example being aminopeptidase A from Lactococcus lactis ssp. cremoris. Interestingly, this aminopeptidase is the only example of an aminopeptidase exhibiting allosteric kinetics (1, 42). PepA adopts a hexameric structure and displays cooperative substrate binding among its six identical subunits. Family 9 includes the pyrrolidone carboxyl peptidase (Pcp) aminopeptidases. Family 9 is yet another example of a family of cysteine aminopeptidases; however, this family exhibits narrow substrate specificity as compared to the PepC family. The Pcp aminopeptidases do not share sequence homology with any other cysteine or serine aminopeptidases. Family 10 is classified as the arginine aminopeptidases and is made up of two members of the cysteine class of aminopeptidases; one member from Streptococcus (43) and the other from E. coli (44). These enzymes are specific for arginine residues in the P1 position and selectively cleave the Arg-X bond. These enzymes differ from previously described cysteine aminopeptidases in that they are reportedly membrane-bound or associated (1). Family 11 consists of the P-type aminopeptidases. PepPs are metalloenzymes that hydrolyze amino acid- proline (X-Pro) bonds when proline is in the penultimate position. The nature of PepP aminopeptidases was uncertain for some time, but they are now believed to be necessary for hormone and signal peptide cleavage (45). Proline protects peptides against non-specific peptide cleavage, and degradation may only be affected by aminopeptidases with very specific hydrolytic activity. PepP aminopeptidases have been characterized in bacteria (E. coli and Streptomyces lividans), plants, and mammals (1, 46). Family 12 contains the proline iminopeptidases or (PIP) aminopeptidases. These are a class of serine aminopeptidases, primarily found in bacteria, which cleave peptides containing an N-terminal proline. Additional varieties of PIP exist that cleave the N-terminal di- and tri-peptides from polypeptides containing an N- terminal proline (47) Family 13 contains di-peptidyl aminopeptidases. This is a broad family containing individual aminopeptidases that cleave X-Pro- or X-Ala-dipeptides from the N- terminus of polypeptides. Many are serine aminopeptidases and have been identified within the Lactobacillus genus and perhaps may have functions related to casein cleavage and cheese ripening (48). Family 14 is a general class of dipeptidases and tripeptidases. Members of this family include PepD, PepE, PepV, PepT, and dipeptidase M. The members of the di- and
7 tripeptidases are classified as metalloenzymes,(1) and many are expressed under anaerobic conditions of phosphate limitation (49-51). Examples of aminopeptidases from this family exist in either monomeric or multimeric forms and have been identified in many bacterial species (52, 53). All are apparently involved in cleavage of di-and tri-peptides that have been released from more complex polypeptides and in this way contribute to the total breakdown of peptides and the recycling of amino acids.
Aminopeptidase Inhibitors. Aminopeptidases in both bacteria as well as in mammalian systems have been implicated in a wide range of pathological conditions (19) as mentioned above. Therefore, there has been considerable effort to identify specific inhibitors of these enzymes (10, 12, 54-63). Bestatin is an analog of the dipeptide substrate PheLeu and is a slow, tight-binding inhibitor of aminopeptidases (Ki value in the nanomolar range) (2). Bestatin
contains a hydroxyl and a carbonyl group in what would be the P1 position of the substrate, and this inhibitor structurally mimics the tetrahedral gem-diolate transition state of the hydrolysis reaction catalyzed by aminopeptidase. Bestatin also has a putative metal chelating ability in addition to its structural basis for inhibition (64). At least in the case of MAP, it appears that bestatin may inhibit by chelating the metal ions in the enzyme active site (65). Several other peptide-derived inhibitors of aminopeptidases are known such as amastatin (66), arphamenine A/B (67), leuhistin (68), and probestin (64). These inhibitors selectively inhibit aminopeptidases (e.g., amastatin inhibits aminopeptidase A and M and leucine aminopeptidase but not aminopeptidase B, and selective inhibition has been used as a criterion to categorize different aminopeptidases (2, 5) . A number of phosphonate, phosphinate (62, 69), and sulfonate (35) inhibitors have also been synthesized and tested as inhibitors of aminopeptidases with mixed results (Figure 2). These compounds are often called transition-state analogs because the tetrahedral phosphorus structurally mimics the putative tetrahedral intermediates that are formed during the aminopeptidase-catalyzed reaction. The sulfonate inhibitor also mimics a tetrahedral transition state, but the more electronegative sulfur atom acts to draw additional electron density away from the sulfonyl oxygens. It was believed this modification would reduce the amount to hydrogen bonding between sulfonyl oxygen atoms and solvent water molecules as compared to the carbonyl oxygens in the phosphonate and phosphinate inhibitors. Such a reduction in
8 potential hydrogen bonding, it was theorized, would reduce the desolvation energy required for binding (35). Crystallographic studies have shown that the phosphonic acid analog of L-leucine, for instance, interacts with several active site residues in a number of leucine aminopeptidases (LAP) (12, 56, 57, 70), and these data have been used to guide subsequent inhibitor design or redesign efforts. Boronic acid analogs of peptides (71) as well as phenylmethylsulfonyl fluoride (PMSF) have been used successfully to inhibit serine proteases (72). Curcumin, a phenolic natural product from the spice turmeric, has been shown to strongly inhibit tumor invasion into surrounding tissues in in vitro and in vivo studies (73). The basis for the inhibitory effect is curcumin’s ability to inhibit aminopeptidase N, which reportedly has an important role in tumor angiogenesis (74).
9 OH
B
OH
H3C Propylphosphonate n-butylboronic acid
Propylphosphinate Sulfonate
CH3
H3C CH3 COOH
H3C O O
NH
H2N NH NH COOH
OH O H3C CH3 Amastatin [(2S, 2R)]-3-Amino-2-hydroxy-5-methylhexanoyl]-Val-Val-Asp-OH
CH3
NH2 O CH3
NH COOH
NH2
Bestatin [(2S, 2R)-3-Amino-2-hydroxy-4-Phenylbutanoyl]-L-Leucine
CH3
OCH3 OCH3 NH2 O CH3
O HO OH NH O OH N COOH
N
OO
Probestin Curcumin
Figure 2. Selected inhibitors of aminopeptidases.
10 Quaternary structures of aminopeptidases. The structural relationships among the various aminopeptidases are quite complex. Aminopeptidases can exist in either monomeric or multimeric forms (1, 2, 5). The most common multimeric aminopeptidases reported thus far exist in di, tetra, or hexameric quaternary structures. In almost all cases, multi-subunit aminopeptidases are homo-multimers. It is not clear why many of the aminopeptidases aggregate; however, it likely is due to the need to minimize hydrophobic interactions with an aqueous environment. Despite the existence of multimeric forms, almost all of the aminopeptidases have been reported to follow Michaelis-Menten kinetics rather than sigmoidal kinetics (75).
Figure 3. Crystal structure of bleomycin hydrolase. Figure rendered using Raswin version 2.7.2.1 and PDB coordinates 1CB5.
Bleomycin hydrolase has a molecular mass of approximately 330 kDa and exists as a homo-hexamer of six identical subunits (Figure 3). Bleomycin hydrolase is classified as cysteinyl aminopeptidase because it possesses an active thiol group that it uses to catalyze peptide hydrolysis (39). It is capable of catalyzing the hydrolysis of a broad range of peptidase substrates but has a preference for basic and hydrophobic residues in the P1 position (34). This aminopeptidase is of great interest to the medical community because of its ability to degrade
11 bleomycin, a glycopeptide antibiotic used as an anti-tumor drug (76). Tumor tissues possessing the ability to express bleomycin hydrolase have proven to be resistant to bleomycin as a chemotherapeutic agent (34). Surprisingly homologs have been found in diverse organisms ranging from lactic acid-utilizing bacteria to yeast and mammals.
Metallo-aminopeptidases. Metallo-aminopeptidases are enzymes that require either one or several metal ions for full activity. Approximately half of all aminopeptidases are reported to be metalloenzymes that require a metal ion cofactor for catalytic activity (2). Aminopeptidase A (PepA) from Escherichia coli has been shown to bind 2 Zn(II) ions and require these metal ions for catalysis (29). Leucine aminopeptidase from Aeromonas protelytica has been reported to be a two-cobalt aminopeptidase (26), and Holz et al. have also described a dinuclear metal ion containing form of MetAp from E. coli (77). Methionyl aminopeptidase in bacteria has attracted many investigators because of its critical role in protein processing and post-translational modification in eukaryotes (78, 79). Some aminopeptidases have been reported to be quite flexible with regard to the metal ions that they incorporate. Several dinuclear metal ion containing aminopeptidases exhibit significant catalytic activities when bound to only one equivalent of metal (5, 25). In these enzymes, the first metal ion is thought to be catalytic, and the second is modulatory (or structural) (25). In most cases, the catalytic metal ion is more tightly bound, and the catalytic site only binds a few different metal ions (25). In contrast, a relatively larger number of metal ions can often bind to the modulatory site. This characteristic has allowed for the preparation of mixed-metal analogs (Co(II)Zn(II), for example) of several metallo-aminopeptidases, and spectroscopic analyses of these analogs have allowed for detailed information about substrate binding and reaction mechanism to be gleaned (25). Not surprisingly, the kinetic characteristics of the enzymes differ vastly depending upon the type of metal ions bound (80-82). Next to iron, zinc is the second most abundant metal in biological systems (83) and is vital to all organisms (84, 85). One specific property of zinc that makes it advantageous to living systems is that it is redox-inactive and remains in a +2 oxidation state (86). Zn(II) typically binds to a very conserved HEXXH motif (87) in metallo-aminopeptidases; however, other metal- binding sequences also exist (88). Tetanus toxin from Clostridium tetani is one example of a
12 protein that contains this specific metal-binding motif and as such has been reported to be a metalloproteinase (19); however to date, the enzyme has not been shown to bind Zn(II). Spectroscopic characterization of metallo-aminopeptidases has required extensive use of metal-substituted analogs of the enzymes. Since Zn(II) has a 3d10 valence electron configuration (S = 0), most of the common spectroscopic techniques cannot be used to characterize proteins containing this metal ion (89). However, cobalt (Co(II)) has a 3d7 valence electron configuration (S = 3/2) and a similar ionic radius as Zn(II). Co(II)-substituted aminopeptidases have been extensively studied by use of EPR, 1H NMR, and UV-Vis spectroscopies, and the information from these studies have revealed important details about the structure and function of these enzymes (89).
Mechanistic studies on aminopeptidases. Many of the mechanistic studies reported for the aminopeptidases have been conducted using either p-nitroanilide or β-naphthylamide amino acid analogs as substrates (90-92). In fact, aminopeptidase N was originally named aminopeptidase “N” because of its ability to cleave naphthylamide amino acid analogs to yield a colored product, which could be monitored using UV-Vis spectrophotometry. The p-nitroanilide substrate analogs are hydrolytically cleaved to yield p-nitroaniline that absorbs at 404 nm. Cleavage of peptides yields individual amino acids or peptides that must be reacted with other chemicals in order to detect them. Prior to the development of continuous assay methods, a ninhydrin assay was used to measure protein cleavage or amino acid production (93). This method utilizes the ninhydrin reagent that forms a colored adduct with free amino acids generated during hydrolytic, peptide cleavage reactions (94). The ninhydrin assay has largely been abandoned to continuous kinetic assays, which typically have lower limits of detection, reduced errors, and the ability to continuously monitor the reaction. The major disadvantage to using artificial substrates, such as p-nitroaniline-containing compounds, for kinetic measurements is that these analogs are typically not hydrolyzed at rates comparable to the native substrates and may involve a different mechanism than is used to hydrolyze the natural substrate(s) (91).
VanX. VanX is an aminopeptidase of great medical importance because it is utilized by vancomycin-resistant bacteria to thwart the effect of the antibiotic vancomycin. Vancomycin has
13 been called the “antibiotic of last resort”(95), and therefore, is typically only employed when all other antibiotics fail to halt the progression of an infection (96). Increased use and misuse of antibiotics have caused an environmental pressure that has selected for antibiotic-resistant forms of bacteria (97). The transfer of genes that lead to the resistance phenotype can be accomplished directly via conjugation or via genetic transposons (98, 99). The VanA phenotype is mediated by the vanA gene cluster that is carried by a 10 kB transposon (Tn1546 or related element), which can be converted into self-transferable plasmids and transferred by conjugation (100). This gene cluster contains 9 genes: orf1, orf2, vanR, vanS, vanH, vanA, vanX, vanY, and vanZ. Orf1 and orf2 are required for transposition; however, only five of the remaining genes are required for high-level vancomycin resistance: vanR, vanS, vanH, vanA, and vanX. The mechanism of high-level vancomycin resistance has been reviewed previously by Walsh and coworkers (95, 101). Briefly, VanR and VanS make up a two- component system that regulates the expression of the other three genes and appears to be stimulated by the presence of vancomycin (102-104). VanH encodes for a D-specific α-ketoacid reductase, which catalyzes the reduction of pyruvate into D-lactate using NADPH (105, 106). VanA is a ligase that uses the D-lactate produced by VanH and generates a D-ala-D-lactate depsipeptide (106, 107). This depsipeptide is added to the sugar-linked tripeptide by MurF, producing a modified bacterial cell wall precursor that can be incorporated into the growing peptidoglycan layer (95, 101, 105). Vancomycin binds 103 less tightly to the D-ala-D-lactate depsipeptide as compared to the normal D-ala-D-ala dipeptide due to the loss of one hydrogen bond and lone pair/lone pair electron repulsions (108). Until late 1995, the role of VanX remained unknown; however, Courvalin and coworkers showed that VanX is a D-ala-D-ala peptidase (98). In the presence of vancomycin or ramoplanin, vancomycin-resistant Enterococcus strain BM4147 produced modified, vancomycin-resistant depsipentapeptides and normal, vancomycin-sensitive pentapeptides at a ratio of 49:1. If the VanX gene was inactivated, this ratio changed to 1:1. This result demonstrated the strict requirement for VanX in vancomycin-resistant bacteria. Without VanX, one-half of the bacterial cell wall precursors are susceptible to vancomycin. Walsh and coworkers subsequently showed that VanX is a Zn(II) dependent enzyme and reported several phosphinate and thiol containing inhibitors (36, 109). The crystal structure of VanX was
14 reported by Bussiere and coworkers in 1998 (99); however, no mechanistic studies have been reported on VanX.
Aminopeptidase N. Another interesting aminopeptidase is aminopeptidase N (PepN). Aminopeptidase N from Lactococcus and Lactobacillus strains has been more thoroughly characterized (40, 110) because of the enzyme’s role in the making of cheese and other dairy products (1, 111). Lactococcus and Lactobacillus bacteria produce a number of proteolytic enzymes that are essential for cheese ripening (112). These enzymes process various free amino acids that contribute to the aroma and also degrade a number of peptides that confer a bitter taste to the cheese (113). Free amino acids and small peptides required for proper bacterial growth are limited in milk cultures. Lactococci must therefore possess the ability to cleave larger casein and other milk proteins in order to free amino acids for anabolic processes (111). Aminopeptidases C and N and X-prolyl-dipeptidylaminopeptidase from Lactococcus lactis have all been cloned and characterized in an effort to understand their structures/functions (32, 40, 87, 110, 112, 114, 115). The mammalian form of aminopeptidase N (CD13) is involved in tumor invasion and angiogenesis (73). Mammalian PepN is membrane-associated and has been identified as a receptor for several viruses and in the activation mechanism of collagenolysis that allows for tumor cell invasion (116-118). Since PepN apparently is involved in a number of biomedically- important processes, a large number of inhibitors of mammalian PepN have been reported. For example, several natural product dipeptides, bestatin, phebestin, probestin, and amastatin, have been shown to be potential inhibitors of mammalian PepN (30, 64, 119). PepN has been identified as the sole alanyl aminopeptidase in E. coli and is the major aminopeptidase involved in ATP-independent processing during cytosolic protein degradation (114). Given these roles, PepN and other bacterial proteases have been recently identified as potential antibiotic targets (19). The first publications on E. coli PepN appeared in the mid-1970’s, and much of this work described kinetic studies on partially-purified enzyme samples (31, 120). E. coli aminopeptidase N is capable of cleaving a broad range of amino acids from peptide substrates with a preference for small and basic residues at the N-terminus (33). The enzyme is localized in the cytosol and exists as a soluble monomer with a pI of 5.7 (33, 121). Based on studies with chemical
15 modification agents, Lazdunski and coworkers hypothesized that E. coli PepN is a cysteinyl aminopeptidase that utlizes an active site cysteine in the nucleophilic attack on substrate (122). The DNA (and amino acid) sequence of E. coli PepN was reported in the mid-1980’s, and the enzyme was predicted to have a molecular weight of 98-99 kDa (33, 123-125). The amino acid sequence revealed that E. coli PepN has an HEXXH motif, which is a common Zn(II)-binding motif found in many proteins. Based entirely on the appearance of this motif, E. coli PepN was re-categorized as a metallo-aminopeptidase (38), although no metal analyses have ever been reported. In 2003, Chandu and Nandi successfully cloned and over-expressed E. coli PepN using an arabinose-induced, over-expression system and purified the enzyme using three chromatographic steps (126). Kinetic studies lead these authors to assert that E. coli PepN is an aminoendopeptidase and the major aminopeptidase in E. coli (114).
Sections of this dissertation Despite 30 years of study, very little is known about E. coli PepN. There is no structural information available on this enzyme other than the primary structure. Although there have been many substrate specificity studies published, little is known about the reaction mechanism of E. coli PepN (114). This dissertation describes our efforts to answer many of the questions that exist about this enzyme. In chapter 2 (L-Alanine-p-nitroanilide is not a Substrate for VanX), evidence is presented that shows that the hydrolysis of artificial substrate L-alanine-p-nitroanilide is due to endogenous aminopeptidase N (PepN) from E. coli and not due to VanX as had been previously reported (91). The material in this chapter has already been published in Anal. Biochem. (127). In chapter 3 (Over-expression, purification, and characterization of E. coli PepN), the over-expression, purification, and partial characterization of PepN is described. This chapter has been accepted for publication in Prot. Express. Purif. (publication in 2006). In chapter 4 (Spectroscopic studies on recombinant E. coli PepN), results from the first known spectroscopic studies on recombinant PepN are described. In chapter 5 (Mechanistic studies on recombinant E. coli PepN), results from mechanistic studies are presented, and the first known reaction mechanism for E. coli PepN is hypothesized. Chapter 6 describes a summary of this work and suggests future directions of this project.
16 References 1. Gonzales, T., and Robert-Baudouy, J. Bacterial aminopeptidases: properties and functions (1996) FEMS Microbiol Lett 18, 319-344. 2. Taylor, A. (1996) Aminopeptidases, R.G. Landes Company, Austin, TX. 3. Taylor, S. W., Kammerer, B., and Bayer, E. New Perspectives in the chemistry and biochemistry of the tunichromes and related compounds (1997) Chem Rev 97, 333-346. 4. Bradshaw, R. A. Aminopeptidases (2004) Encyclopedia Biol. Chemistry 1, 96-98. 5. Taylor, A. Aminopeptidases: structure and function (1993) FASEB J 7, 290-8. 6. Rivett, A. J. High molecular mass intracellular proteases (1989) Biochem J 263, 625-33. 7. Chang, Y.-H., Teichert, U., and Smith, J. A. Purification and characterization of a methionine aminopeptidase from Saccharomyces cerevisiae (1990) J Biol Chem 265, 19892-19897. 8. Hershko, A., and Ciechanover, A. The ubiquitin pathway for the degradation of intracellular proteins (1986) Prog Nucleic Acid Res Mol Biol 33, 19-56, 301. 9. Ashraf, H. Cathepsin enzyme provides clue to SARS infection (2005) Drug Discov Today 10, 1409. 10. Rosenthal, P. J., Olson, J. E., Lee, G. K., Palmer, J. T., Klaus, J. L., and Rasnick, D. Antimalarial effects of vinyl sulfone cysteine proteinase inhibitors (1996) Antimicrob Agents Chemother 40, 1600-1603. 11. Rosenthal, P. J., Sijwali, P. S., Singh, A., and Shenai, B. R. Cysteine proteases of malaria parasites: targets for chemotherapy (2002) Curr Pharm Des 8, 1659-72. 12. Grembecka, J., Mucha, A., Cierpicki, T., and Kafarski, P. The most potent organophosphorus inhibitors of leucine aminopeptidase. Structure-based design, chemistry, and activity (2003) J Med Chem 46, 2641-2655. 13. Li, Q., and Xu, W. Novel anticancer targets and drug discovery in post genomic age (2005) Curr Med Chem Anti-Canc Agents 5, 53-63. 14. Bazan, J., Weaver, L., Roderick, S., Huber, R., and Matthews, B. Sequence and structure comparison suggest that methionine aminopeptidase, prolidase, aminopeptidase P, and creatinase share a common fold (1994) PNAS 91, 2473-2477. 15. Christie, G., Markwell, R. E., Gray, C. W., Smith, L., Godfrey, F., Mansfield, F., Wadsworth, H., King, R., McLaughlin, M., Cooper, D. G., Ward, R. V., Howlett, D. R., Hartmann, T., Lichtenthaler, S. F., Beyreuther, K., Underwood, J., Gribble, S. K., Cappai,
17 R., Masters, C. L., Tamaoka, A., Gardner, R. L., Rivett, A. J., Karran, E. H., and Allsop, D. Alzheimer's disease: correlation of the suppression of beta-amyloid peptide secretion from cultured cells with inhibition of the chymotrypsin-like activity of the proteasome (1999) J Neurochem 73, 195-204. 16. Korant, B. D., and Rizzo, C. J. The HIV protease and therapies for AIDS (1997) Adv Exp Med Biol 421, 279-84. 17. Nelson, D., Potempa, J., Kordula, T., and Travis, J. Purification and characterization of a novel cysteine proteinase (periodontain) from Porphyromonas gingivalis. Evidence for a role in the inactivation of human alpha1-proteinase inhibitor (1999) J Biol Chem 274, 12245-51. 18. Travis, J., Potempa, J., and Maeda, H. Are bacterial proteinases pathogenic factors? (1995) Trends Microbiol 3, 405-7. 19. Travis, J., and Potempa, J. Bacterial proteinases as targets for the development of second- generation antibiotics (2000) Biochim Biophys Acta 1477, 35-50. 20. White, R. J., Margolis, P. S., Trias, J., and Yuan, Z. Targeting metalloenzymes: a strategy that works (2003) Curr Opin Pharma 3, 502-507. 21. Gharbi, S., Bally, M., Lazdunski, A., and Murgier, M. Mutation that affects pepN translation initiation in Escherichia coli (1986) Biochimie 68, 1159-63. 22. Murgier, M., Pelissier, C., Lazdunski, A., and Lazdunski, C. Existence, localization and regulation of the biosynthesis of aminoendopeptidase in gram-negative bacteria (1976) Eur J Biochem 65, 517-20. 23. Vesanto, E., Varmanen, P., Steele, J. L., and Palva, A. Characterization and expression of the Lactobacillus helveticus pepC gene encoding a general aminopeptidase (1994) Eur J Biochem 224, 991-7. 24. Chang, S. Y., McGary, E. C., and Chang, S. Methionine aminopeptidase gene of Escherichia coli is essential for cell growth (1989) J Bacteriol 171, 4071-2. 25. Lowther, W. T., and Matthews, B. W. Metalloaminopeptidases: Common functional themes in disparate structural surroundings (2002) Chem Rev 102, 4581-4607. 26. Holz, R. C. The aminopeptidase from Aeromonas proteolytica: structure and mechanism of co-catalytic metal centers involved in peptide hydrolysis (2002) Coord Chem Rev 232, 5- 26.
