Mutational analysis of avidity and fine specificity of anti-levan antibodies

Using the polyfructose, bacterial levan, as a model polysaccharide, we analyzed how V regions affect binding in anti-polysaccharide mAbs. Previously, panels of mAb were constructed from bacterial levan-immunized BALB/c and CBA/Ca mice. The BALB/c mAb were mostly germline VHJ606:Vkappa11, and a subset contained presumed somatic mutations in the complementarity-determining regions (CDRs) that correlated with increases in avidity for the beta(2-->1) inulin linkage of levan. The CBA/Ca mAb were more heterogeneous in V gene usage, but a subset of inulin-nonreactive mAb were VHJ606:Vlambda and had VH sequence differences in the CDRs from the VHJ606 regions of the BALB/c mAb. In this report, VHJ606 Abs containing various combinations of specifically mutated H and L chains were produced by engineered transfectants and tested for inulin avidity and levan binding. Two presumed somatic mutations seen in CDRs of the BALB/c hybridomas were shown to directly cause marked increases in avidity for inulin (VH N53H, 9-fold; VL N53I, 20-fold; together, 46-fold) but not for beta(2-->6) levan. Exchange of either positions 50 or 53 in VH or the H3 loop between the BALB/c and CBA/Ca mAb resulted in either fine specificity shift or total loss of bacterial levan binding. Three-dimensional models of the V regions suggested that residues that affect binding to inulin alone are near the edge of the CDR surface, while residues involved with binding both forms of levan and affecting fine specificity are in the VH:VL junctional area.     

VH gene family utilization in colonies derived from B and pre-B cells detected by the RNA colony blot assay.

The immunoglobulin heavy chain variable region (VH) genes of the mouse have been categorized into families based upon sequence homology. Utilizing the RNA colony blot assay we have determined the expression of eight of these families in B cell colonies derived from either surface immunoglobulin positive (sIg+) adult spleen B cells or sIg- fetal liver pre-B cells. We demonstrate, based upon the analysis of greater than 6000 individual colonies, that VH gene usage is a characteristic of the mouse strain studied. C57BL/6 mice most frequently (45%) utilize family VHJ558, the largest VH family, whereas BALB/c mice most frequently (22%) utilize family VH7183, the most JH proximal family in BALB/c mice. Moreover, colonies derived from sIg- fetal liver derived precursors show similar patterns, suggesting that selection based on exogenous antigen is not an important parameter in determining VH gene family usage.     

Repertoire diversity of antibody response to bacterial antigens in aged mice. II. Phosphorylcholine-antibody in young and aged mice differ in both VH/VL gene repertoire and in specificity

Aging of mice is accompanied by both quantitative and qualitative changes in antibody responses to phosphorylcholine (PC), an immunodominant epitope of Streptococcus pneumoniae R36a strain (Pn). In order to study these changes at the molecular level, we generated PC-specific hybridomas from young (3 to 4 mo) and aged (20 to 24 mo) mice of different strains after primary immunization with S. pneumoniae R36a strain. These mAb were tested for Ig VH and VL gene family utilization, idiotopic repertoire, and cross-reactivity with unrelated Ag. Hybridomas from young mice (BALB/c, C57BL/6, and D1.LP) uniformly expressed the VH-S107 and V kappa-22 genes as well as most idiotopes of the T15 family, which were identified with different anti-T15 mAb. In contrast, the PC-reactive mAb from aged mice were quite heterogeneous: only 2 out of 13 utilized VHS107, 1 of 13 used VH7183, and 3 of 13 used VHJ558 gene family. Moreover, none of these mAb used L chain encoded by V kappa 22(0/13), but surprisingly they frequently expressed some of the T15 idiotope. In addition, the PC-binding mAb from aged mice showed broad cross-reactivity with various mouse and foreign proteins, whereas the mAb from young mice did not. These results demonstrate the genetic shift in antibody response of aging mice to PC, which is accompanied by a change in the antibody specificity. Interestingly, the qualitative repertoire change appears to be unrelated to the magnitude of antibody response, for the aged BALB/c mice maintain a very high reactivity to PC.     

