Structural and biochemical characterization of a human adenovirus 2/12 penton base chimera Chloe Zubieta 1, *, Laurent Blanchoin 2 and Stephen Cusack 1 1 European Molecular Biology Laboratory, Grenoble Outstation, France 2 Laboratoire de Physiologie Cellulaire Vegetale, Commissariat a l’Energie Atomique, Centre National de la Recherche Scientifique, Institut National de la Recherche Agronomique, Universite Joseph Fourier, Unite Mixte de Recherche 5168, Grenoble, France Adenoviruses are nonenveloped double-stranded DNA viruses found in mammalian and non-mammalian vertebrates. Human adenoviruses are divided into six subgroups (A–F) based on genetic organization, hema- gluttination patterns, immuno-crossreactivity, and nuc- leotide content. Over 50 serotypes have been identified in humans; these cause generally mild respiratory, enteric and ocular disease. However, in immunocom- promised, very young, or elderly individuals, adenoviral infections can lead to serious illness or death [1,2]. Apart from their role as a common human pathogen, adenoviruses are one of the most studied vectors for gene delivery due to extensive knowledge of their biology and the ability to manipulate the adenoviral genome [3–5]. The adenoviral T ¼ 25 icosahedral capsid consists of three major polypeptides: the trimeric hexon, which forms the facets of the particle, the pentameric penton base (pb), which forms the vertices, and the trimeric fiber protein, which extends from the penton base at the vertex positions. Additionally, cementing proteins such as IIIa, VI, VIII and IX help stabilize the capsid, reinforcing the penton–hexon and hexon–hexon inter- actions. The capsid contains 720 copies of the hexon Keywords adenovirus; crystal structure; fiber; fluorescence anisotropy; penton Correspondence C. Zubieta, SLAC ⁄ SSRL, 2575 Sandhill Road, Menlo Park, CA 94025, USA Fax: +1 650 926 3292 Tel: +1 650 926 2992 E-mail: czubieta@slac.stanford.edu *Present address Stanford Synchrotron Radiation Facility, Menlo Park, CA, USA (Received 6 June 2006, revised 11 July 2006, accepted 27 July 2006) doi:10.1111/j.1742-4658.2006.05430.x The vertex of the adenoviral capsid is formed by the penton, a complex of two proteins, the pentameric penton base and the trimeric fiber protein. The penton contains all necessary components for viral attachment and entry into the host cell. After initial attachment via the head domain of the fiber protein, the penton base interacts with cellular integrins through an Arg-Gly-Asp (RGD) motif located in a hypervariable surface loop, trigger- ing virus internalization. In order to investigate the structural and func- tional role of this region, we replaced the hypervariable loop of serotype 2 with the corresponding, but much shorter, loop of serotype 12 and com- pared it to the wild type. Here, we report the 3.6 A ˚ crystal structure of a human adenovirus 2 ⁄ 12 penton base chimera crystallized as a dodecamer. The structure is generally similar to human adenovirus 2 penton base, with the main differences localized to the fiber protein-binding site. Fluorescence anisotropy assays using a trimeric fiber protein mimetic called the minifiber and wild-type human adenovirus 2 and chimeric penton base demonstrate that fiber protein binding is independent of the hypervariable loop, with a K d for fiber binding estimated in the 1–2 lm range. Interestingly, competi- tion assays using labeled and unlabeled minifiber demonstrated virtually irreversible binding to the penton base, which we ascribe to a conforma- tional change, on the basis of comparisons of all available penton base structures. Abbreviations hAd, human adenovirus; MPD, 2-methyl-2,4-pentane diol; NCS, noncrystallographic symmetry averaging; pb, penton base; TMR, tetramethylrhodamine. 4336 FEBS Journal 273 (2006) 4336–4345 ª 2006 The Authors Journal compilation ª 2006 FEBS and 60 copies of the penton base monomers. In addi- tion to its role as a critical component of the viral cap- sid, the penton base also contains an Arg-Gly-Asp (RGD) motif that acts as a trigger for endocytosis of the adenovirus into the host cell. Furthermore, the architecture of the entire virus, including the positions of the cementing proteins, has been revealed in detail by recent high-resolution cryoelectron microscopy reconstructions [6–11]. Initial cell attachment occurs via interactions of the fiber protein C-terminal head domain. For the major- ity of serotypes, including adenoviruses 2 and 12, the Coxsackie- and adenovirus receptor (CAR), an integral membrane protein, shown to be a component of the tight junction in epithelial cells [12], is the primary cel- lular receptor. Upon virus attachment, the RGD motif of the penton base binds to a V b 3 , a V b 5 ,ora V b 1 inte- grin, which acts as a secondary receptor [13–15]. This interaction of the RGD motif with the host cell causes clustering of integrins and activates a signaling cas- cade, allowing viral entry into clathrin-coated pits and endosomes [14,16]. It is thought that upon integrin binding and before endocytosis, the fiber protein is shed from the virus [13,17–19]. One hypothesis for the mechanism of fiber removal from the virus involves a conformational change triggered by the penton base binding to integrin. The process by which integrin binding may be coupled to structural changes in the penton base, allowing fiber release, is not clear, but could involve changes in the hypervariable loop upon integrin binding being communicated to the fiber-bind- ing region of the penton base. Although the adenovirus penton base exhibits a high degree of sequence homology (typically over 70% homology between different subgroups of human adenoviruses), the RGD motif is located in a hyper- variable loop region of the penton base protein. This region can vary in length from approximately eight residues in human