1 STRUCTURAL STUDIES OF AUTONOMOUS PARVOVIRUS CAPSIDS TOWARDS UNDERSTANDING THE MECHANISMS OF CELL RECOGNITION AND ENDOSOMAL TRAFFICKING By SUJATA HALDER A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012
2 2012 Sujata Halder
3 To my parents Debasree Saha Halder and Gokul Chandra Saha Halder for always believing in my abilities
4 ACKNOWLEDGMENTS I would like to begin by thanking my parents; Debasree and Gokul Chandra Saha Halder for the sacrifices they made to provide me with great opportunities and I am forever grateful because they continue to love, support and encourage me ev en after my countless mistakes. I thank my brother, Kunal Halder and grandparents, especially Dr. Rabindranath Basu for timely advice and much needed financial help. I would like to thank my uncle and aunt, Hari Peri and Rama Peri for being the ideal guard ians and for their warmth and words of advice. I feel extremely indebted and honored to be mentored by Dr. Mavis Agbandje McKenna who believed in my abilities and patiently nurtured the scientist in me. I am always grateful for the encouragement and opport unities she has provided me and the personal counseling sessions. I would also like to thank Dr. Robert McKenna for being an inspiring teacher and his great sense of humor My committee members, Dr. James B. Flanegan, Dr. Nicholas Muzyczka, Dr. David C. Bl oom, and Dr. Grant McFadden have been extremely encouraging and helpful and provided critical input. I would also like to thank my collaborators, Dr Peter Tattersall (Yale University) Dr Susan F. Cotmore (Yale University) Dr. Nathalie Salome (Program I nfection and Cancer, Germany) Dr. David F. Smith (Emory University School of Medicine) and the beamline scientists at CHESS (Cornell High Energy Synchrotron Source) for the resources and time invested in my work. I greatly appreciate the help offered by t he administrative staff of the IDP (Interdisciplinary Program) BMB (Biochemistry and Molecular Biology) and International Student Center. from home. Dr. Lakshmanan Govindasamy Dr. Hyun Joo Nam, Dr. Antonette Bennett, Dr. Brittney L. Gurda deserve special thanks for teaching me the techniques used in the
5 lab and also for invaluable advice; Balendu Avvaru, Bala V enkat, Harald Messer, Lawrence Tartaglia and Lauren Drouin for it h elps to have great friends in the lab; Robert Ng and Yue Liu for always being there when I needed your help and for rgraduate students that I have taught over the years, especially Omar Shakeel, Caroline Zapiec and Jay Jackson for being patient with me. I also thank my IDP buddy Mansi Parekh who helped me immensely in my first year in IDP. I am grateful to have been ble ssed with amazing friends; Swati Debroy for being present at all the special events in my life including the first day I landed in Gainesville and for undoubtedly being the perfect elder sister; Priyanka Roy for her fun filled company in my initial years a t U niversity of F lorida ; Subhajit Sengupta, Kiranmoy Das, Arkendu Chatterjee, Souvik Bhattacharya for pampering me like your younger sister; Dipti Patel for the invaluable help in adjusting to a new country; Debapriya Dutta, Sugata Hazra, Ranjana Sarma and Deepak Kar for always being there when I needed you. Last but definitely not the least, my heartfelt thanks to my friend and husband Sourish Dasgupta whose unconditional love and support has helped me tremendously in this memorable journey.
6 TABLE OF CO NTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 9 LIST OF FIGURES ................................ ................................ ................................ ........ 10 LIST OF ABBREVIATIONS ................................ ................................ ........................... 12 ABSTRACT ................................ ................................ ................................ ................... 15 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 17 Overview of Viruses ................................ ................................ ................................ 17 Parvoviridae Family ................................ ................................ ................................ 19 Taxonomic Classification ................................ ................................ .................. 20 Genome Architecture ................................ ................................ ....................... 21 Capsid Composition and Structure ................................ ................................ ... 23 Infectious Pathway of P arvoviruses ................................ ................................ ........ 28 Receptor Binding ................................ ................................ .............................. 29 Intracellular Trafficking and Structural Transitions Induced in the Virion .......... 32 Tissue Tropism and Pathogenicity Determinants ................................ .................... 40 Receptor Binding and Virulence Determination ................................ ...................... 43 Parvoviruses as Gene Therapy Vectors ................................ ................................ 46 Significance ................................ ................................ ................................ ............ 48 2 ANALYSIS OF MINUTE VIRUS OF MICE (MVM) RECEPTOR COMP LEXES TOWARDS UNDERSTANDING THE MECHANISMS OF TISSUE TROPISM ....... 60 Background ................................ ................................ ................................ ............. 60 Experimental Methods ................................ ................................ ............................ 62 Cell Lines ................................ ................................ ................................ .......... 62 Generation of Full and Empty MVM capsids ................................ .................... 63 Recombinant Virus Production and Purification ................................ ............... 64 Crystallization and Data Collection ................................ ................................ ... 65 Structure Solution ................................ ................................ ............................. 66 Sialylated Glycan Microarray (SGM) Preparation and Virus Screening ............ 68 Glycan Profiling of Permissive Cell Lines ................................ ......................... 70 Results ................................ ................................ ................................ .................... 70 Virus Purification and Crystallization ................................ ................................ 70 Structural Analysis of MVM Capsid Receptor Interactions ............................... 71 Binding of MVM Viruses to SGM ................................ ................................ ...... 74 Glycan Profiling of Cell Lines ................................ ................................ ............ 75
7 Discussion ................................ ................................ ................................ .............. 80 Capsid Receptor Interactions Dictate Infectious Outcome ............................... 80 Recognition of Sialic Acid Derivatives by MVM Viruses ................................ ... 85 Cellular Glycan Profiling Validates Glycan Screening Data .............................. 88 Summary ................................ ................................ ................................ .......... 92 3 STRUCTURAL CHARACTERIZATION OF H 1PV: INSIGHTS INTO CAPSID GENOME INTERACTIONS AND RECEPTOR BINDING ................................ ..... 106 Background ................................ ................................ ................................ ........... 106 Experimen tal Methods ................................ ................................ .......................... 109 Cell Lines ................................ ................................ ................................ ........ 109 Virus Production and Purification ................................ ................................ ... 109 Glycan Array Analysis ................................ ................................ .................... 110 Crystallization ................................ ................................ ................................ 112 Diffraction Data Collection and Processing ................................ .................... 112 Structure Determination ................................ ................................ .................. 114 Structural Alignment Comparison ................................ ................................ ... 118 Results ................................ ................................ ................................ .................. 119 Virus Purification and Crystallization ................................ .............................. 119 Glycans Recognized by H 1PV ................................ ................................ ...... 119 H 1PV C apsid Structure ................................ ................................ ................. 120 Structural Studies of H 1PV and Glycan Complexes ................................ ...... 124 Discussion ................................ ................................ ................................ ............ 125 Recognition of Sialylated Glycans by H 1PV ................................ .................. 125 Structure of H 1PV ................................ ................................ ......................... 127 Structural Comparison to Other Aut onomous Parvoviruses ........................... 128 DNA Binding Pocket ................................ ................................ ....................... 130 Conformational Variation Between Full and Empty Capsids .......................... 134 Cis Peptide Bonds ................................ ................................ .......................... 135 Glycan Binding Site on H 1PV Capsid ................................ ........................... 136 Summary ................................ ................................ ................................ ........ 137 4 PARVOVIRUS CAPSID DYNAMICS ASSOCIATED WITH ENDOSOMAL TRAFFICKING ................................ ................................ ................................ ...... 153 Background ................................ ................................ ................................ ........... 153 Experimental Methods ................................ ................................ .......................... 155 Virus Production and Purification ................................ ................................ ... 155 Crystallization, Data Collection and Processing ................................ ............. 156 Structure Determination ................................ ................................ .................. 157 Results ................................ ................................ ................................ .................. 159 Discussion ................................ ................................ ................................ ............ 160 5 SUMMARY AND FUTURE DIRECTIONS ................................ ............................ 166 LIST OF REFERENCES ................................ ................................ ............................. 173
8 BIOGRAPHICAL SK ETCH ................................ ................................ .......................... 203
9 LIST OF TABLES Table page 1 1 Parvoviruses, their receptors, and hosts ................................ ............................ 50 1 2 VP2 amino acid differences between MVMp and MVMi ................................ ..... 51 1 3 Forward mutations in the MVMi strain with a fibrotropic phenotype .................... 51 2 1 Data processing and refinement statistics ................................ .......................... 96 3 1 Data processing and refinement statistics ................................ ........................ 139 3 2 RMSD in C alpha positions f or the parvovirus structures ................................ 140 3 3 Amino acid sequence comparison for autonomous parvoviruses at nucleotide binding pocket ................................ ................................ ................................ .. 140 3 4 Amino acid sequence comparison in cytosine binding pocket .......................... 141 3 5 The ssDNA sequence present in H 1PV, CPV and MVMi ................................ 141 4 1 Data processing and refinement statistics ................................ ........................ 163
10 LIST OF FIGURES Figure page 1 1 Genome architecture of the parvovirus genus ................................ .................... 52 1 2 The parvovirus capsid VP structure and capsid surface topology. ..................... 53 1 3 Capsid features of Parvovirinae subfamily members ................................ .......... 54 1 4 A schematic of the life cycle of parvovirus, MVM ................................ ................ 55 1 5 Sialic acid binding site in MVMp ................................ ................................ ......... 56 1 6 Structural clustering of MVMp/i amino acid differences ................................ ...... 57 1 7 MVM capsid structures at ~3.5 resolution ................................ ....................... 58 1 8 G lycans recognized by the MVM viruses ................................ ............................ 59 2 1 Virus purification ................................ ................................ ................................ 97 2 2 Glycan binding sites on the MVMp and MVMp K/I capsids ................................ 98 2 3 Glycan binding sites on the MVMi capsids ................................ ......................... 99 2 4 Superposition of glycans on MVM capsid ................................ ......................... 100 2 5 Sialylated derivatives recognized by the MVM viruses ................................ ..... 101 2 6 Data from the SGM array ................................ ................................ ................. 102 2 7 N glycan expression on A9, EL4 T and NB324K cell lines ............................... 103 2 8 Polar and non polar glycolipid expression on A9, EL4 T and NB324K cell lines ................................ ................................ ................................ .................. 104 2 9 O glycan expression on A9, EL4 T and NB324K cell lines. .............................. 105 3 1 Virus purification, crystallization and diffraction ................................ ................ 142 3 2 Glycan array data for H 1PV ................................ ................................ ............ 143 3 3 Structure of H 1PV ................................ ................................ ........................... 144 3 4 ssDNA observed in H 1PV virions ................................ ................................ .... 145 3 5 ssDNA topology observed in members of the parvovirus genus ...................... 146
11 3 6 ssDNA protein interactions in autonomous parvoviruses ................................ 147 3 7 Cytosines observed in the full H 1PV structure ................................ ................ 148 3 8 Conformational variation between empty and full capsids ................................ 149 3 9 Location of the cis peptide bonds observed in H 1PV structure ....................... 150 3 10 Sialic acid binding site on the H 1PV capsid ................................ .................... 151 3 11 Comparison of VP2 structures of H 1PV, MVM, PPV, FPV and CPV. .............. 152 4 1 MVMp capsid dynamics on the exterior surface ................................ ............... 164 4 2 Capsid dynamics in the interior of the MVMp capsid ................................ ........ 165
12 LIST OF ABBREVIATION S AAV adeno associated ADV aleutian mink d isease v irus APS Advanced Photon Source BNL Brookhaven National L ab BPV1 bovine parvovirus 1 CAR coxsackie and adenovirus receptor CFG consortium for f unctional g lycomics CHESS Cornell High Energy Synchrotron Source CryoEM cryo electron microscopy CP capsid protein CPV canine parvovirus CsCl cesium chloride DMEM D ulbecc e agle medium DNA deoxyribonucleic acid EDTA ethylene diamine tetraacetic acid EGFR epidermal growth factor receptor EM electron microscopy FGFR1 fibroblast growth factor receptor 1 FPV feline panleukopenia virus H 1PV human tumor isolated 1 pa rvovirus HBoV human bocavirus HGFR hepatocyte growth factor receptor HIV human immunodeficiency virus HS hepar a n sulfate
13 HSPG heparan sulfate proteoglycan HSV herpes simplex virus ITR inverted terminal repeats kDa kilodalton LamR laminin receptor mRNA mess enger ribonucleic acid MVM minute virus of mice Neu5Ac N acetylneuraminic acid NES nuclear export signal NHS N hydroxysuccinimide NLM nuclear localization motif NLS nuclear localization signal NPC nuclear pore complex NS non structural NT nucleotides ORF o pen reading frame PBS phosphate buffer saline PDGFR platelet derived growth factor receptor PEG polyethylene glycol PLA 2 phospholipase A 2 PPV porcine parvovirus RFU relative fluorescence units RMSD root mean square deviation RNA ribonucleic acid RT room te mperature
14 S IA sialic acid SGM sialylated glycan microarray SDS PAGE sodium dodecyl sulfate polyacrylamide gel electrophoresis ssDNA single stranded deoxyribonucleic acid SV40 simian virus 40 TfR transferrin receptor VLPs virus like particle VP viral protei n VP1u VP1 unique N terminal region WT wild type
15 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy STRUCTURAL STUDIES O F AUTONOMOUS PARVOVIRUS CAPSIDS TOWARDS UNDERSTANDING THE MECHANISMS OF CELL RECOGNITION AND ENDOSOMAL TRAFFICKING By Sujata Halder August 2012 Chair: Mavis Agbandje McKenna Major: Medical Sciences Biochemistry and Molecular Biology Minute Virus of Mi ce (MVM) and H 1 Parvovirus (H 1PV), members of the Parvoviridae share 6 6 % capsid viral protein sequence identity, are oncotropic and utilize cell surface sialic acid (SIA) for infection. Here, MVM and H 1PV served as models to characterize the molecular determinants of cell recognition and host range for the parvovirus capsid. In addition, MVM capsid dynamics associated with endosomal trafficking were probed. MVM virus strains screened on a derivatized glycan microarray identified 9 O methylation as an a dditional component of their SIA recognition and g lycan cell profiling confirmed a role for sialylated glycans in infection Structural studies of MVM strains complexed with sialylated glycan receptors w ere used to define the role of specific amino acids in th e interaction that dictates cell recognition These residues were observed in the depression at the capsid icosahedral twofold axis and are associated with host pathogenicity determination. The structure of H 1PV determined to aid the mapping of SIA site(s), is similar to that of other parvoviruses, especially MVM, except for differences at previously defined variable loops. H 1PV capsids screened on a
16 glycan array revealed specificity for 2 3 linked SIA glycans, including the SIA Le x tumor cell mark er motif, similar to previous observations for MVM. Structural studies of the H 1PV complex ed with SIA glycans, including one with a 3 SIA Le X motif, also identified the same twofold depression as MVM, as its SIA binding site This observation indicates th at parvovirus capsids utilize common regions as determinants of cell recognition pathogenicity, and host range. Ordered ssDNA in a conserved binding site inside the H 1PV virion pointed to a potential role in packaging and capsid stability. Structural stu dies of MVM capsids at endosomal pHs showed two co nformational changes : one in the capsid interior which alter s capsid DNA interaction(s), possibly in preparation for DNA uncoating ; and the other on the capsid exterior at the fivefold axis channel region increasing its diameter possibly for VP1 N terminus externalization Information arising from these studies on cellular recognition and the specific capsid receptor interactions that dictate oncotropism could aid the improvement of the efficacy of viral v ectors based on these two viruses.
17 CHAPTER 1 INTRODUCTION Overview of Viruses Viruses are obligate intracellular parasites that are the causative agents of severe diseases and can infect all domains of life, from bacteria archaea to eukaryotes. They ha ve evolved a variety successful infection. Viruses may be of different shapes and sizes, but the basic structure called the nucleocapsid consists of the genomic material (DNA or RNA) encl osed within a protective protein shell (or capsid). Enveloped viruses incorporate lipids (and glycoproteins) from the host cell membranes during assembly as either an external envelope surrounding the nucleocapsid or internal envelope surrounding the genet ic material, while non enveloped viruses lack any lipid bilayer membrane The shapes of viruses range from helical and icosahedral to more complex structures. The packaged genome encodes for the structural capsid/viral proteins (CPs or VPs) that make up th e capsid and the auxiliary replication and regulatory proteins that are required for its propagation. Viruses may be classified on the basis of their morphology, including size, shape, capsid symmetry and presence or absence of an envelope; and biological properties, including genetic material, host range, replication strategy, mode of transmission and pathogenicity. The key initial step of virus infection is virus attachment onto cell surface receptors, followed by penetration through (non enveloped viru ses) or fusion with (enveloped viruses) the cellular plasma membrane to enter the cytosol, and culminates in the release and targeting of the viral genome to the host cell replication compartment. Many viruses with a DNA genome must enter the nucleus, wher eas RNA viruses, with a
18 few exceptions, replicate in the cytosol. A number of factors, such as virulence, host permissivity and the host immune response dictate the outcome of infection. Viruses, in particular those that undergo rapid mutation such as RN A viruses are also known to be able to switch hosts which is a major health risk. Various host defense mechanisms exist that limit viral replication but some viruses have evolved to circumvent these defenses and commandeer factors involved in host defen se or host signaling pathways as part of their infection strategy. The VPs that assemble the capsid perform a wide variety of functions required during the viral life cycle, including host recognition, internalization, intracellular trafficking, genome rep lication, capsid assembly, genome encapsidation, and progeny virion release for re infection. The high resolution three dimensional (3D) images of viruses infecting phylogenetically diverse organisms or with different replication strategies and genome stru ctures surprisingly show common structural features, implying conservation of commonly utilized structural motifs (reviewed in (162) ) The study of the virus life cycle and the intricate relationship s with their host s ha s resulted in several key developments. They have identifie d targets for the development of anti viral drugs such as the drug acyclovir that b locks transcription of H erpesvirus (HSV) DNA ; and also vaccines such as the Sabin vaccine that st imulate s the host immune system to develop long term adaptive immunity against Poliovirus (107, 245) Much of the basi c concepts and tools of molecular biology have been derived from the study of viruses, such as the discovery by Alfred Hershey and Martha Chase in 1952 that genes are composed of DNA, and the use of vir uses as sources of enzymes and vectors for protein pro duction The integration of viral genome into host genome plays a
19 major role in host evolution and provides opportunities for new viruses to emerge that are optimized for replication in the host (reviewed in (103) ) The relative ease of genetic manipulation of viruses has also provided in sights into the host cell machinery. Viruses such as H uman P apillomavirus (HPV) that induce host cell transformation, also serve as model systems to explore the role of complex signal transduction pathways in cell growth and differentiation, while viruses (e.g., Myxoma virus) that control abnormal cell growth are being developed as vectors for anti cancer therapy (284, 292) Understanding the mechanisms of various capsid mediated interactions during an infection suc h as cell surface receptor recognition, genome packaging and host immune response would provide information necessary to design virus based vectors that have specific tissue targeting capability package a therapeutic gene and have low immunogenic respon se. For example, mutating the capsid residues involved in binding to a specific receptor and inserting a peptide that recognizes a different receptor would allow the vector to bind alternate receptors or mutating the residues that interact with a neutrali zing antibody would now make the vector less immunogenic (10, 37) For the single stranded (ss) DNA Parvoviridae little is known about the structural determinants of the steps involved in infection, particularly fo r members of the Parvovirus genus. This project aims to begin to fill this gap by employing two related members Minute Virus of Mice (MVM) and H 1PV, as models to dissect the molecular mechanisms of tissue tropism and capsid dynamics associated with endos omal trafficking Parvoviridae Family The Parvoviridae consists of small (260), non enveloped viruses that package an ~4 6 kb linear ssDNA genome within a T=1 icosahedral capsid (27) The parvovirus biology is dominated by its small size that restricts it s coding capacity. They lack
20 accessory proteins that might induce resting cells to enter S phase and also lack a duplex transcription template and therefore, must wait for the host to enter the S phas e DNA replication machinery for their own replication. Although limited by their size and coding capacity and lack of any motility they can infect a wide range of hosts, including insects, rodents, dogs, and humans. The Pa rvoviridae is highly diverse and consists of both pathogenic and non pathogenic members. Taxonomic Classification On the basis of host range, this broad family is divided into two subfamilies: the Parvovirinae which infect vertebrates and include the fiv e genera Amdovirus Bocavirus Dependovirus Erythrovirus and Parvovirus ; and the Densovirinae which infect insects and other arthropods a nd subdivided into four genera Iteravirus Brevidensovirus Densovirus and Pefudensovirus This study will focus on the members of the Parvovirinae and especially the parvovirus genus. The type species of each of the Parvovirinae genera are; amdovirus: Aleutian mink Disease Virus (ADV); bocavirus: Bovine Parvovirus 1 (BPV1); dependovirus: Adeno Associated Virus seroty pe 2 (AAV2); erythrovirus: Human Parvovirus B19 (B19V); and parvovirus: Minute Virus of Mic e (MVM) (26) Members of the d ependovirus genus rely on co infection with a compl ex helper virus from a different taxonomic family (such as A denovirus (Ad) HSV HPV or V accinia V irus ) for productive infection, and in the absence of the helper virus establish a latent infection in the host that can be re activated by the introduction o f helper virus (38) replicate independently of helper viruses but requi re cellular factors expressed transiently during the S ph ase for their DNA replication. Since these viruses replicate
21 productively only in actively dividing host cell populations, lethal infections occur in fetal or neonatal hosts, or involve tissues that remain actively dividing in adult life such as cells of the gut epithelium or leukocyte lineages. Pathogenic members, such as B19 V and MVM cause severe disease in young and immunocompromised adults, while nonpathogenic members establish asymptomatic but pe rsistent infections (232) The severity of disease depends on the virus and host factors, such as age and susceptibility. The parvovirus genus cont ains four distinct subgroups : species that contains 3 clades (a) MVM, (b) Mouse Parvovirus 1 (MPV1), (c) Rat virus group that includes Rat Minute Virus 1 (RMV1), H 1 Parvo virus (H 1PV) Kilham Rat Virus (KRV), and LuIII us ; (2) an outlying Rat Parvovirus 1 (RPV1) branch ; (3) Feline Panleukopenia Virus (FPV) and Canine Parvovirus (CPV) ; and (4) Porcine Parvovirus (PPV) (293) As mentioned above, t his study utilize d MVM and H 1PV Genome Architecture The common genomic structure of pa rvoviruses consists of two open reading frames (ORFs) flanked by palindromic sequences (120 to ~550 nucleotides (NT) in length) that can fold into hairpin structures and are required for DNA replication and packaging. The terminal hairpins may be related i n sequence and structure, and are referred to as inverted terminal repeats (ITRs) (e.g. members of the dependovirus and erythrovirus genera) or may be different in sequence and structure (e.g. members of the parvovirus genus). Unlike the dependoviruses which package both strands of the ssDNA genome with equal frequency into different capsids, some parvoviruses package predominantly the negative strand which is complimentary to the mRNA (e.g. MVM, CPV, FPV, PPV) while others encapsidate strands of eithe r polarit y in equimolar (e.g.
22 AAVs, LuIII) or different proportions (e.g. BPV1) (55) This differential encapsidation is dictated by the terminal hairpins. The rep or ns ) in the complementary DNA strand encodes non structural proteins (referred to as Rep in the dependoviruses and NS in the autonomous parvoviruses) that are important for genome replication and packaging, while cap ) encodes structural viral proteins (VP) which assemble the capsid (76) The MVM and H 1PV genome are organized into two t ranscription units with promoters located at map units (m.u) 4 (P4) and 38 (P38) (Figure 1 1) (14, 78, 239) The early P4 promoter codes for the non structural regulatory proteins NS1 and NS2 derived from R1 and R2 transcripts, respectively NS1 ( 83 kDa ) is a nuclear phosphoprotein that has helicase, ATPase, endonuclease, and sequence specific DNA binding activities essential for genome amplification and progeny excision (reviewed in (70, 218) ) and is also the major mediator of cytotoxicity (12, 48, 80, 217) NS1 upregulates the P4 promoter itself in a p ositive feedback mechanism. The NS2 protein con sists of three isoforms, NS2P, NS2Y, and NS2L (23 to 28 kDa) derived from alternate splicing of R2 transcript (71) NS2 is not essential for infection in certain non mu rine permissive cells (such as transformed human cell lines) but is indispensable for the infection of natural murine hosts (51, 206) Multiple functions have been attributed to NS2 such as capsid assembly (66) DNA replication, message trans lation virus production (206, 207) and nuclear egress of progeny virions (33, 96, 200, 220) NS2 also appears to be important in viral pathogenesis and tropism, since mutat ion s in the MVM genome that resulted in increased NS2 levels contributed to host range switching by MVM strains (63, 79) and because an NS2 defective virus was unable to infect
23 newborn mice (44) In addition, NS2 enhances NS1 mediated cytotoxicity (39, 176) T he late P38 promoter is transactivated (~100 fold) by NS1 and drives the synthesis of R3 transcripts that encode VP1 and VP2 (58, 89, 295) VP1 (83 kDa) and VP2 (64 kDa) are produced by alternative splicing from the same mRNA and expressed at a ratio of 1:5, and capsid maturation results in the generation of VP3 ( 61 kDa) by post translationa l cleavage of ~20 amino acids from the N terminus of VP2 following genome packaging (76) The entire sequence of VP3 is contained within VP2 which is in turn contained within VP1 that has a unique N terminal region (VP1u) of 142 amino acids (295, 296) VP1u harbors an active phospholipase A 2 (PLA 2 ) activity required for endoso mal exit and important for infectivity (91, 102, 115, 191, 335) The cleavage of VP2 to VP3 can be mimicked by trypsin digest ion although peptide mapping indicates that this cleavage site is not the same as that in vivo The same proteolytic site is present in VP1 but is inaccessible to cleavage, suggesting different structure disposition of these common sequences (296) Capsid Composition and Structure Capsid composition The number of VPs encoded by the cap gene and used to assemble the capsid differs between members of the Parvovirinae and also depends on the capsid maturation stage. ADV and B19V contain only V P1 and VP2; AAV2 and MVM virions contain VP1, VP2, and VP3 ; and BPV1 contains VP1, VP2, VP3 and VP4 (9, 68, 76, 143, 175, 256) Sixty copies of these VPs assemble the capsid with T=1 icosahedral symmetry (53) VP1 is always the minor component while the smallest VP is always the major component in all virus capsids MVM empty capsids (devoid of DNA) contain VP1 and VP2 in a predicted ratio of 1:5, while full capsids (DNA containing) are composed of VP1, VP2 and VP3 in a ratio of 1:1: 10 (295) Empty,
24 recombinant MVM virus like particles (VLPs) that are morphologically and antigenically similar to native, empty capsids can be produced by baculovirus expression of the VP 2 in Sf9 insect cells (132) The large scal e production of the VLPs enables their exploitation as vehicles for antigen presentation in vaccine development and also for various structural and biophysical assays (16, 42, 168, 187, 260, 267, 268) The parvovir us VP is capable of performing a variety of functions during the viral life cycle ranging from cell surface receptor binding, endosomal entry and trafficking, cytoplasmic processing, nuclear entry, capsid self assembly, genome encapsulation capsid maturat ion, nuclear export of infectious virus progeny, and host immune response evasion (reviewed in (5, 6) ) The relatively small genome size has allowed the use of genetic manipulation for the structure function annotat ion of the VPs/capsid. The parvoviruses have evolved to utilize specific functional motifs in the N terminal regions of VPs such as the phospholipase A 2 ( P LA 2 ) function in the VP1u for en dosomal escape during infection (91, 102, 115, 191, 335) nuclear localization s ignal (NLS) in the VP1u and VP2 N terminal region (in dependoviruses) for nuclear entry (122, 145, 186, 279, 310, 312) and nuclear export signal (NES) in the VP2 N terminal region of certain autonomous parvoviruses for nuclear exit of progeny virions (194) The common VP1/2/3 core is involved in receptor attachment, tropism, nuclear entry, capsid assembly, and host antigenic response (5) Capsid str ucture Towards understanding the capsid structural features that dictate the multiple functions of the VP/capsid during cellular infection, the three dimensional structures for several members of the Parvoviridae family have been determined using X ray cr ystallography and/or cryo electron microscopy ( cryo EM) and
25 image reconstruction (cryo reconstruction) (3, 4, 7, 53, 119, 123, 158, 171, 177, 184, 198, 210, 214, 225, 275, 276, 306, 314, 328, 3 30) In all these structures only ~520 residues of the C terminal overlapping VP region have been observed. The signal rich N terminal extensions of the VPs, including VP1u, ~40 60 residues of VP2 and the first ~15 24 residues of VP3 have bee n un resolved except for B19 V Low copy numbers of VP1 and VP2 in the mature virions or differential conformations adopted by the VP1/2/3 N termini, which is incompatible with the 60 fold icosahedral ave raging applied during structur e determination, could result in th e lack of N terminal VP ordering (53) A conserved glycine rich po rtion in the VP2 N terminus has been modeled in the electron containing CPV and MVM virions (7, 306, 329) The un observed N terminal sequences are proposed to be pos itioned inside the capsid consistent with the location of the first N terminal residue observed in crystal structures. Cryo reconstruction studies in which the structures of empty AAV capsids containing VP1 3 were compared to capsids assembled without VP1 and/or VP2 and to infectious virions suggest that the N terminal regions of VP1 and VP2 are located in the interior of the capsid underneath the icosahedral twofold axis (170, 225, 280) For B19 V cryo EM reconstruc tion of wild type (wt) virions and empty particles and comparison to recombinant VP2 containing VLPs showed VP2 N termini exposed on the capsid surface (155, 158) The N te rminal peptides perform essential functions at different steps in the infectious life cycle as mentioned before and become sequentially exposed in response to successive cellular signals following virus internalization (35, 191, 254)
26 The structure of the ordered common VP region is highly conserved for the Parvoviridae even for members that are only ~20% identical at the amino acid sequence level, such as AAV2 and B19 V The core is composed of an eight stranded terminu helix 2A and barrel motif is also observed in other icosahedral viruses with very low sequence similarity (24) Parvovirinae structures. In the Densovirinae strand undergoes domain swapping and interacts in an antiparallel fashion with the B of the twofold re lated monomer (156, 275) barrel strands, helices form the remainder of the VP structure and contribute to capsid s urface topology. The loops are named after the strands between which they are inserted, for example, the GH loop is structure between members in the same and different genera, a nd have been defined as variable regions (VRs) 1 8 for the members of the parvovirus genus (168) and as VR I IX for members of the dependovirus genus (119) sheet forms the contiguous shell while the elaborate loop insertions form characteristic features at and around the icosahedral two three and five fold symmetry axes (Figure 1 2 C). There is (I) a depression (dimple) at each twofold axis, (II) a single protrusion (mound) at the threefold axis of members of the parvovirus genus or three separate protrusions surrounding a depression at the threefold axis in members of the amdovirus, bocavirus dependovirus, and erythrovirus genera, (III) a cylin drical channel at each fivefold axis,
27 between the two fold and fivefold depressions (Figure 1 3). Exceptions to this general surface topology are seen in the Densovirinae members that have smoother capsids as strands (53) The twofold axis (in all the parvo helix forms the wall of this depression. The floor of the twofold depression is the thinnest region of the capsid, being only one polypeptide c hain thick and contains the smallest number of intermonomer interactions (Figure 1 2C) The single threefold protrusion in members of the parvovirus genus are created from six loops (within the GH loop), two from each threefold symmetry related VP monomer; each of the three separate protrusions in the other Parvovirinae members are created by three loop regions (within the GH loop) from two VP monomers. The channel at the fivefold axis that connects the inside and the outside of the capsid is created by clu ribbons in the DE loop of five symmetry related monomers. The floor of the depression around the channel is lined by the HI loop that resembles a flower petal extending from the adjacent fivefold symmetry related monomer and forms the most ext ensive fivefold related VP contacts. A DNA binding pocket, conserved in the sheet in the interior of the capsid. Based on the differences in the surface topologies, the parvoviruses were div ided into three structural groups (225) and member s of the Parvovirinae subfamily fall into group I and III, while group II comprised of members of the Densovirinae subfamily. Group I capsids have a mound like protrusion at the threefold axis and a wider and shallow twofold depression and are comprised of members of the parvovirus
28 genus. The group III capsids have three distinct spike like protrusions at the threefold axis and narrower and deeper depression at the twofold axis and are comprised of members of the amdovirus, bocavirus, dependovirus and eryt hrovirus genera. The protrusions are more pronounced in ADV and the AAVs co mpared to B19V and H uman Bocavirus 1 (( HBoV 1 ), member of the bocavirus genus) B19 V and HBoV 1 appear to share characteristics of members of the parvovirus and dependovirus genera a nd have flatter protrusions. The f ive fold channel is conserved in the group I and III capsid structures. In B19 V and HBoV 1 the wall between the twofold and fivefold depressions are at almost the same height as its threefold protrusions, and form a continu ous rim surrounding the canyon. Mutagenesis, biochemical and structural studies have shown that the VRs play important roles in the viral life cycle, such as receptor recognition, tissue tropism, pathogenic outcome and antigenic response (reviewed in (5, 6) ) Infectious Pathway of Parvoviruses The parvovirus capsids serve as robust delivery vehicles that penetrate two cellular barriers, the plasma membrane and nuclear membrane, and deliver the encapsidated genome to the nucleus for replication (Figure 1 4 ) Virions are stable in the presence of lipid solvents and survive exposure to pH between 3 and 9 and incubation at ~60C for one hour (132, 168, 192, 213, 334) However, biochemical studies demonstrate that the capsid is metastable, and undergoes minor conformational transitions following receptor attachment and internalization to display functional motifs required to maneuver through the endocytic pathway to the nucleus f or genome replication (35, 178, 191) Several research groups have characterized various steps in the life cycle of the parvoviruses using biochemical and cell based assays but there is very little 3D information on the capsid structural dynamics involved in the infectious
29 process. The parvovirus structur es discussed in the previous section represent only low energy state conformations. The data presented in chapter 4 aims to fill this dearth by structural characteri zation of the capsid dynamics associated with endocytic trafficking. The various steps in the life cycle of the parvoviruses have been reviewed in the following sections. Receptor Binding Recognition of cell surface receptor by a virus is the first step of infection and a key paramete r of tropism and pathogenesis. The primary receptors and co receptors utilized for cellular recognition and internalization, respectively, during parvoviral infection are listed in Table1 1. For most of the parvoviruses only th e glycan component of the glycoprotein or glycolipid receptor is known, hence the focus of this study will be on the glycoconjugate receptors. Glycans (carbohydrate polymers) are the major constituents of the cell surface and may be conjugated to cell surf ace proteins or membrane lipid head groups to form glycoproteins and glycolipids, respectively, or are present as glycosaminoglycan (GAG) chains attached to proteoglycans (222) The immense variability of the glycan structures expressed between different species and also between different tissues in the same species creates vast diversity in viral tissue tropism. The most common glycoepi topes contain terminal sialic acid ( SIA ) or sulfated oligosaccharide motifs of GAGs ( e.g., heparan sulfate (HS)). Sialic acids comprise a family of structurally diverse monosaccharides derived from neuraminic acid, a nine carbon sugar. There are more than 50 natural analogues of SIA that result from modifications to the carbohydrate backbone of which N acetylneuraminic acid (Neu5Ac) is the most common In general, the amino group attached to C 5 is N acetylated (Neu5Ac), N glycolylated (N glycolylneuramini c acid; Neu5Gc) or removed