Sequences of variable regions of hybridoma antibodies to alpha (1----6) dextran in BALB/c and C57BL/6 mice

The variable region sequences of light and heavy chains of three hybridoma antibodies to alpha (1----6) dextran, two from BALB/c and one from C57BL/6 mice, were determined by cloning and sequencing their cDNA. The three kappa-light chains are identical in nucleotide and amino acid sequences, except for the use of different J by BALB/c and C57BL/6; all three had the germ-line sequence of antibodies to 2-phenyloxazolone (20). Nevertheless, 2-phenyloxazolone BSA did not cross-react in gel with antidextrans, nor did dextran react with anti-2-phenyloxazolone ascitic fluids. The heavy chains differed, the BALB/c hybridomas having only three amino acid differences in CDR2 and two in CDR3; the C57BL/6 hybridoma differed throughout the variable region. All three VH are members of the J558 family. The three identical V kappa sequences suggest a significant role in dextran binding, with the differences in CDR of VH and the various J mini-genes of VL and VH being responsible for only fine differences in specificity. Alternatively, the role of V kappa might be minor, with most of the complementarity ascribable to VH. Additional sequences are needed to evaluate whether these data are typical of the repertoire of anti-alpha (1----6) dextran-combining sites.     

Amino acid substitutions in VH CDR2 change the idiotype but not the antigen-binding of monoclonal antibodies to alpha(1----6)dextrans

An idiotype defined by mAb and polyclonal antibodies to 10.16.1, an anti-alpha(1----6) dextran was previously reported to be expressed on most BALB/c anti-alpha(1----6)dextrans with groove-type sites and to involved CDR3 and probably CDR2. By comparing amino acid sequences of VH and VL derived from cDNA of idiotype+ and idiotype- anti-alpha(1----6)dextran hybridoma proteins, an idiotope was assigned to VH CDR2. Substitution of phenylalanine for leucine at residue 52 in CDR2 coupled with amino acid changes at either residue 58 or residues 57 and 60 abolished expression of this idiotype without affecting Ag binding.     

A molecular and structural analysis of the VH and VK regions of monoclonal antibodies bearing the A48 regulatory idiotype

The results presented in this paper explore the molecular basis for expression of the A48 regulatory Id (RI). A48 RI+ mAb derived from idiotypically manipulated mice molecularly resembled the A48 and UPC 10 prototypes of this system by utilizing a VHX24-Vk10 combination. Id expression by these antibodies was not restricted by a particular D region sequence, JH, or JK segment, but quantitative differences in Id expression were associated with utilization of different members of the VK10 germ-line gene families. The VL sequences of these A48 RI+ mAb has identified amino acid residues lying in four different idiotope-determining regions which may contribute to the structural correlate of this Id. A comparative sequence analysis of the VH regions of these VHX24 utilizing A48 RI+ mAb with several A48 RI+ mAb utilizing VHJ558 or VH7183 VH genes as well as a hybrid transfectoma antibody derived from two A48 RI-, VHJ558 utilizing hybridomas, all suggested that four nonconsecutive positions which lie outside the idiotope-determining regions may contribute structural elements toward expression of this Id. The VH and VL regions of the A48RI+, VHX24-Vk 10+ mAb showed low to moderate levels of somatic mutation which showed different patterns of distribution between the complementary determining region (CDR) and framework regions in the H and L chains. Although the VK sequences contained 50% of the replacement mutations in the CDR, with a replacement/silent mutation ratio of 10, the CDR of the VH sequences contained only 31% of the replacement mutations with a replacement/silent mutation ratio of 0.69.     

Variable region cDNA sequences and antigen binding specificity of mouse monoclonal antibodies to isomaltosyl oligosaccharides coupled to proteins. T-dependent analogues of alpha(1----6)dextran

