Antigenic Characteristics of Rhinovirus Chimeras Designed ...€¦ · Antigenic Characteristics of Rhinovirus Chimeras Designed in silico for En5hanced Presentation of HIV-1 gp41

doi:10.1016/j.jmb.2010.01.064 J. Mol. Biol. (2010) 397, 752–766

Available online at www.sciencedirect.com

Antigenic Characteristics of Rhinovirus ChimerasDesigned in silico for En5hanced Presentation of HIV-1gp41 Epitopes

Mauro Lapelosa1,2,3, Gail Ferstandig Arnold2,3⁎, Emilio Gallicchio1,2,Eddy Arnold2,3 and Ronald M. Levy1,2

1BioMaPS Institute forQuantitative Biology, RutgersUniversity, Piscataway, NJ08854, USA2Department of Chemistry andChemical Biology, RutgersUniversity, Piscataway, NJ08854, USA3Center for AdvancedBiotechnology and Medicine,Rutgers University, Piscataway,NJ 08854, USA

Received 27 October 2009;received in revised form26 January 2010;accepted 27 January 2010Available online4 February 2010

*Corresponding author. DepartmenChemical Biology, Rutgers Universi08854, USA. E-mail address: gfarnoAbbreviations used: cHRV, chime

antigen-binding fragment;HIV, humvirus; HRV, human rhinovirus; mAbantibody; MPER, membrane-proximNIm-II, neutralizing immunogenic stemperature replica exchange molec

0022-2836/$ - see front matter © 2010 E

The development of an effective AIDS vaccine remains the most promisinglong-term strategy to combat human immunodeficiency virus (HIV)/AIDS.Here, we report favorable antigenic characteristics of vaccine candidatesisolated from a combinatorial library of human rhinoviruses displaying theELDKWA epitope of the gp41 glycoprotein of HIV-1. The design principlesof this library emerged from the application of molecular modelingcalculations in conjunction with our knowledge of previously obtainedELDKWA-displaying chimeras, including knowledge of a chimera with oneof the best 2F5-binding characteristics obtained to date. The molecularmodeling calculations identified the energetic and structural factorsaffecting the ability of the epitope to assume conformations capable offitting into the complementarity determining region of the ELDKWA-binding, broadly neutralizing human mAb 2F5. Individual viruses wereisolated from the library following competitive immunoselection and weretested using ELISA and fluorescence quenching experiments. Dissociationconstants obtained using both techniques revealed that some of the newlyisolated chimeras bind 2F5 with greater affinity than previously identifiedchimeric rhinoviruses.Molecular dynamics simulations of two of these samechimeras confirmed that their HIV inserts were partially preorganized forbinding, which is largely responsible for their corresponding gains inbinding affinity. The study illustrates the utility of combining structure-based experiments with computational modeling approaches for improvingthe odds of selecting vaccine component designs with preferred antigeniccharacteristics. The results obtained also confirm the flexibility of HRV as apresentation vehicle for HIV epitopes and the potential of this platform forthe development of vaccine components against AIDS.

© 2010 Elsevier Ltd. All rights reserved.

Keywords: vaccine design; monoclonal antibody 2F5; chimeric virus;reorganization free energy; protein binding
Edited by I. Wilson
t of Chemistry andty, Piscataway, [email protected] rhinovirus; Fab,an immunodeficiency, monoclonalal external region;ite II; REMD,ular dynamics.

lsevier Ltd. All rights reserve

Introduction

An effective acquired AIDS vaccine would be acrucial component of any long-term strategy tocombat human immunodeficiency virus (HIV)/AIDS. However, designing immunogens that canelicit effective, broadly neutralizing responses againstHIV has been a surprisingly difficult task. Thishighlights the need for novel alternative approachesaimed at the discovery of effective AIDS vaccines.1

Some of the most promising target epitopes of HIVare located in the membrane-proximal externalregion (MPER) of the transmembrane componentof the envelope glycoprotein gp41 of the virus. The

d.

mailto:[email protected]

http://dx.doi.org/10.1016/j.jmb.2010.01.064

753Antigenic Characteristics of Rhinovirus Chimeras

MPER plays an important role in the process of HIVfusion to the host cell membrane,2,3 as well as inpermittingCD4-independent transcytosis of the virusacross epithelial cells at mucosal surfaces.2 Thesefunctions likely explain sequence conservation of thisregion and the efficacy of antibodies directed againstit.3 Potent responses against theMPER are associatedwith stronger and broader neutralizing capabilities ininfected individuals,4 illustrating the value of includ-ing an MPER immunogen in an AIDS vaccine orvaccine cocktail.A set of epitopes has been identified in the MPER

region. These are the ELDKWAepitope (HIV-1HxB2gp41 residues 662–667), recognized by the particu-larly broadly neutralizing human mAb 2F55 and theadjacent residues that bind the broadly neutralizinghuman mAb 4E106 epitope (HxB2 residues 672–679)as well as the phage display-derivedmAb Z13e1 (theIgG version of an affinity-enhanced antigen-bindingfragment (Fab) of Z13).7 More recently, two newIgMs, WR316 and WR320, have been identified thatbind to this region (HxB2 residues 668–673 and 661–679, respectively) and show some degree of neutral-izing activity as well.8

In this study, we focused specifically on theELDKWA epitope. It has been reported that infectedindividuals producing neutralizing antibodies di-rected against the ELDKWA epitope exhibit healthbenefits.9,10 While none of the immunogen-inducedimmune responses generated against this region hasbeen both broadly reactive11 and potent thus far, webelieve that presentations in which the MPERepitopes are preorganized for binding should pro-duce valuable immunogens that can contribute to asuccessful vaccine.A promising approach for epitope presentation

consists of grafting gp41 MPER epitopes of HIV-1onto the surface of the safe and highly immunogenichuman rhinovirus (HRV),12 a picornavirus thatcauses approximately 50% of common colds.4 HRVis likely to be particularly favorable as a vaccinevehicle due to its ability to stimulate potent humoralimmune responses, including mucosal immune res-ponses13 as well as T-cell help responses.14,15 Wehave shown that HRV can accommodate a variety offoreign sequences9,10,12 in a surface loop of the viralcoat protein 2 of HRV, designated the VP2 puff.11,14

This surface loop is, in fact, part of one of HRV'sown immunogenic sites, constituting the largest ofthree loops forming the neutralizing immunogenicsite II (NIm-II15). We demonstrated recently thatELDKWA-derived epitopes can be displayed onHRV in ways that stimulate neutralizing immuneresponses directed against diverse pseudoisolates ofHIV-1.13 These recombinant immunogens have beenamong the few reported that elicit broad, albeitmodest, neutralization of HIV, and we are activelypursuing their further development with the goal ofproviding protective AIDS vaccine components.In an earlier study,16 we used computational

modeling techniques to study the effect of presen-tation modality on the conformational propensity ofthe ELDKWA epitope on the surface of HRV. We

hypothesized that those presentation constructswith the highest fraction of epitope conformationscompatible with antibody complexation wouldpresent the highest binding affinity for 2F5, makingthem the most antigenic.The crystallographic structure of an ELDKWA-

based peptide complexed with the Fab fragment ofthe 2F5 monoclonal antibody (mAb)17 provided thestructural target against which we compared theensemble of conformations generated by modeling.These studies indicated that the length, the hydro-phobic character, and the precise site of insertion ofthe epitope are crucial for achieving the greateststructural similarity to the target structure.16In the present study we have used these insights to

formulate a new combinatorial library of chimericrhinoviruses (cHRVs) in an effort to isolate con-structs with improved binding affinity with respectto the 2F5 mAb. Immunoselection of the library witha combination of mAb 2F5 and competitiveELDKWA-based peptides resulted in the isolationof specific cHRVs displaying greater affinity for 2F5than previously designed constructs with respect toboth ELISA assays and direct spectroscopic deter-mination of the equilibrium constants for antibodybinding. Computer modeling calculations showedthat the increased binding affinity of these con-structs for 2F5 is correlated directly with their morefavorable binding reorganization free energies, ameasure of their ability to assume conformationscompatible with antibody complexation based onthe structural criteria we developed earlier.16