18 27. Walker, K. W., and Bradshaw, R. A. Yeast Methionine Aminopeptidase I (1999) J Biol Chem 274, 13403-13409. 28. Suzuki, H., Kamatani, S., Kim, E. S., and Kumagai, H. Aminopeptidases A, B, and N and dipeptidase D are the four cysteinylglycinases of Escherichia coli K-12 (2001) J Bacteriol 183, 1489-90. 29. Stirling, C. J., Colloms, S. D., Collins, J. F., Szatmari, G., and Sherratt, D. J. xerB, an Escherichia coli gene required for plasmid ColE1 site-specific recombination, is identical to pepA, encoding aminopeptidase A, a protein with substantial similarity to bovine lens leucine aminopeptidase (1989) EMBO J 8, 1623-7. 30. Wilkes, S. H., and Prescott, J. M. The slow, tight binding of bestatin and amastatin to aminopeptidases (1985) J Biol Chem 260, 13154-13162. 31. Murgier, M., Pellissier, C., Lazdunski, A., Bernadac, A., and Lazdunski, C. Aminopeptidase N from Escherichia coli. Unusual interactions with the cell surface (1977) Eur J Biochem 74, 425-33. 32. Foglino, M., Gharbi, S., and Lazdunski, A. Nucleotide-Sequence of the pepN gene encoding aminopeptidase-N of Escherichia-coli (1986) Gene 49, 303-309. 33. McCaman, M. T., and Gabe, J. D. The nucleotide-sequence of the pepN gene and its over- expression in Escherichia-coli (1986) Gene 48, 145-153. 34. Mistou, M. Y., and Gripon, J. C. Catalytic properties of the cysteine aminopeptidase PepC, a bacterial bleomycin hydrolase (1998) Biochim Biophys Acta 1383, 63-70. 35. Yang, K. W., Golich, F. C., Sigdel, T. K., and Crowder, M. W. Phosphinate, sulfonate, and sulfonamidate dipeptides as potential inhibitors of Escherichia coli aminopeptidase N (2005) Bioorg Med Chem Lett 15, 5150-3. 36. Wu, Z., and Walsh, C. T. Phosphinate analogs of D-,D-dipeptides: slow-binding inhibition and proteolysis protection of VanX, a D-, D-dipeptidase required for vancomycin resistance in Enterococcus faecium (1995) Proc Natl Acad Sci USA 92, 11603-11607. 37. McCafferty, D. G., Lessard, I. A. D., and Walsh, C. T. Mutational analysis of potential zinc- binding residues in the active site of the Enterococcal D-Ala-D-Ala dipeptidase VanX (1997) Biochemistry 36, 10498-10505. 38. Taylor, A. Aminopeptidases: towards a mechanism of action (1993) Trends Biochem Sci 18, 167-172.
19 39. Chapot-Chartier, M. P., Rul, F., Nardi, M., and Gripon, J. C. Gene cloning and characterization of PepC, a cysteine aminopeptidase from Streptococcus thermophilus, with sequence similarity to the eukaryotic bleomycin hydrolase (1994) Eur J Biochem 224, 497-506. 40. Chapot-Chartier, M. P., Nardi, M., Chopin, M. C., Chopin, A., and Gripon, J. C. Cloning and sequencing of pepC, a cysteine aminopeptidase gene from Lactococcus lactis subsp. cremoris AM2 (1993) Appl Environ Microbiol 59, 330-3. 41. Lowther, W. T., and Matthews, B. W. Structure and function of the methionine aminopeptidases (2000) Biochim Biophys Acta 1477, 157-67. 42. Niven, G. W. Purification and characterization of aminopeptidase A from Lactococcus lactis subsp. lactis NCDO712 (1991) J Gen Microbiol 137, 1207-1212. 43. Floderus, E., Linder, L. E., and Sund, M.-L. Characterization of membrane-associated arginine aminopeptidase in Streptococcus sanguis 903 (1990) Curr Microbiol 21, 145- 149. 44. Ishino, Y., Shinagawa, H., Makino, K., Amemura, M., and Nakata, A. Nucleotide sequence of the iap gene, responsible for alkaline phosphatase isozyme conversion in Escherichia coli, and identification of the gene product (1987) J Bacteriol 169, 5429-33. 45. Yoshimoto, T., Orawski, A. T., and Simmons, W. H. Substrate specificity of aminopeptidase P from Escherichia coli: comparison with membrane-bound forms from rat and bovine lung (1994) Arch Biochem Biophys 311, 28-34. 46. Butler, M. J., Bergeron, A., Soostmeyer, G., Zimny, T., and Malek, L. T. Cloning and characterisation of an aminopeptidase P-encoding gene from Streptomyces lividans (1993) Gene 123, 115-9. 47. Klein, J. R., Schmidt, U., and Plapp, R. Cloning, heterologous expression, and sequencing of a novel proline iminopeptidase gene, pepI, from Lactobacillus delbrueckii subsp. lactis DSM 7290 (1994) Microbiol 140 ( Pt 5), 1133-9. 48. Magboul, A. A., and McSweeney, P. L. PepN-like aminopeptidase from Lactobacillus curvatus DPC2024: purification and characterization (1999) Lait 79, 515-526. 49. Henrich, B., Monnerjahn, U., and Plapp, R. Peptidase D gene (pepD) of Escherichia coli K- 12: nucleotide sequence, transcript mapping, and comparison with other peptidase genes (1990) J Bacteriol 172, 4641-51.
20 50. Miller, C. G., Miller, J. L., and Bagga, D. A. Cloning and nucleotide sequence of the anaerobically regulated pepT gene of Salmonella typhimurium (1991) J Bacteriol 173, 3554-8. 51. Strauch, K. L., Lenk, J. B., Gamble, B. L., and Miller, C. G. Oxygen regulation in Salmonella typhimurium (1985) J Bacteriol 161, 673-80. 52. Plancot, M. T., and Han, K. K. Purification and characterization of an intracellular dipeptidase from Mycobacterium phlei (1972) Eur J Biochem 28, 327-33. 53. Mierau, I., Haandrikman, A. J., Velterop, O., Tan, P. S., Leenhouts, K. L., Konings, W. N., Venema, G., and Kok, J. Tripeptidase gene (pepT) of Lactococcus lactis: molecular cloning and nucleotide sequencing of pepT and construction of a chromosomal deletion mutant (1994) J Bacteriol 176, 2854-61. 54. Bartlett, P. A., and Kezer, W. B. Phosphinic acid dipeptide analogues: potent, slow-binding inhibitors of aspartic peptidases (1984) J Am Chem Soc 106, 4282-4283. 55. Giannousis, P. P., and Bartlett, P. A. Phosphorus amino acid analogues as inhibitors of leucine aminopeptidase (1987) J Med Chem 30, 1603-1609. 56. Grembecka, J., Sokalski, W. A., and Kafarski, P. Computer-aided design and activity prediction of leucine aminopeptidase inhibitors (2000) J Comput Aided Mol Des 14, 531- 44. 57. Grembecka, J., Mucha, A., Cierpicki, T., and Kafarski, P. Structure-based design and synthesis of dipeptide analogues as new inhibitors of leucine aminopeptidase (2002) Phosphorus, Sulfur and Silicon 177, 1739-1743. 58. Leung, D., Abbenante, G., and Fairlie, D. P. Protease inhibitors: Current status and future prospects (2000) J Med Chem 43, 305-341. 59. Otto, H. H., and Schirmeister, T. Cysteine proteases and their inhibitors (1997) Chem Rev 97, 133-171. 60. Ripka, A. S., Satyshur, K. A., Bohacek, R. S., and Rich, D. H. Aspartic protease inhibitors designed from computer-generated templates bind as predicted (2001) Org Lett 3, 2309- 2312. 61. Roques, B. P. Zinc Metallopeptidases: Active site structure and design of selective and mixed inhibitors: new approaches in the search for analgesics and anti-hypertensives (1993) Biochem Soc Trans 21, 678-685.
21 62. Yang, K. W., Brandt, J. J., Chatwood, L. L., and Crowder, M. W. Phosphonamidate and phosphothioate dipeptides as potential inhibitors of VanX (2000) Bioorg Med Chem Lett 10, 1087-1089. 63. Webb, J. L. (1963) Enzyme and Metabolic Inhibitors, Vol. 1, Academic Press, New York. 64. Righi, G., D'Achille, C., Pescatore, G., and Bonini, C. New stereoselective synthesis of the peptidic aminopeptidase inhibitors bestatin, phebestin, and probestin (2003) Tetrahed Lett 44, 6999-7002. 65. Vaughan, M. D., Sampson, P. B., and Honek, J. F. Methionine in and out of proteins: targets for drug design (2002) Curr Med Chem 9, 385-409. 66. Rich, D. H., Moon, B. J., and Harbeson, S. Inhibition of aminopeptidases by amastatin and bestatin derivatives - effect of inhibitor structure on slow-binding processes (1984) J Med Chem 27, 417-422. 67. Harbeson, S. L., and Rich, D. H. Inhibition of arginine aminopeptidase by bestatin and arphamenine analogues. Evidence for a new mode of binding to aminopeptidases (1988) Biochemistry 27, 7301-10. 68. Yoshida, S., Aoyagi, T., and Takeuchi, T. Biosynthetic study of leuhistin, a new inhibitor of aminopeptidase M (1991) J Antibiot (Tokyo) 44, 683-4. 69. Lejczak, B., De Choszczak, M. P., and Kafarski, P. Inhibition of aminopeptidases by phosphonic acid and phosphinic acid analogues of aspartic and glutamic acids (1993) J Enzyme Inhib 7, 97-103. 70. Grembecka, J., and Kafarski, P. Leucine aminopeptidase as a target for inhibitor design (2001) Mini Rev Med Chem 1, 133-44. 71. Coutts, S. J., Kelly, T. A., Snow, R. J., Kennedy, C. A., Barton, R. W., Adams, J., Krolikowski, D. A., Freeman, D. M., Campbell, S. J., Ksiazek, J. F., and Bachovchin, W. W. Structure-activity relationships of boronic acid inhibitors of dipeptidyl peptidase IV. 1. Variation of the P2 position of Xaa-boroPro dipeptides (1996) J Med Chem 39, 2087- 94. 72. DePaola, C. C., Bennett, B., Holz, R. C., Ringe, D., and Petsko, G. A. 1-butaneboronic acid binding to Aeromonas proteolytica aminopeptidase: a case of arrested development (1999) Biochemistry 38, 9048-9053.
22 73. Shim, J. S., Kim, J. H., Cho, H. Y., Yum, Y. N., Kim, S. H., Park, H. J., Shim, B. S., Choi, S. H., and Kwon, H. J. Irreversible inhibition of CD13/aminopeptidase N by the antiangiogenic agent curcumin (2003) Chem Biol 10, 695-704. 74. Xu, W., and Li, Q. Progress in the development of aminopeptidase N (APN/CD13) inhibitors (2005) Curr Med Chem Anti-Canc Agents 5, 281-301. 75. Gonzalez, J. C., Peariso, K., Penner-Hahn, J. E., and Matthews, R. G. Cobalamin- Independent Methionine Synthase from Escherichia coli: A Zinc Metalloenzyme (1996) Biochemistry 35, 12228-12234. 76. Hecht, S. M. The Chemistry of Activated Bleomycin (1986) Acc Chem Res 19, 383-391. 77. D'Souza V, M., Bennett, B., Copik, A. J., and Holz, R. C. Divalent metal binding properties of the methionyl aminopeptidase from Escherichia coli (2000) Biochemistry 39, 3817-26. 78. Lowther, W. T., Orville, A. M., Madden, D. T., Lim, S., Rich, D. H., and Matthews, B. W. Escherichia coli methionine aminopeptidase: implications of crystallographic analyses of the native, mutant, and inhibited enzymes for the mechanism of catalysis (1999) Biochemistry 38, 7678-88. 79. Cosper, N. J., D'Souza, V. M., Scott, R. A., and Holz, R. C. Structural evidence that the methionyl aminopeptidase from Escherichia coli is a mononuclear metalloprotease (2001) Biochemistry 40, 13302-13309. 80. Van Wart, H. E., and Lin, S. H. Metal binding stoichiometry and mechanism of metal ion modulation of the activity of porcine kidney leucine aminopeptidase (1981) Biochemistry 20, 5682-9. 81. Lin, W. Y., Lin, S. H., and Van Wart, H. E. Steady-state kinetics of hydrolysis of dansyl- peptide substrates by leucine aminopeptidase (1988) Biochemistry 27, 5062-8. 82. Allen, M. P., Yamada, A. H., and Carpenter, F. H. Kinetic parameters of metal-substituted leucine aminopeptidase from bovine lens (1983) Biochemistry 22, 3778-83. 83. Williams, R. J. P., and Fraústo da Silva, J. J. R. (2001) The biological chemistry of the elements: the inorganic chemistry of life, 2nd ed., Oxford University Press, Oxford, New York. 84. Hooper, N. M. Families of Zinc Metalloproteases (1994) FEBS Letters 354, 1-6. 85. Vallee, B. L., and Galdes, A. The Metallobiochemistry of Zinc Enzymes (1984) Adv Enzymol Relat Areas Mol Biol 56, 283-430.
23 86. Blencowe, D. K., and Morby, A. P. Zn(II) Metabolism in Prokaryotes (2003) FEMS Microbiol Lett 27, 291-311. 87. Tan, P. S., van Alen-Boerrigter, I. J., Poolman, B., Siezen, R. J., de Vos, W. M., and Konings, W. N. Characterization of the Lactococcus lactis pepN gene encoding an aminopeptidase homologous to mammalian aminopeptidase N (1992) FEBS Lett 306, 9- 16. 88. Jernigan, R., Raghunathan, G., and Bahar, I. Characterization of interactions and metal ion binding sites in proteins (1994) Curr Opin Struct Biol 4, 256-263. 89. Maret, W., and Vallee, B. L. (1993) in Methods in enzymology pp 52-71, Academic Press, New York. 90. Hawthorne, S. J., Harriott, P., Lim, J., Turner, A. J., Walker, B., and Williams, C. H. Evaluation of some fluorogenic substrates for continuous assay of aminopeptidase P (1997) Anal Biochem 253, 13-17. 91. Brandt, J. J., Chatwood, L. L., Yang, K. W., and Crowder, M. W. Continuous assay for VanX, the D-Alanyl-D-Alanine dipeptidase required for high-level vancomycin resistance (1999) Anal Biochem 272, 94-99. 92. Zhou, Y., Guo, X. C., Yi, T., Yoshimoto, T., and Pei, D. Two continuous spectrophotometric assays for methionine aminopeptidase (2000) Anal Biochem 280, 159-165. 93. Wu, Z., Wright, G. D., and Walsh, C. T. Overexpression, purification, and characterization of VanX, a D-, D-dipeptidase which is essential for vancomycin resistance in Enterococcus faecium BM4147 (1995) Biochemistry 34, 2455-2463. 94. Doi, E., Shibata, D., and Matoba, T. Modified colorimetric ninhydrin methods for peptidase assay (1981) Anal Biochem 118, 173-84. 95. Walsh, C. T. Vancomycin Resistance: Decoding the molecular logic (1993) Science 261, 308-309. 96. Neu, H. C. The crisis in antibiotic resistance (1992) Science 257, 1064-1073. 97. Hamilton-Miller, J. M. Vancomycin-resistant Staphylococcus aureus: a real and present danger? (2002) Infection 30, 118-24. 98. Reynolds, P. E., Depardieu, F., Dutka-Malen, S., Arthur, M., and Courvalin, P. Glycopeptide resistance mediated by Enterococcal Transposon Tn1546 requires production of VanX for hydrolysis of D-Alanyl-D-Alanine (1994) Mol Microbiol 13, 1065-1070.
24 99. Bussiere, D. E., Pratt, S. D., Katz, L., Severin, J. M., Holzman, T., and Park, C. H. The structure of VanX reveals a novel amino-dipeptidase involved in mediating transposon- based vancomycin resistance (1998) Mol Cell 2, 75-84. 100. Courvalin, P. Transfer of antibiotic resistance genes between Gram positive and Gram negative bacteria (1994) Antimicrob Agents Chemother 38, 1447-1451. 101. Walsh, C. T., Fisher, S. L., Park, I. S., Prahalad, M., and Wu, Z. Bacterial resistance to vancomycin: five genes and one missing hydrogen bond tell the story (1996) Chem Biol 3, 21-28. 102. Holman, T. R., Wu, Z., Wanner, B. L., and Walsh, C. T. Identification of the DNA-binding site for the phosphorylated VanR protein required for vancomycin resistance in Enterococcus faecium (1994) Biochemistry 33, 4625-4631. 103. Haldimann, A., Fisher, S. L., Daniels, L. L., Walsh, C. T., and Wanner, B. L. Transcriptional regulation of the Enterococcus faecium BM4147 vancomycin resistance gene cluster by the VanS-VanR two component regulatory system in Escherichia coli K- 12 (1997) J Bacteriol 179, 5903-5913. 104. Ulijasz, A. T., Kay, B. K., and Weisblum, B. Peptide analogues of the VanS catalytic center inhibit VanR binding to its cognate promoter (2000) Biochemistry 39, 11417- 11424. 105. Bugg, T. D. H., and Walsh, C. T. Intracellular steps of bacterial cell wall peptidoglycan biosynthesis: enzymology, antibiotics, and antibiotic resistance (1992) Nat Prod Rep, 199-215. 106. Bugg, T. D. H., Wright, G. D., Dutka-Malen, S., Arthur, M., Courvalin, P., and Walsh, C. T. Molecular basis for vancomycin resistance in Enterococcus faecium BM4147: Biosynthesis of a depsipeptide peptidoglycan precursor by vancomycin resistance proteins VanH and VanA (1991) Biochemistry 30, 10408-10415. 107. Fan, C., Moews, P. C., Walsh, C. T., and Knox, J. R. Vancomycin resistance: Structure of D-Alanine:D-Alanine ligase at 2.3 Å resolution (1994) Science 266, 439-443. 108. McComas, C. C., Crowley, B. M., and Boger, D. L. Partitioning the loss in vancomycin binding affinity for D-ala-D-lac into lost H-bond and repulsive lone pair contributions (2003) J Am Chem Soc 125, 9314-9315.
25 109. Wu, Z., and Walsh, C. T. Dithiol Compounds: Potent, time-dependent inhibitors of VanX, a zinc-dependent D,D-Dipeptidase required for vancomycin resistance in Enterococcus faecium (1996) J Am Chem Soc 118, 1785-1786. 110. van Alen-Boerrigter, I. J., Baankreis, R., and de Vos, W. M. Characterization and overexpression of the Lactococcus lactis pepN gene and localization of its product, aminopeptidase N (1991) Appl Environ Microbiol 57, 2555-61. 111. Klein, J. R., Klein, U., Schad, M., and Plapp, R. Cloning, DNA sequence analysis and partial characterization of PepN, a lysyl aminopeptidase from Lactobacillus delbrukii ssp. lactis DSM7290 (1993) Eur J Biochem 217, 105-114. 112. Magboul, A. A., and McSweeney, P. L. Purification and characterization of an X-prolyl- dipeptidyl aminopeptidase from Lactobacillus curvatus DPC2024 (2000) Lait 80, 385- 396. 113. Chavagnat, F., Casey, M. G., and Meyer, J. Purification, characterization, gene cloning, sequencing, and overexpression of aminopeptidase N from Streptococcus thermophilus A (1999) Appl Environ Microbiol 65, 3001-3007. 114. Chandu, D., and Nandi, D. PepN is the major aminopeptidase in Escherichia coli: insights on substrate specificity and role during sodium-salicylate- induced stress (2003) Microbiology-Sgm 149, 3437-3447. 115. Murgier, M., and Gharbi, S. Fusion of the lac genes to the promoter for the aminopeptidase- N gene of Escherichia-coli (1982) Mol Gen Genet 187, 316-319. 116. Yeager, C. L., Ashmun, R. A., Williams, R. K., Cardellichio, C. B., Shapiro, L. H., Look, A. T., and Holmes, K. V. Human aminopeptidase N is a receptor for human coronavirus 229E (1992) Nature 357, 420-422. 117. Saiki, I., Fujii, H., Yoneda, J., Abe, F., Nakajima, M., Tsuruo, T., and Azuma, I. Role of aminopeptidase N (CD13) in tumor-cell invasion and extracellular matrix degradation (1993) Int J Cancer 54, 137-143. 118. Delmas, B., Gelfi, J., L'Haridon, R., Vogel, L. K., Sjostrom, H., Noren, O., and Laude, H. Aminopeptidase N is a major receptor for the entero-pathogenic coronavirus TGEV (1992) Nature 357, 417-420.
26 119. Nagai, M., Kojima, F., Naganawa, H., Hamada, M., Aoyagi, T., and Takeuchi, T. Phebestin, a new inhibitor of aminopeptidase N, produced by Streptomyces sp. MJ716- m3 (1997) J. Antibiotics 50, 82-84. 120. McCaman, M. T., and Villarego, M. R. Structural and catalytic properties of peptidase N from Escherichia coli K-12 (1982) Arch Biochem Biophys 213, 384-394. 121. Yoshimoto, T., Tamesa, Y., Gushi, K., Murayama, N., and Tsuru, D. An Aminopeptidase- N from Escherichia-coli-HB101 - purification and demonstration that the enzyme possesses arylamidase and peptidase activities (1988) Agric Biol Chem 52, 217-225. 122. Lazdunski, C., Busuttil, J., and Lazdunski, A. Purification and properties of a periplasmic aminoendopeptidase from Escherichia coli (1975) Eur J Biochem 60, 363-9. 123. McCaman, M. T., and Gabe, J. D. Sequence of the promoter and 5' coding region of pepN, and the amino-terminus of peptidase-N from Escherichia-coli K-12 (1986) Mol Gen Genet 204, 148-152. 124. Bally, M., Foglino, M., Bruschi, M., Murgier, M., and Lazdunski, A. Nucleotide sequence of the promoter and amino-terminal encoding region of the Escherichia coli pepN gene (1986) Eur J Biochem 155, 565-9. 125. Bally, M., Murgier, M., and Lazdunski, A. Cloning and orientation of the gene encoding aminopeptidase-N in Escherichia-coli (1984) Mol Gen Genet 195, 507-510. 126. Chandu, D., Kumar, A., and Nandi, D. PepN, the major Suc-LLVY-AMC-hydrolyzing enzyme in Escherichia coli, displays functional similarity with downstream processing enzymes in archaea and eukarya. Implications in cytosolic protein degradation (2003) J Biol Chem 278, 5548-56. 127. Golich, F. C., Sigdel, T., Breece, R. M., Detar, L., Herron, L. R., and Crowder, M. W. L- alanine-p-nitroanilide is not a substrate for VanX (2004) Anal Biochem 331, 398-400.
27 Chapter 2
L-Alanine-p-nitroanilide is not a Substrate for VanX
The material in this chapter has been published: Golich, F.C.; Sigdel, T.; Breece, R.M.; Detar, L.; Herron, L.R.; Crowder, M.W. “L-alanine-p-nitroanilide is not a substrate for VanX” Anal. Biochem. (2004) 331 398-400.
28 Introduction Vancomycin is a glycopeptide antibiotic that is used currently to treat Gram-positive bacterial infections, particularly those caused by methicillin-resistant Staphylococcus aureus (MRSA) and Enterococcus faecalis (1). This antibiotic, called the antibiotic of last resort, is also used extensively in patients that are allergic to β-lactam containing antibiotics and who have antibiotic resistant bacterial infections (2). High-level, clinical resistance to vancomycin has emerged due to the production of five proteins that work in concert to modify the bacterial cell wall and to reduce the binding affinity of vancomycin (2-5). One of the proteins, VanX, is a Zn(II)-containing D-D-specific dipeptidase that depletes the intracellular pool of D-ala-D-ala for bacterial cell wall synthesis (6). Instead, D-ala-D-lactate, which is produced by two of the other vancomycin resistance proteins and does not bind vancomycin strongly, is incorporated into the bacterial cell wall (2). Previous studies by Reynolds showed that insertional inactivation of VanX reestablished vancomycin sensitivity to a resistant strain of Enterococcus faecalis (6). This result demonstrates that VanX is an excellent target for inhibitors, which, given in combination with vancomycin, could be used as an effective treatment for vancomycin resistance bacterial infections. One strategy to prepare inhibitors is to characterize a target protein and design inhibitors based on unique structural or mechanistic properties of the protein. The crystal structure of VanX has been reported and revealed that VanX is a mononuclear Zn(II)-containing aminopeptidase that has a substrate binding pocket of very limited size (7). Initially, a ninhydrin-based assay was used to conduct steady-state kinetic and inhibition studies on VanX (8-10). Due to the ninhydrin assay being difficult to perform and to its being non-continuous and not very useful in probing mechanism, a number of groups have reported alternative activity assays. These assays include the use of capillary electrophoresis to monitor the formation of D-ala (11), a coupled enzyme assay in which the VanX-generated D-ala is oxidized to pyruvate that is then reduced (with oxidation of NADH) to lactate (12), and an assay in which a VanX-generated product is quantitated by using Ellman’s reagent (13). We reported that L-ala-p-nitroanilide is a substrate for VanX, and one of the resulting products, p-nitroaniline, could be continuously monitored by Vis spectrophotometry (14). Recent unpublished work from our lab suggested, however, that the observed L-ala-p-nitroanilide hydrolytic activity may be due to a contaminating protein in our purified recombinant VanX samples. This work describes our efforts at showing that VanX does
29 not in fact hydrolyze L-ala-p-nitroanilide and that the hydrolytic activity observed in our samples is due to aminopeptidase N.
