|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 278, Issue 28, 26208-26215, July 11, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Structural Basis for Variation in Adenovirus Affinity for the Cellular Coxsackievirus and Adenovirus Receptor*
From the Biology Department, Brookhaven National Laboratory, Upton, New York 11973
Received for publication, February 11, 2003 , and in revised form, April 15, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
The majority of adenovirus serotypes can bind to the coxsackievirus and adenovirus receptor (CAR) on human cells despite only limited conservation of the amino acid residues that comprise the receptor-binding sites of these viruses. Using a fluorescence anisotropy-based assay, we determined that the recombinant knob domain of the fiber protein from adenovirus serotype (Ad) 2 binds the soluble, N-terminal domain (domain 1 (D1)) of CAR with 8-fold greater affinity than does the recombinant knob domain from Ad12. Homology modeling predicted that the increased affinity of Ad2 knob for CAR D1 could result from additional contacts within the binding interface contributed by two residues, Ser408 and Tyr477, which are not conserved in the Ad12 knob. Consistent with this structural model, substitution of serine and tyrosine for the corresponding residues in the Ad12 knob (P417S and S489Y) increased the binding affinity by 4- and 8-fold, respectively, whereas the double mutation increased binding affinity 10-fold. X-ray structure analysis of Ad12 knob mutants P417S and S489Y indicated that both substituted residues potentially could form additional hydrogen bonds across the knob-CAR interface. Structural changes resulting from these mutations were highly localized, implying that the high tolerance for surface variation conferred by the stable knob scaffold can minimize the impact of antigenic drift on binding specificity and affinity during evolution of virus serotypes. Our results suggest that the interaction of knob domains from different adenovirus serotypes with CAR D1 can be accurately modeled using the Ad12 knob-CAR D1 crystal structure as a template.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Interaction of animal viruses with their cellular receptors is of particular interest because the viral proteins that function in molecular recognition are subject to immunoselective pressure to vary their antigenic structure. The potentially negative impact of antigenic variation on recognition of the cellular receptor is minimized in some virus families by protein structural features that shield receptor-binding sites from attack by host-neutralizing antibodies (1). In adenoviruses, by contrast, receptor-binding sites are exposed and highly variable in sequence. Therefore, adenovirus-receptor interaction presents an opportunity to observe the range of structural plasticity that can be tolerated without loss of binding specificity.
Receptor-binding function in adenoviruses is associated with viral fibers (2, 3), rod-shaped homotrimeric proteins (4) that protrude from each vertex of the icosahedral virus capsid. The distal end of viral fibers consists of a globular domain, the head or knob domain, which has receptor binding activity (5, 6). Human adenoviral fiber proteins have evolved to recognize several different host cell receptors (7–9); however, to date only one of these receptors has been molecularly characterized, a 46-kDa membrane glycoprotein known as the coxsackievirus and adenovirus receptor (CAR)1 (10, 11). Representative serotypes from adenovirus subgroups A, C, D, E, and F have been shown to interact with CAR (12), suggesting that CAR may be the most common receptor used by human adenoviruses. The CAR protein spans the cell plasma membrane once and has two extracellular Ig domains derived from the N-terminal region of the polypeptide. The CAR N-terminal Ig variable-type domain (D1) alone is necessary and sufficient for interaction with the fiber protein knob domain (13–15). CAR has a broad tissue distribution (16) and is found predominantly on the basolateral surfaces of epithelial cells (17, 18). CAR is localized specifically within tight junctions (19, 20), which hold lateral cell membranes in close apposition to form an effective seal close to the apical surface. Although interaction of viruses with receptors such as CAR that have low expression levels on cell apical surfaces appears paradoxical, recent studies suggest that this arrangement could facilitate the spread of virus infection within tissues and transmission of virus particles to new hosts, because progeny virus particles are released preferentially from the basal-lateral surfaces of infected cells (19, 21).
X-ray structures of recombinant knob domains from the fiber proteins of
adenovirus serotypes 2, 3, 5, and 12 have been determined
(22–25)
and in all cases show that knob monomers fold into similar 8-stranded
-sandwich conformations. Knob monomers assemble into remarkably stable
trimers (26), resulting from a
large number of hydrogen bonds and hydrophobic interactions in the noncovalent
trimer interface (23).
Variations in polypeptide chain length are accommodated by changes in the
length of surface loops, whereas other sequence variations appear to have had
minimal effects on the underlying -sandwich scaffold. Structures of CAR
D1 alone and in complex with the Ad12 knob also have been solved
(24,
27). CAR D1 has a
-sandwich fold that is characteristic of Ig variable domains
(28,
29). The x-ray structure of
the Ad12 knob-CAR D1 binary complex indicates that no conformational
rearrangement occurs in either molecule upon complex formation and that over
50% of the Ad12 knob-CAR D1 interfacial contacts involve residues within the
knob AB loop, with the knob DE and EG loops also providing additional contact
residues (24). Mutational
analysis of the CAR binding activities of Ad5 and Ad12 knobs is consistent
with the Ad12 knob-CAR D1 x-ray structure, suggesting that the location of the
CAR D1-binding sites on knob domains from different serotypes is conserved
(24,
30). The trimeric knob domain
has three identical binding sites for CAR D1 that can be occupied
simultaneously, as observed in the x-ray structure of the binary complex and
in surface plasmon resonance experiments
(24,
31).
