Crystal Structures of Neuraminidase-Antihody
`Complexes
`
`W.R. TULIP,* J.N. VARGHESE,* R.G. WEBSTER,** G.M. AIR,t W.G. LAVER,* AND P.M. COLMAN*
`*C.S.l.R.O. Division of Biotechnology, Parkville 3052, Australia; **St. Jude Children's Research Hospital,
`Memphis, Tennessee 38101; tDepartment of Microbiology, University of Alabama, Birmingham, Alabama 35294;
`Wohn Curtin School of Medical Research, Australian National University, Canberra 2601, Australia
`
`The nature of antigen-antibody interactions has be(cid:173)
`come clearer in the last few years as crystal structures of
`antigen-antibody complexes have been elucidated.
`Such structures have permitted a visualization of the
`interface between antigen and antibody. Three-dimen(cid:173)
`sional structures of five complexes have now been re(cid:173)
`ported, two containing the influenza virus neuraminid(cid:173)
`ase as antigen and the Fabs NC41 and NClO (Colman
`et al. 1987, 1989), and three with hen egg-white
`lysozyme and the Fabs Dl.3, HyHEL-5, and HyHEL-
`10 (Amit et al. 1986; Sheriff et al. 1987b; E.A. Padlan
`et al., in prep.). The major parameters that define the
`scope of the antigen-antibody interface are emerging as
`more structures are solved and refined to greater preci(cid:173)
`topic
`this
`sion. Several reviews have dealt with
`(Mariuzza et al. 1987; Colman 1988; Davies et al.
`1988).
`Influenza is an orthomyxovirus and possesses an
`outer lipid membrane. Embedded in this membrane
`are two protein antigens, a hemagglutinin and a
`neuraminidase, the latter being a tetrameric glycopro(cid:173)
`tein comprising a head and an amino-terminal stalk that
`provides the attachment to the virus. Heads with full
`enzymatic and antigenic properties may be removed
`from the stalk with pronase and thus made soluble.
`Such heads have been used in all crystallographic inves(cid:173)
`tigations of the enzyme. The structure of neuraminid(cid:173)
`ase of N2 subtype is known (Varghese et al. 1983) and
`has a structure similar to that of subtype N9 (Baker et
`al. 1987). N9 neuraminidase is the antigen in both of
`the complexes discussed here.
`In this paper, we describe the crystal structure of the
`neuraminidase-NC41 Fab complex as refined at 2.9 A
`resolution and give a structural basis for the observed
`lack of binding of NC41 antibody with escape mutants
`of N9 neuraminidase. The main features of this com(cid:173)
`plex and the neuraminidase-NClO Fab complex are
`also compared with those of the lysozyme complexes.
`
`METHODS
`
`The isolation (Laver et al. 1984) and sequencing (Air
`et al. 1985) of N9 neuraminidase, production of mono(cid:173)
`clonal antibodies, and properties of the neuraminidase(cid:173)
`antibody complexes (Tulloch et al. 1986; Webster et al.
`1987) have all been reported previously. Sequences of
`antibodies were obtained from RNA (G.M. Air, un-
`
`Cold Spring Harbor Symposia on Quantitative Biowgy, Volume LIV © 1989 Cold Spring Harbor Laboratory Press 0-87969-057 • 7 /89 $1.00
`
`pub!.). The crystallization and structure determination
`were described for complexes containing two of these
`antibodies, namely, NC41 (Colman et al. 1987; Laver
`et al. 1987) and NClO (Air et al. 1987; Colman et al.