adenovirus serotype 12 (hAd12) to over 70 residues in human adenovirus 2 (hAd2). Previ- ous crystallographic studies of the hAd2 penton base revealed this region to be very flexible, with the hyper- variable loop almost completely disordered. At a unique crystal contact, a helix turn from the hypervari- able loop was identified; however, the amino acid sequence could not be definitively assigned [20]. Cryo- electron microscopy reconstructions of adenovirus 2 [21] with an RGD-recognizing antibody and with a soluble a V b 5 integrin [16] show only diffuse density for the hypervariable loop region; however, the resolution was limited to  21 A ˚ . In contrast, cryoelectron micro- scopy reconstructions of complexes with soluble inte- grin and hAd12 gave more defined density in the region of the RGD motif due to the better order of the shorter hypervariable loop in this serotype [16]. Higher-resolution cryoelectron microscopy reconstruc- tions (12 A ˚ ) of dodecahedral penton and penton base particles formed by the hAd3 serotype revealed addi- tional density in the region of the hypervariable loop, implying some favored conformations of the region [22]. The electron microscopy data for serotypes hAd3 and hAd12 and the crystallographic data for hAd2 raise the possibility that the hypervariable loop could possess secondary structural elements potentially important for integrin recognition or binding. We constructed an hAd2pb ⁄ 12pb chimera by repla- cing the long hypervariable loop ( 73 residues) of hAd2pb with the short hypervariable loop ( 8 resi- dues) of hAd12pb. Note that the sequence identity between hAd2 and hAd12 penton bases outside the hypervariable region is approximately 78%. We hypo- thesized that the shorter hypervariable loop of hAd12 would be more structured relative to the longer hAd2 hypervariable loop and allow for crystallographic characterization of this region of the penton base. Additionally, the chimeric construct would enable us to assess its possible role in the binding of the fiber protein in serotypes such as hAd2, which contain a long hypervariable region. The long hypervariable loop could directly interact with the fiber protein, but the shorter hAd12 hypervariable loop would be unable to reach the fiber protein-binding site. Thus, by dra- matically shortening the hypervariable loop to the hAd12 sequence, we eliminated any possible direct interactions between the hypervariable loop of the penton base and the fiber protein. We present here the 3.6 A ˚ crystal structure of an hAd2pb ⁄ 12pb chimera and fluorescence anisotropy binding data of the wild type and the chimeric construct with a fiber protein mimetic. Results and Discussion The 73-residue hypervariable loop of hAd2pb was replaced with the eight-residue hypervariable loop from hAd12 (Fig. 1). Owing to the proteolytic sensi- tivity of the construct, a more stable N-terminally truncated (49-TGGR ) version was designed. All structural studies presented here used the ) 49 N-ter- minally truncated hAd2pb ⁄ 12pb chimera. Analogous to the hAd2 penton base, the hAd2 ⁄ 12 chimeric construct formed well-behaved pentamers in solution, as confirmed by gel filtration and negative- stain electron microscopy (data not shown). As with hAd2pb, the solvent environment has marked effects on the equilibrium between pentamer and dodecamer C. Zubieta et al. Human adenovirus 2 ⁄ 12 penton base chimera FEBS Journal 273 (2006) 4336–4345 ª 2006 The Authors Journal compilation ª 2006 FEBS 4337 (regular arrangements of 12 pentamers), and the chi- mera formed dodecahedral particles under crystalliza- tion conditions. These crystallization conditions were, however, different from those for hAd2. For hAd2pb, high concentrations of ammonium sulfate (1.6 m) and dioxane (10%) were necessary for dodecahedra forma- tion [20], whereas hAd2 ⁄ 12pb crystallized in the dodecahedral form in high (50% v ⁄ v) 2-methyl-2,4- pentane diol (MPD) solvent conditions. The ionic strength of the solvent environment and ⁄ or presence of small organic compounds (dioxane and MPD) seems to favor formation of dodecahedra over isolated pentamers. Previous structural analysis of the residues at the penton–penton interface of the hAd2pb dodeca- hedron showed that there is no obvious pattern char- acterizing the penton bases of those serotypes which do or do not form dodecahedra during adenoviral infection [20]. Indeed, we suspect that all penton bases are capable of forming dodecahedra, depending on two factors: first, the solvent conditions affecting the dodecamer–pentamer equilibrium, and second, proteo- lysis or truncation of the penton base N-terminus. It has recently been shown that formation of hAd3 dodecahedra is favored by truncation of the N-ter- minal region [16], and crystals of the hAd2 and hAd2 ⁄ 12 chimera dodecahedra were obtained with protein N-terminally truncated by 49 residues [19]. The structure shows that the truncated N-terminus of the protein faces the interior of the particle, and small solvent channels between pentamers are not extensive enough to provide egress for the N-termini from the particle while retaining the dodecahedral structure. Calculations of the interior volume of the dodeca- hedral particle based on a cavity radius of 40 A ˚ in hAd2 ⁄ 12pb and hAd2pb give a volume of  300 000 A ˚ 3 , too small to accommodate an extra 2940 residues corresponding to the N-terminal exten- sion from the 60 copies of the full-length protein. The formation of small subviral particles by truncated viral capsid proteins is not limited to adenovirus. For exam- ple, this phenomenon of spontaneous assembly into regular particles has been noted for recombinant human papilloma virus L1 protein, a pentameric cap- sid protein. Only upon a 10-residue N-terminal trunca- tion does the L1 protein form small virus-like particles of icosahedral symmetry [23]. The physiologic role, if any, of penton base dode- camer formation in infected cells has not been defined for adenovirus. However, the N-terminal extremity of the penton base contains two PPxY motifs, shown to interact with the WW domains of cellular ubiquitin ligases [24,25]. As these motifs are presumably import- ant for adenovirus infection, it would be undesirable to incorporate truncated penton bases into virions, and dodecahedron formation could be a mechanism to avoid this. Fig. 1. Sequence alignment of adenovirus penton base. hAd2 (accession number P03276), hAd12 (accession number P36716) and the hAd2 ⁄ 12 chimera are aligned with the consensus secondary structure above the sequence. b-Strands are in orange and helices are in green. The hypervariable loop is highlighted in yellow, with the RGD motif in red. Human adenovirus 2 ⁄ 12 penton base chimera C. Zubieta et al. 4338 FEBS Journal 273 (2006) 4336–4345 ª 2006 The Authors Journal compilation ª 2006 FEBS Overall structure The hAd2⁄ 12pb construct crystallized in space group I222 with unit cell dimensions a ¼ 266.50, b ¼ 292.92, and c ¼ 307.30, and three pentamers (1 ⁄ 4 dodecahed- ron) per asymmetric unit (Table 1). The complete do- decahedral particle of 12 pentamers was formed by the space group symmetry operators (Fig. 2). The original hAd2pb, with a full-length hypervariable loop, crystal- lized in a C2 cell with one dodecahedron per asymmet- ric unit, but exhibited a similar packing arrangement to the chimeric protein [20]. The structure was phased by molecular replacement using the hAd2pb pentamer as a model and searching for three copies in the asym- metric unit. Initial maps were improved by rigid body refinement followed by 15-fold noncrystallographic symmetry averaging (NCS) of the electron density. Owing to the high degree of NCS, the electron density maps were of high quality and readily interpretable. The hAd2 ⁄ 12pb forms a stable pentamer under physiologic conditions. Pentamer formation involves contacts throughout the polypeptide chain and hides a large amount of the hydrophobic surface area. Surface area calculations for the pentamer give a total surface area of  81 000 A ˚ 2 with 30% ( 24 000 A ˚ 2 ) as con- tact area. Thus, a large amount of the available surface area of the molecule is buried upon pentamerization, increasing the stability of the protein. The hAd2 ⁄ 12 penton base monomer can be divided into a basal jellyroll domain formed by two b-sheets and a distal domain formed by insertions between the strands of the jellyroll domain (Fig. 3). Antiparallel b-sheets made up of strands CHEF and BIDG pack against each other, forming the jellyroll topology, a typical viral capsid protein fold [26]. The distal domain is formed by an insertion of  230 residues (residues 133–367) between strands D and E and a smaller insertion of  50 residues (residues 404–458) between strands F and G. The hypervariable loop, Table 1. Data collection and refinement statistics for hAd2/12 pen- ton base. Ad2 ⁄ 12 penton base chimera Data collection statistics Space group I222 Cell dimensions (A ˚ ) a ¼ 266.50 b ¼ 292.92 c ¼ 307.30 Oscillation range 0.5° Contents of asymmetric unit 3 pentamers Observations Total measured reflections 363 715 Unique reflections 124 622 Resolution (A ˚ ) 19.8–3.6 Completeness a (3.69–3.6) 0.90 (0.92) R factor a,b 0.13 (0.64) I ⁄ r 4.12 (1.33) Average measurement redundancy 2.9 NCS redundancy 15-fold Refinement statistics Resolution (A ˚ ) 15.0–3.6 Working reflections 115 706 Test reflections 3041 Number of nonhydrogen atoms 53 065 NCS averaging statistics c Correlation ⁄ R fac d 0.90 ⁄ 0.32 Model refinement R cryst ⁄ R free (%) e 27.5 ⁄ 32.8 Model geometry RMSD bond lengths (A ˚ ) 0.015 RMSD bond angles (°) 1.68 Ramachandran plot Favored + additional regions 96.7% Disallowed (number) 3.3% (3) a Values in parentheses refer to the highest resolution shell. b R factor ¼ S|I ) <I> | ⁄SI,where<I> is the average value of a reflection, I. c Averaging with RAVE to high resolution limit with 15-fold sym- metry. d Correlation between the densities of all noncrystallographic symmetry averaging (NCS)-related points. R fac ¼ S|F obs ) F map | ⁄ SF obs , where F map is the Fourier coefficient of the back-transformed averaged map. e Tight NCS restraints were used for all backbone atoms and relaxed NCS restraints were used for side chains during final refinements. R cryst ¼ S|F obs ) F calc | ⁄SF obs , where summation is over data used in the refinement. R free was calculated using 2.5% of the observed reflections excluded from refinement. Exclu- ded data were randomly selected. Fig. 2. Oligomerization of hAd2 ⁄ 12 penton base. (A) Monomer, pentamer, and dodecamer. The orientation of the monomer in red is kept throughout. (B) Asymmetric unit consisting of three pentam- ers (left) and the full dodecamer formed by the space group sym- metry (right). The pentamers in the asymmetric unit are in red. C. Zubieta et al. Human adenovirus 2 ⁄ 12 penton base chimera FEBS Journal 273 (2006) 4336–4345 ª 2006 The Authors Journal compilation ª 2006 FEBS 4339 containing the RGD motif, is located at the top of the distal domain within the first insertion, and faces the solvent-exposed exterior of the particle. As in the hAd2pb structures, even the shorter Ad12 hypervaria- ble loop is highly flexible, as demonstrated by relat- ively poor electron density and high temperature factors. In all monomers, residues 297–317 were disor- dered and not modeled. The second insertion into the jellyroll domain con- tains part of the putative fiber protein-binding site. This portion of the protein undergoes a conformation- al change upon fiber protein binding, with helix 7 kinking almost 45° to form a binding cleft for the N-terminus of the fiber protein [20]. Structural align- ments of the hAd2 ⁄ 12pb chimera with the hAd2pb fiber