| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Molecular Immunology Group, Tenovus Laboratory, Southampton University Hospitals Trust, Southampton, United Kingdom; and
Randall Center, King’s College London, Guy’s Hospital Campus, London, United Kingdom
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-strand A and the AVY motif (residues 23–25) on -strand B. In
this study we used mutagenesis to show that each of the strand
components of the hydrophobic patch is required for binding the I
carbohydrate Ag. In addition, the crystal structure reveals that amino
acids in the carboxyl-terminal region of CDRH3 form a
surface region adjacent to the hydrophobic patch. We propose that the I
carbohydrate Ag interacts simultaneously with the entire hydrophobic
patch in FR1 and with the outside surface of CDRH3. This
interaction could leave most of the conventional binding site available
for binding other Ags. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1
4GlcNAc) which are linear in the i structure and branched in
I (6). The strong association of a single
VH gene segment with specificity for the I/i Ag
indicated that the particular structure formed by the V4-34-encoded
portion of the Ab is required for I/i binding. It also suggested that
the structural interaction may be of an unconventional nature involving
framework regions (FR) and therefore independent of the conventional Ag
binding site (7, 8).
Staphylococcal protein A (SpA) binding to
VH3-encoded Ig was shown by mutagenesis to
interact with residues in FRH1,
FRH3, and the adjacent residues of
complementarity-determining region (CDR)H2
(9). SpA has been described as the prototypical B cell
superantigen that binds most members of the VH3
family without a L chain preference or influence of
CDRH3 (10). The full extent of this
interaction has recently been elucidated in a crystal structure of a
Fab/SpA complex (11), which rationalizes the
VH family binding specificity of SpA and shows
some similarity to the binding mode of a T cell superantigen to the
V FR of the TCR (12). Therefore, B cell superantigens
may be defined in structural terms as molecules that bind to a subset
of VH family-encoded Ig through interactions
outside of the conventional Ag-binding site.
It is usage of this single V4-34 gene segment that probably accounts
for the cross-reactive Id detected in early studies (13).
The Id was identified more clearly when the rat 9G4 mAb was found to
react with Igs encoded only by the V4-34 gene segment, regardless of D,
JH, or L chain sequence and isotype (14, 15). The Id is not detected on isolated H chains but is present
when the V4-34 H chain is paired with any L chain (14, 16). Mutational analysis revealed that the 9G4 Id is located in
FR1 of V4-34-encoded Ig (16). The Id is discontinuous,
involving W at residue 7 on -strand A (17) and AVY at
residues 23–25 on -strand B (16). There was an early
suspicion that FR1 might also be involved in recognition of the I Ag,
because the mAb inhibited RBC agglutination (14, 18). This was confirmed when replacement of FR1 of a CA with the
FR1 sequences from six VH families resulted in
loss of I specificity (19). All of the recombinant Abs
with FR1 replacements had the natural CA CDRH3
sequence. It was evident from these data that CA
CDRH3 in the absence of FR1 was not sufficient to
mediate I binding. However, a modulatory role of
CDRH3 in Ag recognition had been indicated by the
observations that Id-positive V4-34-encoded mAbs with specificity for
DNA were unable to bind the I Ag (20, 21). This binding
pattern appeared to be associated with basic amino acids in
CDRH3 that were apparently nonpermissive for I
binding. The role of CDRH3 in I binding was also
tested by mutational analysis (19). Converting the six
CDRH3 carboxyl-terminal amino acids of CA FS-6 to
alanines abolished I binding, while replacement of the amino-terminal
seven amino acids by alanines only moderately reduced binding
(19). These results implicated CDRH3
in I recognition.