The results obtained confirm the flexibility of HRVas a presentation vehicle for HIV epitopes and thepotential of this platform for the development ofvaccine components against HIV. In the next sectionwe present the results concerning the design,production, and immunoselection of a combinatori-al library of cHRVs designed according to thestructural insights we presented earlier.16 This isfollowed by the results of ELISA assays used togauge the affinity of specific isolates for 2F5 and theresults of more rigorous binding constant determi-nation using spectroscopic measurements for themost promising cHRVs obtained including, forcomparison, an ELDKWA-based cHRV character-ized previously without the use of computationalmodeling.13 The results of computer simulations ofthe cHRVs are described and discussed, illustratingthe importance of the binding reorganization freeenergy in determining the antigenic characteristicsof these constructs.

Results

Design, production, and immunoselection of anew combinatorial library

A focused combinatorial library was generated onthe basis of cHRV3, a chimeric HRV predicted bycomputer modeling to have a favorable affinity for2F5.16 The library encoded the ELDKWAS core

Table 1. Chimeric libraries investigated in this work

Virus HRV N-linker 2F5 epitope C-Linker HRV

14-C40-1a DLS PCG ALDKWAS SPDCS VGGPcHRV3b DLSSAN — ELDKWAS — EVGGPLibraryc DLSS XXd ELDKWAS XX GGP

Individual immunoselected viruses from libraryB1 DLSS HG ELDKWAS PN GGPC1 DLSS PG ELDKWAS IP GGPD1 DLSS SP ELDKWAS LP GGPF1 DLSS PP ELDKWAS SP GGPB2 DLSS GK ELDKWAS QP GGP

a One of the previous best binders to 2F5.b The modeled chimera predicted to bind 2F5 with greater affinity than 14-C40-1.c The library composition (this work).d X, any of the 20 amino acids.

Fig. 1. Competitive ELISA titers illustrating the com-petition of a 14-mer ELDKWA-containing peptide and anumber of chimeric viruses for binding immobilized mAb2F5. OD is the optical density (absorbance) and [peptide]is the concentration of the competitor peptide. Error barsrepresent the standard errors of the mean.

754 Antigenic Characteristics of Rhinovirus Chimeras

sequence flanked on either side by amino acidsrandomized to encode any of the 20 commonlyoccurring amino acids, enhancing the chances ofidentifying chimeras with favorable growth charac-teristics and well presented ELDKWA inserts. Theentire cassette (Table 1)was insertedbetween residuesS158 and G163 of the VP2 protein of HRV14. Theselection of the genetic modification site was madeaccording to previous studies directed to displayforeign sequences on loop 2 of the neutralizingimmunogenic site II of HRV14.9 A library of plasmidswas obtained from the pST-LIC vector and a library ofoligonucleotides. Subsequently, in vitro transcriptionallowed us to produce infectious RNAs. A pool ofchimeric viruses was produced by transfection intoH1-HeLa cells, and the transfection efficiency(2.4×105 plaque-forming units/μg) was determinedby counting the number of plaques obtained aftertransfection. The chimeric pool was purified so thatselected viruses might be more easily purified at laterstages. Viral RNAs from three recombinant virusesselected at random were sequenced, confirming thepresence of ELDKWAS sequences (data not shown).No HRV residue N-terminal or C-terminal to theELDKWAS core were either mutated or deleted.Chimeric viruses obtained from the library were

propagated, purified, and immunoselected on thebasis of their ability to bind to immobilized mAb 2F5in the presence of 0–16 pmol of competing peptide(Ac-EQELLELDKWASSLW-NH2; described inMate-rials andMethods). The chimeric viruses thus selectedgrew at rates comparable to that of wild-type HRV14,indicating that the presence of the inserts was notdeleterious to their growth. We observed a non-random preponderance of proline and glycine resi-dues among the flanking residues on both sides of theinsert (Table 1). This might be related to structuralfeatures of the engineered loop that affect the stabilityor acceptable folding of the chimeric viruses that arecompatible with the ability to bind 2F5.

Assessment of ability of cHRVs to bind mAb 2F5

ELISA experiments

To assess the ability of individual, immunose-lected viruses to recognize 2F5, we used indirect

competitive ELISAs (detailed in Materials andMethods) and tested the ability of the cHRVs tocompete with a 14-mer ELDKWA-containing pep-tide capable of binding to immobilized 2F5. The B–Dand F virus pools were isolated after immunoselec-tion using 16 pmol (B), 8 pmol (C), 4 pmol (D), and2 pmol (F) of peptide per well. Viruses, B1, C1, D1,F1, and B2 (Table 1) have shown peptide concen-tration-dependent binding to 2F5, unlike the HRV14control (Fig. 1), demonstrating that these viruseswere able to compete with the 14-mer ELDKWApeptide for binding to 2F5. Indeed, the ELISA results(Fig. 1) indicate that the majority of these viruseshave significantly better affinity for 2F5 at allconcentrations of peptide tested than the 14-C40-1chimeric virus, one of the most promising vaccineconstructs of this kind characterized to date. B1,selected from the most competitive immunoselec-tion conditions, showed the best binding to 2F5 inthe competitive ELISA of all of the individualchimeric viruses tested. Thus, B1 effectively com-peted with the 14-mer peptide in preferably binding2F5. C1 and B2 also showed high 2F5 bindingaffinity compared to the other chimeric virusestested, which also follows the pattern that thegreatest competition with peptide yielded the bestbinders to 2F5.

Fig. 2. Fluorescence intensity curve for the B1 con-struct, determined as a function of the 2F5 mAb concen-tration, used to calculate the equilibrium dissociationconstants, Kd (as described inMaterials andMethods). Thenon-linear fitting curve is shown in red: Q=(I – I0)/(I∞ –I0). The corresponding curves for the C1 and 14-C40-1constructs are similar in shape but shifted to higherconcentrations of antibody, reflecting their lower affinitiescompared to B1. The goodness of fit is measured as χ2 perdegree of freedom (B1, 1.01; C1, 0.98; 14-C40-1, 0.8).


The better binding ability of chimeras derivedfrom the combinatorial library relative to 14-C40-1 isconsistent with our earlier computational study inwhich we compared the conformational propensi-ties of the 14-C40-1 insert to that of an insert similarin size and electrostatic character to that of B1.16 Weobserved that B1 had greater solvent exposure andfit better in a docking model with the 2F5 Fab(indicating it has a higher propensity to adopt aconformation competent for binding to 2F5 thandoes the 14-C40-1 chimera). In contrast, the 14-C40-1chimera was found to be likely to form intramolec-ular hydrophobic interactions, constraining theexposure of the epitope on the HRV surface.9

Fluorescence quenching experiments

We conducted fluorescence quenching experi-ments (described in Materials and Methods) toobtain a more quantitative measure of the bindingaffinity than could be obtained from the competitiveELISA experiments. Using this technique, weinvestigated the B1 and C1 isolates, which wereseen to have two of the three most promising ELISAprofiles (using the 14-C40-1 virus as a reference).Fluorescence quenching allows accurate thermody-namic analysis of antigen–antibody binding.Quenching efficiency is a measure of the change inthe environment in the vicinity of optically activeresidues (tryptophans) as the complex is formed.