`1989). Diffraction data sets extending to resolutions of
`2.9 A and 3.0 A, respectively, were collected on film
`from a rotating anode source, and a 2.5 A synchrotron
`data set was obtained for the NC41 complex. Refine(cid:173)
`ment of these two crystal structures proceeded initially
`by the stereochemically restrained least-squares pro(cid:173)
`gram PROLSQ (Hendrickson and Konnert 1981). In
`later stages, the molecular dynamics program X-PLOR
`(Brunger 1988) was used . The NC41 structure will be
`further refined against the 2.5 A resolution data, and
`full details of the refinements will be presented
`elsewhere. Between successive rounds of refinement,
`the atomic models were rebuilt on an Evans and
`Sutherland PS300, using FRODO (Jones 1985) and
`- Fe electron density map (where F0 is
`displaying a 2F0
`the observed structure factor and Fe is the calculated
`structure factor). The mean positional error of the
`models was estimated by a statistical method (Luzzati
`1952). In the description of the structure, maximum
`distances for van der Waals contacts, hydrogen bonds,
`and salt links were as defined in Table 1 (see also
`Sheriff et al. 1987a). The molecular surface program
`MS (Connolly 1983) was used to calculate buried sur(cid:173)
`face area with a probe radius of 1.7 A. MS calculates
`the molecular surface area, rather than the solvent(cid:173)
`accessible surface area (Lee and Richards 1971; Con(cid:173)
`nolly 1983). Kabat numbering was used for the anti(cid:173)
`bodies (Kabat et al. 1987). Amino acid residue num(cid:173)
`bers are prefixed by L (light chain) and H (heavy
`chain). Complementarity-determining regions (CDRs)
`were defined as in Kabat et al. (1987), and their se(cid:173)
`quence numbers are light chain 24-34, 50-56, and
`89-97 and heavy chain 31-35, 50-65 , and 95-102.
`
`RESULTS
`
`NC41 Complex
`The R-factor (E i1F0I - IFJ IE IF0 1) for 20,065 reflec(cid:173)
`tions from 6.0 to 2.9 A resolution is 0.187 for the
`current model. Virtually all of the main-chain is ob(cid:173)
`served in continuous electron density, as are most of
`the side-chains, but the orientations of only about half
`25 7
`
`1 of 7
`
`BI Exhibit 1101
`
`

`

`~
`C1I
`00
`
`Table 1. Contact Residues in the Interface between Neuraminidase and NC41 Fab
`NC41 Fab
`
`light chain
`
`heavy chain
`
`S52 T53 H55
`
`156
`
`Y92
`
`S93
`
`P94 W96
`
`N31 Y32
`
`N52 N53
`
`E96 097 N98
`
`F99
`
`SlOOA
`
`LIOOB
`
`x
`
`x
`
`x
`
`x
`x
`
`x
`
`x
`
`x
`
`x
`
`x.
`x
`
`x
`
`x
`
`x
`
`x
`
`x
`x
`x
`
`x
`
`x
`
`x
`
`x
`x
`x
`x
`
`x
`
`x
`
`N9
`neuraminidase Y 49 W50
`x
`
`P326
`R327
`P328
`N329
`
`x
`x
`
`x
`
`x
`
`x
`
`x
`
`G343
`N344
`N345
`N347
`
`1366
`S367
`1368
`A369
`S370
`S372
`
`L399
`N400
`T401
`0402
`W403
`
`x
`
`x
`
`x
`
`P431
`x
`K432
`x
`0434
`x
`Residues are grouped in segments of polypeptide chain. Crosses indicate residue pairs that have at least one pair of atoms within van der Waals contact distance.
`Maximum contact distances (Sheriff et al. 1987a) are as follows, where C =carbon, N =nitrogen, 0 =oxygen: (CC) 4.107 A; (NN) 3.441 A; (00) 3.33 A; (CN) 3.774 A;
`(CO) 3.719 A; (NO) 3.386A. Residues partially buried from solvent by the interaction, but not in direct contact, are Thr-L31, Arg-L54, His-L91, Pro-L95. Thr-H28. Thr(cid:173)
`H30, Gly-H33, Trp-H50, Thr-H52A, Thr-H54, Glu-H56, and Asp-HlOl on the Fab, and lle-149, Asp-330, Asn-346, Thr-396, Arg-430, Glu-433, Asp-457, Pro-459,
`Lys-463, and lle-464 on the neuraminidase. In addition, two manoose residues. designated 2000 and 200F, which are in the carbohydrate attached to Asn-200 of a
`neighboring neuraminidase subunit, are partially buried.
`
`x
`
`x
`
`2 of 7
`
`BI Exhibit 1101
`
`

`

`CRYSTAL STRUCTURES OF NA-ANTIBODY COMPLEXES
`259
`of the peptide carbonyl oxygen atoms are unambigu(cid:173)
`Interactions between antigen and antibody include 1
`ously determined in electron density at this resolution.