peptide bound and unbound structures reveal a high degree of overall conservation (Figs 3B and 4). Surprisingly, the hAd2 ⁄ 12pb structure most closely resembles the fiber peptide-bound form of hAd2pb; however, no fiber peptide was present during crystal- lization, and the fiber-binding site is empty. A possible explanation for the observed conforma- tion of helix 7 is the effect of the solvent environment, specifically high concentrations (50%) of MPD, favor- ing a conformation mimicking the fiber protein-bound state. HAd2pb and the hAd2pb ⁄ fiber peptide complex were both crystallized from ammonium sulfate ⁄ diox- ane solutions, not MPD solutions [20]. The use of high concentrations of a small organic alcohol will affect the solvent structure of the protein. Studies quantifying the effects of MPD on protein structure have demon- strated that the alcohol tends to bind in hydrophobic sites, particularly at leucine side chains [27]. The fiber Fig. 3. The hAd2 ⁄ 12 and wild-type hAd2 penton base monomers. (A) Ribbon diagram of the hAd2 ⁄ 12 penton base monomer is in blue (left) and rainbow (right). The N-terminus and C-terminus are labeled in both monomers. On the left, the b-strands of the jellyroll domain are labeled, with the BIDG sheets in yellow and the CHEF sheet in cyan. (B) Left, ribbon diagram of the hAd2 penton base (PDB code 1X9P) (green) with a portion of the hypervariable loop (red) and helix 7 (blue). Middle, ribbon diagram of the hAd2 ⁄ 12 chi- mera (green) and helix 7 (blue). Right, hAd2 fiber peptide complex (PDB code 1X9T) with helix 7 in blue and the fiber peptide in magenta. Fig. 4. Comparison of the hAd2, hAd2 ⁄ 12 and hAd2–fiber peptide complex structures. (A) Stereo overlay of hAd2 penton base (yel- low), hAd2 ⁄ 12 (blue) and hAd2 fiber peptide complex (pink). The hAd2 monomers are partially transparent, for clarity. The gray box outlines helix 7. (B) Stereo overlay of helix 7 with hAd2 penton base (yellow), hAd2 ⁄ 12 (blue), and hAd2 fiber peptide complex (pink). The fiber peptide is drawn as ball-and-stick and colored by atom. Human adenovirus 2 ⁄ 12 penton base chimera C. Zubieta et al. 4340 FEBS Journal 273 (2006) 4336–4345 ª 2006 The Authors Journal compilation ª 2006 FEBS protein-binding site consists primarily of hydrophobic residues, including Leu193. It is possible that these hydrophobic interactions lead to a conformational change of helix 7. Although the modest 3.6 A ˚ resolu- tion of the hAd2 ⁄ 12pb crystal structure precludes the location of any ordered MPD molecules within the fiber protein-binding site, this is a likely possibility. A second hypothesis for the observed conformation of helix 7 entails conformational coupling of the hyper- variable loop to the fiber protein-binding region. It has been shown that the fiber protein is shed from the virus upon integrin binding by the penton base hypervariable loop and most likely before endocytosis [13,17–19]. Upon addition of soluble RGD peptides that interfere with penton base binding to integrin, the virion exhib- ited impaired endocytosis. Interestingly, in this experi- ment, the fibers were not shed from the virus [19]. These data raise the possibility of coupling of fiber release to binding of the RGD motif of the hypervaria- ble loop to cellular integrins. Alterations in hypervaria- ble loop topology, in this case engendered by shortening the loop to the Ad12pb sequence, could have allosteric effects on the fiber-binding site. In order to investigate this possibility, we performed fluores- cence-based binding assays of the wild-type hAd2pb and the hAd2 ⁄ 12 chimera with a fiber protein mimetic to determine whether fiber binding or fiber release was coupled to changes in the hypervariable loop. Binding assays Owing to difficulties in expressing and purifying the full-length fiber protein, a 75 amino acid ‘minifiber’ construct was used as a fiber protein mimetic. Previous structural studies of hAd2pb with a 22 amino acid peptide from the N-terminus of the fiber protein have mapped the fiber protein interaction region to a hydro- phobic stretch of amino acids at the N-terminus of the fiber protein [20]. In order to emulate the oligomeriza- tion state of the fiber, a trimerization domain was fused to an N-terminal fiber protein construct (gift of G Nemerow). This construct consists of the first 44 residues of the hAd2 fiber coupled to a trimerizing foldon domain from T-4 bacteriophage fibritin [28] (Fig. 5A). The foldon domain has been shown to induce trimerization in adenovirus fiber protein con- structs [29] and was used in the minifiber to mimic the proper oligomerization state of the full-length protein. Proper trimerization of the minifiber construct was confirmed by gel filtration and native gel (data not shown). The minifiber was labeled with tetramethyl- rhodamine (TMR) at C-terminal cysteine residues and purified by size exclusion chromatography. Changes in fluorescence anisotropy were used to directly measure the binding of the TMR-labeled mi- nifiber to the penton base. The initial reaction between the minifiber and the penton base was assumed to be reversible with F þ PB $ P  where F is the minifiber, PB the penton base, and P* the initial fiber–penton base complex. Owing to the con- formational change of the penton base after fiber bind- ing, the full binding expression can be represented as F þ PB $ P  ! P with the formation of P being virtually irreversible. The apparent K d measured here is the K d for the initial bind- ing of minifiber. The fluorescence emission of the TMR was assumed to be independent of the binding of the minifiber to the penton base, because it is attached at the C-terminal extremity of the foldon domain and is remote from the penton base-binding region. The system was particularly amenable to this tech- nique, due to the relatively small size of the