Table 2. Relative binding free energies of the B1 and C1 chimeasured by competitive ELISA (Fig. 1) and fluorescence que

ΔGobs fluorescence (kcal/mol) ΔΔGobs fluorescence

B1 –11.82±0.1 –3.51±1.0C1 –10.36±0.1 –2.05±1.014-C40-1 –8.31±0.9 0a

a Reference chimera.

The fluorescence quenching titration curve reflectsthe strength of the interaction of the antigen–antibody complex. Variations in the titration curvescan be attributed to differences of affinities of thechimeras for 2F5, which in the cases tested (seebelow) followed the same trends as in the compet-itive ELISA (Fig. 1).Fluorescence quenching measurements were sup-

ported by two control assays: increasing concentra-tions of 2F5 did not alter the backgroundfluorescence of the buffer (data not shown); andthe addition of 2F5 did not influence the emissionfluorescence of wild-type HRV14. The fluorescencequenching titration data (e.g., see Fig. 2 for B1), werewell represented by a non-linear function (describedin Materials and Methods) that determines thedissociation constant assuming monovalent bind-ing. Bivalent binding of the antibody at the NIm-IIloop has been excluded on the basis of distances andmutual orientation of the insertion sites on HRV1413

and the modeled structure of the Fab–virus complex(see below) in which the radial orientation of the Fabwith respect to the viral surface directs the other Fabarm of the antibody away from the virus particleand out into solution.The dissociation constants (Kd) for B1 (2.0±

0.68 nM), C1 (23.81±12 nM) and 14-C40-1 (769.23±125 nM) were estimated from the fluorescence data.The relative binding free energies (Table 2) wereobtained from the dissociation constants as:

DDGobs = RTln K2d = K

1d

� � ð1Þwhere Kd

2 and Kd1 are the dissociation constants of the

given (B1 or C1) and reference virus (14-C40-1),respectively. For comparison, Table 2 gives theapproximate binding free energies for the B1 andC1 viruses relative to 14-C40-1 obtained from thecompetitive ELISA assays. The ELISA binding freeenergy values are estimated on the basis of thepeptide concentration corresponding toA450=0.5 (asdescribed in Materials and Methods). As expected,the ELISA relative binding free energies (Table 2)slightly underestimate the relative binding freeenergies measured by fluorescence quenching be-cause the immobilization of the antibody in theELISA assay disables some of the virus–antibodyinteractions.18 Nevertheless, the ELISA and fluores-cence quenching measurements agree well in termsof the rank order of affinities to 2F5. Among the threechimeras, both measurements indicate that B1 is thestrongest binder followed closely by C1, and 14-C40-1 consistently shows a significantly reduced affinitycompared to B1 and C1.

meric viruses to the 2F5 mAb with respect to 14-C40-1 asnching (Fig. 2), and as calculated

(kcal/mol) ΔΔGobs ELISA (kcal/mol) ΔΔGcalc (kcal/mol)

–2.95±0.21 –2.01±0.51–1.90±0.25 –1.82±0.520a 0a


Reorganization free energy model of observedbinding affinity differences

Conformational reorganization is an importantfactor in protein recognition and binding. Bothbinding partners incur a free-energy penalty forreorganizing the ensembles of conformations pres-ent in their unbound forms to those compatible withcomplexation. Binding affinity is enhancedwhen thebinding partners are preorganized to assume bind-ing-competent conformations in their unboundforms. Minimizing the binding reorganization freeenergy is particularly important in this applicationbecause, as opposed to drug design applicationswhere the composition of the inhibitor is optimizedrelative to the protein target, the epitope sequence isfixed and cannot be varied to enhance the bindingaffinity.Earlier, we hypothesized that the binding affinity

of the ELDKWAepitope inserted ontoHRV for 2F5 isincreased when preferentially presenting this epi-tope in a conformation predisposed for binding.16

We were able to identify a set of characteristics ofchimeric HRV constructs that display the ELDKWAepitope preferentially in the same β-turn conforma-tion as that seen in the crystal structure in complexwith 2F5.19 These characteristics, including thehydrophobic nature and length of the insert as wellthe central positioning of the DKW epitope coremotif within the insert, form the design principles ofthe combinatorial libraries of chimeric virusesstudied in this work.The aim of the present modeling work was to

establish whether the earlier modeling predictions16

are consistent with the measurements of the chi-mera:2F5 binding affinities described here. In orderto generate the corresponding conformational en-sembles, the B1 and C1 chimeric viruses weremodeled by atomistic parallel molecular dynamicssimulations as described (also summarized in Mate-rials and Methods).16 The ensembles were analyzedin terms of the fraction of conformations compatiblewith 2F5 binding (described in Materials andMethods). As discussed below, a large populationof binding-competent conformations can be relatedquantitatively to a high level of antibody affinity.The computational models for the B1 and C1

viruses were generated by homology modeling asdescribed in Materials and Methods. Moleculardynamics simulations were conducted using theparallel REMD conformational sampling method(described in Materials and Methods).20 The salientsimulation results are illustrated in Fig. 3, whichshows the Cα RMSD values of the ELDKWAmotif ofthe conformers generated at 310 Kwith respect to thesolution NMR structure of a 13-mer peptide contain-ing the ELDKWA motif (1LCX)21 and the X-raystructure of the MPER-derived 7mer peptide com-plexed with 2F5 (1TJG).19 The 1LCX peptide NMRstructure was reported by Biron et al. to be helical insolution,21 whereas the peptide-Fab crystal structurecontains a β-turn conformation.19 For comparison,Fig. 3 shows the same kind of analysis for the 14-C40-

1 chimeric virus which, as described,16 assumed amore extended, less β-turn-like conformation.The simulation results (Fig. 3) indicate that the

ELDKWA epitope is quite flexible on the surface ofthe chimeric HRV and that the distribution ofconformations is substantially affected by thenature and number of the flanking amino acids(compare Fig. 3a, b, and c, whose sequences aregiven in Table 1). The majority of conformationsfrom the simulated ensembles of B1 and C1 (Fig. 3aand b) are found in the left region of the plots,indicating that they adopt conformations similar tothe target β-turn conformation more often than 14-C40-1 (Fig. 3c), which is rarely found close to thetarget crystal structure.These results mirror the binding affinity measure-

ments given in Table 2. In agreement with thereorganization free energy argument discussedabove, the chimeras with the greatest populationof conformers resembling the target mAb-bound β-turn conformation (B1 and C1) are those with thestrongest affinity to 2F5. The ensemble of confor-mers of C1 is more heterogeneous than that of B1(demonstrated by the spread of points along the x-axis of Fig. 3a and b). Furthermore, consistent withthe weaker affinity of the C1 conformers for 2F5compared to the B1 conformers, C1 presents aslightly smaller fraction of conformers that aresimilar to the bound conformation compared to B1.To quantify these observations, we estimated the

2F5 binding free energy differences for the threeconstructs, based on the corresponding fractions ofconformers in proximity to the target-bound confor-mation. To this end, we adopted a strict criterion ofproximity (described in Materials andMethods) thattakes into account, through a backbone anglesimilarity measure, the correct orientations of theside chains of the DKW epitope core motif that arecritical for the formation of the proper antigen–antibody interactions. To estimate the relativebinding affinities of the chimeras from the simula-tion, we express the free energy of binding, ΔGbind,as the sum of the antigen–antibody interactionenergy, ΔEint, and the binding reorganization freeenergy, ΔGreorg:

DGbind = DEint + DGreorg ð2Þwhere:

DGreorg = − kTlnP ð3Þand P is the fraction of conformers that arestructurally similar to the bound ELDKWA peptidein complex with 2F5. Assuming thatΔEint variationsare small in this case, based on the fact that theresidues involved in the epitope/2F5 paratope arethe same in all three chimeric viruses, the relativebinding free energies are mostly reduced to thedifference of reorganization free energies. It followsthat the relative binding free energy between viruses1 and 2 is given by