Materials and Methods Preparation of VanX: VanX was over-expressed by using either the pET5bvanX or the pIADL14 over-expression plasmids. When using the pET5bvanX system, the previously published purification protocol (14, 15) or a protocol with a modified Q-Sepharose elution gradient was used. The modified elution gradient consisted of holding the NaCl concentrations constant at 100 and 200 mM for 120 mL each. pIADL14, containing the gene that encodes for a maltose binding protein (MBP)-VanX fusion protein (MBP-VanX), was a gift from Professor Christopher Walsh at Harvard Medical School (16). The plasmid was transformed into BL21(DE3) E. coli cells, and colonies containing the plasmid were selected on LB agar plates containing 25 μg/mL kanamycin. A single colony from this plate was used to make an overnight preculture in LB media containing 25 μg/mL kanamycin. Three 1L flasks, each containing 333 mL of LB, were inoculated with 10 mL of the overnight preculture and were allowed to grow to
an O.D.600 nm of 1.6 to 1.8 at 30 °C. Protein induction was induced by making the cultures 1 mM in IPTG. The cells were then collected by centrifugation for 15 minutes at 8,275 g, flash-frozen in liquid N2, and stored at –80 °C overnight. The cell pellets were thawed on ice and were resuspended in 30 mM Tris-HCl, pH 7.6, containing 200 mM NaCl. Cells were lysed by passing the resuspension two times through a french press at 16,000 psi, and the cellular debris was collected by centrifugation for 15 minutes at 32,583 g. The cell lysate was then loaded on an 60 mL amylose column (26 mm X 20 cm), and protein was eluted with a 0-10 mM maltose gradient over 300 mL in 30 mM Tris-HCl, pH 7.6, containing 200 mM NaCl. Fractions containing MBP-VanX were identified by SDS PAGE gels, and pure fractions were pooled and concentrated to approximately 10 mL with an Amicon ultrafiltration concentrator equipped with a YM-10 membrane. The maltose binding protein from the fusion MBP-VanX protein was removed by thrombin digestion. Three units of restriction grade thrombin were added for every one milligram of uncleaved MBP-VanX, and the cleavage was allowed to incubate for 10 hours in a water bath at 25 °C. The cleaved proteins were then separated by two methods: a 120 mL Sephacryl S-200 HR size-exclusion column (16 mm X 60 cm) or a 25 mL Q-Sepharose anion exchange column (16 mm X 20 cm).
30 Pure fractions, as judged by SDS-PAGE, were pooled and concentrated using an Amicon ultrafiltration concentrator equipped with an YM10 membrane. VanX concentrations were
determined by using the sample’s absorbance at 280 nm, the extinction coefficient (ε280 nm) of 51,263 M-1cm-1 calculated by the Edelhoch method (17) and Beer’s Law.
Identification of contaminating enzyme. BL21(DE3) E. coli cells were transformed with pET5b and plated on LB-agar plates containing 200 μg/mL ampicillin. A single colony was used to start a 10 mL preculture, and this culture was used to inoculate 4X1L of LB containing 200 μg/mL ampicillin. The previously published over-expression and purification protocol was used, and fractions from the Q-Sepharose that exhibited L-ala-p-nitroanilide hydrolase activity were pooled and concentrated using an Amicon equipped with a YM-10 membrane. The resulting solution was made 125 mM in (NH4)2SO4, centrifuged, and loaded onto a XK 16/20 phenyl Sepharose column (25 mL bed volume), which was equilibrated with 50 mM Tris, pH 7.0,
containing 1 M (NH4)2SO4. Bound proteins were eluted by running a 1.0 to 0 M (NH4)2SO4 gradient, and eight milliliter fractions were collected. Fractions that exhibited L-ala-p- nitroanilide hydrolase activity were pooled and concentrated using an Amicon equipped with a YM-10 membrane. The concentrated sample from the phenyl Sepharose column was subjected to native gel electrophoresis (18). Efforts to use L-ala-p-nitroanilide as a substrate for activity staining of the resulting gel were unsuccessful. Therefore, the native gel lane containing the concentration sample from the phenyl Sepharose column was cut into 0.5 cm pieces, and the pieces were incubated in 50 mM Hepes, pH 7.5, containing 1.0 mM L-ala-p-nitroanilide. One gel piece contained an enzyme that hydrolyzed L-ala-p-nitroanilide, and this gel piece was subjected to in- gel trypsin digestion (see below).
Coupled assay for VanX. Steady-state kinetics were performed using a coupled assay developed for VanX by Badet and coworkers (12). This assay involves the use of other enzymes, such as D- lactate dehydrogense (LDH)(Sigma), catalase (Cat)(Sigma), and D-amino acid oxidase (DAAO) (Calzyme Laboratories). All enzymes other than VanX were kept in excess so the rate-limiting process in the coupled reactions is the formation of D-Ala by VanX. All reactions were run at 37 ˚C on an Agilent 8453 UV-Vis spectrophotometer. Concentrations of reagents used in the
31 coupled assay were as follows: VanX at 1-5 nM, LDH at 57 units/mL, Cat at 520 units/mL, DAAO at 1.2 units/mL, NADH at 10 mM, and D-Ala-D-Ala was varied from 50 μM to 2 mM with all solutions being made in a 50 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (Hepes), pH 8.0, containing 200 mM NaCl. The reactions were monitored at 340 nm, and the reactions were designed to ensure the loss of ca. 1 AU over the course of ten minutes. Absorbance values for the linear portion of the curves were converted into velocities using the -1 -1 molar extinction coefficient for NADH (ε340 nm = 6,220 M cm ) and the time interval of the absorbance change. These values were then plotted against the corresponding substrate
concentrations, and the resulting plots were fitted to the Michaelis-Menten equation, vo =
Vmax[S]/(Km+[S]) using Igor Pro version 4.05A.
Assays with protease inhibitors. Kinetic assays were performed in 50 mM Hepes, pH 8.0 using an Agilent 8453 UV-Vis spectrophotometer thermostatted at 25 oC. The total protein concentration in the reactions was 5 µM, and the concentration of substrate L-ala-p-nitroanilide was 400 µM. Bestatin (up to 25 µM), pepstatin A (up to 5 µM), leupeptin (up to 25 µM), amastatin (up to 2.5 µM), and PMSF (up to 3 µM) were tested as inhibitors of L-ala-p- nitroanilide hydrolase activity. The formation of p-nitroaniline was followed at 404 nm over a three-minute period.
In-gel trypsin digestions/MALDI-TOF mass spectrum. The bands from the SDS-PAGE gel (after
the native gel) were cut from the gel and washed twice with 100 μL of 25 mM NH4HCO3 / 50% acetonitrile by vortexing the tube for 10 min. After discarding the supernatant, the pieces were dried, and 12.5 ng/μL trypsin (proteomics grade, Sigma) was added to completely cover the gel pieces. After incubating the gel pieces with trypsin overnight, the digest solutions were extracted into clean eppendorf tubes. A volume of 30 μL (or enough to cover) of 50% acetonitrile / 5% formic acid was used to further extract the proteins by vortexing for 30 min. The extracts from trypsin digestions were pooled and concentrated to 10 μL by using a Speed Vac. The extracts were then resuspended in 10 μL of 0.1% TFA, and 2.0 μL of these extracts were mixed with 2 μL α-cyano-4-hydroxycinammic acid in 50% acetonitrile / 0.05%TFA (10 mg/mL). Two microliters of these mixtures were spotted onto the sample target. MALDI-TOF analysis was carried out by using reflectron mode on a Bruker Reflex III MALDI-TOF mass spectrometer. The peptide masses obtained from the MALDI analysis were used to search for the identity of the proteins. Protein identifications were made using three criteria. Firstly, the peak lists from the MALDI-TOF mass spectra were entered into a search engine called MASCOT (Matrix Science:
32 http://www.matrixscience.com/search_form_select.html). A positive protein identification required at least 5 peptide matches and a total MASCOT score outside of the “gray” area. Secondly, different search engines, Profound (http://prowl.rockefeller.edu) and Protein Prospector (http://prospector.ucsf.edu/), were used to verify the results obtained from MASCOT. Thirdly, the predicted molecular weights of the identified proteins were ascertained by using the protcalc program at: http://www.scripps.edu/~cdputnam/protcalc.htm. These data were compared to the position of the gel spots on the corresponding SDS-PAGE gel.
Results L-ala-p-nitroanilide hydrolase activity due to a contaminating enzyme. Recently, efforts to prepare Co(II)-substituted VanX in our lab for spectroscopic studies using a published pET-5b based over-expression system (15) required an alternative elution gradient for the Q-Sepharose chromatography step to obtain highly-purified samples of Co(II)-substituted VanX (R.M. Breece and M.W. Crowder, unpublished results) (19). When using this modified protocol, two distinct fractions were recovered from the Q-Sepharose column: (1) a highly-purified, concentrated, pink-colored sample that is Co(II)-substituted VanX and (2) an orange-colored sample that contains Co(II)-substituted VanX and several other proteins. When these two fractions were tested for activity, the latter sample is the only one that hydrolyzed L-ala-p-nitroanilide. These results suggested that the continuous, UV-Vis assay on which we previously reported (14) may have been based on activity of a contaminating protein. To further probe this possibility, E. coli BL21(DE3) cells were transformed with pET- 5bvanX (14), and modified purification protocols were used to prepare Zn(II)-VanX. We found that the fractions that eluted earliest from the column were purest; however, these samples did not have any L-ala-p-nitroanilide hydrolase activity. On the other hand, fractions that eluted at higher salt concentrations did exhibit L-ala-p-nitroanilide hydrolase activity, and an SDS-PAGE gel of the fraction with highest activity is shown in Figure 1(lane 3). Although this fraction clearly has significant amounts of a protein with a mass of. ca. 23 kDa, there are other contaminating proteins in the lane. E. coli BL21(DE3) cells were then transformed with pET-5b (empty vector), and the resulting cells were used in our modified purification protocol. An SDS-PAGE gel was run on the fractions that eluted at the same salt concentrations as with the fraction with highest L-ala-p-
33 nitroanilide activity as described above (Figure 1(lane 4)). Surprisingly, there was a band that corresponded to ca. 23 kDa and several other bands present in the gel. This fraction exhibited significant L-ala-p-nitroanilide activity. This result indicates that a constitutively-expressed E. coli enzyme exhibits L-ala-p-nitroanilide hydrolase activity. As a final control, the over-expression plasmid for a MBP-VanX fusion protein was obtained from Professor Christopher Walsh, and VanX was over-expressed, purified with an amylose FPLC column, proteolytically-digested with thrombin, and purified as previously described (20). The resulting protein was shown to be pure by SDS-PAGE (Figure 1(lane 2)). This sample of VanX did not hydrolyze L-ala-p-nitroanilide; however, it was shown to bind 0.75 -1 equivalents of Zn(II) and exhibit steady-state kinetic constants of kcat = 156 + 16 s and Km = 109 + 40 μM, when using the coupled assay (12) and D-ala-D-ala as the substrate.
Aminopeptidase N is the contaminating enzyme. To identify the contaminating enzyme, BL21(DE3) E. coli cells containing the empty pET5b vector were grown to an optical density of 0.6, and the cells were harvested by centrifugation and lysed by using a French press. The cleared, soluble protein-containing fraction was loaded onto a Q-Sepharose column, and bound proteins were eluted with a 0 – 500 mM NaCl gradient. The fractions that contained L-ala-p- nitroanilide hydrolase activity were pooled, and the resulting protein solution was loaded onto a
phenyl Sepharose column. Bound proteins were eluted using a 500 to 0 mM (NH4)2SO4 gradient, and fractions with L-ala-p-nitroanilide hydrolase activity were pooled and concentrated. The resulting protein solution contained a number of proteins but was used in the subsequent experiments in this form. The first attempt to identify the contaminating enzyme was to test several classical protease inhibitors on the L-ala-p-nitroanilide hydrolase activity of the contaminant (Table 1). Phenylmethylsulfonyl fluoride (PMSF) did not inhibit the activity even at concentrations up to 3 μM, suggesting that the contaminant is not a serine protease. On the other hand, the incubation of the enzyme sample with leupeptin, pepstatin A, and bestatin resulted in 33%, 56%, and 88% inhibition of L-ala-p-nitroanilide hydrolase activity, respectively. Amastatin, a relatively specific inhibitor of aminopeptidase N and A (21), was subsequently tested and was shown to inhibit L- ala-p-nitroanilide hydrolase activity by 99.5%. This result suggests the contaminant is an
34 aminopeptidase, although the possibility that the contaminant is a cysteine, trypsin, or aspartic acid protease could not be ruled out unequivocally with this study. The next attempt to identify the contaminating enzyme was to run the partially-purified protein mixture on a non-denaturing PAGE gel, and the resulting gel was analyzed by testing gel slices for enzymatic activity. A single gel slice was clearly identified as being responsible for L- ala-p-nitroanilide hydrolase activity; however, the running of the gel slice in a second dimension (SDS-PAGE) showed that there were two proteins (approximate masses of 100 and 75 kDa) in this gel slice (data not shown). In-gel trypsin-digestion of the proteins in this gel slice, MALDI- TOF MS analysis of peptide fragments, and databank searching identified the two proteins as GTP-binding protein chain elongation factor EF-G (molecular weight of 77.6 kDa (22)) and aminopeptidase N (molecular weight of 98-99 kDa (23, 24)).
Discussion Previously, we reported that L-ala-p-nitroanilide is a substrate of VanX, and a continuous, UV-Vis assay was developed for monitoring the activity of VanX (14). To validate this assay, three independent experiments were conducted, and the results from these studies were compared to those previously reported using the D-ala-D-ala/ninhydrin assay. (1) Apo- VanX did not exhibit L-ala-p-nitroanilide hydrolase activity. (2) Apo-VanX could be activated by Co(II), Zn(II), Fe(II), and Ni(II), and the relative trends in activation were similar to those published previously (9). (3) The phosphonate analog of D-ala-D-ala (O-[(1S)- aminoethylhydroxyphosphinyl]-D-lactic acid) was shown to competitively inhibit the L-ala-p- nitroanilide hydrolase activity of VanX with a Ki value of 400 μM, which is the same mode of
inhibition and a similar Ki value as previously reported (9). We subsequently used this assay to test other potential inhibitors and began work on probing the reaction mechanism of VanX (25). Meanwhile, we started a series of spectroscopic studies to probe the Co(II)-substituted enzyme during inhibitor binding and catalysis. As mentioned in the Results section, our efforts to prepare Co(II)-substituted VanX alerted us to the possibility that VanX may not hydrolyze L-ala- p-nitroanilide. Two main lines of evidence demonstrated that the observed L-ala-p-nitroanilide hydrolase activity was due to a constitutively-expressed, contaminating enzyme rather than to VanX. First, VanX, which was alternatively purified or produced in a fusion-construct
35 purification protocol, did not exhibit L-ala-p-nitroanilide hydrolase activity. Second, E. coli cells containing pET5b but no vanx gene yielded a protein sample that eluted from the Q-Sepharose column similarly as VanX, and the sample exhibited L-ala-p-nitroanilide hydrolase activity. To determine the identity of the contaminating enzyme that has L-ala-p-nitroanilide hydrolase activity, two approaches were undertaken. First, protease inhibitors were used to probe whether the contaminating enzyme was a serine protease (PMSF), a cysteine or trypsin- like protease (leupeptin), an aspartic acid protease (pepstatin A), or an aminopeptidase (bestatin). The greatest inhibition of L-ala-p-nitroanilide hydrolase activity was observed with bestatin, suggesting that the contaminating enzyme is an aminopeptidase. Supporting this conclusion, amastatin, a aminopeptidase N or A-specific protease inhibitor (21, 26), was shown to almost completely inhibit L-ala-p-anilide hydrolase activity. Second, a partially-purified protein sample, which contained the enzyme that hydrolyzes L-ala-p-nitroanilide, was subjected to native and SDS-PAGE electrophoresis. By using a proteomics approach, the contaminating enzyme was identified as aminopeptidase N. Aminopeptidase N, a.k.a. lysyl aminopeptidase in bacteria, peptidase N, or alanyl aminopeptidase, is encoded by the pepN gene and is found in bacteria, fungi, and mammals (23, 24, 27, 28). The enzyme is typically monomeric with a molecular weight of 95-100 kDa, has a pH optimum of 5.0 to 7.5, and rapidly hydrolyzes a wide range of peptides, particularly those with bulky or aliphatic N-terminal residues. Previously, L-ala-p-nitroanilide was reported to be a substrate for the enzyme from E. coli (23, 27). This result supports our conclusion that the contaminating enzyme is aminopeptidase N. Aminopeptidase N is completely inhibited by EDTA or 1,10 phenanthroline, and activity can be restored by addition of Mn(II), Co(II), and Zn(II) (23, 28). These characteristics of aminopeptidase N are identical to those of VanX (9). The calculated pI value for aminopeptidase N is 5.3 (27), which explains why it elutes from a Q- Sepharose column slightly after VanX (pI value of 5.8). The similar characteristics of aminopeptidase N to VanX offers an explanation of why our validation experiments for the continuous assay (see above) yielded the false positive results. However, the third validation experiment, inhibition studies with the phosphonate analog of D-ala-D-ala, can only be explained if this compound also inhibits aminopeptidase N, which preliminary studies already suggest. Efforts are underway in the lab currently to over-express, purify, and perform inhibition studies on aminopeptidase N from E. coli.
36 The characterization of aminopeptidase N has far reaching implications since bacterial proteinases are currently being targeted for the development of second-generation antibiotics (29). Aminopeptidase N has been shown to play a vital role in bacterial metabolism, specifically in nitrogen supply and protein turnover (30). The inactivation of aminopeptidase in Lactococcus lactis resulted in significantly lower bacterial growth rates (31). The rapid emergence of antibiotic resistance in bacteria necessitates the discovery of new antibiotic targets, towards which there is no known resistance phenotypes. The better understanding of the structure and mechanism of aminopeptidase N will allow for mechanism-specific or –based inhibitors of this enzyme and hopefully a new way to combat antibiotic resistant bacterial infections.
References
1. Walsh, C. T. Reconstructing vancomycin (1999) Science 284, 442-443. 2. Walsh, C. T. Vancomycin resistance: Decoding the molecular logic (1993) Science 261, 308- 309. 3. Wright, G. D., and Walsh, C. T. D-alanyl-D-alanyl ligases and the molecular mechanism of vancomycin resistance (1992) Acc Chem Res 25, 468-473. 4. Roper, D. I., Huyton, T., Vagin, A., and Dodson, G. The molecular basis of vancomycin resistance in clinically relevant Enterococci: Crystal structure of D-alanyl-D-lactate ligase (VanA) (2000) Proc Natl Acad Sci 97, 8921-8925. 5. Arthur, M., Reynolds, P. E., Depardieu, F., Evers, S., Dutka-Malen, S., Quintiliani, R., and Courvalin, P. Mechanisms of glycopeptide resistance in Enterococcus (1996) J Infect 32, 11-16. 6. Reynolds, P. E., Depardieu, F., Dutka-Malen, S., Arthur, M., and Courvalin, P. Glycopeptide resistance mediated by Enterococcal transposon Tn1546 requires production of VanX for hydrolysis of D-alanyl-D-alanine (1994) Mol Microbiol 13, 1065-1070. 7. Bussiere, D. E., Pratt, S. D., Katz, L., Severin, J. M., Holzman, T., and Park, C. H. The structure of VanX reveals a novel amino-dipeptidase involved in mediating transposon- based vancomycin resistance (1998) Mol Cell 2, 75-84. 8. Wu, Z., and Walsh, C. T. Phosphinate analogs of D-,D-dipeptides: Slow-binding inhibition and proteolysis protection of VanX, a D-, D-dipeptidase required for vancomycin resistance in Enterococcus faecium (1995) Proc Natl Acad Sci USA 92, 11603-11607.
37 9. Wu, Z., Wright, G. D., and Walsh, C. T. Overexpression, purification, and characterization of VanX, a D-, D-dipeptidase which is essential for vancomycin resistance in Enterococcus faecium BM4147 (1995) Biochemistry 34, 2455-2463. 10. Wu, Z., and Walsh, C. T. Dithiol compounds: Potent, time-dependent inhibitors of VanX, a zinc-dependent D,D-dipeptidase required for vancomycin resistance in Enterococcus faecium (1996) J Am Chem Soc 118, 1785-1786. 11. Tu, J., and Chu, Y. H. Vancomycin resistance in Enterococcus faecium: A capillary electrophoresis-based for VanX enzyme (1998) Anal Biochem 264, 293-296. 12. Yaouancq, L., Anissimova, M., Badet-Denisot, M. A., and Badet, B. Design and evaluation of mechanism-based inhibitors of D- alanyl-D-alanine dipeptidase VanX (2002) Eur J Org Chem, 3573-3579. 13. Anissimova, M., Yaouancq, L., Badet-Denisot, M. A., and Badet, B. New chromogenic dipeptide substrate for continuous assay of the D-alanyl-D-alanine dipeptidase VanX required for high-level vancomycin resistance (2003) J Peptide Res 62, 88-95. 14. Brandt, J. J., Chatwood, L. L., Yang, K. W., and Crowder, M. W. Continuous assay for VanX, the D-alanyl-D-alanine dipeptidase required for high-level vancomycin resistance (1999) Anal Biochem 272, 94-99. 15. Brandt, J. J., Chatwood, L. L., and Crowder, M. W. Analysis of three overexpression systems for VanX, the Zn(II)-dipeptidase required for high-level vancomycin resistance in bacteria (2000) Prot Express Purif 20, 300-307. 16. McCafferty, D. G., Lessard, I. A. D., and Walsh, C. T. Mutational analysis of potential zinc- binding residues in the active site of the Enterococcal D-ala-D-ala dipeptidase VanX (1997) Biochemistry 36, 10498-10505. 17. Edelhoch, H. Spectroscopic determination of tryptophan and tyrosine in proteins (1967) Biochemistry 6, 1948-1954. 18. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular cloning - a laboratory manual, Vol. 1, Second ed., Cold Spring Harbor Laboratory Press. 19. Breece, R. M., Costello, A., Bennett, B., Sigdel, T. K., Matthews, M. L., Tierney, D. L., and Crowder, M. W. A five-coordinate metal center in Co(II)-substituted VanX (2005) J Biol Chem 280, 11074-11081.