minifiber ( 24 kDa for the trimer) in conjunction with the large size of the penton base ( 300 kDa for the pentamer). K d values for minifiber binding to wild-type hAd2pb B A Fig. 5. Sequence alignment and binding data for the minifiber–pen- ton base complex. (A) Alignment of the N-terminal region of hAd2 fiber (accession number CAJ29207), hAd12 fiber (accession num- ber CAJ29196), and minifiber. The residues known to interact with the penton base are in yellow and the T4 fibritin foldon domain of the minifiber is in red. Derivatization with 2-methyl-2,4-pentanediol (TMR) was performed on the C-terminal cysteine residue of the minifiber. (B) Left: fluorescence anisotropy measurement for wild- type (wt) (red circles) and hAd2 ⁄ 12 (blue squares) with labeled minifiber at 33 n M for all measurements. Varying amounts of pen- ton base were titrated in and plotted against anisotropy. Right: competition assay for labeled versus unlabeled minifiber. The labe- led minifiber at 15 n M was preincubated with 3 lM penton base. Unlabeled minfiber was titrated into the solution, and the anisotro- py values were recorded. No dissociation of labeled minifiber was noted with the concentrations of minifiber used. C. Zubieta et al. Human adenovirus 2 ⁄ 12 penton base chimera FEBS Journal 273 (2006) 4336–4345 ª 2006 The Authors Journal compilation ª 2006 FEBS 4341 and chimeric hAd2 ⁄ 12pb were 1.0 ± 0.3 lm and 1.85 ± 0.7 lm, respectively (Fig. 5B). During viral assembly, concentrations of the fiber protein within the nucleus of an infected cell are likely to be at least in the low micromolar range, based upon the concentra- tion of virions [30] and excess fiber protein [31] pro- duced during adenoviral infection; thus, the K d values measured are physiologically reasonable. In order t o investigate the r ole o f the pr oposed con- formational change that occurs upon fiber binding, we attempted to disso ciate the bound minifiber from the penton base . Previous studies have s hown that in vitro removal of the fiber protein from adenovir us requires the use of chaotropic salts and increased temperatures [32,33], and no release of fiber protein has been noted even after long-term storage of dilute solutions of virus [34]. To address these observations, we performed competi- tion experiments using unlabeled minifiber. Unlabeled minifiber was titrated into samples of TMR-labeled minifiber bound to wild-type and chimeric penton base. No decrease in fluorescence anisotropy was observed even upon addition of 50-fold excess of un- labeled minifiber over the calculated K d (Fig. 5B). These data support the hypothesis that after fiber binding, the conformational change occurring in the penton base locks the fiber protein into place. Thus, as noted previously, a full representation of the binding equilibrium of the interaction is F þ PB $ P  ! P where F is the fiber or minifiber protein, PB is the pen- ton base, P* is the initial penton, and P is the penton after conformational change. The stability of the com- plex results from the effect of the crystallographically observed cooperative conformational change in the fiber protein-binding site. Based on the structure of hAd2pb with an N-terminal fiber peptide, conforma- tional changes will occur in the penton base upon interaction of the fiber protein, essentially locking in the fiber [20]. Studies with adenovirus 2 have shown that fiber loss occurs at the cell surface and prior to endocytosis [19]. Although the mechanism of fiber dissociation from the penton is not clear, the extreme stability of the penton once formed and the relatively weak micromolar affin- ity for initial complex formation has important impli- cations for adenovirus infectivity. For successful infection, adenovirus needs the fiber protein as a pri- mary points of attachment to the host cell. Studies with fiberless adenoviral particles demonstrated a sev- eral thousand-fold decrease in infectivity [35]. Thus, premature loss of the fiber would result in marked decreases in infection. Once the virus is attached to the cell surface, however, the fiber protein is not neces- sary for the later steps in the viral cycle. After fiber loss, the low micromolar affinity of the penton base for fiber protein minimizes reattachment of the fiber to the viral particle. The altered conformation of the fiber-binding site in the hAd2 ⁄ 12 chimeric structure led us to investigate the possibility of coupling between the conformation of the hypervariable loop, which contains the integrin-binding site, and the fiber- binding site of the penton base. Based on the fluores- cence anisotropy data presented here for the wild-type and chimeric protein showing similar binding profiles with relatively weak (1–2 lm) initial affinity of the penton base for the fiber and virtually no release of fiber once bound, adenovirus modulates fiber attach- ment by a two-step fiber-binding process that is inde- pendent of the conformation of the hypervariable loop. The mechanism for fiber release is not coupled to the hypervariable loop size or conformation but may be influenced by the solvent environment of the virion. In this structural study, the fiber-binding site was in a ‘bound’ conformation, although no fiber was present. The most likely explanation for this observa- tion is that the switch from the fiber ‘bound’ to ‘unbound’ states of the penton base (or vice versa) is not only dependent on the presence or absence of the fiber itself but can also be triggered by solvent envi- ronment effects. Owing to the cooperativity of this switch [19], all penton bases will be in the same con- formation. Triggering the switch to the unbound state would clearly favor fiber release. Our results suggest that this is not dependent on the hypervariable loop directly, but could be due to other interactions with the bound integrin or the particular