DDGbindc − kTlnP2 = P1 ð4Þ

Fig. 3 (legend on next page)


Fig. 3. Scatter plot of the Cα RMSD values from the ELDKWAmotif from the REMD ensembles of the chimeric virusesin solution at 310 K with respect to the same motif of the peptide in the 1TJG crystal structure (x-axis) and in the 1LCXNMR structure (y-axis). The B1 and C1 structures display primarily β-turn conformations, and the 14-C40-1 structuresdisplay turn conformations that are less similar to a canonical β-turn conformation.


where P1 and P2 are the populations of the binding-competent macrostates of viruses 1 and 2, respec-tively. The population of the binding-competentmacrostates is computed using a definition involv-ing both RMSD and backbone dihedral angledeviations (see Materials and Methods). We haveconfirmed that changing the threshold values of thisdefinition has a small effect on the computed relativebinding free energy from Eq. (4). We have confirmedthat the members of the defined binding-competentensemble closelymatch the target crystal structure ofthe complex (1TJG) in terms of both structural andenergetic features.The computed relative binding free energies

from Eq. (4) are compared to the experimentalrelative binding free energies from the fluorescencequenching measurements in Table 2. The modelingresults, based on the computed reorganization freeenergies (Eq. (4)), reproduce the superior bindingaffinity of the B1 and C1 chimeric viruses for 2F5relative to 14-C40-1. In general agreement with themeasured values, the binding affinities of the C1and B1 viruses are predicted to be approximately2 kcal/mol more favorable than that of 14-C40-1,with the B1 virus slightly more favored than C1(Table 2).

Model of B1:2F5 complex

We have constructed a structural model of thecomplex of the B1 virus with the 2F5 Fab using thecomputed conformation of B1 closest to the targetcrystal structure. We superimposed the ELDKWAresidues of this conformation on the correspondingresidues of the 1TJG crystal structure of thepeptide:2F5 Fab complex (Fig. 4). In this model theELDKWA epitope exhibits a remarkable shapecomplementarity with the 2F5 paratope, asevidenced by the amount of buried surface area ofthe ELDKWA motif in the complex. The ELKDWAmotif is folded in a type I β-turn conformation thatsuperimposes well on the crystal structure of thecomplex of the ELDKWA peptide with the 2F5.The interactions between the B1 chimeric virus andthe antibody are mostly concentrated at the center ofthe epitope where the DKW core makes contactswith the corresponding paratope of 2F5 (Fig. 4). Theinterface between the two binding partners ispredominantly hydrophobic with interaction be-tween W666 of B1 and P98 of 2F5 and, to a lesserextent, with V100. A salt bridge between D664 of B1and R95 of 2F5 is also present. The main chain ofD664 interacts with the aromatic ring of Y94 of the

Fig. 4 (legend on next page)


Fig. 4. Amodel of the B1 chimera bound to the 2F5 Fab. (a) One of the VP2 subunits of the protomeric unit is shown ingreen. The Fab has a ribbon representation, with the light chain shown in orange and the heavy chain shown in cyan. TheDKW motif of the 2F5 epitope is shown in a stick representation. (b) Close-up view of the binding region of the modeledB1:2F5 Fab complex. Residues of the ELDKWA motif and residues of the 2F5 within 7 Å are shown in stickrepresentations, and the rest are depicted as ribbons, with VP2 in green and the heavy chain, including the H3 region, andthe light chain of 2F5 shown in cyan and orange, respectively. The HIV-1 residues are numbered according to their HIV-1HxB2 amino acid numbers. (c) As (b) with the modeled conformation of the ELDKWA peptide (magenta) in the crystalstructure (1TJG) superimposed.


light chain. Notably, the residues at the C-terminalend of ELDKWA assume an α-helical conformation,previously shown to be crucial for interaction with2F5.21,22 The model for the B1:2F5 complex showshow binding can occur without interference with thelong CDR H3 finger of 2F5, which would otherwiseprevent the formation of a high-affinity complex byclashing with the virus capsid. A similarly good fitwas observed for the model of the cHRV3 chimericvirus in complex with the 2F5 Fab,16 indicating that,in both cases, the flanking residues induce epitopeconformations suitable for 2F5 recognition.

Discussion and Conclusions

We have analyzed the antigenic properties ofisolates from a combinatorial library of HRV:HIVchimeric viruses displaying the ELDKWA epitopeof the MPER of the gp41 transmembrane glycopro-tein of HIV. The design of the library was based onthe results of modeling calculations that identified anumber of energetic and steric factors affecting theconformational propensities of the ELDKWA insert

onto the HRV virus and its ability to assumeconformations capable of fitting into the comple-mentary determinant region of 2F5. From thislibrary, we have been able to isolate a number ofpromising chimeras that bind 2F5 with greateraffinity than that obtained earlier.Remarkably, in a competitive ELISA, four of the

five chimeras tested from this library bound 2F5withgreater affinity than one of the best binding chimerasfrom a previous library (14-C40-1, which also eliciteddiverse neutralizing anti-HIV antibodies whenimmunized with an ELDKWA-based peptide).9

Fluorescence quenching measurements showedthat the best among the chimeras analyzed, B1,binds to 2F5 400 times more strongly than does 14-C40-1. The difference in epitope sequence betweenthe two constructs (ALDKWA in 14C40-1 andELDKWA in B1 and C1) is likely unrelated to thevariations in binding affinities observed in this workbecause the mutation of a pseudovirus from E to Adid not interfere with neutralization.3 Moleculardynamics simulations of the B1, C1 and 14-C40-1chimeras confirm the crucial role of the favorableconformational propensities of the inserts that are


responsible, in part, for the corresponding gain inbinding affinity.Our results for the B1 chimera indicate a dissoci-

ation constant in the nanomolar range, similar tothat measured for the free ELDKWA peptide.17,23

This correspondence is consistent with the reorga-nization free energy we propose. The presentcomputational result, together with our earlieranalysis,16 suggests that the free peptide and theB1 epitope have similar propensities for adoptingbinding-competent conformations. The key result ofthis work is to show that it is possible to display theepitope on HRV in conformations favoring high-affinity binding despite the structural restraintspresent at the insertion site. It is useful to comparethe binding affinities to 2F5 measured for the HRVchimeras studied in this work with those corres-ponding to systems closer to the native environ-ment. The affinity of B1 for 2F5 (Kd 2 nM) is similarto those (Kd 0.35–1.39 nM) between the 2F5 Fab andtrimeric MPER constructs designed to mimic theintermediate conformation of gp41, which is be-lieved to be the target of the 2F5 mAb in vivo.24,25

This correspondence between the affinity of 2F5 forthe engineered HRV chimeras and the trimericMPER suggests, according to the conformationalreorganization free energy model proposed, that theELDKWA epitope is displayed on B1 in a confor-mation similar to that in the native environment.The value of this work lies in the demonstration

that, when properly employed, computational toolscan be a useful complement to more experimentalmethods of protein design26 by making it possible totest hypotheses in silico and pursue the hypotheseswith subsequent focused experimental testing invitro. The agreement between computed and ob-served binding affinities highlights the usefulness ofthe free energy reorganization model in rationaliz-ing the epitope binding trends and confirms thevalidity of atomistic models as a tools to design denovo chimeric viruses that are likely to bind well invitro and might lead to more timely identification ofuseful candidates for vaccine development.The relative binding affinities from fluorescence

quenching are greater than the correspondingcomputed values (relative free energy differencesof –3.50 versus –2.01 kcal/mol for B1 and –2.02versus –1.82 kcal/mol for C1 relative to 14-C40-1).This quantitative discrepancy can be attributed to anumber of methodological factors, including limita-tions of the accuracy of the potential model used inthe simulations. It could be interpreted also as anindication that conformational reorganization of theepitope, which is the only quantity probed by thecalculations, is quantitatively important but not theonly factor influencing the observed binding freeenergy differences. Other factors, such as binding-interaction energies and antibody reorganization,can contribute to variations in binding affinity. Ourresults indicate that in this case differences inconformational reorganization are more importantfor determining relative antibody affinities thanother factors that are not modeled in the simulations.