`salt link from Lys-432 to Asp-H97, 10 hydrogen bonds,
`Six water molecules were included in the model with
`and 115 additional van der Waals contacts, although
`full occupancy. The root mean square difference be(cid:173)
`these latter numbers may change slightly after the high(cid:173)
`tween ideal and observed values is 0.022 A for bonds
`angle refinement. The buried surface area was calcu(cid:173)
`and 4.5° for angles, whereas the mean positional error
`lated to be about 878 A 2 on the surface of neuramini(cid:173)
`of the atoms was estimated to be 0.3 A.
`dase and 885 A 2 on the Fab. The nonpolar (carbon and
`Atoms at the interface between neuraminidase and
`sulfur) components of the total buried surface area of
`antibody are all in convincing electron density, with the
`neuraminidase and the NC41 variable module are 61 %
`exception of side-chains in the neuraminidase segment
`and 53%, respectively. These numbers are not signifi(cid:173)
`431-435. There appear to be no water molecules buried
`cantly different from those of the exposed surfaces of
`from solvent in the interface, but at least one water
`the two proteins (58% and 59% ), which implies that
`molecule is in contact with both antigen and antibody.
`the epitope and paratope are not particularly hydro(cid:173)
`The outline of the buried surface resembles a three-leaf
`phobic or hydrophilic.
`clover in shape, as shown in Figure 1, with the active
`The epitope on neuraminidase is extensive (see Fig.
`site in between two of the "leaves." A remarkable
`1; Table 1). It involves 5 segments on the upper surface
`shape complementarity exists over most of the inter(cid:173)
`of the enzyme, which contribute 22 residues in direct
`face, with protuberances on one protein matched by
`contact and 5 that are partially buried. In addition,
`depressions on the other. There is a large groove be(cid:173)
`there are partially buried atoms in 5 residues from 2
`tween the variable domains of the light and heavy
`other segments and in 2 sugar residues from the carbo(cid:173)
`chains (VL and VH) that accommodates a large ridge
`hydrate of a neighboring subunit.
`traversing the epitope made up of residues Asp-434,
`The binding site of the Fab is composed of 20 contact
`Lys-432, Ser-370, Ala-369, lle-368, and Asn-329. Dis(cid:173)
`residues and 12 partially buried residues (see Fig. l ;
`tances of 27 A, 30 A, and 27 A separate the tips of the
`Table 1). As CDR L1 is at least 5 A from neuramini(cid:173)
`three parts of the interface.
`dase, only 5 of the 6 CDRs are in direct contact; as a
`
`Figure 1. Stereoviews of the NC41 complex. (Top) N9 neuraminidase; (bottom) the NC41 VLVH dimer (light chain uppermost).
`The two proteins are shown opened out from a view, with the neuraminidase at left and the Fab at right, so that the interfaces are
`viewed en face. A Ca trace is drawn in thin Jines for noninteracting residues. Contact and partially buried residues are shown in
`toto by bold lines. (For pairs of residues in contact, see Table 1.)
`
`3 of 7
`
`BI Exhibit 1101
`
`

`

`260
`
`TULIP ET AL.
`
`result, the heavy chain has more interaction than the
`light chain. The comparative data for heavy and light
`chains, respectively, are 506 A 2 and 383 A 2 for the
`buried surface area; 6 and 5 for the number of ion pairs
`and hydrogen bonds; and 72 and 43 for the number of
`other contacts. Among the hypervariable loops, CDR
`H3 has the most buried surface area (267 A 2
`), repre(cid:173)
`senting 30% of the total, and it also has the greatest
`number of contacts (60). There are three framework
`region (FR) residues in the paratope, namely, Tyr-L49,
`Thr-H28, and Thr-H30, all of which are in segments
`adjacent to CDRs.
`
`Possible Conformational Changes in Antigen
`and Antibody
`Colman et al. ( 1987) reported that the structure of
`both the antigen and NC41 antibody may have changed
`as a result of antibody binding (for review, see Colman
`1988). The refinement of the NC41 complex is essen(cid:173)
`tially complete, whereas the refinement of uncomplex(cid:173)
`ed N9 neuraminidase (Baker et al. 1987) at 2.2 A
`resolution has not yet converged. A detailed com(cid:173)
`parison of the free and liganded neuraminidase struc(cid:173)
`tures will be possible in the near future. On the other
`hand, the structure of the free Fab has not been de(cid:173)
`termined; hence, there can be no direct comparison
`with the uncomplexed antibody. Comparisons with
`other Fab structures may suggest possibilities.