solvent environ- ment of the virion at the cell surface or in the initial stages of endocytosis. Experimental procedures For baculovirus expression, the SF21 and Hi5 cell lines and the Bac-to-Bac expression system and vectors were from Invitrogen (Carlsbad, CA, USA). Protease inhibitors were from Roche (Basel, Switzerland) and Ni-NTA resin was obtained from Bio-Rad (Hercules, CA, USA). The minifiber construct was a gift from the Laboratory of Glen Nemerow at The Scripps Research Institute (TSRI) in La Jolla, CA, USA. An MOS450 fluorimeter (Biologic, SA, Claix, France) was used in the fluorescence anisotropy assays. Protein expression and purification cDNA encoding a 49-residue N-terminal truncation of hAd2pb was cloned into a pFastbac vector as described Human adenovirus 2 ⁄ 12 penton base chimera C. Zubieta et al. 4342 FEBS Journal 273 (2006) 4336–4345 ª 2006 The Authors Journal compilation ª 2006 FEBS previously [20]. The hAd2 ⁄ 12pb chimera was constructed by annealing complementary overhangs from two fragments of the hAd2 gene. The forward oligomer 5¢-TATTT TCAGGGCGCCATGGGATCCCCCTTCGATGCTCCC-3¢ and the reverse oligomer 5¢-TGGTTTCGGAGCGGCC GCATTATCGCCCCTCC CGCCCTGTTCG G-3¢ were used to introduce a Not1 site C-terminal to the RGD motif of hAd2 in addition to an N-terminal Nco1 cloning site. The forward o ligomer 5 ¢-TCCGAAACCAGCG GCCGCTT TATCGCGTTA AAAC CGGT GATCAA ACCCC -3¢ and the reverse oligomer 5¢-GTAGGCCTT TGAATTCCTCAAAA AGTGCGGCTCGAT-3¢ were used to introduce a Not1 site followed by the hAd12pb hypervariable loop sequence into hAd2 in addition to a C-terminal EcoR1 site. The resulting fragments were digested with Not1, Nco1orEcoR1, puri- fied, and annealed. The Not1 site in the resulting gene was removed using the complementary oligomers 5¢-GGCGGGAGGGGCGATAATTTTATCGCGTTAAAA CCG-3¢ (forward) and 5¢-CGGTTTTAACGCGATAAAAT TATCGCCCCTCCCGCC-3¢ (reverse). Virus amplification was performed in monolayer SF21 cells, and protein expression was performed in HighFive cells in shaker flasks (135 r.p.m., 27 °C, 3 days). For protein expression, a multiplicity of infection of 5–10 was used. After 3 days of expression, cells were pelleted at 195 g for 5 min at 4 °C using a Jouan CR3 centrifuge with a T-40 rotor, and stored at ) 80 °C. For protein purification, cells were resus- pended in 25 mm Tris (pH 7.5) ⁄ 100 mm NaCl, plus protease inhibitors (Roche) and lysed by sonication. Cell debris was pelleted at 43 000 g for 45 min at 4 °C using an Avanti J-25 centrifuge with a JA25.50 rotor (Beckman Coulter, Fuller- ton, CA, USA), and the supernatant collected. The protein was precipitated with 30% ammonium sulfate and the precipitant collected. After resuspension in 25 mm Tris (pH 7.5) ⁄ 100 mm NaCl, the protein was dialyzed overnight against the same buffer. The protein solution was concentra- ted to 5mgÆmL )1 and applied to a MONO Q column (Pharmacia, Uppsala, Sweden) with a linear gradient of 100 mm to 1 m NaCl in 25 mm Tris (pH 7.5). The protein eluted at approximately 220 mm NaCl. Fractions of interest were buffer exchanged, concentrated to 5–10 mgÆmL )1 and stored at ) 80 °C. The minifiber construct was cloned into a pEtM11 expres- sion vector between the Kpn1 and Nco1 sites using a forward oligomer 5¢-CTTTATTTTCAGGGCGCCATGAAGCGCG CAAGACCGTCTGAA-3¢ and a reverse oligomer 5¢-AGCT CGAATTCG GATCCGGTACCTCAGAAGGTAGACAG CAGAACC-3¢. For derivitization with TMR, a Gly-Gly-Cys sequence was introduced at the C-terminus using oligomers 5¢-CTGCTGTCTACCTTTGGAGGTTGCTGATCCGAA TTCGAG-3¢ (forward) and 5¢-GCTCGAATTCGGATCAG CAACCTCCAAAGGTAGACAGCA-3¢ (reverse). BL21 cells were transformed with the pETM11 construct and grown until a D of 0.8 (600 nm) in Terrific Broth supple- mented with phosphate b uffer and kanamycin at 37 °C. The temperature was reduced to 27 °C and the culture induced with 0.5 mm isopropyl-b-d-thiogalactopyrano- side. Approximately 6 h postinduction, the cells were harvested at 5800 g for 15 min using an Avanti J-25 cen- trifuge with a JA-10 rotor. Cells were resuspended in 25 mm Tris (pH 7 .5) ⁄ 100 mm NaCl ⁄ 10 mm b-mercapto- ethanol ⁄ 1 ·protease inhibitors. Cells were lysed by soni- cation and cell debris pelleted at 47 000 g using an Avanti J-25 centrifuge with a J A25.50 rotor. The super- natant was collected and applied to a 2 mL Ni-NTA column. The column was washed with lysis buffer con- taining 10 mm imidazole and the protein eluted with lysis buffer containing 200 mm imidazole. The His tag was cleaved overnight at 4 °C using Tev protease, and the His-tag and Tev were depleted by running the solu- tion over the same Ni-NTA column. The protein was concentrated and applied to an S200 size exclusion column using 25 mm Tris (pH 7.5) ⁄ 100 mm NaCl ⁄ 10 mm Tris(2-carboxyethyl)phosphine hydrochloride. Protein frac- tions of interest were then concentrated and stored at ) 80 °C. TMR labeling and binding assays A five-fold excess of malemide-derivatized TMR was added to a solution of purified minifiber and allowed to react for  2 h. Excess TMR was removed by dialysis against 25 mm Tris ⁄ 100 mm NaCl, and the protein was concentrated. The labeled protein was then applied to an S75 column for fur- ther purification. A standard protocol for fluorescence anisotropy measure- ments is as follows. Upon excitation at 546 nm, the fluores- cence anisotropy of a 33 nm solution of labeled minifiber was measured at an emission wavelength of 575 nm perpen- dicular to the excitation vector. Either wild-type hAd2pb or hAd2 ⁄ 12pb was titrated into the solution in  200–300 lm increments. Anisotropy values were recorded for 60 points over 60 s, and an average value was taken as a data point. Titrations were continued until a stable anisotropy value was obtained. Fluorescence anisotropy, A, is defined as A ¼ðI V  I H Þ=ðI V þ 2I H Þ where I V and I H refer to the parallel and perpendicular components