The prediction accuracy of the atomistic moleculardynamics simulations we have used is limited bytheir ability to properly sample conformationalspace and by the reliability of the energy function.Assumptions inherent in the model concerning thesolution environment (pH, salt and cosolvent con-centrations, etc.) can further limit the fidelity ofcomputational models.We recognize that strong binding affinity between

HRV chimeric viruses and the 2F5 mAb (antigenic-ity) does not guarantee the ability of the chimeras toelicit the production of neutralizing antibodiesagainst HIV (immunogenicity).25 For instance, theviral membrane, which has not been considered inthis work, has been hypothesized to play an impor-tant role in the binding of the epitope to 2F5.27

Separate immunization and HIV neutralizationassays in animal models are needed to probe theimmunogenicity of these chimeric viruses in elicitingneutralizing responses against HIV.In conclusion, this study presents the application

of a computational approach to quantitativelyestimate binding affinities of chimeric viruses tothe 2F5 mAb. The simulated binding affinities agreewell with both forms of spectroscopic measure-ments used in this study. In addition, we haveidentified a chimeric HRV:HIV virus, B1, with asignificantly enhanced binding affinity (ΔΔG –3.5 kcal/mol) relative to one of the best bindingchimeras (14-C40-1) previously characterized. Theimproved antigenic properties of B1 virus make thisantigen a very attractive candidate for immunoge-nicity studies. The molecular design of B1 hasproved that it is possible to improve the ability ofantigens to bind specifically with complementaryantibody based on structural insights, which is animportant part of rational vaccine design.

Materials and Methods

Cells, viruses, plasmids, and media

H1-HeLa cells5 were used for the production, propaga-tion, titering, and immunoselection of both wild-typeHRV14 and recombinant HRV14:HIV-1 gp41 ELDKWASchimeric viruses.28 Media M and PA11 were used with 1–10% (v/v) fetal bovine serum for plaque isolations andvirus propagations. Dulbecco's modified Eagle medium(DMEM;Gibco, BRL, Carlsbad, CA)was used forHeLa celltransfection experiments. The bacterial strain used fortransformation of the engineered plamids was Escherichiacoli DH10B ElectroMax (Gibco, BRL, Carlsbad, CA). ThepST-LIC plasmid, which was used for all genetic engi-neering, differs from the HRV-encoding p3IIST plasmid29

used earlier in the engineered region (i.e. the regionencoding the VP2 puff of the NIm-III region). In pST-LIC,the unique ApaI site was replaced with a BseRI site thatcould be digested first with BseRI and then with bacte-riophage T4 DNA polymerase. This polymerase has 3′→5′exonuclease activity in the absence of dNTPs, whichallowed us to silently engineer the HRV sequences thatflank the recombinant inserts to have an 11–15 sequencedevoid of C or G residues (depending on the DNA strand)to allow the generation of 11–15 base sticky ends (in the


presence of dCTP or dGTP). This plasmid promotesefficient mutagenesis with foreign inserts with comple-mentary sticky ends.

Monoclonal antibodies and peptides

The broadly neutralizing human mAb 2F5 (PolyMun,Inc., Vienna), which binds the gp41 MPER epitopeELDKWA,30 was used for immunoselection and antige-nicity assays. Murine mAb 17, which binds to theneutralizing immunogenic site IA (NIm IA) of HRV14,15

was used for ELISAs of chimeric HRV. Horseradishperoxidase-conjugated goat-anti-mouse IgG (CappelICN, Irvine, CA) was used for detection of chimericHRVs. Ac-EQELLELDKWASSLW-NH2 peptide, synthe-sized by NeoMPS, Inc. (San Diego, CA), was used forcompetition in ELISAs and for elution of the chimericviruses in competitive immunoselections.

Design and generation of chimeric HRV14:ELDKWAlibrary

The combinatorial library described was designedfollowing the molecular simulation study that led to theinception of the cHRV3 construct.16 A ligation-indepen-dent cloning method was used in which the pST-LICplasmidwas digestedwith BseRI and T4DNApolymeraseto generate long sticky ends. A complementary recombi-nant insert was prepared for hybridization with theplasmid by first hybridizing two DNA oligonucleotides(synthesized by IDT, Piscataway, NJ) overlapping in theELDKWAS-encoding region using codons preferred by theHRV14 RNA polymerase (Fig. 5). To optimize the bindingof the foreign epitope to 2F5, the non-overlapping regionsof the DNA oligomers encoded two randomized residueson each side of the ELDKWAS sequence (Table 1). Beyondthese sequences were the sequences encoding the adjacentHRV14 residues to provide the sticky ends for hybridizingwith the plasmid. The hybridized oligonucleotides wereextended (using each of the dNTPs and Klenow DNApolymerase I (New England Biolabs, Beverly, MA)),treated with T4 DNA polymerase (and dCTP or dGTP togenerate the complementary sticky ends), and allowed tohybridize in a 10:1 ratio with the prepared vector.Electroporation of DH10B E. coli cells was performedusing a Gene Pulser system (Bio-Rad laboratories, Hercu-les, CA) according to the manufacturer's instructions(25 μF, 200 V, 200 A). Transformed cells were grown inbulk liquid cultures at 30 °C to late log phase and thenharvested using Qiagen Mini-Prep Spin Kits (Valencia,CA). After linearizing the pools of recombinant plasmidswith MluI, plasmids were transcribed in vitro usingAMBION MegaScript kits (Austin, TX). Full-length RNAtranscripts were transfected into H1-HeLa cells by electro-poration (10 μL of the infectious RNAs added to 1×107 of

Fig. 5. The DNA oligonucleotides for the forward and reverN, equimolar fractions of A, G, C, and T; and S, equimolar fra

H1-HeLa cells in 400 μL of D-MEM and then placed in a0.4 cm Gibco-BRL electroporation cuvette and electropo-rated at 250 V, 960 μF. The resultant chimeric viruses wereharvested as described.11

Partial purification of virus libraries and viral isolates

A protocol developed by Zhang et al.31 was followed forthe partial purification of the combinatorial virus libraryas well as individual viruses. After three cycles of freezingand thawing, infected cells were harvested, concentratedand the suspension was centrifuged (to remove cell walldebris), after which the clarified lysates were subjected totreatment with DNase I followed by ultracentrifugation(with a 30% (w/v) sucrose cushion) at 42,000 rpm for 2.5 hat 15 °C in a Beckman 45 Ti rotor. Pellets were suspendedin 20 mM Tris–HCl, pH 7.4 and contained the chimericviruses as well as some cell-derived ribosomes.

Immunoselection of viruses

Immunosorp plates (96 well; Nunc, Rochester, NY)were coated with mAb 2F5 to immunoselect chimericrhinovirus pools from the new library, essentially asdescribed.12 After incubation at 4 °C overnight, the plateswere washed and blocked with 3% (w/v) gelatin inphosphate-buffered saline (PBS). After incubation for 1 hat 37 °C, the plates were washed six times with PBS-T (PBScontaining 0.05% (v/v) Tween 20). Serial dilutions of thecompeting peptide Ac-EQELLELDKWASSLW-NH2,yielding 0, 2, 4, 8, 16 pmol/well, were prepared and3×105 plaque-forming units of the chimeric virus librarywere added. The plates were incubated for 2 h at 37 °C,washed six times with PBS-T and then 2×104 H1-HeLacells in mediumMwith 10% (v/v) FBSwere added to eachwell. The plates were incubated for 70 h at 34.5 °C in anatmosphere of 2.5% (v/v) CO2 and cells were harvestedwhen they displayed 80% cytopathic effect (CPE). Theharvested cells were lysed by three cycles of freezing andthawing and then centrifuged to produce clarified viruslysates.