Alignment of the Ad12 knob amino acid sequence with sequences of knob domains from other CAR-binding serotypes indicated that contact residues are not well conserved. For example, of the 15 total contact residues in Ad12 knob, only six are identical in the Ad5 knob, whereas eight are identical in the knob from Ad41L (32). This extensive variation of contact residues implies that the mechanism of knob-CAR interaction could differ significantly among adenovirus serotypes. Individual contact residues at structurally analogous positions might differ considerably in their relative contributions to binding affinity, as suggested recently in a comparison of the interaction of recombinant knob domains from Ad5 and Ad9 with CAR D1 (31, 32). One question arising from such studies is whether the relative positions of knob and CAR D1 in binary complexes changes significantly with the variation of contact residues in the knob component. This question could be addressed directly by solving crystal structures of knob-CAR binary complexes, but so far we have been unable to obtain diffracting crystals of binary complexes involving knob domains from any adenovirus serotypes other than Ad12, for reasons not presently understood. In this study, we investigated the structural basis for the difference in affinity between Ad2 and Ad12 knob for CAR D1 by first constructing homology models of Ad2 knob-CAR D1 binary complexes to predict important differences in contact residues between the two serotypes. Ad12 knob mutants were then constructed to test these predictions, both in functional binding assays and ultimately in crystal structure determination of binary complexes of the mutant knob proteins with CAR D1. Together, our results suggest that mutation of individual contact residues, as might occur during antigenic drift, in many cases only incrementally decreases affinity for the receptor. Partial retention of binding affinity may enable mutant viruses to successfully infect host cells and subsequently acquire secondary mutations that either restore high affinity binding or shift binding specificity to a novel receptor. Avidity effects resulting from the trivalent binding mechanism of knob may further compensate for mutations that weaken binding affinity, as suggested previously (31, 33, 34).
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Protein Expression and Purification—Recombinant knob domain from Ad12 fiber protein and the N-terminal domain of CAR (CAR D1) were prepared as described previously (15). The recombinant knob domain from Ad2 fiber protein was prepared following the same protocol used for Ad12 knob. Briefly, PCR products encoding the Ad2 and Ad12 fiber knob domains and several flanking amino acids from the fiber shaft (Ad2 knob residues 387–582 and Ad12 knob residues 401–587) were cloned between the NdeI and BamHI restriction endonuclease cleavage sites of vector pET15b (Novagen Inc., Madison, WI). The constructs were transformed into strain BL21-DE3 (Novagen Inc., Madison, WI) for expression of the hexahistidine-tagged knob proteins. All of the knob variants were constructed by primer-directed PCR mutagenesis (mutagenic primers: Ad12 knob P417S, GCAGTTTGGTGATGGGTCAGGAG; Ad12 knob S489Y, TCTATATCCCCAGTAAGCTTGAGGAA; Ad2 knob S408P, TGCAGTTAGGAGGTGGGTCTGGG; and Ad2 knob Y477S, TTCTAAAGTTCCAAGAATGTTTTTTAAGT) and cloned as above. The proteins were purified via ion exchange and nickelnitriloacetic acid affinity chromatography (Qiagen Inc., Valencia, CA), and the molecular weight of each construct was confirmed by mass spectrometry. Each soluble expressed knob protein formed stable trimers, as assessed by nondenaturing (native) polyacrylamide gel electrophoresis and SDS-polyacrylamide gel electrophoresis of unheated protein samples. PCR products encoding CAR D1 (CAR residues 21–143, numbering from the N-terminal signal peptide) with and without the S46C mutation were cloned between the NcoI and XhoI restriction endonuclease cleavage sites of pET15b. A stop codon was omitted from the reverse primer to extend the CAR D1 C terminus with a 22-residue peptide encoded by vector sequences. This peptide extension was found to increase folding of CAR D1 in the Escherichia coli cytoplasm (15). CAR D1 proteins were expressed in strain BL21-DE3 and were purified by ion exchange chromatography. CAR D1 mutation S46C was constructed by primer-directed mutagenesis, using the mutagenic oligonucleotide primer GTCTTCGGGACAAAGCGTAAATTTG.
Labeling of CAR D1 S46C—Serine 46 of CAR D1 was converted to cysteine by primer-directed mutagenesis, and the resulting protein was purified by ion exchange chromatography. CAR D1 S46C protein was covalently labeled with 6-iodoacetamidefluorescein through the introduced reactive thiol group of C46. The S46C protein (2.4 mg/ml) was reduced for 90 min in buffer A (20 mM NaPO4, pH 8.0, 200 mM NaCl) containing 20 mM dithiothreitol. Excess dithiothreitol then was removed on a Sephadex G-10 spin column equilibrated in buffer A. The reduced protein was collected directly into a receiving tube containing 1.5 times molar excess of 6-iodoacetamide fluorescein, and the mixture was incubated for 4 h at 25 °C in the dark with end-over-end tumbling. Excess 6-iodoacetamide fluorescein was removed by gel filtration on a Sephedex G-10 column (60 x 15 mm) in buffer A. The degree of labeling was determined spectrophotometrically by measuring the absorbance of the labeled protein at 280 and 492 nm. Wild type CAR D1 control protein remained unlabeled after treatment by this procedure, indicating that only the introduced cysteine group of the S46C mutant was reactive.