`At the level of local change, it appears that the
`main-chain conformations of five CDRs follow the
`canonical structures as predicted ( Chothia and Lesk
`1987; C. Chothia et al., in prep.). Hence, the local fold
`of each of these five loops is unlikely to have changed
`substantially upon the formation of a complex. How(cid:173)
`ever, CDRs can be flexible, for example, CDR L1 in
`Fab McPC603 (Satow et al. 1986) and CDR H3 in Fab
`19.9 (Lascombe et al. 1989). As the Fab engages an(cid:173)
`tigen, structural rearrangements might take place in
`both proteins. Such changes could be similar to those
`seen in the regions of crystal contacts (Wlodawer et al.
`1987) or in enzyme-inhibitor interactions (Greenblatt
`et al. 1989).
`On the basis of observed variability in VLVH interac(cid:173)
`tions in antibodies (Davies and Metzger 1983), it has
`been suggested that antigen might cause small rear(cid:173)
`rangements in the VLHH interface (Colman et al.
`1987). Whether or not this has happened in the NC41
`complex remains unclear. The analysis of that complex
`showed that its VLVH pairing was an outlier among
`other uncomplexed Fab structures. The angular differ(cid:173)
`ences in the VLVH pairing (see Table 1 in Colman et al.
`1987) of NC41 relative to New, McPC603, Kol, and
`J539 are currently 7.4°, 7.7°, 10.3°, and 11.5°. Given
`that such differences of up to 12.8° have since been
`found between uncomplexed Fabs (Lascombe et al.
`1989), NC41 can no longer be considered as an outlier.
`Direct evidence for this putative quaternary change can
`only come from comparison of complexed and uncom-
`
`plexed antibodies. A reported rotation of the side(cid:173)
`chain of Trp-H47 (Colman et al. 1987) has not been
`substantiated by the structure refinement. That tryp(cid:173)
`tophan now appears to have a canonical structure.
`
`Escape Mutants of N9 Neuraminidase
`Monoclonal antibodies directed against neuramini(cid:173)
`dase have selected a number of escape mutants of N9
`(see Tables 2 and 3 in Webster et al. 1987). Nine
`different mutants have been so selected, eight of them
`possessing single-site sequence changes in the contact
`surface with NC41. Figure 2 shows the positions of the
`nine mutants on a schematic diagram of neuraminidase.
`Several of these mutants crystallize isomorphously with
`wild-type N9, and their structures are being analyzed.
`Two, in particular, Oxl (370 Ser-Leu) and Ox2 (329
`Asn-Asp), have now been studied by difference
`Fourier methods, and preliminary results indicate local
`changes only for the Oxl substitution, and no structural
`change associated with the Ox2 substitution.
`The possible effect of the nine substitutions on the
`binding of NC41 antibody can be considered now that
`the structure of the interface is known in detail. The
`inhibition of neuraminidase activity by NC41 antibody
`was measured for these nine substitutions (see Fig. 2 in
`Colman et al. 1987). Wild-type neuraminidase is inhib(cid:173)
`ited, but there is no loss of activity for most of the
`mutants, indicating that the mutation has lowered the
`binding affinity of NC41. A decrease in activity, as a
`result of antibody binding, could be explained by two
`mechanisms. First, the access of substrate to and/or
`release of product from the enzyme's active site may be
`sterically hindered. Second, the active site could be
`
`Figure 2. Schematic diagram of one of the four subunits of
`neuraminidase viewed down the molecular fourfold rotation
`axis (bottom right). Arrows represent /3-sheet strands. Sites of
`mutation in N9 neuraminidase are numbered. All of the muta(cid:173)
`tions except 220 occur at residues that lie in the epitope
`recognized by NC41 Fab and reduce its binding affinity.
`
`4 of 7
`
`BI Exhibit 1101
`
`

`

`261
`A) and two ordinary contacts ( < 3.9 A) would be lost,
`including the only ion pair in the interface. The charge
`on Asp-H97 would then be unbalanced, even if a water
`molecule were bound in the hole.