Plaque isolation and sequencing

Individual viral isolates were obtained as described.17

Plaque purification was done twice to minimize contam-ination by additional viruses. QIAamp viral RNA purifi-cation kits were used to extract the viral RNA of chimericviruses according to the manufacturer's instructions. Theresulting DNA was analyzed by electrophoresis in a 2%(w/v) agarose gel. QIAquick gel extraction kits (Valencia,CA) were used to purify the products, which weresequenced at the UMDNJ-RWJMS DNA Core Facility(Piscataway, NJ).

se primers used to generate the new combinatorial library.ctions of C and G.


Competitive enzyme-linked immunosorbent assays(ELISAs) of chimeric viruses

Competitive ELISAs were done in triplicate usingimmobilized mAb 2F5 essentially as described9 with theexception that 3×105 plaque-forming units of virus weremixed with the 14-mer ELDKWA-containing peptide inserial dilutions before being added to the immobilized 2F5for 120 min at 37 °C. The anti-HRV14 mAb 17 was thenadded, incubated, and washed before treatment withhorseradish peroxidase-conjugated goat anti-mouse IgG(at a 1:10,000 dilution). This was incubated, washed, andtreatedwith, peroxidase substrate (0.3mg/mL tetramethyl-benzidine dissolved in 10% dimethyl sulfoxide, 0.18 Msodium citrate, pH 3.95). The color reactionwas obtained byadding H2O2 to 0.001% (v/v), and then stopped by addingan equal volume of 4 M H2SO4, and the absorbance at450 nm (A450) was measured. It was expected that thepresence of the competing peptide would lead to morestringent conditions, promoting the binding of 2F5 to onlythose chimeras displaying ELDKWA in conformationsbetter resembling those of the native epitope than thepeptide. The binding free energies of theB1 andC1chimerasrelative to 14-C40-1 were estimated from the ELISA data as:

DDGi = − kTlnCi = C14−C40−1

where Ci is the concentration of the peptide at A450=0.5for the corresponding chimera. The statistical uncertain-ties of the ELISA-derived ΔΔG's were obtained by errorpropagation.

Fluorescence measurements

Fluorescence quenching experiments were used tomeasure the binding constants between the mAb 2F5 andselected chimeric viruses. The fluorescence spectra wereobtained using a spectrofluorimeter equippedwith a quartzcuvette of 1 cm path length and a thermally controlled cellholder.32 The fluorescence emission (λem) spectra wererecorded from 300 nm to 450 nm, with the path length forfluorescence excitation and emission of 5 mm, at anexcitation wavelength (λex) of 280 nm. Virus–antibodybinding was monitored by changes in the fluorescenceemission intensity recorded at 370 nm. Changes in intrinsicfluorescence of viruses and antibody fragments were usedto monitor ligand binding at 25 °C by titration with mAb2F5. All virus and antibody solutions were prepared in20 mM Tris–HCl, pH 7.2 made with deionized H2O. Alltitrations were done with the same amounts of 2F5 addedto a fixed concentration of protein, holding the concentra-tion of chimeric viruses B1, C1, and 14-C40-1 at 1 nM. Tocover the broad range of ligand concentration, stocksolutions with concentrations of 2F5 ranging from6.74×10–9 to 1.04×10−6 M were used. Freshly prepared2mL solutions of B1, C1, and 14-C40-1 (in 20mMTris–HCl,pH 7.2) at the desired concentration were placed into thecuvette, 1 μL of 2F5 solution was added and the mixturewas stirred. Values are reported as the average of twoindependent measurements for the experiments done in thepresence of chimeric virus (Fig. 5b). We calculated thedissociation constant (Kd) by fitting the 2F5-inducedchanges in fluorescence using the following formula:32

I = I0 +0 I∞ − Ioð Þ

2 D½ �tot ½ D½ �tot + R½ �tot + Kd� �

−ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiD½ �tot + R½ �tot + Kd

� �2 − 4 D½ �tot R½ �totq �

In this relationship, I0 and I are the fluorescence emissionintensity of the virus in the absence and in the presence ofantibody, respectively; I∞ is the fluorescence emissionintensity of the virus at saturation; and [D]tot and [R]totare the total concentrations of epitope and antibody,respectively. The statistical standard error estimates forKd were obtained by non-linear regression analysis basedon the error bounds of the fluorescence intensities.

Replica exchange molecular dynamics

The temperature replica exchange molecular dynamics(REMD) method33–35 as implemented in the IMPACTcomputational package36 was used in this work to explorethe predicted conformational space of several peptidesequences derived from the MPER of HIV gp41 andinserted into the HRV14 viral capsid. REMD is anadvanced canonical conformational sampling algorithmdesigned to help overcome the sampling problem en-countered in biomolecular simulations. It consists ofrunning a series of MD simulations of the molecularsystem in parallel over multiple processors at differenttemperatures. Periodically, a Monte Carlo move isattempted, aimed at swapping atomic coordinates of onereplica with those of another. Typically, the lowesttemperature replica corresponds to the temperature ofinterest and the high temperature replicas allow a rapidexploration of the extent of conformational space. In thetemperature replica exchange method, efficient intercon-version among low-energy conformations occurs wherebythe temperature switches from low to high, allowing theconformation to overcome potential energy barriers andthen transition to another conformation. A new low-energy conformation can then be established when thisconformation transfers back to low temperature replicas.

Force-field and implicit solvation model

All simulations employed the OPLS-AA fixed chargeall-atom force field37,38 and the AGBNP19 implicit solventmodel to mimic the water environment. Simulations usingimplicit solvent models are not as computationallyintensive as those using explicit solvent models. This isparticularly true for REMD simulations because far fewerreplicas are needed to achieve reasonable replica exchangerates due to the smaller system and broader overlap of theenergy distributions among replicas at different tempera-tures. AGBNP is an implicit solvent model based on anovel implementation of the pairwise descreening schemeof the generalized Born model39,40 for the electrostaticcomponent, and a novel non-polar hydration-free energyestimator. The non-polar term consists of an estimator forthe solute solvent van der Waals dispersion energydesigned to mimic explicit solvent solute-solvent van derWaals interaction energies in addition to a surface areaterm corresponding to the work required for cavityformation. It has been used to study protein conforma-tional equilibrium and protein allostery.41

Model preparation

The B1 and C1 HRV14:HIV chimeric constructs inves-tigated in this work were obtained by insertion of gp41-derived epitopes into theβE-βF region of the so-called VP2puff of the VP2 subunit of the HRV14 protomeric unit, thelargest of three loops forming the neutralizing immuno-genic site II (NIm-II) of HRV14 (Table 1).13,14 The chimericcHRV3 construct16 contains the sequence ELDKWAS


inserted between the N160 and E161 residues of VP2,without the removal of any HRV14 residue or the additionof any flanking residue. The new chimeric virus libraryfrom which the B1 and C1 chimeras were isolated differsfrom the cHRV3 chimera in that the two HRV residuesflanking the ELDKWAS insert were randomized instead ofbeing limited to the HRV14 AN and EV residues (Table 1).Starting structures were obtained by homology modelingusing the program Prime42 (Schrodinger, Inc.) and thestructure of wild-type HRV14 (PDB ID 4RHV) as astructural template (97% sequence identity). The energy-minimized models were assessed for quality usingProCheck, indicating 84.6%, 14.6%, 0.7%, and 0.1% in themost favored regions, allowed regions, generouslyallowed regions, and disallowed regions of the Ramachan-dran plot, respectively.43