Fluorescence Anisotropy Experiments—Equilibrium binding
experiments were performed on a Tecan Ultra 384 plate reader (Tecan, Salzburg,
Austria) using flat bottomed, black, 96-well, untreated microplates (Nalge
Nunc Int., Naperville, IL). The reaction mixtures contained 5 nM of
fluorescein-labeled CAR D1 S46C in binding buffer (10 mM Tris-HCl,
pH 8.0, 0.1 mM EDTA, 0.0125% Nonidet P-40) with varying
concentrations of knob proteins (10 nM to 5 µM). The
total reaction volume was 300 µl, and the plates were incubated at 25
°C for 30 min. The fluorometer was set up with excitation and emission
wavelength filters of 485 and 535 nm, respectively, and with a fluorescein
dichroic mirror. The integration time was 40 µs, with 10 lamp
flashes/measurement. The gain and optimal Z-position were determined manually
using a control prior to the start of the experiment. The anisotropy values
were calculated as the ratio of the difference between vertical and horizontal
emission intensities (I// and I
)
normalized to the total intensity (Equation 1). Factor G is the ratio
of the sensitivities of the detection system for vertically and horizontally
polarized light.
 | (Eq. 1) |
The equilibrium binding curves were analyzed using Table Curve (SPSS Inc.,
Chicago, IL). Binding curves obtained for the knob-CAR interaction were fitted
to an A + B 
AB model that describes single-site independent binding of
ligand to the receptor. This model assumes that complexes form with no
multivalent avidity effects, because both the receptor and the ligand are free
in solution during the analysis.
Competitive displacement studies were performed by titrating 5
nM of labeled CAR D1 S46C to 
90% saturation with 650
nM knob domain from Ad2 fiber protein and then back titrating the
complexes with unlabeled wild type CAR D1 while measuring the decrease in
fluorescence anisotropy. The results of the concentration-dependent decrease
in anisotropy were fitted according to the equations of Huff et al.
(35). Because the
concentration of CAR D1 in the anisotropy assays (5 nM) was well
below the reported dissociation constant for CAR D1 homodimers (16
µM) (27), the
protein can be regarded as monomeric under the conditions used.
Modeling—Homology modeling of Ad2 knob in complex with CAR
D1 was performed using Swiss PDB Viewer
(36). The Ad12 knob-CAR D1
binary complex was used as a template (PDB code 1KAC
[PDB]
), and Ad2 knob was
superimposed on the structure, 1.04 Å root mean square deviation for
-carbon atoms. Visual inspection of this model indicated potential
contacts at the interface of Ad2 knob and CAR D1 that are not observed in the
Ad12 knob-CAR D1 structure. The residues at the complex interface that were
not conserved between Ad2 and Ad12 knob were modeled in the context of Ad12
knob in all rotomer conformations. Positive and negative potential
interactions were scored using the Swiss PDB Viewer mutation tool.
Crystallization—Crystals of purified knob-CAR D1 complexes were grown at room temperature by the sitting drop vapor diffusion method. All of the reagents used were obtained from Hampton Research (Hampton Research, Laguna Niguel, CA). Protein drops contained 2 µl of protein solution and 2 µl of reservoir solution, and the wells contained 500 µl of crystallization buffer. Crystals of Ad12 knob P417S-CAR D1 were grown in 3.2 M ammonium sulfate in 100 mM HEPES, pH 7.0. The crystals of the Ad12 knob S489Y-CAR D1 were grown in 0.8 M ammonium sulfate in 100 mM HEPES, pH 7.0. The crystals were flash cooled at 99 K with 50% ethylene glycol as a cryoprotectant.
Data Collection and Model Refinement—In each case, full data sets were collected from single crystals using a 4 cell CCD on National Synchrotron Light Source Beamline X25 at Brookhaven National Laboratory (Upton, NY). The data were processed using the HKL Program suite (37) as summarized in Table II. The protein coordinates of 1KAC [PDB] were subjected to rigid body refinement in CNS after a 5 Å region around the mutant site had been omitted. A cycle of torsion angle refinement at 5000 K was performed prior to initial Fo - Fc and 2Fo - Fc electron density map calculation. The electron density for the substituted amino acids could be assigned unambiguously. The final refinement statistics are shown in Table II. The coordinates for the P417S-CAR D1 and S489Y-CAR D1 structures were deposited in the PDB and assigned PDB codes 1P69 and 1P6A, respectively.