30 (deaminoneuraminic acid; Kdn ). The hydroxyl groups may be free, esterified (acetylated, lactylated, sulfated, phosphorylated), or etherified (methylated), thus leading to increased chemical diversity (149) Structural heterogeneity a mong sialylated glycans can arise from variations in the glycosidic linkage positions associated with the Neu5Ac residues, which may be linked 6, principally to galactose (Gal) or N acetylgalactosamine (GalNAc) residues, or linked 9 t o adjacent SIA residues (11, 309) The carboxyl group at position 1 confers a negative charge on SIA under physiological conditions. They commonly occur at the non reducing termini of oligosaccharide chains that are attached to glycoconjugates. HS is a highly sulfated, linear polysaccharide built up from disaccharide units of glucosamine and uronic acid. HS proteoglycan s (HSPG) are the major components of extracellular matrix. The sulfate groups impart high charge de nsity to HS and contribute to non specific virus binding through electrostatic interactions (183) Binding to HS often involves positively charged patches on capsid surface (219, 223) Ho wever, in some cases, HS acts as a specific receptor as in the case of HSV and Foot and Mouth Disease Virus (FMDV) (109, 273) A few examples of neutral virus receptor glycans, such as histo blood group epitopes ha ve also been identified. B iochemical studies utilizing neuraminidase and proteinase K treatment have shown that MVM requires SIA containing glycoproteins for cell recognition and infection (78, 188) SIA is also an important attachment factor for other parvoviruses infecting different species, such as BPV1 (28, 142, 297) H 1PV (10) PPV (34) AAV1 (324, 325) AAV4, AAV5 (148, 270, 315) CPV and FPV although t he SIA CPV and SIA FPV interactions are not essentia l for infection in certain cell types (19) CPV and FPV use
31 their respective transferrin receptor (TfR) on canine and feline cells for infection (19, 136, 230) Eliminating the glycosylation site on the canine TfR e xpanded its ability to bind both FPV and CPV (116, 117, 229) H SPG serves as the cell surface receptor for AAV2 and the closely related AAV3b (290) AAV6, which is closely related to AAV1 is able to utilize HS or SIA depending on the cell type being infected (324, 325) AAV 9 utilizes terminal galactose as a receptor (23, 272) Bovine AAV (B AAV) utilizes gangliosides for transduction (264) 4 linked N acetyl glucosamine found on gp96, for cellular transcytosis (85) B19V binds to the glycolipid 5 1 integrin and Ku80 as co receptor s for cellular entry (43, 205, 316, 317) The protein receptors/co recep tors that have been identified for the parvoviruses, such as the V 5 integrin and growth factor receptors are listed in Table 1 1 and reviewed in detail in Halder et al (126) The recognition sites for the glycoconjugate receptor s, conformed onl y on the assembled capsid have been identified using biochemical, molecular and str uctural approaches for several parvoviruses, such as MVM, CPV, FPV, B19V, AAV2, AAV5 and AAV1/ AAV6 The depression at the twofold axis, also known as the dimple was identif ied as the SIA binding site on MVM, and this is proximal to the CPV and FPV determinants of SIA binding to erythrocytes (188, 303) (Figure 1 5 ). Significantly, the residues determining in vitro tropism and in vivo p athogenicity for MVM, such as K368 and I362 are localized in the vicinity of this SIA binding pocket. The CPV TfR attachment site has been mapped to the wall between the two and fivefold depressions (118, 135, 136) Structural studies utilizing cryo reconstruction mapped the globoside
32 receptor attachment site for B19 V to the depression at the threefold axis (57, 154) The AAV2 HS binding site has been mapped to basic residues located at the wall of the protrusions surrounding the threefold axis (160, 178, 219, 223) B inding of B19V virions to globoside triggers the externalization of VP1u (35) Structural studies on the AAV2 HS complex report ed structural rearrangement of the HI loop which was proposed to be related to the opening of the fivefold channel (178) Protease susceptibility assays did not detect any changes in the CPV capsid structure upon TfR binding (213) The 3D structure determination studies of wt and mutant MVM capsid glycan receptor complexes presented in chapter 2 attempts to provide a structural understanding of the role of the specific amino acid residues in receptor recognition and specificity Data presented in chapter 3, identifies the sialic acid motifs recognized by H 1PV and examines the structural interactions involved in receptor attachment. This information allows for a comparison of the mechanisms of host recognition between the two members of the parvovirus genus and will also aid in the development of these viruses as vectors with specific tissue tropism. Intracellular Trafficking a nd Structural Transitions Induced in the Virion MVM CPV, AAV2 and AAV5 are the best characterized viruses with respect to endosomal trafficking MVM will be the main focus of this section and will be compared to CPV, AAV2 and AAV5 when variations are observed in their pathways Although the parvovirus es use different cellular receptors and the attachment sites vary, most of them are internalized by receptor mediated endocytosis, predominantly by dynamin dependent clathrin mediated endocytosis (34, 93, 231) (Figu re 1 4 ). Alternate mechanisms of entry might be used ; for example, AAV2 utilizes the clathrin independent carriers / GPI anchored protein enriched endosomal compartment (CLIC /
33 GEEC) pathway and AAV5 utilizes the caveolar endocytic pathway (17, 215) Some of the AAVs have been demonstrated to use selective receptor mediated vesicle transcytosis to penetrate barrier cell layers (83) AAV2 integrin binding and clustering has been shown to activate cell signaling path ways that enhance virus uptake (150, 261) Particle to infectivity ratio of most parvoviruses seem s to be high, implying that most particles that bind and enter the cell fail to navigate to the nucleus. Endosomal t rafficking of parvovirus capsids is reported to be a slow and rate limiting process in viral transduction in several cell types (94, 127, 128, 131) Following uptake, parvoviral particles are trafficked through the endocytic pathway and delivered to early endosomes, late endosomes, recycling endosomes and lysosomes (20, 86, 87, 130, 191, 231, 253, 288) CPV capsids appear to remai n associated with the TfR in recycling endosome s (230, 231, 288) This suggests that in vitro biochemical and structural studies characterizing the role of the capsid in the infectious life cycle be conducted in the presence of the receptors (if known). Confocal microscopy studies have shown that in addition to trafficking through the endocytic pathway, AAV2 can localize to the Golgi and endoplasmic reticulum prior to nuclear entry (144, 227) and similarly AAV5 has been d etected in the Golgi compartment (17, 18) Further studies are required to characterize the role of capsid host cellular factor(s) interactions in facilitating these alternative trafficking routes. Structural studie s of capsid receptor complexes in chapter 2 an d chapter 3 are the first step towards fulfilling this aim CPV and MVM move through the cytoplasm to the nucleus by microtubule mediated processes, as shown by the disruption of the microtubule network with no codazole or microinjection of antibodies
34 against the microtubule based motor protein (129, 133, 191, 231, 253, 261, 286, 288, 311, 313) MVM and PPV have also been shown to be dependent on an intact microfilament ne twork as disruption of the actin filament network with cytochalasin B/D and latrunculin treatment blocks cytoplasmic trafficking (34, 253, 261) Endosomal acidification is known to be essential for the infection of all parvoviruses, since lysosomotropic drugs such as bafilomycin A1 and chloroquine, and NH 4 Cl that interfere with the endosomal pH, block infection (20, 21, 92, 128, 231, 253, 311) The low pH environment of the e ndosome triggers the externalization of the VP1u for its PLA 2 activity essential for endosomal exit while the capsid remains intact (65, 91, 102, 115, 191, 283, 335) The PLA 2 activity is reported to induce a tran sient pore formation or permeability change in endosomal membranes rather than complete endosomal lysis as smaller dextrans (molecular weight of 3000) coendocytosed with CPV were released from the vesicles while large dextrans (molecular weight of 10,000) and sarcin were retained in the vesicles (231, 287) Treatment of CPV particles in vitro with the acidic pH of the endo some induce s VP1u exposure (287) However, for MVM, cleavage of VP2 N termini to VP3 is a prerequisite for VP1u externalization at low pH in vitro (65, 100) And, in the case of AAV low pH alone is not sufficient to mediate VP1u extrusion suggesting a requirement for yet un identified cellular factors in this capsid transition (170) The VP1u externalization in MVM CPV and AAV2 may be mimicked in vitro by exposur e to heat or urea (65, 67, 170, 312) H eat treatment of AAV virions to 65 C and empty capsids to 75 C is able to mimic endosomal conditions required to release the VP1u (170) For MVM VP1u evic tion is only seen in full capsids but not in empty particles when heated in vitro however a recent report observed VP1u
35 extrusion in empty capsids by in vivo immunofluorescence staining and in situ hybridization (65 191) The differences observed in the in vitro and in vivo studies for MVM, suggests that artificial heat treatment experiments cannot directly reproduce the in vivo stimuli but have enabled biochemical and structural characterization of capsid dynamics (65, 67, 170, 312) A combination of factors, such as pH, receptor binding or interaction with other host factors is implicated to play a role in these capsid conformational changes. It was earlier believed that th e VP1u of B19 V is always exposed on the capsid surface but recent studies show that it is true only for B19 V VLPs and the VP1u of B19 V virions is buried but can be exposed in vitro by heat or low pH treatments (254, 257) Although the extrusion of VP1 N terminus is a common feature among all parvoviruses and is essential for infectivity, the exact mechanism that triggers this event is still unclear. As has been reported for other PLA 2 domains, the VP1u is predicted to be helical in nature (170) Mutagenesis and biochemical studies indicate that the channel at the fivefold axis serves as the VP1u extrusion route for AAV2 and MVM (32, 100, 101) Such dynamic flex ibility has been reported for other viruses, such as P oliovirus where externalization of VP1 N terminus upon receptor binding is essential for infectivity (36, 108) and the myristoylated VP4 polypeptide found on th e inside surface of the virion is released early in infection while the particles remain intact (179) Structural studies of AAV 8, CPV and MVM capsids (in chapter 4 of this study) under pH conditions that mimic the environment encountered in the endocytic pathway have detected conformational changes which will be discussed in detail in chapter 4 (208, 274)
36 Following endosomal release, there is likely further processing in the cytoplasm because microinjection of virions with/without exposed N termini does not confer a nuclear translocation phenotype (2 31, 279, 287, 288) It has been shown that p hosphorylation of capsid surface exposed tyrosines followed by ubiquitination targets AAV2 and AAV5 capsid s for proteosomal degradation (332, 336, 337) Thus, co adminstr ation of proteasome inhibitors enhance s AAV2 and AAV5 transduction efficiency (92, 94, 203, 331 333) However, for MVM, CPV and PPV, treatment with proteasome inhibitors was detrimental to infection, and specificall y, the chymotrypsin like activity of proteasome appeared necessary for infection (253, 255) PPV capsid proteins were ubiquitinated early during infection but no evidence of particle ubiquitinylation or degradation was observed for CPV and MVM (34, 253, 255) T he exact mechanism of genom e translocation into the nucleus following endosomal release is unclear for the parvoviruses, as are the determinants of capsid uncoating. Th e VP1 N termini in MVM (and the VP1/2 N termini in AAV s ) contains nuclear localization signals (NLSs) required for nuclear trafficking (122, 145, 186, 279, 312, 323) Theoretically, the 260 diameter capsids of th e parvoviruses should be able to pass through the nuclear pore complex (NPC) intact, and this has been seen for CPV capsids during nuclear entry and for newly synthesized MVM capsids during viral egress (194, 200, 31 3) Other studies report that virus uncoating occurs during endosomal trafficking and the viral DNA enters the nucleus devoid of viral protein (130, 191) There is also evidence indicating that MVM might not enter the nucleus through the NPC, and an alternative nuclear entry strategy involving partial disruption of the nuclear membrane by host caspase 3 has been proposed (59 61) Similarly, for AAV2,
37 majority of the studies i ndicate that genome uncoating occurs in the nucleus (20, 146, 261, 279, 326) however there are reports that uncoating may occur before or during nuclear entry (190) MVM is the best characterized member with respect to capsid requirements of genome release. Studies with MVM suggest that the low endosomal pH and divalent to uncoating thus exp end of the viral DNA to polymerases (67) The pH dependence of the simultaneous externalization of VP1 u and viral DNA from intact capsids without particle disassembly has been observed in vivo (191) Similarly, in vitro heat or urea treatment induces VP1u exposure and genome release from intact capsids (65, 100, 312) Mutational studies of the residues at the base of the fi vefold channel control genome uncoating thus implicating the cylinder as the likely genome extrusion portal (74, 100) Following genome delivery, NS1 (for autonomous parvoviruses) or Rep78/68 (for AAVs) initiated replication proceeds via a series of duplex intermediates in a rolling hairpin mechanism primed by replication origins at each end of the linear genome (70) The replicated genome is transcribed, and then translated in the cytoplasm and the VPs are reported to be transported back into the nucl eus in the form of intermediates for capsid assembly. Mutagenesis and structural studies suggest that VP trimers are the stable assembly intermediates for MVM capsids (236, 246, 248) The NLS in the VP1 u and the str uctural nuclear localization motif barrel region of the MVM capsid, target the expressed VPs to the nucleus (185, 186) In the case of AAVs, a transiently expressed 23kDa protein called assembly activating protein
38 (AAP ) targets VPs into the nucleolus where capsid assembly is proposed to occur (280, 281) M utagenesis studies have identified charged residues involved in symmetry interface interactions (6) and residues in the HI loop to be important for capsid assembly (88, 122) MVM and AAV2 are the best characterized parvoviruses with respect to genome packaging. empty capsid using energy provided probably by NS1 or Rep52/40 and requires the terminal hairpins as packaging signa ls (31, 54, 69, 95, 164, 327) It has been shown that is outside the virions and covalently attached to NS1 or Rep ( 69, 241) M ut a tional analysis on MVM implicate the fivefold channel as the portal for genome encapsidation in addition to VP1u exposure and genome release (32, 101, 240, 323) For AAV2, in addition to the fivefold channel, the DE and HI loop are also shown to be involved in genome packaging (32, 323) Crystal structures of MVM and CPV infectious virions have shown the presence of ordered genomic ssDNA inside the capsid with the bases interacting with side chains of the capsid amino acid residues (7, 306) The DNA capsid interactions involve residues that surround the interior capsid surface near the two and threefold axes and mutagene sis of these residues impaired infectivity (247) Unlike in AAVs, there is no ordered DNA in empty capsid structures in MVM or CPV which suggests that the capsid genome interactions in MVM and CPV are specific (7, 119, 177, 210, 214, 329) T his explains the low tolerance of autonomous parvoviruses, such as MVM and H 1PV as compared to the AAV, to replacement of wild type genome sequence with a foreign sequence for vector production (161, 174) In chapt er 3, the ssDNA capsid interactions
39 in the context of H 1PV will be analyzed in an effort to provide an understanding of genome packaging which would also aid in the engineering of improved H 1PV based gene therapy vectors. Post capsid assembly and genome packaging, the virions must exit the nucleus and traffic to the cell surface for a second round of infection. The exit mechanism for the parvoviruses is best characterized for MVM. The VP2 N termini possesses a phosphoserine rich nuclear export signal (NE S) that interacts with the exportin molecule chromosome region maintenance 1 protein (CRM1) to traffick packaged virions out of the nucleus (193, 194) Full capsids are released from the cell with all their VP2 N te rmini intact but capsid maturation in the extracellular environment or during re entry into a new host cell involves the proteolytic removal of ~25 amino acids from most of the VP2 N termini to generate VP3, thus removing the nuclear export signal (100, 194) It has been shown that t he VP2 N terminus is exposed in DNA containing capsids prior to any treatment and is susceptible to trypsin digestion However, in empty capsids and VLPs this sequence is sequestered but heat treatment can trigger its exposure (65, 132, 191) It has been suggested that following DNA packaging a structural shift occurs in the capsid leading to the exposure of VP2 N termini. The VP2 to VP3 cleavage primes the pH dependent externalization of the VP1u and genome as shown by the studies on the fivefold channel base mutants of MVM (100) suggesting that the channel serve s as the site for VP2 externalization for cleavage to VP3 in autonomous parvoviruses that undergo a maturation step. Inte restingly, cryo electron EM and X ray crystallographic studies of B19 V reveal that its fivefold channel is narrower, consistent
40 with the observation that B19 V does not undergo VP2 to VP3 cleavage and that the VP2 N termini is exposed on the capsid surface (3, 57, 155) From the available parvovirus capsid structures at physiological pH it is understood that the N terminal extensions of VP1 and VP2 cannot be externalized without structural rearrangement. Although it i s appreciated tha t capsid dynamics play a central role in parvovirus cell binding, entry, trafficking, genome release, and egress following assembly, few studies have addressed the three dimensional visualization of these structural transitions. Chapters 2 and 3 focus on identifying the capsid regions involved in receptor binding and chapter 4 characterizes the structural transitions encountered in the endocytic pathway. Tissue Tropism and Pathogenicity Determinants The parvovirus genus includes viruses that infect many different species, including mice, cats, dogs, and pigs, and share medium (~50%) to very high (~98%) sequence identity. For several members of the parvovirus genus, there are distinct tissue tropism and pathogenic differences between highl y homologous strains. FPV infects cats but not dogs (136, 230, 235, 304) CPV 2 emerged as a host range variant of FPV in 1978 and could infect dogs but not cats (136, 304, 30 5) In 1979, CPV 2 was replaced by an antigenic variant called CPV 2a that could infect both dogs and cats (137, 228, 233, 234) Since the 1980s, several antigenic variants of CPV 2a, designated CPV 2b and CPV2c th at share the same host range have been identified (233) CPV can bi nd to both canine and feline TfR while FPV can only bind feline TfR. However, alteration of the glycosylation site (N383K) in the apical domain of canine TfR enabled it to bind to FPV (228, 229) Several PPV strains have been distinguished by pronounced differences in their tissue tropism and in vivo pathology, although their VP2 proteins are
41 ~99% identical (25) NADL 2 is the attenuated vaccine strain of PPV that is non pathogenic, but can be lethal i f injected in utero while Kresse, IAF A54, IAF 76, and NADL 8 are the virulent strains (25) In the case of ADV, the highly pathogenic ADV Utah 1 strain replicates poorly in cell culture, whereas the non pathogenic ADV G strain replicates p ermissively in Crandell feline kidney (CrFK) cells. MVM, the focus of this study, has two well characterized strains, the prototype strain (MVMp) and the immunosuppressive strain (MVMi) that are reciprocally restricted despite sharing 97% sequence identity and being serologically indistinguishable MVMp was originally isolated from a murine adenovirus stock and replicates efficiently in A9 mouse fibroblasts, whereas MVMi, which was recovered from an EL 4 T cell lymphoma replicates in mouse T lymphocytes and hematopoietic precur sors (294) However, both v iruses can replicate efficiently in the human transformed cell line, NB324K ( SV40 transformed human newborn kidney fibroblast cells ) The viruses differ by only 14 amino acids (of 587) in VP2, all of which are ordered in the crystal structure except residu e 10 (Table 1 2). Twelve of the 13 ordered residues are located throughout the primary amino acid sequence but are clustered on the assembled 3D capsid at and around the icosahedral twofold axis and on the shoulder of the threefold protrusions (Figure 1 6) These regions also differ between other highly homologous parvoviral strains, such as wt CPV and wt FPV, or wt CPV and its host range mutants. Mutagenesis and selective plaque assays map the MVM in vitro tropism determinant to residues 317 and 321 in VP 2 (15, 112, 196) and forward second site mutations conferring fibrotropism to MVMi to residues 399, 460, 553, and 558 when
42 either 317 or 321 are mutated (7) (Table 1 3). The switch to fibrotropism for MVM i requires both an equivalent region of the MVMi capsid protein gene and a segment of the non structural protein genes that re sults in an increase in NS2 levels (63, 79) The MVMi strain, inoculated by the oronasal route in newborn mice is pathogenic and replicates in endothelia, neuroblasts (244) and hematopoietic stem cells, an d in adult Severe Combined Immunodeficient (SCID) mice, causes acute leucopenia (269) while MVMp infection is asymptomatic (163) However, the MVMp intravenously inoculated into SCID mice, evolved into virulent variants which carried one of the three mutations (V325M, I362S, and K368R) in the VP capsid protein, which caused a systemic lethal d isease when introduced by the oronasal route (259) Residues 362 and 368 differ between MVMp and MVMi, but these variants remain ed fibrotropic in vitro without any genetic changes in the capsid gene that control MVM tropism in vitro The rei ntroduction of the highly virulent MVMp mutants into SCID mice caused lethal leu k openia, reflecting the pattern of MVMi infection (189) Significantly, t he residues determining in vitro tropism (317 and 321), in viv o pathogenicity (325, 362, 368) conferring fibrotropism on MVMi (399, 460, 553, 558), and those associated with the development of leukopenia (321, 551, 575) show local surface structure variability between the strains and are localized in the vicinity of the sialic acid binding pocket at the twofold axis (188, 189) (Figure 1 5 1 6 and 1 7) For PPV, CPV and FPV tissue tropism and pathogenicity determinants have also been mapped to this region (reviewed in (5) ) thus highlighting the util ization of common autonomous capsid regions for similar functions Also, the residues controlling CPV and
43 FPV host range and virulence show local structural variations on the capsid surface and colocalize with the pr edicted transferrin receptor binding sit e (135, 136) The co lo c alization of structurally variable tropism/pathogenicity determinants with receptor recognition sites suggests that the disparities in phenotypes involves differential MVM glycan receptor in teraction and/or utilization Previous cell binding studies have shown that both MVM strains compete for binding to either cell type, arguing against restriction at the level of a cell surface receptor (282) Hybrid cells from fibroblasts and lymphocytes can propagate both strains of MVM (294) suggesting that the block is due to the lack of a differentiation dependent cellular factor in restrictive cells. It was suggested that the point of restriction was post entry and conversion of genomic ssDNA to replicative form (RF) intermediates, but prior to viral genome transcription (134) It has also been suggested that the block is prior to RF DNA replication and likely due to a block in uncoating (242) However, glycan array screening and affinity assays have demonstrated (as detailed in the next section) that MVMp, MVMi and the virulent MVMp mutants show differences in SIA receptor intera ctions and specificity (188, 189, 209) The tissue tropism and pathogenicity of a virus may be regulated by virus receptor interactions or post entry virus host cellular factor interactions. The work in chapter 2 is aimed at structurally characterizing the role of minor MVM ca p sid variations in regulating receptor recognition and dictating cell recognition and pathogenicity. Receptor Binding and Virulence Determination The MVMp and MVMi strains have served as ideal model s to address questions related to the role of cell surface receptor recognition in determination of tissue tropism and pathogenicity between highly homologous parvovirus strains. These two viruses
44 have been shown to require cell surface SIA for infect ion as well as to compete with each other for cell binding even in the restrict ive cell line (77, 282) Thus, to identify possible differences in SIA structures or other carbohydrates recognized by the MVM viruses, the interactions of VLPs of MVMp, MVMi, and the three MVMp virulent mutants, MVMp I362S, MVMp K368R, and MVMp I362S/K368R (MVMp K / I) were studied on a glycan microarray with 180 different naturally occurring and synthetic glycans (209) All 3 sialylated glycans linked to a co mmon 4GlcNAc (3' SIA LN LN or 3' SIA (LN ) 2 ) 1 4GlcNAc (3' SIA LN LN LN or 3' SIA (LN ) 3 ), 3) 3Gal 3)GlcNAc ( SIA Le x Le x Le x or 3' SIA (Le x ) 3 ). In addition MVMi showed expanded recognition to multisialylated glycans with terminal SIA ) 3 8Ne 4Glc (GT3), and 4Glc (GD2) (Figure 1 8) Interestingly, the virulent MVMp K368R mutant that contains the MVMi residue type also recognized GT3, indicating that the residue 368 plays a role in the multi sia lic acid recognition observed for MVMi. Notably, other than the extended recognition of multi sialic acid by MVMi and MVMp 3 linked terminal SIA and not 6 linked SIA SIA LN motif is commonly found in N and O linked glycoproteins, which are abundant in most cell surface glycoproteins. The recognition of the 3' SIA L e x motif, which is a known tumor cell marker by all MVM viruses, explains
45 (139, 151, 152, 182) Interactions with the 8 glycans that are abundant in brain glycoproteins likely mediate (138, 262, 263) Also, GD3 gangliosides and 9 O acetylated form of GD3 are present on these cells (105) Based on cell binding assays and neuraminidase treatments, in vitro cytotoxicity in A9 fibroblasts, glycan array screening and BIAcor e surface plasmon resonance (SPR) studies, it was shown that the virulent MVMp mutants had a lower affinity for the sialic acid component of the receptor, than MVMp or MVMi (188, 209, 259) The double recombinant mu tant MVMp I362S/K368R showed the lowest affinity but increased pathogenicity. This suggests that the affinity/avidity of the interactions with the SIA containing receptor modulates parvovirus virulence. Similarly, in Polyomavirus single amino acid changes that reduced SIA receptor affinity increased viral spread and disease severity (22) There are other virus es including Influenza Virus and V irus where a single or few amino acid changes modulate recep tor binding and virulence. For I nfluenza virus, a single amino acid mutation in the receptor binding pocket of hemagglutinin causes the virus to switch specificity from Neu 5 3Gal to Neu 5 6Gal terminal residues (250) and capsid mutations affecting SIA binding V irus (141) Since, the K368R mutant bound to GT3 with lower affinity than MVMi, but the MVMp bind to GT3, it indicated that more than a single amino acid substitution and the other 14 amino acids that differ between MVMp and MVMi near the receptor binding site, may 8 linked sialic acids by MVMi. Also, the
46 virulent MVMp mutants remained fibrotropic in vitro suggesting a difference in virulence pattern between the MVMp mutants and MVMi. In a recent study to characteriz e the adaptive host range of MVM, capsid adaptations at the SIA binding site (mutations at residues 334, 384, 554, 578) resulted in the emergence of a host range variant, called F1 which was able to infect rat fibroblasts (Figure 1 6) (99) This observation suggests that in addition to controlling virulence and pathogenic outcome, the MVM receptor binding pocket also plays a role in adaptations to a new host. Chapter 2 presents data on the structural verification of th e binding sites of the glycan receptors recognized by MVMp, MVMi and the MVMp virulent mutants in an effort to define the specificity and role of the capsid glycan interactions in the infectious life cycle. Also, the glycan composition of the three cell lines that are permissive or restrictive for infection by MVMp or MVMi was analyzed in chapter 2 to correlate the natural expression of the glycans identified in the glycan array screening to the differences in tissue tropism. Parvoviruses as Gene Therap y Vectors The autonomously replicating parvoviruses predominantly propagate in rapidly dividing cells due to their dependence on both cellular proliferation factors expressed transiently during the S phase and the differentiation of host cells. These virus es for example, MVMi and H 1PV, were first isolated from tumor tissue and were then believed to be oncogenic (300) But, it was later observed that the rodent parvoviruses display oncopreferential cytotoxic activity in vitro and also possess an oncosuppressive potential, inhibiting the formation of spontaneous and chemical or virus induced tumors in vivo and in vitro (226, 251) The cytotoxicity has been attributed, in part, to NS1 (39) The intracellular oncotropism is mainly based on enh anced P4 promoter activity via
47 binding of ATF/CreB, Ets and SP1 transcription factors in transformed cells (110, 237) Since transformed cells are intrinsically deficient in antiviral mechanisms, MVM 's oncotropism m ight be related to its failure to mount an antiviral type I interferon (IFN) response in transformed cells, similar to other oncotropic viruses (120, 147) Thus, it is already known that specific host cellular facto rs are involved in the oncotropic properties of MVM and H 1PV but the mechanisms underlying tumor cell recognition which is the first step towards a successful infection, and oncosuppression are not yet fully understood. It has been shown that MVM viruses recognize the 3' SIA L e x tumor cell marker and in chapter 2 the interactions of this glycan with the MVM capsid surface residues were analyzed (209) Chapter 3 presents data towards identifying the glycan receptor that explains H 1PV's recognition of tumor cells and structural characterization of the capsid glycan interact ions. The rodent parvoviruses, such as MVMp H 1PV and LuIII c an persistently infect their natural hosts, do not integrate their genome into cellular chromosomes, are non pathogenic in adult animals, and efficiently infect human cell lines (reviewed in (30) ). These characteristics make these viruses attractive candidate vectors for anticancer gene therapy, particularly for cytoreductive and immunogene therapy approaches to target tumor cells. H 1PV based vectors are currently being tested for antineoplastic effects in preclinical studies (1, 172) T he AAVs have low toxicity, no known pathology, broad tropism and can establish long term transgene expression thus making them als o an excellent choice for gene therapy vectors. Currently clinical trials are underway with AAV vectors packaging therapeutic genes for the treatment of several diseases,
48 di (204, 211, 285) Significance The structures of the capsid VPs that are assembled into virions control various steps in the infectious life cycle, such as, host cell receptor recognition, endosomal trafficking, capsid assembly, and genome encapsidation. For the homologous MVM strains minor structural variations due to the amino acid differences in the capsid VPs result in altered virus SIA receptor interactions and utilization which results in pronounced differences in their host range, tissue tropism and pathogenic outcomes of infection. Several research groups have characterized some of the steps in the infectious life cycle but there is very littl e structural information on the capsid dynamics or the capsid regions involved in the structural transitions associated with the life cycle This project aims to utilize MVM and H 1PV as models to decipher the capsid dependent mechanisms for host cell reco gnition and structural transitions required for a successful infection In chapter 2 structure determination studies on the capsid receptor complexes for the MVM strains and mutants was conducted to map the receptor binding site on the MVM capsids and pro vide further insight into the differential mechanisms of receptor recognition by these highly homologous strains and single site mutants that results in differences in tissue tropism and pathogenicity Cellular glycan profiling analysis on the cell lines p ermissive for the MVM viruses was conducted to verify the expression of the glycans that had been shown to interact with MVM by glycan array screening and SPR studies Also, MVM viruses were screened on an array consisting of SIA derivatives to provide fur ther chemical information on the identity of the MVM glycan receptor MVM
49 serves as a tractable model for studying emerging viral pathogens allowing us to track subtle capsid changes that confer a disease phenotype as well as adaptations to a new host. Ch apter 3 presents data on the identification of the glycan receptor for H 1PV and structural characterization of the capsid glycan interactions and also capsid genome interactions. Chapter 2 and 3 provide information on the capsid regions utilized for recog nition of tumor cells by MVM and H 1PV and its role in dictating onco tropism. So, chapter 2 focuses on understanding the mechanism of differential tropism of MVMp (fibrotropic) and MVMi (fibrotropic and neurotropic), and chapter 2 and 3 utilize MVM and H 1 PV as models to examine the mechanism of their oncotropism. A detailed understanding of MVM and H 1 PV viruses for tumor cell targeted gene delivery applications. In chapter 4, t he dynamic character of the parvovirus capsid that allows for the sequential exposure of domains required to mediate successive interactions with the host, was probed by structure determination studies of MVM capsids at pHs (pH 6.0, 5.5, and 4.0) that mimic the endocytic pathway Th e parallel study of these two members of the family and the homologous strains provides a comprehensive view of the mechanisms of infection and enables us to identify the common or divergent links. This study stems from the need to structurally characteriz e the mechanisms of host cell recognition and capsid dynamics associated with endosomal trafficking to fill in the information gaps in this field.
50 Table 1 1. Parvoviruses, their receptors, and hosts Virus Receptors Co receptors Host References AAV1 6 N linked sialic acid Human (325) AAV2 HSPG a In FGFR1 b HGFR c LamR d Humans (8, 13, 153, 243, 289, 290) AAV3 HSPG a HGFR c LamR d FGFR1 b Humans (8, 29, 180) AAV4 2 3 O linked sialic acid NHP e (148) AAV5 6 N linked sialic acid PDGFR f Humans (84, 148, 270, 315) AAV6 6 N linked sialic acid, HSPG a EGFR g Humans (319, 324, 325) AAV8 LamR d NHP e (8) AAV9 Galactose LamR d Humans (8, 23, 272) Bovine AAV Gangliosides, Chitotriose Bovine (8 5, 264) ADV ADV binding protein Mink (106) BPV1 Sialic acid Glycophorin A Bovine (28, 142, 297) B19V Erythrocyte P antigen Human (43, 205, 316, 317) MVM Sialic acid Rodent (77, 209) CPV and FPV Sialic acid Transferrin receptor Cat, dog (77, 136, 230) PPV Sialic acid Swine (34, 77) a HSPG=hepar a n sulfate proteoglycan; b FGFR1=fibroblast growth factor receptor1; c HGFR=hepatocyte growth factor receptor; d LamR=37/67 kDa laminin receptor; e NHP=non human primate; f PDGFR=platelet derived growth factor receptor; g EGFR=epidermal growth factor receptor This table was adapted from Halder et al (126)
51 Table 1 2. VP2 amino acid differences between MVMp and MVMi 10 160 232 317 321 362 366 368 388 402 410 440 455 551 MVMp Ser Leu Val Thr Gly Ile Val Lys Ser Ser Lys Asn Ala Ala MVMi Gly Ser. Ile Ala Glu Val Met Arg Ala Asn Arg Asp Thr Val Position N S S SA S S S S S B S S B B *S=Surface; B=Buried; SA=Solvent accessible; N=not ordered in structure Table 1 3. Forward mutations in the MVMi strain with a fibrotropic phenotype Site directed mutation A317T A317T A317T E321G E321G E321G Selected forward mutation D399G D399A D553N A317T S460A Y558H
52 Figure 1 1. Genome architecture of the parvovirus genus. The single stranded, negative sense DNA genome of the parvovirus genome with the terminal folded hairpin structures is shown. The two viral promoters, P4 and P38 are shown by rightward arrows, and the mature, cytoplasmic transcript s R1, R2, and R3 are displayed belo and AAA denoting their polyadenylated tails. ORFs specifying the viral gene products, named on the right are displayed in different shades according to their reading phase, and their spliced out introns are represented by the thin lined carets. The dashed box denotes the VP1u region involved in entry functions.
53 Figure 1 2. The parvovirus capsid VP structure and capsid surface topology. (A) Ribbon diagram of MVMp VP2 superimposed on a semi transparent surface representation helix A, and the first N terminal residue modeled (39) and the C terminal residue (587) are labeled. The approximate icosahedral twofold (filled oval) threefold (filled triangle), and fivefold (filled pentagon) axes are shown. (B) Conserved secondary structure superposition of VP for one member from every genus in Parvovirinae subfamily is shown: ADV (red),HBoV (yellow), AAV2 (blue), B19V (green) and M VMp (orange). Atomic coordinates for AAV2, MVMp, and B19V were obtained from RCSB protein database (PDB accession numbers 1lp3, 1z14, and 1s58, respectively). The ADV and HBoV images were generated from pseudo atomic coordinates built into cryo reconstruct ions (123, 198) The N terminus (N), C terminus (C), variable regions (VRI IX, VR1 8), DE, and HI loops are labeled. The boxed in all parvovirus VP structures determined to date. (C) Surface representation of MVMp used to illustrate the topological features of the parvovirus capsid surface. The image is depth cued ( blue cyan green yellow orange red ) to show regions at the shortest radial distance to capsid center in blue and those at the furthest radial distance in red A viral asymmetric unit is depicted by a black triangle bound by a fivefold axis and two threefold axes divided by a line drawn through the twofold axis. (A) and (B ) were generated using the PyMol program (82) and (C) was gener ated using the UCSF Chimera program (238)
54 Figure 1 3. Capsid features of Parvovirinae subfamily members. Depth cued (blue red yellow white) capsid surface representation of representative members of the five genera of Parvovirinae viruses, an d one member from the Densovirinae is shown. The virus and the genus to which it belongs are labeled. A viral asymmetric unit (white triangle) is shown on the AAV2 image. A horizontal scale bar (100) for diameter measurement is shown on the right hand sid e and vertical color bar for radial distance () from the center of the particle is also shown on the right hand side. These figures were generated using the UCSF Chimera program (238) The coordinates used were obtained as described in the legend o f Figure 1 2. This figure was adapted from (123)
55 Figure 1 4 A schematic of the life cycle of parvovirus, MVM. (1) Virus binding to receptor followed by receptor mediated endocytosis, (2) NES cleavage and exposure of the NLS of VP1 during trafficking, (3) Nuclear entry, (4) Nuclear translocation of assembly intermediates, (5) Viral DNA enc apsidated into preformed empty capsids (6) Nuclear exit of newly formed virions. Figure inset shows the nuclear transport sequences identified in the VP proteins of MVM. NLS: Nuclear Localization Signal; NLM : Nuclear Localization Motif; NES: Nuclear Expor t Signal; NPC: Nuclear Pore Complex.
56 Figure 1 5 Sialic acid binding site in MVMp. (A) Surface representation of a n assembled MVMp capsid showing the SIA binding site (region enclosed in red box) in context of the whole capsid with the 60 VP2 monomer s in different colors (B) Surface representation of the c lose up of the enclosed region in red box in (A) showing the depression at the icosahedral twofold axis of the MVMp capsid The reference VP2 monomer (ref, in yellow ), and icosahedrally related twof old (2f, in magenta), threefold (3f1 and 3f2, in cyan and orange ), and fivefold (5f, in green ) monomers are shown The surface positions of residues I362 K368 and D399 are highlighted in red, blue and grey, respectively. The S IA model (colored according t o atom type ; carbon, nitrogen and oxygen in green, blue and red, respectively ) is shown inside a 2F o F c map (dark grey mesh) the icosahedral twofold axes is shown as a filled black oval. This figure was generated using the Pymol program (82)
57 Figure 1 6. Structural clustering of MVMp/i amino acid differences. (A) Close up view of the MVMp icosahedral twofold axes, with the positions of surf ace MVMp/i amino acid differences colored and labeled: blue for p/i differences grey for forward fibrotropic mutations and black for host range switch mutations (B) Close up view of the MVMi icosahedral twofold axes, with the residues colored and labeled as in panel (A), except that the host range mutations are not shown A viral asymmetric unit is shown in the panels (A) and (B). White ovals represent approximate icosahedral twofold axes. This figure was generated using the PyMol program (82)
58 Figure 1 7 MVM capsid structures at ~3.5 resolution. (A) Ribbon diagram of VP2 of MVMp, MVMi, MVM p K368R, MVM p I362S, and MVM p K/I sh owing the side chain variations at the icosahedral twofold axis. (B F) The side chains of the differing amino acid residues are shown as sticks and colored according to atom type. The 2F o F c generated using the PyMol program (82)
59 Figure 1 8. Glycans recognized by the MVM viruses. A schematic representation of the glycans recognized by the MVM viruses is shown. Figure inset depicts the glycan symbols.