of the polarized fluorescence emission. The changes in anisotropy are a linear function with A ¼ A F þðA B  A F Þð½Lig B =½Lig tot Þ where A is the measured anisotropy value, A F is the aniso- tropy of the free minifiber, A B is the anisotropy of the mini- fiber bound to the penton base, [Lig B ] is the concentration of the bound ligand (minifiber), and [Lig tot ] is the total con- centration of the ligand. At any total concentration of lig- and, the anisotropy depends on the total concentration of C. Zubieta et al. Human adenovirus 2 ⁄ 12 penton base chimera FEBS Journal 273 (2006) 4336–4345 ª 2006 The Authors Journal compilation ª 2006 FEBS 4343 penton base and the dissociation equilibrium constant for the complex, K d , as shown by A ¼ A F þ DAðfa ða 2 ð4½PB tot ½Lig tot ÞÞ 1=2 g=2½Lig tot Þ where a ¼ ([PB] tot + [Lig] tot + K d ), DA ¼ (A B ) A F ), and [PB] is the concentration of the penton base. Fitting the anisotropy data to the previous equation gives a value for the K d. Binding curves were generated in kaleidagraph and fit- ted to the above equation. All measurements were done in triplicate with at least eight data points per experiment. Crystallization Crystals were grown in hanging drops against 50% MPD ⁄ 0.2 m ammonium phosphate⁄ 0.1 m Tris (pH 7.5) at 4 °C. A 1 lL drop of the protein solution at a concentra- tion of 5–10 mgÆmL )1 was added to 1 lL of well solution. Small crystals grew within 2–3 months, reaching dimensions of 50 · 50 · 50 lm. Microseeding of equilibrated drops yielded bigger crystals of dimensions 100 · 100 · 100 lm after a few weeks. Because the high MPD concentration acted as a cryoprotectant, crystals were taken directly from the drop and flash frozen prior to data collection. Data collection and structure determination Crystallographic data for hAd2 ⁄ 12pb were collected on the microfocus ESRF beamline ID13 with the EMBL microdif- fractometer (Maatel, Voreppe, France) [36]. Owing to the large unit cell, an oscillation range of 0.5° was used during data collection. Radiation damage was readily apparent after a few exposures, which limited the amount of usable data per exposed position to 2° (four 0.5° frames). Owing to the small beam size used (10 lm), the crystal was reposi- tioned after each 2°, permitting data collection from an undamaged volume of crystal. Two large crystals (100 · 100 · 100 lm 3 ) provided 104 frames (52°) and 128 frames (64°) of usable data to a maximum resolution of  3.5 A ˚ . A direct beam was taken in order to have a highly precise value for the beam center. The beam center was found and refined in mosflm [37]. All data were integrated and scaled with xds. molrep [38] was used to find a molecular replacement solution with the hAd2 pentamer as a search model (pdb code 1X9P). Based on the unit cell dimensions and Mat- thews coefficient, three pentamers were expected per asym- metric unit, with the full dodecahedral particle formed from the space group symmetry operators. A clear solution with three pentamers was found and used to generate initial maps. Subsequent 15-fold NCS averaging and gradual phase extension from 15 A ˚ to 3.6 A ˚ with rave [39] allowed the generation of high-quality electron density maps. Model building was performed in o [40] and refinements carried out with refmac. Initially, strict NCS restraints were applied to the model; however, in the later stages of refine- ment, tight NCS constraints were only applied to the back- bone, and the side chains were refined using medium restraints. Unaveraged maps were examined to locate devia- tions between monomers. According to procheck [41] (Table 1), only three (Thr66, Glu172, and His237) residues were in disallowed regions of the Ramachandran plot. These generally lie in poorly ordered loop regions. Residues 507–508 (C-terminus) and residues 297–317 of the hypervariable loop were disordered and not modeled. All figures were generated using pymol [42]. Coordinates and structure factors have been deposited in the RCSB Protein Data Bank under accession code 2C6S. References 1 Hierholzer JC (1992) Adenoviruses in the immunocom- promised host. Clin Microbiol Rev 5, 262–274. 2 Kojaoghlanian T, Flomenberg P & Horwitz MS (2003) The impact of adenovirus infection on the immunocom- promised host. Rev Med Virol 13, 155–171. 3 McConnell MJ & Imperiale MJ (2004) Biology of adenovirus and its use as a vector for gene therapy. Hum Gene Ther 15, 1022–1033. 4 Danthinne X & Imperiale MJ (2000) Production of first generation adenovirus vectors: a review. Gene Ther 7, 1707–1714. 5 Amalfitano A & Parks RJ (2002) Separating fact from fiction: assessing the potential of modified adenovirus vectors for use in human gene therapy. Curr Gene Ther 2, 111–133. 6 Furcinitti PS, van Oostrum J & Burnett RM (1989) Adenovirus polypeptide IX revealed as capsid cement by difference images from electron microscopy and crys- tallography. EMBO J 8, 3563–3570. 7 Stewart PL, Burnett RM, Cyrklaff M & Fuller SD (1991) Image reconstruction reveals the complex molecular organization of adenovirus. Cell 67, 145–154. 8 Stewart PL, Fuller SD & Burnett RM (1993) Difference imaging of adenovirus: bridging the resolution gap between X-ray crystallography and electron microscopy. EMBO J 12, 2589–2599. 9 Stewart PL & Burnett RM (1995) Adenovirus structure by X-ray crystallography and electron microscopy. Curr Top Microbiol Immunol 199, 25–38. 10 Fabry CM, Rosa-Calatrava M, Conway JF, Zubieta C, Cusack S, Ruigrok RW & Schoehn G (2005) A quasi- atomic model of human adenovirus type 5 capsid. EMBO J 24, 1645–1654. 11 Saban SD, Nepomuceno RR, Gritton LD, Nemerow GR & Stewart PL (2005) CryoEM structure at 9A reso- lution of an adenovirus vector targeted to hematopoietic cells. J Mol Biol 349, 526–537. 