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Fluorescence Anisotropy Assay and Design of Knob Variants—The residues involved in the recognition of CAR D1 are not strictly conserved within the knob domains of fiber proteins from different adenovirus serotypes (Fig. 1). To assess the effects of this variation on binding affinity, we developed a fluorescence anisotropy-based assay that measures interaction of knob and CAR D1 in solution. This assay limits multivalent avidity effects that can arise when either knob or CAR D1 is immobilized on a two-dimensional surface, because it measures the affinity of individual binding sites on the trivalent knob molecule. For this assay, a reactive sulfhydryl group was introduced into CAR D1 to permit site-specific labeling with fluorescein. Serine 46 was chosen for conversion to cysteine based on its surface exposure and its location outside of the knob-binding surface of CAR D1 (Fig. 2C). Spectroscopic analysis showed that the CAR D1 S46C mutant protein was quantitatively labeled after reaction with 6-iodoacetamide fluorescein, whereas no incorporation of fluorescein was detected in control reactions containing wild type CAR D1 protein (data not shown). Formation of knob-CAR D1 complexes was monitored by the increase in fluorescence anisotropy upon incubation of fluorescein-labeled CAR D1 S46C with increasing, subsaturating concentrations of knob protein (Fig. 2A). To determine whether the fluorescein label influenced the interaction of CAR D1 with the knob proteins, a competitive displacement experiment was performed. Fluorescein-labeled CAR D1 S46C was titrated to
90% saturation with Ad2 knob protein, and the complexes were
then back titrated with wild type unlabeled CAR D1
(Fig. 2B). The data
were fitted to a displacement curve to determine the inhibition constant
(Ki). The dissociation constant
(Kd) of the forward titration and the
Ki of the back titration were equivalent to
within 3% error, indicating that the labeled CAR D1 S46C protein retained
binding activity similar to that of wild type CAR D1.
|
|
Using this anisotropy assay, we measured the binding affinities of two
different serotypes of adenovirus knob, Ad2 and Ad12, toward CAR D1.
Equilibrium binding affinities were derived from binding isotherms, as
reported in Table I. Ad2 knob
bound to CAR D1 with a Kd of 35 nM,
whereas Ad12 knob bound to CAR D1 with a Kd of
295 nM. To investigate the basis for this 8-fold difference in
affinity, we constructed a homology model (rigid body interaction) of the Ad2
knob-CAR D1 complex (Fig. 3).
Because the main chain conformations of Ad2 and Ad12 knob are similar
(-carbon atoms have root mean square deviation of 1.04 E), Ad2 knob was
built into the homology model in the same relative orientation as that of Ad12
knob in the crystallographically determined Ad12 knob-CAR D1 complex. No
steric clashes were observed at the binding interface, suggesting that the
homology model reasonably approximated the actual structure of the Ad2
knob-CAR D1 complex.
|
|
Six contact residues of Ad12 knob are strictly conserved in Ad2 knob, and all of these appear to contact CAR D1 in the homology model of Ad2 knob-CAR D1. Seven contact residues of Ad12 knob are conservatively substituted in Ad2 knob, and three of these Ad2 knob residues (Asn482, Thr486, and Thr507) appear to contact CAR D1, whereas three others (Ser442, Lys475, and Gln508) are more distant from CAR D1 and thus may contribute less to binding affinity. The final Ad2 knob residue in this group, Ser408, appeared to form an additional hydrogen bond with CAR D1 that is not observed in the Ad12 knob-CAR D1 complex. The remaining two contact residues of Ad12 knob (Leu426 and Glu523) correspond to nonconservative substitutions in Ad2 knob. Ad12 knob residue Leu426 forms a main chain hydrogen bond with CAR D1, and the corresponding Ad2 knob residue, Asn417, also is within hydrogen bonding distance from CAR D1 in the homology model. Ad12 knob Glu523 forms hydrophobic contacts with CAR D1, but the corresponding Ad2 knob residue (Thr511) is too distant from CAR D1 to make similar contacts. However, small rearrangements of the EG loop compared with its position in the model potentially could enable Thr511 to contact CAR D1. In summary, homology modeling suggested that 4 of the 15 residues of Ad2 knob that correspond to contact residues in Ad12 knob may not be oriented favorably to contact CAR D1. Consequently, we searched the homology model further for additional potential contact residues that could account for the greater affinity of Ad2 knob for CAR D1.
It was observed that one Ad2 knob residue in particular, Tyr477, was within hydrogen bonding distance of CAR D1 and that its side chain was able to make additional contacts at the complex interface. Tyr477 is positioned within a region that forms a cavity in the Ad12 knob-CAR D1 complex (Fig. 3). Modeling of tyrosine at the corresponding position in Ad12 knob, Ser489, indicated that a tyrosine residue at this position could be accommodated in the Ad12 knob-CAR D1 interface and could form two additional hydrogen bonds with CAR D1. Therefore, the overall results of homology modeling suggested that the increased affinity of Ad2 knob for CAR D1 could result from additional contacts contributed by Ad2 knob residues Ser408 and Tyr477.
To test the importance of Ser408 and Tyr477 in the
interaction of Ad2 knob with CAR D1, these residues were mutated to the
corresponding residues in Ad12 knob, proline and serine, respectively, and the
interaction of the resulting Ad2 knob mutants S408P and Y447S with CAR D1 was
characterized in the fluorescence anisotropy assay
(Fig. 4). The affinity of Ad2
knob S408P was decreased by approximately a factor of 2, whereas the affinity
of Ad2 knob Y477S was decreased 100-fold
(Table I). The converse mutants
also were constructed to determine whether additional contacts contributed by
these residues would increase the binding affinity of Ad12 knob. Equilibrium
binding studies showed that the resulting Ad12 knob mutants P417S and S489Y
both had increased affinity for CAR D1
(Table I and
Fig. 4), consistent with the
model that the side chains of both mutant residues make positive contacts with
CAR D1. A double mutant, introducing both point mutations into Ad12 knob,
further increased the binding affinity to 28 nM, greater than the
observed affinity of wild type Ad2 knob for CAR D1
(Fig. 4).