60 CHAPTER 2 ANALYSIS OF MINUTE VIRUS OF MICE ( MVM ) RECEPTOR COMPLEXES TOWARDS UNDERSTANDIN G THE MECHANISMS OF TISSUE TROPISM Background MVM serves as an ideal model for studying the capsid determinants of tissue tropism and pathogenicity dictated by receptor interactions due to several reasons (i) the small ssDNA genome size and the simple T=1 icosahedral capsid, formed from the assembly of three VPs, enables genetic manipulation and structural studies (ii) it is known that MVM SIA recognition is essential for infection and the depression at the twofo ld axis has been mapped as the SIA binding site on the MVMp capsid, (iii) two highly homologous strains of MVM, MVMp and MVMi have pronounced differences in tissue tropism and in vivo pathogenicity, (iv) the pronounced differences are associated with one o r two VP amino acid differences localized at the SIA binding site and result in local structural variations that alter MVM SIA receptor interactions and utilizations (v) virulent MVMp variants with one or two VP changes associated with altered receptor sp ecificity and affinity which confers a pathogenic phenotype have also been observed. In addition, capsid adaptations at the SIA binding site resulted in a host range variant which also makes MVM a useful model for the study of emerging pathogens. The cry stal structure of wt empty MVMp and empty MVMi capsids, MVMp VLPs, MVMi and MVMp virions as well as for VLPs of virulent MVMp mutants, MVMp I362S, MVMp K368R, and MVMp I362S/K368R, have been determined to ~3.5 resolution by X ray crystallography ( (7, 168) and unpublished data). The structures of the mutant viruses are similar to each other and to MVMp and MVMi, except for local main and side chain differences resulting in the loss or gain of intra subunit interac tions at or close to the mutated site (Figure 1 7). A chain of weak intra and intersubunit amino acid
61 interactions involving the MVMp/i differing residues from the wall (E321) toward the floor (R368, D399) of the dimple are observed in MVMi, which are not possible in MVMp. A hydrogen bonding interaction between residues R368 and D399 was observed for MVMp K368R and MVMi but is not possible in MV M p. Significantly, mutation of D399 to G or A in MVMi, which abolishes this hydrogen bonding interaction, confers fibrotropism to MVMi (7, 79) It has also been proposed that MVM 's tissue tropism and pathogenicity is likely controlled by an alteration of the surface charge in the vicinity of the twofold depression (168) These structural differences between MVM viruses colocalize with the SIA bindi ng site of MVM as well as the tropism and pathogenicity determinants for other autonomous parvoviruses such as CPV, FPV, PPV and ADV Prior biochemical studies on CPV, FPV and MVM tissue tropism and pathogenicity differences suggested that a block in init ial cell receptor attachment is not a restriction to infection, but interactions post entry with intracellular factors was a determinant. However, the colocalization of tropism/pathogenicity determinants that result in local structural variability with rec eptor recognition sites close to the twofold depression and the shoulder of the threefold axis on the capsids suggests a role for receptor interaction in these phenotypes. The structures for the native capsids provide s the 3D platform for comparison wit h resulting structures of capsid glycan complexes for identification of the glycan binding site(s) as well as any capsid conformational changes resulting from the glycan interaction. Previous cell binding, glycan array screening and SPR studies provided in formation on the specificity and affinity of the sialylated glycans but did not define the role of the capsid surface amino acids in th ese interaction s Here, structural studies of
62 the MVM viruses complexed with the glycans identified in the glycan array s creening were conducted to characterize the nature of the altered MVM SIA receptor and o ther carbohydrate interactions that confer an infectious phenotype and the recognition of transformed tumor and neuronal cells. Cellular glycan profiling analysis on th e three cell types differentially infected by the MVM strains: EL4 T lymphocytes, A9 fibroblasts and NB324K transformed fibroblast cell line, was conducted to correlate the MVM interacting glycans identified in the glycan array screening with their express ion on the different cell types and to investigate the possibility that differences in the glycan composition of these cells results in the differential tropism between MVMp and MVMi Finally, t o understand the potential for recognition of different modif ications of sialic acid by MVM, three types of capsids, VLPs wt empty particles (Empties) and DNA packaged virions (Fulls) were screened on a newly developed sialylated glycan microarray (SGM). The three types of capsids were analyzed on the SGM to inves tigate if they exhibit differences in SIA recognition and affinity and also to justify the use of VLPs in lieu of infectious virions to study MVM capsid glycan receptor interactions. Experimental Methods Cell Lines A9 cells are the ouabain resistant deriva tive of the HGPRT defective cell line A9 and are perm issive host cells for MVMp. EL4 T is an adherent variant of the T cell lymphoma line EL4 and is permissive for MVMi. NB324K is a clone of SV40 transformed human newborn kidney fibroblast cells and is per missive for both MVMp and MVMi. For the glycomic profiling, all the three mammalian cell lines were cultured in inactivated fetal calf serum and they were grown to a density of 1x10 7 cells. Sf9 insect cells were
63 grown in suspension culture 900 II SFM media (Gibco/Invitrogen Corporation) supplemented with 1% Antibiotic antimycotic (ABAM) at 27C. Generation of Full and Empty MVM capsids A9 ouabr11 cells were grown in spinner culture in DMEM containing 5% fetal bovi ne serum and antibiotics to a density of ~6x10 5 cells/ml. The cultures were infected with predetermined titers of transfection derived parvovirus MVM (MVMp) stocks (GenBank accession number J02275) and expanded until cell counts indicated a progressive ris e in numbers of dead cells. The cells were harvested by centrifugation, washed in phosphate buffered saline without Ca 2+ or Mg 2+ (PBS; Invitrogen, Carlsbad, CA), pelleted, and resuspended in 10 ml of TE8.7 ( 50 mM Tris HCl pH 8.7 0.5 mM EDTA) per liter of i nfected cells. Following three cycles of freeze thaw at 37C the pellets underwent repeated centrifugation at 2,000 rpm (800 g ) to clarif y extracts, and were stored at 20C. For purification, 6 rpm in a Sorvall SS34 rotor at 4C and then f loated on top of a 6 ml iodixanol (OptiPrep; Axis Shield, Oslo, Norway) step gradie nt (1 ml 55% and 2 ml 45% in TE 8.7, followed by 2 ml 35% and 1 ml 15% in PBS plus 1 mM MgCl 2 and 2.5 mM KCl). Samples were centr ifuged at 35,000 rpm for 18 h at 18C in a Beckman SW41 rotor. Fractions were collected from the bottom of the gradient and full and empty particle concentrations were assessed by hemagglutination and also sodium dodecyl sulfate polyacrylamide gel electrop horesis (SDS PAGE) followed by staining with Coomassie Blue (Sigma). The capsid integrity was checked by negative stain electron microscopy (EM) (Figure 2
64 spotted onto a 4 00 mesh carbon coated copper grid (Ted Pella, Inc., Redding, CA, negatively stained with NanoW for 1 min twice, blotted dry and viewed on a Hitachi 3000 electron microscope MVMi was recovered in the same way as MVMp except that the transfection was done in NB324K cells. For SGM screening, all viruses were finally buffer exchanged into 1XPBS. Recombinant Virus Production and Purification Recombinant baculovirus constructs e xpressing the VP2 of wt MVMi wt MVMp and the virulent MVMp mutants (MVMp I362S, MVMp K368R, and MVMp I362S/K368R) that self assemble into VLPs were constructed as described (188) Sf9 insect cells grown in suspensi on culture at 27C were infected with a titered baculovirus construct at a multiplicity of infection (m.o.i) of 5.0 plaque forming units per cell. Following incubation at 27C for 72 h, the cells were spun down for 20 min at 1 500 rpm and then the pellet r esuspended in lysis buffer (50 mM Tris HCl pH 8.0, 150 mM NaCl, 0.2% Triton X 100, 10 mM MgCl 2 ). The MVMp and MVMi VLPs were purified based on published procedures (132) with some modifications. The virus capsids we re released from the cells by three cycles of rapid freeze thaw with the addition of Benzonase (Merck KGaA, Germany) after the second cycle. The cellular debris was removed by low speed centrifugation in a JA 20 rotor (10,000 rpm, 15 min, 4C ). The superna tant was diluted with TNET buffer (50 mM Tris HCl pH 8.0, 100 mM NaCl, 1 mM EDTA, 0.2% Triton X 100) and pelleted through a 20% (w/v) sucrose cushion by ultracentrifugation in 70Ti rotor at 45 000 rpm for 3 h at 4C The resulting pellet w as resuspended ov ernight at 4C in TNETM buffer (25 mM Tris HCl pH 8.0, 100 mM NaCl, 1 mM EDTA, 0.2% Triton X 100, and 2 mM MgCl 2 ). The resuspended sample was then
65 subjected to a low speed 2 000 rpm spin to remove particulate material and further purified by ultracentrifug ation on a sucrose step gradient (5 40% w/v in TNETM) at 35 000 rpm for 3 h at 4C in a Beckman SW41Ti rotor. A visible blue fraction containing VLPs, sedimenting at ~20 25% sucrose, was extracted and dialyzed against TE buffer (25 mM Tris HCl pH 7.5, 1 mM EDTA). Next, CsCl was added to the dialyzed sample to a final density of 1.40 g/cm 3 and then subjected to equilibrium centrifug ation in a Beckman SW41Ti rotor at 35 000 rpm for 24 h at 4C Visible virus fractions were extracted and extensively dialyzed into 1XPBS for SGM screening studies or into 10 mM Tris HCl pH 7.5, 150 mM NaCl for crystallization screens. The concentration of the viruses were estimated from optical density measurements (assuming an extinction 10 mg/ml using Ultrafree 100 kDa cut off centrifugal filter units (Millipore, Billerica, MA). The purity and integrity of the vir us capsids were monitored using SDS PAGE and negativ e stain EM respectively (Figure 2 1). Crystallization and Data Collection Crystallization drops were setup using the hanging drop vapor diffusion method (199) with VDX 24 well plates and siliconized cover slips (Hampton Research, Laguna Niguel, CA, USA). Crystals were grown under the same con ditions previously used for MVMp VLP s (132) with the virus at a concent ration of 5 10 mg/ml in 10 mM Tris HCl pH 7.5, 150 mM NaCl buffer with polyethylene glycol (PEG) 8000 (1% w/v), 8 mM CaCl 2 .2H 2 0 and 150 mM NaCl as precipitants. The glycans 3' SIA (Le x ) 3 3' SIA (LN) 3 and GT3 were provided by Core D of the Consortium for Functional Glycomics (CFG) ( http://www.functionalglycomics.org/ ) and resuspended in 10 mM Tris H Cl pH 7.5 to give a concentration of 10 mg/ml. The co
66 virus solution (10 s against 1 ml of reservoir solution at RT Crystallization drops containing virus alone were also setup in preparation for glycan soaking experiments if the co crystallization trials did not yield crystals amenable to diffraction. These drops ( 4 mg/ml) ml of reservoir solution at RT. Certain virus glycan co crystallization trials yielded no useful crystals, and in these cases, VLP crystals were soaked with the resp ective glycan for 1 2 h prior to data collection. The virus glycan complex es for which diffraction data was collected are summarized in Table 2 1 X ray diffraction data was collected from crystals incubated in cryoprotectant solution containing 10 mM Tris HCl pH 7.5, 150 mM NaCl and 8 mM CaCl 2 .2H 2 0 with 10% PEG 8000 and 30% glycerol for 30 s and flash cooled in liquid nitrogen vapor. Diffraction data for the various capsid receptor complex crystals w ere collected at t hree beamline facilities, APS (Advanced Photon Source), BNL (Brookhaven National Lab) and CHESS (Cornell High Energy Synchrotron Source). Data was collected with a crystal to detector distance of 300 mm, an oscillation angle of 0.3 per image and exposure time s of 30 45 s per image. The reflect ions were indexed and integrated with the HKL2000 suite of programs (224) and scaled and merged with SCALEPACK (224) The space group for the crystals was determined to be C2. The data collection and processing statistics are su mmarized in Table 2 1. Structure Solution The diffraction intensity data set s were converted to structure factor amplitudes using the TRUNCATE program from CCP4 (Collaborative Computational Project,
67 Number 4) (62) M ) for all the complex data sets was calculated to be ~3.5 3 Da c orresponding to solvent content of 6 5 % (195) The C2 unit cell contains two half particles with different orientations in the crystallographic asymmetric unit, as described for t he previously solved MVMi and MVMp VLP structures (7, 168, 184) The MVMp VP2 VLP structure coordinates (PDB accession no. 1 Z 14 ) (168) was used as the phasing model to initiate molecular replacement using the CNS program (46, 258) Iterative cycle s of model refinement were performed using the simulated annealing, energy minimization and individual temperature factor (B factor) refinement options in the CNS program while applying strict 60 fold non crystallographic symmetry (NCS) operators (46) Five percent of the total data set was par titioned for monitoring of the refinement process (45) The refinement cycles were alternated with manual model building using the COOT program (97) into the sig ma weighted averaged Fourier electron density maps (2F o F c and F o F c maps where F o represents the structure factors for the capsid glycan complex data and F c represents the calculated structure factors from the MVMp VLP model ) that were generated in the C NS program using a molecular mask while applying strict 60 fold NCS operators (46) The coordinate files for the glycan molecules were obtained from the HIC Up server (166) and the geometry restraints and dictionary file s were generate d using the subroutine phenix.elbow from PHENIX (2) T he glycans models were docked into the F o F c density using interactive rigid body rotations and translations in COOT (97) The topology and dictionary files generated for the glycans in PHENIX (2) were then used for subsequent refinement in CNS (46) The refinement process was deemed to have converged when there was no further improvement in the agreement betwe en the observed F o and
68 calculated F c structure factors ( R factor where R factor = o | |F c o |)100 ). The quality of the refined structures was analyzed using COOT (97) and MOLPROBITY (56) The refinement statistics are given in Table 2 1. The figures were generated using the program PYMOL (82) Sialylated Glycan Microarray ( SGM ) Preparation and Virus Screening Terminally sialylated glycans with various sialic acid modi fications were synthesized as previously described (277, 278) Briefly, a s ynthetic strategy that combines the bifunctional fluorescent tag 2 amino (N aminoethyl) benzamide (AEAB) and the one pot three enzyme sialyl ation reaction was utilized for generating the sialylated glycans The GAEAB (Glycan AEAB conjugates) with terminal galactose residues were used as precursors and included AEAB conjugates of lactose, lacto N neo tetraose (LNnT), lacto N tetraose (LNT) and asialo galactosylated biantennary oligosaccharide (NA2), which were prepared from natural glycans and purified prior to the multi enzyme sialylation step. Aliquots (50 g to 1 mg) of the GAEABs were subjected to enzymatic sialylation in a combinatorial fas hion. The addition of different modified ManNAc, ManNGc, or Man precursors in the reactions allowed formation of corresponding terminal Neu5Ac, Neu5Gc, or Kdn derivatives, respectively ( these SIA derivatives were introduced in chapter 1) The products were purified by HPLC and the structure of each sialylated glycan was confirmed by MALDI TOF and HPLC analysis. 77 sialylated structures were generated incorporating 16 different terminal sialic acids on 4 different underlying structures that were tagged with a fluorescent linker. Each glycan was quantified based on its fluorescence and printed on to N hydroxysuccinimide (N HS ) activated glass slides in replicates of n =4 to generate the SGM. The array also
69 included three controls (LNnT, NA2, and Man5), correspond ing to chart ID numbers 78 80 The MVM samples (buffer exchanged into 1XPBS) were screened on the SGM in collaboration with Core H of CFG. The printed slides were washed and blocked with 50 mM ethanolamine in 0.1 M Tris HCl buffer (pH 9.0) for 1 h and pr ior to screening the slides were rehydrated for 5 min in TSM buffer (20 mM Tris HCl pH 7.4 150 mM NaCl, 2 mM CaCl 2 and 2 mM MgCl 2 ) The virus samples were diluted with Binding Buffer (TSM buffer plus 1%BSA and 0.05% Tween 20) to give a final volume of 70 100 l ( 0.05 mg/ml ) which was added to the slide and incubated at room temperature for 1 h The slide was then washed with Wash Buffer (TSM buffer plus 0.05% Tween 20) and an anti MVM capsid antibody (Tatt 2; polyclonal from rabbit) was added at a dilutio n of 1:5 000. The slide was then washed again with Wash buffer and incubated with Cy5 labeled goat anti rabbit IgG at 5 The slide was scanned with a Perkin Elmer ProScanarray microarray scanner and for Cy5 fluorescence, the wavelengths 649 nm (Excit ation) and 670 nm (Emission) were used. The scanned images were analyzed with the ScanArray Express software to determine the average relative fluorescence units (RFU) and standard deviation (S.D) of the four replicates. To analyze the results, all glycans were ranked according to their signal to noise (S/N) ratio by dividing their mean RFU from four replicates by the mean background generated in the control wells lacking sialylated glycans Variation within the 4 replicates was assessed as the coefficient of variation (%CV), which was calculated as 100 x S.D/Mean. A ny value with a %CV of >30, was considered unacceptable as reported in other studies (278) 4 Glc;
70 4 GlcNAc. Glycan Profiling of Permissive Cell Lines A9 fibroblasts (infected by MVMp but not MVMi), EL4 lymphocytes (i nfected by MVMi but not MVMp) and NB324K transformed cells (infected by both MVMp and MVMi) were grown in suspension to a density of 110 7 cells/ml The c ells were harvested by centrifugation at 250g for 5 min, washed with 1XPBS three times and then pelle ted. The cell pellets were resuspended in 1 ml of 1XPBS, transferred to a m icr ofuge tube, and pelleted at 800 g. The supernatant was aspirated, and the cell pellet s w ere purify and analyze the glycans by mass spectrometry was carried out in collaboration with Core C of CFG as described in North et al (216) Results Virus Purification and Crystallization T he purity and integrity of the MVM capsids used in this study ( Full MVMp, Empty MVMp, MVMp VLP, Full MVMi, Empty MVMi, MVMi VLP, MVMp I362S VLP, MVMp K368R VLP and MVMp I362S/K368R VLP) were verified by SDS PAGE (Figure 2 1( 1 )) and negative stain EM (Figur e 2 1( 2 )), prior to cry stallization and SGM screening. The SDS PAGE a nalysis showed that the full MVMp and MVMi capsids contained VP1 (83 kDa), VP2 (64 kDa) and VP3 (61 kDa); empty MVMp and MVMi capsids contained VP1 and VP2; and all the VLPs contained onl y VP2. The negative s tain EM analysis showed the full and empty capsids an d VLPs to be intact. Co crystals for the MVMp 3' SIA (Le x ) 3 complex and MVMp K/I 3' SIA (LN) 3 complex and VLPs were obtained in approximately 3 4 weeks and they grew to dimensions of 0.3 x 0.2 x 0. 1 mm
71 Structural Analysis of MVM Capsid Receptor Interactions Diffraction data sets were obtained for co crystals of MVMp 3' SIA (Le x ) 3 complex and MVMp K/I 3' SIA (LN) 3 complex. For the MVMp 3' SIA (LN) 3 MVM i 3' SIA (LN) 3 MVMi 3' SIA (Le x ) 3 and MVMi GT3 complexes, the VLP crystals were soaked with the respective glycans and the diffraction data was collected (Table 2 1). The co crystals of the MVM viruses with glycans or MVM crystals soaked with glycans diffracted X rays t o ~3.3 to 3.9 resolution ( Table 2 1). The crystals were isomorphous to those of wt MVM i and MVMp VLP for which structures are available (7, 168) The completeness and quality of the complex data sets was consistent with those used to solve t he crystal structures of MVMp and MVMi. Residues 39 to 587 of the VP2 were built into the MVM electron density maps. The averaged density maps were not interpretable beyond N terminal residue 39 of VP2, as was previously reported for the MVMp VLP structure (168) A comparison of all the MVM receptor complex structures showed th at all the glycan receptors, whether linked 8, utilized the same receptor binding pocket at the icosahedral twofold axis identified previously (188) on both the strains (Figure 2 2 and Figure 2 3 ). In all the complex structures, the glycans were mo deled into the F o F c difference density map Although longer sialylated gl yc ans were utilized for these structural studies, all the carbohydrate components of a particular complexed glycan were not observed in the crystal structure. For the capsid gl ycan complexes, there was variability in the size of the oligosaccharides that were observed even for the complexes where the similar glycan was used (Figure 2 2 C and Figure 2 3C ) The MVMp 3' SIA (Le x ) 3 and MVMi 3' SIA (Le x ) 3 complex data allowed for the visu alization of a longer ordered oligosaccharide ( 3) : tetrasac c hharide) in the pocket at the icosahedral twofold axis as compared to when only sialic acid was
72 soaked into the MVM capsids or the other complex data sets that were analyzed in this study (188) (Figure 2 2A and Figure 2 3A ). For the MVMp 3' SIA (LN) 3 MVMi 3' SIA (LN) 3 and MVMp K/I 3' SIA (LN) 3 SIA LN, terminal SIA SIA Gal were ordered, respectively (Figure 2 2 C Figure 2 3C and Figure 2 2E ). For the MV Mi GT3 complex only densities for the terminal trisaccharides, ( SIA ) 3 was observed and densities for the other two sugars, 4Glc were not ordered (Figure 2 3E ). In these complexes, the ring structure of the ordered carbohydrate molecules were placed with at this sigma level. At lower si gma contour levels the map extends over some of the side groups but their exact conformation could not be assigned In the MVMp 3' SIA (Le x ) 3 MVMi 3' SIA (Le x ) 3 and MVMp 3' SIA (LN) 3 complexes where longer oligosaccharides were ordered, the conf ormation of the oligosaccharide reduced the degrees of freedom and allowed the modeling of some side groups. The side groups such as N acetyl, glycerol or carboxyl, that could not be assigned were modeled in the best possible conformation, i.e ., to avoid clashes with other atoms and maximize hydrogen bonding and ionic interactions. Comparative analysis of the MVMp VLP and MVMi virion structures (168) with the MVM capsid glycan complex structures did not reveal any detectable conformational changes upon receptor binding. In the MVMp 3' SIA (Le x ) 3 complex structure, the glycan exte nds from the pocket at the twofold depression to wards the wall of the threefold protrusion that faces the pocket (Figure 2 2A and B ) The glycan interacts (contact distance of 2.4 4.5 ) with the residues M243, N323, K368, R375, E394, Y396, T397, W398, D 399 T401, Y558 and V575 The glycan contact region contains residues that differ between MVMp and
73 MVMi, as well as the residues associated with in vitro tropism and pathogenicity, e.g., 362, 368, 399, 558 575 etc. In the MVMp 3' SIA (LN) 3 complex, the gly can extends from the pocket to the wall of the threefold protrusion in a manner similar to that observed for the 3' SIA (Le x ) 3 glycan in the MVMp 3' SIA (Le x ) 3 complex structure but in this case the glycan follows the contour of the pocket wall and only SIA LN is observed in the structure (Figure 2 2 C and D and Figure 2 4A ) For the MVMi 3' SIA (Le x ) 3 complex, the glycan lies further away from the twofold pocket or the wall of the pocket and interacts with the wall of the threefold protrusion that faces the p ocket (Figure 2 3A and B ) This glycan confo r mation was different from that observed in the MVMp glycan complexes, described above (Figure 2 4C) Residues N323, R368, R375, Y378, E394, T397, Y558, K563 and V575 interact with th e 3'SIA(Le x ) 3 glycan on the MVMi capsid The MVMi 3' SIA (LN) 3 complex data showed the terminal SIA binding deeper into the twofold pocket compared to the MVMp 3' SIA (Le x ) 3 MVMp 3' SIA (LN) 3 and MVMi 3' SIA (Le x ) 3 complex structures, described above (Figure 2 3C and D ). The residues invol ved in this interaction are E 552, Y558 and I578. The GT3 binds along the wall of the twofold pocket interacting with residues E321, N323, M366, R368, R375, E394 Y396, T397, D399, Y558 and V575 on the MVMi capsid (Figure 2 3E and F ). The binding pattern of 3' SIA (Le x ) 3 3' SIA (LN) 3 and GT3 glycans on the MVMi capsid is different wherein 3' SIA (Le x ) 3 binds closer to the wall of the threefold protrusion that faces the pocket, the 3' SIA (LN) 3 glycan binds deep into the twofold pocket, and GT3 binds along the wall of the twofold pocket (Figure 2 4B) On the MVMp K/I capsid the 3' SIA (LN) 3 glycan binds further down in the twofold pocket as compared to the MVMp or MVMi glycan
74 complexes and interacts with the residues D218, K241, R349, R368 and D553 (Figure 2 2E and F ). Binding of MVM Viruses to SGM T hree types of MVMp and MVMi capsids ; VLPs empty particles and virions were screened on the SGM For the particular strain, MVMp or MVMi, the binding profile of the VLPs was similar to that of the full and empty particl es 3 and 6 linked sialylated glycans and three asialoglycans as controls, but MVMp and MVMi 3 linked sialylated derivatives which is consistent with the previous glycan array screening (Figure 2 5 and Figure 2 6 ) (209) MVMp and MVMi showed a preference for bindin g to SIA that was methylated at C 9 and not C 8. Out of a total of 4 glycans that were derivatized with Neu5Ac9 O methyl SIA MVMp and MVMi 3 linked SGM 23 ( 2 1 1 1 4Glcitol ) and SGM 55 ( 2 1 4 1 1 1 1 1 1 1 4GlcNAcitol ) glycans. The other two glycans, SGM 8 ( 2 6 1 1 1 4Glcitol ) and SGM 40 ( 2 6 1 1 1 1 1 1 1 1 4GlcNAci tol ) 6 linked and were not recognized by the MVM viruses. All the MVMp viruses also showed slight binding to glycans with the 9 O acetylated (SGM 52 : 2 1 1 1 1 4Gl 1 1 1 1 4GlcNAcitol ) and the 9 O lactoyl ated (SGM 60 : N 2 1 1 1 2 1 1 1 1 1 4GlcNAcitol ) SIA derivative (Figure 2 6A B and C ) MVMi viruses showed the same profi le as the MVMp viruses, except that they did not bind to 9 O acetylated (SGM 52) and the 9 O
75 lactoyl ated (SGM 60) SIA derivative (Figure 2 6D, E and F ) The binding profile for the genome containing viruses, especially MVMp virions showed some non specific binding with high %CV to other sialylated glycans, as compared to the empty capsids (Figure 2 6A) In addition, all the MVM viruses bound to the biantennary glycan, SGM 48 ( 2 3 1 1 1 3( 2 3 1 1 1 1 1 4GlcNAcitol ) on the SGM array. The affinity ranking of the derivitized glycans recognized by MVMp was in the order 9 O methylated monosialylated (SGM 23) > 9 O methylated biantennary (SGM 55) > 9 O acetylated O lactolylated SIA biantennary (SGM 48) (Figure 2 6) The affinity ranking for the MVMi viruses was 9 O methylated monosialylated (SGM 23) > 9 O methylated biantennary ( SGM 55) > SIA biantennary (SGM 48) (Figure 2 6) The glycans that were recognized by the MVM viruses had different SIA modifications but they all had 4GlcNAc ) as the core structure (consistent with previous glycan array screening result s) and were derived from either LNnT or NA2 precursor glycans. Also, consistent with the previous glycan array screening studies, none of the MVM viruses bound to the Neu5Gc or K dn sialylated derivatives. Glycan Profiling of Cell Lines T hree cell lines, A9 fibroblasts (infected by MVMp) EL4 T lymphocytes (infected by MVMi) and NB324K transformed cells (infected by both MVMp and MVMi) were subjected to glycomic profiling by MALDI MS analysis. The profiling provided information on the carbohydrate compositio n of the glycans present on these cells but not the linkages ( 3 or 6 ) present in the glycans However, SIA SIA (multisialylated) can only be linked 8 or 9 to each other. T he mass spectr a
76 obtained from the peptide N glycosidase F released and p ermethylated N glycans from the three cell lines are shown in Figure 2 7 For the N glycan (glycan that is covalently attached to protein at asparagine (Asn) residues by an N glycosidic bond and consists of a common core sequence ( 4Gl Asn ) fraction for all three cell lines, a full complement of oligomannose ( Man residues are attached to the core ) as well as complex type N glycans (multi branched glycopeptides containing outer chains of SIA galactose (Gal) or N acetylgl ucosamine (GlcNA c)) with bi tri or tetra antennary (branches) structures were observed. Hybrid N glycans ( two branches from the core, one that terminates in Man and the other that terminates in a sugar of the complex type ) were not observed. The N glyc ans on these cell lines terminate d in Man, Gal, GlcNAc or SIA Also, a ll the N linked glycans have GlcNAc at their reducing end ( which is attached to protein) which is the most common type of N glycan linkage that has been reported. The presence of bisec ting GlcNAc (the GlcNAc in complex N glycans is also observed for all three cell lines. The majority of the sialylated N glycans observed are branched. For the low mass N glycans, there are two (biantennary) to three (triantennary) branches, and as the molecular mass increases, there is an increase in the number of branches instead of an increase in the chain length (Figure 2 7 ) All the three cell lines also possessed core fucosylated sialoglycans (i.e., fucose attached to GlcNAc at the reducing end) with the A9 cells showing the least expression of these glycans. Only in the NB324K cells, fuc ose is linked to GlcNAc in the antennae as part of the 3' SIA Le x motif ( 3) ) ( e.g., m/z=3142.4, m/z= 3952.3 etc ; enclosed in red box in Figure 2 7 C and F) The glycan composition for the A9 and EL4
77 T cells were very similar for the low molecular mass ( Figure 2 7 A and B) but for the medium mass t here were no similar glycans present ( Figure 2 7 D and E) T he NB324K cell line expresses more glycans than A9 and EL4 T, especially in the medium mass fraction ( Figure 2 7 F). The low mass glycans present on the NB324K cells are similar to that expressed on the A9 and EL4 T cells ( Figure 2 7 A, B and C) but the medium mass fraction of the NB324K cells shares only one common glycan (m/z=3603.5 ) with the A9 cells ( enclosed in green box in Figure 2 7 D and F) and one common glycan (m/z=3776.5) with the EL4 T ce lls (enclosed in blue box in Figure 2 7 E and F) The profiles for the polar glycolipids and non polar glycolipids are shown in Figure 2 8 For the ease of identification the common names (if known ) of the identified glycolipids are included in Figure 2 8 The glycolipids present in the spectra were; GM3 (m/z = 855.5) GM2 (m/ z = 1100.7) GM1a (m/ z = 1304.9, major species ) LST1d (m/ z = 1304.9 minor spe c ies), GT3 (m/z=1577.8), GD1a (m/ z = 1665.9, major species), GD1c (m/ z = 1665.9, minor species), GT1c/GT1a (m/ z = 2027 .0) and GQc/b (m/ z = 2387.1) (265) A9 cells and NB324K cells have the sam e glycan profile for the non polar glycolipid fraction although the relative abundance of the glycans in the NB324K cells is higher ( Figure 2 8 D and F). In the EL4 T cells, no ne of the known non polar glycolipids could be assigned to the spectra observed ( Figure 2 8 E). A9 cells showed the expression of more types of polar glycolipi ds than EL4 T or NB324K cells ( Figure 2 8 A, B and C). All the glycolipids had SIA at the non reducing end and glucose (Glc) at the reducing end in the polar and non polar glycolipid fraction for all the cells Also, there are three lower molecular weight s ialylated glycans ( GM3, GM2, m/z= 1130.7 ; enclosed in red box in Figure 2 8 ) that are present on both the polar and non polar glycolipids
78 fraction. In all the three cell lines, t here is an increased expression of the monosialylated glycolipids as compared t o the multisialylated glycolipids, which is more evident in the polar glycolipids fraction s All the glycolipids in the polar and non polar fraction for the three cell lines were identified to be gangliosides amide motif ) except fo r LST1d ( enclosed in black box in Figure 2 8 A) ) which belongs to the neolactoseries ( amide motif ) (265) The multisialylated glycans 9) are present in the polar glycolipid fraction for all the three cell lines. The ganglioside GT3 which was recognized by MVMi in the previous glycan array screening was present on ly on A9 cells at low abundance ( enclosed in gree n box in Figure 2 8 A ) (209) However, the core motif in GT3 ( Neu5Ac Neu5Ac Gal Glc ) is also present in GT1c that is expressed in EL4 T cells albeit at low abundance (enclosed in blue box in Figure 2 8 A and B). Branching of glycans was also observed for the polar glycolipid fraction. Comparing the presence of the two deri vatives of SIA Neu5Ac is the major SIA in these cell lines, but Neu5Gc is present in trace amounts in the N glycan fraction for A9 and EL4T cells and in the glycolipid fraction for the A9 and NB324K cells (enclosed in dashed black box in Figure 2 7 A, B a nd D ; and Figure 2 8 A C, D and F ) The mass spectr a of the permethylated O glycans (an O glycan is covalently attached via an N acetylgalactosamine (GalNAc) moiety to proteins at serine or th reonine residues by an O glycosidic bond) are shown in Figure 2 9 While the A9 and EL4 T cells are shown to express only one type of O glycan, the NB324K cells express four types of O glycans, of which one is similar to the one expressed on A9 and EL4 T cells. The glycan which is expressed in all the three cell lines at the highest abundance (m/ z = 534.5 ; enclosed in red box in Figure 2 9 ) can be classified as T antigen and
79 belongs to either Core 1 O glycan ( ) or Core 8 O glycan ( ), depending on whether (41) The only SIA containing O glycan (enclosed in green box in Figure 2 9C ) is expressed on the NB324K cells and belongs to Core 1. The remaining O glycans expressed on the NB324K cells were identified to belong to Core 2 ( ) (m/ z = 779.4, m/z = 983.6) The N glycan ( l ow mass and medium mass), polar and non polar glycolipid composition for all the three cell lines showed the presence of terminal SIA glycans that were identified in the glycan array conducted previously and the SGM screening cond ucted in this study to recognize MVM capsids (209) In addition, gl ycans containing the SIA LN motif ( Neu5Ac Gal GlcNAc ) were present in the N glycan fraction for all the cells and in the polar glycolipid fraction of A9 cells (for example, present in the glycans enclosed in the green or blue box in Figure 2 7D, E and F or black box in Figure 2 8A ), which was also consistent with the glycan array and SGM screening data. The biantennary glycan with the terminal SIA LN motifs (i.e ., SGM 48; 2 3 1 1 1 3( 2 3 1 1 1 1 1 4GlcNAcitol ) was present in the low mass fraction of A9 cells and NB324K cells, but not in the EL4T cells (m/z=2792.2, enclosed in dashed red box in Figure 2 7 A and C ). B ased on the linkage combinations possible for the glycan with m/z= 4850.2 (enclosed i n a dashed green box in Figure 2 7 F) in the m edium mass N glycan fraction of the NB324K cells it possibly contains the 3' SIA LN LN motif which w as recognized by the MVM viruses in the glycan array screening (209)
80 Discussion Capsid Receptor Interactions Dictate Infectious Outcome The MVM capsid has been shown to play a c rucial role in tropism and in the onset of infection, though the essential interactions between the capsid and host cell receptor in pathogenesis are poorly understood (15, 112, 189, 196) Structural characterizatio n of the MVM capsid glycan interactions were conducted in this study utilizing glycans 3' SIA (Le x ) 3 and 3' SIA (LN ) 3 that were recognized by MVMp, MVMi and the MVMp virulent mutants, and GT3 that was specifically recognized by MVMi (209) Also, the recognition of 3' SIA (Le x ) 3 glycan (a tumor cell marker) by MVM, and the reco gnition of GT3 (neuronal marker) by MVMi explained MVM's tropism for transformed cells and MVMi's tropism for neuronal cells (209) In all the MVM receptor complex structures determined in this study, the SIA glycans bound in the vicinity of the same site at the icosahedral twofold axis that was identified previously wh en SIA alone was soaked into MVMp VLP crystals (Figure 1 5 and Figure 2 2) (188) The MVM SIA binding pocket is shallow and surrounded by charged and hydrophobic residues (VP2 residues in MVMp: K241, M243, I362, K36 8, R375, Y396, W398, D399, D553, Y558, and T578) that would accommodate and stabilize a long chain of sugar molecules. Significantly, the residues determining in vitro tropism and in vivo pathogenicity (residues 317, 321, 362, 368, 399, 553 and 558) are lo calized in the vicinity of this SIA binding pocket (7, 15, 112, 196) Such a pocket profile is a common structural feature for other virus SIA interactions (47, 90, 111, 338) In the capsid glycan complex es all the carbohydrate moieties that compose a particular oligosaccharide were not observed in the crystal structure. Th is observation suggest s that the glycans at the reducing termini that were not observed in the crystal s tructure
81 may not be forming tight interactions with the capsid surface or that only the SIA conservatively interacts at all the sixty binding sites of the capsid, with the remaining glycans adopting different orientations. These two possible scenarios wo uld be inconsistent with the icosahedral symmetry imposed during the structure determination, and would lead to lack of ordering of the density for these molecules. Due to these aforementioned reasons, the side groups for some of the carbohydrates structur es were not visible in the density map but were modeled based on clashscore and bonding interactions. In the MVMp 3' SIA (Le x ) 3 complex, the glycan extends towards the wall of the threefold protrusions and it can be speculated that the carbohydrate molecule s at the reducing end that are not ordered in the structure could interact with the V325 which is on the wall of the threefold protrusion (Figure 2 2B ) Similar possibilities occur in the MVMi 3' SIA (Le x ) 3 and MVMp 3' SIA (LN) 3 complexes (Figure 2 2D and Figu re 2 3B ). The MVMp V325M is a virulent mutant and V325 has been shown to modulate SIA binding in a manner similar to I362 and K368 (188, 209, 259) The 3' SIA (LN) 3 and 3' SIA (Le x ) 3 glycans bind along the wall of the p ocket on the MVMp capsid and although the terminal SIA can be superposed, the rest of the molecule adopts a slightly different conformation as evident in Figure 2 4A and Figure 2 4 D. The differences in the conformation could be attributed to the difference s in glycan composition The 3' SIA (Le x ) 3 glycan contains additional fucose group as compared to the 3' SIA (LN) 3 glycan. The 3' SIA (Le x ) 3 glycan contains the 3' SIA Le x tumor cell marker that is suggested to be utilized by the MVM viruses to bind to cancer cel ls (139, 151, 152, 182) D ifferences in the glycan
82 composition might also explain the different binding pattern observed in the MVMi 3' SIA (Le x ) 3 and MVMi 3' SIA (LN) 3 complexes (Figure 2 3 and 2 4 ). Comparative analy sis of all the complex structures shows that the binding pocket can accommodate the various glycans in different conformations (Figure 2 4C). Since the terminal SIA is recognized by all the MVM viruses, the SIA from all the capsid glycan structures determi ned in this study were superimposed to understand the specificity of SIA binding to these viruses (Figure 2 4 F). The SIA conformation is similar for the MVMp complexes (Figure 2 4D) but not for the MVMi complexes (Figure 2 4E) which makes it difficult to define the role of specific amino acid residues in dictating the glycan binding specificity. The heterogeneity in the terminal SIA binding observed for these structures could be due to the promiscuous nature of the binding pocket and /or limitations of the crystal structure determination studies, such as the merging of separate crystal diffraction data sets to achieve more data completeness, probable twinning of the MVM crystals and icosahedral averaging. Although, the capsid glycan interactions were not co nserved, a few capsid surface residues made non specific contacts with the glycans on all the MVM capsids. Such as the residue 558 which is a cell tropism determinant and makes contacts with all the glycans whether on MVMp or MVMi capsid. I nteraction with the residue 368 which is also a cell tropism determinant is stronger for the MVMp glycan complexes (2.5 3.0 ) as compared to the MVMi complexes (greater than 4.0 ). Also, R368 is involved in hydrogen bonding interactions with D399 and E321 in MVMi; an d with D399 in MVMp K368R, which are not present in MVMp (which has K368, D399 and G321) and MVMp K/I (D399 points away from R368) (Figure 1 7) (7, 168) Residue 399 interacts with the glycans on the MVMp capsid
83 but 2 8 linked GT3 glycan. In the previously solved MVMp SIA glycan complex structure, residue D399 was the only residue within interaction distance (3.2 ) (Figure 1 5). These observations suggest that D399 is n ot necessary for binding to 2 3 linked SIA glycans on the MVMi capsid but is required for binding to 2 8 linked glycans It is known that a forward second site mutation that confers fibrotropism to MVMi occurs at D399 (D339 mutated to A or G results in loss of hydrogen bond ing interaction with R368 ) when either 317 or 321 are mutated; and D399 is essential for infection in lymphocytes (7, 63, 79) Also, mutation at D399 (D399A or D399G) in MVMi accompanied with a second non coding change in the NS gene (that results in the accumulation of NS2) confers fibrotropism (7, 63, 79) Towards dissecting the role of D399 in tissue tropism, the MVMi D399A mutant was screened on the CF G Mammalian Printed Array Version 4.1 alongside wt MVMp and MVMi (data not shown). The data show ed that this mutant binds to 3 linked SIA but not to 8 linked SIA confirming the suggestion that D399 is not necessary for binding of 3 linked SIA on MVMi capsid, but it is important for 8 linked SIA recognition Residue 368 might have a direct effect on the SIA binding or an indirect effect through interaction with D399. The D399 R368 interaction present in MVMi and MVMp K368R mutant must configure 8 linked SIA Apart from residue 368 (conserved between MVMi and MVMp K368R which bind to GT3) and 399, there are other residues that are involved in unique interactions with GT3 such as M36 6 and E321 (also cell tropism determinants) that might be neurotropic determinants. The interaction between D399 and R368 is present in the MVMp K368R mutant but not the R368 and E321 interaction (G321 in MVMp),
84 which might explain its lower affinity to G T3 than MVMi and its fibrotropic phenotype in vitro (which is different from MVMi 's lymphotropic phenotype ) (188, 209) Apart from Y558, there are other conserved residues that make contacts with the glycans on bot h the viruses, such as R375, E394, Y396, T397 and 575 (a leukopenia associated mutation). The binding affinity for the 3' SIA (LN) 3 was MVMp=MVMi>>MVMp K368R>MVMp I362S=MVMp K/I (209) Also, the MVMp K/I mutant showed the most reduced affinity to the 3' SIA (LN) 3 glycan than MVMp, MVMi or MVMp K368R, but was the most pathogen ic in vivo (188, 209) In the MVMp K/I mutant, the mutation at residue 362 that changes Ile to Ser (instead of Valine as in MVMp K368R and MVMi) and t he different conformation adopted by D399 in the MVMp K/I mutant compared to the MVMi or MVMp K368R mutants (although R368 is common to these three viruses) seems to be involved in its non recognition of GT3 MVMp I362S and MVMp K/I have the same reduced affinity for the 3' SIA (LN) 3 glycan which implicates residue 362 in regulating the binding specificity to the 3 linked SIA glycans, while residue 368 regulates the binding specificity for both the 3 linked and 8 l inked SIA glycans Although, no binding interactions with residue 362 were observed in the complex structures analyzed in this study or in the SIA soaked MVMp VLP structure solved previously this residue lies in the vicinity of the pocket (188) T hese structure determination studies structurally verified the role of certain residues such as, 368, 399 and 558 in glycan binding but could not identify the determinants of SIA binding specificity. The SIA receptor binding pocket is flexible and colocalizes with the tropism and pathogenicity determinants and also allows for differential engagement of the SIA rec eptor The differences in the