The crystal structure of the Fab from the human I binding CA KAU
has been determined (22). The KAU H chain is encoded by
the V4-34-(D unidentifiable)-JH4 gene segments,
while the L chain is A11
(kv305,V
IIIb)-J4
encoded (3). There is only one amino acid difference at
position 31 in CDR1 (Gly (G) to Asp (D)) in KAU compared with the
germline V4-34 sequence. The reported analysis of the crystal structure
indicated that the KAU combining site has an extended cavity and a
neighboring pocket. Residues from the hypervariable loops
CDRH3, CDRL1, and
CDRL3 form the cavity, while the pocket is
defined by residues from CDRH1,
FRH2, and CDRH2. The
authors concluded that the nature of the site was consistent with
polysaccharide binding and proposed that the I Ag is binding via the
conventional combining site. Because residues in FR1 are clearly
outside this conventional combining site, they explained the previous
reports of inhibition of I/i binding by mAb 9G4 in terms of a
conformational change or steric blocking of the conventional site by
the anti-Id. However, such an explanation cannot account for the
inhibition of I binding by mutagenesis of individual residues in FR1
(Ref. 19 and this paper).
To resolve this paradox, we extend the previous mutational analysis and find that both the W at position 7 and the AVY motif at positions 23–25 are simultaneously required for I binding. On the basis of the crystal structure we find that these amino acids form an extensive hydrophobic patch in FR1. We are now able to correlate the crystallographic data with the molecular data in terms of the relative locations of FR1 and CDRH3 required for I binding, and we propose a superantigenic mode of recognition that involves an unconventional Ag binding site modulated by the structure of the CDRH3.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Recombinant viral stocks of Autographa californica nuclear polyhedrosis virus were produced and assayed in the Spodoptera frugiperda cell line Sf9 (ATCC CRL1711; American Type Culture Collection, Manassas, VA) grown in TC-100 medium (Invitrogen, Paisley, U.K.) supplemented with 10% (v/v) FCS plus 100 U/ml penicillin, 100 µg/ml streptomycin, and 0.25 µg/ml fungizone. Recombinant Abs were produced in Trichoplusia ni BTI-TN-5B1-4 (High Five cells; Invitrogen) grown in Ex-Cell 420 serum-free medium (AMS Biotechnology, Abingdon, U.K.).
Gene construction
Genes for the CA mAb FS-6 H and L chains were cloned from an immortalized human B cell line generated from a patient with CA disease (2). Recombinant FS-6 Ab and mutant FR1 (V4-39), in which the FR1 of FS-6 is replaced with the FR1 sequence of the V4-39 gene segment, were expressed in insect cells as previously described (19).
Mutagenic PCR conditions
All mutations to generate the FS-6(W7S) and FS-6(23–25TVS)
mutants were conducted using the Quick Change kit (Stratagene,
Amsterdam Zuidoost, The Netherlands) as per the manufacturer’s
instructions. The template was double-stranded FS-6 DNA isolated from
Escherichia coli strain XL-1-Blue (Stratagene) using the
QIAprep Miniprep kit (Qiagen, Crawley, U.K.). The PCR used Pfu turbo
DNA polymerase with mutagenic oligonucleotides. The sequences for the
mutagenic primers are shown in Table I
with the mutated bases underlined.
|
Transfection and isolation of recombinant viruses
Miniprep DNA was used to transfect Sf9 cells using linearized BacPAK6 viral DNA and Bacfectin (Clontech Laboratories, Basingstoke, U.K.) as per the manufacturer’s instructions. Viral supernatants were harvested at 3 days posttransfection.
Recombinant Ab production
Recombinant Abs were synthesized in BTI-TN-5B1-4 (High Five) cells grown in EX-CELL 420 serum-free medium by coinfecting cells with equal amounts of H and L chain-producing viruses at a multiplicity of infection of 5 for each virus. Three days postinfection, the supernatants were centrifuged at 1600 rpm for 5 min to remove cellular material. Tris-HCl (2 M; pH 8.8) was added to neutralize the supernatants to pH 7. The resulting precipitate was removed by centrifugation at 4000 x g for 10 min. One tablet of complete protease inhibitor mixture (Roche, Lewes, U.K.) was added per 20 ml of supernatant. The clarified supernatants were concentrated between 10- and 80-fold using Vivaspin concentrators (5000 MW; Sartorius, Epsom, U.K.). It has previously been determined that recombinant IgM Abs are predominantly produced as four-chain monomers and not as pentamers or hexamers in insect cells, even in the presence of J chain (19).