`The only mutant that potentially preserves the shape
`complementarity but changes the chemical complemen(cid:173)
`tarity of the epitope is 329 Asn-+ Asp, designated Ox2.
`Its neuraminidase activity is decreased, indicating that
`NC41 does bind to Ox2 but not as strongly as to wild
`type. The complex between Ox2 and NC41 Fab crys(cid:173)
`tallizes isomorphously to wild type, and a data set has
`been collected from the crystals, but the model has not
`as yet been refined against it. A difference map of the
`wild-type complex with the Ox2 complex confirms CJUr
`earlier conclusion (Colman et al. 1987) that the binding
`of the antibody to these two antigenically distinct
`neuraminidases is isoteric. An X-ray refinement of the
`mutant complex is needed to confirm and clarify any
`small structural rearrangements that may be a con(cid:173)
`sequence of accommodating the aspartate at 329 in the
`interface.
`
`CRYSTAL STRUCTURES OF NA-ANTIBODY COMPLEXES
`deformed statically or dynamically. Recent studies
`(J.N. Varghese, unpubl.) have shown the orientation of
`sialic acid in the active site, which implies that the
`approach of substrate into the site requires the aglycon
`part of the substrate to be in the vicinity of the epitope
`recognized by NC41.
`Six of the nine substitutions replace smaller side
`chains by larger ones, namely, 367 Ser-+ Asn, 368
`Ile-+ Arg, 369 Ala-+ Asp, 370 Ser-+ Leu, 372
`Ser-+ Tyr, and 400 Asn-+ Lys. All of these substitu(cid:173)
`tions would be expected to diminish the shape com(cid:173)
`plementarity of the antigen-antibody interface leading
`to low binding affinity, and the results indicate that this
`is true for all but the 368 substitution, designated
`NC24Vl. Ile-368 is spatially near the middle of the
`epitope, and yet the mutation has little effect on the
`binding of NC41. An inspection shows that it is sitting
`at the base of a solvent-accessible pocket, which has
`both Fab and neuraminidase residues on its sides, and
`that an arginine side-chain could be accommodated if it
`followed the CG2 path of the isoleucine. This would
`first entail the loss of contacts made by CGl and CDl
`of the isoleucine and the creation of a hole where those
`two atoms were, and then the possible disturbance of
`other interacting residues by the arginine side-chain,
`such as Asn-329 and mannose-200F. Tentative support
`for such a placement of the arginine side-chain is given
`by a difference Fourier with (F0 b,NC24Vl - F 0 b,Nati)
`as coefficients, but the data are sparse and only extend
`to 3.8 A resolution.
`One mutant, 432 Lys-+ Asn, involves a decrease in
`the size of the side-chain, and although this would not
`produce a clash as above, the inhibition result indicates
`that the affinity of NC41 for that mutant is very low.
`This is not surprising because two good contacts ( < 3.0
`
`NCIO Complex
`At this stage, the refinement has not progressed
`significantly since the structure was reported (Colman
`et al. 1989). The R-factor for 12,674 reflections in the
`resolution range 6.0-3.0 A is 0.20, but segments of
`chain still remain out of density. In particular, the
`details of the interface, including buried surface area
`and numbers of contacts, are not clear. Nevertheless, it
`is apparent that the epitopes recognized by NClO and
`NC41 are largely overlapping (Colman et al. 1989). In
`addition, the mode of attachment of the two antibodies
`is completely different. As shown in Figure 3, the
`
`Figure 3. (Left) The tern N9 neuraminidase-NC41 Fab complex; (right) the whale N9 neuraminidase-NClO Fab complex. On
`both diagrams, the neuraminidase (lower left) and the VLVH dimer (upper right) are shown, with the neuraminidases in the same
`orientation. Ca traces of neuraminidase and heavy chains are indicated by thin lines, and light chains, by bold lines.
`
`5 of 7
`
`BI Exhibit 1101
`
`

`

`262
`
`TULIP ET AL.
`
`relative positions of VL and VH place equivalent CDRs
`a long distance from each other with respect to the
`antigen, and only four of the six CDRs on NClO are in
`contact with the antigen, namely, CDRs Ll, L3, H2,
`and H3. The similarity of the shape and charge distribu(cid:173)
`tion of the surfaces of the NClO and NC41 combining
`sites remains to be seen.