8 5 capsid receptor interactions suggest a role in dictating pathogenicity but could not explain the differences in cell tropism for the MVM strains and mutants. Recognition of S ialic Acid Derivatives by MVM Viruses Previous glyca n array screening studies focused on identifying the SIA motif(s) recognized by MVM among a large number of different glycans (209) while screening on the SGM array enabled the explor ation of the potential role of SIA derivatives in dictating differential MVM strain recognition. Three types of MVM capsids, VLPs (no genom e and assembled from VP2 alone), Empties (no genome and assembled from VP1 and VP2), Fulls (contains wt genome and assembled from VP1, VP2 and VP3) were screened on the SGM array to investigate if the ir different composition results in differences in glyca n binding properties. The diversity of SIA modifications on this array allowed for the detection of novel MVM glycan interactions. On this array, the gly cans 2 3 2 6 linkages, (b) derivatives of Neu5Ac, Neu5Gc or Kdn (defined in c hapter 1) and (c) underlying structures derived from lactose, LNnT, NA2 or LNT (defined in the experimental methods in this chapter ) Depending on the strain, MVMp or MVMi, the glycan binding profile for VLPs, empty and full capsids was similar ( Figure 2 5 and Figure 2 6 ). T he data for t he full MVMp viruses however, exhibited high background noise as evident by the high %CV ( Figure 2 6 A ) The recognition of similar glycans by full and empty capsids of the particular strain indicates that the encapsidation of DNA by the full capsids followed by its maturation i.e cleavage of VP2 to give VP3 does not affect its receptor binding ability com pared to empty capsids, and this agrees with previous observations (181, 282) Also, the similar recognition profile for empty capsids and VLPs validates previous claims that they are antigenically and structurally equivalent (132)
86 Complex N glycans are reported to have tw o types of LacNAc (Gal GlcNAc) 4GlcNAc (Type2) of which T ype2 repeats are more common. All the MVM viruses bound to SIA 3 linkages to T ype 2 LacNAc (Gal 1 4GlcNAc) of LNnT and NA2 precursor which is cons istent with the previous glycan array screening data (209) ( Figure 2 5 and Figure 2 6 ) This suggests that the type of Gal GlcNAc linkage plays a role in the receptor capsid interaction. Preliminary modeling of the Type 1 LacNAc glycan (such as 3 GlcNAc ) in the MVM SIA binding site ( guided by the capsid glyc an complex structures determined in this study and the MVMp VLP SIA complex structure solved previously (188) indicates that the 3GlcNAc linkage creates steric hindrance in the binding site. The MVM p and MVMi viruses showed a strong preference for binding to SIA that is methylated at position C 9 of Neu5Ac (SGM 23 and SGM 55) compared to the other d erivatives (acetylated or lacto ylated N eu5Ac such as SGM 52 and SGM60, respectively, o r Neu5Gc) and also non derivatized Neu5Ac glycan (SGM 48) ( Figure 2 6 ) The hydrophobic methyl group specifically at position C 9 and not at C 8 of Neu5Ac is preferred by both MVMp and MVMi. Among the SIA glycans methylated at position C 9, binding affinit y of SGM 23 (methylated, monosialylated but single antennae) is higher than SGM 55 (methylated, monosialylated but biantennary). The specific recognition of Neu5Ac9 O methyl derivative of sialylated glycans by both MVMp and MVMi is interesting but since th is is a synthetic derivative that has not yet been isolated in nature, no further conclusions can be made. Also, preliminary modeling studies of SIA glycans methylated at C 9 or C 8 into the MVMp glycan binding site did not provide any
87 clues to this differ ential specificity. The C 8 or C 9 of N eu5Ac modeled in the complex structures ( i.e., structures determined in this study and the MVMp VLP SIA complex structure solved previously (188) ) is not in close proximity to any capsid surface amino acid residues to have any effect on the glycan binding interaction Even though the binding for the biantennary glycan SGM 48 by MVMp and MVMi viruses is low as compared to the Neu5Ac methylated derivatives it is still the highest among the Neu5Ac non derivatized glycans which suggests a strong preference of MVM p and MVMi for this glycan. MVMp viruses showed binding to the 9 O acetylated Neu5Ac derivative (SGM 52) ( Figure 2 5 and Figure 2 6 A, B and C ) Binding to glycans containi ng Neu5,9Ac is significant, because it has been seen that SIA 9 O acetylation is upregulated (e.g., 9 O acetylated GD3) in melanoma cells in humans (249) The enveloped viruses, Bovine Coronavirus (BCoV) and Influenza C virus also require 9 O acetylated S IA containing receptor for a successful infection. The se viruses possess 9 O acetyl esterase activity that promotes escape of virion progeny (266) None of the MVMp o r MVMi viruses bound to Neu5Gc or its derivatives even though it exists in mammals and other parvoviruses such as AAV5 and CPV have been reported to bind it ( (19, 303) and unpublished data) So, an N glycolyl modi fication at position C 5 of Neu (as in Neu5Gc ) is not preferred for binding as compared to an N acetyl modification at the same position ( as in Neu5Ac ) Also, the K dn derivative in which the amino group at C 5 of Neu5Ac is removed is not recognized. The MV Mp viruses also bound to the 9 O lactoylated Neu5Ac derivative (SGM 60), which are present in serum glycoproteins (165) This implies that the binding of MVM viruses to the sialylated derivatives is highly
88 dependent on the type of substitutions at position C 5 and C 9 of the sialic acid The MVMi viruses did not bind to 9 O acetylated (SGM 52) and the 9 O lactoyl (SGM 60) SIA derivatives ( Figure 2 5 and Figure 2 6 ). In the case of MVMp viruses, the SIA modification s that are recognized are: N acetylation at C 5 (Neu5Ac) alone or with O acetylation at C 9, or O lactoylation at C 9 or O methylation at C 9 For MVMi, the SIA modifications that are recognized are : N acetylation at C 5 (Neu5Ac) alone or with O methylation at C 9. These observations suggest that th e MVMp capsid allows for binding to more derivatives than the MVMi capsid, but preliminary modeling studies of these SIA derivatives on the MVMp or MVMi SIA binding site showed that the glycerol chain containing C 8 or C 9 is not involved in any capsid int eractions that could explain these differences. SGM screening for the first time identifies the SIA derivatives specifically recognized by the MVM viruses and also shows recognition for the SIA LN motifs that are prese nt on the cell types permissive for MVM viruses (results of the glycan profiling experiment ) Cellular Glycan Profiling Validates Glycan Screening Data The glycan profiling data from the three cell lines A9 fibroblasts, EL4 T lymphocytes and NB324K trans formed cells, correlates with the previous glycan array screening and current SGM array results since the glycan motifs that were recognized by MVM viruses in these arrays are also expressed on these cell types ( Figure 2 7 and Figure 2 8 ) The abundance an d variability of SIA containing glycans on these cell surfaces, makes them an obvious choice for receptors, as exemplified by the huge number of viruses that utilize sialylated receptors. The glycan arrays attempt to mimic the diversity of glycans present on a cell surface and if a virus specifically recognize s a particular type of glycan on these arrays, it is suggested that the particular virus
89 probably recognizes similar glycans in its natural environment. As shown by the glycan profiling of the three c ell lines infected by MVM viruses, there are many glycans with carbohydrates other than SIA at the non reducing end, but MVM viruses specifically recognized the SIA glycans, which is consisten t with the glycan array screening and SGM data (209) In addition, glycans with the SIA derivative Neu5Gc ( recognized by AAV5 and CPV) are also prese nt in all the three cell lines as in the glycan array and SGM but MVM virus es did not recognize these derivatives in either array ( (19, 209, 303) and unpublished data) Humans (but not mice) l ack the hydroxylase required to produce Neu5Gc from Neu5Ac but we do ob serve a Neu5Gc containing glycan (enclosed in dashed black box in Figure 2 8C and F ) in the glycolipids fraction of NB324K cells which are human transformed cells (309) This might be explained by the ability of human cells to incorporate Neu5Gc from mamm alian foods, particularly red meat and milk (309) Also, cancer cells have been found to express higher amounts of Neu5Gc on their surface (3 09) C ontamination due to fetal bovine serum glycoproteins might also explain this observation. Glycans containing the SIA LN (Neu5Ac Gal GlcNAc) motif (for example, present in the glycans enclosed in the green or blue box in Figure 2 7 D, E and F, or bla ck box in Figure 2 8A ) were recognized by both MVMp and MVMi in glycan array and SGM screening, and their expression in all the three cell lines would explain the common recognition of these glycans by both MVMp and MVMi ( Figure 2 5 ) (209) Also, the biantennary complex N glycan SGM 48 which was recognized by both MVMp and MVMi viruses in the SGM array is present in the low mass fraction of A9 cells and NB324K cells (m/z=2792.2, enclosed in dashed red box in Figure 2 7A and C) The
90 glycan SIA LN LN that was recognized in previous glycan array screening is possibly (base d on the linkage combinations allowed) present in the NB324K medium mass N glycan fraction ( m/z= 4850.2 ) (enclosed in a dashed green box in Figure 2 7F ) The gangliosides although broadly distributed, are predominantly expressed in the brain, whereas the n eolacto series glycolipids are common on certain hematopoietic cells such as leukocytes (265) The presence of the three common glycan structures (GM 3, GM2 and a Neu5Gc containing glycan ; enclosed in red box in Figure 2 8 ) in both the polar and non polar glycolipids fraction in all the three cell lines suggests that these glycans do not differentiate between the polar and non polar glycolipids present on these cell surfaces. The GT3 ganglioside present in the A9 fibroblasts (enclosed in green box in Figure 2 8A) was recognized by MVMi in the previous glycan array screening (209) There was no binding to monosialylated or GM gangliosides in the previous glycan array screening (209) even though these are the most highly expressed gangliosides in the glycolipids fraction for all the th ree cell lines (Figure 2 8 ) All the 8 multisialylated glycans that were recognized by MVMi viruses in the glycan array screening are present only in the polar glycolipid fractions in these cell lines. However, previous experiments have shown that a SIA containing glycoprotein (and not glycolipid) is utilized as a receptor for MVM infection (78, 188) Although the presence of multisialylated glycans on glycolipids has been well documented in the past, several rece nt studies have shown that 8 multisialylated glycans especially GT3 and GD3 are also present on glycoproteins in mouse brain (138, 262, 263) which would be consistent with MVMi viruses utilizing a n 8 multisialylated glyc oprotein for i nfection in neuronal cells M ultisialylated glyc oproteins 8 link age s such as Neural Cell
91 A dhesion M olecule (NCAM) and CD166 (or A ctivated L eukocyte C ell A dhesion M olecule (ALCAM)) are commonly found expressed on the neuronal cells and may serve as receptors for MVMi on these cells (263, 271) The O glycans have GalNAc at their reducing end and are as such classified as the mucin type ( Figure 2 9 ) (41) The sialylated T antigen (m/z 895.5) is annotated only in the NB324K cell line which is consistent with the observation of this motif on the cell surfaces of many leukemia and tumor cells (41) Results from previous biochemical and screening assays could not conclude whether the sialylated glycans recognized by MVM were N linked or O linked (209) But the absence of sialylated O linked glycans and the abundanc e of N linked sialylated glycans on these cell surfaces would imply that MVM binds to N linked sialylated glycoproteins on these cell surfaces. The NB324K cell line is a transformed cell line that expresses sialylated N glycans with the SIA Le x motif ( e.g ., m/z=3142.4, m/z= 3952.3 etc ; enclosed in red box in Figure 2 7 C and F) that was recognized previously by the MVM viruses on the glycan array (209) and is a known carbohydrate marker for cancer cells (139, 151, 152, 182) The expression of this motif has also been specifically observed in the N glycan profile of other transformed cell lines such as THP 1 (monocytic leukemia cell line), HL 60 (promyelocytic leukemia cells) and K562 (erythromyeloblastoid leukemia cell line) and this data is available at the CFG database ( http://www.functionalglycomics.org/glycomics/publicdata/glycoprofiling new.jsp ) This validates the suggestion that the SIA Le x motif on the tumor cell lines is utilized by the oncotropic MVM virus to bind to these cells. NB324K cells can be infected by both MVMp and MVMi, while A9 fibroblasts can be infected only by MVMp and EL4 T
92 lymphocytes can be infected only by MVMi. Comparing the SIA glycan profile (because it has been shown that MVM viruses only bind to SIA containing glycans) of these three cell lines, two glycans were identified to be present in A9 cells and NB324K cells but not in the EL4 T cells (m/z=2792.2, enclosed in dashed red box in Figure 2 7 A and C; and m/z=3603.5, enclosed in green box in Figure 2 7 D and F). The glycan with m/z=2792.2 (enclosed in dashed red box in Figure 2 7 A and C) was also recognized by both MVMp and MVMi in the SGM screening (SGM 48) which suggests that EL4 T cells might restrict in fection by MVMp due to the absence of this glycan on their cell surface Since this glycan was recognized by MVMi and is present on A9 cells, it still unclear why MVMi cannot bind and infect these cells. Both A9 and EL4 T cells have a similar glycan profil e which fails to explain the differences in the cell tropism between these viruses based on receptor interaction. However, the glycan profiling data does confirm the role of sialylated glycans in MVM infection. Summary Structural studies of the homologous MVM strains and mutants in complex with the respective receptors provided an understanding of how a single or few amino acids substitutions affect receptor binding specificity. A common receptor binding pocket at the twofold axis was utilized by all the M VM viruses but the glycan conformation and the capsid receptor interactions were different. In the MVMp 3' SIA (Le x ) 3 MVMp 3' SIA (LN ) 3 MVMp 3' SIA (Le x ) 3 and M V Mi GT3 complexes, longer oligosaccharides were visualized. In the complexes where longer oligosacch arides were observed, interaction with the residue V325 (virulence determinant) was possible. Residue D399 is not required for binding of 3 linked SIA on MVMi capsid but is a determinant for binding to 8 linked SIA glycans. This study also suggested that residues 558, 375, 394 could dictate
93 SIA binding, and residues 321 and 366 might dictate the neurotropic properties of MVMi virus. The binding profile for VLPs, empty and full particles was similar on the SGM which validates the use of VLPs in lieu of infectious virions for structural and biochemical studies tha t examine receptor interactions. MVMp and MVMi specifically 3 linked sialylated derivatives which is consistent with the previous glycan array screening Both MVMp and MVMi showed a preference for binding to sialic acid that was methylated at C 9 compared to the other derivatives All MVMp viruses also bound to 9 O acetylated SIA and 9 O lactoylated SIA derivatives, but MVMi did not bind to these glycans. In addition, all the MVM viruses bound to a biantennary SIA glycan with 3' SIA LN motif that is also present in the A9 and NB324 K cells. To summarize the SGM data MVMp can tolerate SIA modifications such as: N acetylation at C 5 (Neu5Ac) alone or with O acetylation at C 9, or O lactoylation at C 9, or O methylation at C 9 ; while for MVMi only N acetylation at C 5 (Neu5Ac) alone or with O methylation at C 9 is recognized. G lycomic profiling of A9 fibroblasts (infected by MVMp but not MVMi), EL4 T lymphocytes (infected by MVMi but not MVMp) and NB324K cells (infected by both MVMp and MVMi) validate d the presence of the glycans that w ere recognized by the MVM viruses in the previous glycan array and the current SGM screening such as glycans w ith SIA LN motif and SIA Le x motif multisialylated glycans, and the biantennary glycans. The expression of sialylated N glycans with the SIA Le x motif on ly on the NB324K transformed cell line implies that the MVM viruses utilize the SIA Le x motif on the tumor cell lines to bind t o these cells. The biantennary glycan that was recognized by MVM viruses in the SGM screening is
94 present in A9 and NB324K cells but not in EL4 T cells. This suggests that EL4 T cells might restrict infection by MVMp due to the absence of this glycan on the ir cell surface. A common SIA binding pocket is utilized by both the MVM strains and the capsid SIA receptor interactions involved residues known to play a role in dictating differences in cell tropis m and pathogenicity which suggests that the flexible re ceptor binding pocket does play a role in cell recognition, especially cancer cell recognition. The differences in the capsid SIA receptor interactions might dictate the differences in pathogenicity but could not explain the differences in cell tropism bet ween the MVM strains. The differences in cell tropism and pathogenicity have been mapped to the VP2 at the twofold depression in the vicinity of the receptor binding pocket (15, 112, 196) Previous studies have show n that both MVM strains bind and enter the same cells (282) and the glycan profiling dat a also showed that similar glycans are expressed on both A9 and EL4 T cells. Also, hybrid cells from fibroblasts and lymphocytes can propagate both strains of MVM (294) suggesting that the block is due to the lack of a differentiation dependent cellular factor in restrictive cells. This would suggest that the block in infection is post entr y but regulated by the capsid twofold pocket It is possible that post cellular entry, the virus utilizes the twofold pocket to interact with cellular host proteins once the SIA receptor is recycled and the absence of such a cellular protein in the non per missive cell line restricts virus infection. Another possibility is that differential utilization of the SIA receptor by MVMp and MVMi could trigger different cell signaling pathways in the permissive and non permissive cell lines resulting in different ou tcomes of virus infection. Also, SIA receptor binding and/or the low pH encountered during endocytic trafficking might trigger structural transitions in the capsid that facilitate
95 interactions with cellular proteins that lead to successful infection. There are a number of other viral systems in which tropism appears to be controlled by the coat protein, yet restriction appears to act at an intracellular stage, such as Hamster polyomavirus (81) and Human Immunodeficiency Virus (HIV) for which the tropism determinant maps to the env coding sequence but appears to act after receptor binding and prior to reverse transcription (291)
96 Table 2 1. Data processing and refinement statistics Parameter MVMp s(LN) 3 Soak MVMp s(Le x ) 3 Co crys tal MVMi s(LN) 3 Soak MVMi s(Le x ) 3 Soak MVMi GT3 Soak MVMp K/I s(LN) 3 Co crys tal Wavelength ( 1.000 1.000 0.918 0.918 0.918 0.918 Space Group C2 C2 C2 C2 C2 C2 Unit cell parameters (, ) a=448.7, b=416.5, c=306.1, a=441.9, b=410.4, c=301.2, a=440.8, b=409.7, c=301.1, a=440.7, b=408.7, c=300.3, a=444.1, b=41 1.9, c=302.7, a=441.9, b=409.9, c=301.2, Resolution () 40 3.7 (3.8 3.7) a 40 3.3 (3.4 3.3) a 50 3.9 (4.0 3.9) 50 3.5 (3.6 3.5) 40 3.6 (3.7 3.6) 40 3.5 (3.6 3.5) Completeness (%) 73.4 76.1 71.1 78.4 71.4 66.7 Redundancy 1.7 2.3 2.0 2 .0 2.4 2.0 R sym b (%) 13.7 12.9 18.6 10.3 15.6 10.0 R factor c / R free d (%) 34.6/34.8 31.3/31.6 39.6/39.5 39.2/39.4 35.0/35.1 34.6/35.3 a Values in parenthesis are for highest resolution shell. b R sym where I is the intensity of a r eflection with indices h, k, l and is the average intensity of all symmetry equivalent measurements of that reflection. c R factor o | |F c o |)100 where F o and F c are the observed and calculated structure factor amplitudes, respectively. d R free is calculated the same as R factor except it uses 5% of reflection data omitted from refinement.
97 Figure 2 1. Virus purification. Coomassie stain ed SDS P AGE showing the VP2 ( 1 ) for (a) MVMp VLP, (b) Empty MVMp (c) Full MVMp (d) MVMi VLP, (e) Empty MVMi, (f) Full MVMi, (g) MVMp I362S VLP, (h) MVMp K368R VLP and (i) MVMp I362S/K368R VLP; and negative stain EM showing intact capsids ( 2 ) imaged at magnifi cation of 100,000 times for ( a ) ( d ) (i) ( e ) and ( f ); 60,000 times for (b), (c) and (g); and 40 ,000 times for (h).
98 Figure 2 2. Glycan binding sites on the MVM p and MVMp K/I capsids. (A to F ) Surface representation of the MVM capsids showing the glycan binding site The glycans are shown in the stick form inside a 2F o F c density map (gray mesh ) in (A, C and E ) and colored according to atom type (the carbon atoms are colored in magenta cyan and grey for the MVMp 3' SIA(Le x ) 3 MVMp 3'SIA(LN) 3 and MVMp K/I 3'SIA(LN) 3 complex es respectively ; nitrogen in blue and oxygen in red). The symmetry relat ed monomers are colored differently and labeled. The amino acid residues interacting with the glycan and/or tropism determinants are labeled. The filled oval denotes the twofold axis.
99 Figure 2 3. Glycan binding sites on the MVMi capsids. (A to F) Sur face representation of the MVM i capsids showing the glycan binding site The glycans are shown in the stick form inside a 2F o F c density map (gray mesh) in (A, C and E) and colored according to atom type (the carbon atoms are colored in blue, orange and gr een for the MVMi 3' SIA(Le x ) 3 MVMi 3'SIA(LN) 3 and MVMi GT3 complexes, respectively; nitrogen in blue and oxygen in red). The symmetry related monomers are colored differently and labeled. The amino acid residues interacting with the glycan and/or tropism d eterminants are labeled. The filled oval denotes the twofold axis.
100 Figure 2 4 Superposition of glycans on MVM capsid (A to F ) Surface representation of the MVM capsids with the bound glycans are shown as in Figure 2 2 and Figure 2 3 (A) The 3' SIA(Le x ) 3 and 3'SIA(LN) 3 glycan are superimposed on MVMp capsid (B) The 3' SIA(Le x ) 3 3'SIA(LN) 3 and GT3 glycan are superimposed on MVMi capsid, (C) all the glycans are superimposed, (D) SIA of the 3' SIA(Le x ) 3 and 3'SIA(LN) 3 are superimposed on MVMp capsid (E) SIA of the 3' SIA(Le x ) 3 3'SIA(LN) 3 and GT3 are superimposed on MVMi capsid (F) all the SIA are superimposed.
101 Figure 2 5 Sialylated derivatives recognized by the MVM viruses. The glycan number and the schematic representation is shown.
102 Figure 2 6 Data from the SGM array. (A) Full MVMp, (B) Empty MVMp, (C) MVMp VLPs, (D) Full MVMi (E) Empty MVMi empties, (F) MVMi VLPs. The bars represent relative fluorescence for a given glycan. The glycans showing specificity to the MVM viruses are labeled.
103 Figure 2 7 N glycan expression on A9, EL4 T and NB324K cell lines Low mass (A, B and C) and medium mass (D, E and F) N linked glycans are shown. The glycans are annotated in the cartoon form. Figure inset depicts the cartoon legend. The y axis repre sents relative glycan abundance.
104 Figure 2 8 Polar and non polar glycolipid expression on A9, EL4 T and NB324K cell lines P olar glycolipids (A, B and C) and non polar glycolipids (D, E and F) derived from A9 (A and D), EL4 T (B and E) and NB324K cells (C and F) are shown. The glycans are annotated in the cartoon form. Derivatives (DR) were made to differentiate between symmetrical molecules. M and m refer to major and minor species respectively. Figure inset depicts the cartoon legend.
105 Figure 2 9 O glycan expression on A9, EL4 T and NB324K cell lines MALDI TOF MS profiles of the O glycans derived from (A) A9, (B) EL4 T and (C) NB324K cells are shown. The glycans are annotated in the cartoon form. Figure inset depicts the cartoon legend. The y ax is represents relative glycan abundance.
106 CHAPTER 3 STRUCTURAL CHARACTERIZATION OF H 1PV: INSIGHTS INTO CAPSID GENOME INTERACTION S AND RECEPTOR BINDING Background H 1 parvovirus (H 1PV) is a member of the rodent subgroup of the parvovirus genus of the P arvoviridae and shares 6 6 % VP2 sequence identity with both MVMp and MVMi that also belong to the same subgroup and for which 3D structures are known (26) H 1PV was isolat ed from rats transplanted with H EP 1 (human liver adenocarcinoma cell ) and from aborted human fetuses (301) H 1PV, like MVM and LuIII (member of the rodent subgroup) has oncolytic properties and is selectively cytotoxic for transformed or tumor derived cells of various species including human (discussed in chapter 1) However, it is non pathogenic in humans and do es not integrate its genome into cellular chromosomes which makes it an excellent choic e as a vector for gene therapy The first clinical study of H 1PV was conducted in 1965 with the injection of wt virus into two osteosarcoma patients (302) Although this treatment did not completely ablate tumor development, neutralizing antibodies against the virus were found. In a recent preclinical study, H 1PV was assessed for the killing of human neuroblastoma cells and hepatoma cells and demonstrated tumor selective lytic effects and low toxicity for non transformed cells (173, 201) Local or systemic treatment of advanced rat and human gliomas in rat models with H 1PV has also been observed to ind uced regression (114) These studies have set the stage for the first phase I /IIa clinical trial using H 1PV vectors in subjects suffering from glioblastoma multiforme (113) Most of these recombinant parvoviral vectors (based on, H 1PV, MVM etc) targeted for tumor therapy utilize a double edged strategy that take s advantage of their inherent oncotropism and selective cytotoxicity plus their ab ility to deliver therapeutic
107 genes that code for toxins such as thymidine kinase or cytokines/chemokines that enhance host immune response against the tumor cells (98, 124, 202, 221, 320) The cytotoxicity has been attributed, in part, to NS1 (39) It has been shown tha t interaction with host cellular factors expressed during S phase, or interferon response mounted by the tumor cells dictate s successful infection by H 1PV or MVM. However, the mechanisms underlying tumor cell recognition which is the first key step toward s a successful infection, and regulates tissue tropism and pathogenesis are not yet fully understood (110, 120, 237) Glycan array screening studies of MVM viruses have suggested that their recognition and binding t o cancer cells is likely mediated via interactions with glycans containing the 3' SIA Le x motif which is a known tumor cell marker (139, 151, 152, 182, 209) Similar to MVM, H 1PV also utilizes the S I A component of a glycoprotein receptor for cell surface attachment (10, 77) and it is likely (based on its high VP2 sequence identity with MVM) that H 1PV also utilizes 3' SIA Le x containing glycans to infect tumor cells However, despite their common oncotropic properties, NB324K transformed cells are more permissive for infection by H 1PV than MVM (64) In a recent study (226) it was shown that the autonomous parvoviruse s H 1 PV, LuIII, and MVM exhibit variation in tumor tropism and the oncolytic activity of LuIII was mapped to the capsid VP2 gene using viral chimeras generated between LuIII and MVM. The capsid was shown to assist a post entry and pre DNA amplification ste p in the viral life cycle that lead to successful infection and transformed cell killing (226) The H 1PV capsid has also been suggested to play a role in cellular tropism in another study in which different tumor c ell lines were transd uced with H 1PV or H 1PV/AAV hybrid vectors (172) These viruses exhi bited variable tropism between tumor types. It has also
108 been demonstrated that the autonomous parvoviruses, such as MVM and H 1PV are less tolerant to replacement of wild type genome sequence with a foreign sequence for vector production, as compared to AA Vs (161, 174) It has been suggested that t he specificity of the capsid genome interactions which was observed to be different for MVM and AAVs might affect packaging of recombinant genomes and vector infectivity (7, 119, 177, 210, 214) The need to understand the role of H 1PV capsid in cell tropism efficient therapeutic gene packaging, and neutralization by antibodies has thus generated a need for the structural characte rization of H 1PV and the mapping of capsid regions involved in these steps in the infectious life cycle. In this study the specific carbohydrate motifs recognized by H 1 PV were identified using glycan array analysis and structural studies of H 1PV alone and in complex with the identified glycan s were conducted to examine the structural interactions involved in receptor attachment and possibly oncotropism The capsid gene has been mapped as the oncotropism determinant, so comparison of H 1PV and MVM capsid structures would identify structurally variable regions that may be responsible for the differences in their oncotropism. Structure comparison of H 1PV and MVM capsids might aid in understanding the mechanism of host range switching by the MVMp virus F1 t hat adapted in tissue culture to infect the non permissive rat fibroblasts (H 1PV also infects rat fibroblasts) and acquired mutations at the twofold depression on the MVMp capsid (99) In addition, the structures o f wt H 1PV virions and wt H 1PV empty capsids were determined and compared in this study to investigate the capsid genomic DNA interactions. Information on the capsid dependent mechanism of receptor recognition, genome packaging and host immune response wo uld aid in the engineering of MVM or
109 H 1PV b ased gene therapy vectors with specific tissue targeting capability, improved therapeutic gene packaging efficiency and low immunogenicity. Experimental Methods Cell Lines NB324K cells (simian virus 40 transforme d human newborn kidney fibroblast cells) BRL) with 5% fetal calf serum, glutamine, and antibiotics. Virus Production and Purification NB324K cells grown as monolayers to 50% confluency (cell density of 4 x 10 6 /10 cm plate) were infected with wt H 1PV virus for 1 h at 37C with m.o.i of 0.1 PFU per cell and followed by occasional rocking of the plates. The cells were incubated for an additional 5 7 days till cytopathic effect was observed (~80% cell lysis). Cells were harvested by scraping and pelleted by low speed centrifugation T he cell pellet was resuspended in TE buffer (50 mM Tris HCl pH 8.7, 0.5 mM EDTA) and stored at 20C. The virus capsids were released from the cells by three cycles of rapid freez e thaw ing The cellular debris was removed by low speed centrifugation in a JA 20 rotor (10,000 rpm, 15 min, 4 C ). The supernatant was then treated with micrococcal nuclease. Next, CsCl was added to the solution to a final density of 1.40 g/cm 3 and the sa mple subjected to equilibrium centrifugation at 35,000 rpm in a Beckman SW41Ti rotor for 24 h at 4 C The bands corresponding to empty capsids (1.32 g/cm3) and full virions (1.41 1.46 g/cm 3 ) were removed from the tubes by side puncture. The fractions were subjected to another round of CsCl equilibrium centrifugation as in the previous step, bands corresponding to empty and full capsids were extracted and the fractions dialyzed into Tris HCl Buffer (10 mM Tris HCl pH 7.5, 150 mM NaCl, 8 mM
110 CaCl 2 .2H 2 O) The d ialyzed fraction were subjected to a third round of CsCl equilibrium centrifugation as described previously. The virus bands were extracted and finally dialyzed into Tris HCl Buffer. T he particle concentrations were determined by hemagglutination of sheep erythrocytes as well as absorbance measurements (assuming an extinction coefficient of 1.0 and 7.0 for calculations in mg/ml for empty and full particles, respectively) and adjusted to 10 mg/ml using Ultrafree centrifugal filter units (Millipore, Billerica MA). The purity and integrity of the viral capsids was monitored using SDS PAGE with Coomassie blue staining and negative stain EM, respectively (Figure 3 1A, B and C) For the EM visualization, 5 l of purified virus sample at an estimated concentration of 2.0 mg/ml was spotted onto a 400 mesh carbon coated copper grid (Ted Pella, Inc., Redding, CA, USA) for 1 min before blotting with filter min twice blotted dry, and vie wed on a HITACHI 3000 electron microscope. Glycan Array Analysis The H 1PV empty capsids were screened to determine their glycan recognition specificity on the Mammalian Printed Array Version 3.1 containing 377 natural and synthetic glycans in collaboratio n with the Core H of CFG. The glycans present on this array are listed on the CFG website at http://www.functionalglycomics.org/static/consortium/resources/ resourcecoreh11.shtml The experiment was conducted as described previously (209) but with several modifications as mentioned below. To generate the printed array, a library of natural and synthetic mammalian glycans with amino linkers was printed onto N hydroxysuccinimide (NHS) activated glass microscope slides (SCHOTT Nexterion) in replicates of n=6. H 1PV was diluted with Binding Buffer (10 mM Tris HCl pH 7.5, 150
111 mM NaCl, 8 mM CaCl 2 .2H 2 O, 1% BSA 0.05% Tween 20) of this diluted sample was applied to the surface of the printed array slide, a cover slip was used to cover the slide, and then the slide was incubated at RT in a dark, humidified chamber for 1 h. The cover slip was removed the slide was then rins ed 4 times in W ash B uffer ( B inding B uffer minus BSA) and 4 times in Wash Buffer without Tween 20. Seventy microliters of anti H 1PV mouse antibody in hybridoma media (diluted 1:10 in 1XPBS) was added to the slide and it was covered by a cover slip for incu bation at room temperature for 1 h. Following incubation, the cover slip was removed and the slide was washed as above mouse IgG Alexa488 in 1XPBS at a concentration of 5 g/ml for 1 h. The slide was washed as above, followed by 4 washes in distilled water. The fluorescence intensity was detected using a ScanArray 5000 confocal scanner (Perkin Elmer, Waltham, MA). The image obtained was analyzed using the IMAGENE image analysis software (BioDiscovery, El Segundo, CA). The mean fluorescence intensity for each glycan in the printed array was obtained by removal of the t he highest and lowest values from each set of six replicates followed by averaging of the four remaining readings and the standard deviation (S.D) was also calculated To analyze the results, all the glycans were ranked according to their signal to nois e (S/N) ratio by dividing their mean relative fluorescence units (RFU) from the four replicates by the mean background generated in the control wells lacking glycosides. Variation within the 4 replicates was assessed as the coefficient of variation (%CV), calculated as 100 x S .D /Mean and was considered low if it was less than 3 0 as reported in other studies (278) The selection of glycans with specific binding affinity was based on two independent criteria: high ove rall total binding as measured by RFU
112 and low variation among the 4 replicates as assessed by %CV. Three sets of data with varying concentrations of H Crystallization Crystals of empty (no DNA) and f ull (DNA containing) H 1PV particles were grown using the hanging drop vapor diffusion method (199) with VDX 24 well plates and siliconized cover slips (Hampton Research, Laguna Niguel, CA, USA). The reservoir solution contained 1 3% (w/v) PEG 8000, 150 mM NaCl, and 8 mM CaCl 2 .2H 2 O as precipita nt s in 10 mM Tris solution (10 mg/ml) in Tris HCl Buffer (10 mM Tris HCl pH 7.5, 150 mM NaCl, 8 mM CaCl 2 .2H 2 O) solutio n at RT. Crystals of empty H 1PV capsids complexed with the glycans were obtained by co crystallization using the same crystallization conditions as above. The 3' SIA (LN) 2 and 3' SIA (Le x ) 3 glycan s provided by Core D of the CFG were resuspended in 10 mM Tris pH 7.5 to give a stock solution at 10 mg/ml. The co crystallization drops concentration that results in capsid : glycan ratio of 1: 180 or 1: 600 r solution The drop was equilibrated against 1 ml of reservoir solution at RT. Crystals obtained were soaked for 30 s in cryoprotectant solution containing the precipitant solution (10 mM Tris HCl pH 7.5, 150 mM NaCl, 8 mM CaCl 2 .2H 2 O) with 10% PEG 8000 an d 30% glycerol and flash cooled in liquid nitrogen vapor prior to X ray diffraction data collection. Diffraction Data Collection and Processing Diffraction data were collected for crystals of full H 1PV particles at the F1 beamline at the Cornell High Ene rgy Synchrotron Source (CHESS) on an ADSC
113 Quantum 270 CCD detector and empty H 1PV particles at the X29 beamline at Brookhaven National Laboratory (BNL) on an ADSC Quantum 315 CCD detector. For the H 1PV fulls, a total of 392 usable images were collected f rom 2 crystals with a crystal to detector distance of 230 and 300 mm, an oscillation angle of 0.3 per image, and usable images were collected from a single crystal of empty H 1PV particles at a to detector distance of 400 mm, an oscillation ang le of 0.3 per image and an exposure time of 20 s per image. The crystals diffracted X rays to beyond 2.7 and 3.2 resolution for the full and empty H 1PV particles respectively (Figure 3 1E) The measured diffraction intensities were indexed and inte grated with the HKL2000 suite of programs (224) and scaled and merged with SCALEPACK (224) The crystal system was determined to be primitive monoclinic. Inspection of the 0k0 class of reflections (for k=2n) showed systematic abs ences for the odd reflections indicating the presence of a 2 1 screw axis and thus the crystals belong to the space group P2 1 The full and empty H 1PV data sets scaled with an R sym of 13.2% (93.1% completeness) and 11.4% (61.3% completeness), respectively The details of the data collection strategy and processing statistics are summarized in Table 3 1. For the virus glycan receptor studies, two complex structures H 1PV 3' SIA (LN) 2 and H 1PV 3' SIA (Le x ) 3 were determined. X ray diffraction images were coll ected for empty H 1PV 3' SIA (LN) 2 co crystal s at the CHESS F1 beamline on an ADSC Quantum 270 CCD detector with the crystal to detector distance set at 350 and 400 mm. A total of 303 images were collected from two empty H 1PV 3' SIA (LN) 2 co crystal s with an
114 0.917 The crystals diffracted X rays to 2.7 resolution, and the R sym and completeness for this dataset were 12.8% and 75.8%, respectively. The data processing stat istics are given in Table 3 1 X ray diffraction data on the empty H 1PV 3' SIA (Le x ) 3 co crys tals was also collected at the CHESS F1 beamline at a wavelength of = 0.918 oscillation angle of 0.3, distance of 300 mm and exposure time of 45 s per image The total number of usable images was 234 The crystals diffracted X rays to 2.9 resolution, and the R sym and completeness for th is dataset were 12.1% and 68.7%, respectively. The data processing statistics are given in Table 3 1. Structure Determinatio n The X ray diffraction intensity data set collected was converted to structure factor amplitudes using the TRUNCATE program from CCP4 (Collaborative Computational Project, Number 4) (62) coefficient (V M ) was calculated to be 3.07 3 Da for the full H 1PV data and 2.95 3 Da for empty H 1PV data corresponding to solvent contents of 64% and 58%, respectively with two particles per unit cell related by a 2 1 screw axis, with one particle occupying a crystallographic asymmetric unit (195) The orientations of the two H 1PV virus particles in the crystal unit cell were determined with a self rotation function using the General Lock Rotation Function ( GLRF ) program (299) computed with 10% of the observed data between 10.0 and 5.0 180 to search for the twofold, threefold and fivefold icosahedral symmetry axes, respectively. The radius of integration was set to 120 established the orientation of the i cosahedral symmetry axes relative to the crystal axes
115 and confirmed that the two particles in the unit cell are related by the crystallographic twofold screw axis (b axis). The structure of the full H 1PV particle was determined by molecular replacement (258) PDB accession no. 1Z14) (168) was generated by the MOLEMAN program (167) and expanded to 60 subunits (one capsid) using the Oligomer Generator subroutine available at the VIPER database (50) for use as a molecular replacement phasing model. Structure factors in the 10.0 5.0 resolution range were calculated for the phasing model using the SFALL program from CCP 4 (Collaborative Computational Project, Number 4) (62) and used to calculate cross rotation and translation functions, using the AMoRe program (212) The highest peak obtained for the cross rotation function was used in the translation function calcula tion which was performed in both primitive monoclinic space group s P2 and P2 1 using the structure factor correlation coefficient (CC) as a key parameter for determining the correct solution A clear solution to the translation search using the oriented M VMp model was found for space group P2 1 which gave the highest correlation coefficient (CC) of 35.4% and lowest R factor ( o | |F c o |)100 ) where F o and F c are the observed and calculated structure factors, respectively ) of 46.5%. A similar approach w as used to phase the empty H 1PV data set which gave similar molecular replacement solutions to those for the full H 1PV data Rigid body refinement (FITTING in AMoRe ) improved the CC and R factor to 40.3% and 44.9%, respectively (212) The MVMp pa rticle was then rotated and translated into the full H 1PV data set unit cell according to the final molecular replacement solutions obtained, with an orientation, in Eulerian angles f 0.2510,
116 0.0000, and 0.2489. The MVMp all atom model (PDB accession no. 1Z14) was then superimposed individually on the 60 subunits of the oriented and positioned capsid, to get an all atom phasing model to be used in the subsequent refinement steps using the CNS program (46) For the full H 1 PV data initial phases for the oriented and positioned all atom MVMp model were calculated to 2.7 resolution and further improved by refinement in the CNS program using simulated annealing, energy minimization, conventional positional, and individual te mperature factor (B factor) refinement ( 46) This procedure was followed by real space electron density averaging using a molecular mask. The refinement and averaging procedures were conducted while applying strict 60 fold noncrystallographic symmetry (NCS) in CNS program (46) Five percent of the total data set was partitioned for monitoring of the refinement process with an R free calculation ( o | |F c o |)100 calculated with a 5% randomly selected fraction of the reflection data not included in the refinement) (45) Electron density that would be interpreted as residues 38 to 593 (last C terminal residue, VP2 numbering) of the H 1PV VP2 was built into the averaged sigma weighted 2F o F c electron density map by interactive substitution, insertion, and deletion of amin o acids relative to the MVMp model using the COOT program (97) Following model building, new phases were calculated and improved by several alternating cycles of refinement, real space electron density averaging, and rebuilding. To improve the quality of the maps density map modifica tion was carried out using the Density Modification subroutine in the CNS program which performed solvent flattening and NCS averaging (46)
117 Following the building of VP2 amino acids 38 to 593 into the averaged density maps, ordered unassigned density in the full H 1PV capsid interior, located in a pocket created by the reference VP monomer and an icosahedral twofold related VP monomer was modeled as nine ssDNA nucleotides (NTs) using the ssDNA modeled in the previously solved structures of MVM i and CPV as templates (PDB accession no. 1Z1C, 4D PV, and 1P5W) (118, 168, 329) Density interpreted as two Mg 2+ metal ions which coordinate the phosphate backbone of the ssDNA in MVM i and CPV were also ordered in the H 1PV pocket and modeled. In addition to the st retch of nine nucleotides, three more regions of unassigned density were observed in the F o F c density map inside the capsid and were modeled as cytosine base, with one of these located in a conserved AAV DNA binding pocket (119, 177, 210, 214) Finally, 98 water molecules and two molecules of ethylene glycol were built into the rest of the unassigned averaged positive F o F c density (at density threshold of bonding distance geometry. Significantly large peaks (F o F c refined with very low temperature factors (Bfactors), were re assigned as metal ions, with a test of B factors for Na + Ca 2 + or Mg 2+ Several cycles of refinement and averaging procedures interspersed by visual inspection of the 2F o F c and F o F c electron density maps and model building in COOT (97) were performed till there was no longer any improvement in the agreement between the observed (F o ) and ca lculated (F c ) structure factors as given by R factor (Table 3 1). The CONTACT subroutine in the program CNS was used to analyze the ssDNA VP2 interaction within the binding pocket (46)