REMD simulations of the chimeric constructs

REMD simulations of the B1 and C1 viral constructswere done to estimate the conformational propensities ofthe inserts. The production REMD runs, consisting of 30replicas with temperatures in the range 310–495 K, were5 ns per replica in length for a total of 60 ns for each system.The production runs were preceded by minimization andequilibration calculations starting with the comparativemodels described above. To reduce computational com-plexity, the model included only the HRV14 residues(including those from the VP1 and VP3 subunits) thatcontained atoms within 20 Å from any atom of the foreigninsert and, furthermore, only the 176 residues closest to theinsert were allowed to move. The positions of Cα atoms ofthe residues at the ends of dangling protein chains wereharmonically restrained using an isotropic force constantof 0.3 kcal mol–1Å–2. This setup prevents unfolding of thesimulated protein region at high temperature whileproviding a reasonable description of the protein environ-ment surrounding the inserted sequence. This methodol-ogy was tested by applying it to the chimeric HRVdisplaying part of the HIV-1 V3 loop of the MN-III-2chimera studied by Smith et al.12 and by Ding et al.44 Weconfirmed that the main cluster of conformations of theloop generated by REMD structurally matched that of thecrystal structure.

Algorithm to detect conformer similarity

We have used a structural criterion to select conformersof the simulated ensembles of B1, C1, and 14-C40-1 thatmatch closely to the ELDKWA peptide conformationbound to the 2F5 Fab (1TJG). A conformation belongs tothis binding-competent macrostate if it is within 2 Å Cα

RMSD and within 30° of the backbone angles of DKWrelative to the peptide in 1TJG. This analysis allowed us todefine the population of the bound state and to estimatethe free energy of reorganization (ΔΔGcalc) from theREMDsimulation. Statistical uncertainty for the populations wasobtained by means of a Bayesian inference analysis basedon the counts of binding-competent conformations and thetotal numbers of samples as described.45

Acknowledgements

This work was supported, in part, by NIH grantGM30580 (to R.M.L.) and by NIH grant AI071874

(to G.F.A.). The computer simulations for this workwere done at the BioMaPS High PerformanceComputing Center at Rutgers University andfunded, in part, by National Institute of Healthshared instrumentation grant no. 1 S10 RR022375.We are grateful to Thomas Mariano and RachelBradley for their assistance with cell cultures, andto Chris Barbieri for assistance with the fluores-cence experiment.

Supplementary Data

Supplementary data associated with this articlecan be found, in the online version, at doi:10.1016/j.jmb.2010.01.064

References

1. Karlsson Hedestam, G. B., Fouchier, R. A., Phogat,S., Burton, D. R., Sodroski, J. & Wyatt, R. T. (2008).The challenges of eliciting neutralizing antibodies toHIV-1 and to influenza virus. Nat. Rev. Microbiol. 6,143–155.

2. Alfsen, A. & Bomsel, M. (2002). HIV-1 gp41 enveloperesidues 650-685 exposed on native virus act as a lectinto bind epithelial cell galactosyl ceramide. J. Biol.Chem. 277, 25649–25659.

3. Zwick,M. B., Jensen, R., Church, S.,Wang,M., Stiegler,G., Kunert, R. et al. (2005). Anti-human immunodefi-ciency virus type 1 (HIV-1) antibodies 2F5 and 4E10require surprisingly few crucial residues in themembrane-proximal external region of glycoproteingp41 to neutralize HIV-1. J. Virol. 79, 1252–1261.

4. Penn-Nicholson, A., Han, D. P., Kim, S. J., Park, H.,Ansari, R., Montefiori, D. C. & Cho, M. W. (2008).Assessment of antibody responses against gp41 inHIV-1-infected patients using soluble gp41 fusionproteins and peptides derived from M group consen-sus envelope. Virology, 372, 442–456.

5. Muster, T., Guinea, R., Trkola, A., Purtscher, M.,Klima, A., Steindl, F. et al. (1994). Cross-neutralizingactivity against divergent human immunodeficiencyvirus type 1 isolates induced by the gp41 sequenceELDKWAS. J. Virol. 68, 4031–4034.

6. Cardoso, R. M., Zwick, M. B., Stanfield, R. L., Kunert,R., Binley, J. M., Katinger, H. et al. (2005). Broadlyneutralizing anti-HIV antibody 4E10 recognizes ahelical conformation of a highly conserved fusion-associated motif in gp41. Immunity, 22, 163–173.

7. Pejchal, R., Gach, J. S., Brunel, F. M., Cardoso, R. M.,Stanfield, R. L., Dawson, P. E. et al. (2009). Aconformational switch in human immunodeficiencyvirus gp41 revealed by the structures of overlappingepitopes recognized by neutralizing antibodies.J. Virol. 83, 8451–8462.

8. Matyas, G. R., Wieczorek, L., Beck, Z., Ochsenbauer-Jambor, C., Kappes, J. C., Michael, N. L. et al. (2009).Neutralizing antibodies induced by liposomal HIV-1glycoprotein 41 peptide simultaneously bind to boththe 2F5 or 4E10 epitope and lipid epitopes. AIDS, 23,2069–2077.

9. Resnick, D. A., Smith, A. D., Gesiler, S. C., Zhang, A.,Arnold, E. & Arnold, G. F. (1995). Chimeras from ahuman rhinovirus 14-human immunodeficiency virustype 1 (HIV-1) V3 loop seroprevalence library induce




neutralizing responses against HIV-1. J. Virol. 69,2406–2411.

10. Smith, A. D., Geisler, S. C., Chen, A. A., Resnick, D. A.,Roy, B. M., Lewi, P. J. et al. (1998). Human rhinovirustype 14:human immunodeficiency virus type 1 (HIV-1) V3 loop chimeras from a combinatorial libraryinduce potent neutralizing antibody responses againstHIV-1. J. Virol. 72, 651–659.

11. Arnold, E. & Rossmann, M. G. (1990). Analysis of thestructure of a common cold virus, human rhinovirus14, refined at a resolution of 3.0 Å. J. Mol. Biol. 211,763–801.

12. Arnold, G. F., Velasco, P. K., Holmes, A. K., Wrin, T.,Geisler, S. C., Phung, P. et al. (2009). Broad neutraliza-tion of human immunodeficiency virus type 1 (HIV-1)elicited from human rhinoviruses that display theHIV-1 gp41 ELDKWA epitope. J. Virol. 83, 5087–5100.

13. Gallichan, W. S., Johnson, D. C., Graham, F. L. &Rosenthal, K. L. (1995). Mucosal immunization with arecombinant adenovirus vector induces local andsystemic immunity and protection from herpessimplex virus. Adv. Exp. Med. Biol. 371, 1581–1585.

14. Rossmann, M. G., Arnold, E., Erickson, J. W.,Frankenberger, E. A., Griffith, J. P., Hecht, H. J.et al. (1985). Structure of a human common cold virusand functional relationship to other picornaviruses.Nature, 317, 145–153.

15. Sherry, B. & Rueckert, R. (1985). Evidence for at leasttwo dominant neutralization antigens on humanrhinovirus 14. J. Virol. 53, 137–143.

16. Lapelosa, M., Gallicchio, E., Arnold, G. F., Arnold, E.& Levy, R. M. (2009). In silico vaccine design based onmolecular simulations of rhinovirus chimeras pre-senting HIV-1 gp41 epitopes. J. Mol. Biol. 385, 675–691.