38 20. Lessard, I. A. D., and Walsh, C. T. Mutational analysis of active-site residues of the Enterococcal D-ala-D-ala dipeptidase VanX and comparison with Escherichia coli D- ala-D-ala ligase and D-ala-D-ala carboxypeptidase VanY (1999) Chem Biol 6, 177-187. 21. Rich, D. H., Moon, B. J., and Harbeson, S. Inhibition of aminopeptidases by amastatin and bestatin derivatives - effect of inhibitor structure on slow-binding processes (1984) J Med Chem 27, 417-422. 22. Kim, K. K., Min, K., and Suh, S. W. Crystal structure of the ribosome recycling factor from Escherichia coli (2000) EMBO J 19, 2362-2370. 23. McCaman, M. T., and Villarego, M. R. Structural and catalytic properties of peptidase N from Escherichia coli K-12 (1982) Arch Biochem Biophys 213, 384-394. 24. McCaman, M. T., McPartland, A., and Villarego, M. R. Genetics and regulation of peptidase N in Escherichia coli K-12 (1982) J Bact 152, 848-854. 25. Yang, K. W., Brandt, J. J., Chatwood, L. L., and Crowder, M. W. Phosphonamidate and phosphothioate dipeptides as potential inhibitors of VanX (2000) Bioorg Med Chem Lett 10, 1087-1089. 26. Yoshimoto, T., Tamesa, Y., Gushi, K., Murayama, N., and Tsuru, D. An aminopeptidase-N from Escherichia-coli-HB101 - purification and demonstration that the enzyme possesses arylamidase and peptidase activities (1988) Agric Biol Chem 52, 217-225. 27. Klein, J. R., Klein, U., Schad, M., and Plapp, R. Cloning, DNA sequence analysis and partial characterization of PepN, a lysyl aminopeptidase from Lactobacillus delbrukii ssp. lactis DSM7290 (1993) Eur J Biochem 217, 105-114. 28. Chavagnat, F., Casey, M. G., and Meyer, J. Purification, characterization, gene cloning, sequencing, and overexpression of aminopeptidase N from Streptococcus thermophilus A (1999) Appl Environ Microbiol 65, 3001-3007. 29. Travis, J., and Potempa, J. Bacterial proteinases as targets for the development of second- generation antibiotics (2000) Biochim Biophys Acta 1477, 35-50. 30. Barrett, A. J., Rawlings, N. D., and Woessner, J. F. (1998) Handbook of proteolytic enzymes, 1st ed., Academic Press, New York. 31. Mierau, I., Haandrikman, A. J., Velterop, O., Tan, P. S., Leenhouts, K. L., Konings, W. N., Venema, G., and Kok, J. Tripeptidase gene (pepT) of Lactococcus lactis: Molecular
39 cloning and nucleotide sequencing of pepT and construction of a chromosomal deletion mutant (1994) J Bacteriol 176, 2854-61. 32. Taylor, A. (1996) Aminopeptidases, R.G. Landes Company, Austin, TX. 33. Hernandez, A. A., and Roush, W. R. Recent advances in the synthesis, design and selection of cysteine protease inhibitors (2002) Curr Opin Chem Biol 6, 459-465. 34. Ripka, A. S., Satyshur, K. A., Bohacek, R. S., and Rich, D. H. Aspartic protease inhibitors designed from computer-generated templates bind as predicted (2001) Org Lett 3, 2309- 2312. 35. Leung, D., Abbenante, G., and Fairlie, D. P. Protease inhibitors: Current status and future prospects (2000) J Med Chem 43, 305-341.
40 1 2 3 4
Figure 1: SDS-PAGE gel of samples characterized in this work. Lane 1: Novagen perfect protein molecular weight markers (from top to bottom: 150, 100, 75, 50, 35, 25, and 15 kDa); Lane 2: VanX over-expressed and purified using the pIADL14 over-expression plasmid and the procedure of McCafferty et al. (16); Lane 3: VanX over-expressed and purified using the pET5b over-expression plasmid and the procedure of Brandt et al. (14); Lane 4: Q-Sepharose column fraction of proteins produced from BL21 (DE3) E. coli cells containing the empty pET5b over- expression plasmid.
41 Table 1: Protease inhibitors tested. Inhibitor % inhibition reference Targeted protease
bestatin aminopeptidase B 88 ± 3 (32) leucine aminopeptidase aminopeptidase N leupeptin cysteine proteases 33 ± 15 (33) pepstatin A aspartic proteases 56 ± 5 (34) amastatin leucine aminopeptidase >99.5 (26, 32) aminopeptidase A aminopeptidase N PMSF serine protease 0 (35) Assays were conducted as described in Materials and Methods.
42 Chapter 3
Over-expression, Purification, and Characterization of Aminopeptidase N
(PepN) from Escherichia coli
The material in this chapter has been accepted for publication: Frank C. Golich, Maria Han, and Michael W. Crowder. “Over-expression, Purification, and Characterization of Aminopeptidase N (PepN) from Escherichia coli” Prot. Express. Purif. (2006).
43 Introduction The roles of aminopeptidases in prokaryotes are quite varied, with proposed functions such as protein degradation, nutrient utilization, and down-stream processing (1-3). Within eukaryotes, it is becoming clear that disruption in the regulation of even a single aminopeptidase may be linked to a host of diseases (4). A similar link is less clear in bacteria, but it is known that many if not most aminopeptidases are necessary for bacterial survival (5, 6). In contrast to the situation in eukaryotes, prokaryotes display redundancy in the expression of degradative proteins, possibly indicative of their absolute necessity for survival (3). Recently, aminopeptidases have been identified as potential therapeutic and antibiotic targets (7). Aminopeptidase N (PepN), a.k.a. lysyl aminopeptidase (8), has been reported to be a broad specifity metalloaminopeptidase found in bacteria, fungi, and mammals (6). In mammals, PepN is membrane-associated and has been implicated as a receptor for several viruses and in the activation mechanism of collagenolysis that allows for tumor cell invasion (9-11). In bacteria, E. coli PepN has been studied for nearly 30 years; however, the Lactococcus and Lactobacillus enzymes have been more thoroughly studied because of their roles in the cheese and dairy industries (12-15). Bacterial PepN enzymes are apparently cytosolic, and the enzymes are sensitive to chelating agents (8). Aminopeptidase N in E. coli has been studied relatively less; however, Chandu and Nandi have recently reported that PepN is the major aminopeptidase in E. coli that is involved in ATP-independent processing during cytosolic protein degradation (1, 2). In addition, it is the only known alanyl aminopeptidase in E. coli (2). Early knock-out lines of E. coli PepN did not demonstrate a discernable phenotype (16, 17); nonetheless, Chandu and Nandi recently reported that PepN is a negative regulator of the sodium salicylate induced stress (2). Travis et al. have conducted a review of bacterial proteinases and their potential as targets for second-generation antibiotics (7). Aminopeptidases were identified as one enzyme class that could serve as an antibiotic target. Our interest in PepN is based upon its proposed important role as the sole alanyl aminopeptidase present in E. coli (18, 19) and focuses upon its possible susceptibility as an adjunctive target for reversal of antibiotic resistance. In order to conduct future mechanistic, spectroscopic, and inhibition studies, we anticipate the need for larger quantities of the purified enzyme. This work describes our efforts to develop a reproducible and efficient over-expression and purification protocol for recombinant PepN from E. coli.
44
Materials and Methods Cloning of the pepN gene. The polymerase chain reaction (PCR) was used to amplify the pepN open reading frame (20) from genomic E. coli DNA, strain W3110. The primers used were AAAAAAACATATGACTCAACAGCCAC and AAAAAGCTTCAAGCCAGTTTAGT, which introduced unique NdeI and HindIII restriction sites around the pepN gene. The 2.6 kb PCR fragment was digested with NdeI and HindIII and ligated into pET26b, which was digested with the same restriction enzymes, to form the over-expression plasmid pET26bpepN. The over- expression plasmid was transformed into competent BL21(DE3) E. coli cells, and the resulting mixture was plated on LB agar plates containing 25 μg/mL kanamycin. The gene was confirmed by DNA sequencing.
Over-expression and purification of recombinant PepN. An overnight pre-culture of E. coli BL21(DE3) containing the pET26bpepN plasmid was used to inoculate 3 x 1L flasks of LB containing 25 μg/mL kanamycin, and the cultures were allowed to grow with shaking at 37 °C until they reached an optical density of 0.6–0.8 (OD600nm). Protein production was induced by making the cell cultures 1 mM in IPTG, and the cells were shaken at 37 °C for an additional 2-3 hours. Cells were collected by centrifugation (15 minutes at 8,200xg) and were resuspended in 15 mL of cold, 50 mM Tris-HCl, pH 8.5. The resuspended cells were were lysed by sonication using a Fisher Scientific model 100 sonic dismembrator. Cell debris was removed from the sample by centrifugation (15 minutes at 23,400xg), and the crude protein solution was dialyzed overnight at 4 °C versus 2 L of 50 mM Tris-HCl, pH 8.5. The dialyzed, crude protein solution was then centrifuged for 15 minutes at 23,400xg to remove any precipitated proteins or lipids, and the cleared supernatant was loaded onto a 16 x 20 mm Q-Sepharose column that had been previously equilibrated with 300 mL of 50 mM Tris-HCl, pH 8.5. Bound proteins were eluted from the column with a linear gradient of 0–500 mM NaCl in 50 mM Tris–HCl, pH 8.5, at a flow rate of 2 mL/min. The gradient increased at a rate of approximately 1% per minute, and 8 mL fractions were collected. Fractions containing PepN were identified by SDS–PAGE, after staining the gels with Coomassie blue. Those fractions exhibiting bands at approximately 99 kDa with greater than 95% purity were pooled and concentrated in an Amicon ultrafiltration cell equipped with an YM-10 cellulose membrane. Protein concentrations were determined by using
45 an extinction coefficient of 147,595 M-1 cm-1 which was determined by amino acid analysis and confirmed using a Bradford assay (21). The authors thank Dilip Chandu and Dipankar Nandi for providing us with the PepN knock-out line (DH5αΔpepN) and the pBAD/EcPepN over- expression construct.
Metal analyses. The metal content of PepN samples was determined using a Varian Liberty 2 Inductively Coupled Plasma spectrometer with atomic emission spectroscopy detection (ICP- AES). The final dialysis buffer (50 mM Tris-HCl, pH 8.5) was used as a blank, and the enzyme concentration was 10 μM. Calibration curves for all metals tested were based on four standards, and all calibration curves had correlation coefficients of 0.9950 or greater. Emission lines at 213.856, 238.892, 324.754, 259.940, 257.610, and 231.604 nm, the most intense emissions for zinc, cobalt, copper, iron, manganese, and nickel, respectively, were used to determine metal content for each enzyme preparation.
Steady-state kinetics. Steady-state kinetic studies were conducted using L-alanine-p-nitroanilide purchased from Sigma (St. Louis, MO). Kinetic assays were carried out on an Agilent 8453 UV/Vis spectrophotometer, using an Isotemp circulator to maintain the reactions at 37 °C. The buffer was 50 mM Tris-HCl, pH 8.5. Enzyme concentrations in the cuvette were 1-10 nM, and substrate concentrations were varied from 50 μM to 1 mM. The reaction of PepN and L-alanine- p-nitroanilide was followed at 404 nm, and absorbance changes were converted to concentration changes using the molar extinction coefficient of liberated p-nitroaniline (ε404nm = 10,700 -1 -1 M cm ). Km and kcat values were determined by least-squared fitting of the kinetic data to the Michaelis Menten equation. Reported errors reflect fitting errors.
Gel Filtration chromatography. Gel filtration chromatography was performed on an 1.6 X 60 cm Amersham Pharmacia Biotech Sephacryl S-200 column with a flow rate of 0.8 mL/min. The column was equilibrated with 50 mM phosphate buffer, pH 7.2, containing 150 mM NaCl. Catalase (232 kDa), ovalbumin (43 kDa), and bovine serum albumin (66 kDa) were used as molecular weight standards, and Blue Dextran 2000 was used to measure the void volume of the column.
46 Amino Acid Analysis. Amino acid analysis was performed at the Protein Separation and Analysis Laboratory, Purdue University, West Lafayette, Indiana, using a Beckman 166 HPLC system and Waters AccQTag amino acid analysis column and buffers. The PepN enzyme sample was purified using a Centri-Spin 40 spin column, Princeton Separations (Adelphia, NJ) prior to amino acid analysis. Molar extinction coefficient (ε280nm) was calculated to be 147,595 M-1cm-1.
MALDI-TOF MS. Mass spectra of purified PepN samples were acquired on a Bruker Reflex III time-of-flight (TOF) mass spectrometer operating in the linear mode as previously described (22). In-gel trypsin digestions and protein identifications were performed as described previously (23).
Results The pepN gene was ligated into pET26b allowing for protein over-expression under the control of a lac promoter and using kanamycin antibiotic selection. The pET26bpepN over- expression vector was then transformed into BL21(DE3) E. coli cells, which were used to over- express soluble protein. Over-expression levels were optimized using test cultures (22). The greatest levels of over-expressed PepN were obtained by allowing the bacterial cultures to grow to an O.D.600nm of 0.6-0.8 before induction and by inducing protein production for 2-3 hours at 37 oC with 1 mM IPTG. Recombinant PepN was able to be purified to greater than 95% purity by using a single Q-Sepharose column equilibrated to pH 8.5 (Figure 1 and Table 1), and the yield was 53 mg from a 3 L culture. In some of the preparations, two proteins bands (<5% of total protein) with apparent molecular weights of approximately 55 kDa and 42 kDa were observed in the SDS–PAGE gels (not shown). In-gel trypsin digestions, MALDI-TOF mass spectra, and database searches (23) identified the bands as the endogenously-expressed E. coli proteins glycerol kinase and elongation factor-tu (EF-Tu) chain G, respectively. A review of the literature revealed no reports of either of the proteins possessing L-alanine-p-nitroanilide hydrolase activity. In order to further ensure that the L-alanine-p-nitroanilide hydrolase activity identified in these cultures was solely from either endogenously- or recombinantly-expressed PepN, an E. coli knockout line (DH5αΔpepN) was used (2). After growth and induction, the cultures were
47 handled as described above, and the same fractions in which PepN eluted from the FPLC column were pooled and concentrated. Activity assays demonstrated that there was no L-alanine-p- nitroanilide hydrolase activity in these fractions. The amount of endogenously-expressed PepN in our BL21(DE3) E. coli cells was ascertained by performing our over-expression and purification protocol using E. coli cells transformed with empty (pET26b) vector. The FPLC fractions in which recombinant PepN elutes did demonstrate some hydrolase activity (<1 % of the fractions obtained in the over-expressed protocol). Recombinant PepN is isolated containing <0.1 equivalents of Zn(II) and 0.5 equivalents of iron. The as-isolated enzyme was incubated with up to 2 equivalents of Zn(II), and the resulting enzyme solution was exhaustively dialyzed versus 3 changes of 50 mM Tris-HCl, pH 8.5, containing 100 mM NaCl. ICP-AES analyses of these samples revealed that PepN tightly binds only 0.08 equivalents of Zn(II). However if as-isolated PepN is incubated with 10 equivalents of ferrous ammonium sulfate and the samples are dialyzed against three changes of two liters each of 50 mM Tris-HCl, pH 8.5, containing 100 mM NaCl (or passed through a Sephadex G-25 column (24)), ICP-AES analyses showed that the resulting enzyme tightly binds 5 equivalents of iron. Size exclusion chromatography revealed that recombinant E. coli aminopeptidase N is a monomeric protein with an approximate mass of 99 kDa. MALDI-TOF mass spectrometry revealed that PepN has a molecular weight of 98,750 m/z, which is very similar to the predicted molecular mass as predicted by the deduced amino acid sequence (25).
Steady-state kinetic studies demonstrated that as-isolated PepN exhibits a kcat of 354 + 11 -1 s and a Km of 376 + 39 μM when using L-alanine-p-nitroanilide as the substrate. Addition of 1.0 equivalent of Zn(II) to this enzyme resulted a 60% loss in activity. PepN containing 5 -1 equivalents of Fe exhibited a kcat of 236 + 23 s and a Km of 231 + 91 μM. To determine whether PepN is a metalloenzyme, apo-PepN was prepared by dialyzing as-isolated enzyme versus 50 mM Tris-HCl, pH 8.5, containing 10 mM EDTA. After removal of the chelating agent by using dialyses, the resulting apo-enzyme was shown to contain < 0.1 equivalents of Fe or Zn -1 and exhibited a kcat of 123 + 80 s and a Km of 217 + 33 μM.
48 Discussion E. coli aminopeptidase N was first reported by Lazdunski et al. in 1975 (26), and numerous manuscripts have been reported on this enzyme and the corresponding enzymes in other organisms (7, 9-12, 14, 16, 18, 27-34). Nonetheless, there appears to be a great deal of controversy about E. coli PepN. For example, the enzyme has been reported to be an aminoexopeptidase (18) and an aminoendopeptidase (1, 2, 26, 27). E. coli aminopeptidase N was initially thought to be a cysteine active site peptidase (18, 26, 27); however, recent reports suggest that the enzyme is a Zn(II)-containing peptidase (6, 8). There are conflicting reports as to the localization of the enzyme in vivo (18, 26, 35), and there is no structural information available for the enzyme. While there are a number of reports describing the over-expression and purification of aminopeptidase N’s from various organisms (12, 28, 30) , almost all of the studies conducted on E. coli PepN have used constitutively-expressed PepN that was purified using multiple chromatography steps (18, 26, 27, 36). These procedures yield only small amounts of purified enzyme from cell cultures as large as 100 L, and typically only kinetic studies were performed. Although McCaman and Gabe reported an over-expression construct (no purification or characterization) in 1986 (17), the first reported over-expression/purification protocol for E. coli aminopeptidase N was reported in 2003 by Chandu et al. (1, 2). In this procedure E. coli aminopeptidase N was cloned into pBAD24, which places the over-expression of the target protein under the control of an arabinose-regulated promoter, and the resulting over- expressed protein was purified using a three-step chromatography procedure (Chandu, et al reported a yield of 77 mg purified protein following their three-step purification protocol; however, no initial culture volume was provided). We obtained this over-expression plasmid; however, we were unable to over-express and purify suitable quantities of PepN using this construct. Therefore, we cloned E. coli PepN into pET26b, which places protein over-expression under the control of a lac promoter that is inducible with IPTG. As shown in Figure 1 and Table 1, PepN is over-expressed at high levels (53 mg from a 3L culture) when using this over- expression plasmid and the protocol described in Materials and Methods. Recombinant PepN can be purified to >95% purity by using a single chromatography step. The resulting enzyme exhibits -1 -1 a steady-state kcat of 354 s , which is very similar to the value (370 s ) recently reported by Chandu et al., when using L-ala-p-nitroanilide as substrate (1). No other papers on E. coli PepN
49 reported activities as kcat values. Previously, the steady-state Km value for L-ala-p-nitroanilide hydrolysis by E. coli PepN has been reported to range between 200-400 μM (2, 18, 27) (one
group reported a Km of 2.86 μM (36)). Another reported characteristic of E. coli PepN is that the enzyme is inhibited by Zn(II) (27, 36), despite predictions that the E. coli PepN is a Zn(II)- metalloenzyme ((6) and references within). Recombinant E. coli PepN was strongly inhibited by Zn(II) (60% inhibition at 1 equivalent of Zn(II)). These steady-state kinetic results along with the gel filtration data support that recombinant E. coli PepN described herein is similar to the enzymes that were previously partially-characterized. Metal analyses demonstrate that recombinant E. coli PepN is not a Zn(II)- metalloenzyme. Based on chemical modification and pH dependence studies (18, 26, 27), E. coli PepN was initially predicted to be a cysteine-active site aminopeptidase. Based entirely on amino acid sequence comparisons, the enzyme was subsequently predicted to be a Zn(II)- containing aminopeptidase (6). In support of this prediction was the fact that E. coli PepN is inhibited by phenanthroline and that activity can be restored by addition of certain divalent metal ions (2, 36). However, E. coli PepN is not inhibited by EDTA (18, 26, 36), which is a better Zn(II) chelator than phenanthroline. The inhibition of PepN by phenanthroline must be due to the chelator binding to and inhibiting the enzyme and not strictly by metal chelation. The restoration of activity by addition of metal ions could be explained by metal-phenanthroline complexes not being inhibitors of the enzyme. The inhibition by Zn(II) could be explained by Zn(II) coordinating the catalytic cysteine at high Zn(II) concentrations. Recently, Laermans et al. have reported that metal ions can modulate the activities of a cystinyl aminopeptidase (37). The behavior of PepN from Lactobacillus delbruckii (30), Streptococcus thermophilus (28), and mammals (38) towards chelators and Zn(II) is very different than that of E. coli PepN. Therefore, it is possible that aminopeptidase N’s from other organisms are metalloenzymes. One of the most surprising results in this work is that recombinant E. coli PepN tightly binds 5 equivalents of iron and that iron binding is not essential for catalytic activity. In fact, the presence of iron may result in a slightly less active enzyme. Recently, Liu et al. reported the crystal structure of PhoU, which contains a novel multinuclear iron cluster of unknown function (39). The iron clusters are coordinated by conserved E(D)XXXD motifs; there are six such motifs in E. coli PepN, suggesting that this enzyme may also contain a multinuclear iron cluster.
50 Experiments are underway to fully characterize the structure of E. coli PepN, particularly the iron centers.
References 1. Chandu, D., Kumar, A., and Nandi, D. PepN, the major suc-llvy-amc-hydrolyzing enzyme in Escherichia coli, displays functional similarity with downstream processing enzymes in archaea and eukarya (2003) J Biol Chem 278, 5548-5556. 2. Chandu, D., and Nandi, D. PepN is the major aminopeptidase in Escherichia coli: Insights on substrate specificity and role during sodium-salicylate- induced stress (2003) Microbiol- Sgm 149, 3437-3447. 3. Taylor, A. Aminopeptidases: Towards a mechanism of action (1993) Trends Biochem Sci 18, 167-172. 4. Bradshaw, R. A. Aminopeptidases (2004) Encyclopedia Biol Chemistry 1, 96-98. 5. Jankiewicz, U., and Bielawski, W. The properties and functions of bacterial aminopeptidases (2003) Acta Microbiol Polon 52, 217-231. 6. Gonzales, T., and Robert-Baudouy, J. Bacterial aminopeptidases: Properties and functions (1996) FEMS Microbiol Lett 18, 319-344. 7. Travis, J., and Potempa, J. Bacterial proteinases as targets for the development of second- generation antibiotics (2000) Biochim Biophys Acta 1477, 35-50. 8. Klein, J. R., and Henrich, B. (1998) in Handbook of proteolytic enzymes (Barrett, A. J., Rawlings, N. D., and Woessner, J. F., Eds.) pp 1-4, Academic Press, London. 9. Yeager, C. L., Ashmun, R. A., Williams, R. K., Cardellichio, C. B., Shapiro, L. H., Look, A. T., and Holmes, K. V. Human aminopeptidase N is a receptor for human coronavirus 229e (1992) Nature 357, 420-422. 10. Saiki, I., Fujii, H., Yoneda, J., Abe, F., Nakajima, M., Tsuruo, T., and Azuma, I. Role of aminopeptidase N (CD13) in tumor-cell invasion and extracellular matrix degradation (1993) Int J Cancer 54, 137-143. 11. Delmas, B., Gelfi, J., L'Haridon, R., Vogel, L. K., Sjostrom, H., Noren, O., and Laude, H. Aminopeptidase N is a major receptor for the entero-pathogenic coronavirus TGEV (1992) Nature 357, 417-420.
51 12. Van Allen-Boerrigter, I. J., Baankreis, R., and De Vos, W. M. Characterization and overexpression of the Lactococcus lactis pepN gene and localization of its product, aminopeptidase N (1991) Appl Environ Microbiol 57, 2555-2561. 13. Magboul, A., and Sweeney, P. L. H. Pepn-like aminopeptidase from Lactobacillus curvatus DPC2024: Purification and characterization (1999) Lait 79, 515-526. 14. Tan, P. S. T., Van Allen-Boerrigter, I. J., Poolman, B., Siezen, R. J., De Vos, W. M., and Konings, W. N. Characterization of the Lactococcus lactis pepN gene encoding an aminopeptidase homologous to mammalian aminopeptidase N (1992) FEBS Lett 306, 9- 16. 15. Sanz, Y., and Toldra, F. Purification and characterization of an arginine aminopeptidase from lactobacillus sakei (2002) Appl Environ Microbiol 68, 1980-1987. 16. Bally, M., Murgier, M., and Lazdunski, A. Molecular cloning and amplification of the gene for aminopeptidase N of Escherchia coli (1983) FEMS Microbiol Lett 19, 261-265. 17. McCaman, M. T., and Gabe, J. D. The nucleotide-sequence of the pepN gene and its over- expression in Escherichia-coli (1986) Gene 48, 145-153. 18. McCaman, M. T., and Villarego, M. R. Structural and catalytic properties of peptidase N from Escherichia coli K-12 (1982) Arch Biochem Biophys 213, 384-394. 19. Wilkes, S. H., and Prescott, J. M. The slow, tight binding of bestatin and amastatin to aminopeptidases (1985) J Biol Chem 260, 13154-13162. 20. Bally, M., Foglino, M., Bruschi, M., Murgier, M., and Lazdunski, A. Nucleotide sequence of the promoter and amino-terminal coding region of the Escherchia coli pepN gene (1986) Eur J Biochem 155, 565-569. 21. Bradford, M. M. A rapid and sensitive method for the quantification of microgram quantities of protein utilizing the principle of protein-dye binding (1976) Anal Biochem 72, 248- 254. 22. Crowder, M. W., Walsh, T. R., Banovic, L., Pettit, M., and Spencer, J. Overexpression, purification, and characterization of the cloned metallo-b-lactamase L1 from Stenotrophomonas maltophilia (1998) Antimicrob Agents and Chemother 42, 921-926. 23. Sigdel Tara, K., Cilliers, R., Gursahaney Priya, R., and Crowder Michael, W. Fractionation of soluble proteins in Escherichia coli using DEAE-, SP-, and Phenyl sepharose chromatographies (2004) J Biomol Tech: JBT 15, 199-207.