Quantitative ELISA
All recombinant Ig samples were first quantitated by capture
ELISA using known amounts of IgM as a standard. Relative concentrations
were calculated from the absorbances and dilutions of the samples.
Microtiter plate wells (Maxisorb; Fisher, Loughborough, U.K.)
were first coated with goat anti-human IgM (Sigma, Poole, U.K.) as
capture Ab and goat anti-human -chain Ab (Sigma) as
detecting Ab.
ELISA for 9G4 reactivity
The reactivity of the recombinant Abs with mAb 9G4 was measured by capture ELISA as previously described (19). Reactivities of recombinant Abs were measured using equivalent amounts of Abs.
I binding ELISA
Recombinant IgM monomers only weakly agglutinate RBCs in a hemagglutination assay. As an alternative to hemagglutination, we used an I-active extract from RBCs to determine the interaction between I Ag and Ab by ELISA. The I Ag was isolated by Dr. D. Roelcke (University of Heidelberg, Heidelberg, Germany), as previously described (19). This phenol/saline extract was shown to correlate with I and not i activity.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The germline amino acid sequence encoded by the V4-34 gene segment
compared with that of representative members of the
VH4 family is shown in Fig. 1. It can be seen that the amino acid
differences between V4-34-encoded VH and other
members of the VH4 family are primarily located
in FR1, CDR1, and CDR2. Of the nine amino acid differences in FR1, five
are common to all of the non-V4-34-encoded sequences which encode Igs
that neither express the 9G4 Id nor are involved in I/i binding
(14, 15, 16). Based on the fact that FR1 appears to be
critical for I binding (19), two mutant Ig were produced
by PCR mutagenesis, derived from FS-6 (Fig. 2). FS-6 is an IgM/ CA with
predominantly I specificity, based on a strong ability to agglutinate
adult erythrocytes and weak agglutination of cord erythrocytes
(2, 24, 25), and is unreactive with ssDNA, dsDNA, lipid A,
and IgG (21). The V4-34 gene segment is 96% homologous to
germline (2), and the net charge of the
CDRH3 is -2 (21). FS-6 had
previously demonstrated strong reactivity in ELISA using the I extract
(19). One mutant had W7 converted to S on -strand A,
while the other had the AVY(23, 24, 25) motif on -strand B converted to
TVS (Fig. 2). By mutating these residues to amino acids found in the
other VH4 family sequences, correct folding of
the VH domains was ensured. The recombinant Abs
were isolated from tissue culture supernatants and quantitated by
capturing with anti-µ chain Abs and detecting with anti-
chain Abs. This approach ensured that only H:L chain pairs were being
measured.
|
|
Reactivity of the recombinant Abs with the anti-Id rat mAb 9G4
was measured by capture ELISA using rabbit anti-rat IgG to coat the
wells and peroxidase-conjugated rabbit anti-human IgM Ab to detect
bound Abs. Mutation of either site in FR1 severely disrupted the
structure of the Id (Fig. 2
A), as demonstrated by the
significant reduction in 9G4 reactivity of both the W7S and
AVY(23, 24, 25)TVS constructs compared with FS-6. Some residual Id
expression was detected in the AVY(23, 24, 25)TVS construct compared with
the W7S construct. Hybrid Ig FR1(V4-39), in which the FR1 sequence of
V4-39 (Fig. 2) replaced the V4-34-encoded FR1 in the FS-6 CA, had
previously been constructed (19) and was used as a
negative control.
Ability of mutants W7S and AVY(23, 24, 25)TVS to bind the I Ag
I binding was detected by ELISA using wells coated with I Ag. The
reactivities of the recombinant Abs were conducted in the cold and
detected using peroxidase-conjugated goat anti-human IgM Ab. Fig. 2B shows that when either the W7 was mutated to S or the
AVY(23, 24, 25) motif was replaced with TVS, I Ag binding by both mutants
was severely reduced compared with the parental FS-6 recombinant Ab.