Four mouse hybridomas specific for alpha(1----6)dextran, 16.4.12E (IgA kappa, C57BL/6), 28.4.10A (IgM kappa, BALB/c), 35.8.2H (IgG1 kappa, BALB/c), and 36.1.2D (IgM kappa, BALB/c) were obtained by immunization with the T-dependent Ag isomaltohexaose or isomaltotriose coupled to keyhole limpet hemocyanin or to BSA. Immunochemical characterization of the hybridoma antibodies showed that 16.4.12E and 36.1.2D had cavity-type combining sites, recognizing the terminal non-reducing end of alpha(1----6)dextran, whereas 28.4.10A and 35.8.2H had groove-type sites, recognizing internal linear segments of the dextran. The V region cDNA of the H and L chains of the antibodies were cloned and sequenced. VH of 16.4.12E and VH of 36.1.2D belonged to the X24 and Q52 germ-line gene families, respectively. The VH and V kappa sequences of 16.4.12E and V kappa sequence of 36.1.2D were highly homologous to those of W3129, the only anti-alpha(1----6)dextran mAb with a cavity-type site thus far sequenced; 16.4.12E differed from W3129 in the D, JH, and J kappa. VH genes of 28.4.10A and 35.8.2H were homologous to those of several anti-alpha(1----6)dextrans with groove-type sites, but belonged to the J558 germ-line gene family, differed from the other J558 anti-alpha(1----6)dextrans, probably representing a different germ-line subfamily. The L chain sequence of 28.4.10A encoded by V kappa-Ars and J kappa 2 was almost identical to other groove-type anti-alpha(1----6)dextrans obtained by immunizing with the T-independent glycolipid Ag, stearyl-isomaltotetraose. Use of T-dependent Ag such as isomaltosyl oligosaccharide-protein conjugates provides an additional parameter for probing the fine structure of antibody combining sites and evaluating the V-gene repertoire of anti-alpha(1----6)dextrans.     

Structure, map position, and evolution of two newly diverged mouse Ig VH gene families.

We have characterized two novel mouse VH gene families, VH3609N and VHSM7. These VH families have recently diverged from previously defined VH families. The VH3609N family, which may contain only one member in most inbred strains of mice, shares sequence similarity with the VHJ606 family and is located to the 3' side of VHJ606. VHSM7, with at least three members, is related to the VHJ558 family but maps 3' of VHJ558. These findings suggest that physical displacement of VH sequences may facilitate their subsequent divergence. During the early stages of VH gene family evolution that are exemplified by these new families, amino acid replacements have been selected against in frame-work regions and selected for in complementarity-determining regions. This pattern of nucleotide substitution appears to reflect evolutionary pressures to maintain germ-line VH diversity and, possibly, to select for new antibody specificities, as well as to select against mutations resulting in aberrant Ig. The classification of VH sequences with borderline similarity to previously defined VH families is discussed.     

Two families of monoclonal antibodies to alpha(1----6)dextran, VH19.1.2 and VH9.14.7, show distinct patterns of J kappa and JH minigene usage and amino acid substitutions in CDR3.

Nine groove-type mAb to alpha(1----6)dextran were cloned and sequenced. Together with previous reports from this laboratory, the VH and VL of 34 mAb have been sequenced, in which 10 VH19.1.2 and 11 VH9.14.7 combined with the V kappa-Ox1 gene to form two major families of anti-alpha(1----6)dextrans. The same D minigene (DFL16) was used by all VH19.1.2 and VH9.14.7 mAb; however, the patterns of JH and J kappa usage are quite different. VH19.1.2 mAb used only JH3 and J kappa 2, whereas VH9.14.7 mAb used three JH (JH1, JH2, and JH3) and all four active J kappa (J kappa 1, J kappa 2, J kappa 4, and J kappa 5). Relative uniformity in the lengths of VH CDR3 and the junctional sequences is seen in both families. Some mAb from different mouse strains share common structural features. The differences in idiotypic specificities and in the amino acid sequences suggest that VH19.1.2 and VH9.14.7 may differ in the conformation of CDR1 and CDR2. Combining with V kappa-Ox1 gene to generate groove-type combining sites to the single site-filling epitope of alpha(1----6)dextran, the two VH chains may require certain conformations of CDR3. Whether such conformational requirements influence the choice of J minigenes, the selection of the length of VH CDR3 and the sequences at junctions, are discussed.     

The Ig VH repertoire of fetal liver-derived pre-B cells is influenced by the expression of a gene linked to X-linked immune deficiency.

Ig VH repertoire differences between normal and x-linked immune deficiency- (xid) expressing mice are well established. To test the hypothesis that such differences might exist as early as the pre-B stage of ontogeny we generated panels of xid fetal liver derived Abelson murine leukemia virus transformants with H chain Ig VDJ rearrangements. Cells from CBA/Tufts.xid mice used VH genes from many families, with no demonstrable preference for 3' genes. Analysis of cells derived from (CBA/Tufts.xid X CBA/Tufts)F1 mice showed preferential usage of 3' family genes in the phenotypically normal females, even though V to DJ joins were made in vivo. The defective male mice did not show this marked preferential usage. A similar, but less marked, effect on VH gene usage was seen in mice with X-linked immune deficiency and a BALB/c background. Taken together, these results show that either X-linked immune deficiency, or a closely linked gene, affects fetal pre-B cells such that the usual pattern of predominant usage of 3' family genes is altered.     