12 Cohen CJ, Shieh JT, Pickles RJ, Okegawa T, Hsieh JT & Bergelson JM (2001) The coxsackievirus and adeno- Human adenovirus 2 ⁄ 12 penton base chimera C. Zubieta et al. 4344 FEBS Journal 273 (2006) 4336–4345 ª 2006 The Authors Journal compilation ª 2006 FEBS virus receptor is a transmembrane component of the tight junction. Proc Natl Acad Sci USA 98, 15191–15196. 13 Wickham TJ, Mathias P, Cheresh DA & Nemerow GR (1993) Integrins alpha v beta 3 and alpha v beta 5 pro- mote adenovirus internalization but not virus attach- ment. Cell 73, 309–319. 14 Li E, Stupack D, Bokoch GM & Nemerow GR (1998) Adenovirus endocytosis requires actin cytoskeleton reor- ganization mediated by Rho family GTPases. J Virol 72, 8806–8812. 15 Li E, Brown SL, Stupack DG, Puente XS, Cheresh DA & Nemerow GR (2001) Integrin alpha(v)beta1 is an adenovirus coreceptor. J Virol 75, 5405–5409. 16 Chiu CY, Mathias P, Nemerow GR & Stewart PL (1999) Structure of adenovirus complexed with its inter- nalization receptor, alphavbeta5 integrin. J Virol 73, 6759–6768. 17 Greber UF, Willetts M, Webster P & Helenius A (1993) Stepwise dismantling of adenovirus 2 during entry into cells. Cell 75, 477–486. 18 Medina-Kauwe LK (2003) Endocytosis of adenovirus and adenovirus capsid proteins. Adv Drug Deliv Rev 55, 1485–1496. 19 Nakano MY, Boucke K, Suomalainen M, Stidwill RP & Greber UF (2000) The first step of adenovirus type 2 disassembly occurs at the cell surface, independently of endocytosis and escape to the cytosol. J Virol 74, 7085– 7095. 20 Zubieta C, Schoehn G, Chroboczek J & Cusack S (2005) The structure of the human adenovirus 2 penton. Mol Cell 17, 121–135. 21 Stewart PL, Chiu CY, Huang S, Muir T, Zhao Y, Chait B, Mathias P & Nemerow GR (1997) Cryo-EM visuali- zation of an exposed RGD epitope on adenovirus that escapes antibody neutralization. EMBO J 16, 1189–1198. 22 Fuschiotti P, Schoehn G, Fender P, Fabry CM, Hewat EA, Chroboczek J, Ruigrok RW & Conway JF (2006) Structure of the dodecahedral penton particle from human adenovirus type 3. J Mol Biol 356, 510–520. 23 Chen XS, Garcea RL, Goldberg I, Casini G & Harrison SC (2000) Structure of small virus-like particles assembled from the L1 protein of human papillomavirus 16. Mol Cell 5, 557–567. 24 Chroboczek J, Gout E, Favier AL & Galinier R (2003) Novel partner proteins of adenovirus penton. Curr Top Microbiol Immunol 272, 37–55. 25 Galinier R, Gout E, Lortat-Jacob H, Wood J & Chro- boczek J (2002) Adenovirus protein involved in virus internalization recruits ubiquitin-protein ligases. Biochemistry 41, 14299–14305. 26 Chelvanayagam G, Heringa J & Argos P (1992) Anat- omy and evolution of proteins displaying the viral cap- sid jellyroll topology. J Mol Biol 228, 220–242. 27 Anand K, Pal D & Hilgenfeld R (2002) An overview on 2-methyl-2,4-pentanediol in crystallization and in crys- tals of biological macromolecules. Acta Crystallogr D Biol Crystallogr 58, 1722–1728. 28 Guthe S, Kapinos L, Moglich A, Meier S, Grzesiek S & Kiefhaber T (2004) Very fast folding and association of a trimerization domain from bacteriophage T4 fibritin. J Mol Biol 337, 905–915. 29 Papanikolopoulou K, Forge V, Goeltz P & Mitraki A (2004) Formation of highly stable chimeric trimers by fusion of an adenovirus fiber shaft fragment with the foldon domain of bacteriophage t4 fibritin. J Biol Chem 279, 8991–8998. 30 Green M & Pina M (1963) Biochemical studies on ade- novirus multiplication. IV. Isolation, purification, and chemical analysis of adenovirus. Virology 20, 199–207. 31 Pettersson U, Philipson L & Hoglund S (1968) Struc- tural proteins of adenoviruses. II. Purification and char- acterization of the adenovirus type 2 fiber antigen. Virology 35, 204–215. 32 Boudin ML, Moncany M, D’Halluin JC & Boulanger PA (1979) Isolation and characterization of adenovirus type 2 vertex capsomer (penton base). Virology 92, 125–138. 33 Pettersson U & Hoglund S (1969) Sructural proteins of adenoviruses. 3. Purification and characterization of the adenovirus type 2 penton antigen. Virology 39, 90–106. 34 Boudin ML & Boulanger P (1981) Antibody-triggered dissociation of adenovirus penton capsomer. Virology 113, 781–786. 35 Von Seggern DJ, Chiu CY, Fleck SK, Stewart PL & Nemerow GR (1999) A helper-independent adenovirus vector with E1, E3, and fiber deleted: structure and infectivity of fiberless particles. J Virol 73, 1601–1608. 36 Perrakis A, Cipriani F, Castagna JC, Claustre L, Burg- hammer M, Riekel C & Cusack S (1999) Protein micro- crystals and the design of a microdiffractometer: current experience and plans at EMBL and ESRF ⁄ ID13. Acta Crystallogr D Biol Crystallogr 55, 1765–1770. 37 Leslie AGW (1992) Recent changes to the MOSFLM package for processing film and image plate data. Joint CCP4 + ESF-EAMCB Newsl Protein Crystallogr 26. 38 Vagin A & Teplyakov A (2000) An approach to multi- copy search in molecular replacement. Acta Crystallogr D Biol Crystallogr 56, 1622–1624. 39 Kleywegt GJ & Jones TA (1999) Software for handling macromolecular envelopes. Acta Crystallogr D Biol Crystallogr 55, 941–944. 40 Jones TA, Zou JY, Cowan SW & Kjeldgaard M (1991) Improved methods for building protein models in elec- tron density maps and the location of errors in these models. Acta Crystallogr A 47, 110–119. 41 Laskowski RA, MacArthur MW, Moss DS & Thornton JM (1993) PROCHECK: a program to check the stereo- chemical quality of protein structures. J Appl Crystal- logr 26, 283–291. 42 DeLano WL (2002) The PyMOL Molecular Graphics System. DeLano Scientific, San Carlos, CA. C. Zubieta et al. Human adenovirus 2 ⁄ 12 penton base chimera FEBS Journal 273 (2006) 4336–4345 ª 2006 The Authors Journal compilation ª 2006 FEBS 4345 . Structural and biochemical characterization of a human adenovirus 2/12 penton base chimera Chloe Zubieta 1, *, Laurent Blanchoin 2 and Stephen Cusack 1 1. forward o ligomer 5 ¢-TCCGAAACCAGCG GCCGCTT TATCGCGTTA AAAC CGGT GATCAA ACCCC -3¢ and the reverse oligomer 5¢-GTAGGCCTT TGAATTCCTCAAAA AGTGCGGCTCGAT-3¢