The reduced reactivities of the single mutants were comparable with the
hybrid Ig FR1(V4-39), in which both the W and AVY motif have been
mutated. These data indicate that both portions of the FR1 Id on
-strands A and B are critical for I Ag binding.
Structural modeling of FR1 and its proximity to CDRH3
An examination of FR1 in the crystal structure of the anti-I
CA KAU by Cauerhff et al. (22) reveal that the W at
position 7 and the AVY motif at positions 23–25 lie close together.
Furthermore, the side chains of W7 and the Y25 both lie flat along the
surface of the domain and are oriented in such a way that, together
with the side chain of A23, they form a continuous and very extensive
hydrophobic patch. This can be seen in Fig. 3A. The side chain of V24 is
totally buried in the hydrophobic core of the VH
domain and does not contribute to the patch. Further graphic
representations of the Ab/idiotope can be viewed in color at
http://www.gene-memes.com/VH4-34profile.html.
|
A. Fig. 3B shows an orthogonal view, looking
onto the conventional combining site, with all CDRs indicated. Thus, an
Ag binding to residues of FR1 could certainly contact the
outward-pointing residues of CDRH3. Some
CDRH3 sequences could therefore permit, or even
enhance, binding in this mode, while others could have side chains,
especially in the carboxyl-terminal region, that would prevent
interaction with the adjacent FR1 region. |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Mutational data from this current study indicate that when the amino acid at W7 is mutated to S or the AVY(23, 24, 25) motif to TVS, I Ag binding is severely reduced. It is extremely unlikely that mutation of these particular exposed residues will have any long-range structural effects, such as in the CDRs. Although it is now well documented that mutations of FR residues immediately adjacent to the CDRs can modulate Ag binding (26, 27, 28), there are no examples of FR residues further away having any such effect. The effects of these adjacent FR residues are readily understood in terms of local conformational effects upon residues in contact with Ag and, furthermore, their effects upon Ag binding affinity are relatively small (26). In contrast, residues W7, A23, and Y25 make no such contacts with other residues as they are exposed on the surface, and yet they have a considerable effect upon Ag affinity.
Further evidence for the fact that mutation of these FR residues is
unlikely to affect CDR conformation comes from the observation that the
conformational changes observed in Fab crystal structures upon
conventional Ag binding do not extend beyond the CDRs themselves
(although the relative alignment of VL and
VH can be affected) (29, 30).
Similarly, the crystal structures of superantigens bound to their
targets have shown that they do not induce major conformational
changes. Binding of SpA to VH3-encoded Fab did
not alter the conformation of the combining site (11), and
complex formation between Staphylococcus aureus enterotoxin
B and HLA-DR1 did not induce significant conformational changes in
either itself or DR1 (31). In addition, no major
conformational changes were observed in either the TCR -chain or
enterotoxin C3 upon complex formation (12). These data
indicate that there are no long-distance structural effects either by
Ag binding through the conventional combining site or by Ag binding
unconventionally outside of the combining site. Thus, we conclude that
the mutations of W7, and A23 with Y25, are directly responsible for the
reduced affinity for I Ag, and we infer from this that both FR1
-strands A and B are involved in this interaction.
Cauerhff et al. (22) argue that residues from the
V4-34-encoded hypervariable loops CDRH1 and
CDRH2 that form the wall of the pocket are
conserved in anti-I/i CAs. They report that the KAU
CDRH1 loop (residues 26–32 of the
VH region) folds with the group 1 canonical
structure. The residues at key sites for the H1 canonical loop
structure 1 are 24, 26, 27, 29, 34, and 94 (32). However,
residues 24, 26, 27, 34, and 94 are present in all members of the
VH4 family, as shown in Fig. 1, and are not
specific to V4-34-encoded Ig. Clothia et al. (32) indicate
that the amino acid at position 29 can be F, I, or V in canonical
structure 1. Although F is present only in V4-34-encoded Ig, canonical
structure 1 is maintained and is present in V4-34-encoded Ig with both
anti-I and anti-DNA specificities (21). Therefore,
it does not appear that any of the residues at key sites for the H1
loop structure can be critical for I binding.