52 24. Yang, K. W., and Crowder, M. W. Method for removing EDTA from apo-proteins (2004) Anal Biochem 329, 342-344. 25. Foglino, M., Gharbi, S., and Lazdunski, A. Nucleotide-sequence of the pepN gene encoding aminopeptidase-N of Escherichia-coli (1986) Gene 49, 303-309. 26. Lazdunski, C., Busuttil, J., and Lazdunski, A. Purification and properties of a periplasmic aminoendopeptidase from Escherchia coli (1975) Eur J Biochem 60, 363-369. 27. Chappelet-Tordo, D., Lazdunski, C., Murgier, M., and Lazdunski, A. Aminopeptidase N from Escherichia coli: Ionizable active-center groups and substrate specificity (1977) Eur J Biochem 81, 299-305. 28. Chavagnat, F., Casey, M. G., and Meyer, J. Purification, characterization, gene cloning, sequencing, and overexpression of aminopeptidase N from Streptococcus thermophilus A (1999) Appl Environ Microbiol 65, 3001-3007. 29. Femfert, U. On the mechanism of amide bond cleavage catalyzed by aminopeptidase M. Kinetic studies in deuteriumoxide (1971) FEBS Lett 14, 92-94. 30. Klein, J. R., Klein, U., Schad, M., and Plapp, R. Cloning, DNA sequence analysis and partial characterization of pepN, a lysyl aminopeptidase from Lactobacillus delbrukii ssp. lactis DSM7290 (1993) Eur J Biochem 217, 105-114. 31. Lazdunski, A., Murgier, M., and Lazdunski, C. Phospholipid synthesis-dependent activity of aminopeptidase N in intact cells of Escherichia coli (1979) J Mol Biol 128, 127-141. 32. McCaman, M. T., McPartland, A., and Villarego, M. R. Genetics and regulation of peptidase N in Escherichia coli K-12 (1982) J Bact 152, 848-854. 33. Mierau, I., Kunji, E. R. S., Leenhouts, K. J., Hellendoorn, M. A., Haandrikman, A. J., Poolman, B., Konings, W. N., Venema, G., and Kok, J. Multiple-peptidase mutants of Lactococcus lactis are severely impaired in their ability to grow in milk (1996) J Bacteriol 178, 2794-2803. 34. Nagai, M., Kojima, F., Naganawa, H., Hamada, M., Aoyagi, T., and Takeuchi, T. Phebestin, a new inhibitor of aminopeptidase N, produced by Streptomyces sp. Mj716-m3 (1997) J Antibiot 50, 82-84. 35. Murgier, M., Pellissier, C., Lazdunski, A., and Lazdunski, C. Existence, localization and regulation of the biosynthesis of aminoendopeptidase in gram negative bacteria (1976) Eur J Biochem 65, 517-520.
53 36. Yoshimoto, T., Tamesa, Y., Gushi, K., Murayama, N., and Tsuru, D. An aminopeptidase-N from Escherichia-coli-HB101 - purification and demonstration that the enzyme possesses arylamidase and peptidase activities (1988) Agric Biol Chem 52, 217-225. 37. Laeremans, H., Demaegdt, H., De Backer, J. P., Le, M. T., Versemans, V., Michotte, Y., Vauguelin, G., and Vanderheyden, P. M. L. Metal ion modulation of cystinyl aminopeptidase (2005) Biochem J 390, 351-357. 38. Wang, J., and Cooper, M. D. (1996) in Zinc metalloproteases in health and disease (Hopper, N. M., Ed.) pp 131-151, Taylor & Francis Inc., Bristol, PA. 39. Liu, J., Lou, Y., Yokota, H., Adams, P. D., Kim, R., and Kim, S. H. Crystal structure of a PhoU protein homologue (2005) J Biol Chem 280, 15960-15966.
54 Table 1: Purification of recombinant E. coli PepN Step Total protein Total activityb Specific Yield (%) Purification (mg)a activityc level Sonicationd 255 64,500 253 100 1 Overnight 218 56,200 257 87 1.02 dialysise Q-Sepharosef 53 51,100 964 79 3.8 aTotal protein was determined by using a Bradford assay (21). Typically, 20 g of wet cells were collected from a 3 L culture. bTotal activity was determined by monitoring L-alanine-p-nitroanilide hydrolysis. The units for total activity are nmoles of L-alanine-p-nitroanilide hydrolyzed per second. cThe units for specific activity are nmoles of L-alanine-p-nitroanilide hydrolyzed per second per milligram of total protein. dSample was taken after disrupting E. coli cells from a 3 L culture with sonication and centrifuging the sample (see Materials and Methods). eSample was taken after the soluble protein extract was dialyzed overnight and centrifuged to remove precipitated proteins (see Materials and Methods) fSample was taken from Q-Sepharose fractions that had been pooled and concentrated (see Figure 1).
55 1 2 3 4 5 6
Figure 1. SDS–PAGE gel (5.2% acrylamide) of recombinant PepN purification from pET26b- PepN over-expression system. Lane 1, Novagen Perfect Protein molecular weight markers (shown from top to bottom: 150, 100, and 75 kDa); lane 2, boiled cell fraction of BL21(DE3) E. coli cells containing the pET26b-PepN plasmid before induction; lane 3, boiled cell fraction of BL21(DE3) E. coli cells containing the pET26b-PepN plasmid after a 2 h induction period with 1mM IPTG; lane 4, crude protein after sonication and centrifugation; lane 5, crude protein following overnight dialysis; and lane 6, purified PepN.
56 Chapter 4
Spectroscopic studies on recombinant E. coli PepN
Introduction
Aminopeptidases are a ubiquitous class of peptidases that catalyze the removal of N- terminal amino acids from peptides and proteins (1, 2). Early studies on aminopeptidases focused on aminopeptidase N (PepN) and its role in peptide degradation and its importance to the commercial dairy industry (3-5). Another major area of study involved the role of PepN in cellular housekeeping functions such as degradation of signaling peptides and nutrient homeostasis within the cell (6). The mammalian form of aminopeptidase N (CD13) is involved in tumor invasion and angiogenesis (7-9). Mammalian PepN is membrane-associated and has been identified as a receptor for several viruses and in the activation mechanism of collagenolysis that allows for tumor cell invasion (1, 7). E. coli aminopeptidase N is a monomeric, cytosolic enzyme with a molecular weight of 98-99 kDa (10) that cleaves a broad range of amino acids from peptide substrates with a preference for small hydrophobic residues at the N-terminus (11). Chandu and Nandi have asserted that PepN in the sole alanyl aminopeptidase in E. coli (6, 11, 12). Chappelet-Tordo and coworkers previously used chemical modification agents on partially-purified enzyme to hypothesize that E. coli PepN is a cysteinyl aminopeptidase that utilizes an active site cysteine in the nucleophilic attack on substrate (13). Subsequently, the DNA (and amino acid) sequence of E. coli PepN was reported (14) and revealed that E. coli PepN has an HEXXH motif, which is a common Zn(II) binding motif found in many proteins. Based essentially on the presence of this motif, E. coli PepN was re-categorized as a metalloaminopeptidase (15) and assumed to bind Zn(II), although no metal analyses were ever reported. The inhibition by classical metalloaminopeptidase inhibitors such as bestatin and amastatin was consistent with the reclassification of E. coli PepN as a metalloaminopeptidase (16-18). However, recent studies on recombinant E. coli PepN have demonstrated that the enzyme does not bind significant amounts of Zn(II) and that Zn(II) is in fact inhibitory (Chapter 3). Metal analyses demonstrated that E. coli PepN tightly binds 5 equivalents of iron per protein; however, the metal ions were not required for catalytic activity (Chapter 3). This surprising
57 result prompted a study of the role of iron in the structure of E. coli PepN. In this chapter, fluorescence emission, UV-Vis, 1H NMR, and EPR spectroscopic studies were conducted to determine if the addition of iron affects the structure of the enzyme and to characterize the iron binding sites in the enzyme.
Materials and Methods E. coli strain BL21(DE3) was obtained from Novagen. Luria-Bertani (LB) media was made following published procedures (19). Isopropyl-β-D-thiogalactoside (IPTG), Biotech grade, was procured from Anatrace. Protein solutions were concentrated with an Amicon ultrafiltration cell equipped with YM-10 DIAFLO membranes from Amicon, Inc. Dialysis tubing was prepared using Spectra/Por regenerated cellulose molecular porous membranes with a molecular weight cut-off of 6-8,000 g/mol. Q-Sepharose Fast Flow was purchased from Amersham Pharmacia Biotech. L-alanine-p-nitroanilide was purchased from Sigma. Ferrous ammonium sulfate was purchased from Aldrich, and ferrous sulfate (99.999%) was purchased
from American Elements. Sodium hydrosulfite (Na2S2O4) was purchased from Aldrich. All buffers and media were prepared using Barnstead NANOpure ultrapure water (18 MΩ cm-1).
Preparation of PepN. Large-scale (4 L) preparations of recombinant E. coli PepN were prepared as described in Chapter 3. Briefly, a 10 mL overnight culture of these cells in LB medium was used to inoculate 4 x 1 L of LB medium containing 25 mg/mL kanamycin. The cells were allowed to grow at 37 °C with shaking until the cells reached an optical density at 600 nm of 0.6- 0.8. Protein production was induced by making the cultures 1 mM in isopropyl-β-D- thiogalactopyranoside (IPTG), and the cells were shaken at 37 °C for 3 h. The cells were collected by centrifugation (15 min at 8,200 x g) and resuspended in 15 mL of 50 mM Tris, pH 8.5. The resuspended cells were lysed by sonication using a Fisher Scientific model 100 sonic dismembrator. Cell debris was removed from the sample by centrifugation (15 minutes at 23,400 x g), and the crude protein solution was dialyzed overnight at 4 °C versus 2 L of 50 mM Tris- HCl, pH 8.5. The dialyzed, crude protein solution was then centrifuged for 15 minutes at 23,400 x g to remove any precipitated proteins or lipids, and the cleared supernatant was loaded onto a 16 x 20 mm Q-Sepharose column that had been previously equilibrated with 300 mL of 50 mM Tris-HCl, pH 8.5. Bound proteins were eluted with a 0 to 500 mM NaCl gradient in 50 mM Tris,
58 pH 8.5, at 2 mL/min. Fractions (8 mL) containing PepN were pooled and concentrated with an Amicon ultrafiltration cell equipped with a YM-10 membrane. Protein purity was ascertained by sodium dodecyl sulfate (SDS) polyacrylamide gel electrophoresis. The concentration of PepN was determined by measuring the protein’s absorbance at 280 nm and using the extinction -1 -1 coefficient of ε280nm = 147,595 M •cm . The metal content of PepN samples was determined as described in Chapter 3.
Preparation of Apo-PepN and Fe-containing PepN. Apo-PepN was prepared by three dialysis steps of recombinant PepN against a 300-fold excess of tris-(hydroxymethyl)aminomethane (Tris) buffer at pH 8.5, containing 10 mM EDTA (12 h each dialysis at 4 °C). The EDTA was removed completely by three dialysis steps against the same buffer containing 150 mM NaCl, followed by passage through a 1.5 x 68 cm column of Sephadex G-25 (bed volume 120 mL), equilibrated with 50 mM Tris pH 8.5, containing 100 mM NaCl (20). The column flow rate was adjusted to 1 mL/min, and 6 mL fractions were collected and monitored by absorbance at 280 nm. Column fractions containing PepN were pooled and concentrated using ultrafiltration, as described above. ICP-AES was used to demonstrate that metal had been removed from the enzyme samples. To prepare Fe-containing PepN, 1-10 molar equivalents of ferrous ammonium sulfate,
Fe(NH4)2(SO4)2, were added directly to solutions of apo-PepN or to as-isolated PepN. The samples were incubated on ice for 1 hour and then used in spectroscopic studies.
Steady-state kinetics. Steady-state kinetic studies were carried out in 50 mM Tris buffer, pH 8.5, on a Hewlett-Packard 5480A UV / Visible spectrophotometer, using an Isotemp circulator to maintain the reactions at 37 ºC. Steady-state kinetic assays were performed using L-alanine-p- nitroanilide substate in a concentration range of 50 μM – 8 mM. The PepN concentration for the
kinetic assays was maintained at 1 nM. Steady-state kinetics constants, Km and kcat, were determined by fitting initial velocity versus substrate concentration data directly to the Michaelis equation using CurveFit v.1.0 (21). The errors reported are generated by CurveFit as a result of Chi square minimization. All steady-state kinetic studies were performed in triplicate.
59 Circular dichroism spectroscopy. Circular dichroism samples were prepared by dialyzing the purified enzyme samples versus 3 x 2 L of 5 mM phosphate buffer, pH 7.0, over 6 h. The samples were diluted with 5 mM phosphate buffer, pH 7.0, to a final concentration of approximately 75 μg/mL. A JASCO J-810 CD spectropolarimeter operating at 25 ºC was used to obtain CD spectra.
Fluorescence emission spectroscopy. Fluorescence emission spectra were obtained on a Perkin- Elmer LS 55 luminescence spectrometer using 2.5 μM PepN in 50 mM Tris, pH 8.5, at 25 oC. The excitation wavelength was 284 nm, and the emission spectra from 300 – 400 nm were
obtained. Aliquots of FeSO4 were added to as-isolated PepN to generate samples containing 1, 2, 3, 4, 5, 6, and 10 equivalents of Fe(II).
Electronic absorption spectroscopy. Electronic absorption spectra were obtained on a Hewlett- Packard 5480A spectrophotometer at 37 oC. Samples were diluted in 50 mM Tris, pH 8.5.
NMR spectroscopy. NMR spectra were collected on a Bruker Avance 500 NMR spectrometer operating at 500.13 MHz, 298 K, and a magnetic field of 11.7 T. The spectra were obtained by using a modified presaturation pulse sequence (zgpr) for water suppression and the following parameters: recycle delay (AQ), 41 ms; sweep width, 400 ppm; receiver gain, 128; line broadening, 80 Hz. The samples for NMR studies were made by concentrating apo-, as-isolated, and Fe(II)-added PepN using an Amicon equipped with a YM-10 membrane or a Centricon-10 to final concentrations of ca. 1 mM in buffer containing approximately 10 % D2O. The sample volumes were approximately 0.4 mL. Samples in 90 % D2O were prepared by concentrating the protein to minimal volume using an Centricon-10 and diluting with 50 mM Tris, pH 8.5, prepared with 100 % deuterium oxide. This step was repeated three times, and the sample was
finally diluted in sufficient 100 % D2O buffer to provide a final PepN concentration of ~ 1 mM. Samples were loaded into Wilmad 5-mm tubes for NMR studies. Protein chemical shifts were
calibrated by assigning the H2O signal the value of 4.70 ppm at 298 K.
EPR spectroscopy. EPR spectra were recorded at the National Center for Biomedical EPR, Medical College of Wisconsin with the assistance of Professor Brian Bennett, G. Periyannan,
60 and A. Kumar, using a Bruker EleXsys E500 EPR spectrometer equipped with an ER-4116DM dual mode TE102/TE012 resonant cavity, operating at 9.63 GHz in perpendicular mode and 9.37 GHz in parallel mode. An Oxford Instruments ESR900 helium flow cryostat and an Oxford Instruments ITC502 temperature controller facilitated sample temperature variation from 3.6 - 80 K. EPR samples were made by pipetting 400 μL of ca. 1 mM PepN into a 4 mm o.d. quartz EPR tube. Oxidized samples were prepared by the addition of 1-5 equivalents of H2O2 directly to the PepN samples. Reduced samples were prepared by addition of dithionite crystals to the EPR sample.
Results and Discussion Effect of Fe binding. Previous studies on E. coli PepN have yielded contradictory results regarding the requirement of metal ions for activity. Initially categorized as a cysteinyl aminopeptidase that did not require the presence of metal ions for activity (13), E. coli PepN was later re-categorized as a Zn(II)-containing aminopeptidase based on amino acid sequence similarities to known Zn(II)-containing metalloproteins (10). Our studies on recombinant E. coli PepN demonstrate that this enzyme is not a Zn(II)-containing aminopeptidase (Chapter 3) and suggests that the enzyme is indeed a cysteinyl aminopeptidase (Chapter 5), in agreement with early reports (13). Our studies further demonstrated that recombinant E. coli PepN tightly binds 5 equivalents of iron, and surprisingly, that these metal ions are not required for catalysis. Specifically, the as-isolated enzyme, which contains 0.5 equivalents of iron and < 0.1 equivalents -1 of Zn(II), exhibited a kcat of 354 s , while PepN containing 5 equivalents of Fe exhibited a kcat of 236 s-1. We speculated that the drop in activity may be due to a change in the structure of PepN upon metal ion binding. This unusually large number of iron equivalents further sparked our interest in this study and lead us to speculate that there may be unique, multinuclear iron centers in PepN. Recombinant E. coli PepN was over-expressed and purified as previously described (Chapter 3), and the as-isolated enzyme was shown to bind <0.1 equivalents of Zn(II) and 0.5 equivalents of iron. Our first approach to probe for any structural change to PepN upon binding Fe was to obtain CD spectra of the enzyme as we titrated the as-isolated enzyme with Fe (Figure 1). Our results clearly showed no significant changes in the shapes or intensities of the CD
61 spectra of PepN containing up to 10 equivalents of Fe. This result indicates that there are no changes in the secondary structure of PepN with binding of Fe. The second approach used to probe for structural changes to PepN upon binding of Fe involved the use of fluorescence emission spectroscopy. The most important, intrinsic fluorophores in proteins are tryptophan, tyrosine, and phenylalanine residues, with tryptophan being most important (22). Small changes in the local environments of tryptophan residues can often result in large changes in the fluorescence of a protein (23). In fact, tryptophan fluorescence has been used to probe the tertiary structure of proteins (24). Therefore, fluorescence emission spectroscopy using an excitation wavelength of 284 nm was used to probe for tertiary structural changes in PepN upon addition of Fe. The amino acid sequence of PepN demonstrates that there are 13 tryptophans in the enzyme (14). The fluorescence emission spectra of as-isolated PepN and PepN containing 1, 2, 3, 4, 5, and 10 equivalents of Fe(II) are shown in Figure 2. There was a slight drop in emission intensity upon the addition of the first equivalent of Fe(II) possibly due to dilutional effects; however, there were no further changes with additional Fe(II). Furthermore, there does not appear to be any evidence of a red-shift in the emission maximum, which would suggest unfolding of the protein (25). These data strongly suggest that there is no significant change in the tertiary structure of PepN with added Fe. However, we cannot unequivocally rule out the possibility that small changes in tertiary structure occurred in portions of the enzyme where there were no tryptophans present. The studies above indicate that Fe does not play a structural role in E. coli PepN; however, the relatively large number of equivalents of Fe that bind to PepN suggest the presence of a multinuclear Fe center in the enzyme. Although there have been numerous spectroscopic studies published on Fe-containing proteins and model compounds, Fe has a number of characteristics that complicates interpretation of data gleaned from these studies. For example, there are two common oxidation states for Fe in biological molecules (Fe(II) and Fe(III)); however, there are numerous examples of Fe(IV) and possibly Fe(V) centers (26). Fe(III) has a valence shell electron configuration of 3d5 and can exist in either high- (S = 5/2) or low-spin (S = ½) states depending on the ligand field and geometry. Fe(II) has a valence shell electron configuration of 3d6 and can exist in either high- (S = 2) or low- (S = 0) spin states. Fortunately, the spin state of Fe in samples is largely governed by ligand field strengths. All amino acids that coordinate metal ions are weak-field ligands (27), resulting in high-spin Fe centers. Hemes (Fe-
62 porphyrins) are the only common Fe centers in biological molecules that are low-spin (27). Fe(III) and Fe(II) have been shown to coordinate a wide variety of biologically-relevant ligands, - such as His, Asp, Glu, Tyr, Asn, Arg, Gln, CO, CN , O2, porphyrins, etc., and exist in a number of coordination geometries (27). This flexibility in coordination environments leads to widely varying zero-field splittings (ZFS) and a wide range of values for E/D (28, 29), resulting in EPR spectra that are difficult to interpret. The presence of spin-coupled, multinuclear Fe-containing centers in proteins, such as FeS and oxo-bridged, dinuclear Fe centers, further complicates spectroscopic characterization of unknown Fe centers in proteins. In spite of these complicating factors, UV-Vis, EPR, and 1H NMR spectroscopic studies were used to probe the Fe binding sites in E. coli PepN.
UV-Vis spectrophotometry. The three most common features observed in the UV-Vis spectra of iron-containing proteins are π−π* Soret transitions, ligand field transitions (d-d bands), and ligand to metal charge transfer bands (LMCT). These transitions can be distinquished in most cases by the observed molar absorptivities of the bands: Soret bands typically have ε’s > 10,000 M-1cm-1, LMCT normally have ε’s > 1,000 M-1cm-1, and d-d bands normally have ε’s < 500 M- 1cm-1. Soret bands are only observed when there are heme groups in the protein. High-spin 5 Fe(II) has a T2 ground energy state; therefore, there is one possible “spin-allowed” transition, and d-d bands for any Fe(II) centers are expected. Unfortunately, the energy gap between the ground and excited energy levels results in d-d bands for octahedral, high-spin Fe(II) that exist in the near-IR region of the electromagnetic spectrum (30). In contrast, high-spin Fe(III) normally 6 has a A1 ground state energy term, and no “spin-allowed” transitions are expected. The presence of a d-d band in the UV-Vis/near IR spectrum of PepN could determine whether there is Fe(II) or Fe(III) in the protein. Both Fe(II) and Fe(III) centers can exhibit LMCT’s if there are coordinated Tyr or Cys ligands (30). The UV-Vis spectra of PepN titrated with Fe(II) are shown in Figure 3. The spectrum of as-isolated PepN (0.5 equivalents of Fe) consists of a relatively sharp feature at 350 nm, a broad, low intensity feature at 435 nm, and possibly a very low intensity feature at 560 nm. The addition of 0.5 equivalents of Fe(II) to as-isolated PepN results in a slight increase in the intensity of all three bands (Figure 3). As more equivalents of Fe(II) are added, there are slight increases in the intensities of all three features and a red shift of the peak at 350 nm to 382 nm.
63 There are no peaks in these spectra that are intense enough to be labeled as Soret bands, indicating that there are no heme centers in PepN. The peak at 350 nm has a molar absorptivity of 1,600 M-1cm-1 per Fe and thus can be attributed to a LMCT band. Similar peaks have been reported for rubredoxin and ferredoxin:thioredoxin reductase (FTR) (31, 32). These proteins contain a mononuclear FeS center (4 cys), suggesting that Fe is binding to a site that contains at least 1 cysteine group. The red-shift in the peak position upon addition of Fe(II) suggests that there is a change in the oxidation state of the Fe during the titration. Similar shifts have been observed previously with rubredoxin and FTR as the oxidation state of Fe changed; however, red-shifting does not automatically correlate with Fe oxidation (31-34). The spectra of PepN were obtained from 200 – 1100 nm; however, there was no evidence of any other bands, such as d-d bands. The absence of a d-d band in these spectra does not necessarily mean that the sample contains Fe(III), since the peak position for a d-d band from Fe(II) could appear at wavelengths ranging from 200 to >1100 nm depending upon the ligands and coordination geometry (30).