118 For the empty H 1PV data set, the refined full H 1PV VP2 structure, without the modeled nucleotides (NTs), so lvent molecules metal ions and with the residues interacting with the NTs mutated to alanine was used as a starting model for refinement since the molecular replacement solutions were identical. The refinement procedure was carried out as described abov e for the full H 1PV data For the H 1PV co crystal data sets, the refined full H 1PV VP2 structure alone was used for molecular replacement and the structure refinement followed the procedure as outlined above for the full H 1PV capsid structure determin ation The F o F c electron density maps revealed well defined electron densit ies at the twofold axis for the S I A component of the complexed glycans The coordinate files for S I A were obtained from the HIC Up server (166) and the geometry restraints and dictionary file s were generated using the subroutine phenix.elbow in the PHENIX program (2) T he S I A was docked into F o F c density using interactive rigid body rotations and translations in COOT. The topology and dictionary files generate d for S I A in the PHENIX program were then used for subsequent refinement in the CNS program (46) The quality of the refined structures were analyzed using COOT and MOLPROBITY (56, 97) The values of root mean square deviations (rmsd) from ideal bond len gths and angles were obtained from the CNS program (46) and average B factor for VP2 models, solvent molecules NTs and ions were calculated using MOLEMAN (167) The refinement statistics are given in Table 3 1. The figures were generated using the PYMOL program (82) Structural Alignment Comparison The VP2 coordinates of MVM, CPV, FPV, and PPV (PDB accession no. 1Z1C, 1P5W 1C8F and 1K3V, respectively ) were aligned with the refined H 1PV structure using the secondary structure matching (SSM) program available in PDBefold (169)
119 positions aligned and also a list of the atomic distances (in ) between each aligned posit ion. It also provides information on amino acid residues that are structurally equivalent and gap regions The output from this program was used to identify variable regions that the superimposed structures as previously defined (168) Results Virus Purif ication and Crystallization The analysis of the purified empty and full H 1PV particles by SDS PAGE showed the expected two bands corresponding to VP1 and VP2 ; and three bands corresponding to VP1, VP2, and VP3, respectively (Figure 3 1A). The abundance of VP2 is approximately same as VP3 in the H 1PV full virions indicating that in these particles half of the VP2 proteins have undergone cleavage to VP3, although the process may not be complete Examination by negative stain EM showed intact particles (stai n penetrated and stain excluded for empty and fulls, respectively) for both samples (Figure 3 1B &C). Thin plate shaped and thin rod shaped c rystals grew in ~ 4 6 weeks from the 10 mM Tris HCl pH7.5, 150 mM NaCl, 8 mM CaCl 2 .2H 2 O and 3% PEG 8000 condition (F igure 3 1D) The approximate crystal dimensions were 0.3 x 0.15 x 0.005 mm for the thin plate shaped crystals and 0.15 x 0.01 x 0.005 mm for the thin rod shaped crystals. Diffraction data was collected on the thin rod shaped crystals. Glycans Recognized by H 1PV The empty H 1PV particles were screened at three different concentrations of 200 g/ml, 400 g/ml and 900 g/ml on the glycan array Depending on the concentration of H 1PV analyzed the RFU signal and the %CV for each glycan was different and takin g
120 into account these variables the highest affinity glycans were identified (Figure 3 2). Of the 377 glycans present on the array, H 1PV recognized only three 3 sialylated glycans linked to a common Gal 4 GlcNAc motif: SIA Le x Le x Le x or 3' SIA (Le x ) 3 ; glycan 227) SIA LN LN or 3' SIA (LN ) 2 ; glycan 236) and a biantennary glycan H 1PV Capsid Structure The structure of the full and empty H 1PV particles were determined to 2. 7 and 3.2 resolution, respectively, by X ray crystallography. The refinement values are within the range reported for other virus structures determined at a comparable resolution, as calculated by the Polygon subroutine (307) in the program PHENIX (2) The similarity of R f actor and R free for virus structures is a result of the high noncrystallographic icosahedral sy mmetry of the capsid. The map quality and refinement statistics show that the refined structure is not biased by the MVMp phasing model (Figure 3 3A). While the empty H 1PV particles were assembled from VP1 and VP2 and full H 1PV virions were assembled f rom VP1, VP2 and VP3, only the C terminal 554 amino acid (aa) residues, corresponding to residues 38 to 593 (VP2 numbering) of the VP2 (and VP3 in virions) common sequence were traceable in the icosahedrally averaged electron density map (Figure 3 3B) Thi s lack of N terminal ordering is also reported for all other parvovirus structures determined to date (exception B19 V (155) ). In the full H 1PV virion structure under the icosahedral fivefold axis, but it is not connected to the first interpretable N
121 terminal residue (aa 38) and was thus not modeled. In addition to N terminal residues 1 37 two amino acids at strands D and E see below ), which together with four five fold symmetry related DE loops assemble the fivefold channel, were not ordered in electron density maps for both full and empty H1 PV structures. The VP2 structure of the full and empty H 1 PV particles conserved the parvovirus capsid VP topology. It conta ins the core eight stranded jelly roll motif, consisting of two sheets, BIDG and CHEF, common in most virus structures, with long loops between the strands, and an the icosahedral twofold axis (Figure 3 strands are observed in the loops between the strands as previously described for other parvoviruses (reviewed in (53) barrel motif represents 17% of the ordered VP2 amino acid sequence, and forms the contiguous shell. The loops clustered from icosahedral symmetry related VP2 monomers form th e capsid surface. Two of the small stretches of strand structure, ribbon (DE loop) which ribbons from the fivefold symmetry related monomers to form the conserved cylindrical channel at the fivefold axes (Fig ure 3 3C and D ). Depressions are observed at and surrounding the twofold axis and surrounding the fivefold channel (Figure 3 3D). The floor of the depression around the channel is lined by the HI loop a flower pet al extending from the adjacent fivefold symmetry related monomer (Figure 3 3 D and E). The intertwining of six large surface loops, two from each threefold symmetry related monomer, forms the mound like protrusion at the icosahedral threefold axes (Figure 3 3 D). These loops are
122 3B). The shoulder of the In the full H 1PV structure, ordered density for nine ssDNA nucleotides was observed in the capsid interior at the twofold axis in a conserved DNA binding pocket previously reported for other members of the parvovirus genus (Figure 3 4A) (7, 306) The chain direction of nucleotides was evident from th e positive difference density in the F o F c map for the phosphate groups, and the relative position of the sugars, phosphates and bases also aided in the model building. The good quality of the icosahedrally averaged electron density maps for the H 1PV data set enabled the clear distinction between purine and pyrimidine bases and for some nucleotides also between A or G, and C or T The DNA sequence 5' TGCCTTCAA 3' was built into the well ordered density (NTs numbered from 2 10 to correspond to the NT numbe ring in the CPV structure) (Figure 3 4A). The overall conformation adopted by the ordered ssDNA stretch in the H 1PV virions is similar to that found in MVMi and CPV, with the bases pointing outward to interact with the capsid protein while the phosphate d eoxyribose backbone is on the inside of the loop with the negatively charged phosphates held together by Mg 2+ ions (Figure 3 4A and Figure 3 5 ) (7, 306) The 5' end of the DNA strand is close to the fivefold axis C apsid protein DNA interactions include non specific (with backbone and deoxyribose sugar) and specific (nucleotide base atoms) van der Waals and polar interactions. The binding site contains several residues that are within acceptable distances (2.4 to 4 .0 ) for hydrogen bonding interactions such as L146, Q148, L184, S186, N187, I189, T272, Y275, I276, D480, N497, N498, P500 (from the reference monomer), D57, T285, P545, L547, H590 (from the three fold related
123 monomer), and T51, Y52, K53 and F54 (from th e fivefold related monomer) which are highly conserved among the autonomous parvoviruses (Table 3 3 ) and (Figure 3 6 A) The DNA strand is stabilized by base base stacking interactions, and there are two distinct stacks: one consisting of T6 T7 C8 and A1 0 and the other consisting of G3 and A9 ). The protein DNA complex structure is stabilized by extended DNA DNA base stacking as well as NT protein stacking with aromatic amino acids (Figure 3 6A ). For example, C 5 base stacks with H590. Apart from the ssDNA sequence, three cytosine molecules were also modeled into the full H 1PV structure (Figure 3 7 ) The capsid amino acid residues interacting with the cytosines are H483, R486 (Cyt1), W59, S140, N142, P281, K540 (Cyt2), and W59, K540 (Cyt3). The VP2 struct ures for the full and empty H 1PV particles structures superimposed with an overall rmsd of 0.47 for all 55 4 residues. In the empty H 1PV crystal structure, there was no density observed for the nucleotides, which is consistent with the data for other em pty autonomous parvovirus structures (168, 276, 322) The full and empty H 1PV structures were almost identical except for side chain conformation differences observed at the DNA binding pocket (Figure 3 8A ). In ful l H 1PV structure, H590 exists in two different conformations, interacting with the C5 nucleotide in one conformation. However, in the empty H 1PV structure it exists in only one conformation that faces slightly away from either of the two conformations ad opted in the full H 1PV structure. The other three detectable residue level changes were at K53, G56, and D57 Four cis peptide bonds ( 3 proline: A423 P424 I427 P428 Y469 P470 and one non proline cis peptide bond H348 D349) were observed in the H 1PV v irion structure ( Figure 3 9 ) The cis peptide bonds in the full H 1PV structure were identified based on
124 the backbone geometry and better fit in to the electron density map Residue P470 is located at the twofold helix The other cis peptide bonds (2 proline: A423 P424; I427 P428 and one non proline cis peptide bond: H348 D349) are at the threefold and interact with loops from the symmetry related molecules Two of the cis peptide bonds (423 424 and 427 428 ) are present in the same loop Structural Studies of H 1PV and Glycan Complexes Two of the glycans ( 3' SIA (LN ) 2 and 3' SIA (Le x ) 3 ), recognized by empty H 1PV capsids when screened on the glycan array were co crystallized with empty H 1PV particles and X ray diffr action data was collected on the co crystals for structure determination The structure of H 1PV 3' SIA (LN ) 2 and H 1PV 3' SIA (Le x ) 3 co crystal structures were determined to 2.7 and 2.9 resolution, respectively. The data refinement statistics are summariz ed in Table 3 1. For the H 1PV 3' SIA (Le x ) 3 complex data, positive density was observed in the F o F c difference maps at at the depression at the icosahedral twofold axis of symmetry This site was also identified as the S I A binding site for MVM in previous studies (188) (Figure 3 10A B and D ) and was modeled as a single molecu le of SIA R esidues, L248, N249, R355, D367, I368, D405, A406, a nd S409 interact (2.4 4.0 ) with the S I A molecules (Figure 3 10C) I368 and D405 in H 1PV are homologous to the MVMp residue s I362 and D399 which are known to be involved in tissue tropism and pathogenicity determination for MVMp (188, 259) For the modeled SIA the N acetyl group interacts with D367, I368 D405 and A406; the glycerol group interacts with S409 L248, and N249 ; and the carboxyl group interacts with R355. For the H 1PV 3' SIA (LN ) 2 complex data set, the F o F c difference map showed density that could be modeled as a S I A in a pocket adjacent to the
125 previously identified MVMp S I A binding site (Figure 3 10 C ) (188) This binding site on the H 1PV capsid is narrower than the MVMp glycan binding site. The mod eled S I A in this site interacts with charged and hydrophobic residues such as Q326, N327, D373, H374, T415, A423, P424 and R437 Residues N327 and H3 74 are homologous to MVMp residues G321 and K368, respectively, which are known tissue tropism and pathogenicity determinant s for MVM (15, 112, 188, 196, 259) The gly cerol chain of SIA sits in a pocket lined with D 373, T415, R437 (D367, T409, R431 in MVMp) the ri ng structure interact s with H374, A423 and Q 424 (K368, A417, P418 in MVMp), and the N acetyl group interact s with r esidues Q326 and N327. There were no detectable changes observed in the capsid upon recepto r attachment. Discussion Recognition of Sialylated Glycans by H 1PV Empty H 1PV particles were screened on a glycan array containing 377 different glycan motifs and specifically recognized only three structures with terminal sialic acid linked 3 to a co mmon 4GlcNAc motif; SIA Le x Le x Le x or 3' SIA (Le x ) 3 ; glycan 227) SIA LN LN or 3' SIA (LN ) 2 ; glycan 236) and a bianten nary glycan (Figure 3 2). Complex N glycans are reported to have two types of LacNAc (Gal 4GlcNAc ( T ype2 ) of which T ype2 repeats are more common. Specifically H 1PV did not recognize any sialylated glycans linked to T ype 1 LacNAc motif and only bound to glycans with the common motif Neu SIA LN) similar to observations with
126 MVM (209) This suggests that the type of Gal GlcNAc linkage plays a rol e in the receptor capsid interaction for H 1PV and MVM, and preliminary modeling studies with 3 GlcNAc ( Type1 ) in the SIA binding site of H 1PV and MVM indicates that the 3GlcNAc linkage creates steric hindranc e in the binding site H 1PV bound with high specificity to 3' SIA (Le x ) 3 ( SIA Le x Le x Le x ; glycan 227) which was also recognized by MVM in a previous glycan array screening (209) and is a known tumor cell marker. Another oncotropic rodent parvovirus, LuIII also showed recognition for the 3' SIA (Le x ) 3 in glycan array scre ening (unpublished data). This suggests a common theme of utilization of the 3' SIA Le x motif by the oncotropic rodent parvoviruses, MVM, H 1PV and LuIII to bind and infect transformed cells The LN LN linkage (GlcNAc 3Gal) in glycan 236 that was recognized by H 1PV SIA LN LN glycan ( 4GlcNAc ) recognized by MVM in previous glycan array studies but is unfortunately not present in the current array (209) The biantennary glycan (glycan 316) has two terminal sialic acids; 3 is on the 3 branch of 6 is on the 6 branch of the 6 and also when 3 or both 6 b inding to H 1PV is abolished. This suggests that in the context of a biantennary glycan, H 1PV recognizes 3 SIA only on the 3 branch and 6 SIA only on the 6 branch, while it won't bind to a single chain glycan with 6 linked SIA However, MVM virus es screened on the SGM (data reported in chapter 2) identified a biantennary glycan with terminal SIA linked only 3 to both the 3 branch and 6 branch. H 1PV and MVM have some
127 common properties, including their recognition of similar glycans, but there a re some differences in their properties which may be capsid mediated. Structure of H 1PV The H 1PV VP2 structure has the conserved parvovirus capsid viral protein topology, consisting of a core eight stranded anti barrel motif with large loop i nsertions between the strands (53) (Figure 3 3B) The VP1u and the VP2 N terminal residues (1 37) were not observed in the crystal structure. Low copy numbers of VP1 and VP2 in the mature virions or different conformations adopted by the VP1/2/3 N termini, which is incompatible with the 60 fold icosahedral averaging appl ied during structural determination, are postulated to result in the lack of N terminal VP ordering (53) In the full virions of other autonomous parvoviruses, such as MVM, CPV and also for the dependoviru s AAV8, low sigma threshold electron density was observed inside the capsid, under the twofold axis, running between fivefold related VP monomers, and also within the fivefold channel (4, 7, 210, 306) In MVM and CPV, this density was modeled as the conserved glycine rich N terminal region of one VP2 subunit spanning residues 28 t o 38 to represent the externalized VP2 In MVM, additional density corresponding to residues 36 to 38 from the remaining four fivefold related VP was modeled as extending back into the particle interior at the base of the fivefold cylinder (7) The lack of ordering of the residues that form the tip of the DE loop in full and empty H 1PV particles might be due to the r eported dynamic nature of the fivefold cylinder (168) as is required for its reported function as a portal for VP2 and VP1 u externalization and genome encapsidation (75, 100 102) The structural transitions that occur during these processes might not be compatible with icosahedral avera ging carried out during structure determination process
128 Structural Comparison to Other Autonomous Parvoviruses A structural superposition of the ordered VP2 region of H 1PV with the analogous VP2 structures of other autonomous parvoviruses MVM, CPV, FPV a nd PPV shows that topologies (Figure 3 11 ). While the VP2 sequence identit y between these parvovirus members range s from 52% to 68%, the structural homology ranges from 7 5 % t o 9 0 % (Table 3 2) with an overall rmsd 1PV VP2 and these viruses ranging from 0.63 0.94 (Table 3 2). The high structural homology is due to the high sequence identity for the barrel core while the surface loops are t he most varied in sequence and show structural variations in between the members. In Kontou et al (168) of as much as 5 ) between these viruses (except for H 1PV) were identified and numbered 1 to 8. These regions also differ between the highly homo logous parvovirus strains, such as MVMi and MVMp, CPV and FPV, and CPV and its host range mutants. These variable loop regions are clustered on the assembled 3D capsid at and around the icosahedral twofold axis, on the shoulder of the threefold protrusion and at the fivefold axis. Prominent differences between H 1PV and the other parvovirus structures are also observed at the threefold protrusions, twofold depression and at the fivefold axis (Figure 3 11 A and B). The majority of these differences are locat ed in the previously defined variable regions, including VR1 (152 176), VR2 (231 253), VR3 (297 311) VR4a (314 340), VR4b (421 433), VR5 (361 380), VR6 (388 398), VR7 (508 525) and VR8 (553 572). In addition, H 1PV differs from MVM in a surface loop cons isting of residues 88 100 located at the tip of the threefold protrusion due to a four amino acid insertion in H 1PV. This VR was not identified in the previous study by Kontou et al
129 (168) a nd is labeled as VRX. Variable region 2, VR4a and VR4b contribute to the threefold protrusions and have been mapped as antigenic sites on the MVM and CPV capsids (125, 157, 321) Differences in VR8 (between H 1PV and MVM) and VR5 (for all parvoviruses compared), contribute to capsid surface variation at the floor of the twofold depression, and at th e wall of the twofold depression, respectively. These capsid regions serve as the SIA receptor binding sites in MVM and CPV in addition to being involved in tissue tropism and pathogenicity determination for these and other parvoviruses (5, 19, 188) In a recent study (226) it was shown that the rodent parvoviruses H 1 PV, LuIII, and MVM exhibit variation in tumor tropism and oncolytic activity of LuIII was mapped to the capsid VP2 gene. These obs ervations suggest that the VP2 structural variations observed between MVM and H 1 PV might dictate the differences in tumor tropism In an effort to correlate the tropism of H 1PV with that of the MVMp F1 hot switch mutants that adapted to infect rat fibro blasts, the conformation of the residues in H 1PV that are homologous to the residues mutated in the MVMp F1 virus (334, 384, 554, 578 in MVMp correspond to 340, 390, 560 and 584 in H 1PV) were analyzed Out of the four host range switch mutations in vitro in MVMp that conferred the ability to grow in rat fibroblasts (similar to H 1PV's cell tropism), only two of those (residues 560 and 584 in H 1PV) are structurally variable between MVMp and H 1PV and are located at the twofold depression and might play a role in the virus adaptation to a new host. A comparison of the available dependovirus capsid structures also identified similar variable regions that are involved in various functions such as transduction, receptor recognition and antibody binding (119) Comparative analysis of the parvovirus structures provides inform ation on the capsid interactions at the
130 intermonomer interfaces that are required for efficient assembly, and interdigitation of the loops which enables identification of capsid regions that can tolerate loop insertions and deletions. Recent developments i n gene therapy vector design include engineering of mosaic/chimeric vectors by shuffling capsid genes from related viruses in an effort to design vectors with specific tissue tropism and reduced immunogenicity Structure guided vector design identifies the capsid regions of parvoviruses that are not conducive to peptide insertions because they would either create steric hindrance at the symmetry axes due to mismatched loop sizes or lack residues involved in stabilizing inter monomer interfaces. DNA Binding Pocket The ordered density for ssDNA in the full H 1PV structure was observed in the capsid interior in a cavity present at the interface of three VP monomers (reference, threefold related and fivefold related monomer). Ordered density for ssDNA has also b een observed in an analogous region of the capsid structures of other autonomous parvoviruses such as CPV and MVM thus providing additional proof of the importance and existence of this conserved DNA binding pocket (Figure 3 5 ) (7, 306) However, a s mentioned in previous reports, the ordering of the DNA density is unusual because the structure determination procedure assumes icosahedral symmetry (or 60 equivalent positions) Since the capsid packages only one copy o f the genome and it has been seen that the genome sequence observed in such viruses has a low degree of repetition within the genome it is not possible that the exactly same sequence is present in 60 equivalent sites in the capsid (329) Electron density for a si ngle nucleotide has been observed in the capsid structures of AAV4, AAV6 and AAV8 at the icosahedral threefold axis (119, 177, 210, 214) While in the autonomous parvoviruses, ~12 to 25% of the
131 genome is ordered, in the dependoviruses only ~1% is ordered. This is suggested to be due to the packaging of predominantly the negative sense strand (or a single polarity) of the genome by the autonomous parvoviruses (except for LuIII) and packaging of both sense strands by t he dependoviruses inside different capsids. Therefore, the AAV crystals would contain different populations of capsids, which would result in the heterogeneity of DNA sequence in the crystallized virus, thus decreasing the probability of visualizing a long er DNA sequence in the crystal structure after icosahedral and crystal averaging applied during structure determination The amino acid residues in the nucleotide binding pocket of the parvovirus genus members and dep endoviruses differs, so specific capsid DNA interactions could also dictate the amount of ordered DNA. Structure determination of LuIII, a member of the parvovirus genus that packages either sense strand in different particles, would aid in determining if the structural ordering of ssDNA is dic tated by genome sequence and packaging constraints and/or a conserved DNA binding pocket in the capsid. Similar to the autonomous parvoviruses, i cosahedrally ordered ssDNA was also observed in the B X174 (197) Icosahedrally ordered RNA has been found in Flock House Virus where non specific interactions were observed between the sug ar phosphate backbone and the capsid protein (104) and in bacteriophage MS2 where the capsid protein i s involved in extensive contacts with nucleotide bases (308) The nucleotides in H 1PV capsid are involved in specific interactions with capsid surface amino acid residues that a re highly conserved among the autonomous parvoviruses (Table 3 3 ). H590 exists in a dual conformation in full H 1PV particles and in one conformation base stacks with C5 T6 T7, C8 and A10 form a stacking ladder, so
132 for C5 to flip out of the stabilizing stacking ladder to interact with H590, indicates the high specificity of this interaction and an important role of H590 in genome packaging (Figure 3 6 A ) One of the modeled cytosine s ( labeled Cyt1 in Figure 3 7 B ) is located in the conserved AAV nucleotide binding pocket under the icosahedral threefold axis and interacts with the conserved residue, R486 (Table 3 4) Although the DNA binding pocket in parvovirus genus members is highly conserved, c omparison of the ssDNA sequence observed in the autonomous pa rvoviruses, H 1PV, MVMi and CPV, shows that t here are less nucleotides (NTs) ordered in the H 1PV structure (9 NTs; Chain C) compared to the MVMi (21 NTs; Chain B and C) and CPV (11 NTs; Chain B) structures (Figure 3 4) (7, 306) There were also differences in the ssDNA sequence observed in H 1PV, MVM and CPV structures (Table 3 5 ). T2 (base followed by nucleotide number) interacts with highly conserved residues (L146, Q148, p500, T272, Y275), and Y275 interacts with the methyl group, so T is observed in this site for H 1PV, CPV and MVMi. Y275 and N497 interact with the amino group of G3 (guanine 3 ) and I276 interacts with the carboxyl group, so only a G or C can be accommodated at this site. Since I276 differs betwee n these viruses at this site, it allows the accommodation of A in CPV and C in MVM. C4 is present in the three viruses as it interacts only with the highly conserved N497. At nucleotide position 5 only a pyrimidine is allowed because the amino acid residu es (T285, P545, l547, H590) interact with the carboxyl group and the carbon at position 2 and 4 of the pyrimidine ring. Also, since the residues at thi s site differ between the three viruses, either a C or T could be accommodated P545 interacts with methy l group, T51 with the carboxyl group, Y52 with the amine and F54 with the carboxyl group at position 4 of T6 Due to the recognition of methyl group by
133 P545, only a T can fit into the H 1PV pocket at this site For the other viruses, this site contains var iable residues. K53 interacts with the carboxyl group at position 2 of T7 so only a C or T can be accommodated at this site (C in CPV). D57 and K53 interact only with the amino group in C8 so a lot of variation is allowed. D57 interacts with the amino gr oup of A10 so even at this site either A G or C can be accommodated An analysis of the nucleotides and binding pocket of these capsids that are ~50% identical at the nucleotide and amino acid level, shows that the common interactions allow for DNA packa ging and stabilization while the differences between the structures dictate the packaging of a specific genome sequence. It has been demonstrated using a MVM and LuIII chimeric virus that secondary structure elements, including stem loops and guanidine ric h regions, can interrupt packaging (72, 73) Further analysis of the DNA binding pocket reveals that the amino acids interacting with the DNA are associated for other non viral proteins which contain oligonucleotide binding motif (OB fold) (298) The conservation of the DNA binding pocket suggests that this DNA protein interaction may play an important role in the virus life cycle, such as in capsid assembly or stabilization. An an alysis of the mechanical properties of the empty and full MVM capsids suggested that the bound genome reinforces the stiffness of the particle, especially at the twofold symmetry axis (49) In the study by Reguera et al (247) alanine mutants of t hree of the ssDNA interacting residues in MVM (Y270, D273, D474 corresponding to F2 66, D269, D475 in CPV, and Y275, N278 and D480 in H 1PV respectively ) displayed drastic reduction in infectivity, with the Y270A defective in capsid assembly and D273A unable to encapsidate genome, while there were other mutants
134 (D 58, W 60, N 183, T 267, or K 471 ) that were less stable. These results further propose the role of capsid DNA non covalent interactions in maintaining capsid stability. The same study showed that at least one acidic residue at each DNA binding pocket is required for genome packaging. Conformational Variation Between Full and Empty Capsids To examine possible conformational modifications imposed on the capsid proteins upon ssDNA binding the full and empty H 1PV capsid structures were superimposed for comparison. The structures superim posed with an overall rmsd of 0.47 for all 554 residues. The structures were almost identical except for side chain conformation differences observed at the DNA binding pocket at H590, K53, G56 and D57 (Figure 3 8 A ). Comparisons of full and empty structur es of CPV and MVM show variation at the wherein the loop rearrangement increases the diameter of the opening in the empty capsid structure (Figure 3 8B and C) (7, 168, 322) D ue to the lack of ordered density for this loop in both the empty and full capsids of H 1PV, no conclusions can be made. Large main chain and side chain displacements were observed at the ssDNA binding site in full and empty CPV particles, compared to small displacements in the full and empty MVM structures. No ordered density for DNA was observed in the empty H 1PV capsids which is consistent with the structures of empty particles of other members of this genus (168, 276, 322) However, AAV VLPs package small amounts of cellular DNA and in crystal structure of AAV VLP, ordered density for a single nucleotide is observed, which suggests that in the case of the dependoviruses, capsid DNA int eractions might play a role in capsid assembly (168, 276, 322)
135 Cis Peptide Bonds An analysis of the protein structures in the PDB database conducted in Weiss et al (318) and Jabs et al (140) found only 0.03% of all X Xnp, and 5.2% of all X Pro peptide bonds to occur in the cis conformation (where X=any residue, Xnp=any non Pro residue). A Pro Pro bond has the highest frequency to be in the cis for m, followed by Tyr Pro. The cis peptide bonds, especially non Pro cis peptides are rare due to the steric strain but when present they are usually located near functional sites and have been suggested to be important to a s a kind of energy reservoir that drives conformational changes (40) As an example, an Ala Phe cis peptide bond at the binding site for the antiviral assembly inhibitor drug CAP 1 in the HIV capsid protein was suggested to facilitate capsid assembly (159) Residue P470 was also shown to exist in the cis conformation in the CPV capsid structure (322) It is conserved among the members of the parvovirus genus but was not refined as a cis peptide in any of the other structures and is located at the twofold (Figure 3 9) The other cis peptide bonds ( 2 proline: A423 P424; I427 P428 and one non proline cis peptide bond: H348 D349) are at the threefold and interact with loops from the symmetry related molecules (Figure 3 9). Two of the cis peptide bonds (423 424 and 427 428 ) are present in the same loop The loops that are involved in the interactions at the threefold axis of symmetry are very closely interdigitated and the presence of the cis peptide bonds coul d impart flexibility to the loop. As suggested in Riolobos et al (248) if the MVM capsid assembles via trimeric intermediates then these cis peptides could play a major role in folding /unfolding of the loops to al low efficient capsid assembly.
136 Glycan Binding Site on H 1PV Capsid Preliminary structural studies o f the empty H 1PV capsids complexed with 3' SIA (Le x ) 3 glycan showed that the SIA receptor binding site is the same as identified for MVMp in previous studies and in the complex structures determined in c hapter 2 (Figure 3 10 A ) (188) Only the SIA component of the glycan was observed in the crystal structure and modeled into the F o F c moieties at the reducing end may not be forming tight interactions with the capsid surface or do not bind to all the sites with an equal occupancy or they are very flexible and dynamic or adopt different conformations. These possible scenarios would be inconsistent with the icosahedral symmetry imposed during structure determination, and would lead to lack of ordering of the density for these molecules. Although structurally homologous residues (D405, I368 in H 1PV correspond to D399 and I362 in MVMp respectively) were involved in the SIA recognition, the surface of the pocket and the side chain conformation of the interacting residues is different from that observed in MVMp. The pocket in H 1PV is wider a nd shallower in comparison to MVMp (Figure 3 10). Comparative analysis of the structures of 3' SIA (Le x ) 3 complexed to MVM ( data presented in chapter 2) and H 1PV, shows that although interaction with I368 and D405 (I362 and D399 in MVMp) is conserved, the g lycan binds deeper into the H 1PV pocket and interacts with residues, D367, R355, A406, L248, N249 and S409 (correspond to residues D361, R349, E400, M243, N244 and F403 in MVMp) that it does not interact with on the surface of MVMp or MVMi capsid. It has been shown that H 1PV and MVM vary in their oncotropic properties, and this could be regulated by the differences in their binding interactions with 3' SIA (Le x ) 3 the tumor cell marker (226) M VMp K368R and MVMp I362S virulent mutants ha ve been shown to have a reduced affinity for the sialic
137 acid receptor, and a recent study reported that analogous substitutions introduced into the H 1PV capsid (H374R and I368S) resulted in an even more drastic reduction in cell binding as compared to the MVMp mutants (10, 188) This suggests that SIA is an important determinant of H 1PV and MVM's infectivity, and common residues regulate their SIA binding but there are differences in receptor affinity ( demonstrate d by differences in sensitivity to neuraminidase treatment) that results in differences in their in vivo tropism. Also, interactions with other amino acid residues in the pocket may affect the receptor binding affinity. For the H 1PV 3' SIA (LN ) 2 complex, th e shape and composition of the SIA binding site was different from that identified for MVMp or the H 1PV 3'SIA (Le x ) 3 complex (Figure 3 10D ). This binding site is narrower and occluded. In this site, the glycan interacts with residues homologous to MVMp, su ch as H374 and Q326 (corresponds to residues K368 and G321 in MVM) that are known to dictate tissue tropism and pathogenicity in MVMp. The refinement of the H 1PV co crystal structure s needs to be completed to be able to fully define the role of the speci fic amino acids in dictating the mechanism of cell recognition Summary Glycan array screening of empty H 1PV particles showed that it recognized only 3 sialylated glycans linked to a common T ype2 LacNAc ( Gal 4 GlcNAc ) motif, similar to observations with MVM (209) The recognition of 3' SIA Le x a tumor cell marker, indicates that H 1PV like MVM also utilizes 3' SIA Le x motif on cancer cells to recognize and bind to these cells. The overall topology of VP of H 1PV con tains a highly conserved eight stranded anti barrel core with large loop insertions between the strands that forms the capsid surface, as is previously reported for other parvovirus capsid structures. The H 1 PV full and empt y capsid structures a re identical except for
138 slight conformational variation s in side chain amino acids at a conserved nucleotide binding pocket previously reported in other autonomous parvoviruses. Electron density interpretable as nine nucleotides was observed at the icosahe dral twofold axis in a conserved DNA binding pocket previously reported for other members of the parvovirus genus The capsid surface amino acid residues are involved in specific interactions with the nucleotide. Four cis peptide bonds ( 3 proline and one n on proline cis peptide bond s ) were observed in the H 1PV virion structure that may play a role in capsid assembly A comparison to other autonomous parvoviruses identified the most significant structural differences on the capsid surface at the loop region s surrounding the icosahedral two three and fivefold axes at residues reported to control receptor binding, tissue tropism, pathogenicity, DNA packaging, and antibody recognition. For the H 1PV 3' SIA (Le x ) 3 complex, SIA bound at the same site as SIA bin ding site on MVMp, however, there were differences in the shape and composition of the pocket which might dictate the differences in the oncotropism between these two viruses. For the H 1PV 3' SIA (LN ) 2 complex, SIA was modeled into a pocket adjacent to the previously identified MVMp SIA binding site The SIA binding on H 1PV capsid involves interactions with residues I368, H374, Q326 ( corresponds to residues I362, K368 and G321 in MVM), that are known tissue tropism and pathogenicity determinants for MVM.
139 Table 3 1. Data processing and refinement statistics Parameter Empty H 1PV Full H 1PV H 1PV 3' SIA (Le x ) 3 H 1PV 3' SIA (LN ) 2 1.0895 0.9186 0.918 0.917 Space Group P2 1 P2 1 P2 1 P2 1 Unit cell parameters (, ) a=255.2, b=350.1, c=272 .1, a=255.4, b=350.4, c=271.6, a=258.0, b=348.0, c=272.1, a=258.4, b=347.9, c=272.4, Resolution () 40 3.2 (3.3 3.2 ) a 50 2.7 (2.8 2.7) 40 2.9 (2.9 2.8) 50 2.7 (2.8 2.7) No. of unique reflections 479,478 (45,513) 1,2 10,268 (123,000) 725,031 (63004) 993,289 (102,849) Completeness (%) 61.3 (58.3) 93.1 (94.7) 68.7 (59.8) 75.8 (78.7) I/ 4.9 (1.3) 6.9 (2.4) 6.5 7.2 (2.9) R sym b (%) 11.4 (40.0) 13.2 (37.8) 12.1 12.8 (33.9) No. of atoms (protein/H 2 0/DNA/ion/othe r solvent) 4371/69/0/0 /0 4375/98/205/4/8 4371 /0/0/0/21 4360 / 0 / 0/2 /21 Average Bfactor ( 2 )(protein/H 2 0/DNA/ion/other solvent) 6 2 /57 / / / 30/30/73/46/44 38.9/ / / / 27.6/ / /30 / R factor c / R free d (%) 25.0/25.4 21.5/21.6 27.5/27.7 20.1/20.2 Rmsd bonds () and angles () 0.007, 1.35 0.006, 1.36 0.006, 1.38 0.006, 1.38 Ramachandran Statistics (%) Most favoured/Allowed/Outliers 91.8/ 8 96.5 3.5 94.7 2 96.5/ 3.5 a Values in parenthesis are for highest resolution shell. b R sym where I is the intensity of a reflection with indices h, k, l and is the average intensity of all symmetry equivalent measurements of that reflection. c R factor o | |F c o |)100 where F o and F c are the observed and calculat ed structure factor amplitudes, respectively. d R free is calculated the same as R factor except it uses 5% of reflection data omitted from refinement.
140 Table 3 2. RMSD in C alpha positions for the parvovirus structures Autonomous Parvovirus % Identity Max imum RMSD in C position between H 1PV and the other autonomous parvoviruses for VR regions() Overall VRX VR1 VR2 VR3 VR4a VR4b VR5 VR6 VR7 VR8 MVMp 68 a (89) b 0.63 1.7 +gap 1.5 +gap 3.3 +gap 1.7 1.2 +gap 1.8 1.5 1.7 2.2 1.6 +gap MVMi 68 a (78) b 0.68 2.2 +gap 2.5 +gap 3.9 +gap 2.8 1.4 +gap 1.6 1.0 1.2 2.4 2.6 +gap CPV 53 a (89) b 0.94 1.2 +gap 2.2 +gap 2.7 +gap 2.9 +gap 2.7 +gap 1.7 3.6 +gap 2.2 2.8 +gap 3.3 +gap FPV 52 a (81) b 0.87 0.9 +gap 4.3 2.8 +gap 3.0 +gap 2.6 +gap 2.2 3.7 +gap 2.2 +gap 1.9 +gap 3.5 +gap PPV 53 a (7 4) b 0.94 2.0 +gap 2.3 +gap 3.3 +gap 3.5 +gap 1.8 +gap 1.7 1.9 3.9 +gap 3.9 +gap 3.2 +gap a Values are for VP2 sequence identity b Values in parenthesis are for VP2 structural identity Table 3 3. Amino acid sequence comparison for autonomous parvovirus es at nucleotide binding pocket Virus 51 52 53 54 57 146 148 184 186 187 189 272 275 276 285 480 497 498 500 545 547 590 H 1PV T Y K F D L Q L S N I T Y I T D N N P P L H MVM H Y R F D L Q V S N I T Y Y T D N N P A T R CPV E F K F N F Q L S N T T F F T D N N P A H R FPV E F K F N F Q L S N T T F F T D N N P A H R PPV E F Q Y E F Q L T N T T Y H T D N N P S N R *Identical residues are shown in red, similar in green and residues that are different in black.
141 Table 3 4. Amino acid sequence comparison in cytosine binding pocket Virus 59 140 142 281 540 483 486 H 1PV W S N P R H R MVM W Q N P R H R CPV W E H P K L R FPV W E H P K L R PPV L E N S T L R AAV2 G H AAV4 G H AAV8 G H *Similar residues are shown in green and residues that differ between these viruses are shown in black. Table 3 5. The ssDNA sequence present in H 1PV, CPV and MVMi Parvovirus Nucleotide sequence H 1PV (9 ntds) TGCCTTCAA CPV (11 ntds) ATACCTCTTGC MVMi (21 ntds) ATCCTCTATCAC chain B ACACCAAAA chain C
142 Figure 3 1. Virus purification, crystallization and diffraction. (A) An SDS PAGE of the purified empty H 1PV and full H 1PV particles Electron microscopy images of (B) empty H 1PV viewed at 60,000X magnification and (C) full H 1PV viewed at 100,000X magnification. (D) Optical photograph of H 1PV crystals showing the rod shaped and plate shaped crystals at 4X magnification. ( E ) A diffraction image of the full H 1PV crystal. The concentric rings indicate the 10.0 5.0 and 3.0 r esolution shells.
143 Figure 3 2. Glycan array data for H 1PV. (A) The glycan array data for H 1PV is shown. The bars represent relative fluorescence for a given glycan. The y axis represents relative fluorescence units (RFU) and the x axis represents the glycan numbers on the array (1 377). The glycans showing specificity to H 1PV are annotated in the cartoon form. (B) Cartoon representation of the glycans recognized by H 1PV. (C) Glycan symbols.
144 Figure 3 3. Structure of H 1PV (A) Section of the 2F o F c electron density map contoured at 1.5 (gray mesh) of H 1PV virions ( orange ) for residues F220 to P222 that c orrespond to the MVMp ( red ) sequence C216 to D218. The H 1PV and MVMp coor dinates are shown in stick form (B) Ribbon diagram representation o f H barrel (yellow) containing strand structure (orange) t he first N terminal residue observed (38), the C terminal r e sidue (593), DE and HI loops are labeled. Approximate positions of icosahedral two three and fivefold symmetry axes are depicted as filled ovals, triangles, and pentagons, respectively. (C) H 1PV VP2 related to the r ef erence (Ref) monomer by fivefold ( 5f2 5f5), threefold (3f1 3f2) and twofold (2f) symmetry relationships. The black triangle depicts a viral asymmetric unit. ( D) Depth cued s urface representation of the H 1PV capsid showing the topological features. (E) Ribbon diagram showing the interactio n between fivefold related monomers. The HI loop is shown in blue.
145 Figure 3 4. ssDNA observed in H 1PV virions. (A) The ssDNA sequence interpreted as a stretch of 9 nucleotides is shown inside the 2F o F c electron density map (gray mesh) and contoured bases and the phosphodiester backbone is nucleotides are labeled. The Mg 2+ ions are shown as green spheres and labeled ( B ) A cross section of the interior surface of full H 1PV capsid is s how n in surface representation ( orange ) viewed down the two fold axis The ssDNA cage that bind s to the cavities at the two fold axis is shown in blue spheres (C) A 180 side view of the image in (B) showing the ssDNA cage. Approximate positions of icosahed ral two three and fivefold symmetry axes are depicted as red filled ovals, triangles, and pentagons, respectively. These figures were generated with PyMol program (82)
146 Figure 3 5. ssDNA topology observed in members of the parvovirus genus. (A and B) The conserved nucleotide binding pocket showing the ssDNA interacting with the VP2 of H 1PV ( orange ), MVM i ( red ) and CPV ( green ) at the capsid interior at the two fold axis. B ) Close up view of the superposition of the ssDNA observed in H 1 PV MVMi and CPV virions. Approximate posi tions of icosahedral two three and fivefold symmetry axes are depicted as black filled ovals, triangles, and pentagons, respectively. These figures were generated with PyMol program (82)
147 Figure 3 6 ssDNA protein interactions in autonomous parvoviruses ssDNA protein interactions in ( A ) H 1PV, (B) CPV and ( D ) MVMi are shown Amino acids within 2.4 to 4.0 of the ordered DNA density are depicted in the stick model and labeled. The interacting residues are colored differently in yellow, pink and blue to show the contribution from the symmetry related monome rs. The sugars and bases are colored according to virus type : H 1PV (orange), CPV (green) and MVM i (red), and the phosphodiester backbone colored The Mg 2+ ions are depicted as magenta colored sph eres.