In addition, V4-34 and VH4 family-encoded Ig have
different amino acids at positions 52 and 53 in the
CDRH2 loop (residues 52–56). The V4-34 gene
segment encodes NH at these positions, while they are either YY or YH
in the rest of the VH4 family (Fig. 1). However,
when the CDRH2 of a V4-61-encoded Ig replaced the
V4-34 CDRH2 sequence (Fig. 1), the NH was
replaced by YY and there was no effect on I binding (19).
This indicates that the amino acids NH at positions 52 and 53 are not
important for I binding. The CDRH2 exchange also
removed the potential N-linked glycosylation site at N52
(NHS), which is unique among the VH4 family to
V4-34-encoded Ig. Retention of the ability to bind I Ag indicated that
this glycosylation site is not a requirement for I binding. Because
there are no amino acid differences that distinguish V4-34 from the
other VH4 family members outside of FR1, it is
difficult to conclude that sequences other than FR1 are involved in the
V4-34 restriction among CAs.
When we analyzed the FR1 in the KAU anti-I crystal structure, we
discovered that W7 and AVY(23, 24, 25) lie immediately adjacent to each
other and form an extensive hydrophobic patch (Fig. 3A). The
nature of this hydrophobic patch, with aromatic ring structures lying
flat along the surface of the domain, and flanked by hydrophilic
residues capable of hydrogen bonding, is similar to that of the
oligosaccharide-binding hydrophobic patch on the C
2 domains of IgG
Fc (33). In this structure, phenylalanine rings lie flat
along the surface and the hexose residues of the oligosaccharide chains
stack on top of them, while surrounding residues provide hydrogen
bonding to the hydroxyl groups of the sugar rings. The limited database
of crystal structures of Abs complexed with carbohydrate Ags also
reveals hydrophobic stacking interactions and hydrogen bonding
(34, 35, 36).
If I binding occurs at the conventional combining site, as suggested by
Cauerhff et al. (22), it is difficult to see how the
hydrophobic patch in FR1 can also be involved, unless the Ag makes very
extensive interactions across a large region of the Ab surface; if the
latter is true, then the restriction to V4-34 is difficult, if not
impossible, to explain. The spatial relationship between the
conventional Ag binding site (defined by the CDRs) and the hydrophobic
patch in FR1 can be seen in Fig. 3A. Further examination of
the KAU crystal structure reveals a number of surface-exposed residues
of CDRH3 that point away from the conventional Ag
binding site and lie in close proximity to surface residues of FR1. In
particular, the exposed carboxyl-terminal residues of
CDRH3 lie adjacent to the hydrophobic patch
determinants W7 and AVY(23, 24, 25) (Fig. 3A), and it is
precisely these C-terminal residues that have been shown to affect I/i
binding (19). If the I Ag binds outside of the
conventional binding site and contacts the outer surface of the Ig V
domains, then both CDRH3 and FR1 can indeed
affect binding. This structural analysis provides further evidence that
the I carbohydrate Ag is binding to the V4-34-encoded Abs outside of
the conventional combining site, in a mode analogous to a superantigen,
and is consistent with our hypothesis that the shape and/or charge
characteristics of the surface-exposed CDRH3
residues adjacent to FR1 are able to influence the ability of the I Ag
to associate simultaneously with CDRH3 and
FR1.
The I/i Ags are relatively simple structures. Both are composed of N-acetyllactosamine, which is linear in the i structure and branched in the I structure. Hence, the I Ag has determinants in common with the i Ag (6). Most CA have some level of reactivity toward both I/i Ags and are assigned an I or i specificity based on the stronger reactivity (25). It is impossible to completely rule out that both the conventional site involving the pocket and the hydrophobic patch modulated by CDRH3 might both be used in binding the branched I Ag. However, this seems unlikely in view of the sequence restriction with I binding. Using Kabat’s functional classification of anti-carbohydrate binding sites, the pocket site is thought to interact most strongly with the terminal nonreducing ends of sugar chains and could potentially bind the nonbranched portion of the I Ag. However, the antigenic determinants recognized by anti-I and anti-i Abs are thought to be larger than trisaccharides (37).