The use of a battery of strains of mice in a factorial design to study the induction of dominant lethal mutations

The induction of dominant lethality following oral dosing of males with 200 mg/kg of cyclophosphamide was investigated using a factorial experimental design. Males from 3 genotypes, BALB/c, CBA/Ca and CBA/Ca X C57BL/6JF1 hybrid (CBB6F1) were mated to 6 females of the same genotype as the males over 3 weeks. Cyclophosphamide reduced the mating frequency of the BALB/c and CBA/Ca males. The total number of implants/female was reduced in all 3 genotypes with the greatest effect in the first 2 weeks after the males were treated. The proportion of early deaths/litter was significantly increased in CBA/Ca and CBB6F1 but the increase was smaller and non-significant with BALB/c. There was a high incidence (29.8%) of early deaths in the control BALB/c females. Statistical analysis of the ratio of early deaths to total implants in a litter using either the Freeman-Tukey binomial or the arc-sine transformation gave similar and satisfactory results. Analysis of early death data rather then the ratio of early deaths: total implants would have led to misleading conclusions. The implications of the use of a factorial design in dominant lethal assays for the detection of strain variation in mutagenic response without an increase in animal usage is discussed.     

Nucleotide and translated amino acid sequences of cDNA coding for the variable regions of the light and heavy chains of mouse hybridoma antibodies to blood group A and B substances

This is the first report of nucleotide and translated amino acid sequences of the variable region light (VL) and heavy (VH) chains of mouse monoclonal hybridoma anti-blood group A and B substances, the combining sites of which have been mapped. Monoclonal hybridoma anti-A and anti-B produced in BALB/c mice by immunization with A or B blood group substances, with A1 erythrocytes, and water-soluble blood group A substance or with synthetic B determinants coupled to bovine serum albumin or to O erythrocytes have been characterized immunochemically. To relate the immunochemical properties of the monoclonals to their primary structures, we have cloned and sequenced cDNAs of variable regions of light and heavy chains of two anti-A and two anti-B. The anti-A hybridomas have very similar combining site specificities and have almost identical VH sequences belonging to the J558 germ-line family, but their VL are from different germ-line VK gene families. The two anti-B hybridomas have different combining site specificities and use the same VL which differs completely from the anti-A VL; their VH are derived from different VH germ-line genes belonging to the J606 family. The results suggest that the heavy chains play a major role in determining the specificities of the antibody combining sites, with only minor contribution of VL. Additional sequence data on monoclonal antibodies of defined specificity for blood group substances are needed for further insights into the genetic and structural basis for their specificities.     

Organization of the murine immunoglobulin VH complex in the inbred strains.

Deletion mapping analyses have been employed to order the heavy chain variable region (VH) gene families in three inbred murine strains. These nine VH gene families have been positioned with respect to the J558 and 3660 VH families in A/J (Ighe) as follows: 3609-J558-(J606,VGAM3-8,S107)-3660-(X24,Q52,7183 )-DH. Maps generated with respect to J558 in the BALB/c (Igha) and C57BL/6 (Ighb) strains are consistent with these results. The organization of the VH complex produced by deletion mapping is quite different from the accepted map generated by other methods, particularly in that J558 is more DH distal and 3660 is more DH proximal than previously thought. The order presented here is compatible with VH rearrangement frequencies suggesting preferential utilization of DH-proximal VH gene segments. Our data also indicate that interspersion of some VH family members may be a common feature of the murine VH complex since the 3609 VH family is interdigitated in the three strains and a Q52 VH gene segment is interspersed in C57BL/6.     