148 Figure 3 7 Cytosines observed in the full H 1PV structure. The amino acid residues interacting with the three cytosines are shown as green stick model colored according to atom type The cytosines are colored pink. In (B) the cytosine found i n AAV6 structure is superposed on the full H 1PV structure and is colored yellow. The nucleotides are shown inside the 2F o F c difference density map (gray mesh) contoured at 0.8 nucleotides are labeled.
149 Figure 3 8 Conformational variation between empty and full capsids. (A) Conformational variation observed between VP2 structure of the full H 1PV (yellow) and empty H 1PV capsids ( cyan ) at th e nucleotide binding pocket are shown. The sugars bases and the phosphodiester backbone are colored orange. In (B) and (C) the full (green) and empt y capsids (yellow) of CPV and MVM are compared, respectively. The sugars and bases are colored blue and t he phosphodiester backbone is colored pink. The amino acid residues and the nucleotides are labeled The Mg 2+ ions are depicted as magenta colored spheres.
150 Figure 3 9 Location of the cis peptide bonds observed in H 1PV structure. Ribbon diagram sh owi ng the secondary structure of H 1PV VP2 related to the reference (Ref) monomer by fivefold (5f2 5f5), threefold (3f1 3f2) and twofold (2f) symmetry relationships. The amino acid residues involved in cis peptide bonds are depicted as spheres: P470 (orange), P424 (blue), P428 (red) and H348 (cyan). The black triangle depicts a viral asymmetric unit. Approximate positions of icosahedral two three and fivefold symmetry axes are depicted as filled ovals, triangles, and pentagons, respectively. This figure w a s generated with PyMol program (82)
151 Figure 3 10 Sialic a cid binding site on the H 1PV capsid. Surface representation of a close up of the depression at the icosahedral twofold axes of the capsid showing the SIA binding site on H 1PV. In (A and B ) the S I A binding site identified in the H 1PV 3'SIA (Le x ) 3 complex is shown. The residues that interact with S I A are shown in stick model and are in the same color as the surface. The SIA is shown in stick model (carbon atoms in green) inside a 2F o F c density map (dark grey) at contoured at In ( C ) the S I A modeled (ca rbon atoms in blue) for the H 1PV 3'SIA (LN ) 2 complex is shown. In (D), the SIA (carbon in purple) modeled on the MVMp capsid surface is superimposed on the H 1PV capsid surface. The approximate location of the icosahedral twofold axes is shown by the fille d oval.
152 Figure 3 11 Comparison of VP2 structures of H 1PV, MVM, PPV, FPV and CPV (A) Superimposition of the coil representation of VP2 monomer structures for members of parvovirus genus, H 1 PV (orange), MVM (red), PPV ( yellow ), FPV (pink), and CPV ( green ). The common variable regions are labeled. The DE and HI loops are labeled. The approximate icosahedral twofold (filled oval), threefold (filled triangle), and fivefold (filled pentagon) axes are shown. (B) Close up views of the regions inside dashed boxes These figures were generated with PyMol program (82)
153 CHAPTER 4 PARVOVIRUS CAPSID DY NAMICS ASSOCIATED WI TH ENDOSOMAL TRAFFICKING Background The ssDNA parvoviruses enter host cells through receptor mediated endocytosis, and undergo endosomal pH mediated processing in the early to late endosomes as well as i n lysosomes prior to nuclear entry for replication ( infectious life cycle of parvoviruses described in detail in Chapter 1). It is known that the capsid undergoes structural transitions during the entry process so as to enable the (i) externalization of V P 1u that harbors PLA 2 activity essential for during endosomal escape and also nuclear localization signals (NLS) required for nuclear entry, (ii) exposure of the VP2 N terminus that contains nuclear export signal (NES) required to traffick newly formed vir ions out of nucleus and the cleavage of VP2 N terminus to form VP3 in the autonomous parvoviruses that undergo a maturation step, and (iii) ex trusion of the genome Mutagenesis studies on AA V 2 and MVM have predicted the fivefold channel to be the portal fo r these exposures (32, 74, 100, 101, 121, 170, 240, 279, 323) These extrusion events can be mimicked using limited heat shock and/or coupled with pH treatment with the capsid remain ing intact (65, 67, 100, 132, 170, 191, 252, 312) However, the mechanisms of capsid endosomal processing during trafficking to the nucleus for genome replication are poorly understood. The available parvovirus capsid structures show that the resid ues within these N terminal regions cannot be accommodated through the fivefold channel without structural re ribbons and immunological studies support the "breathing" of this pore (6 5) To gain insight into the capsid structural transitions during infection, the crystal structures of AAV8 were determined at pH 4.0, pH 5.5, pH 6.0, and pH 7.5 after incubation at pH 4.0,
154 to mimic the conditions encountered during trafficking to early e ndosomes, late endosomes, lysosomes and endosomal escape to the cytosol prior to nuclear entry (208) Significant amino acid side chain conformational transitions were observed inside the capsid at the threefold axis and on the ca psid exterior at the icosahedral twofold axis At low pH, c onformational changes at the capsid interior disrupted the interaction of the VP with the ordered nucleic acid density observed in crystal structures and this was suggested to prime genomic uncoati ng. Structural changes in the capsid exterior at the icosahedral twofold axis result ed in a reduction in the number of inter monomer interactions at the twofold as p H was reduced. This of the interface was proposed to be involved in capsid dest abilization events that enable AAV VP1u externalization without c apsid disassembly. Similar conformational changes were observed in an equivalent site in AAV1 (unpublished data). The structural transitions induced in AAV8 at low pH were observed to be reve rsible that would stabilize the capsid following VP1u exposure. This was the first study on the dependoviruses characterizing the capsid dynamics at endosomal pHs It has been observed that the structurally variable capsid surface residues dictate the diff erences in tissue tropism and transduction efficiency observed among the AAV serotypes (unpublished data) For the autonomous parvoviruses, capsid dynamics have been structurally characterized only for CPV and FPV. T he capsid structure of CPV was determin ed at pH 7.5 and pH 5.5 to mimic the endosomal pH. Structures for CPV and FPV were also determined at pH 7.5 and pH 6.2 in the presence or absence of Ca 2+ ions to understand the differences in the hemagglutination ability of CPV and FPV (274) While, CPV is able to hemagglutinate at pH 6.2 and pH 7.5, FPV can only do so at pH 6.2 (52) This
155 study showed that the largest structural difference occurred in a surface loop (also ca lled the 'flexible' or hemagglutinating' loop in previous studies and consists of CPV VP2 residues 359 375) located at the wall of the twofold depression and colocalizes with the S I A binding site and host range determinants (274) This flexible loop was disordered in CPV at pH 5.5 and pH 6.2, and in FPV in the presence of EDTA. This loop binds to Ca 2+ ion in FPV (only at pH 7.5) but not in CPV ( at any of the studied pH s ) and is adjacent to a double Ca 2+ binding site which is conserved in both FPV and CPV. These results showed that the capsid dynamics in CPV and FPV are influenced by the Ca 2+ ion concentration and are related to their hemagglutinating activity (274) Utilizing p roteinases, minor structural variations in the surface loops of the CPV capsids were observed at pH 4.5 but not at pH 5.5 (213) However, these studies on CPV and FPV did not provide information on the capsid dynamics at pH 4.0 where the conformational changes observed in AAV8 were more pronounced. The capsi d surface residues at the twofold depression differ between CPV, FPV and MVM, thus it is likely that the local surface charge variation among the autonomous parvoviruses might result in different structural transitions during endosomal trafficking To unde rstand the mechanism of structural transitions observed in MVM capsids during endosomal trafficking, the structures of MVM capsids were determined in this study at pH 4.0, pH 5. 6 and pH 6. 0 to mimic the pH encountered in the early endosomes, late endosomes and lysosomes respectively Experimental Methods Virus Production and Purification The MVMp VLPs were produced and purified as described in the Experimental Methods section of c hapter 2. For the pH experiments the purified virus samples were
156 dialy zed i nto the appropriate buffers; pH 4.0 and pH 5.6 (150 mM Sodium Acetate buffer ), pH 6.0 (150 mM NaCl, 10 mM bis Tris HCl) T he purity and integrity of the capsids were verified by SDS PAGE and negative stain EM, respectively (as in chapter 2) Crystallizatio n, Data Collection and Processing Crystals of MVM p VLPs at the various pHs were grown using the hanging drop vapor diffusion method (199) with VDX 24 well plates and siliconized cover slips (Hampton Research, Laguna Niguel, CA, USA) in screening solutions containing PEG 8000 (1 3% w/v) and 150 mM NaCl in 10 mM b is Tris HCl (pH 6.0 ) or 150 mM Sodium Acetate buffer (pH 4.0 and 5.6). The drops reservoir solution at RT. For pH 6.0 and 5.6, crystals isomorphous to those for MVMp VLPs at pH 7.5 were obtained but for pH 4.0, Crystals obtained were soaked for 30 s in cryoprotectant solution cont aining the precipitant solution with 10% PEG 8000 and 30% glycerol and flash cooled in liquid nitrogen vapor prior to X ray diffraction data collection. Diffraction data for the MVMp pH 6.0 crystals was collected at the F1 beamline at CHESS on an ADSC Quantum 270 CCD detector for the MVMp pH 5.6 crystals at the SERCAT 22ID beamline at APS and for the MVMp pH 4.0 crystals at the X29 beamline at BNL on an ADSC Quantum 315 CCD detector. For MVMp pH 6.0, a total of 200 usable images were collected from one crystal with a crystal to detector distance of 300 mm, an oscillation angle of 0.3 per image, and an exposure time of 25 s per image at a = 0.918 0 A total of 500 usable images were collected from three crystal s of MVMp pH 5.6 00 with a crystal to detector
157 distance of 350 mm, an oscillation angle of 0.3 per image and an exposure time of 4 s per image. For MVM p pH 4.0, a total of 230 usable images were collected from two crystals with a crystal to detector distance of 300 mm, an oscillation angle of 0.3 per image, and an exposure time of 8 1.0809 The crystals diffracted X rays to beyond 3. 8 3.3 and 3. 2 resolution for pH 6 .0, 5.6 and 4 .0, respectively. The measured diffraction intensities were indexed and integrated with the HKL2000 suite of programs (224) and scaled and merged with SCALEPACK (224) The space group for the MVM pH 6.0, pH 5.6 and pH 4.0 MVMp crystals was determined to be C2. The unit cell volume for the pH 4.0 crystals was smaller and half of that of pH 6.0 and pH 5.6 crystals. The MVMp pH 6.0, pH 5.6 and pH 4.0 data sets scal ed with an R sym of 16.0 % ( 48.4 % completeness) 13.3 % ( 94.6 % completeness), and 11.9% (72% completeness), respectively. The data processing statistics are summarized in Table 4 1. Structure Determination The diffraction intensity data set s were converted to structure factor amplitudes using the TRUNCATE program from CCP4 (Collaborative Computational Project, Number 4) (62) For the MVMp pH 6.0 and pH 5.6 data, the C2 unit cell contains two half particles with different orientations in the crystallographic asymmetric unit, as described for the previously solved MVMi and MVMp VLP structures (7, 168, 184) The MVMp VP2 VLP structure coordinates (PDB accession no. 1 Z 14) (168) was used as the phasing model to initiate molecular replacement and refinement was carried out using the CNS program (46) as detailed in the Experimental Methods section of c hapter 2 The refinement statistics are given in Table 4 1. The figures were generated using program PYMOL (82)
158 For the MVMp pH 4.0 data, t s coefficient (V M ) was calculated to be 3.16 3 Da corresponding to a solvent contents of 61 % with two particles in the unit cell related by a crystallographic twofold axis and the icosahedral twofold axis coincident with the crystallographic twofold axis (i.e ., one half capsid or 30mer occupying a cr ystallographic asymmetric unit ) (195) The particle orientation in the crystal unit cell was determined with a self rotation function using the General Lock Rotation Function ( GLRF ) p rogram (299) computed with 10% of the observed data between 10.0 and 5.0 the twofold, threefold and fivefold icosahedral symmetry axes, respectively. The radius of integration was set to 120 orientation of the of the icosahedral symmetry axes relative to the crystal axes and confirmed that the two particles in the unit cell are related by the crystallographic twofold axis (b axis). A polyalanine model of the MVMp VP2 crystal structure ( PDB accession no. 1Z14) (168) was generated by the MOLEMAN program (167) and expanded to 3 0 subunits ( half capsid or 30 mer ) using the Oligomer Generator subroutine available at the VIPER database (50) for use as a molecular replacement phasing model (258) Mole cular replacement was carried out using the AutoMR subroutine in PHENIX (2) The best solution gave a L og L ikelihood Gain (LL G ) score of 2208 z score for rotation function of 42.3 and z score for translation function of 50.2 and R factor of 53.8%, with an orientation, in Eulerian angles 179.99 65.64 270.04 and fractional coordinate position of 0. 5003 0.0000, and 0. 0000 The initial phases for the oriented and positioned MVMp polyalanine model were further improved by refinement in the CNS program using simulated annealing, energy m inimization, conventional positional,
159 and individual temperature factor (B factor) refinement (46) while applying strict 3 0 fold noncrystallographic symmetry (NCS) The coordinates of the MVMp VLP structure at pH 7.5 (PDB accession no. 1 Z 14) (168) was superposed onto the refined MVMp structures at pH 6.0, pH 5.6 and pH 4.0 using the LSQ subroutine in COOT (97) to identify conformational changes Results Diffraction data sets were obtained for MVMp VLP at pH 6.0, pH 5.6 and pH 4.0. The MVMp crystals at pH 5.6 and pH 6.0 were isomorphous to those of MVMp VLP and wt MVMi for which structures are available (7, 168) Residues 39 to 587 of the MVMp VP2 were built into the MVM p electron density maps. The averaged density maps were not interpretable beyond N terminal residue 39 of VP2, as was previously reported for the MVMp and MVMi VLP structures (168) The resolution for the MVMp data at pH 6.0 is low ( 3.8 ) and the refinement stat istics ( R factor /R free is 40.2/40.6) are also not comparable to previously solved structures for MVM (7, 168, 184) The refined structure at resolution of 3.8 did not show any detectable differences in main chain o r side chain conformations in comparison to MVMp structure determined at pH 7.5. The MVMp pH 5.6 structure was determined from a very complete data set (94.6% complete) and the current refinement R factor /R free is 30.5 /30.6 The overall topology of VP2 at p H 5.6 is very similar to that at pH 7.5 with no main chain differences but significant amino acid side chain conformational changes at four regions on the capsid surface. The first is on the exterior capsid surface at the interface between the fivefold mo nomers and involves amino acid residue E79. Reducing the pH to 5.6 disturbs the interaction of E79 with the neighboring residues, R500 and R518 from the same monomer and K204 from a fivefold related monome r (Fig ure 4 1 ) The second change is on the capsid exterior on
160 the wall of the fivefold channel At pH 5.6, ribbon in DE loop orients away from E157 of the same monomer and towards D507 of a five fold related monomer (Fig ure 4 1 ) The third change involves residue Q158, which is involved in fivefold symmetry related interactions at the neck of the f ivefold channel. At pH 5.6, Q158 moves away by 1 towards the wall of the channel, thus widening the pore diameter by 2 (Fig ure 4 1 ) The fourth change occurs inside the capsid at the icosahedral twofold axis near the nucleotide binding pocket identified for CPV and MVMi (7, 306) where the side chain of R54 orients towards E62 from the same monomer at pH 5.6 (Figure 4 2A) The structure refinement for MVMp pH 4.0 data is not complete yet. However, p reliminary anal ysis of the averaged density map and comparison to MVMp structure at pH 7.5 shows main chain differences at the HI loop. Discussion The structures of MVMp VLPs have been determined at pHs that mimic the environment encountered by MVMp during endosomal tra fficking. The data for MVMp at pH 5.6 is the most complete. Comparison to MVMp structure at pH 7.5, identified amino acid side chain conformational changes mainly at subunit interfaces and in the capsid interior at the twofold axis near the nucleotide bind ing pocket. The conformational change of E79 reduces the intersubunit contacts between the VP mono mers at the fivefold interface. The change in the orientation of K166 to interact with D507 might contribute to the DE loop flexibility at reduced pH. The Q15 8 of five fivefold symmetry related monomers act as sentries at the opening of the fivefold channel. The side chain amino acid changes observed for E79, K166 and Q158 at or near the fivefold axis of symmetry might be involved in capsid destabilization that primes the capsid for VP1u or genome extrusion through the fivefold channel. Conformational changes that 'weaken'
161 the intersubunit interfaces was also observed in the AAV8 structures solved at low pHs (208) Residues K166 and E79 are conserved amongst the members of parvovirus genus. Q158 is conserved for the rodent parvoviruses (S158 in CPV, FPV and PPV). In full MVMi capsids (genome containing), R54 interacts with an adenine at pH 7.5 (Figure 4 2B) The R54 conformation observed in the MVMp VLP at pH 5.6 would prevent its interaction with the nucleotide. The weakening of the capsid DNA interaction at pH 5.6 might prepare the genome for its release from the capsid. Similar observations were made for the AAV8 structure solved at pH 4.0, where the conformation change in H632 resulted in loss of interaction with the ordered nucleotide (208) R54 is replaced with K in CPV, FPV, H 1PV, LuIII and with Q in PPV. While in MVMi, R54 interacts with a denine, in CPV a nd H 1PV the interacting nucleotide is a cytosine and thymine, respectively ( (7, 306) and data from studies conducted in chapter 3) It is possible that due to the differences in capsid DNA interactions the autonom ous parvoviruses might respond differently to the endosomal pH and exhibit different genome uncoating kinetics. The conformational change of the HI loop observed at pH 4.0 might be related to the opening of the fivefold channel as was observed in AAV2 upon HS binding (178) The structural transitions observed for the MVMp structures at the different pHs were subtle compared to that observed in CPV and FPV where a loop adopts a different conformation or becomes disor dered depending on the pH (274) The MVMp structures determined at pH values of 6.0, 5.6 and 4.0 to mimic the endosomal trafficking environment demonstrated that the although the capsid VP topologies of all the str uctures were very similar, significant amino acid side chain
162 conformational changes were observed at reduced pH on (a) the interior surface of the capsid at the icosahedral twofold axis near the nucleotide binding pocket, and (b) the exterior capsid surfac e at or n ear the fivefold channel. These structural transitions that disrupt intersubunit contacts and capsid DNA interactions and widen the fivefold channel are consistent with capsid destabilization events that likely facilitate VP1u extrusion for PLA2 a ctivity, VP2 N terminus exposure for capsid maturation and capsid priming for genome release.
163 Table 4 1. Data processing and refinement statistics Parameter MVMp pH 6.0 MVMp pH 5.6 MVMp pH 4.0 0.918 1.000 1.08 09 Space Group C2 C2 C2 Unit cell parameters (, ) a=437.9, b=408.7, c=299.6, a=442.0, b=411.7, c=301.8, a=250.7, b=433.9, c=250.7, Resolution () 50 3.8 (3.9 3.8) a 40 3.3 (3.4 3.3) a 40 3.2 ( 3.3 3.2 ) Completeness (%) 48.4 94.6 72.0 Redundancy 1.9 2.6 2.0 R sym b (%) 16.0 13.3 13.1 R f actor c / R free d (%) 40.2/40.6 30.5/30.6 33.1/33.3 a Values in parenthesis are for highest resolution shell. b R sym c R factor = o | |F c o |)100 d R free is calculated the same as R factor except it uses 5% of reflection data omitted from refinement.
164 Figure 4 1. MVMp capsid dynamics on the exterior surface. The amino acid residues that show side chain conformational changes in the MVMp structure at pH 5.6 are colored as red (E79), bl ue (Q158) and green (K166), respectively. The symmetry related VP2 monomers are colored differently and labeled. The inset provides a close up view of the structural variations; At the shoulder of the canyon, K166 at pH 5.6 (green) moves closer to D507 (wh eat) of a five fold related monomer and moves away from E157 ( light blue ) of the same reference monomer as compared to at pH 7.5 ( K166 in brown); inside the five fold pore of the MVMp capsid Q158 from the five fold related monomers at pH 5.6 (in blue) and p H 7.5 (in yellow) is shown. The side chains of the interacting residues are shown in stick model. The approximate icosahedral twofold (filled oval), threefold (filled triangle), and fivefold (filled pentagon) axes are shown. These figures were generated wi th PyMol program (82)
165 Figure 4 2. Capsid dynamics in the interior of the MVMp capsid. (A) The amino acid residues that show side chain differences at pH 5.6 are colored as red (E79), blue (Q158), green (K166), and black (R54), respectively. Inset shows a close up view of the two fold axis. The ref erence (Ref) mon omer, the twofold (2f) related monomer and the residue R54 are depicted in wheat, green and black, respectively. (B) The side chains of R54 and D58 interact with ordered nucleotides in infectious virions. Nucleotides and amino acid side chains are shown as stick representation and colored according to atom type. The side chain of R54 interacts with DNA at pH 7.5 ( carbon atoms in magenta, nitrogen in blue and oxygen in red ) but moves away at pH 5.6 ( R54 colored in black)
166 CHAPTER 5 SUMMARY AND FUTURE D IRE CTIONS The objective of this study was to utilize MVM and H 1PV, autonomous members of the Parvoviridae family as models to map the capsid regions that dictate surface glycan receptor attachment, cell tropism host range and endosomal trafficking of these very similar viruses. The structure dete rmination studies of MVM capsid SIA glycan complexes, showed that all the glycans bound in the vicinity of the twofold depression on the MVM capsid that had been identified as the S I A binding site in previous struct ure studies (188) In this study we were able to visualiz e a longer oligosaccharide in the binding pocket as compared to the previous studies. The binding pocket seemed to accommodate the various glycans in differen t conformations and the capsid interactions with the terminal SIA were not conserved, except in the MVMp 3'SIA (Le x ) 3 and MVMp 3'SIA (LN ) 3 complexes Although, the capsid glycan interactions were not conserved, a few capsid surface residues made non specific contacts with the glycans on all the MVM capsids. This study structurally verified the role of residues 362, 368 399 and 558 that are also cell tropism and host pathogenicity determinants for MVM, in the SIA glycan binding. It also identified the possibl e role of residues 321 and 366 in the recognition of 8 linked S I A glycans by MVMi. However, the heterogeneity of the terminal SIA binding observed for these structures, which could be due to the promiscuous nature of the binding pocket or the inherent limitations of the crystallography technique, preclude d the identification of capsid surface determinants involved in the specificity of the terminal SIA binding To define the specificity of the MVM capsid interactions with the commonly recognized terminal SIA and to minimize the heterogeneity inherent in th e crystallographic technique, multiple diffraction data
167 sets for native MVMp capsid and for MVMp complexed with SIA alone need to be separately solved (no merging of data from different crystals) and F o F o difference Fourier maps calculated. If this strate gy s hows SIA binding on the MVMp capsid in one particular conformation, then it c ould be compared to SIA binding on the MVMi capsid to analyze the differences in receptor binding In addition to the structural studies residues E321 and M366 in MVMi could be mutated to verify their suggested role in recognition of 8 linked S I A glycans. Also, the mutant viruse s could then be screened on the glycan array to determine any changes in binding specifi ci ties. C ell binding competition assays using these re cognized glycans to correlate the glycan screening, glycomic profi ling and structural data have been initiated and will be completed. The utilization of a common SIA binding pocket by all MVM viruses that colocalizes with tropism and pathogenicity determinants suggests the role of this receptor binding pocket in cell rec ognition and the differences in SIA receptor interactions could explain the differences in pathogenicity but the mechanism for differences in cell tropism is still unclear. Screening of MVM viruses on the derivatized SGM showed that both MVMp and MVMi spec 3 linked sialylated derivatives which was consistent with the previous glycan array screening (209) It also identified 9 O methylation (and 9 O acetylation and 9 O lactoylation for MVMp viruses) as an additional component of their SIA recognition. However, preliminary modeling of these SIA derivatives in the glycan binding pocket failed to provide an understanding of the mechanism of specific recognition of these derivatives. The 9 O acetylated and 9 O lactoylated S I A derivatives are present in nature, but the 9 O methylated derivative has n ot yet been isolated from natural sources so the significance of this recognition is unknown The binding profile
168 for VLPs, empty and full particles on the SGM was similar which validates the use of VLPs in lieu of infectious virions for structural and bi ochemical studies that examine receptor interactions. In addition, all the MVM viruses bound to a biantennary S I A glycan with 3'SIA LN motif that was not present on the previously used glycan arrays G lycomic profiling of the cell lines permissive for MVMp and/or MVMi validate d the presence of the glycans that were recognized by the MVM viruses in the previous glycan array and the current SGM screening such as glycans w ith S IA LN motif and SIA Le x m The SIA Le x cancer cell motif is only expressed on the NB324K cell surface which is a SV40 transformed kidney fibroblast cell line and this was The simi lar glycan profile of the A9 and EL4 T cells failed to explain the differences in cell tropism between these MVM viruses based on differential receptor recognition. Based on previous published data (15, 112, 196) an d the results of this study it is evident that the capsid twofold pocket plays a role in dictating differences in cell tropism post cellular entry. The post entry event could be interaction with a cellular host factor present only in permissive cell line or the correct cell signaling pathway initiated in the permissive cell line triggered by ambient utilization of SIA receptor. Covalent tagging of MVM capsids with peptides such as poly Histidine FLAG or GST (Glutathione) followed by immunoprecipitation e xperiments using cell lysate could be conducted to identify the cellular host proteins that interact with the capsid post entry and during trafficking. The differences in protein expression or abundance between the permissive and restrictive cell lines fol lowing initiation of cell signaling pathways could be studied by Two Dimensional Difference Gel Electrophoresis (2D DIGE), wherein the proteins
169 expressed in the two cell lines would be labeled with different fluorescent dyes and then the samples mixed and separated based on isoelectric point and molecular weight in the same gel. This experiment could also be utilized to compare protein expression before and after virus infection The role of cell signaling could be probed by conducting a genome wide microar ray before and after virus infection in the different cell lines to identify differences in cellular transcription. Also, small molecule inhibitors of specific cell signaling pathways could be used to confirm the role of a particular pathway in MVM infecti on. Recently developed techniques, such as Fluorescent Cell Barcoding (where samples are labeled with different intensities of a single fluorophore ) in combination with phospho epitope specific flow cytometry (flow cytometry based analysis of different ph osphorylated proteins using specific antibodies) would allow for high throughput detection of differences in cell signaling networks. The structure of H 1PV determined to aid the mapping of SIA site(s), is similar to that of other parvoviruses, especiall y MVM, except for differences at previously defined variable loops. H 1PV capsids screened on a glycan microarray revealed specificity for 2,3 linked SIA glycans, including the SIA Le x tumor cell marker motif, similar to previous observations for MVM. The common recognition of the 3 SIA Le x tumor cell marker by MVM and H 1PV explains their oncotropism. Structural studies of the H 1PV complex ed with SIA glycans, including one with a 3 SIA Le x motif, also identified the same twofold depression as MVM, as its SIA binding site This observation indicates that parvovirus capsids utilize common regions as determinants of cell tropism, pathogenicity and host range. Out of the four host range switch mutations in vitro in MVMp that conferred the ability to grow in rat fibroblasts (similar to H 1PV's cell
170 tropism), only two of those (residues 560 and 584 in H 1PV) are structurally variable between MVMp and H 1PV and are located at the twofold depression and might play a role in the virus adaptation to a new host. There are other residues that differ between these viruses at the SIA binding site and adopt different side chain conformations and might be h ost range determinants. The H 1PV capsid glycan complex structures need to be completely refined and only preliminary conclusions can be made The glycans on the H 1PV capsid are involved in non specific interactions with homologous residues on MVMp that a re also involved in the cell tropism and pathogenicity determination and SIA binding for MVM. Especially, D405 (D399 in MVMp) seems to be a n 2,3 linked SIA binding determinant on both the viruses. In a recent study, it was shown that the H 1PV capsid mut ants, H374R and I368S demonstrated much reduced cell binding as compared to the homologous MVMp mutants (10, 188) The residue 374 (H374 in H 1PV and K368 in MVMp), although a conservative mutation between the two capsids could contribute to the differences in the neuraminidase sensitivity or SIA binding affinity, similar to observations with the MVMp virulent mutants T he structural differences observed between MVM and H 1PV at the S I A receptor binding site might d ictate the differences in oncotropic properties observed for these viruses. The H37 4 R and I36 8 S H 1PV mutant capsids (constructs available) could be produced in large quantities t o re screen on the glycan array and also tested in cell binding competition a ssays with the 3'S IA (Le x ) 3 glycan to verify the glycan binding specificity of the mutant capsids. LuIII, another oncotropic rodent parvovirus has been shown to bind to 3'S IA (Le x ) 3 glycan which highlights the common theme of utilization of this motif by the se oncotropic viruses to bind to cancer cells. Structure of LuIII complexed to
171 3'S IA (Le x ) 3 would provide structural verification for this common property amongst the rodent parvoviruses and provide information that would aid in elucidating the mechanism of differential tumor tropism between these viruses. Ordered ssDNA in a conserved binding site inside the H 1PV virion suggested a potential role in genome packaging and capsid stability. The differences in the genome sequence observed in the H 1PV, CPV and MVM viruses could be attributed to the slight differences in the amino acid composition of the nucleotide binding pocket. Mutagenesis of the amino acids that are involved in unique interactions with the H 1PV genome could be done to investigate the signif icance of these capsid DNA interactions in genome specific packaging. Also, amino acids involved in conserved capsid DNA interactions could be mutated (apart from the ones made in a previous study (247) ) to understand their role in DNA packaging and capsid stability. To study the m echanism of DNA ordering, the structure of LuIII, a member of the parvovirus genus that packages minus strand and plus strand into different particles, could be determined. The information from this study would tell us if the structural ordering is due to a specific genome sequence of the parvovirus genus or because of a conserved DNA binding pocket. Also, the f our cis peptide bonds identified in the H 1PV virion structure might play a role in capsid assembly and this could be verified by mutating the virus at those positions and monitoring intact capsid production The information on capsid regions involved in cellular specificity and genome interactions would aid in the development of oncotropic parvovirus based gene therapy vectors with improved cancer ce ll targeting and therapeutic genome packaging efficiency.
172 Structural studies of the MVMp capsids at the pHs encountered during endosomal trafficking provided insights into the capsid dynamics. Conformational changes were observed, especially at the icosah edral two fold regions, which alter capsid DNA interaction(s) inside the capsid, possibly in preparation for DNA uncoating, and at the fivefold region, consistent with the opening of the channel, possibly for VP1 externalization. The structure of the MVMp V LPs at pH 4.0 needs to be completely refined. Diffraction data for DNA containing infectious virions at low pHs has been collected and will be processed to provide structural verification for the disruption of capsid DNA interactions at the twofold during endosomal trafficking prior to genome uncoating. Since it is not known whether receptor remains bound or dissociates from the capsid following endocytosis diffraction data for MVMi VLP and H 1PV crystals grown in the presence of glycan receptor at the var ious pHs have been collected and will be solved to compare any differences in the capsid dynamics for the receptor bound and unbound state. Time resolved limited proteolysis with/without heat shock and site specific labeling in combination with peptide mas s spectrometry has also been initiated to provide information on the most dynamic and accessible regions of the capsid.
173 LI ST OF REFERENCES 1. Abschuetz, A., T. Kehl, R. Geibig, B. Leuchs, J. Rommelaere, and A. Regnier Vigouroux. 2006. Oncolytic murine autonomous parvovirus, a candidate vector for glioma gene therapy, is innocuous to normal and immunocompetent mouse glial cells. Cell Tissue Res 325: 423 436. 2. Adams, P. D., P. V. Afonine, G. Bunkoczi, V. B. Chen, I. W. Davis, N. Echols, J. J. Headd, L. W. Hung, G. J. Kapral, R. W. Grosse Kunstleve, A. J. McCoy, N. W. Moriarty, R. Oeffner, R. J. Read, D. C. Richardson, J. S. Richardson, T. C. Terwilliger, and P. H. Zwart. 2010. PHENIX: a comprehensive Python based system for macromolecula r structure solution. Acta Crystallogr D Biol Crystallogr 66: 213 221. 3. Agbandje, M., S. Kajigaya, R. McKenna, N. S. Young, and M. G. Rossmann. 1994. The structure of human parvovirus B19 at 8 A resolution. Virology 203: 106 115. 4. Agbandje, M., R. McKenn a, M. G. Rossmann, M. L. Strassheim, and C. R. Parrish. 1993. Structure determination of feline panleukopenia virus empty particles. Proteins 16: 155 171. 5. Agbandje McKenna, M., and M. S. Chapman. 2006. Correlating structure with function in viral capsid, p. 125 139. In J. R. Kerr, S. F. Cotmore, M. E. Bloom, R. M. Linden, and C. R. Parrish (ed.), Parvoviruses. Hodder Arnold, London. 6. Agbandje McKenna, M., and J. Kleinschmidt. 2011. AAV capsid structure and cell interactions. Methods Mol Biol 807: 47 92. 7. Agbandje McKenna, M., A. L. Llamas Saiz, F. Wang, P. Tattersall, and M. G. Rossmann. 1998. Functional implications of the structure of the murine parvovirus, minute virus of mice. Structure 6: 1369 1381. 8. Akache, B., D. Grimm, K. Pandey, S. R. Yant, H. Xu, and M. A. Kay. 2006. The 37/67 kilodalton laminin receptor is a receptor for adeno associated virus serotypes 8, 2, 3, and 9. J Virol 80: 9831 9836. 9. Alexandersen, S., M. E. Bloom, and S. Perryman. 1988. Detailed transcription map of Aleutian mink di sease parvovirus. J Virol 62: 3684 3694. 10. Allaume, X., N. El Andaloussi, B. Leuchs, S. Bonifati, A. Kulkarni, T. Marttila, J. K. Kaufmann, D. M. Nettelbeck, J. Kleinschmidt, J. Rommelaere, and A. Marchini. 2012. Retargeting of Rat Parvovirus H 1PV to Can cer Cells through Genetic Engineering of the Viral Capsid. J Virol 86: 3452 3465. 11. Angata, T., and A. Varki. 2002. Chemical diversity in the sialic acids and related alpha keto acids: an evolutionary perspective. Chem Rev 102: 439 469.
174 12. Anouja, F., R. Wattiez, S. Mousset, and P. CailletFauquet. 1997. The cytotoxicity of the parvovirus minute virus of mice nonstructural protein NS1 is related to changes in the synthesis and phosphorylation of cell proteins. Journal of Virology 71: 4671 4678. 13. Asokan, A ., J. B. Hamra, L. Govindasamy, M. Agbandje McKenna, and R. J. Samulski. 2006. Adeno associated virus type 2 contains an integrin alpha5beta1 binding domain essential for viral cell entry. J Virol 80: 8961 8969. 14. Astell, C. R., M. Thomson, M. Merchlinsky and D. C. Ward. 1983. The complete DNA sequence of minute virus of mice, an autonomous parvovirus. Nucleic Acids Res 11: 999 1018. 15. Ball Goodrich, L. J., and P. Tattersall. 1992. Two amino acid substitutions within the capsid are coordinately required for acquisition of fibrotropism by the lymphotropic strain of minute virus of mice. J Virol 66: 3415 3423. 16. Bansal, G. P., J. A. Hatfield, F. E. Dunn, A. A. Kramer, F. Brady, C. H. Riggin, M. S. Collett, K. Yoshimoto, S. Kajigaya, and N. S. Young. 1993. Candidate recombinant vaccine for human B19 parvovirus. J Infect Dis 167: 1034 1044. 17. Bantel Schaal, U., I. Braspenning Wesch, and J. Kartenbeck. 2009. Adeno associated virus type 5 exploits two different entry pathways in human embryo fibroblasts. J Gen Virol 90: 317 322. 18. Bantel Schaal, U., B. Hub, and J. Kartenbeck. 2002. Endocytosis of adeno associated virus type 5 leads to accumulation of virus particles in the Golgi compartment. J Virol 76: 2340 2349. 19. Barbis, D. P., S. F. Chang, and C. R. Parri sh. 1992. Mutations adjacent to the dimple of the canine parvovirus capsid structure affect sialic acid binding. Virology 191: 301 308. 20. Bartlett, J. S., R. Wilcher, and R. J. Samulski. 2000. Infectious entry pathway of adeno associated virus and adeno a ssociated virus vectors. J Virol 74: 2777 2785. 21. Basak, S., and H. Turner. 1992. Infectious entry pathway for canine parvovirus. Virology 186: 368 376. 22. Bauer, P. H., R. T. Bronson, S. C. Fung, R. Freund, T. Stehle, S. C. Harrison, and T. L. Benjamin. 1995. Genetic and structural analysis of a virulence determinant in polyomavirus VP1. J Virol 69: 7925 7931. 23. Bell, C. L., L. H. Vandenberghe, P. Bell, M. P. Limberis, G. P. Gao, K. Van Vliet, M. Agbandje McKenna, and J. M. Wilson. 2011. The AAV9 recepto r and its modification to improve in vivo lung gene transfer in mice. J Clin Invest 121: 2427 2435.
17 5 24. Benson, S. D., J. K. Bamford, D. H. Bamford, and R. M. Burnett. 2004. Does common architecture reveal a viral lineage spanning all three domains of life? Mol Cell 16: 673 685. 25. Bergeron, J., B. Hebert, and P. Tijssen. 1996. Genome organization of the Kresse strain of porcine parvovirus: identification of the allotropic determinant and comparison with those of NADL 2 and field isolates. J Virol 70: 2508 25 15. 26. Bergoin, M., J. Kleinschmidt, J. M. Almendral, K. Hedman, Y. Li, M. Agbandje McKenna, P. Tattersall, P. Tijssen, D. J. Pintel, and T. W. Flegel. 2011. Ninth Report of the International Committee on Taxonomy of Viruses. In A. M. Q. King, Adams, M.J. Carstens, E.B., Lefkowitz, E.J. (ed.), Virus Taxonomy. Elsevier. 27. Berns, K., and C. R. Parrish. 2007. Parvoviridae, p. 2437 2478. In D. M. Knipe and P. M. Howley (ed.), Fields Virology, Fifth ed, vol. Two. Lippincott Williams and Wilkins. 28. Blackbur n, S. D., S. E. Cline, J. P. Hemming, and F. B. Johnson. 2005. Attachment of bovine parvovirus to O linked alpha 2,3 neuraminic acid on glycophorin A. Arch Virol 150: 1477 1484. 29. Blackburn, S. D., R. A. Steadman, and F. B. Johnson. 2006. Attachment of ad eno associated virus type 3H to fibroblast growth factor receptor 1. Arch Virol 151: 617 623. 30. Blechacz, B., and S. J. Russell. 2004. Parvovirus vectors: use and optimisation in cancer gene therapy. Expert Rev Mol Med 6: 1 24. 31. Bleker, S., M. Pawlita, and J. A. Kleinschmidt. 2006. Impact of capsid conformation and Rep capsid interactions on adeno associated virus type 2 genome packaging. J Virol 80: 810 820. 32. Bleker, S., F. Sonntag, and J. A. Kleinschmidt. 2005. Mutational analysis of narrow pores at the fivefold symmetry axes of adeno associated virus type 2 capsids reveals a dual role in genome packaging and activation of phospholipase A2 activity. J Virol 79: 2528 2540. 33. Bodendorf, U., C. Cziepluch, J. C. Jauniaux, J. Rommelaere, and N. Salom. 19 99. Nuclear export factor CRM1 interacts with nonstructural proteins NS2 from parvovirus minute virus of mice. J Virol 73: 7769 7779. 34. Boisvert, M., S. Fernandes, and P. Tijssen. 2010. Multiple pathways involved in porcine parvovirus cellular entry and t rafficking toward the nucleus. J Virol 84: 7782 7792.