It has been observed that V4-34-encoded anti-DNA Igs do not have anti-RBC reactivity (20, 21). This means that a V4-34-encoded Ig expressing the 9G4 Id and with an intact hydrophobic patch is not necessarily a CA. Anti-DNA Igs characteristically display positively charged residues in CDRH3, and it may be that these residues required for binding DNA inhibit binding to I/i by virtue of their proximity to the hydrophobic patch. While mAb 9G4 inhibits I/i binding, it does not inhibit ssDNA binding to V4-34-encoded Ig (21), further suggesting that DNA binding activity occurs in the conventional binding site. This concept is supported by the fact that D5Trp, a W7S mutant with a structurally disrupted hydrophobic patch, retains specificity for ssDNA (17).
The interaction of an autoantigen such as I/i with the CDRs, and in particular CDRH3, outside of the conventional binding site has a precedent in the crystal structure of the complex between rheumatoid factor RF-AN and IgG-Fc (38). This structure revealed a mode of binding in which Ag recognition was achieved through CDR residues that lay along one edge of, and pointed away from, the conventional binding site. Despite this unconventional site, CDRH3 residues dominated the interaction, as in virtually all Ab-Ag interactions. However, as a result of this binding to the side of the VH:VL domain pair, almost all of the conventional binding site remained free for the potential recognition of a different, as-yet-unidentified Ag.
Thus, the pocket and cavity described in the three-dimensional structure of KAU may not be intended for the I Ag. The ultimate proof for these interactions will be in the crystal structures of V4-34-encoded Abs combined with their Ags.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Kathleen N. Potter, Tenovus Laboratory, Southampton University Hospitals Trust, Tremona Road, Southampton SO16 6YD, U.K. E-mail address: kp1{at}soton.ac.uk
3 Abbreviations used in this paper: CA, cold agglutinin; FR, framework region; SpA, staphylococcal protein A; CDR, complementarity-determining region.
Received for publication March 12, 2002. Accepted for publication July 22, 2002.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-chain complexed with a superantigen. Nature 384:188.[Medline]
M globulins. Science 161:379.This article has been cited by other articles:
|
|
 |
|
  D. Capello, M. Martini, A. Gloghini, M. Cerri, S. Rasi, C. Deambrogi, D. Rossi, M. Spina, U. Tirelli, L. M. Larocca, et al. Molecular analysis of immunoglobulin variable genes in human immunodeficiency virus-related non-Hodgkin's lymphoma reveals implications for disease pathogenesis and histogenesis Haematologica, August 1, 2008; 93(8): 1178 - 1185. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Murray, N. Darzentas, A. Hadzidimitriou, G. Tobin, M. Boudjogra, C. Scielzo, N. Laoutaris, K. Karlsson, F. Baran-Marzsak, A. Tsaftaris, et al. Stereotyped patterns of somatic hypermutation in subsets of patients with chronic lymphocytic leukemia: implications for the role of antigen selection in leukemogenesis Blood, February 1, 2008; 111(3): 1524 - 1533. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Froyland, K. M. Thompson, S. J. Thorpe, S. S. Sahota, T. Gedde-Dahl, and B. Bogen A VH4-34+ myeloma protein with weak autoreactivity Haematologica, May 1, 2007; 92(5): 690 - 693. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Merlini and M. J. Stone Dangerous small B-cell clones Blood, October 15, 2006; 108(8): 2520 - 2530. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. A. Gertz Cold Hemolytic Syndrome Hematology, January 1, 2006; 2006(1): 19 - 23. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R J Looney, J Anolik, and I Sanz B lymphocytes in systemic lupus erythematosus: lessons from therapy targeting B cells Lupus, May 1, 2004; 13(5): 381 - 390. [Abstract] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and mutants FR1(V4-39) (
), W7S (•), and
AVY(23–25)TVS (
) with rat anti-Id mAb 9G4. B,
Reactivity of recombinant parental FS-6 and mutants with I Ag (symbols
as in A). Error bars indicate mean ± SD.