|
X-ray Structures of Ad12 Knob Mutants in a Binary Complex with CAR
D1—To further investigate the interactions that contribute to the
increased binding affinity of the Ad12 knob mutants P417S and S489Y, x-ray
structures of the two mutant proteins in binary complexes with CAR D1 were
solved to 3 Å resolution (Table
II). Data sets were essentially isomorphous to the native data
sets, resulting in corresponding structures that are very similar to the
native structure. Backbone atoms of wild type Ad12 knob-CAR D1 and Ad12 knob
P417S-CAR D1 were superimposable with a root mean square deviation of 0.1
Å, indicating that the overall main chain structures are identical and
that CAR D1 was bound in the same conformation and orientation in both the
mutant and wild type complexes. Backbone atoms of wild type Ad12 knob-CAR D1
and Ad12 knob S489Y-CAR D1 also were superimposable with a root mean square
deviation of 0.1 Å, again indicating identical conformations and
orientation of bound CAR D1. The -hydroxyl group of the serine residue
introduced into the AB loop of Ad12 knob (P417S) is within hydrogen bonding
distance of the O
1 atom of CAR D1 residue Glu56
(Fig. 5b) and thus
potentially could form an additional hydrogen bond within the knob-CAR D1
interface. No other conformational changes, with the exception of a rotation
of the side chain of CAR D1 residue Asp54 at the C atom, were
observed compared with the structure of the wild type binary complex,
suggesting that the increase in affinity of this mutant may be a direct
consequence of the serine residue at position 417. The crystal structure of
the Ad12 knob S489Y-CAR D1 complex indicated that the extended length of the
tyrosine residue side chain at residue 489 increased its potential to form
additional hydrogen bonds with CAR D1 that cannot form by the corresponding
serine residue in wild type Ad12 knob. The -OH group of knob residue
Tyr489 was within hydrogen bonding distance of the backbone oxygen
and nitrogen atoms of CAR D1 residues Pro52 and Ala125
(Fig. 6b). Rotation of
the side chain of CAR D1 residue Asp54 again was the only observed
change in conformation as compared with the wild type Ad12 knob-CAR D1
structure. The substantial increase in affinity of Ad12 knob mutant S489Y for
CAR D1 therefore likely results from formation of two additional hydrogen
bonds and burial of the tyrosine aromatic side chain.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Evolution of adenoviruses under immunoselective pressure has resulted in antigenic drift of surface residues, including those within the receptor-binding sites of the fiber protein knob domain. Despite this high degree of surface variation, many adenovirus serotypes attach to cells through interaction with CAR, a membrane glycoprotein that also serves as a cellular receptor for group B coxsackieviruses. Investigation of the interaction of fiber proteins from different adenovirus serotypes with CAR D1 therefore presents an opportunity to estimate the range of solutions to the molecular recognition problem that occur in a natural system. With the exception of Ad12 knob, it has so far not been possible to directly view the interaction of knobs from other adenovirus serotypes with CAR by x-ray crystallography. Here we circumvented this problem by using the x-ray structure of Ad12 knob-CAR D1 binary complex as a template to model contact residues on the surface of other knob serotypes and predict how binding affinity and specificity might be affected by variation of contact residues.
Ad2 knob was chosen for this initial modeling study because it binds to CAR D1 with significantly greater affinity than does Ad12 knob (Table I). Equilibrium binding studies were performed using fluorescence anisotropy measurements of CAR D1 labeled with fluorescein at a specific site. Affinity values calculated from the anisotropy measurements were lower (weaker) than have been previously observed in other in vitro assays, such as surface plasmon resonance (31, 32). These differences in affinity probably can be attributed to the absence of avidity effects in complex formation. In binding assays such as surface plasmon resonance, either the receptor or the ligand is immobilized, thereby limiting the association event to a two-dimensional surface. By contrast in the fluorescence anisotropy assay used here, both interacting species are in solution, and the affinity values therefore reflect individual binding events. Although the absolute affinity values obtained by fluorescence anisotropy probably underestimate the actual overall affinity of knob for membrane-associated CAR, the values obtained by this method are useful for characterizing the contribution of individual amino acid residues within the knob-CAR interface. Equilibrium binding affinity measurements using the fluorescence anisotropy assay showed that Ad2 knob bound CAR D1 with an 8-fold greater affinity compared with Ad12 knob (Table I). The increased affinity of Ad2 knob indicates that its receptor-binding sites overall make more favorable contacts with CAR D1 than do the binding sites of Ad12 knob. Binding site fitness could be improved either by increasing the number of favorable contacts or by eliminating contacts that interfere with stable binding.