`
`DISCUSSION
`
`It appears that the neuraminidase-NC41 Fab inter(cid:173)
`face is somewhat more extensive than the interfaces
`between lysozyme and the Fabs Dl.3, HyHEL-5, and
`HyHEL-10. The surface area buried by complex forma(cid:173)
`tion is larger, and with the exception of HyHEL-10,
`there are more van der Waals contacts. However, the
`number of hydrogen bonds and salt bridges is about the
`same. The extent of the interacting surface is also
`reflected in the numbers of contacting and partially
`buried residues (see Table 2) . This result is achieved
`through contact with only five antibody CDRs because
`more of CDR L2 and CDR H3 are used in this case
`than in the lysozyme examples.
`There are two significant differences between the
`interfaces of the neuraminidase and lysozyme complex(cid:173)
`es. First, the NC41 and NClO complexes demonstrate
`that an antibody does not require the use of all six
`CDRs in the interaction to achieve the number of
`contacts and buried surface area needed to bind effec(cid:173)
`tively. In the NC41 complex, the only noncontacting
`CDR, Ll, is "hanging over" the active site . In this
`position, it could possibly interfere with the access of
`substrate to the enzyme. Second, the outline of the
`interface of the NC41 complex has a peculiar trefoil
`shape, which is partly due to the presence of the active(cid:173)
`site pocket and the fact that CDR L1 is not in contact.
`In contrast, the interfaces of the lysozyme complexes,
`
`which all have six CDRs in contact, have simpler out(cid:173)
`lines.
`An important similarity between the neuraminidase
`and lysozyme complexes is the presence of FR residues
`in the paratope. Berek and Milstein (1987) reported
`that somatic hypermutation occurred at framework
`sites in the maturation of the immune response against
`a hapten. There is a priori no reason to exclude FR
`residues from the antigen-binding site of an antibody.
`Their capacity to participate in such binding remains a
`property of the adjacent CDR surface.
`The epitopes in the two neuraminidase complexes
`are composed of more segments than the epitopes in
`the three lysozyme complexes, which is in contrast with
`the number of antibody CDRs in contact (Table 2).
`The highly discontinuous nature of the neuraminidase
`epitope is largely determined by the tertiary structure
`of the enzyme (Varghese et al. 1983; Baker et al. 1987),
`which has many short loops on its upper surface con(cid:173)
`necting the /3-sheet strands. The lengths of the
`neuraminidase segments in contact are correspondingly
`shorter than in the lysozyme complexes.
`Chothia et al. ( 1986) reported that in the lysozyme-
`01.3 complex, the observed main-chain conformations
`of four of the CDRs were close to their predictions.
`Similarly, predictions were made for the NC41 CDRs.
`Five hypervariable loops follow the canonical struc(cid:173)
`tures and are close to the predicted conformations (C.
`Chothia et al., in prep.). However, there are errors
`between 0.4 A and 3.0 A in the predicted spatial ar(cid:173)
`rangement of the CDRs with respect to each other.
`
`ACKNOWLEDGMENTS
`
`We thank Peter Tulloch for helpful discussions on the
`manuscript, Paul Davis for assistance with computing,
`and Bert van Donkelaar for technical support. Steven
`
`Table 2. Number of Contacts and Buried Surface Areas in Antigen-antibody Complexes
`
`Antigen
`
`segments
`
`NC41'
`NCIO'
`
`5
`5+1CHO
`
`residuesb
`CD Rs
`Neuraminidase
`22 /34
`5
`4
`?
`
`20/32
`?
`
`878/885
`?
`
`Antibody
`residuesb
`
`Buried surface
`area (A 2
`) '
`antigen I antibody
`
`Lysozymed
`6
`17/22
`16/27
`2
`690/680
`Dl.3
`14/24
`3
`HyHEL-5
`6
`17/28
`7501745
`4
`15/27
`HyHEL-10
`6
`19/30
`7741721
`' In conjunction with S. Sheriff, who provided invaluable help, the buried surface areas were calculated
`with the same parameters so that the numbers of buried residues and surface areas are directly
`comparable.
`bX / Y denotes that X residues are in direct contact and Y residues are fully or partially buried.
`Maximum contact distances are defined in Table I (also see Sheriff et al. 1987a), with the exception of
`the Dl.3 complex in which a 4.0 A cutoff was used.