EPR spectroscopy. EPR spectroscopy is a powerful tool for characterizing the electronic properties of paramagnetic centers in proteins. EPR spectra have been reported for mononuclear Fe(III) centers, numerous FeS clusters, and oxo-bridged dinuclear Fe centers (35). As discussed above, the Fe in PepN is most likely high-spin and is either Fe(II) or Fe(III). High-spin Fe(III) has five unpaired electrons (S = 5/2) and is expected to yield an observable EPR spectrum using a typical perpendicular-mode, X-band instrument. On the other hand, high-spin Fe(II) has four unpaired electron resulting in an integer spin state of S = 2. Integer spin systems do not yield EPR spectra in perpendicular-mode instruments. Therefore, we hypothesized that we would be able to discern the oxidation state of Fe in PepN by using EPR spectroscopy. The perpendicular-mode, X-band EPR spectra of as-isolated PepN (0.5 equivalents of Fe) at 5 K and 15 K are shown in Figure 4A and 4B. The signals at g ~ 9 and g = 4.1 - 4.3 are consistent with the presence of high-spin Fe(III) in the sample. The loss of the g ~ 9 signal as
the temperature is raised to 15 K (Figure 4B) suggests strongly that the Ms = |+/- 1/2> doublet is
the ground state and that the Ms = |+/-3/2> and Ms = |+/-5/2> doublets are excited states (36). The signals associated with high-spin Fe(III) were not spin-integrated because problems with obtaining reliable integrations for Fe(III) (B. Bennett, personal communication); however, relative amounts of Fe(II) and Fe(III) in as-isolated PepN can be estimated (see below). It is
64 very likely that the g = 4.1 – 4.3 signal is due to protein-bound Fe(III) and adventitiously-bound (junk) Fe(III) (sharp signal at g = 4.1). Two other, unknown features were observed in the spectra of as-isolated PepN: a low intensity, broad g ~ 2.3 signal and a very sharp g = 2 signal. Given the position of the g ~ 2.3 signal, we cannot attribute this signal to the presence of an antiferromagnetically-coupled, mixed valent, oxo-bridged (or FeS) dinuclear Fe center. It is possible that this signal arises from a small amount of contaminating Cu(II) either in the sample or in the EPR cavity. The g = 2 signal probably arises from a radical; however, we are uncertain whether this radical is enzyme-based or whether an oxygen radical was produced during the freezing of the sample. In order to further probe the Fe in as-isolated PepN, EPR spectra were obtained on samples containing 1 and 5 equivalents of dithionite. We reasoned that the addition of this reductant would result in all of the Fe(III) to be reduced to Fe(II) and the complete loss of an EPR signal. However, the intensity of the EPR spectrum of as-isolated PepN containing 1 equivalent of dithionite is the same as that of as-isolated enzyme (Figure 4A and 4C), and there was only a small drop in intensity of the EPR signal of the sample containing 5 equivalents of dithionite (Figure 4D). This result suggests that the Fe(III) giving rise to the g = 4.1 - 4.3 signals is not readily redox active, possibly because it is solvent-inaccessible. Spectra of as-isolated PepN containing 5 equivalents of dithionite were obtained at 20, 60, and 70K (Figure 4E, F, and G). A very unusual signal emerged in these spectra that shifted upon the increase in temperature. We are unable to interpret these temperature-dependent spectra at the present time. It is possible
that with higher temperatures we have thermally-populated excited Ms manifolds; however, we cannot rule out the possibility of some spin-coupling between Fe ions or with an enzyme- associated spin-radical. The corollary experiment to that with dithionite is to add hydrogen peroxide to as- isolated PepN. We reasoned that any solvent-accessible Fe(II), which is EPR-silent in our EPR studies, would be oxidized to Fe(III), which gives rise to a readily observable EPR signal. The EPR spectrum of as-isolated PepN containing 1 equivalent of hydrogen peroxide is shown in Figure 4H. This spectrum was reduced by a factor of two so that all of the signals in Figure 4 would be of comparable intensities. The spectrum of as-isolated PepN containing 1 equivalent of hydrogen peroxide shows very intense g = 4.1 – 4.3 signals, especially in the region of the g = 4.3 signal. The signals corresponding the g = 4.3 and g ~ 9 are due to protein-bound Fe(III) as
65 these signals have narrow zero-field splitting distributions as compared to those of adventitious Fe(III) (B. Bennett, personal communication). The small signals at g = 2.3 and 2 are nearly unobservable in Figure 4H due to the relatively large increase in intensity of the Fe(III) signals. It is also possible that hydrogen peroxide may have oxidized the species giving rise to these features. The addition of 5 equivalents of hydrogen peroxide to as-isolated PepN resulted in a further increase in intensity of the Fe(III)-associated features (Figure 4I), particularly of the g = 4.1 signal, which is probably due to adventitious Fe(III). It is possible that the excess hydrogen peroxide caused some protein degradation and the loss of protein-bound Fe(III). Interestingly, the signals at g = 2.3 and 2 are again observable in this sample (Figure 4I). These data demonstrate that the majority of the protein-bound Fe in as-isolated PepN is high-spin Fe(II). EPR spectra were also obtained for PepN containing 5 equivalents of Fe (Figure 5). The spectrum of PepN containing 5 equivalents of Fe is shown in Figure 5A. This spectrum shows typical g = 4.1 – 4.3 and g ~ 9 signals that are attributable to high-spin Fe(III). There is also a significant multi-line signal centered at g = 2, which is due to high-spin Mn(II). The Mn(II) was introduced into the sample from our stock of ferrous iron. The addition of 1 and 5 equivalents of dithionite to PepN containing 5 equivalents of Fe resulted in EPR spectra that were similar in appearance to those of as-isolated PepN (Figure 5B and 5C). The addition of 1 equivalent of dithionite resulted in little change in the EPR spectrum of PepN containing 5 equivalents of Fe (Figure 5B). However as observed with the spectrum of as-isolated PepN, there was a marked change in the appearance of the spectrum of PepN containing 5 equivalents of Fe upon addition of 5 equivalents of dithionite (Figure 5C). The broadened features in this spectrum (Figure 5C) are similar to those observed on the sample of as-isolated PepN obtained at 5 K (Figure 4D). The EPR signals attributed to Mn(II) were unaffected by dithionite, and in the spectrum of PepN containing 5 equivalents of Fe and 5 equivalents of dithionite, there is an intense, sharp signal at g = 2 that is probably due to a radical (Figure 5C). It is known that dithionite reacts with oxygen in aqueous solutions (37), and as these solutions were prepared aerobically, the presence oxygen could not be entirely avoided. The spectrum of PepN containing 5 equivalents of Fe and 1 equivalent of hydrogen peroxide shows an increase in the intensity of the g ~ 9 signal, suggesting, as before with as-isolated PepN, that a significant portion of the Fe in this sample is Fe(II) (Figure 5D).
66 1H-NMR spectroscopy. Paramagnetic 1H NMR spectroscopy is technique that has been used for over 30 years to study the structure of metal binding sites in proteins. Paramagnetic metal ions, such as Co(II), Ni(II), or Fe(II), when bound to a protein, can induce isotropically-shifted NMR resonances of any protons near the paramagnet (38). The magnitude of the shift depends on the type of interaction between the unpaired electrons of the paramagnetic metal ion and the protons. Larger shifts are most often caused by a contact mechanism, which is a through-bond interaction. The second mechanism involves through-space interactions (39). In addition to the isotropic shift of resonance positions, line broadening of proton peaks is observed due to the presence of paramagnetic species. NMR resonances are broadened due to two major factors: the rotational
correlation time and the T1e of the paramagnet (39). The rotational correlation time involves how fast the protein rotates in the NMR tube. Large proteins (molecular masses of > 50 kDa) have relatively slow rotational correlation times and yield very broad NMR lines. The T1e is the electron relaxation rate for the unpaired electron in the paramagnet. Samples containing metal -10 -11 1 ions with T1e’s faster than 10 -10 seconds exhibit detectable, isotropically-shifted H NMR -10 resonances. Samples containing metal ions with T1e’s slower than 10 seconds yield very broad NMR resonances that are often too broad for detection (39). 1 The paramagnetic H NMR spectrum of as-isolated PepN in 10% D2O is shown in Figure 6. Two very broad resonances at 70 and 90 ppm were observed, and both of these signals
disappear when the sample is in 100% D2O or when there is no metal bound to the enzyme (Figure 6A, B, and E). Given the solvent-substitution and resonance positions, these peaks can - be assigned to NH protons on Fe(II)-bound histidines (40-43). High-spin Fe(II) has a T1e of 10 11 -12 -10 -11 -10 seconds, while high-spin Fe(III) has a T1e of 10 – 10 seconds. Although high-spin
Fe(III) exhibits a T1e that is borderline for inducing observable, paramagnetically-shifted resonances, the large rotational correlation time associated with PepN and the borderline T1e most likely results in proton resonances shifted by high-spin Fe(III) to be too broad for detection (41, 42). In fact, there have been no reports of paramagnetically-shifted resonances observed for high-spin Fe(III)-containing proteins (L. Que, personal communication). There have been many reports of paramagnetically-shifted resonances for proteins containing coupled Fe(III)Fe(II) centers (oxo- or sulfido-bridged) (44); however, our EPR spectra clearly show that there are no such centers present in as-isolated PepN. The EPR studies do strongly suggest the presence of high-spin Fe(II), which would be expected to yield observable, paramagnetically-shifted NMR
67 resonances. Samples containing high-spin Fe(II) normally exhibit narrow NMR lines; however, the large molecular weight of PepN probably contributes to the broad NMR resonances seen in Figure 6. To address whether any of the Fe bound to the 2 histidine site was Fe(III), we added one equivalent of dithionite to as-isolated PepN and collected the 1H NMR spectrum (Figure 6). We reasoned that if there was any Fe(III) present, the addition of dithionite would result in the reduction of Fe(III) to Fe(II) and more intense histidine NH proton resonances. Surprisingly, the NMR spectrum of as-isolated PepN containing 1 equivalent of dithionite shows a marked decrease in the intensities of the two peaks at 70 and 90 ppm, suggesting that the addition of dithionite causes a slight denaturation of the protein. In the EPR section, we argued that the reduction in EPR signal intensity of the as-isolated PepN samples containing dithionite was due to a reduction of Fe(III) in the enzyme (Figure 4D). However, it is possible that the reduced EPR intensities were due to protein denaturation. It is important to note that the NMR and EPR results strongly suggest that the predominant metal ion in as-isolated PepN is high-spin Fe(II). In an effort to probe the other Fe binding sites in PepN, an NMR spectrum was obtained for PepN containing 5 equivalents of Fe (Fe(II) was the added metal). The only two peaks observed were the same resonances observed with as-isolated PepN. The intensities of the peaks at 70 and 90 ppm were much lower than those in as-isolated PepN. We attribute this reduced intensity to oxidation of the Fe(II) at the histidine site upon addition of Fe(II). The resulting Fe(III) in the histidine site would not yield sufficiently-narrow NMR resonances for detection. The oxidation of Fe(II) to Fe(III) is supported by our previous UV-Vis titrations, which showed a red shift in the LMCT band upon addition of Fe(II) (Figure 3). In addition, the EPR studies showed very intense g = 4.1 – 4.3 signals for the PepN samples containing 5 equivalents of Fe, which indicates the presence of significant amounts of high-spin Fe(III).
Conclusions The discovery that E. coli PepN tightly binds 5 equivalents of Fe sparked our interest in understanding the structure of metal binding sites in the enzyme. However, our subsequent results that showed that the Fe ions are not required for catalysis or structure, which initially dampened our enthusiasm to characterize the metal centers. Recently however, Liu et al. published a crystal structure for PhoU (45). It was shown that PhoU contains two novel, multi-
68 iron clusters; one containing three and a second containing four iron atoms. The Fe ions are coordinated extensively by Asp and Glu residues, which serve as terminally-bound and bridging ligands. There is no known function for these unique metal-binding clusters. Since E. coli PepN tightly binds 5 equivalents of Fe, we speculated that PepN may have a similar cluster to one of those reported in PhoU. The metal binding ligands in PhoU are primarily contained in a repeating E(D)XXXD motif (45). Examination of the amino acid sequence of E. coli PepN revealed the existence of six E(D)XXXD motifs. We are collaborating currently with Professor John Peters at Montana State University to determine the crystal structure of E. coli PepN. Concurrently, we conducted several spectroscopic studies to probe the metal binding site(s) in PepN. The detection of Asp/Glu metal binding ligands is very difficult with most spectroscopic techniques. It is possible under some 1 circumstances to observe β-CH2 protons of metal bound Asp/Glu residues with paramagnetic H NMR spectroscopy as these protons typically yield resonances at 15 – 30 ppm. In our NMR studies however, water suppression was not sufficient to detect peaks in this region. Our spectroscopic studies did reveal a number of important results. Firstly, our data strongly indicates that the metal ion found in as-isolated PepN is high-spin Fe(II). UV-Vis and 1H NMR studies suggest that this Fe(II) is bound in a site containing two histidines and at least 1 cysteine as ligands (Figure 7A). There numerous reports in the literature of 1Fe-0S centers (1 Fe and 4 Cys as ligands) found in proteins such as rubredoxin, and these centers have been extensively characterized (31, 32). Several FeS centers have also been reported to contain mixed Cys/His ligand groups, such as desulfoferrodoxin, which contains a 1Fe-0S center (4 Cys) and a 1Fe-0S center (1 Cys and 4 His) (31). The remaining Fe’s that bind to PepN may indeed form a cluster similar to that observed in PhoU (Figure 7B). Preliminary EPR studies on PhoU demonstrate that the clusters in this protein do not yield a perpendicular-mode EPR signal showing spin-coupling (expected g values < 2) (F. Golich, M.W. Crowder, B. Bennett, unpublished results). The identities of the other Fe binding sites in PepN await the crystal structure.
69 References 1. Taylor, A. (1996) Aminopeptidases, R.G. Landes Company, Austin, TX. 2. Taylor, A. Aminopeptidases: Structure and function (1993) FASEB J 7, 290-8. 3. Arora, G., and Lee, B. H. Comparative studies on peptidases of Lactobacillus casei subspecies (1990) J Dairy Sci 73, 274-279. 4. Tan, P. S., van Alen-Boerrigter, I. J., Poolman, B., Siezen, R. J., de Vos, W. M., and Konings, W. N. Characterization of the Lactococcus lactis pepN gene encoding an aminopeptidase homologous to mammalian aminopeptidase N (1992) FEBS Lett 306, 9-16. 5. van Alen-Boerrigter, I. J., Baankreis, R., and de Vos, W. M. Characterization and overexpression of the Lactococcus lactis pepN gene and localization of its product, aminopeptidase N (1991) Appl Environ Microbiol 57, 2555-61. 6. Chandu, D., and Nandi, D. PepN is the major aminopeptidase in Escherichia coli: Insights on substrate specificity and role during sodium-salicylate-induced stress (2003) Microbiol- Sgm 149, 3437-3447. 7. Li, Q., and Xu, W. Novel anticancer targets and drug discovery in post genomic age (2005) Curr Med Chem Anti-Canc Agents 5, 53-63. 8. Xu, W., and Li, Q. Progress in the development of aminopeptidase N (APN/CD13) inhibitors (2005) Curr Med Chem Anti-Canc Agents 5, 281-301. 9. Shim, J. S., Kim, J. H., Cho, H. Y., Yum, Y. N., Kim, S. H., Park, H. J., Shim, B. S., Choi, S. H., and Kwon, H. J. Irreversible inhibition of CD13/aminopeptidase N by the antiangiogenic agent curcumin (2003) Chem Biol 10, 695-704. 10. Gonzales, T., and Robert-Baudouy, J. Bacterial aminopeptidases: Properties and functions (1996) FEMS Microbiol Lett 18, 319-344. 11. McCaman, M. T., and Villarego, M. R. Structural and catalytic properties of peptidase N from Escherichia coli K-12 (1982) Arch Biochem Biophys 213, 384-394. 12. Chandu, D., Kumar, A., and Nandi, D. PepN, the major suc-LLVY-AMC-hydrolyzing enzyme in Escherichia coli, displays functional similarity with downstream processing enzymes in archaea and eukarya. Implications in cytosolic protein degradation (2003) J Biol Chem 278, 5548-56.
70 13. Chappelet-Tordo, D., Lazdunski, C., Murgier, M., and Lazdunski, A. Aminopeptidase N from Escherichia coli: Ionizable active-center groups and substrate specificity (1977) Eur J Biochem 81, 299-305. 14. Foglino, M., Gharbi, S., and Lazdunski, A. Nucleotide-sequence of the pepN gene encoding aminopeptidase-N of Escherichia-coli (1986) Gene 49, 303-309. 15. Bradshaw, R. A. Aminopeptidases (2004) Encyclopedia Biol Chem 1, 96-98. 16. Kim, H., and Lipscomb, W. N. X-ray crystallographic determination of the structure of bovine lens leucine aminopeptidase complexed with amastatin: Formulation of a catalytic mechanism featuring a gem-diolate transition state (1993) Biochemistry 32, 8465-78. 17. Luciani, N., Marie-Claire, C., Ruffet, E., Beaumont, A., Roques, B. P., and Fournie-Zaluski, M.-C. Characterization of Glu350 as a critical residue involved in the N-terminal amine binding site of aminopeptidase N (ec 3.4.11.2): Insights into its mechanism of action (1998) Biochemistry 37, 686-692. 18. Taylor, A. Aminopeptidases: Towards a mechanism of action (1993) Trends Biochem Sci 18, 167-172. 19. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular cloning - a laboratory manual, Vol. 1, Second ed., Cold Spring Harbor Laboratory Press. 20. Yang, K. W., and Crowder, M. W. Method for removing EDTA from apo-proteins (2004) Anal Biochem 329, 342-344. 21. Crowder, M. W., Walsh, T. R., Banovic, L., Pettit, M., and Spencer, J. Overexpression, purification, and characterization of the cloned metallo-b-lactamase L1 from Stenotrophomonas maltophilia (1998) Antimicrob Agents Chemother 42, 921-926. 22. Garrity, J. D., Pauff, J. M., and Crowder, M. W. Probing the dynamics of a mobile loop above the active site of L1, a metallo-β-lactamase from Stenotrophomonas maltophilia, via site-directed mutagenesis and stopped-flow fluorescence spectroscopy (2004) J Biol Chem 279, 39663-39670. 23. Campbell, I. D., and Dwek, R. A. (1984) Biological spectroscopy, Benjamin Cummings, Menlo Park. 24. Zang, T. M., Hollman, D. A., Crawford, P. A., Crowder, M. W., and Makaroff, C. A. Arabidopsis glyoxalase II contains a zinc/iron binuclear metal center that is essential for substrate binding and catalysis (2001) J Biol Chem 276, 4788-4795.
71 25. Periyannan, G., Shaw, P. J., Sigdel, T., and Crowder, M. W. In vivo folding of recombinant metallo-β-lactamase L1 requires the presence of Zn(II) (2004) Prot Sci 13, 2236-2243. 26. Valentine, A. M., Tavares, P., Pereira, A. S., Davydov, R., Krebs, C., Hoffman, B. M., Edmondson, D. E., Huynh, B. H., and Lippard, S. J. Generation of a mixed-valent Fe(III)Fe(IV) form of intermediate Q in the reaction cycle of soluble methane monooxygenase, an analog of intermediate X in ribonucleotide reductase R2 assembly (1998) J Am Chem Soc 120, 2190-2191. 27. Lippard, S. J., and Berg, J. M. (1994) Principles of Bioinorganic Chemistry, University Science Books, Mill Valley, CA. 28. Hendrich, M. P., and Debrunner, P. G. Integer spin electron paramagnetic resonance of iron proteins (1989) Biophys J 56, 489-506. 29. Hagen, W. R. EPR of non-Kramers doublets in biological systems. Characterization of an S=2 system in oxidized cytochrome c oxidase (1982) Biochim Biophys Acta 708, 82-98. 30. Solomon, E. I., Decker, A., and Lehnert, N. Non-heme iron enzymes: Contrasts to heme catalysis (2003) Proc Natl Acad Sci USA 100, 3589-3594. 31. Staples, C. R., Ameyibor, E., Fu, W., Gardet-Salvi, L., Stritt-Etter, A. L., Schurmann, P., Knaff, D. B., and Johnson, M. K. The function and properties of the iron-sulfur center in spinach ferredoxin:thioredoxin reductase: A new biological role for iron sulfur clusters (1996) Biochemistry 35, 11425-11434. 32. Lovenberg, W., and Sobel, B. E. Rubredoxin: A new electron transfer protein from Clostridium pasteurianum (1965) Proc Natl Acad Sci U S A 54, 193-9. 33. Chasteen, N. D., Grady, J. K., Skorey, K. I., Neden, K. J., Riendeau, D., and Percival, M. D. Characterization of the non-heme iron center of human 5-lipoxygenase by electron paramagnetic resonance, fluorescence, and ultraviolet-visible spectroscopy: Redox cycling between ferrous and ferric states (1993) Biochemistry 32, 9763-71. 34. Lombardi, A., Marasco, D., Maglio, O., Di Costanzo, L., Nastri, F., and Pavone, V. Miniaturized metalloproteins: Application to iron-sulfur proteins (2000) Proc Natl Acad Sci U S A 97, 11922-7. 35. Bertini, I., Gray, H. B., Lippard, S. J., and Valentine, A. M. (1994) Bioinorganic Chemistry, University Science Books, Mill Valley, CA.
72 36. Copik, A. J., Waterson, S., Swierczek, S. I., Bennett, B., and Holz, R. C. Both nucleophile and substrate bind to the catalytic Fe(II)-center in the type II methionyl aminopeptidase from Pyrococcus furiosus (2005) Inorg Chem 44, 1160-1162. 37. Morello, J. A., Craw, M. R., Contantine, H. P., and Forster, R. E. Rate of reaction of dithionite ion with oxgyen in aqueous solution (1964) J Appl Physiol 19, 522-525. 38. Bertini, I., Banci, L., and Luchinat, C. (1989) in Meth Enzymol pp 246-263, Academic Press, New York. 39. Drago, R. S. (1992) Physical Methods for Chemists, 2nd ed., WB Saunders, San Deigo, CA. 40. Bertini, I., Luchinat, C., and Rosato, A. The solution structure of paramagnetic metalloproteins (1996) Prog Biophys Mol Biol 66, 43-80. 41. Borovik, A. S., Papaefthymiou, V., Taylor, L. F., Anderson, O. P., and Que, L. Models for iron-oxo proteins. Structures and properties of FeIIFeIII, ZnIIFeIII, and FeIIGaIII complexes with (μ-phenoxo)bis(μ-carboxylato)dimetal cores (1989) J Am Chem Soc 111, 6183-6195. 42. Lauffer, R. B., Antanaitis, B. C., Aisen, P., and Que, L. 1H NMR studies of porcine uteroferrin (1983) J Biol Chem 258, 14213-14218. 43. Wang, Z., Ming, L. J., Que, L., Vincent, J. B., Crowder, M. W., and Averill, B. A. 1H NMR and NOE studies of the purple acid phosphatases from porcine uterus and bovine spleen (1992) Biochemistry 31, 5263-5268. 44. Bertini, I., and Luchinat, C. (1986) NMR of Paramagnetic Molecules in Biological Systems, Benjamin Cummings Publishing Company, Menlo Park, CA. 45. Liu, J., Lou, Y., Yokota, H., Adams, P. D., Kim, R., and Kim, S. H. Crystal structure of a PhoU protein homologue: A new class of metalloprotein containing multinuclear iron clusters (2005) J Biol Chem 280, 15960-6.
73 120
)
-1 100
residue 80
-1
d mol 60 2
40 (deg cm
-3 20
0
-20
Molecular ellipticity X ellipticity Molecular 10 -40
-60 180 200 220 240
wavelength (nm)
Figure 1: CD spectra of as-isolated PepN and Fe-added PepN. The data were collected at 25 oC on a JASCO J-810 CD spectropolarimeter. The enzymes were dialyzed into 5 mM phosphate buffer, pH 7.0, and were 75 μg/mL. Figure was created using Sigmaplot v.6.1. Inset: ellipticity values plotted versus the amount of added Fe. Black – as-isolated PepN; red – 1 eq Fe; green – 2 eq Fe; blue – 3 eq Fe; pink – 4 eq Fe; cyan – 5 eq Fe; dark red – 10 eq Fe.