176 35. Bonsch, C., C. Zuercher, P. Lieby, C. Kempf, and C. Ros. 2010. The globoside receptor triggers structural changes in the B19 virus capsid that facilitate virus internalization. J Virol 84: 11737 11746 36. Bothner, B., X. F. Dong, L. Bibbs, J. E. Johnson, and G. Siuzdak. 1998. Evidence of viral capsid dynamics using limited proteolysis and mass spectrometry. J Biol Chem 273: 673 676. 37. Bowles, D. E., S. W. McPhee, C. Li, S. J. Gray, J. J. Samulski, A. S. Camp, J. Li, B. Wang, P. E. Monahan, J. E. Rabinowitz, J. C. Grieger, L. Govindasamy, M. Agbandje McKenna, X. Xiao, and R. J. Samulski. 2012. Phase 1 gene therapy for Duchenne muscular dystrophy using a translational optimized AAV vector. Mol Ther 20: 4 43 455. 38. Bowles, D. E., J. E. Rabinowitz, and S. R.J. 2006. The Genus Dependovirus, p. 15 24. In J. R. Kerr, Cotmore, S.F., Bloom M.E., Linden, R.M., Parrish, C.R. (ed.), Parvoviruses. Hodder Arnold, London. 39. Brandenburger, A., D. Legendre, B. Avalos se, and J. Rommelaere. 1990. NS 1 and NS 2 proteins may act synergistically in the cytopathogenicity of parvovirus MVMp. Virology 174: 576 584. 40. Brandts, J. F., H. R. Halvorson, and M. Brennan. 1975. Consideration of the Possibility that the slow step in protein denaturation reactions is due to cis trans isomerism of proline residues. Biochemistry 14: 4953 4963. 41. Brockhausen, I., H. Schachter, and P. Stanley. 2009. O GalNAc Glycans. In A. Varki, R. D. Cummings, J. D. Esko, H. H. Freeze, P. Stanley, C. R Bertozzi, G. W. Hart, and M. E. Etzler (ed.), Essentials of Glycobiology. The Consortium of Glycobiology Editors, La Jolla, California, Cold Spring Harbor NY. 42. Brown, C. S., S. Welling Wester, M. Feijlbrief, J. W. Van Lent, and W. J. Spaan. 1994. Chim eric parvovirus B19 capsids for the presentation of foreign epitopes. Virology 198: 477 488. 43. Brown, K. E., S. M. Anderson, and N. S. Young. 1993. Erythrocyte P antigen: cellular receptor for B19 parvovirus. Science 262: 114 117. 44. Brownstein, D. G., A. L. Smith, E. A. Johnson, D. J. Pintel, L. K. Naeger, and P. Tattersall. 1992. The pathogenesis of infection with minute virus of mice depends on expression of the small nonstructural protein NS2 and on the genotype of the allotropic determinants VP1 and V P2. J Virol 66: 3118 3124. 45. Brunger, A. T. 1992. Free R value: a novel statistical quantity for assessing the accuracy of crystal structures. Nature 355: 472 475.
177 46. Brunger, A. T., P. D. Adams, G. M. Clore, W. L. DeLano, P. Gros, R. W. Grosse Kunstleve, J. S. Jiang, J. Kuszewski, M. Nilges, N. S. Pannu, R. J. Read, L. M. Rice, T. Simonson, and G. L. Warren. 1998. Crystallography & NMR system: A new software suite for macromolecular structure determination. Acta Crystallogr D Biol Crystallogr 54: 905 921. 47. Burmeister, W. P., D. Guilligay, S. Cusack, G. Wadell, and N. Arnberg. 2004. Crystal structure of species D adenovirus fiber knobs and their sialic acid binding sites. J Virol 78: 7727 7736. 48. Caillet Fauquet, P., M. Perros, A. Brandenburger, P. Spege laere, and J. Rommelaere. 1990. Programmed killing of human cells by means of an inducible clone of parvoviral genes encoding non structural proteins. EMBO J 9: 2989 2995. 49. Carrasco, C., A. Carreira, I. A. Schaap, P. A. Serena, J. Gomez Herrero, M. G. Ma teu, and P. J. de Pablo. 2006. DNA mediated anisotropic mechanical reinforcement of a virus. Proc Natl Acad Sci U S A 103: 13706 13711. 50. Carrillo Tripp, M., C. M. Shepherd, I. A. Borelli, S. Venkataraman, G. Lander, P. Natarajan, J. E. Johnson, C. L. Bro oks, 3rd, and V. S. Reddy. 2009. VIPERdb2: an enhanced and web API enabled relational database for structural virology. Nucleic Acids Res 37: D436 442. 51. Cater, J. E., and D. J. Pintel. 1992. The small non structural protein NS2 of the autonomous parvovir us minute virus of mice is required for virus growth in murine cells. J Gen Virol 73 ( Pt 7): 1839 1843. 52. Chang, S. F., J. Y. Sgro, and C. R. Parrish. 1992. Multiple amino acids in the capsid structure of canine parvovirus coordinately determine the cani ne host range and specific antigenic and hemagglutination properties. J Virol 66: 6858 6867. 53. Chapman, M. S., and M. Agbandje McKenna. 2006. Atomic structure of viral particles, p. 107 123. In J. R. Kerr, S. F. Cotmore, M. E. Bloom, R. M. Linden, and C. R. Parrish (ed.), Parvoviruses. Hodder Arnold, London. 54. Chejanovsky, N., and B. J. Carter. 1989. Mutagenesis of an AUG codon in the adeno associated virus rep gene: effects on viral DNA replication. Virology 173: 120 128. 55. Chen, K. C., B. C. Shull, M. Lederman, E. R. Stout, and R. C. Bates. 1988. Analysis of the termini of the DNA of bovine parvovirus: demonstration of sequence inversion at the left terminus and its implication for the replication model. J Virol 62: 3807 3813. 56. Chen, V. B., W. B. Are ndall, 3rd, J. J. Headd, D. A. Keedy, R. M. Immormino, G. J. Kapral, L. W. Murray, J. S. Richardson, and D. C. Richardson. 2010. MolProbity: all atom structure validation for macromolecular crystallography. Acta Crystallogr D Biol Crystallogr 66: 12 21.
178 57. Chipman, P. R., M. Agbandje McKenna, S. Kajigaya, K. E. Brown, N. S. Young, T. S. Baker, and M. G. Rossmann. 1996. Cryo electron microscopy studies of empty capsids of human parvovirus B19 complexed with its cellular receptor. Proc Natl Acad Sci U S A 93: 7502 7506. 58. Clemens, K. E., and D. J. Pintel. 1988. The two transcription units of the autonomous parvovirus minute virus of mice are transcribed in a temporal order. J Virol 62: 1448 1451. 59. Cohen, S., A. R. Behzad, J. B. Carroll, and N. Pante. 2006. Parvoviral nuclear import: bypassing the host nuclear transport machinery. J Gen Virol 87: 3209 3213. 60. Cohen, S., A. K. Marr, P. Garcin, and N. Pante. 2011. Nuclear envelope disruption involving host caspases plays a role in the parvovirus replication cy cle. J Virol 85: 4863 4874. 61. Cohen, S., and N. Pante. 2005. Pushing the envelope: microinjection of Minute virus of mice into Xenopus oocytes causes damage to the nuclear envelope. J Gen Virol 86: 3243 3252. 62. Collaborative Computational Project, N. 199 4. The CCP4 suite: programs for protein crystallography. Acta Crystallogr D Biol Crystallogr 50: 760 763. 63. Colomar, M. C., B. Hirt, and P. Beard. 1998. Two segments in the genome of the immunosuppressive minute virus of mice determine the host cell speci ficity, control viral DNA replication and affect viral RNA metabolism. J Gen Virol 79 ( Pt 3): 581 586. 64. Cornelis, J. J., P. Becquart, N. Duponchel, N. Salome, B. L. Avalosse, M. Namba, and J. Rommelaere. 1988. Transformation of human fibroblasts by ioni zing radiation, a chemical carcinogen, or simian virus 40 correlates with an increase in susceptibility to the autonomous parvoviruses H 1 virus and minute virus of mice. J Virol 62: 1679 1686. 65. Cotmore, S. F., M. D'Abramo A, Jr., C. M. Ticknor, and P. T attersall. 1999. Controlled conformational transitions in the MVM virion expose the VP1 N terminus and viral genome without particle disassembly. Virology 254: 169 181. 66. Cotmore, S. F., A. M. D'Abramo, Jr., L. F. Carbonell, J. Bratton, and P. Tattersall. 1997. The NS2 polypeptide of parvovirus MVM is required for capsid assembly in murine cells. Virology 231: 267 280. 67. Cotmore, S. F., S. Hafenstein, and P. Tattersall. 2010. Depletion of virion associated divalent cations induces parvovirus minute virus of mice to eject its genome in a 3' to 5' direction from an otherwise intact viral particle. J Virol 84: 1945 1956.
179 68. Cotmore, S. F., V. C. McKie, L. J. Anderson, C. R. Astell, and P. Tattersall. 1986. Identification of the major structural and nonstructu ral proteins encoded by human parvovirus B19 and mapping of their genes by procaryotic expression of isolated genomic fragments. J Virol 60: 548 557. 69. Cotmore, S. F., and P. Tattersall. 1989. A genome linked copy of the NS 1 polypeptide is located on the outside of infectious parvovirus particles. J Virol 63: 3902 3911. 70. Cotmore, S. F., and P. Tattersall. 2006. A Rolling hairpin Strategy: Basic Mechanism of DNA Replication in the Parvoviruses, p. 171 188. In J. R. Kerr, Cotmore, S.F., Bloom M.E., Linden R.M., Parrish, C.R. (ed.), Parvoviruses. Hodder Arnold, London. 71. Cotmore, S. F., and P. Tattersall. 1990. Alternate splicing in a parvoviral nonstructural gene links a common amino terminal sequence to downstream domains which confer radically differe nt localization and turnover characteristics. Virology 177: 477 487. 72. Cotmore, S. F., and P. Tattersall. 2005. Encapsidation of minute virus of mice DNA: aspects of the translocation mechanism revealed by the structure of partially packaged genomes. Viro logy 336: 100 112. 73. Cotmore, S. F., and P. Tattersall. 2005. Genome packaging sense is controlled by the efficiency of the nick site in the right end replication origin of parvoviruses minute virus of mice and LuIII. J Virol 79: 2287 2300. 74. Cotmore, S. F., and P. Tattersall. 2012. Mutations at the Base of the Icosahedral Five Fold Cylinders of Minute Virus of Mice Induce 3' to 5' Genome Uncoating and Critically Impair Entry Functions. J Virol 86: 69 80. 75. Cotmore, S. F., and P. Tattersall. 2007. Parvov iral host range and cell entry mechanisms. Adv Virus Res 70: 183 232. 76. Cotmore, S. F., and P. Tattersall. 2006. Structure and Organization of the Viral Genome, p. 73 94. In J. R. Kerr, Cotmore, S.F., Bloom M.E., Linden, R.M., Parrish, C.R. (ed.), Parvovi ruses. Hodder Arnold, London. 77. Cotmore, S. F., and P. Tattersall. 1987. The autonomously replicating parvoviruses of vertebrates. Advances in Virus Research 33: 91 174. 78. Cotmore, S. F., and P. Tattersall. 1987. The autonomously replicating parvoviruse s of vertebrates. Adv Virus Res 33: 91 174. 79. D'Abramo, A. M., Jr., A. A. Ali, F. Wang, S. F. Cotmore, and P. Tattersall. 2005. Host range mutants of Minute Virus of Mice with a single VP2 amino acid change require additional silent mutations that regulat e NS2 accumulation. Virology 340: 143 154.
180 80. Daeffler, L., R. Horlein, J. Rommelaere, and J. Nuesch. 2003. Modulation of minute virus of mice cytotoxic activities through site directed mutagenesis within the NS coding region. Journal of Virology 77: 12466 12478. 81. de la Roche Saint Andre, C., F. Harper, and J. Feunteun. 1990. Analysis of the hamster polyomavirus infection in vitro: host restricted productive cycle. Virology 177: 532 540. 82. DeLano, W. L. 2002. The PyMOL Molecular Graphic System. DeLano Sc ientific, San Carlos, CA. 83. Di Pasquale, G., and J. A. Chiorini. 2006. AAV transcytosis through barrier epithelia and endothelium. Mol Ther 13: 506 516. 84. Di Pasquale, G., B. L. Davidson, C. S. Stein, I. Martins, D. Scudiero, A. Monks, and J. A. Chiorin i. 2003. Identification of PDGFR as a receptor for AAV 5 transduction. Nat Med 9: 1306 1312. 85. Di Pasquale, G., N. Kaludov, M. Agbandje McKenna, and J. A. Chiorini. 2010. BAAV transcytosis requires an interaction with beta 1 4 linked glucosamine and gp96 PLoS One 5: e9336. 86. Ding, W., L. Zhang, Z. Yan, and J. F. Engelhardt. 2005. Intracellular trafficking of adeno associated viral vectors. Gene Ther 12: 873 880. 87. Ding, W., L. N. Zhang, C. Yeaman, and J. F. Engelhardt. 2006. rAAV2 traffics through both the late and the recycling endosomes in a dose dependent fashion. Mol Ther 13: 671 682. 88. DiPrimio, N., A. Asokan, L. Govindasamy, M. Agbandje McKenna, and R. J. Samulski. 2008. Surface loop dynamics in adeno associated virus capsid assembly. J Virol 82: 5178 5189. 89. Doerig, C., B. Hirt, P. Beard, and J. P. Antonietti. 1988. Minute virus of mice non structural protein NS 1 is necessary and sufficient for trans activation of the viral P39 promoter. J Gen Virol 69 ( Pt 10): 2563 2573. 90. Dormitzer, P. R., Z. Y. Sun, G. Wagner, and S. C. Harrison. 2002. The rhesus rotavirus VP4 sialic acid binding domain has a galectin fold with a novel carbohydrate binding site. EMBO J 21: 885 897. 91. Dorsch, S., G. Liebisch, B. Kaufmann, P. von Landenberg, J. H. Hoffmann, W. Drobnik, and S. Modrow. 2002. The VP1 unique region of parvovirus B19 and its constituent phospholipase A2 like activity. J Virol 76: 2014 2018. 92. Douar, A. M., K. Poulard, D. Stockholm, and O. Danos. 2001. Intracellular trafficking of adeno associated virus vectors: routing to the late endosomal compartment and proteasome degradation. J Virol 75: 1824 1833.
181 93. Duan, D., Q. Li, A. W. Kao, Y. Yue, J. E. Pessin, and J. F. Engelhardt. 1999. Dynamin is required for recombinant adeno associated virus type 2 infection. J Virol 73: 10371 10376. 94. Duan, D., Y. Yue, Z. Yan, J. Yang, and J. F. Engelhardt. 2000. Endosomal processing limits gene transfer to polarized airway epithelia by adeno associated virus. J Clin Invest 105: 1573 1587. 95. Dubielzig, R., J. A. K ing, S. Weger, A. Kern, and J. A. Kleinschmidt. 1999. Adeno associated virus type 2 protein interactions: formation of pre encapsidation complexes. J Virol 73: 8989 8998. 96. Eichwald, V., L. Daeffler, M. Klein, J. Rommelaere, and N. Salome. 2002. The NS2 p roteins of parvovirus minute virus of mice are required for efficient nuclear egress of progeny virions in mouse cells. J Virol 76: 10307 10319. 97. Emsley, P., and K. Cowtan. 2004. Coot: model building tools for molecular graphics. Ac ta Crystallogr D Biol Crystallogr 60: 2126 2132. 98. Enderlin, M., E. V. Kleinmann, S. Struyf, C. Buracchi, A. Vecchi, R. Kinscherf, F. Kiessling, S. Paschek, S. Sozzani, J. Rommelaere, J. J. Cornelis, J. Van Damme, and C. Dinsart. 2009. TNF alpha and the I FN gamma inducible protein 10 (IP 10/CXCL 10) delivered by parvoviral vectors act in synergy to induce antitumor effects in mouse glioblastoma. Cancer Gene Ther 16: 149 160. 99. Etingov, I., R. Itah, M. Mincberg, A. Keren Naus, H. J. Nam, M. Agbandje McKenn a, and C. Davis. 2008. An extension of the Minute Virus of Mice tissue tropism. Virology 379: 245 255. 100. Farr, G. A., S. F. Cotmore, and P. Tattersall. 2006. VP2 cleavage and the leucine ring at the base of the fivefold cylinder control pH dependent exte rnalization of both the VP1 N terminus and the genome of minute virus of mice. J Virol 80: 161 171. 101. Farr, G. A., and P. Tattersall. 2004. A conserved leucine that constricts the pore through the capsid fivefold cylinder plays a central role in parvovir al infection. Virology 323: 243 256. 102. Farr, G. A., L. G. Zhang, and P. Tattersall. 2005. Parvoviral virions deploy a capsid tethered lipolytic enzyme to breach the endosomal membrane during cell entry. Proc Natl Acad Sci U S A 102: 17148 17153. 103. Fesc hotte, C., and C. Gilbert. 2012. Endogenous viruses: insights into viral evolution and impact on host biology. Nat Rev Genet 13: 283 296. 104. Fisher, A. J., and J. E. Johnson. 1993. Ordered duplex RNA controls capsid architecture in an icosahedral animal v irus. Nature 361: 176 179.
182 105. Fox, D., X. He, A. Abe, T. Hollander, L. Li, L. Kan, A. Friedman, Y. Shimizu, J. Shayman, and K. Kozarsky. 2001. The T lymphocyte structure CD60 contains a sialylated carbohydrate epitope that is expressed on both ganglioside s and glycoproteins. Immunological Investigations 30: 67 85. 106. Fox, J. M., and M. E. Bloom. 1999. Identification of a cell surface protein from Crandell feline kidney cells that specifically binds Aleutian mink disease parvovirus. J Virol 73: 3835 3842. 1 07. Fox, J. P. 1984. Modes of action of poliovirus vaccines and relation to resulting immunity. Rev Infect Dis 6 Suppl 2: S352 355. 108. Fricks, C. E., and J. M. Hogle. 1990. Cell induced conformational change in poliovirus: externalization of the amino ter minus of VP1 is responsible for liposome binding. J Virol 64: 1934 1945. 109. Fry, E. E., S. M. Lea, T. Jackson, J. W. Newman, F. M. Ellard, W. E. Blakemore, R. Abu Ghazaleh, A. Samuel, A. M. King, and D. I. Stuart. 1999. The structure and function of a foo t and mouth disease virus oligosaccharide receptor complex. EMBO J 18: 543 554. 110. Fuks, F., L. Deleu, C. Dinsart, J. Rommelaere, and S. Faisst. 1996. ras oncogene dependent activation of the P4 promoter of minute virus of mice through a proximal P4 eleme nt interacting with the Ets family of transcription factors. J Virol 70: 1331 1339. 111. Gamblin, S. J., L. F. Haire, R. J. Russell, D. J. Stevens, B. Xiao, Y. Ha, N. Vasisht, D. A. Steinhauer, R. S. Daniels, A. Elliot, D. C. Wiley, and J. J. Skehel. 2004. The structure and receptor binding properties of the 1918 influenza hemagglutinin. Science 303: 1838 1842. 112. Gardiner, E. M., and P. Tattersall. 1988. Mapping of the fibrotropic and lymphotropic host range determinants of the parvovirus minute virus of m ice. J Virol 62: 2605 2613. 113. Geletneky, K., J. Huesing, J. Rommelaere, J. R. Schlehofer, M. Dahm, O. Krebs, M. v. Knebel Doeberitz, B. Huber, and J. Hajda. 2012. Phase I/IIa study of intratumoral/intracerebral or intravenous/intracerebral administration of Parvovirus H 1 (ParvOryx) in patients with progressive primary or recurrent glioblastoma multiforme: ParvOryx01 protocol. BMC Cancer 12: 99. 114. Geletneky, K., I. Kiprianova, A. Ayache, R. Koch, Y. C. M. Herrero, L. Deleu, C. Sommer, N. Thomas, J. Romm elaere, and J. R. Schlehofer. 2010. Regression of advanced rat and human gliomas by local or systemic treatment with oncolytic parvovirus H 1 in rat models. Neuro Oncol 12: 804 814.
183 115. Girod, A., C. E. Wobus, Z. Zadori, M. Ried, K. Leike, P. Tijssen, J. A Kleinschmidt, and M. Hallek. 2002. The VP1 capsid protein of adeno associated virus type 2 is carrying a phospholipase A2 domain required for virus infectivity. J Gen Virol 83: 973 978. 116. Goodman, L. B., S. M. Lyi, N. C. Johnson, J. O. Cifuente, S. L. Hafenstein, and C. R. Parrish. 2010. Binding site on the transferrin receptor for the parvovirus capsid and effects of altered affinity on cell uptake and infection. J Virol 84: 4969 4978. 117. Goodman, L. B., S. M. Lyi, N. C. Johnson, J. O. Cifuente, S. L. Hafenstein, and C. R. Parrish. Binding site on the transferrin receptor for the parvovirus capsid and effects of altered affinity on cell uptake and infection. J Virol 84: 4969 4978. 118. Govindasamy, L., K. Hueffer, C. R. Parrish, and M. Agbandje McKenna. 2003. Structures of host range controlling regions of the capsids of canine and feline parvoviruses and mutants. J Virol 77: 12211 12221. 119. Govindasamy, L., E. Padron, R. McKenna, N. Muzyczka, N. Kaludov, J. A. Chiorini, and M. Agbandje McKenna. 2006. S tructurally mapping the diverse phenotype of adeno associated virus serotype 4. J Virol 80: 11556 11570. 120. Grekova, S., R. Zawatzky, R. Horlein, C. Cziepluch, M. Mincberg, C. Davis, J. Rommelaere, and L. Daeffler. 2010. Activation of an antiviral respons e in normal but not transformed mouse cells: a new determinant of minute virus of mice oncotropism. J Virol 84: 516 531. 121. Grieger, J. C., J. S. Johnson, B. Gurda Whitaker, M. Agbandje McKenna, and R. J. Samulski. 2007. Surface exposed adeno associated v irus Vp1 NLS capsid fusion protein rescues infectivity of noninfectious wild type Vp2/Vp3 and Vp3 only capsids but not that of fivefold pore mutant virions. J Virol 81: 7833 7843. 122. Grieger, J. C., S. Snowdy, and R. J. Samulski. 2006. Separate basic regi on motifs within the adeno associated virus capsid proteins are essential for infectivity and assembly. J Virol 80: 5199 5210. 123. Gurda, B. L., K. N. Parent, H. Bladek, R. S. Sinkovits, M. A. DiMattia, C. Rence, A. Castro, R. McKenna, N. Olson, K. Brown, T. S. Baker, and M. Agbandje McKenna. 2010. Human bocavirus capsid structure: insights into the structural repertoire of the parvoviridae. J Virol 84: 5880 5889. 124. Haag, A., P. Menten, J. Van Damme, C. Dinsart, J. Rommelaere, and J. J. Cornelis. 2000. Hi ghly efficient transduction and expression of cytokine genes in human tumor cells by means of autonomous parvovirus vectors; generation of antitumor responses in recipient mice. Hum Gene Ther 11: 597 609.
184 125. Hafenstein, S., V. D. Bowman, T. Sun, C. D. Nel son, L. M. Palermo, P. R. Chipman, A. J. Battisti, C. R. Parrish, and M. G. Rossmann. 2009. Structural comparison of different antibodies interacting with parvovirus capsids. J Virol 83: 5556 5566. 126. Halder, S., R. Ng, and M. Agbandje McKenna. 2012. Parv oviruses: structure and infection. Future Virology 7: 253 278. 127. Hansen, J., K. Qing, H. J. Kwon, C. Mah, and A. Srivastava. 2000. Impaired intracellular trafficking of adeno associated virus type 2 vectors limits efficient transduction of murine fibrobl asts. J Virol 74: 992 996. 128. Hansen, J., K. Qing, and A. Srivastava. 2001. Adeno associated virus type 2 mediated gene transfer: altered endocytic processing enhances transduction efficiency in murine fibroblasts. J Virol 75: 4080 4090. 129. Harbison, C. E., J. A. Chiorini, and C. R. Parrish. 2008. The parvovirus capsid odyssey: from the cell surface to the nucleus. Trends Microbiol 16: 208 214. 130. Harbison, C. E., S. M. Lyi, W. S. Weichert, and C. R. Parrish. 2009. Early steps in cell infection by parvov iruses: host specific differences in cell receptor binding but similar endosomal trafficking. J Virol 83: 10504 10514. 131. Hauck, B., W. Zhao, K. High, and W. Xiao. 2004. Intracellular viral processing, not single stranded DNA accumulation, is crucial for recombinant adeno associated virus transduction. J Virol 78: 13678 13686. 132. Hernando, E., A. L. Llamas Saiz, C. Foces Foces, R. McKenna, I. Portman, M. Agbandje McKenna, and J. M. Almendral. 2000. Biochemical and physical characterization of parvovirus m inute virus of mice virus like particles. Virology 267: 299 309. 133. Heuser, J. 1989. Changes in lysosome shape and distribution correlated with changes in cytoplasmic pH. J Cell Biol 108: 855 864. 134. Horiuchi, M., N. Ishiguro, H. Goto, and M. Shinagawa. 1992. Characterization of the stage(s) in the virus replication cycle at which the host cell specificity of the feline parvovirus subgroup is regulated in canine cells. Virology 189: 600 608. 135. Hueffer, K., L. Govindasamy, M. Agbandje McKenna, and C. R. Parrish. 2003. Combinations of two capsid regions controlling canine host range determine canine transferrin receptor binding by canine and feline parvoviruses. J Virol 77: 10099 10105. 136. Hueffer, K., J. S. Parker, W. S. Weichert, R. E. Geisel, J. Y. Sgr o, and C. R. Parrish. 2003. The natural host range shift and subsequent evolution of canine parvovirus resulted from virus specific binding to the canine transferrin receptor. J Virol 77: 1718 1726.
185 137. Hueffer, K., and C. R. Parrish. 2003. Parvovirus host range, cell tropism and evolution. Curr Opin Microbiol 6: 392 398. 138. Inoko, E., Y. Nishiura, H. Tanaka, T. Takahashi, K. Furukawa, K. Kitajima, and C. Sato. 2010. Developmental stage dependent expression of an alpha2,8 trisialic acid unit on glycoprotei ns in mouse brain. Glycobiology 20: 916 928. 139. Irimura, T., S. Nakamori, Y. Matsushita, Y. Taniuchi, N. Todoroki, T. Tsuji, Y. Izumi, Y. Kawamura, S. D. Hoff, K. R. Cleary, and et al. 1993. Colorectal cancer metastasis determined by carbohydrate mediated cell adhesion: role of sialyl LeX antigens. Semin Cancer Biol 4: 319 324. 140. Jabs, A., M. S. Weiss, and R. Hilgenfeld. 1999. Non proline cis peptide bonds in proteins. J Mol Biol 286: 291 304. 141. Jnaoui, K., M. Minet, and T. Michiels. 2002. Mutations th at affect the tropism of DA and GDVII strains of Theiler's virus in vitro influence sialic acid binding and pathogenicity. J Virol 76: 8138 8147. 142. Johnson, F. B., L. B. Fenn, T. J. Owens, L. J. Faucheux, and S. D. Blackburn. 2004. Attachment of bovine p arvovirus to sialic acids on bovine cell membranes. J Gen Virol 85: 2199 2207. 143. Johnson, F. B., and M. D. Hoggan. 1973. Structural proteins of HADEN virus. Virology 51: 129 137. 144. Johnson, J. S., M. Gentzsch, L. Zhang, C. M. Ribeiro, B. Kantor, T. Kaf ri, R. J. Pickles, and R. J. Samulski. 2011. AAV exploits subcellular stress associated with inflammation, endoplasmic reticulum expansion, and misfolded proteins in models of cystic fibrosis. PLoS Pathog 7: e1002053. 145. Johnson, J. S., C. Li, N. DiPrimio M. S. Weinberg, T. J. McCown, and R. J. Samulski. 2010. Mutagenesis of adeno associated virus type 2 capsid protein VP1 uncovers new roles for basic amino acids in trafficking and cell specific transduction. J Virol 84: 8888 8902. 146. Johnson, J. S., and R. J. Samulski. 2009. Enhancement of adeno associated virus infection by mobilizing capsids into and out of the nucleolus. J Virol 83: 2632 2644. 147. Johnston, J. B., S. H. Nazarian, R. Natale, and G. McFadden. 2005. Myxoma virus infection of primary huma n fibroblasts varies with cellular age and is regulated by host interferon responses, p. 235 248, Virology, vol. 332, United States.
186 148. Kaludov, N., K. E. Brown, R. W. Walters, J. Zabner, and J. A. Chiorini. 2001. Adeno associated virus serotype 4 (AAV4) and AAV5 both require sialic acid binding for hemagglutination and efficient transduction but differ in sialic acid linkage specificity. J Virol 75: 6884 6893. 149. Kamerling, J. P., and G. J. Gerwig. 2006. Structural analysis of naturally occurring sialic acids. Methods Mol Biol 347: 69 91. 150. Kaminsky, P. M., N. W. Keiser, Z. Yan, D. C. Lei Butters, and J. F. Engelhardt. 2012. Directing Integrin linked Endocytosis of Recombinant AAV Enhances Productive FAK dependent Transduction. Mol Ther. 151. Kannagi, R. 2004. Molecular mechanism for cancer associated induction of sialyl Lewis X and sialyl Lewis A expression The Warburg effect revisited. Glycoconj J 20: 353 364. 152. Kannagi, R., M. Izawa, T. Koike, K. Miyazaki, and N. Kimura. 2004. Carbohydrate mediated cell adhesion in cancer metastasis and angiogenesis. Cancer Sci 95: 377 384. 153. Kashiwakura, Y., K. Tamayose, K. Iwabuchi, Y. Hirai, T. Shimada, K. Matsumoto, T. Nakamura, M. Watanabe, K. Oshimi, and H. Daida. 2005. Hepatocyte growth factor receptor is a coreceptor for adeno associated virus type 2 infection. J Virol 79: 609 614. 154. Kaufmann, B., U. Baxa, P. R. Chipman, M. G. Rossmann, S. Modrow, and R. Seckler. 2005. Parvovirus B19 does not bind to membrane associated globoside in vitro. Virology 332: 18 9 198. 155. Kaufmann, B., P. R. Chipman, V. A. Kostyuchenko, S. Modrow, and M. G. Rossmann. 2008. Visualization of the externalized VP2 N termini of infectious human parvovirus B19. J Virol 82: 7306 7312. 156. Kaufmann, B., M. El Far, P. Plevka, V. D. Bowma n, Y. Li, P. Tijssen, and M. G. Rossmann. 2011. Structure of Bombyx mori densovirus 1, a silkworm pathogen. J Virol 85: 4691 4697. 157. Kaufmann, B., A. Lopez Bueno, M. G. Mateu, P. R. Chipman, C. D. Nelson, C. R. Parrish, J. M. Almendral, and M. G. Rossman n. 2007. Minute virus of mice, a parvovirus, in complex with the Fab fragment of a neutralizing monoclonal antibody. J Virol 81: 9851 9858. 158. Kaufmann, B., A. A. Simpson, and M. G. Rossmann. 2004. The structure of human parvovirus B19. Proc Natl Acad Sci U S A 101: 11628 11633.
187 159. Kelly, B. N., S. Kyere, I. Kinde, C. Tang, B. R. Howard, H. Robinson, W. I. Sundquist, M. F. Summers, and C. P. Hill. 2007. Structure of the antiviral assembly inhibitor CAP 1 complex with the HIV 1 CA protein. J Mol Biol 373: 3 55 366. 160. Kern, A., K. Schmidt, C. Leder, O. J. Muller, C. E. Wobus, K. Bettinger, C. W. Von der Lieth, J. A. King, and J. A. Kleinschmidt. 2003. Identification of a heparin binding motif on adeno associated virus type 2 capsids. J Virol 77: 11072 11081. 161. Kestler, J., B. Neeb, S. Struyf, J. Van Damme, S. F. Cotmore, A. D'Abramo, P. Tattersall, J. Rommelaere, C. Dinsart, and J. J. Cornelis. 1999. cis requirements for the efficient production of recombinant DNA vectors based on autonomous parvoviruses. Hum Gene Ther 10: 1619 1632. 162. Khayat, R., and J. E. Johnson. 2011. Pass the jelly rolls. Structure 19: 904 906. 163. Kimsey, P. B., H. D. Engers, B. Hirt, and C. V. Jongeneel. 1986. Pathogenicity of fibroblast and lymphocyte specific variants of minute virus of mice. J Virol 59: 8 13. 164. King, J. A., R. Dubielzig, D. Grimm, and J. A. Kleinschmidt. 2001. DNA helicase mediated packaging of adeno associated virus type 2 genomes into preformed capsids. EMBO J 20: 3282 3291. 165. Klein, A., S. Diaz, I. Ferrei ra, G. Lamblin, P. Roussel, and A. E. Manzi. 1997. New sialic acids from biological sources identified by a comprehensive and sensitive approach: liquid chromatography electrospray ionization mass spectrometry (LC ESI MS) of SIA quinoxalinones. Glycobiolog y 7: 421 432. 166. Kleywegt, G. J. 2007. Crystallographic refinement of ligand complexes. Acta Crystallogr D Biol Crystallogr 63: 94 100. 167. Kleywegt, G. J., and T. A. Jones. 1997. Model building and refinement practice. Methods Enzymol 277: 208 230. 168. K ontou, M., L. Govindasamy, H. J. Nam, N. Bryant, A. L. Llamas Saiz, C. Foces Foces, E. Hernando, M. P. Rubio, R. McKenna, J. M. Almendral, and M. Agbandje McKenna. 2005. Structural determinants of tissue tropism and in vivo pathogenicity for the parvovirus minute virus of mice. J Virol 79: 10931 10943. 169. Krissinel, E., and K. Henrick. 2004. Secondary structure matching (SSM), a new tool for fast protein structure alignment in three dimensions. Acta Crystallogr D Biol Crystallogr 60: 2256 2268. 170. Kronenb erg, S., B. Bottcher, C. W. von der Lieth, S. Bleker, and J. A. Kleinschmidt. 2005. A conformational change in the adeno associated virus type 2 capsid leads to the exposure of hidden VP1 N termini. J Virol 79: 5296 5303.
188 171. Kronenberg, S., J. A. Kleinsch midt, and B. Bottcher. 2001. Electron cryo microscopy and image reconstruction of adeno associated virus type 2 empty capsids. EMBO Rep 2: 997 1002. 172. Kruger, L., H. Eskerski, C. Dinsart, J. Cornelis, J. Rommelaere, U. Haberkorn, and J. A. Kleinschmidt. 2008. Augmented transgene expression in transformed cells using a parvoviral hybrid vector. Cancer Gene Ther 15: 252 267. 173. Lacroix, J., B. Leuchs, J. Li, G. Hristov, H. E. Deubzer, A. E. Kulozik, J. Rommelaere, J. R. Schlehofer, and O. Witt. 2010. Parvo virus H1 selectively induces cytotoxic effects on human neuroblastoma cells. Int J Cancer 127: 1230 1239. 174. Lang, S. I., S. Boelz, A. Y. Stroh Dege, J. Rommelaere, C. Dinsart, and J. J. Cornelis. 2005. The infectivity and lytic activity of minute virus o f mice wild type and derived vector particles are strikingly different. J Virol 79: 289 298. 175. Lederman, M., R. C. Bates, and E. R. Stout. 1983. In vitro and in vivo studies of bovine parvovirus proteins. J Virol 48: 10 17. 176. Legrand, C., J. Rommelaere and P. Caillet Fauquet. 1993. MVM(p) NS 2 protein expression is required with NS 1 for maximal cytotoxicity in human transformed cells. Virology 195: 149 155. 177. Lerch, T. F., Q. Xie, and M. S. Chapman. 2010. The structure of adeno associated virus sero type 3B (AAV 3B): insights into receptor binding and immune evasion. Virology 403: 26 36. 178. Levy, H. C., V. D. Bowman, L. Govindasamy, R. McKenna, K. Nash, K. Warrington, W. Chen, N. Muzyczka, X. Yan, T. S. Baker, and M. Agbandje McKenna. 2009. Heparin b inding induces conformational changes in Adeno associated virus serotype 2. J Struct Biol 165: 146 156. 179. Li, Q., A. G. Yafal, Y. M. Lee, J. Hogle, and M. Chow. 1994. Poliovirus neutralization by antibodies to internal epitopes of VP4 and VP1 results fro m reversible exposure of these sequences at physiological temperature. J Virol 68: 3965 3970. 180. Ling, C., Y. Lu, J. K. Kalsi, G. R. Jayandharan, B. Li, W. Ma, B. Cheng, S. W. Gee, K. E. McGoogan, L. Govindasamy, L. Zhong, M. Agbandje McKenna, and A. Sriv astava. 2010. Human hepatocyte growth factor receptor is a cellular coreceptor for adeno associated virus serotype 3. Hum Gene Ther 21: 1741 1747. 181. Linser, P., H. Bruning, and R. W. Armentrout. 1977. Specific binding sites for a parvovirus, minute virus of mice, on cultured mouse cells. J Virol 24: 211 221.
189 182. Liu, F., H. L. Qi, and H. L. Chen. 2001. Regulation of differentiation and proliferation inducers on Lewis antigens, alpha fucosyltransferase and metastatic potential in hepatocarcinoma cells. Br J Cancer 84: 1556 1563. 183. Liu, J., and S. C. Thorp. 2002. Cell surface heparan sulfate and its roles in assisting viral infections. Med Res Rev 22: 1 25. 184. Llamas Saiz, A. L., M. Agbandje McKenna, W. R. Wikoff, J. Bratton, P. Tattersall, and M. G. Ros smann. 1997. Structure determination of minute virus of mice. Acta Crystallogr D Biol Crystallogr 53: 93 102. 185. Lombardo, E., J. C. Ramirez, M. Agbandje McKenna, and J. M. Almendral. 2000. A beta stranded motif drives capsid protein oligomers of the parv ovirus minute virus of mice into the nucleus for viral assembly. J Virol 74: 3804 3814. 186. Lombardo, E., J. C. Ramirez, J. Garcia, and J. M. Almendral. 2002. Complementary roles of multiple nuclear targeting signals in the capsid proteins of the parvoviru s minute virus of mice during assembly and onset of infection. J Virol 76: 7049 7059. 187. Lopez de Turiso, J. A., E. Cortes, C. Martinez, R. Ruiz de Ybanez, I. Simarro, C. Vela, and I. Casal. 1992. Recombinant vaccine for canine parvovirus in dogs. J Virol 66: 2748 2753. 188. Lopez Bueno, A., M. P. Rubio, N. Bryant, R. McKenna, M. Agbandje McKenna, and J. M. Almendral. 2006. Host selected amino acid changes at the sialic acid binding pocket of the parvovirus capsid modulate cell binding affinity and determin e virulence. J Virol 80: 1563 1573. 189. Lopez Bueno, A., J. C. Segovia, J. A. Bueren, M. G. O'Sullivan, F. Wang, P. Tattersall, and J. M. Almendral. 2008. Evolution to pathogenicity of the parvovirus minute virus of mice in immunodeficient mice involves ge netic heterogeneity at the capsid domain that determines tropism. J Virol 82: 1195 1203. 190. Lux, K., N. Goerlitz, S. Schlemminger, L. Perabo, D. Goldnau, J. Endell, K. Leike, D. M. Kofler, S. Finke, M. Hallek, and H. Buning. 2005. Green fluorescent protei n tagged adeno associated virus particles allow the study of cytosolic and nuclear trafficking. J Virol 79: 11776 11787. 191. Mani, B., C. Baltzer, N. Valle, J. M. Almendral, C. Kempf, and C. Ros. 2006. Low pH dependent endosomal processing of the incoming parvovirus minute virus of mice virion leads to externalization of the VP1 N terminal sequence (N VP1), N VP2 cleavage, and uncoating of the full length genome. J Virol 80: 1015 1024. 192. Mani, B., M. Gerber, P. Lieby, N. Boschetti, C. Kempf, and C. Ros. 2 007. Molecular mechanism underlying B19 virus inactivation and comparison to other parvoviruses. Transfusion 47: 1765 1774.