`'Only segments in direct contact are counted.
`•References of structure reports: DI .3 (Amil et al. 1986); HyHEL-5 (Sheriff et al. 1987b; Davies et al.
`1988; HyHel-10 (Davies el al. 1988). Additional data for HyHEL-10 were kindly provided by E.A.
`Padlan and D.R. Davies prior to publication (E.A. Padlan et al., in pr~p.).
`
`6 of 7
`
`BI Exhibit 1101
`
`

`

`~~~-----------------. . !,·~
`
`CRYSTAL STRUCTURES OF NA-ANTIBODY COMPLEXES
`
`263
`
`Sheriff provided data for buried surface areas. This
`work was supported by a Commonwealth Postgraduate
`Research Award to W.R.T. and by U.S. Public Health
`Service research grants Al-21659, AI-08831, AI-20591,
`and AI-19084 from the National Institute of Allergy
`and Infectious Diseases.
`
`REFERENCES
`
`Air, G .M., L.R. Ritchie, W.G. Laver, and P.M. Colman .
`influenza
`1985. Gene and protein sequence of an
`neuraminidase with haemagglutinin activity. Virology
`14S: 117.
`Air, G.M., R.G. Webster, P.M. Colman, and W.G. Laver.
`1987. Distribution of sequence differences in influenza N9
`neuraminidase of tern and whale viruses and crystallization
`of the whale neuraminidase complexed with antibodies.
`Virology 160: 346.
`Amit, A .G., R.A. Mariuzza, S.E.V. Phillips, and R.J. Poljak.
`1986. Three-dimensional structure of an antigen-antibody
`complex at 2.8 A resolution. Science 233: 747 .
`Baker, A.T., J .N. Varghese , W.G . Laver, G.M . Air, and P.M.
`Colman. 1987. Three-dimensional structure of neura(cid:173)
`minidase of subtype N9 from an avian influenza virus.
`Proteins 2: 111.
`Berek, C. and C. Milstein. 1987. Mutation drift and repertoire
`shift in the maturation of the immune response. Immunol.
`Rev. 96: 23.
`Brunger, A.T. 1988. Crystallographic refinement by simulated
`annealing. Application to a 2.8 A resolution structure of
`aspartate aminotransferase. J. Mo/. Biol. 203: 803.
`Chothia, C. and A.M. Lesk. 1987. Canonical structures for
`the hypervariable regions of immunoglobulins. J. Mo/.
`Biol. 196: 901.
`Chothia, C. , A.M . Lesk, M. Levitt , A.G . Amit, R.A . Mariuz(cid:173)
`za, S.E.V. Phillips, and R .J . Poljak . 1986. The predicted
`structure of immunoglobulin D 1.3 and its comparison with
`the crystal structure . Science 233: 755 .
`Colman, P.M. 1988. Structure of antibody-antigen complexes:
`Implications for immune recognition . Adv. Immunol.
`43:99.
`Colman, P.M., W.G. Laver, J .N. Varghese, A.T. Baker, P.A .
`Tulloch, G.M. Air, and R.G. Webster. 1987. Three(cid:173)
`dimensional structure of a complex of antibody with in(cid:173)
`fluenza virus neuraminidase. Nature 326: 358.
`Colman, P.M., W.R. Tulip , J .N. Varghese, A.T. Baker, W.G.
`Laver, G.M . Air, and R .G. Webster . 1989. 3-D structures
`of influenza virus neuraminidase-antibody structures.
`Philos. Trans. R. Soc. Lond. B 323: 511.
`Connolly , M.L. 1983. Analytical molecular surface calcula(cid:173)
`tion . J. Appl. Crystallogr. 16: 548.
`Davies, D.R. and H. Metzger. 1983. Structural basis of anti(cid:173)
`body function. Annu. Rev. Immunol. 1: 87.
`Davies, D .R., S. Sheriff, and E .A. Padlan. 1988. Minireview:
`Antigen-antibody complexes. J. Biol. Chem. 263: 10541.
`Greenblatt, H.M., C.A. Ryan, and M.N.G. James. 1989.
`Structure of the complex of Streptomyces griseus protein-
`
`ase Band polypeptide chymotrypsin inhibitor-1 from Rus(cid:173)
`set Burbank potato tubers at 2.1 A resolution. J. Mo/.