The homology model of the Ad2 knob-CAR D1 complex predicted that two residues of Ad2 knob, Ser408 and Tyr477, could contribute additional hydrogen bonds and other contacts that may account for the greater affinity of Ad2 knob for CAR D1 (Fig. 3). These predictions were well supported by mutagenesis studies, where substitution of serine and tyrosine for the corresponding residues in Ad12 knob increased its affinity for CAR D1, whereas the converse mutations decreased the affinity of Ad2 knob for CAR D1. Prior studies using kinetic-based measurements indicated that Ad12 knob has a slower on-rate but similar off-rate compared with Ad2 and Ad5 knobs (31, 32). Although kinetic parameters cannot be determined from the equilibrium binding assay used here, it is likely that the P417S and S489Y mutations increase the association rate of Ad12 knob through the formation of more favorable interfacial contacts. The P417S and S489Y mutations are located in the Ad12 knob AB and DE loops, respectively, which is consistent with the earlier conclusion that individual residues within these two loops make substantial contributions to binding affinity (24, 30). Importantly, whereas several earlier studies have predicted contact residues in knob serotypes based on amino acid sequence alignments with Ad12 knob (14, 24, 30, 32, 38), here through the use of homology modeling we were able to identify a novel contact residue in Ad2 knob (Y477) that does not have a functionally equivalent counterpart in Ad12 knob. Substitution of tyrosine for the positionally equivalent residue in Ad12 (Ser489) resulted in an 8-fold increase in binding affinity.
To directly investigate the contribution of individual residues to binding affinity, we crystallized and solved the structures of two Ad12 knob mutants, P417S and S489Y, in complex with CAR D1. No changes in the relative positions of knob and CAR D1 components were detected in the resulting 3 Å resolution structures, as compared with the wild type complex, and no significant perturbations of the Ad12 knob structure were detected for either mutant. The structures support predictions from modeling studies that each substituted residue forms additional hydrogen bonds with CAR D1. Homology modeling also predicted that the protruding aromatic side chain of Ad2 knob residue Tyr477 would reduce the size of a cavity that forms in the Ad12 knob-CAR D1 interface (24), possibly excluding some nonstructural water molecules that become trapped in the Ad12 knob-CAR D1 complex. Burial of an aromatic side chain can play a critical role in stabilizing protein-protein interactions. For example, in the interaction of human immunodeficiency virus with its cellular receptor CD4, insertion of CD4 residue F43 into a cavity on the surface of gp120 accounts for 23% of interatomic contacts at the interface (39). The 100-fold decrease in Ad2 knob binding affinity resulting from mutation Y477S indicates that Tyr477 also makes a large contribution to binding affinity, although the precise role of the residue cannot be fully understood without determining a high resolution structure of the Ad2 knob-CAR D1 complex. The relatively smaller effect of tyrosine at the equivalent position in Ad12 knob (mutant S489Y) suggests that the contribution of individual contact residues to binding affinity may be context-dependent. At 3 Å resolution it was not possible to determine the impact of the S489Y mutation on the size and solvent content of interfacial cavities.
The high degree of variation in contact residues among CAR-binding
serotypes and the observed tight range of binding affinities (variations over
a 10-fold range have been measured by surface plasmon resonance
(31,
32) with the exception of Ad9
knob (32)) suggest that the
knob domain architecture may be specialized to minimize the impact of
mutations on binding site structure and function. The receptor binding sites
of knob consist of surface loops that are supported by an unusually stable,
trimeric scaffold, and in this regard have a similar architecture to the
antigen binding sites of antibodies. Although the range of molecular targets
recognized by knob domains of natural adenoviruses appears limited to a small
number of cellular receptors, this narrow range undoubtedly reflects the
strong selective pressure on the virus to retain sufficient binding affinity
to successfully infect host cells. It is probable that the natural evolution
of binding specificity is a multi-step process where the loss of affinity for
the initial receptor and gain of affinity for a novel receptor occur
simultaneously. Multivalent binding may compensate for relatively weak or low
affinity binding at individual sites, enabling virus to bind tightly enough to
infect host cells (31). In
addition, certain adenovirus serotypes have two independent fiber genes
(40), which may have provided
an alternate means for independent evolution of novel binding specificity.
Changes in fiber protein binding specificity during virus evolution likely
contributed to the overall success of adenoviruses as human and animal
pathogens. The utility of adenovirus-based vectors for use in gene therapy or
as recombinant virus vaccines also might be enhanced by rational modification
of fiber protein binding specificity. The availability of bacterial expression
systems for recombinant knob domain
(5) and of methods for
structure-guided mutagenesis and directed protein evolution
(41) now can be exploited to
investigate the potential of the knob scaffold to interact with a broad range
of molecular targets.
|
FOOTNOTES |
|---|
* This work was supported by Grant R01-AI36251 from the United States Public Health Service and by a grant from the United States Department of Energy Office of Biological and Environmental Research under Contract DE-AC02-98CH10886. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Present address: Dept. of Biological Sciences, Biophysics Section, Blackett
Laboratory, Imperial College London, Prince Consort Road, London SW7 2BW,
UK.
To whom correspondence should be addressed: Biology Dept., Brookhaven National
Laboratory, Upton, NY 11973. Tel.: 631-344-3350; Fax: 631-344-3407; E-mail:
freimuth{at}bnl.gov.
1 The abbreviations used are: CAR, coxsackievirus and adenovirus receptor;
D1, domain 1; Ad, adenovirus serotype; PDB, Protein Data Bank.
|
ACKNOWLEDGMENTS |
|---|
We thank Karen Springer for excellent technical assistance.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
- Rossmann, M. G. (1989) J. Biol.