Complete heavy and light chain variable region sequence of anti-arsonate monoclonal antibodies from BALB/c and A/J mice sharing the 36-60 idiotype are highly homologous

Structural and serologic studies on murine A/J monoclonal anti-arsonate antibodies resulted in the identification of a second idiotype family (Id36-60) in addition to the predominant idiotype family (IdCR). Id36-60, unlike IdCR, is a dominant idiotype in the BALB/c strain but is a "minor" idiotype in the A/J strain. The complete heavy and light chain variable region (VH and VL) amino acid sequences of a representative Id36-60 hybridoma protein from both the A/J and BALB/c strains have been determined. There are only four amino acid sequence differences between the VH of antibody 36-60 (A/J) and antibody 1210.7 (BALB/c). Two of these differences arise from single nucleotide changes in which the A/J and BALB/c Id36-60 VH germline gene sequences differ. The two other differences are the result of somatic mutation in hybridoma protein 36-60. In addition, Id36-60 heavy chains employ the same D and JH3 segments in both strains. The entire Vk2 VL of 36-60 and 1210.7 differ by only two amino acids, suggesting that like the heavy chains, they are derived from highly homologous VL genes. The same Jk segment is used in both antibodies. A comparison of the amino acid sequence data from Id36-60-bearing hybridomas suggests that a heavy chain amino acid difference accounts for the diminished arsonate binding by the 1210.7 hybridoma protein. Because the 1210.7 heavy chain is the unmutated product of the BALB/c VH gene, somatic mutation in VH may be required to enhance Ars affinity in this system.     

A comparison of the IgG isotype expression in the so-called thymus-independent immune responses to alpha(1----3) and alpha(1----6) dextran

This paper presents data on the IgG antibody response against two "thymus-independent" dextran (Dex) antigens from Leuconostoc mesenteroides, alpha(1----3) Dex B 1355S and alpha(1----6) Dex B 512F in BALB/c and C57BL/6 mice, which are considered to be responders or low responders to the respective antigen. The data point toward three common rules governing the two anti-Dex responses despite immunogenetic and antigenic disparities: (1) age dependency of the IgG isotype regulation of the response; (2) down-regulation of IgG isotype expression by T cells; and (3) individually determined preposition for IgG isotype formation in a given animal.     

Molecular basis of a mouse strain-specific anti-hapten response

The response of C57BL/6 and BALB/c mice to immunization with proteins coupled to (4-hydroxy-3-nitrophenyl)acetyl (NP) is dominated by distinctly different sets of antibodies. The VH gene family previously shown to be involved in the C57BL/6 response has now been shown to have highly homologous counterparts in BALB/c but of five sequenced BALB/c VH regions, none appeared likely to be able to encode an NP-binding protein. The active VH region from a BALB/c hybridoma making a characteristic anti-NP antibody was recovered and sequenced and shown to be quite different from the VH gene family involved in the C57BL/6 response. Comparison of the variation of the closely related VH regions between the two mouse strains showed that there are separate types of evolutionary pressures on the framework and complementarity-determining regions. The molecular basis for strain-specific immune responses appears to be that the structural divergence of VH regions between mouse strains is great enough that different strains use different VH regions for making the predominant class of antibodies to a specific hapten.     

Expression of a cross-reactive idiotope on naturally occurring murine antibodies against 3-fucosyllactosamine, levan, and dextran

We recently identified a cross-reactive Id (6C4) that is expressed on the H chain of many BALB/c mAb against the 3-fucosyllactosamine (3-FL) determinant, Gal(beta 1-4) (Fuc(alpha 1-3] GlcNAc-R. The VH segments of seven mAb that we recently sequenced are encoded by VH441, which also encodes VH segments of antibodies against galactan, levan, and dextran. To analyze the expression of the 6C4 Id on naturally occurring anti-carbohydrate antibodies, we isolated 6C4+ antibodies by affinity chromatography from pools of normal BALB/c serum. Approximately 20 to 30% of antibodies against 3-FL and levan, and all antibodies against dextran, were removed from the sera by passage over a column containing mAb 6C4. Absorption of the eluate with 3-FL beads removed anti-3-FL antibodies but not anti-dextran or anti-levan. The expression of a cross-reactive Id on naturally occurring antibodies against several carbohydrate Ag suggests that these antibodies may participate in an Id network. We also reported previously that BALB/c mice have naturally occurring anti-3-FL antibodies and respond well to immunization against this determinant, whereas C57BL/6 mice do neither. To examine the role of the Igh-C allotype in the regulation of the anti-3-FL response, we studied congenic strains of BALB/c and C57BL/6 mice. Both congenic strains produced anti-3-FL antibodies in response to immunization, but only C.B-20 mice exhibited naturally occurring antibodies. These data suggest that the naturally occurring and elicited antibody responses against 3-FL are differentially regulated.     