74
700 600
500
400
300
Fluorescence 200 100
0
300 320 340 360 380 400 wavelength (nm)
Figure 2: Fluorescence emission spectra of as-isolated PepN (dashed line) and PepN after the addition of 1, 2, 3, 4, 5, 6 and 10 equivalents of Fe(II). The spectra were obtained using 2.5 μM PepN in 50 mM Tris, pH 8.5 at 25 oC. The excitation wavelength used in these studies was 284 nm. Black – as-isolated PepN; red – 1 eq Fe; green – 2 eq Fe; blue – 3 eq Fe; dark red – 4 eq Fe; cyan – 5 eq Fe; gray – 6 eq Fe; dark cyan – 10 eq Fe.
75
4
3
2
Absorbance 1
0
300 400 500 600 700 800 wavelength (nm)
Figure 3: UV-Vis difference spectra of as-isolated PepN titrated with Fe(II). As-isolated PepN (1.0 mM) in 50 mM Tris, pH 8.5, was titrated by addition of 1, 2, 3, 4, 5, and 10 molar equivalents of Fe(II) (increasing intensities as Fe(II) as added with bottom curve the spectrum of as-isolated PepN). Black – as-isolated PepN; red – 1 eq Fe; green – 2 eq Fe; blue – 3 eq Fe; pink – 4 eq Fe; cyan – 5 eq Fe; dark red – 10 eq Fe.
76 A
B
C
D
E
F
G
H
I
. Figure 4: EPR spectra of as-isolated PepN collected on a Bruker EleXsys E500 EPR spectrometer equipped with an ER-4116DM dual mode TE102/TE012 resonant cavity, operating at 9.63 GHz in perpendicular mode and an Oxford Instruments ESR900 helium flow cryostat and an Oxford Instruments ITC502 temperature controller facilitated sample temperature variation from 3.6 - 80 K. EPR samples were ca. 1 mM PepN. Spectra A and B: as-isolated PepN (0.5 eq. of Fe) at 5 K and 15 K, respectively; Spectrum C: as-isolated PepN containing 1 eq. of dithionite; Spectrum D: as-isolated PepN containing 5 eq. of dithionite; Spectra E, F, & G: as- isolated PepN containing 5 eq. of dithionite, obtained at 20, 60, and 70K, respectively; Spectrum H: as-isolated PepN containing 1 eq. of H2O2; Spectrum I: as-isolated PepN containing 5 eq. of H2O2.
77
A
B
C
D
Figure 5: EPR spectra of as-isolated PepN containing 5 equivalents of Fe. Spectrum A: PepN containing 5 eq. of Fe at 5K; Spectrum B: PepN containing 5 eq. of Fe and 1 eq. of dithionite; Spectrum C: PepN containing 5 eq. of Fe and 5 eq. of dithionite; Spectrum D: PepN containing 5 eq. of Fe and 1 eq. hydrogen peroxide.
78
A
B
C
D
E
200 180 160 140 120 100 80 60 40 20
ppm
Figure 6: 1H NMR spectra of PepN. A: Spectrum of apo-PepN; B: Spectrum of as-isolated PepN in 90% D2O; C: Spectrum of as-isolated PepN with 5 equivalents Fe(II) added; D: Spectrum of as-isolated PepN reduced with 1 equivalent of dithionite; E: Spectrum of as- isolated PepN. Samples were ~1 mM PepN in 50 mM Tris, pH 8.5, in 10% or 90 % D2O as indicated. The spectra were collected at a temperature of 300 K on a Bruker 500 MHz NMR spectrometer.
79
A. Proposed Fe(II)-center in recombinant E. coli PepN
B. Hypothetical multi-iron-oxo bridged cluster
Figure 7: Proposed iron centers in recombinant E. coli PepN. A - Proposed Fe(II)-center in recombinant E. coli PepN; B - Hypothetical multi-iron-oxo bridged cluster in recombinant E. coli PepN. Figures rendered using CS ChemDraw Ultra v.5.0.
80 Chapter 5 Kinetic and mechanistic studies on recombinant PepN
Introduction Early studies on aminopeptidase N (PepN) focused on the structure, cellular localization, and regulation of the enzyme (1-6). Later, many groups investigated the role of PepN in commercial dairy processes such as cheese ripening and flavor enhancement (7-10). PepN is an exoaminopeptidase that hydrolyzes the N-terminal amino acid from peptides and proteins. Depending on the source, PepN can be cytosolic or membrane-associated and is active across a wide range of pH’s and temperatures (see Chapter 1 for more details). Recently, efforts have focused on attempts to clone and over-express recombinant PepN from E. coli for more basic biomedical research. Nandi and Chandu have hypothesized that PepN is the major aminopeptidase in E. coli and is the sole alanyl aminopeptidase in this organism (11, 12). These characteristics of E. coli PepN have lead to an interest in this enzyme as a possible target for novel antibiotics (13). After many decades of research on PepN however, very little is known about the structure, physiological function, and reaction mechanism of aminopeptidase N from any source. The most recent literature assumes that PepN is a Zn(II)-metalloenzyme, despite the fact that no metal analyses have been reported on a PepN from any source (our manuscript (Chapter 3) is in press at this time). This assumption is entirely based on amino acid sequence homologies with known Zn(II)-containing enzymes that contain a HEXXH motif (5). The assumption that PepN is a Zn(II)-containing enzyme in the absence of experimental proof is surprising, since the only biochemical study probing the mechanism of E. coli PepN concluded that the enzyme is a cysteine aminopeptidase (14). In this study, pH dependence, kinetic studies revealed a bell-
shaped log Km versus pH plot with inflection points due to ionizable groups that had pKa values of 7.2 and 8.2 when using L-ala-p-nitroanilide as the substrate. The latter pKa was assigned to the ionization of the amino group on substrate, while the former pKa was assigned to the ionization of an active site cysteine. Chemical modification studies demonstrated that E. coli PepN was inactivated by N-ethylmaleimide (a cysteine-specific chemical modification agent), and the enzyme was protected from inactivation by incubating the enzyme with competitive
81 inhibitor hexylamine. Interestingly, this group reported in this same study that Zn(II) is an inhibitor of E. coli PepN (14). In this chapter, we conducted a series of biochemical and kinetic studies to probe the reaction mechanism of E. coli PepN. In addition, several novel inhibitors are reported for this enzyme. The information contained in this chapter should aid in the future development of more specific inhibitors of PepN that may be used as potential antibiotics to combat antibiotic-resistant bacterial infections.
Materials and Methods Most of the materials used in this chapter were described in Chapters 3 and 4. L-alanine- p-nitroanilide was purchased from Sigma. Tris-(hydroxymethyl)aminomethane (Tris), 3- cyclohexylamino-1-propanesulfonic acid (CAPS), 2-[N-cyclohexylamino]ethanesulfonic acid (CHES), and 2-(4-morpholino)-ethanesulfonic acid (MES) were purchased from Fisher Biotech. Iodoacetic acid was purchased from Sigma. Ferrous ammonium sulfate was purchased from Aldrich, and ferrous sulfate (99.999%) was purchased from American Elements. Deuterium oxide, 99.9 deuterium atom percent, was purchased from Aldrich. All buffers and media were prepared using Barnstead NANOpure ultrapure water (18 MΩ cm-1). Inhibitors used in this study were synthesized by Dr. Ke-Wu Yang (15) unless otherwise noted.
Preparation of PepN. Recombinant PepN from E. coli was over-expressed, purified, and characterized as described in Chapter 3.
Steady-State Kinetics. Steady-state kinetic assays of PepN were conducted as described in Chapters 2 and 3. Steady-state kinetic assays were conducted in 50 mM Tris buffer, pH 8.5, on a HP 5480A diode array UV-Vis spectrophotometer at 37 oC. Absorbance changes were converted into concentration changes by using Beer’s law and the molar absorptivity of p- nitroaniline at 404 nm (10,700 M-1cm-1). Substrate concentrations were varied between 50 μM –
8 mM. Steady-state kinetics constants, Km and kcat, were determined by fitting initial velocity versus substrate concentration data directly to the Michaelis-Menten equation using CurveFit (16). The errors reported are generated by CurveFit as a result of Chi square minimization. All steady-state kinetic studies with recombinant PepN were performed in triplicate.
82
Chemical modification of PepN with iodoacetate. One molar equivalent of iodoacetic acid was added a solution of 860 μM as-isolated PepN in 50 mM Tris, pH 8.5. The solution was allowed to incubate on ice for 20 minutes, and steady-state kinetic studies were performed to measure the observed change in activity (17). In-gel trypsin digestions and peptide identifications using MALDI-TOF mass spectrometry were conducted as previously described (18).
pH Dependence studies. pH Dependence studies were carried out in a multi-component buffer consisting of 20 mM CHES, CAPS, MES, and Tris containing 100 mM NaCl, pH adjusted to pH 6.5 through pH 10 (19). The concentration of substrate L-alanine-p-nitroanilide was varied
between 0.1 X and 10 X Km, and the concentration of PepN in all assays was 1 nM. The log kcat + or log (kcat/Km) versus pH plots were fitted to the equation, log Y = log (C / (1 + ([H ]/K1) + (K2/ + [H ]))) where Y is kcat or (kcat/Km), C is the pH independent value of Y, and K1 and K2 are the equilibrium constants for groups that take part in the protonation events (20).
Proton inventory studies. Proton inventories were performed by determining the value of kcat in
buffers containing 0 – 100% D2O. The buffer used in this study was 50 mM Tris, pH 7.0 and 8.5. 2 The pH(D) of the D2O-containing solutions was corrected for the lower activity of H as compared to 1H by using the formula pD = pH (reading) + 0.4. The concentrations of L-alanine- p-nitroanilide used in this study were 50 μM, 250 μM, 500 μM, 1 mM, 2 mM, 4mM, and 8 mM, and the PepN concentration was 1 nM for all reactions. Proton inventory data were fitted to the Gross-Butler equation assuming one-proton, two-proton, and multiple-proton in flight models (21).
Stopped-flow kinetics. All of the pre-steady state kinetic experiments were carried out in 50 mM Tris, pH 8.5 at 4 oC. Assays were conducted using 157 μM PepN and 78 μM, 118 μM, 157 μM, 314 μM, and 471 μM L-alanine-p-nitroanilide. Stopped-flow Vis studies on PepN were performed using an Applied Photophysics SX.18MV stopped-flow spectrophotometer equipped with a 1 cm pathlength optical cell. In a typical experiment, a solution of PepN was rapidly mixed with L-alanine-p-nitroanilide in an Applied Photophysics SX.18MV stopped-flow
83 spectrophotometer. Data from at least three reproducible experiments were collected, averaged, and corrected for the instrument dead time (1.5 ms). The molar extinction coefficient of p- -1 -1 nitroaniline (ε404 nm = 10,700 M cm ) was used to convert the single wavelength, stopped-flow absorbance data of product to concentration data. The molar extinction coefficient for substrate was determined experimentally by using known standards of L-alanine-p-nitroanilide and Beer's law.
Kinetic simulations. The kinetic mechanism for PepN hydrolysis of L-ala-p-nitroanilide was simulated using the program KINSIM (22-26). The values for flux and integral tolerances were left as the default values of 0.02 and 1 X 10-6, respectively.
Inhibition studies. Inhibition studies of PepN by L-alanine and p-nitroaniline were conducted in 50 mM Hepes, pH 7.0 using L-alanine-p-nitroanilide as the substrate, L-alanine concentrations of 5, 10, and 15 mM, and p-nitroaniline concentrations of 50, 150, and 300 μM. Inhibition studies using other inhibitors were conducted similarly. All kinetic assays were conducted on a HP 5480A UV-Vis spectrophotometer operating at 25 oC and performed in triplicate. The mode of
inhibition was determined by generating Lineweaver-Burk plots of the data (27), and the Ki values for the inhibitors were determined by fitting initial velocity versus substrate concentration at each inhibitor concentration to vi = Vmax[S] / [S] + Km(1 + [I]/Ki) using SigmaPlot v. 4.0,
where vi = initial velocity, Vmax = maximum velocity, [S] = initial substrate concentration, Km =
Michaelis constant, [I] = inhibitor concentration, and Ki = inhibition constant (27, 28). The
reported Ki values represent the average of all calculations with each inhibitor, and the reported
errors are standard deviations (σn-1, n = 8).
Results and Discussion
Chemical modification. Previous chemical modification studies on E. coli PepN using N- ethylmaleimide suggested that the enzyme has an active site cysteine that is potentially involved in catalysis (14). In these studies a large excess of chemical modification agent was used, and the authors reported a 90% reduction in activity after 60 minutes of incubation. Protection from chemical modification was partially achieved (50%) by including a large excess of a competitive
84 inhibitor. Subsequent to that work, PepN was predicted to utilize an active site Zn(II) in catalysis (29). In order to address this apparent discrepancy, we conducted kinetic assays of PepN in the presence of stoichiometric amounts of iodoacetic acid to verify the previous chemical modification studies. Recombinant PepN was incubated on ice with 1 molar equivalent of iodoacetic acid for 30 minutes, and the resulting enzyme was assayed for catalytic activity. The chemically- modified enzyme was inhibited by 30%, as compared to an unmodified enzyme. This result is consistent with the previous chemical modification studies (14) and suggests that E. coli PepN is a cysteine aminopeptidase. The inability to completely inactivate the enzyme with 1 molar equivalent of iodoacetic acid is probably due to the presence of 7 other cysteines in PepN and the
presence of protonated Cys (ca. 24% if a pKa of 8.0 is assumed for Cys), which would not react as favorably as the anion. We attempted to identify which Cys residue(s) was modified in our studies by using trypsin digestions, MALDI-TOF mass spectrometry, and peptide identifications (18). This procedure yielded good peptide coverage with 25 fragments positively identified as being from PepN (Figure 1). Unfortunately, none of the ionizable fragments contained a cysteine residue. To identify which residues are modified with iodoacetic acid will require cleavage with another protease or with two proteases so that smaller peptides result. It may be also possible to identify which Cys residue(s) is modified when PepN is reacted with iodoacetic acid by preparing site- directed mutants of PepN with each Cys residue sequentially mutated to Ser. This approach may be problematic if any of the Cys residues are involved in disulfide bonds.
pH Dependence studies. In the mechanism paper by Chappelet-Tordo (14), pH dependence studies on E. coli PepN were presented (summarized in the Introduction of this chapter). The
conclusions drawn from these studies suggested that an active site Cys with a pKa of 7.2 was ionized in the free enzyme (inflection point observed in the log Km and log Ki versus pH plots).
These studies also demonstrated that there were no inflection points in the log kcat versus pH plots. These results are surprising since an active site cysteine that is purported to be involved in catalysis would be expected to be differentially active when protonated versus deprotonated.
One would have expected to observe an inflection point in the log kcat versus pH plot
corresponding to the pKa of a catalytically-important Cys. We therefore conducted pH
85 dependence, steady-state kinetic studies to probe for any kinetically-important ionizations in recombinant E. coli PepN.
The log kcat /Km and log kcat versus pH plots are shown in Figure 2. These studies could only be conducted over a relatively narrow pH range since the enzyme precipitates at pH values less than 6.5 and the substrate is insoluble at pH values greater than 9.5. These studies were conducted in a four component buffer system to minimize differential activities of PepN in different buffers. In addition, the buffers used in this study contained 100 mM NaCl to minimize
ionic strength effects. Inflection points in log kcat /Km versus pH plots are normally associated
with ionization of groups on free enzyme or substrate (20). Inflection points in log kcat versus pH plots are attributed to kinetically-important ionizations in the ES complex. In contrast to the data published by Chappelet-Tordo (14), our data clearly shows no discernable inflection points in the
log kcat /Km versus pH plot. In addition, our log kcat versus pH data, which was fitted to the equation in the materials and methods section, resulted in a bell-shaped plot with inflection points corresponding to pKa’s of 6.5 + 0.3 and 9.0 + 0.5. The former pKa is tentatively assigned to an active site His that could be involved in acid/base catalysis. The residue that exhibits this
pKa is most active when deprotonated. The latter pKa is probably due to an active site Cys. The
fact that this pKa appears on the basic limb of the pH dependence plot suggests that the more active form of Cys is the protonated form, and as Cys is deprotonated, the activity drops. These pH dependence data suggest the involvement of an active site Cys in catalysis and that the most active form of this Cys is the protonated form.
Solvent isotope studies. To probe further the mechanism of PepN, solvent isotope studies were conducted at pH 7.0, which is the pH optimum of the enzyme (Figure 2). Solvent isotope studies can be used to determine whether rate-limiting proton transfers (from solvent) occur during catalysis (30). The mass of a deuteron is roughly twice that of a proton. In reactions that have rate-significant proton transfers, it is expected that rate of proton transfer would be different (normally faster) than the corresponding deuteron transfer. PepN exhibited a solvent isotope effect at pH 7.0 of kH/kD = 2.2, which suggests a proton transfer(s) during a nucleophilic reaction (31). A proton inventory at pH 7.0 was also conducted to evaluate the number of protons that are transferred during the rate-limiting step(s) of PepN (Figure 3). The data in proton inventories
86 are usually presented as kcat versus %D2O plots, and data are normally fitted to the Gross-Butler equation (21). If the data fit best to a straight line, the theory predicts that there is one “proton in flight” during the rate-significant step. If the data are fitted best to quadratic or exponential versions of the Gross-Butler equation, the theory predicts 2 or multiple (infinite) “protons in flight”, respectively. As the number of steps, particularly steps involving transition states with differential solvent isotope effects, increase, the complexity of the Gross-Butler equation increases. The data obtained for PepN at pH 7.0 does not fit well to a straight line (Figure 3), suggesting that more than one proton is in flight during the rate-limiting step(s). The scatter in the data does not unequivocally rule out the possibility of good fit to a straight line. However, the data on PepN are best fitted using a complicated version of the Gross-Butler equation (kN = ko(1-
n+nΦT)/(1-n+nΦR) where n is the % D2O, kN is the observed rate, ko is the maximum rate in D H 100% H2O, and ΦT, ΦR are fractionation factors (k /k ) corresponding to transition states in the
mechanism (21). The fitting of the data to this equation yielded values for ΦT, ΦR of 0.186 + 0.052 and 0.418 + 0.095, respectively. A similar proton inventory plot was reported for the serine hydrolase acetylcholinesterase (21). This proton inventory for acetylcholinesterase was initially hypothesized to result from contributions from a normal isotope effect and a reverse isotope effect during the transition states of the reaction. However, these data have been re-attributed to
a change in the rate-determining step when the reaction is catalyzed in D2O versus H2O (21). Taken together, the solvent isotope and proton inventory data on PepN demonstrates at least one rate-limiting proton transfer during catalysis. Further studies are required (reactions at different
%D2O values) to unequivocally determine the number of “protons in flight” during the rate- limiting step.
Inhibition of PepN by L-alanine and p-nitroaniline. The reaction of PepN with L-alanine-p- nitroanilide results in the formation of two very distinct products, L-alanine and p-nitroaniline (32). Product inhibition studies were conducted in an effort to determine which product binds tighter to PepN. The rationale behind these studies is that differences in the binding affinities of the products suggest an ordered release of products during catalysis. Products with similar binding constants would indicate random release of products during catalysis. L-alanine is a
competitive inhibitor of PepN with a Ki of 9.6 + 3.5 mM when using L-ala-p-nitroanilide as substrate. Chapplett-Tordo and coworkers previously reported that L-alanine inhibits E. coli
87 PepN with a Ki greater than 10 mM (14). The second product, p-nitroaniline, is a competitive
inhibitor of PepN with a Ki of 167 + 11 μM. The greater than 10-fold tighter binding of p- nitroaniline suggests that PepN utilizes an ordered, uni-bi (uni – one substrate, bi – two products) hydrolytic mechanism (27) and that L-ala dissociates from the enzyme first.
Stopped-flow kinetic studies. L-Ala-p-nitroanilide is potentially an excellent substrate to use in -1 -1 - kinetic studies because the substrate (ε300nm = 6,700 M cm ) and product (ε404nm = 10,700 M 1cm-1) absorb UV-Vis radiation, thus allowing for simultaneous monitoring of substrate depletion and product formation. Unfortunately, unhydrolyzed L-ala-p-nitroanilide absorbs at 300 nm, and the large absorbance of PepN at 280 nm at high enzyme concentrations partially masks the absorbance of unhydrolyzed substrate. However, the absorption peak of product is far enough removed from the peak at 280 nm to allow for monitoring of product formation even at high enzyme concentrations. Stopped-flow kinetic studies were conducted to probe the reactions of 157 μM PepN with 471 μM, 335 μM, 157 μM, 118 μM, and 78 μM L-ala-p-nitroanilide. The concentration of enzyme and substrate used in these studies allowed for us to monitor the PepN reaction at single turnover and at steady-state conditions. The resulting progress curves for PepN showed a rapid, linear increase in product concentrations, with no apparent burst phases, until substrate was depleted (Figure 4, data points). Efforts were also made to follow the reaction of PepN with L-ala-p-nitroanilide using stopped-flow fluorescence studies; however, there were no changes in the fluorescence properties of PepN during the course of the reaction (data not shown).
Kinetic simulations. By using the product inhibition data and the stopped-flow UV-Vis data, a minimal kinetic mechanism was proposed (Figure 5). This mechanism assumes reversible binding of substrate and end products, an ordered release of products, and a single, rate-limiting chemistry step. KINSIM was used to simulate progress curves based on the proposed kinetic mechanism, and the simulated progress curves are shown as lines in Figure 5. KINSIM is a computer program that simulates reaction progress curves based on the input of a reaction mechanism, rate constants, and initial reactant concentrations (25, 26). In the simulated curves for PepN, substrate and end products were assumed to bind at diffusion-controlled rates of 108 -1 -1 M s , k3 and k4 were determined from the Ki values of L-alanine and p-nitroaniline,
88 respectively. The second step (chemistry step) was assumed to be irreversible and was set to the
experimentally-determined kcat value. The shapes of the simulated curves were highly-dependent
on the values of k-1 and k2. By using the mechanism and rate constants shown in Figure 5, simulated progress curves were generated that reasonably fit the stopped-flow data (Figure 4).
This mechanism suggests that the chemistry step is rate-limiting and that the Ks (dissociation constant for substrate binding) is 30 mM. This value, although very high, is not surprising since
L-ala-p-nitroanilide is not a natural substrate for PepN. As is the case for most enzymes, the Ks is
not equal to Km, which is 350 μM.