Chem. 264,
14587-14590
[Abstract/Free Full Text] - Philipson, L., Lonberg-Holm, K., and Pettersson, U.
(1968) J. Virol.
2,
1064-1075
[Abstract/Free Full Text] - Levine, A. J., and Ginsberg, H. S. (1967)
J. Virol. 1,
747-757
[Abstract/Free Full Text] - van Oostrum, J., and Burnett, R. M. (1985)
J. Virol. 56,
439-448
[Abstract/Free Full Text] - Henry, L. J., Xia, D., Wilke, M. E., Deisenhofer, J., and Gerard,
R. D. (1994) J. Virol.
68,
5239-5246
[Abstract/Free Full Text] - Louis, N., Fender, P., Barge, A., Kitts, P., and Chroboczek, J.
(1994) J. Virol.
68,
4104-4106
[Abstract/Free Full Text] - Wu, E., Fernandez, J., Fleck, S. K., Von Seggern, D. J., Huang, S., and Nemerow, G. R. (2001) Virology 279, 78-89[CrossRef][Medline] [Order article via Infotrieve]
- Stevenson, S. C., Rollence, M., White, B., Weaver, L., and McClelland, A. (1995) J. Virol. 69, 2850-2857[Abstract]
- Defer, C., Belin, M. T., Caillet-Boudin, M. L., and Boulanger, P.
(1990) J. Virol.
64,
3661-3673
[Abstract/Free Full Text] - Tomko, R. P., Xu, R., and Philipson, L. (1997)
Proc. Natl. Acad. Sci. U. S. A.
94,
3352-3356
[Abstract/Free Full Text] - Bergelson, J. M., Cunningham, J. A., Droguett, G., Kurt-Jones, E.
A., Krithivas, A., Hong, J. S., Horwitz, M. S., Crowell, R. L., and Finberg,
R. W. (1997) Science
275,
1320-1323
[Abstract/Free Full Text] - Roelvink, P. W., Lizonova, A., Lee, J. G., Li, Y., Bergelson, J.
M., Finberg, R. W., Brough, D. E., Kovesdi, I., and Wickham, T. J.
(1998) J. Virol.
72,
7909-7915
[Abstract/Free Full Text] - Wang, X., and Bergelson, J. M. (1999) J.
Virol. 73,
2559-2562
[Abstract/Free Full Text] - Kirby, I., Davison, E., Beavil, A. J., Soh, C. P., Wickham, T. J.,
Roelvink, P. W., Kovesdi, I., Sutton, B. J., and Santis, G.
(2000) J. Virol.
74,
2804-2813
[Abstract/Free Full Text] - Freimuth, P., Springer, K., Berard, C., Hainfeld, J., Bewley, M.,
and Flanagan, J. (1999) J. Virol.
73,
1392-1398
[Abstract/Free Full Text] - Bergelson, J. M., Krithivas, A., Celi, L., Droguett, G., Horwitz,
M. S., Wickham, T., Crowell, R. L., and Finberg, R. W. (1998)
J. Virol. 72,
415-419
[Abstract/Free Full Text] - Walters, R. W., Grunst, T., Bergelson, J. M., Finberg, R. W.,
Welsh, M. J., and Zabner, J. (1999) J. Biol.
Chem. 274,
10219-10226
[Abstract/Free Full Text] - Zabner, J., Freimuth, P., Puga, A., Fabrega, A., and Welsh, M. J. (1997) J. Clin. Invest. 100, 1144-1149[Medline] [Order article via Infotrieve]
- Walters, R., Freimuth, P., Moninger, T., Ganske, I., Zabner, J., and Welsh, M. (2002) Cell 110, 789-799[CrossRef][Medline] [Order article via Infotrieve]
- Cohen, C. J., Shieh, J. T., Pickles, R. J., Okegawa, T., Hsieh, J.
T., and Bergelson, J. M. (2001) Proc. Natl. Acad. Sci.
U. S. A. 98,
15191-15196
[Abstract/Free Full Text] - Spear, P. G. (2002) Dev. Cell 3, 462-464[CrossRef][Medline] [Order article via Infotrieve]
- Xia, D., Henry, L. J., Gerard, R. D., and Deisenhofer, J. (1994) Structure 2, 1259-1270[Medline] [Order article via Infotrieve]
- van Raaij, M. J., Louis, N., Chroboczek, J., and Cusack, S. (1999) Virology 262, 333-343[CrossRef][Medline] [Order article via Infotrieve]
- Bewley, M. C., Springer, K., Zhang, Y. B., Freimuth, P., and
Flanagan, J. M. (1999) Science
286,
1579-1583
[Abstract/Free Full Text] - Durmort, C., Stehlin, C., Schoehn, G., Mitraki, A., Drouet, E., Cusack, S., and Burmeister, W. P. (2001) Virology 285, 302-312[CrossRef][Medline] [Order article via Infotrieve]
- Hong, J. S., and Engler, J. A. (1996) J.