Allelic immunoglobulin VH genes in two mouse strains: possible germline gene recombination.

The nucleotide sequence of two germline immunoglobulin heavy chain variable region (VH) genes of mouse BALB/c origin was determined. These two genes are highly homologous to each other. They both have the unusual codon CCT for proline at position 7, which so far has been found only in a specific set of VH genes, called the NPb family. We show that the two VH genes belong to this set. One of our BALB/c genes, VH124, is more homologous to a C57BL/6 NPb VH gene than to any BALB/c VH gene, and we propose that these two genes are alleles. A comparison of the substitutions between these two genes with published sequences of all other BALB/c and C57BL/6 NPb VH genes reveals evidence for past homologous recombination events between related germline VH genes Homologous recombination may play an important role in the diversification of germline immunoglobulin VH genes.     

Northern hybridization analysis of VH gene expression in murine monoclonal antibodies directed to cancer-associated ganglioside antigens having various sialic acid linkages

The expression of the VH genes in 46 murine hybridoma cells that secrete mAb directed to the cancer-associated carbohydrate Ag, especially acidic glycolipids such as gangliosides and sulfated glycoplipids, was analyzed by Northern hybridization of poly(A)+ RNA of hybridoma with cDNA probes for nine VH gene families. Different hybridomas tended to express VH genes of the same family when the cognate Ag had the same or similar carbohydrate structures; i.e., the VH genes of the J558 family (group 1) were preferentially expressed in the mAb directed to various gangliosides that have NeuAc alpha (or NeuGc alpha) 2-3 and/or 2-8 linkage (71%), the most common linkage of sialic acid residues in the gangliosides of higher animals, and the hybridomas directed to sulfated glycolipids also expressed mainly the VH genes of the J558 family (80%). In contrast, the five mAb directed to various gangliosides with NeuAc alpha 2-6 linkage were exclusively encoded by the VH genes of Q52 family (group 2, 100%), and three antibodies directed to gangliosides with a NeuAc alpha 2-9 linkage all expressed genes of J606 family (group 6, 100%). The VH family usage was largely correlated with the linkage of sialic acid residues in the cognate carbohydrate Ag, but was not correlated at all with the difference in the fine specificities toward the core neutral carbohydrate chain, to which the sialic acid residues were attached. These findings suggest that the VH gene family in these anticarbohydrate antibodies is selected, depending primarily on the linkage of the sialic acid residues in carbohydrate Ag; these residues form the immunodominant sugar residue in the respective antigenic determinant.     

Human anti-V3 HIV-1 monoclonal antibodies encoded by the VH5-51/VL lambda genes define a conserved antigenic structure

Preferential usage of immunoglobulin (Ig) genes that encode antibodies (Abs) against various pathogens is rarely observed and the nature of their dominance is unclear in the context of stochastic recombination of Ig genes. The hypothesis that restricted usage of Ig genes predetermines the antibody specificity was tested in this study of 18 human anti-V3 monoclonal Abs (mAbs) generated from unrelated individuals infected with various subtypes of HIV-1, all of which preferentially used pairing of the VH5-51 and VL lambda genes. Crystallographic analysis of five VH5-51/VL lambda-encoded Fabs complexed with various V3 peptides revealed a common three dimensional (3D) shape of the antigen-binding sites primarily determined by the four complementarity determining regions (CDR) for the heavy (H) and light (L) chains: specifically, the H1, H2, L1 and L2 domains. The CDR H3 domain did not contribute to the shape of the binding pocket, as it had different lengths, sequences and conformations for each mAb. The same shape of the binding site was further confirmed by the identical backbone conformation exhibited by V3 peptides in complex with Fabs which fully adapted to the binding pocket and the same key contact residues, mainly germline-encoded in the heavy and light chains of five Fabs. Finally, the VH5-51 anti-V3 mAbs recognized an epitope with an identical 3D structure which is mimicked by a single mimotope recognized by the majority of VH5-51-derived mAbs but not by other V3 mAbs. These data suggest that the identification of preferentially used Ig genes by neutralizing mAbs may define conserved epitopes in the diverse virus envelopes. This will be useful information for designing vaccine immunogen inducing cross-neutralizing Abs.