190 193. Maroto, B., J. C. Ramirez, and J. M. Almendral. 2000. Phosphorylation status of the parvovirus minute virus of mice particle: ma pping and biological relevance of the major phosphorylation sites. J Virol 74: 10892 10902. 194. Maroto, B., N. Valle, R. Saffrich, and J. M. Almendral. 2004. Nuclear export of the nonenveloped parvovirus virion is directed by an unordered protein signal ex posed on the capsid surface. J Virol 78: 10685 10694. 195. Matthews, B. W. 1968. Solvent content of protein crystals. J Mol Biol 33: 491 497. 196. Maxwell, I. H., A. L. Spitzer, F. Maxwell, and D. J. Pintel. 1995. The capsid determinant of fibrotropism for t he MVMp strain of minute virus of mice functions via VP2 and not VP1. J Virol 69: 5829 5832. 197. McKenna, R., L. L. Ilag, and M. G. Rossmann. 1994. Analysis of the Single stranded DNA Bacteriophage [phi] X174, Refined at a Resolution of 3 0 Journal of molecular biology 237: 517 543. 198. McKenna, R., N. H. Olson, P. R. Chipman, T. S. Baker, T. F. Booth, J. Christensen, B. Aasted, J. M. Fox, M. E. Bloom, J. B. Wolfinbarger, and M. Agbandje McKenna. 1999. Three dimensional structure of Aleutian mink diseas e parvovirus: implications for disease pathogenicity. J Virol 73: 6882 6891. 199. McPherson, A. 1982. Preparation and Analysis of Protein Crystals 1ed. Wiley, New York. 200. Miller, C. L., and D. J. Pintel. 2002. Interaction between parvovirus NS2 protein a nd nuclear export factor Crm1 is important for viral egress from the nucleus of murine cells. J Virol 76: 3257 3266. 201. Moehler, M., B. Blechacz, N. Weiskopf, M. Zeidler, W. Stremmel, J. Rommelaere, P. R. Galle, and J. J. Cornelis. 2001. Effective infecti on, apoptotic cell killing and gene transfer of human hepatoma cells but not primary hepatocytes by parvovirus H1 and derived vectors. Cancer Gene Ther 8: 158 167. 202. Moehler, M., M. Zeidler, J. Schede, J. Rommelaere, P. R. Galle, J. J. Cornelis, and M. H eike. 2003. Oncolytic parvovirus H1 induces release of heat shock protein HSP72 in susceptible human tumor cells but may not affect primary immune cells. Cancer Gene Ther 10: 477 480. 203. Monahan, P. E., C. D. Lothrop, J. Sun, M. L. Hirsch, T. Kafri, B. Ka ntor, R. Sarkar, D. M. Tillson, J. R. Elia, and R. J. Samulski. 2010. Proteasome inhibitors enhance gene delivery by AAV virus vectors expressing large genomes in hemophilia mouse and dog models: a strategy for broad clinical application. Mol Ther 18: 1907 1916. 204. Mueller, C., and T. R. Flotte. 2008. Clinical gene therapy using recombinant adeno associated virus vectors. Gene Ther 15: 858 863.
191 205. Munakata, Y., T. Saito Ito, K. Kumura Ishii, J. Huang, T. Kodera, T. Ishii, Y. Hirabayashi, Y. Koyanagi, and T. Sasaki. 2005. Ku80 autoantigen as a cellular coreceptor for human parvovirus B19 infection. Blood 106: 3449 3456. 206. Naeger, L. K., J. Cater, and D. J. Pintel. 1990. The small nonstructural protein (NS2) of the parvovirus minute virus of mice is requir ed for efficient DNA replication and infectious virus production in a cell type specific manner. J Virol 64: 6166 6175. 207. Naeger, L. K., N. Salom, and D. J. Pintel. 1993. NS2 is required for efficient translation of viral mRNA in minute virus of mice in fected murine cells. J Virol 67: 1034 1043. 208. Nam, H. J., B. L. Gurda, R. McKenna, M. Potter, B. Byrne, M. Salganik, N. Muzyczka, and M. Agbandje McKenna. 2011. Structural studies of adeno associated virus serotype 8 capsid transitions associated with en dosomal trafficking. J Virol 85: 11791 11799. 209. Nam, H. J., B. Gurda Whitaker, W. Y. Gan, S. Ilaria, R. McKenna, P. Mehta, R. A. Alvarez, and M. Agbandje McKenna. 2006. Identification of the sialic acid structures recognized by minute virus of mice and t he role of binding affinity in virulence adaptation. J Biol Chem 281: 25670 25677. 210. Nam, H. J., M. D. Lane, E. Padron, B. Gurda, R. McKenna, E. Kohlbrenner, G. Aslanidi, B. Byrne, N. Muzyczka, S. Zolotukhin, and M. Agbandje McKenna. 2007. Structure of a deno associated virus serotype 8, a gene therapy vector. J Virol 81: 12260 12271. 211. Nathwani, A. C., Tuddenham, E.G.D, Rangarajan, S., Rosales, C., McIntosh, J., Linch, D.C., Chowdary, P., Riddell, A., Pie, J., Harrington, C., O'Beirne, J., Smith, K., Pa si, J., Glader, B., Rustagi, P., Ng, C.Y.C., Kay, M.A., Zhou, J., Spence, Y., Morton, C.L., Allay, J., Coleman, J., Sleep, S., Cunningham, J.M., Srivastava, D., Basner Tschakarjan, E., Mingozzi, F., High, K.A., Gray, J.T., Reiss, U.M., Nienhuis, A.W., Davi doff, A.M. 2011. Adenovirus Associated Virus Vector Mediated Gene Transfer in Hemophilia B. the New England Journal of Medicine. 212. Navaza, J. 2001. Implementation of molecular replacement in AMoRe. Acta Crystallogr D Biol Crystallogr 57: 1367 1372. 213. Nelson, C. D., E. Minkkinen, M. Bergkvist, K. Hoelzer, M. Fisher, B. Bothner, and C. R. Parrish. 2008. Detecting small changes and additional peptides in the canine parvovirus capsid structure. J Virol 82: 10397 10407. 214. Ng, R., L. Govindasamy, B. L. Gur da, R. McKenna, O. G. Kozyreva, R. J. Samulski, K. N. Parent, T. S. Baker, and M. Agbandje McKenna. 2010. Structural characterization of the dual glycan binding adeno associated virus serotype 6. J Virol 84: 12945 12957.
192 215. Nonnenmacher, M., and T. Weber. 2011. Adeno associated virus 2 infection requires endocytosis through the CLIC/GEEC pathway. Cell Host Microbe 10: 563 576. 216. North, S., H. Huang, S. Sundaram, J. Jang Lee, A. Etienne, A. Trollope, S. Chalabi, A. Dell, P. Stanley, and S. Haslam. 2010. G lycomics Profiling of Chinese Hamster Ovary Cell Glycosylation Mutants Reveals N Glycans of a Novel Size and Complexity. Journal of Biological Chemistry 285: 5759 5775. 217. Nuesch, J., and J. Rommelaere. 2006. NS1 interaction with CKII alpha: Novel protein complex mediating parvovirus induced cytotoxicity. Journal of Virology 80: 4729 4739. 218. Nesch, J. P. F. 2006. Regulation of non structural protein functions by differential synthesis, modification and trafficking, p. 275 289. In J. R. Kerr, Cotmore, S. F., Bloom M.E., Linden, R.M., Parrish, C.R. (ed.), Parvoviruses. Hodder Arnold, London. 219. O'Donnell, J., K. A. Taylor, and M. S. Chapman. 2009. Adeno associated virus 2 and its primary cellular receptor -Cryo EM structure of a heparin complex. Virology 385: 434 443. 220. Ohshima, T., T. Nakajima, T. Oishi, N. Imamoto, Y. Yoneda, A. Fukamizu, and K. Yagami. 1999. CRM1 mediates nuclear export of nonstructural protein 2 from parvovirus minute virus of mice. Biochem Biophys Res Commun 264: 144 150. 221. Olijsl agers, S., A. Y. Dege, C. Dinsart, M. Voorhoeve, J. Rommelaere, M. H. Noteborn, and J. J. Cornelis. 2001. Potentiation of a recombinant oncolytic parvovirus by expression of Apoptin. Cancer Gene Ther 8: 958 965. 222. Olofsson, S., and T. Bergstrom. 2005. Gl ycoconjugate glycans as viral receptors. Ann Med 37: 154 172. 223. Opie, S. R., K. H. Warrington, Jr., M. Agbandje McKenna, S. Zolotukhin, and N. Muzyczka. 2003. Identification of amino acid residues in the capsid proteins of adeno associated virus type 2 t hat contribute to heparan sulfate proteoglycan binding. J Virol 77: 6995 7006. 224. Otwinowski, Z., and W. Minor. 1997. Processing of X ray diffraction data collected in oscillation mode. Methods in Enzymology 276: 307 326. 225. Padron, E., V. Bowman, N. Kal udov, L. Govindasamy, H. Levy, P. Nick, R. McKenna, N. Muzyczka, J. A. Chiorini, T. S. Baker, and M. Agbandje McKenna. 2005. Structure of adeno associated virus type 4. J Virol 79: 5047 5058.
193 226. Paglino, J., and P. Tattersall. 2011. The parvoviral capsid controls an intracellular phase of infection essential for efficient killing of stepwise transformed human fibroblasts. Virology 416: 32 41. 227. Pajusola, K., M. Gruchala, H. Joch, T. F. Luscher, S. Yla Herttuala, and H. Bueler. 2002. Cell type specific ch aracteristics modulate the transduction efficiency of adeno associated virus type 2 and restrain infection of endothelial cells. J Virol 76: 11530 11540. 228. Palermo, L. M., S. L. Hafenstein, and C. R. Parrish. 2006. Purified feline and canine transferrin receptors reveal complex interactions with the capsids of canine and feline parvoviruses that correspond to their host ranges. J Virol 80: 8482 8492. 229. Palermo, L. M., K. Hueffer, and C. R. Parrish. 2003. Residues in the apical domain of the feline and c anine transferrin receptors control host specific binding and cell infection of canine and feline parvoviruses. J Virol 77: 8915 8923. 230. Parker, J. S., W. J. Murphy, D. Wang, S. J. O'Brien, and C. R. Parrish. 2001. Canine and feline parvoviruses can use human or feline transferrin receptors to bind, enter, and infect cells. J Virol 75: 3896 3902. 231. Parker, J. S., and C. R. Parrish. 2000. Cellular uptake and infection by canine parvovirus involves rapid dynamin regulated clathrin mediated endocytosis, fo llowed by slower intracellular trafficking. J Virol 74: 1919 1930. 232. Parrish, C. R. 2006. Autonomous Parvovirus Variation and Evolution, p. 47 54. In J. R. Kerr, Cotmore, S.F., Bloom M.E., Linden, R.M., Parrish, C.R. (ed.), Parvoviruses. Hodder Arnold, L ondon. 233. Parrish, C. R., C. F. Aquadro, M. L. Strassheim, J. F. Evermann, J. Y. Sgro, and H. O. Mohammed. 1991. Rapid antigenic type replacement and DNA sequence evolution of canine parvovirus. J Virol 65: 6544 6552. 234. Parrish, C. R., P. Have, W. J. F oreyt, J. F. Evermann, M. Senda, and L. E. Carmichael. 1988. The global spread and replacement of canine parvovirus strains. J Gen Virol 69 ( Pt 5): 1111 1116. 235. Parrish, C. R., and Y. Kawaoka. 2005. The origins of new pandemic viruses: the acquisition o f new host ranges by canine parvovirus and influenza A viruses. Annu Rev Microbiol 59: 553 586. 236. Perez, R., M. Castellanos, A. Rodriguez Huete, and M. G. Mateu. 2011. Molecular determinants of self association and rearrangement of a trimeric intermediat e during the assembly of a parvovirus capsid. J Mol Biol 413: 32 40.
194 237. Perros, M., L. Deleu, J. M. Vanacker, Z. Kherrouche, N. Spruyt, S. Faisst, and J. Rommelaere. 1995. Upstream CREs participate in the basal activity of minute virus of mice promoter P4 and in its stimulation in ras transformed cells. J Virol 69: 5506 5515. 238. Pettersen, E. F., T. D. Goddard, C. C. Huang, G. S. Couch, D. M. Greenblatt, E. C. Meng, and T. E. Ferrin. 2004. UCSF Chimera -a visualization system for exploratory research and analysis. J Comput Chem 25: 1605 1612. 239. Pintel, D., D. Dadachanji, C. R. Astell, and D. C. Ward. 1983. The genome of minute virus of mice, an autonomous parvovirus, encodes two overlapping transcription units. Nucleic Acids Res 11: 1019 1038. 240. Plevka P., S. Hafenstein, L. Li, A. D'Abrgamo, Jr., S. F. Cotmore, M. G. Rossmann, and P. Tattersall. 2011. Structure of a packaging defective mutant of minute virus of mice indicates that the genome is packaged via a pore at a 5 fold axis. J Virol 85: 4822 4827 241. Prasad, K. M., and J. P. Trempe. 1995. The adeno associated virus Rep78 protein is covalently linked to viral DNA in a preformed virion. Virology 214: 360 370. 242. Previsani, N., S. Fontana, B. Hirt, and P. Beard. 1997. Growth of the parvovirus minu te virus of mice MVMp3 in EL4 lymphocytes is restricted after cell entry and before viral DNA amplification: cell specific differences in virus uncoating in vitro. J Virol 71: 7769 7780. 243. Qing, K., C. Mah, J. Hansen, S. Zhou, V. Dwarki, and A. Srivastav a. 1999. Human fibroblast growth factor receptor 1 is a co receptor for infection by adeno associated virus 2. Nat Med 5: 71 77. 244. Ramirez, J. C., A. Fairen, and J. M. Almendral. 1996. Parvovirus minute virus of mice strain i multiplication and pathogene sis in the newborn mouse brain are restricted to proliferative areas and to migratory cerebellar young neurons. J Virol 70: 8109 8116. 245. Reardon, J. E., and T. Spector. 1989. Herpes simplex virus type 1 DNA polymerase. Mechanism of inhibition by acyclovi r triphosphate. J Biol Chem 264: 7405 7411. 246. Reguera, J., A. Carreira, L. Riolobos, J. M. Almendral, and M. G. Mateu. 2004. Role of interfacial amino acid residues in assembly, stability, and conformation of a spherical virus capsid. Proc Natl Acad Sci U S A 101: 2724 2729. 247. Reguera, J., E. Grueso, A. Carreira, C. Sanchez Martinez, J. M. Almendral, and M. G. Mateu. 2005. Functional relevance of amino acid residues involved in interactions with ordered nucleic acid in a spherical virus. J Biol Chem 280 : 17969 17977.
195 248. Riolobos, L., J. Reguera, M. G. Mateu, and J. M. Almendral. 2006. Nuclear transport of trimeric assembly intermediates exerts a morphogenetic control on the icosahedral parvovirus capsid. J Mol Biol 357: 1026 1038. 249. Ritter, G., E. Boo sfeld, E. Markstein, R. K. Yu, S. L. Ren, W. B. Stallcup, H. F. Oettgen, L. J. Old, and P. O. Livingston. 1990. Biochemical and serological characteristics of natural 9 O acetyl GD3 from human melanoma and bovine buttermilk and chemically O acetylated GD3. Cancer Res 50: 1403 1410. 250. Rogers, G. N., J. C. Paulson, R. S. Daniels, J. J. Skehel, I. A. Wilson, and D. C. Wiley. 1983. Single amino acid substitutions in influenza haemagglutinin change receptor binding specificity. Nature 304: 76 78. 251. Rommelaer e, J., and J. J. Cornelis. 1991. Antineoplastic activity of parvoviruses. J Virol Methods 33: 233 251. 252. Ros, C., C. Baltzer, B. Mani, and C. Kempf. 2006. Parvovirus uncoating in vitro reveals a mechanism of DNA release without capsid disassembly and str iking differences in encapsidated DNA stability. Virology 345: 137 147. 253. Ros, C., C. J. Burckhardt, and C. Kempf. 2002. Cytoplasmic trafficking of minute virus of mice: low pH requirement, routing to late endosomes, and proteasome interaction. J Virol 7 6: 12634 12645. 254. Ros, C., M. Gerber, and C. Kempf. 2006. Conformational changes in the VP1 unique region of native human parvovirus B19 lead to exposure of internal sequences that play a role in virus neutralization and infectivity. J Virol 80: 12017 120 24. 255. Ros, C., and C. Kempf. 2004. The ubiquitin proteasome machinery is essential for nuclear translocation of incoming minute virus of mice. Virology 324: 350 360. 256. Rose, J. A., J. V. Maizel, Jr., J. K. Inman, and A. J. Shatkin. 1971. Structural pr oteins of adenovirus associated viruses. J Virol 8: 766 770. 257. Rosenfeld, S. J., K. Yoshimoto, S. Kajigaya, S. Anderson, N. S. Young, A. Field, P. Warrener, G. Bansal, and M. S. Collett. 1992. Unique region of the minor capsid protein of human parvovirus B19 is exposed on the virion surface. J Clin Invest 89: 2023 2029. 258. Rossmann, M. G., R. McKenna, L. Tong, D. Xia, J. Dai, H. Wu, H. K. Choi, and R. E. Lynch. 1992. Molecular replacement real space averaging. Journal of applied crystallography 25: 166 18 0. 259. Rubio, M. P., A. Lopez Bueno, and J. M. Almendral. 2005. Virulent variants emerging in mice infected with the apathogenic prototype strain of the parvovirus minute virus of mice exhibit a capsid with low avidity for a primary receptor. J Virol 79: 1 1280 11290.
196 260. Saliki, J. T., B. Mizak, H. P. Flore, R. R. Gettig, J. P. Burand, L. E. Carmichael, H. A. Wood, and C. R. Parrish. 1992. Canine parvovirus empty capsids produced by expression in a baculovirus vector: use in analysis of viral properties an d immunization of dogs. J Gen Virol 73 ( Pt 2): 369 374. 261. Sanlioglu, S., P. K. Benson, J. Yang, E. M. Atkinson, T. Reynolds, and J. F. Engelhardt. 2000. Endocytosis and nuclear trafficking of adeno associated virus type 2 are controlled by rac1 and phos phatidylinositol 3 kinase activation. J Virol 74: 9184 9196. 262. Sato, C., H. Fukuoka, K. Ohta, T. Matsuda, R. Koshino, K. Kobayashi, F. A. Troy, 2nd, and K. Kitajima. 2000. Frequent occurrence of pre existing alpha 2 ->8 linked disialic and oligosialic ac ids with chain lengths up to 7 Sia residues in mammalian brain glycoproteins. Prevalence revealed by highly sensitive chemical methods and anti di oligo and poly Sia antibodies specific for defined chain lengths. J Biol Chem 275: 15422 15431. 263. Sato, C., T. Matsuda, and K. Kitajima. 2002. Neuronal differentiation dependent expression of the disialic acid epitope on CD166 and its involvement in neurite formation in Neuro2A cells, p. 45299 45305, J Biol Chem, vol. 277, United States. 264. Schmidt, M., a nd J. A. Chiorini. 2006. Gangliosides are essential for bovine adeno associated virus entry. J Virol 80: 5516 5522. 265. Schnaar, R. L., A. Suzuki, and P. Stanley. 2009. Glycosphingolipids. In A. Varki, R. D. Cummings, J. D. Esko, H. H. Freeze, P. Stanley, C. R. Bertozzi, G. W. Hart, and M. E. Etzler (ed.), Essentials of Glycobiology. The Consortium of Glycobiology Editors, La Jolla, California, Cold Spring Harbor NY. 266. Schwegmann Wessels, C., and G. Herrler. 2006. Sialic acids as receptor determinants fo r coronaviruses. Glycoconjugate Journal 23: 51 58. 267. Sedlik, C., A. Dridi, E. Deriaud, M. F. Saron, P. Rueda, J. Sarraseca, J. I. Casal, and C. Leclerc. 1999. Intranasal delivery of recombinant parvovirus like particles elicits cytotoxic T cell and neutr alizing antibody responses. J Virol 73: 2739 2744. 268. Sedlik, C., J. Sarraseca, P. Rueda, C. Leclerc, and I. Casal. 1995. Immunogenicity of poliovirus B and T cell epitopes presented by hybrid porcine parvovirus particles. J Gen Virol 76 ( Pt 9): 2361 2368 269. Segovia, J. C., J. M. Gallego, J. A. Bueren, and J. M. Almendral. 1999. Severe leukopenia and dysregulated erythropoiesis in SCID mice persistently infected with the parvovirus minute virus of mice. J Virol 73: 1774 1784. 270. Seiler, M. P., A. D. Mi ller, J. Zabner, and C. L. Halbert. 2006. Adeno associated virus types 5 and 6 use distinct receptors for cell entry. Hum Gene Ther 17: 10 19.
197 271. Seki, T., and Y. Arai. 1993. Highly polysialylated neural cell adhesion molecule (NCAM H) is expressed by new ly generated granule cells in the dentate gyrus of the adult rat. J Neurosci 13: 2351 2358. 272. Shen, S., K. D. Bryant, S. M. Brown, S. H. Randell, and A. Asokan. 2011. Terminal N linked galactose is the primary receptor for adeno associated virus 9. J Bio l Chem 286: 13532 13540. 273. Shieh, M. T., and P. G. Spear. 1994. Herpesvirus induced cell fusion that is dependent on cell surface heparan sulfate or soluble heparin. J Virol 68: 1224 1228. 274. Simpson, A. A., V. Chandrasekar, B. Hebert, G. M. Sullivan, M G. Rossmann, and C. R. Parrish. 2000. Host range and variability of calcium binding by surface loops in the capsids of canine and feline parvoviruses. J Mol Biol 300: 597 610. 275. Simpson, A. A., P. R. Chipman, T. S. Baker, P. Tijssen, and M. G. Rossmann 1998. The structure of an insect parvovirus (Galleria mellonella densovirus) at 3.7 A resolution. Structure 6: 1355 1367. 276. Simpson, A. A., B. Hebert, G. M. Sullivan, C. R. Parrish, Z. Zadori, P. Tijssen, and M. G. Rossmann. 2002. The structure of porc ine parvovirus: comparison with related viruses. J Mol Biol 315: 1189 1198. 277. Song, X., B. Xia, S. Stowell, Y. Lasanajak, D. Smith, and R. Cummings. 2009. Novel Fluorescent Glycan Microarray Strategy Reveals Ligands for Galectins. Chemistry & Biology 16: 36 47. 278. Song, X., H. Yu, X. Chen, Y. Lasanajak, M. Tappert, G. Air, V. Tiwari, H. Cao, H. Chokhawala, H. Zheng, R. Cummings, and D. Smith. 2011. A Sialylated Glycan Microarray Reveals Novel Interactions of Modified Sialic Acids with Proteins and Viruse s. Journal of Biological Chemistry 286: 31610 31622. 279. Sonntag, F., S. Bleker, B. Leuchs, R. Fischer, and J. A. Kleinschmidt. 2006. Adeno associated virus type 2 capsids with externalized VP1/VP2 trafficking domains are generated prior to passage through the cytoplasm and are maintained until uncoating occurs in the nucleus. J Virol 80: 11040 11054. 280. Sonntag, F., K. Kother, K. Schmidt, M. Weghofer, C. Raupp, K. Nieto, A. Kuck, B. Gerlach, B. Bottcher, O. J. Muller, K. Lux, M. Horer, and J. A. Kleinschm idt. 2011. The assembly activating protein promotes capsid assembly of different adeno associated virus serotypes. J Virol 85: 12686 12697. 281. Sonntag, F., K. Schmidt, and J. A. Kleinschmidt. 2010. A viral assembly factor promotes AAV2 capsid formation in the nucleolus. Proc Natl Acad Sci U S A 107: 10220 10225.
198 282. Spalholz, B. A., and P. Tattersall. 1983. Interaction of minute virus of mice with differentiated cells: strain dependent target cell specificity is mediated by intracellular factors. J Virol 4 6: 937 943. 283. Stahnke, S., K. Lux, S. Uhrig, F. Kreppel, M. Hosel, O. Coutelle, M. Ogris, M. Hallek, and H. Buning. 2011. Intrinsic phospholipase A2 activity of adeno associated virus is involved in endosomal escape of incoming particles. Virology 409: 77 83. 284. Stanford, M. M., M. Shaban, J. W. Barrett, S. J. Werden, P. A. Gilbert, J. Bondy Denomy, L. Mackenzie, K. C. Graham, A. F. Chambers, and G. McFadden. 2008. Myxoma virus oncolysis of primary and metastatic B16F10 mouse tumors in vivo. Mol Ther 16: 52 59. 285. Stieger, K., T. Cronin, J. Bennett, and F. Rolling. 2011. Adeno associated virus mediated gene therapy for retinal degenerative diseases. Methods Mol Biol 807: 179 218. 286. Suikkanen, S., T. Aaltonen, M. Nevalainen, O. Valilehto, L. Lindholm, M Vuento, and M. Vihinen Ranta. 2003. Exploitation of microtubule cytoskeleton and dynein during parvoviral traffic toward the nucleus. J Virol 77: 10270 10279. 287. Suikkanen, S., M. Antila, A. Jaatinen, M. Vihinen Ranta, and M. Vuento. 2003. Release of ca nine parvovirus from endocytic vesicles. Virology 316: 267 280. 288. Suikkanen, S., K. Saajarvi, J. Hirsimaki, O. Valilehto, H. Reunanen, M. Vihinen Ranta, and M. Vuento. 2002. Role of recycling endosomes and lysosomes in dynein dependent entry of canine pa rvovirus. J Virol 76: 4401 4411. 289. Summerford, C., J. S. Bartlett, and R. J. Samulski. 1999. AlphaVbeta5 integrin: a co receptor for adeno associated virus type 2 infection. Nat Med 5: 78 82. 290. Summerford, C., and R. J. Samulski. 1998. Membrane associa ted heparan sulfate proteoglycan is a receptor for adeno associated virus type 2 virions. J Virol 72: 1438 1445. 291. Takeuchi, Y., M. Akutsu, K. Murayama, N. Shimizu, and H. Hoshino. 1991. Host range mutant of human immunodeficiency virus type 1: modificat ion of cell tropism by a single point mutation at the neutralization epitope in the env gene. J Virol 65: 1710 1718. 292. Talora, C., D. C. Sgroi, C. P. Crum, and G. P. Dotto. 2002. Specific down modulation of Notch1 signaling in cervical cancer cells is re quired for sustained HPV E6/E7 expression and late steps of malignant transformation. Genes Dev 16: 2252 2263.
199 293. Tattersall, P. 2006. The Evolution of Parvovirus Taxonomy, p. 5 14. In J. R. Kerr, Cotmore, S.F., Bloom M.E., Linden, R.M., Parrish, C.R. (ed .), Parvoviruses. Hodder Arnold, London. 294. Tattersall, P., and J. Bratton. 1983. Reciprocal productive and restrictive virus cell interactions of immunosuppressive and prototype strains of minute virus of mice. J Virol 46: 944 955. 295. Tattersall, P., P J. Cawte, A. J. Shatkin, and D. C. Ward. 1976. Three structural polypeptides coded for by minite virus of mice, a parvovirus. J Virol 20: 273 289. 296. Tattersall, P., A. J. Shatkin, and D. C. Ward. 1977. Sequence homology between the structural polypepti des of minute virus of mice. J Mol Biol 111: 375 394. 297. Thacker, T. C., and F. B. Johnson. 1998. Binding of bovine parvovirus to erythrocyte membrane sialylglycoproteins. J Gen Virol 79 ( Pt 9): 2163 2169. 298. Theobald, D. L., R. M. Mitton Fry, and D. S. Wuttke. 2003. Nucleic acid recognition by OB fold proteins. Annu Rev Biophys Biomol Struct 32: 115 133. 299. Tong, L., and M. G. Rossmann. 1997. Rotation function calculations with GLRF program. Methods Enzymol 276: 594 611. 300. Toolan, H. W. 1990. The rod ent parvoviruses, p. 159 176. In P. Tjissen (ed.), Handbook of parvoviruses. CRC Press, Boca Raton, Fla. 301. Toolan, H. W., and N. Ledinko. 1968. Inhibition by H 1 virus of the incidence of tumors produced by adenovirus 12 in hamsters. Virology 35: 475 478 302. Toolan, H. W., E. L. Saunders, C. M. Southam, A. E. Moore, and A. G. Levin. 1965. H 1 VIRUS VIREMIA IN THE HUMAN. Proc Soc Exp Biol Med 119: 711 715. 303. Tresnan, D. B., L. Southard, W. Weichert, J. Y. Sgro, and C. R. Parrish. 1995. Analysis of the cell and erythrocyte binding activities of the dimple and canyon regions of the canine parvovirus capsid. Virology 211: 123 132. 304. Truyen, U., A. Gruenberg, S. F. Chang, B. Obermaier, P. Veijalainen, and C. R. Parrish. 1995. Evolution of the feline subgr oup parvoviruses and the control of canine host range in vivo. J Virol 69: 4702 4710. 305. Truyen, U., and C. R. Parrish. 1992. Canine and feline host ranges of canine parvovirus and feline panleukopenia virus: distinct host cell tropisms of each virus in v itro and in vivo. J Virol 66: 5399 5408.
200 306. Tsao, J., M. S. Chapman, M. Agbandje, W. Keller, K. Smith, H. Wu, M. Luo, T. J. Smith, M. G. Rossmann, R. W. Compans, and et al. 1991. The three dimensional structure of canine parvovirus and its functional impl ications. Science 251: 1456 1464. 307. Urzhumtseva, L., P. Afonine, P. Adams, and A. Urzhumtsev. 2009. Crystallographic model quality at a glance. Acta Crystallographica Section D Biological Crystallography 65: 297 300. 308. Valegard, K., J. B. Murray, P. G. Stockley, N. J. Stonehouse, and L. Liljas. 1994. Crystal structure of an RNA bacteriophage coat protein operator complex. Nature 371: 623 626. 309. Varki, N. M., and A. Varki. 2007. Diversity in cell surface sialic acid presentations: implications for biol ogy and disease. Lab Invest 87: 851 857. 310. Vihinen Ranta, M., L. Kakkola, A. Kalela, P. Vilja, and M. Vuento. 1997. Characterization of a nuclear localization signal of canine parvovirus capsid proteins. Eur J Biochem 250: 389 394. 311. Vihinen Ranta, M., A. Kalela, P. Makinen, L. Kakkola, V. Marjomaki, and M. Vuento. 1998. Intracellular route of canine parvovirus entry. J Virol 72: 802 806. 312. Vihinen Ranta, M., D. Wang, W. S. Weichert, and C. R. Parrish. 2002. The VP1 N terminal sequence of canine parvo virus affects nuclear transport of capsids and efficient cell infection. J Virol 76: 1884 1891. 313. Vihinen Ranta, M., W. Yuan, and C. R. Parrish. 2000. Cytoplasmic trafficking of the canine parvovirus capsid and its role in infection and nuclear transport J Virol 74: 4853 4859. 314. Walters, R. W., M. Agbandje McKenna, V. D. Bowman, T. O. Moninger, N. H. Olson, M. Seiler, J. A. Chiorini, T. S. Baker, and J. Zabner. 2004. Structure of adeno associated virus serotype 5. J Virol 78: 3361 3371. 315. Walters, R. W., S. M. Yi, S. Keshavjee, K. E. Brown, M. J. Welsh, J. A. Chiorini, and J. Zabner. 2001. Binding of adeno associated virus type 5 to 2,3 linked sialic acid is required for gene transfer. J Biol Chem 276: 20610 20616. 316. Weigel Kelley, K. A., M. C. Yode r, and A. Srivastava. 2003. Alpha5beta1 integrin as a cellular coreceptor for human parvovirus B19: requirement of functional activation of beta1 integrin for viral entry. Blood 102: 3927 3933. 317. Weigel Kelley, K. A., M. C. Yoder, and A. Srivastava. 2001 Recombinant human parvovirus B19 vectors: erythrocyte P antigen is necessary but not sufficient for successful transduction of human hematopoietic cells. J Virol 75: 4110 4116.
201 318. Weiss, M. S., A. Jabs, and R. Hilgenfeld. 1998. Peptide bonds revisited. Nat Struct Biol 5: 676. 319. Weller, M. L., P. Amornphimoltham, M. Schmidt, P. A. Wilson, J. S. Gutkind, and J. A. Chiorini. 2010. Epidermal growth factor receptor is a co receptor for adeno associated virus serotype 6. Nat Med 16: 662 664. 320. Wetzel, K., P. Menten, G. Opdenakker, J. Van Damme, H. J. Grone, N. Giese, A. Vecchi, S. Sozzani, J. J. Cornelis, J. Rommelaere, and C. Dinsart. 2001. Transduction of human MCP 3 by a parvoviral vector induces leukocyte infiltration and reduces growth of human cervica l carcinoma cell xenografts. J Gene Med 3: 326 337. 321. Wikoff, W. R., G. Wang, C. R. Parrish, R. H. Cheng, M. L. Strassheim, T. S. Baker, and M. G. Rossmann. 1994. The structure of a neutralized virus: canine parvovirus complexed with neutralizing antibod y fragment. Structure 2: 595 607. 322. Wu, H., and M. G. Rossmann. 1993. The canine parvovirus empty capsid structure. J Mol Biol 233: 231 244. 323. Wu, P., W. Xiao, T. Conlon, J. Hughes, M. Agbandje McKenna, T. Ferkol, T. Flotte, and N. Muzyczka. 2000. Muta tional analysis of the adeno associated virus type 2 (AAV2) capsid gene and construction of AAV2 vectors with altered tropism. J Virol 74: 8635 8647. 324. Wu, Z., A. Asokan, J. C. Grieger, L. Govindasamy, M. Agbandje McKenna, and R. J. Samulski. 2006. Singl e amino acid changes can influence titer, heparin binding, and tissue tropism in different adeno associated virus serotypes. J Virol 80: 11393 11397. 325. Wu, Z., E. Miller, M. Agbandje McKenna, and R. J. Samulski. 2006. Alpha2,3 and alpha2,6 N linked siali c acids facilitate efficient binding and transduction by adeno associated virus types 1 and 6. J Virol 80: 9093 9103. 326. Xiao, W., K. H. Warrington, Jr., P. Hearing, J. Hughes, and N. Muzyczka. 2002. Adenovirus facilitated nuclear translocation of adeno a ssociated virus type 2. J Virol 76: 11505 11517. 327. Xiao, X., W. Xiao, J. Li, and R. J. Samulski. 1997. A novel 165 base pair terminal repeat sequence is the sole cis requirement for the adeno associated virus life cycle. J Virol 71: 941 948. 328. Xie, Q., W. Bu, S. Bhatia, J. Hare, T. Somasundaram, A. Azzi, and M. S. Chapman. 2002. The atomic structure of adeno associated virus (AAV 2), a vector for human gene therapy. Proc Natl Acad Sci U S A 99: 10405 10410. 329. Xie, Q., and M. S. Chapman. 1996. Canine p arvovirus capsid structure, analyzed at 2.9 A resolution. J Mol Biol 264: 497 520.
202 330. Xie, Q., T. F. Lerch, N. L. Meyer, and M. S. Chapman. 2011. Structure function analysis of receptor binding in adeno associated virus serotype 6 (AAV 6). Virology 420: 10 19. 331. Yan, Z., D. C. Lei Butters, X. Liu, Y. Zhang, L. Zhang, M. Luo, R. Zak, and J. F. Engelhardt. 2006. Unique biologic properties of recombinant AAV1 transduction in polarized human airway epithelia. J Biol Chem 281: 29684 29692. 332. Yan, Z., R. Zak G. W. Luxton, T. C. Ritchie, U. Bantel Schaal, and J. F. Engelhardt. 2002. Ubiquitination of both adeno associated virus type 2 and 5 capsid proteins affects the transduction efficiency of recombinant vectors. J Virol 76: 2043 2053. 333. Yan, Z., R. Zak, Y. Zhang, W. Ding, S. Godwin, K. Munson, R. Peluso, and J. F. Engelhardt. 2004. Distinct classes of proteasome modulating agents cooperatively augment recombinant adeno associated virus type 2 and type 5 mediated transduction from the apical surfaces of hu man airway epithelia. J Virol 78: 2863 2874. 334. Yunoki, M., M. Tsujikawa, T. Urayama, Y. Sasaki, M. Morita, H. Tanaka, S. Hattori, K. Takechi, and K. Ikuta. 2003. Heat sensitivity of human parvovirus B19. Vox Sang 84: 164 169. 335. Zadori, Z., J. Szelei, M C. Lacoste, Y. Li, S. Gariepy, P. Raymond, M. Allaire, I. R. Nabi, and P. Tijssen. 2001. A viral phospholipase A2 is required for parvovirus infectivity. Dev Cell 1: 291 302. 336. Zhong, L., B. Li, G. Jayandharan, C. S. Mah, L. Govindasamy, M. Agbandje Mc Kenna, R. W. Herzog, K. A. Weigel Van Aken, J. A. Hobbs, S. Zolotukhin, N. Muzyczka, and A. Srivastava. 2008. Tyrosine phosphorylation of AAV2 vectors and its consequences on viral intracellular trafficking and transgene expression. Virology 381: 194 202. 3 37. Zhong, L., B. Li, C. S. Mah, L. Govindasamy, M. Agbandje McKenna, M. Cooper, R. W. Herzog, I. Zolotukhin, K. H. Warrington, Jr., K. A. Weigel Van Aken, J. A. Hobbs, S. Zolotukhin, N. Muzyczka, and A. Srivastava. 2008. Next generation of adeno associate d virus 2 vectors: point mutations in tyrosines lead to high efficiency transduction at lower doses. Proc Natl Acad Sci U S A 105: 7827 7832. 338. Zhou, L., Y. Luo, Y. Wu, J. Tsao, and M. Luo. 2000. Sialylation of the host receptor may modulate entry of dem yelinating persistent Theiler's virus. J Virol 74: 1477 1485.
203 BIOGRAPHICAL SKETCH Sujata Halder was born in 1983 in the capital city of New Delhi, India. She spent most of her childhood and schooling years in New Delhi. She completed her high school ed ucation at Birla Vidya Niketan, India in 2001. Towards the later years of her high school education, she became very interested in Biology, and wanted to become a doctor just like her father. However, after graduating from high school she enrolled in Bache lor of Science (B.Sc.) at Sri Venkateswara College, University of Delhi, India and graduated in 2004 with an honors in Biochemistry. During this time she conducted summer research at School of Environmental Sciences, Jawaharlal Nehru University, New Delhi India gaining experience in molecular biology techniques under the guidance of Dr. Sudha Bhattacharya. This experience kindled within her a keen interest for further research She then enrolled in Masters of Science (M.Sc ) in Bio t echnology from the este emed Indian Institute of Technology (Roork e e), India and graduated in 2006 During her master's program Sujata developed deep passion for serious scientific research where under the supervision of Dr Partha Roy she worked on her thesis in developing fun ctional assays for endocrine disruptors present in the Sujata decided to explore new and better opportunities and decided to apply to the graduate schools in the United States of America for a PhD degr ee. However, she was undecided about any particular field of study and wanted to learn new techniques before committing to join a laboratory, and hence only applied to schools with interdisciplinary programs that offered lab rotation experience. In the fa ll of 2006 she joined the I nterdisciplinary P rogram (IDP) at University of Florida, Gainesville (UFL). F ollowing the first year of core courses and lab rotations, she finally began research as a graduate assistant under the supervision of
204 Dr. M a vis Agband je McKenna (Professor, Department of Biochemistry and Molecular Biology UFL) in the field of virus crystallography Her areas of interest include virology, macromolecular crystallography and cancer biology. Her favorite hobby is amateur photography.