5534
`
`Immunology: Chien et al.
`
`Proc. Natl. Acad. Sci. USA 86 (I989)
`
`l-——->vH
`S 1 0 7 TGTCCCAATCTTCACATTCAGAAATCAGCACTCAGTCCTGTCACTATGAAGTTGTGGTTAMCTGGGTTTTI‘CT'l'I‘TAACACTTTTACATGGTATCCAGTGTGAGGTGAAGCTGGTGGM l 2 O
`U10
`I
`HVl
`'
`SI 0 7 TCTGGAGGAGGCTTGGTACAGCCTGGGGGTTCTCTGAGACTCTCCTGTGCAACTTCTGGGTTCACCTTCAGTGATTTCTACATGGAGTGGGTCCGCCAGCCTCCAGGGAAGAGACTGGAG 2 4 0
`U10
`HV2
`Mg
`S 1 0 7 TGGATTGCTGCAAGTAGAAACAAAGCTMTGATTATACMCAGAGTACAGTGCATCTGTGAAGGGTCGGTTCATCGTCTCCAGAGACACTTCCCAAAGCATCCTCTACCTTCAGATGAAT 3 60
`
`U10
`D
`l—’CH1
`|—> JH T831
`SI 0 7 GCCCTGAGAGCTGAGGACACTGCCATTTATTACTGTGCMGAGATTACTACGGTAGTAGCTACTGGTACTTCGATGTCTGGGGCGCAGGGACCACGGTCACCGTCTCCTCAGAGTCTGCG 4 80
`U10 Asp
`(GCT)
`Ala
`S 1 0 7 AGAAATCCCACCATCTACCCACTGACACTCCCACCAGTCCTGTGCAGTGATCCCGTGATMTCGGCTGCCTGATTCACGA'I'I‘ACTTCCCTTTCGGCACGATGAATGTGACCTGGGGAAAG 6 00
`U10
`
`S 1 0 7 AGTGGGAAGGATATMCCACCGTGAACTTTCCACCTGCCCTCGCCTCTGGGGGACGGTACACCATGAGCAGCCAGTTAACCCTGCCAGCTGTCGAGTGCCCAGAAGGAGAGTCCGTGAAA 7 2 0
`
`010
`
`S 1 0 7 TGTTCCGTGCAACATGACTCTAACCCCGTCCAAGAATTGGATGTGAATTGCTCTG
`U1 0
`
`775
`
`FIG. 3. Sequence comparison of the rearranged H-chain genes of U10 and 8107. The nucleic acid sequence includes a portion of the 5 ’
`untranslated region and the leader. The V3, D, JH, and CH] domains of the U10 and 8107 genes are shown labeled at the fist base of each region.
`The solid line representing identity of the two sequences is interrupted by the altered codon (in parentheses) and the resulting amino acid
`substitution in U10. The fust (HVl) and second (HV2) hypervariable regions and the Tan site that is lost due to the base change are also shown.
`
`experiments, suggesting that the loss of antigen binding is
`associated with the H chain (23). Thus, the only sequence
`difference between the $107 and U10 V and adjoining C
`region domains is the substitution of an uncharged Ala for
`negatively charged Asp-101 in the fifth residue of 1H1.
`Significant Conformational Change in the U10 Binding Site.
`To examine the impact of the U10 mutation on V region
`structure, we examined the reactivity of a number of anti-
`idiotypic antibodies with the 8107 and U10 antibodies. Al-
`though eight difl‘erent anti-T15 monoclonal antibodies, spe-
`cific for nonbinding site surfaces of the V domain (29), did not
`distinguish between the U10 and 8107 proteins, polyclonal
`antibodies to the antigen-binding site of 8107 did (Fig. 4).
`Binding of a polyclonal antiserum to 8107 could be inhibited
`by the addition of either $107 or U2U4, a single-site mutant
`of 8107 exhibiting a significant decrease but not a complete
`loss of P-Cho binding (30). In contrast, much higher concen-
`trations of the U10 protein only partially inhibited $107
`binding (Fig. 4). The same results were obtained with three
`difi‘erent monoclonal anti-binding site antibodies, suggesting
`a conformational change in the U10 binding site.