17. Alam, S. M., McAdams, M., Boren, D., Rak, M.,Scearce, R. M., Gao, F. et al. (2007). The role ofantibody polyspecificity and lipid reactivity in bind-ing of broadly neutralizing anti-HIV-1 envelopehuman monoclonal antibodies 2F5 and 4E10 toglycoprotein 41 membrane proximal envelope epi-topes. J. Immunol. 178, 4424–4435.

18. Dekker, E. L., Porta, C. & Van Regenmortel, M. H.(1989). Limitations of different ELISA procedures forlocalizing epitopes in viral coat protein subunits.Arch.Virol. 105, 269–286.

19. Ofek, G., Tang, M., Sambor, A., Katinger, H., Mascola,J. R., Wyatt, R. & Kwong, P. D. (2004). Structure andmechanistic analysis of the anti-human immunodefi-ciency virus type 1 antibody 2F5 in complex with itsgp41 epitope. J. Virol. 78, 10724–10737.

20. Gallicchio, E. & Levy, R. M. (2004). AGBNP: ananalytic implicit solvent model suitable for moleculardynamics simulations and high-resolution modeling.J. Comput. Chem. 25, 479–499.

21. Biron, Z., Khare, S., Samson, A. O., Hayek, Y., Naider,F. & Anglister, J. (2002). A monomeric 3(10)-helix isformed in water by a 13-residue peptide representingthe neutralizing determinant of HIV-1 on gp41.Biochemistry, 41, 12687–12696.

22. Bryson, S., Julien, J. P., Isenman, D. E., Kunert, R.,Katinger, H. & Pai, E. F. (2008). Crystal structure of thecomplex between the F(ab)′ fragment of the cross-neutralizing anti-HIV-1 antibody 2F5 and the F(ab)fragment of its anti-idiotypic antibody 3H6. J. Mol.Biol. 382, 910–919.

23. Tian, Y., Ramesh, C. V., Ma, X., Naqvi, S., Patel, T.,Cenizal, T. et al. (2002). Structure-affinity relationshipsin the gp41 ELDKWA epitope for the HIV-1 neutral-izing monoclonal antibody 2F5: effects of side-chain

and backbone modifications and conformationalconstraints. J. Pept. Res. 59, 264–276.

24. Frey, G., Peng, H., Rits-Volloch, S., Morelli, M., Cheng,Y. & Chen, B. (2008). A fusion-intermediate state ofHIV-1 gp41 targeted by broadly neutralizing anti-bodies. Proc. Natl Acad. Sci. USA, 105, 3739–3744.

25. Hinz, A., Schoehn, G., Quendler, H., Hulsik, D. L.,Stiegler, G., Katinger, H. et al. (2009). Characterizationof a trimeric MPER containing HIV-1 gp41 antigen.Virology, 390, 221–227.

26. Sood, V. D. & Baker, D. (2006). Recapitulation anddesign of protein binding peptide structures andsequences. J. Mol. Biol. 357, 917–927.

27. Brunel, F. M., Zwick, M. B., Cardoso, R. M., Nelson,J. D., Wilson, I. A., Burton, D. R. & Dawson, P. E.(2006). Structure-function analysis of the epitope for4E10, a broadly neutralizing human immunodeficiencyvirus type 1 antibody. J. Virol. 80, 1680–1687.

28. Conant, R. M. & Hamparian, V. V. (1968). Rhino-viruses: basis for a numbering system. II. Serologiccharacterization of prototype strains. J. Immunol. 100,114–119.

29. Smith, A. D., Resnick, D. A., Zhang, A., Geisler, S. C.,Arnold, E. & Arnold, G. F. (1994). Use of randomsystematic mutagenesis to generate viable humanrhinovirus 14 chimeras displaying human immuno-deficiency virus type 1 V3 loop sequences. J. Virol. 68,575–579.

30. Muster, T., Steindl, F., Purtscher, M., Trkola, A.,Klima, A., Himmler, G. et al. (1993). A conservedneutralizing epitope on gp41 of human immunodefi-ciency virus type 1. J. Virol. 67, 6642–6647.

31. Zhang,A.,Nanni, R.G., Li, T., Arnold,G. F.,Oren,D.A.,Jacobo-Molina, A. et al. (1993). Structure determinationof antiviral compound SCH 38057 complexed withhuman rhinovirus 14. J. Mol. Biol. 230, 857–867.

32. Barbieri, C. M., Kaul, M. & Pilch, D. S. (2007). Use of 2-aminopurine as a fluorescent tool for characterizingantibiotic recognition of the bacterial rRNA A-site.Tetrahedron, 63, 3567–6574.

33. Sugita, Y. & Okamoto, Y. (1999). Replica-exchangemolecular dynamics method for protein folding.Chem. Phys. Lett. 314, 141–151.

34. Felts, A. K., Harano, Y., Gallicchio, E. & Levy, R. M.(2004). Free energy surfaces of beta-hairpin and alpha-helical peptides generated by replica exchange mo-lecular dynamics with the AGBNP implicit solventmodel. Proteins: Struct. Funct. Genet. 56, 310–321.

35. Gallicchio, E., Levy, R. M. & Parashar, M. (2007).Asynchronous replica exchange for molecular simula-tions. J. Comput. Chem. 29, 788–794.

36. Banks, J. L., Beard,H. S., Cao,Y., Cho,A. E., Damm,W.,Farid, R. et al. (2005). Integrated modeling program,applied chemical theory (IMPACT). J. Comput. Chem.26, 1752–1780.

37. Kaminski, G. A., Friesner, R. A., Tirado-Rives, J. &Jorgensen, W. L. (2001). Evaluation and reparame-trization of the OPLS-AA force field for proteins viacomparison with accurate quantum chemical calcula-tions on peptides. J. Phys. Chem. B, 105, 6474–6487.

38. Jorgensen, W. L., Maxwell, D. S. & Tirado-Rives, J.(1996). Development and testing of the OPLS all-atom force field on conformational energetics andproperties of organic liquids. J. Am. Chem. Soc. 118,11225–11236.

39. Qiu, D., Shenkin, P. S., Hollinger, F. P. & Still, W. C.(1997). The GB/SA Continuummodel for solvation. Afast analytical method for the calculation of approx-imate born radii. J. Phys. Chem. A, 101, 3005–3014.


40. Bashford, D. & Case, D. A. (2000). Generalized bornmodels of macromolecular solvation effects. Annu.Rev. Phys. Chem. 51, 129–152.

41. Ravindranathan, K. P., Gallicchio, E. & Levy, R. M.(2005). Conformational equilibria and free energyprofiles for the allosteric transition of the ribose-binding protein. J. Mol. Biol. 353, 196–210.

42. Guallar, V., Jacobson, M., McDermott, A. & Friesner,R. A. (2004). Computational modeling of the catalyticreaction in triosephosphate isomerase. J. Mol. Biol. 337,227–239.

43. Laskowski, R. A., MacArthur, M. W., Moss, D. S. &

Thornton, J. M. (1993). PROCHECK: a program tocheck the stereochemical quality of protein structures.J. Appl. Crystallogr. 26, 283–291.

44. Ding, J., Smith, A. D., Geisler, S. C., Ma, X., Arnold,G. F. & Arnold, E. (2002). Crystal structure of ahuman rhinovirus that displays part of the HIV-1 V3loop and induces neutralizing antibodies againstHIV-1. Structure, 10, 999–1011.

45. Gallicchio, E., Andrec, M., Felts, A. K. & Levy, R. M.(2005). Temperature weighted histogram analysismethod, replica exchange, and transition paths.J. Phys. Chem. B, 109, 6722–6731.

Antigenic Characteristics of Rhinovirus Chimeras Designed ...€¦ · Antigenic Characteristics of Rhinovirus Chimeras Designed in silico for En5hanced Presentation of HIV-1 gp41

Documents