`Biol. 20S: 201.
`Hendrickson, W.A. and J .H. Konnert. 1981. Stereochemically
`restrained crystallographic least-squares refinement of
`macromolecule structures. In Biomolecular structure, con(cid:173)
`formation, function, and evolution (ed. R. Srinivasan), vol.
`1, p. 43. Pergamon Press, Oxford.
`Jones, T.A. 1985. Interactive computer graphics: FRODO .
`Methods Enzymol. HS: 157.
`Kabat, E .A ., T .T. Wu, M. Reid-Miller, H .M. Perry, and K.S.
`Gottesman. 1987. Sequences of proteins of immunological
`interest, 4th edition . U.S. Public Health Service, National
`Institutes of Health , Washington, D.C.
`Lascombe, M.-B., P.M. Alzari, G . Boulot, P. Saludjian, P.
`Tougard, C. Berek, S. Haba, E.M. Rosen, A. Nishonhoff,
`and R.J. Poljak. 1989. Three-dimensional structure of Fab
`R19.9, a monoclonal murine antibody specific for the p(cid:173)
`azobenzenearsonate group. Proc. Natl. Acad. Sci. 86: 607.
`Laver, W.G., R.G. Webster, and P.M . Colman. 1987. Crystals
`of antibodies complexed with influenza virus neuraminid(cid:173)
`ase show isosteric binding of antibody to wild-type and
`variant antigens. Virology 1S6: 181 .
`Laver, W.G ., P.M . Colman, R.G . Webster, V.S . Hinshaw,
`and G .M. Air. 1984. Influenza virus neuraminidase with
`haemagglutinin activity. Virology 137: 314.
`Lee, B. and F.M. Richards. 1971. The interpretation of pro(cid:173)
`tein structures: Estimation of static accessibility. J. Mo/.
`Biol. SS: 379.
`Luzzati, V. 1952. Statistical treatment of errors in the determi(cid:173)
`nation of crystal structures. Acta Crystallogr. S: 802.
`Mariuzza, R.A., S.E.V. Phillips, and R.J. Poljak. 1987. The
`structural basis of antigen-antibody recognition. Annu.
`Rev. Biophys. Chem. 16: 139.
`Satow, Y., G.H. Cohen, E.A. Padlan, and D .R. Davies. 1986.
`Phosphocholine binding immunoglobulin Fab McPC603 .
`An X-ray diffraction study at 2.7 A. J. Mo/. Biol. 190: 593.
`Sheriff, S. , W.A. Hendrickson, and J .L. Smith. 1987a. Struc(cid:173)
`ture of myohemerythrin in the azidomet state at 1. 7I1. 3 A
`resolution . J. Mo/. Biol. 197: 273 .
`Sheriff, S., E.W. Silverton, E .A. Padlan, G.H. Cohen, S.J .
`Smith-Gill, B.C. Finzel , and D.R. Davies. 1987b. Three(cid:173)
`dimensional structure of an antibody-antigen complex.
`Proc. Natl. Acad. Sci. 84: 8(175 .
`Tulloch, P.A., P.M. Colman, P.C. Davis, W.G. Laver, R.G.
`Webster, and G.M. Air. 1986. Electron and X-ray diffrac(cid:173)
`tion studies of influenza neuraminidase complexed with
`monoclonal antibodies. J. Mot. Biol. 190: 215.
`Varghese, J.N., W.G. Laver, and P.M. Colman. 1983. Struc(cid:173)
`ture of the influenza virus glycoprotein antigen neura(cid:173)
`minidase at 2.9 A resolution . Nature 303: 35.
`Webster, R .G ., G.M. Air, D.W. Metzger, P.M. Colman, J.N.
`Varghese, A.T . Baker, and W.G. Laver. 1987. Antigenic
`influenza virus N9
`in an
`structure and variation
`neuraminidase . J. Virol. 61: 2910.
`Wlodawer, A., J . Deisenhofer, and R. Huber. 1987. Com(cid:173)
`parison of two highly refined structures of bovine pan(cid:173)
`creatic trypsin inhibitor. J. Mo/. Biol. 193: 145.
`
`7 of 7
`
`BI Exhibit 1101
`
`