`Computer Modeling of U10. The three-dimensional struc-
`ture of the V region ofthe U10 antibody was analyzed starting
`from the two crystallographic structures of the SlO7-like
`antibody McPC603 (10, 31). One had P-Cho bound and the
`
`
`o—o—o—o—.\.
`
`\.\
`
`0.7%-
`
`._.—
`
`0.1. ;\
`/.’//°
`0°405nm 23
`7.7g
`
`”'6
`
`V
`
`0.2
`0.l
`
`\
`
`PC
`
`10'
`IO
`I0'5
`I0‘4
`:04 10'? ur'
`Inhibitor Concentration (pg/ml)
`
`.
`
`I02
`
`:03
`
`FlG. 4. Determination of a conformational change in the binding
`site of U10 by competition ELISA. The ability of U10 (0), 8107 (o),
`and U2U4 (A) proteins to inhibit the binding of a binding-site-specific
`polyclonal antiserum to 8107 protein was tested. Binding is ex-
`pressed in 0D units. At the concentrations used, the polyclonal
`antiserum is inhibited 100% by P-Cho (I) as indicated.
`
`other a sulfate ion. A model of the 5107 V domain was built
`from the sulfate-bound McPC603 structure (Fig. 5A) by
`incorporating appropriate amino acid substitutions. Calcula-
`tions of the electrostatic potential on the molecular surface of
`the hapten and the V regions of the McPC603 structure and
`the 8107 model reveal strong electrostatic complementarity
`between the bound P-Cho hapten and the antibodies. These
`results quantitate local electrostatic stabilization of the hap-
`ten suggested in analyses of the crystal structure (10). Res-
`idues contributing most to the local eleCtrostatic complemen-
`tarity with P-Cho were all in VH2 Arg-52 with the phosphate
`moiety and Glu-35 and Asp-95 with the positively charged
`trimethylammonium group (see Fig. 5A).
`In the McPC603 structures, a hydrogen-bonded salt bridge
`in Va between the side chains of Asp-101 and Arg-94 forms
`at the surface of the antibody, with both side chains extending
`away from the active site, as seen in the model of $107 in Fig.
`5A. Asp-101 is 11.8 A distant from P—Cho, and Arg-94 is 9.3
`A distant. Disruption of this salt bridge due to the Ala-101
`mutation of U10 could affect the conformation of Arg-94 such
`that the Arg-94 side chain could extend into the negatively
`charged pocket that binds the positively charged trimethyl-
`ammonium moiety of P-Cho in the nonmutated antibody.
`Rotation of the whole length of the Arg side chain by 180°
`would allow the terminal nitrogen atoms to extend to within
`2 A of the P—Cho nitrogen in the bound complex.
`To explore this hypothesis, two models of the V region of
`the U10 protein, termed ARG-IN and ARG-OUT, were built
`from the 8107 model by substitution of Asp-101 with Ala by
`side-chain truncation. Computer graphics followed by a
`combination of template forcing, molecular dynamics, and
`full energy minimization, including explicit solvent mole-
`cules, yielded stereochemically reasonable structures. In the
`ARG-OUT model, the side-chain positions were maintained
`as in the McPC603 structure, while in the ARG-IN model, the
`Arg-94 side chain was moved into the P-Cho—binding site
`(Fig. SB). Since Arg-94 and Asp-95 are adjacent on a [3-
`strand, bringing Arg-94 into the binding pocket caused the
`Asp-95 side chain to extend out to the surface away from the
`binding pocket.
`Root-mean-square deviations of both ARG-OUT and
`ARG-IN from McPC603 framework main-chain positions
`were well within the range observed in immunoglobuhn
`crystal structures (21), suggesting that main-chain conforma-
`tional changes made in the ARG-IN model to accommodate
`the Arg-94 side-chain repositioning are plausible. Superpo-
`sition of individual immunoglobulin VL and V" domains onto
`
`