Inhibition studies on PepN. In an effort to probe the rate-limiting step of the reaction mechanism of PepN, several phosphinate, sulfonate, and sulfonamidate analogs of alanyl-alanine were tested as inhibitors of recombinant PepN from E. coli (Figure 2, Chapter 1). Previous studies have demonstrated that similar analogs are transition state analog inhibitors of peptidases,(33-36) and these inhibitors are typically tight binding and specific. We reasoned that -10 if these inhibitors bound tightly (Ki <<10 M) we could argue that the inhibitors were transition state analogs and that the rate-limiting step(s) of the PepN mechanism involves the formation of a tetrahedral intermediate. Propylphosphinate was previously reported to be an inhibitor of the aminopeptidase VanX (37, 38) and was synthesized as previously described (39). This compound was shown to be a competitive inhibitor of PepN from E. coli with a Ki value of 10 + 2 μM (Table 1). The comparable phosphonamidate analog was a partial competitive inhibitor with a Ki of 36 + 2 μM (39), while the phosphonate analogs did not appreciably inhibit PepN at concentrations up to 200 μM (Table 1). We reasoned that the P – X – C bond angle (ideally 109.5o in propylphosphinate, 107o in phosphonamide, and 104.5o in the phosphonate analog) may play a role in the relative binding strengths of these compounds. Therefore, a compound with a large P – X – C bond angle was tested as an inhibitor. The propenylphosphinate was synthesized as described previously
(15) and shown to be a competitive inhibitor with a Ki of 1.0 + 0.2 μM (Table 1). As we hypothesized, the lengthening of the P – X – C bond angle resulted in tighter binding of the inhibitor. In an effort to improve the binding affinity of these phosphinate dipeptide analogs to
PepN, a hydrophobic substituent (C8H17) was attached to the β-carbon of propylphosphinate to
89 yield decylphosphinate. We reasoned that decylphosphinate would be a bifunctional inhibitor in which the substituent would provide additional points of attachment to PepN. Decylphosphinate
was shown to be a competitive inhibitor of PepN with a Ki of 1.1 + 0.1 μM (Table 1). The inclusion of the hydrophobic substituent on propylphosphinate did result in improved binding affinity. Unfortunately, efforts to synthesize a decylphosphinate analog of propenylphosphinate have been unsuccessful. Previously, phosphinate analogs of peptides have been reported to be very tight binding -15 inhibitors of several peptidases with Ki values reported as low as 10 M (35). The relatively weaker binding to PepN suggested that these compounds may not be transition state analogs. Nonetheless, we attempted to improve the binding of the dipeptide analogs by preparing sulfonate and sulfonamide analogs of Ala-Ala. We reasoned that the sulfonate and sulfonamide analogs would require less desolvation than the phosphinate analogs, and the reduced desolvation energy would result in relatively tighter binding. The sulfonate and sulfonamidate analogs did not inhibit PepN at concentrations up to 300 μM. These studies demonstrate that the value of the P – X – C bond angle in phosphinate analogs of Ala-Ala plays a role in the relative binding affinity of these compounds to PepN. In addition, the presence of a hydrophobic substituent on the β-carbon of the inhibitor also improves the binding affinity of the compounds. Recently, Grembecka et al. reported that the phosphinate analog of PheTyr is a tight binding inhibitor (Ki of 36 nM for diastereomeric mixture) of mammalian PepN (40). The work presented herein suggests that the binding affinity of this compound could be improved by ca. 100-fold if a double bond and a hydrophobic substituent is included in the phosphinate dipeptide. These studies also show that the use of sulfonate/sulfonamide analogs to replace phosphinates is not an effective strategy to improve the binding strength of the compounds. Lastly, these studies strongly suggest that phosphinate analogs are not transition state analogs of PepN, which further suggests that the formation of a tetrahedral intermediate is not rate-limiting.
Conclusions Contrary to the most recent literature, the chemical modification and previous inhibition studies with leupeptin (Chapter 2) strongly suggest that E. coli PepN is a cysteine active site
90 aminopeptidase. This conclusion is supported by pH dependence, steady-state kinetic data that also implicated a His residue in the mechanism. By using all of the data presented in this chapter, a potential reaction mechanism of E. coli is now possible (Figure 6). pH Dependence data demonstrated that the most active form of the enzyme has a protonated Cys and a deprotonated His. After substrate binding (top left structure in Figure 6), we have suggested that there is a H-bond between the thiol on the active site Cys with the active site His. The protonated thiol on Cys nucleophilically attacks the carbonyl on substrate, generating a thio-tetrahedral intermediate, whose formation nor breakdown is rate-limiting (see inhibition studies). It is possible that the generation of the nucleophile (i.e., proton transfer from Cys to His) is the rate-limiting proton transfer identified in solvent isotope studies. It is entirely possible that the cleavage of the C-N bond on substrate occurs upon nucleophilic attack by the Cys; however, all proposed reaction mechanisms of peptidases predict C-N bond cleavage upon attack of the tetrahedral intermediate by water (as shown in Figure 6). Our data at this point does not favor either scenario. A solvent molecule is then activated by an unknown pathway to attack the tetrahedral intermediate resulting in the cleavage of the C-N bond on substrate. Product inhibition studies suggest that product p- nitroaniline does not leave the active site at this point. The cleavage of the C-N bond probably occurs simultaneously to protonation of the p-nitroaniline amine group. The proton for this protonation could come from solvent (as shown in Figure 6) or possibly from that active site His. It is entirely possible that this proton transfer could also be rate-limiting. The breakdown of the second tetrahedral intermediate results in the formation of L-ala, which readily dissociates from the active site. Product inhibition studies suggest that then p-nitroaniline dissociates. This proposed reaction mechanism now provides a model to test in future studies. The identification of the active site Cys is critical, and future studies are needed to address issues such as the timing of C-N bond cleavage, the proton donor source, and other active site residues that may be involved in the stabilization of the tetrahedral intermediate and the activation of water.
91 References 1. Lazdunski, C., Busuttil, J., and Lazdunski, A. Purification and properties of a periplasmic aminoendopeptidase from Escherichia coli (1975) Eur J Biochem 60, 363-9. 2. Bally, M., Murgier, M., Tommassen, J., and Lazdunski, A. Physical mapping of the gene for aminopeptidase-N in Escherichia-coli-K12 (1984) Mol Gen Genet 193, 190-191. 3. Bally, M., Murgier, M., and Lazdunski, A. Cloning and orientation of the gene encoding aminopeptidase-N in Escherichia-coli (1984) Mol Gen Genet 195, 507-510. 4. Murgier, M., Pellissier, C., Lazdunski, A., Bernadac, A., and Lazdunski, C. Aminopeptidase N from Escherichia coli. Unusual interactions with the cell surface (1977) Eur J Biochem 74, 425-33. 5. Foglino, M., Gharbi, S., and Lazdunski, A. Nucleotide-sequence of the pepN gene encoding aminopeptidase-N of Escherichia-coli (1986) Gene 49, 303-309. 6. Lazdunski, A., Murgier, M., and Lazdunski, C. Evidence for an aminoendopeptidase localized near the cell surface of Escherichia coli. Regulation of synthesis by inorganic phosphate (1975) Eur J Biochem 60, 349-55. 7. Tan, P. S., van Alen-Boerrigter, I. J., Poolman, B., Siezen, R. J., de Vos, W. M., and Konings, W. N. Characterization of the Lactococcus lactis pepN gene encoding an aminopeptidase homologous to mammalian aminopeptidase N (1992) FEBS Lett 306, 9-16. 8. Tan, P. S. T., and Konings, W. N. Purification and characterization of an aminopeptidase from Lactococcus lactis subsp. cremoris wg2 (1990) Appl Environ Microbiol 56, 526- 532. 9. Arora, G., and Lee, B. H. Comparative studies on pepidases of Lactobacillus casei subspecies (1990) J Dairy Sci 73, 274-279. 10. van Alen-Boerrigter, I. J., Baankreis, R., and de Vos, W. M. Characterization and overexpression of the Lactococcus lactis pepN gene and localization of its product, aminopeptidase N (1991) Appl Environ Microbiol 57, 2555-61. 11. Chandu, D., Kumar, A., and Nandi, D. PepN, the major suc-llvy-amc-hydrolyzing enzyme in Escherichia coli, displays functional similarity with downstream processing enzymes in archaea and eukarya. Implications in cytosolic protein degradation (2003) J Biol Chem 278, 5548-56.
92 12. Chandu, D., and Nandi, D. PepN is the major aminopeptidase in Escherichia coli: Insights on substrate specificity and role during sodium-salicylate-induced stress (2003) Microbiol-Sgm 149, 3437-3447. 13. Travis, J., and Potempa, J. Bacterial proteinases as targets for the development of second- generation antibiotics (2000) Biochim Biophys Acta 1477, 35-50. 14. Chappelet-Tordo, D., Lazdunski, C., Murgier, M., and Lazdunski, A. Aminopeptidase N from Escherichia coli: Ionizable active-center groups and substrate specificity (1977) Eur J Biochem 81, 299-305. 15. Yang, K. W., Golich, F. C., Sigdel, T. K., and Crowder, M. W. Phosphinate, sulfonate, and sulfonamidate dipeptides as potential inhibitors of Escherichia coli aminopeptidase N (2005) Bioorg Med Chem Lett 15, 5150-3. 16. Crowder, M. W., Walsh, T. R., Banovic, L., Pettit, M., and Spencer, J. Overexpression, purification, and characterization of the cloned metallo-b-lactamase L1 from Stenotrophomonas maltophilia (1998) Antimicrob Agents Chemother 42, 921-926. 17. Oesterhelt, D., Bauer, H., Kresze, G. B., Steber, L., and Lynen, F. Reaction of yeast fatty acid synthetase with iodoacetamide. 1. Kinetics of inactivation and extent of carboxamidomethylation (1977) Eur J Biochem 79, 173-80. 18. Sigdel Tara, K., Cilliers, R., Gursahaney Priya, R., and Crowder Michael, W. Fractionation of soluble proteins in Escherichia coli using DEAE-, SP-, and Phenyl sepharose chromatographies (2004) J Biomol Tech : JBT 15, 199-207. 19. Yanchak, M. P., Taylor, R. A., and Crowder, M. W. Mutational analysis of metallo-b- lactamase CcrA from Bacteroides fragilis (2000) Biochemistry 39, 11330-11339. 20. Cornish-Bowden, A. (1995) Analysis of enzyme kinetic data, Oxford University Press, Oxford. 21. Venkatasubban, K. S., and Schowen, R. L. The proton inventory technique (1984) CRC Crit Rev Biochem 17, 1-44. 22. Barshop, B. A., Wrenn, R. F., and Frieden, C. Analysis of numerical methods for computer simulation of kinetic processes: Development of KINSIM -- a flexible, portable system (1983) Anal Biochem 130, 134-145. 23. Dang, Q., and Frieden, C. New PC versions of the kinetic-simulation and fitting programs, KINSIM and FITSIM (1997) Trends Biochemical Sci 22, 317.
93 24. Fierke, C. A., Johnson, K. A., and Benkovic, S. J. Construction and evaluation of the kinetic scheme associated with dihydrofolate reductase from Esherichia coli (1987) Biochemistry 26, 4085-4082. 25. Frieden, C. Numerical integration of rate equations by computer (1993) Trends Biochemical Sci 18, 58-60. 26. Frieden, C. Numerical integration of rate equations by computer: An update (1994) Trends Biochemical Sci 19, 181-182. 27. Segel, I. H. (1993) Enzyme kinetics, John Wiley and Sons, Inc., New York. 28. Yang, K. W. Inhibition studies on the metallo-b-lactamase L1 from Stenotrophomonas maltophilia (1999) Arch Biochem Biophys 368, 1-6. 29. Gonzales, T., and Robert-Baudouy, J. Bacterial aminopeptidases: Properties and functions (1996) FEMS Microbiol Lett 18, 319-344. 30. Schowen, K. B., and Schowen, R. L. (1982) in Methods in enzymology pp 551-606, Academic Press. 31. Garrity, J. D., Carenbauer, A. L., Herron, L. R., and Crowder, M. W. Metal binding Asp-120 in metallo-β-lactamase L1 from Stenotrophomonas maltophilia plays a crucial role in catalysis (2004) J Biol Chem 279, 920-927. 32. Brandt, J. J., Chatwood, L. L., Yang, K. W., and Crowder, M. W. Continuous assay for VanX, the D-alanyl-D-alanine dipeptidase required for high-level vancomycin resistance (1999) Anal Biochem 272, 94-99. 33. Bartlett, P. A., and Kezer, W. B. Phosphinic acid dipeptide analogues: Potent, slow-binding inhibitors of aspartic peptidases (1984) J Am Chem Soc 106, 4282-4283. 34. Bartlett, P. A., and Marlowe, C. K. Phosphoramidates as transition state analogue inhibitors of thermolysin (1991) Biochemistry 22, 4618-4624. 35. Kaplan, A. P., and Bartlett, P. A. Synthesis and evaluation of an inhibitor of
carboxypeptidase A with a Ki value in the femtomolar range (1991) Biochemistry 30, 8165-70. 36. Phillips, M. A., Kaplan, A. P., Rutter, W. J., and Bartlett, P. A. Transition-state characterization: A new approach combining inhibitor analogues and variation in enzyme structure (1992) Biochemistry 31, 959-963.
94 37. Wu, Z., and Walsh, C. T. Phosphinate analogs of D-,D-dipeptides: Slow-binding inhibition and proteolysis protection of VanX, a D-, D-dipeptidase required for vancomycin resistance in Enterococcus faecium (1995) Proc Natl Acad Sci USA 92, 11603-11607. 38. Wu, Z., Wright, G. D., and Walsh, C. T. Overexpression, purification, and characterization of VanX, a D-, D-dipeptidase which is essential for vancomycin resistance in Enterococcus faecium BM4147 (1995) Biochemistry 34, 2455-2463. 39. Yang, K. W., Brandt, J. J., Chatwood, L. L., and Crowder, M. W. Phosphonamidate and phosphothioate dipeptides as potential inhibitors of VanX (2000) Bioorg Med Chem Lett 10, 1087-1089. 40. Grembecka, J., Mucha, A., Cierpicki, T., and Kafarski, P. The most potent organophosphorus inhibitors of leucine aminopeptidase. Structure-based design, chemistry, and activity (2003) J Med Chem 46, 2641-2655.
95
1 TQQPQAK YRH DYRAPDYQIT DIDLTFDLDA QK TVVTAVSQ AVR HGASDAP 51 LRLNGEDLKL VSVHINDEPW TAWKEEEGAL VISNLPERFT LKIINEISPA 101 ANTALEGLYQ SGDALCTQCE AEGFRHITYY LDRPDVLARF TTKIIADK IK 151 YPFLLSNGNR VAQGELENGR HWVQWQDPFP KPCYLFALVA GDFDVLRDTF 201 TTRSGR EVAL ELYVDR GNLD R APWAMTSLK NSMKWDEERF GLEYDLDIYM 251 IVAVDFFNMG AMENKGLNIF NSKYVLAR TD TATDKDYLDI ERVIGHEYFH 301 NWTGNR VTCR DWFQLSLK EG LTVFRDQEFS SDLGSR AVNR INNVRTMRGL 351 QFAEDASPMA HPIRPDMVIE MNNFYTLTVY EKGAEVIR MI HTLLGEENFQ 401 KGMQLYFERH DGSAATCDDF VQAMEDASNV DLSHFRRWYS QSGTPIVTVK 451 DDYNPETEQY TLTISQR TPA TPDQAEKQPL HIPFAIELYD NEGKVIPLQK 501 GGHPVNSVLN VTQAEQTFVF DNVYFQPVPA LLCEFSAPVK LEYKWSDQQL 551 TFLMR HARNDFSRWDAAQSL LATYIK LNVARHQQGQPLSL PVHVADAFRA 601 VLLDEKIDPA LAAEILTLPS VNEMAELFDI IDPIAIAEVR EALTRTLATE 651 LADELLAIYN ANYQSEYR VE HEDIAKR TLR NACLR FLAFG ETHLADVLVS 701 K QFHEANNMT DALAALSAAV AAQLPCRDAL MQEYDDKWHQ NGLVMDKWFI 751 LQATSPAANV LETVRGLLQH R SFTMSNPNR IRSLIGAFAG SNPAAFHAED 801 GSGYLFLVEM LTDLNSRNPQ VASRLIEPLI RLKRYDAKRQ EKMRAALEQL 851 K GLENLSGDL YEK ITKALA
Figure 1: Matrix Science Mascot search results for trypsin digestion of aminopeptidase N (PepN) from E. coli. Matched peptides shown in bold. Mass spectrometric analysis was performed by T. Sigdel, Ph.D.
96
3 Figure 2: ( )
cat k 2
( ) or log
m 1
/K cat k log 0
678910 pH
Figure 2: pH Dependence plots of as-isolated PepN with L-alanine-p-nitroanilide. The log kcat + + versus pH plot was fitted to the equation, log Y = log (C / (1 + ([H ]/K1) + (K2/ [H ]))) where Y is kcat, C is the pH independent value of Y, and K1 and K2 are the equilibrium constants for groups that take part in the protonation events (solid line). The log kcat versus pH plot is bell- shaped with inflection points corresponding to pKa’s of 6.5 + 0.3 and 9.0 + 0.5. Error bars represent standard deviation of nine trials.
97
800
700
600
) -1
(s 500
cat k 400
300
200 0 20406080100
% D2O
Figure 3: Proton inventory of recombinant PepN at pH(D) 7.0. The dashed line represents a simple single proton in flight model. The solid line represents a fit of the data using the Gross- Butler equation (kN = ko(1-n+nΦT)/(1-n+nΦR) where n is the % D2O, kN is the observed rate, ko is D H the maximum rate in 100% H2O, and ΦT, ΦR are fractionation factors (k /k ) corresponding to transition states in the mechanism (21). The fitting of the data to this equation yielded values for ΦT, ΦR of 0.186 + 0.052 and 0.418 + 0.095, respectively.
98
Figure 4: Kinetic simulations of stopped-flow data using KINSIM. Reactions were carried out in 50 mM Tris, at pH 8.5, at 4 ºC. PepN concentration was 157 μM and substrate (L-ala-p- nitroanilide) concentrations were 471 μM (hexagons), 335 μM (diamonds), 157 μM (triangles), 118 μM (squares), and 78 μM (circles). The lines are the simulated progress curves using KINSIM and the mechanism (with rate constants) in Figure 5.
99
Figure 5: Proposed kinetic mechanism of PepN.
100
Figure 6: Proposed reaction mechanism of E. coli PepN
101
Table 1: Inhibitors of PepN.
Name Structure Ki (μM) P-X-C bond Mode of inhibition anglea Propylphosphonate 10 + 2 109.5o Competitive
Phosphonamidateb 36 + 2 107o Partial competitive
Phosphonate >200 104.5o Competitive
Propenylphosphinate 1.0 + 0.2 120o Competitive
Decylphosphinate 1.1 + 0.1 109.5o Competitive
Sulfonate >300 NDc
Sulfonamidate >300 NDc
aP-X-C bond angle is approximated using small molecule models: for phosphonate, angle of water is used, for phosphonamidate, angle of NH3 is used, for propylphosphinate, angle of CH4 is used, and for propenylphosphinate, sp2 hydridized carbon angle is used. bReference (39). cNot determined.
102 Chapter 6 Summary and Conclusions
This dissertation describes the first detailed structural and mechanistic characterization of aminopeptidase N from E. coli. As discussed throughout these chapters, E. coli PepN has been studied since the 1970’s. Many of the earlier studies were conducted on partially-purified forms of the enzyme However, the report of the DNA (and amino acid) sequence (1, 2) allowed for the cloning of E. coli pepN, and ~20 years later, for the over-expression, purification, and partial characterization of the recombinant enzyme (3, 4). Most studies on PepN since the 1980’s have focused on enzymes from Lactobacillus strains since PepN has been implicated in dairy processes (5) and on enzymes in humans, which have been implicated in a number of diseases and health conditions (6). The hypotheses that PepN is the sole alanyl aminopeptidase and the major aminopeptidase in E. coli (3, 4, 7, 8) and the assertion that bacterial aminopeptidases are potential targets for the generation of novel antibiotics (9) prompted our study of E. coli PepN. Chapter 3 describes the best available over-expression and purification protocol for E. coli PepN. As discussed in Chapter 3, we obtained the over-expression plasmid described by Chandu (3, 4); however, we were unable to obtain any purified enzyme using the arabinose- induced system. The over-expression of PepN is now under control of a lac promoter, and this system yields ~18 mg of purified recombinant PepN per liter of growth culture. These levels of enzyme were sufficient to conduct the first ever detailed structural and mechanistic studies on E. coli PepN. In contrast to previous work (5), our structural studies indicate that E. coli PepN is not a Zn(II)-metalloenzyme. At this point, it is not clear whether PepN from any source requires Zn(II) for activity. Since the mammalian PepN’s are membrane-associated while the bacterial enzymes appear to be cytosolic (6), it is possible that the mammalian enzymes, which exhibit different responses than E. coli PepN to metal chelators and metal ions, may require Zn(II). Future metal analyses are required to address this issue. Our structural studies also, for the first time, demonstrate that E. coli PepN tightly binds 5 equivalents of Fe. Interestingly, kinetic studies show that this Fe is not required for activity, and CD and fluorescence studies indicate that this Fe is not required for the maintenance of structure. EPR, NMR, and UV-Vis studies strongly suggest that the five Fe ions are partitioned into at least 2 distinct sites. At one site,
103 Fe(II) is coordinated by 2 histidines and at least 1 cysteine. These spectroscopic techniques could not yield any information about the identity of the other Fe binding site(s). We have speculated that PepN contains an oxo-bridged, multi-nuclear Fe center similar to that reported in PhoU (10); however, additional studies in the future are required to test this hypothesis. We have established a collaboration with Professor John Peters at Montana State University to solve the crystal structure of recombinant E. coli PepN; however, suitable crystals have not yet been obtained. We also plan to conduct additional EPR studies in which we titrate PepN with Fe(II) in an effort to observe any new EPR signals associated with this proposed cluster. Mössbauer studies are also planned on PepN samples containing 1-5 equivalents of 57Fe to probe each Fe binding site as it is populated. Performing parallel studies with PhoU may aid in the interpretation of these undoubtedly complex spectroscopic studies. Our conclusion that E. coli PepN is not a metalloenzyme prompted a study of the reaction mechanism of the enzyme. The first reaction mechanism for any PepN is proposed in Chapter 5, and this mechanism accounts for all of the available kinetic and mechanistic data published to date on E. coli PepN. This reaction mechanism serves as a starting point for a complete understanding of PepN’s hydrolytic reaction. Future studies will require the use of more physiological substrates, since it is not certain that the mechanism used to hydrolyze L-ala-p- nitroanilide is also used for peptides. These proposed studies are complicated by the fact that there are few, if any, chromophoric peptides available that would allow for the continuous monitoring of the reaction (a desirable feature for mechanistic studies). Efforts will be made in the future to prove the involvement of an active site cysteine, and the proposed studies were discussed at the end of Chapter 4. A crystal structure with bound substrate or bound competitive inhibitor would provide vital information about the role of cysteine in the mechanism. The ever-increasing presence of antibiotic resistance in the clinic has necessitated the search for novel drug targets. The data presented herein can be used immediately for the design of novel inhibitors of E. coli, and possibly other bacterial, PepN’s. The structural studies on E. coli PepN demonstrate that the targeting of an inhibitor to Zn(II) or another metal ion is not an effective strategy for inhibitor design. The mechanistic studies strongly suggest that inhibitors that affect proton transfer during hydrolysis may be effective. In addition, inhibitors that react with thiols to form covalently-bound intermediates may also be useful. Rational design and re-
104 design (possibly via a combinatorial approach) of inhibitors will await the proposed studies discussed in the dissertation.
References 1. Foglino, M., Gharbi, S., and Lazdunski, A. Nucleotide-sequence of the pepn gene encoding aminopeptidase-n of escherichia-coli (1986) Gene 49, 303-309. 2. McCaman, M. T., and Gabe, J. D. The nucleotide-sequence of the pepn gene and its over- expression in escherichia-coli (1986) Gene 48, 145-153. 3. Chandu, D., Kumar, A., and Nandi, D. Pepn, the major suc-llvy-amc-hydrolyzing enzyme in escherichia coli, displays functional similarity with downstream processing enzymes in archaea and eukarya. Implications in cytosolic protein degradation (2003) J Biol Chem 278, 5548-56. 4. Chandu, D., and Nandi, D. Pepn is the major aminopeptidase in escherichia coli: Insights on substrate specificity and role during sodium-salicylate- induced stress (2003) Microbiology-Sgm 149, 3437-3447. 5. Gonzales, T., and Robert-Baudouy, J. Bacterial aminopeptidases: Properties and functions (1996) FEMS Microbiol. Lett. 18, 319-344. 6. Taylor, A. (1996) Aminopeptidases, R.G. Landes Company, Austin, TX. 7. Lazdunski, A., Murgier, M., and Lazdunski, C. Evidence for an aminoendopeptidase localized near the cell surface of escherichia coli. Regulation of synthesis by inorganic phosphate (1975) Eur J Biochem 60, 349-55. 8. Chappelet-Tordo, D., Lazdunski, C., Murgier, M., and Lazdunski, A. Aminopeptidase n from escherichia coli: Ionizable active-center groups and substrate specificity (1977) Eur. J. Biochem. 81, 299-305. 9. Travis, J., and Potempa, J. Bacterial proteinases as targets for the development of second- generation antibiotics (2000) Biochim. Biophys. Acta 1477, 35-50. 10. Liu, J., Lou, Y., Yokota, H., Adams, P. D., Kim, R., and Kim, S. H. Crystal structure of a phou protein homologue: A new class of metalloprotein containing multinuclear iron clusters (2005) J Biol Chem 280, 15960-6.
105