Virol. 70,
7071-7078
[Abstract/Free Full Text] - van Raaij, M. J., Chouin, E., van der Zandt, H., Bergelson, J. M., and Cusack, S. (2000) Struct. Fold. Des. 8, 1147-1155[Medline] [Order article via Infotrieve]
- Harpaz, Y., and Chothia, C. (1994) J. Mol. Biol. 238, 528-539[CrossRef][Medline] [Order article via Infotrieve]
- Williams, A. F., and Barclay, A. N. (1988) Annu. Rev. Immunol. 6, 381-405[Medline] [Order article via Infotrieve]
- Roelvink, P. W., Mi Lee, G., Einfeld, D. A., Kovesdi, I., and
Wickham, T. J. (1999) Science
286,
1568-1571
[Abstract/Free Full Text] - Lortat-Jacob, H., Chouin, E., Cusack, S., and van Raaij, M. J.
(2001) J. Biol. Chem.
276,
9009-9015
[Abstract/Free Full Text] - Kirby, I., Lord, R., Davison, E., Wickham, T. J., Roelvink, P. W.,
Kovesdi, I., Sutton, B. J., and Santis, G. (2001) J.
Virol. 75,
7210-7214
[Abstract/Free Full Text] - Baranowski, E., Ruiz-Jarabo, C. M., and Domingo, E.
(2001) Science
292,
1102-1105
[Abstract/Free Full Text] - Wang, J. (2002) Trends Biochem. Sci. 27, 122-126[CrossRef][Medline] [Order article via Infotrieve]
- Huff, S., Matsuka, Y. V., McGavin, M. J., and Ingham, K. C.
(1994) J. Biol. Chem.
269,
15563-15570
[Abstract/Free Full Text] - Guex, N., and Peitsch, M. C. (1997) Electrophoresis 18, 2714-2723[CrossRef][Medline] [Order article via Infotrieve]
- Otwinowski, Z., and Minor, W. (1997) Methods Enzymol. 276, 307-326
- Kirby, I., Davison, E., Beavil, A. J., Soh, C. P., Wickham, T. J.,
Roelvink, P. W., Kovesdi, I., Sutton, B. J., and Santis, G.
(1999) J. Virol.
73,
9508-9514
[Abstract/Free Full Text] - Kwong, P. D., Wyatt, R., Robinson, J., Sweet, R. W., Sodroski, J., and Hendrickson, W. A. (1998) Nature 393, 648-659[CrossRef][Medline] [Order article via Infotrieve]
- Yeh, H. Y., Pieniazek, N., Pieniazek, D., Gelderblom, H., and Luftig, R. B. (1994) Virus Res. 33, 179-198[CrossRef][Medline] [Order article via Infotrieve]
- Patten, P. A., Howard, R. J., and Stemmer, W. P. (1997) Curr. Opin. Biotechnol. 8, 724-733[CrossRef][Medline] [Order article via Infotrieve]
- Thompson, J. D., Gibson, T. J., Plewniak, F., Jeanmougin, F., and
Higgins, D. G. (1997) Nucleic Acids Res.
25,
4876-4882
[Abstract/Free Full Text]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  N. M. Beach, R. B. Duncan, C. T. Larsen, X.-J. Meng, N. Sriranganathan, and F. W. Pierson Comparison of 12 turkey hemorrhagic enteritis virus isolates allows prediction of genetic factors affecting virulence J. Gen. Virol., August 1, 2009; 90(8): 1978 - 1985. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. C. Russell Adenoviruses: update on structure and function J. Gen. Virol., January 1, 2009; 90(1): 1 - 20. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Seiradake, H. Lortat-Jacob, O. Billet, E. J. Kremer, and S. Cusack Structural and Mutational Analysis of Human Ad37 and Canine Adenovirus 2 Fiber Heads in Complex with the D1 Domain of Coxsackie and Adenovirus Receptor J. Biol. Chem., November 3, 2006; 281(44): 33704 - 33716. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Seiradake and S. Cusack Crystal Structure of Enteric Adenovirus Serotype 41 Short Fiber Head J. Virol., November 15, 2005; 79(22): 14088 - 14094. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Marttila, D. Persson, D. Gustafsson, M. K. Liszewski, J. P. Atkinson, G. Wadell, and N. Arnberg CD46 Is a Cellular Receptor for All Species B Adenoviruses except Types 3 and 7 J. Virol., November 15, 2005; 79(22): 14429 - 14436. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Zhang and J. M. Bergelson Adenovirus Receptors J. Virol., October 1, 2005; 79(19): 12125 - 12131. [Full Text] [PDF]
|
|
|
|
 |
|
 M. Qin, B. Escuadro, M. Dohadwala, S. Sharma, and R. K. Batra A Novel Role for the Coxsackie Adenovirus Receptor in Mediating Tumor Formation by Lung Cancer Cells Cancer Res., September 15, 2004; 64(18): 6377 - 6380. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. P. Burmeister, D. Guilligay, S. Cusack, G. Wadell, and N. Arnberg Crystal Structure of Species D Adenovirus Fiber Knobs and Their Sialic Acid Binding Sites J. Virol., July 15, 2004; 78(14): 7727 - 7736. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. Awasthi, G. Meinken, K. Springer, S. C. Srivastava, and P. Freimuth Biodistribution of Radioiodinated Adenovirus Fiber Protein Knob Domain after Intravenous Injection in Mice J. Virol., June 15, 2004; 78(12): 6431 - 6438. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| All ASBMB Journals | Molecular and Cellular Proteomics |
| Journal of Lipid Research | ASBMB Today |