HYBRIDOMA AND HYBRIDOMICS
Volume 22, Number 2, 2003
© Mary Ann Liebert, Inc.
Development of Functional Human Monoclonal Single-Chain
Variable Fragment Antibody Against HIV-1 from
Human Cervical B Cells
JODY D. BERRY,1,2,6 JOHN RUTHERFORD,2 GREGG J. SILVERMAN,3 RUPERT KAUL,4
MARIKKA ELIA,1 SARAH GOBUTY,1 ROBERTA FULLER,1 FRANCIS A. PLUMMER,2,5
and CARLOS F. BARBAS, III1
ABSTRACT
A panel of novel recombinant single-chain variable fragment (scFv) antibody against human immunodeficiency virus type-1 (HIV-1) was isolated and characterized. We generated human scFvs using RNA harvested
from cervical B lymphocytes of Kenyan prostitutes who are highly exposed to HIV-1, but remain persistently
seronegative. The variable regions of the heavy (VH) and light (VL) chain antibody genes were selected as
hybrids using guided-selection with the VL and VH, respectively, of a derivative of IgGb12 using the phagemid
vector pComb3X. IgGb12 is a previously well-characterized HIV-1 neutralizing human monoclonal antibody
(MAb). One of the hybrid scFv, IgA6/4L, neutralizes HIV-1 infectivity in in vitro cell culture assay. The cervical VH and VL chain antibody genes were connected by a DNA linker and subcloned in pComb3X. The
cervical scFv clones were functional in recognizing HIV-1 gp120 by enzyme-linked immunosorbant assay
(ELISA) and on cells in flow cytometry. Whole IgGb12 does not inhibit binding of clones IgA6/5k nor IgA6/30
to gp120, which suggests that they bind different epitopes. Nucleotide sequence analysis of the cervical scFv
show the clones are unique and reveal interesting characteristics of human cervical V gene pools. This work
demonstrates, for the first time, cloning of a functional scFv MAb to a sexually transmitted disease pathogen
from local cervical B-cell pools in exposed humans.
prostitutes(6) and discordant partners of HIV-1 infected individuals(7) also possess HIV-1 reactive IgA in the absence of detectable serum IgG, respectively. Recently, purified plasma and
mucosal IgA from Kenyan HEPS were shown to be neutralizing to a primary HIV-1 isolate in vitro(8) and to inhibit transcytosis of HIV-1 in an epithelial cell model.(9) Thus, mucosal
antibodies may contribute to the apparent broad HIV-1 resistance found in HEPS individuals.
Antibodies have played a useful role in defining virus–host
interactions and in revealing the antigenic structure of viruses.
As antibodies are the only host defense capable of completely
preventing an infection,(10) the prophylactic induction of locally
protective antibodies should therefore be an attractive goal for
INTRODUCTION
UMAN IMMUNODEFICIENCY
VIRUS TYPE-1 (HIV-1) has become the modern plague of mankind. The design of vaccines capable of preventing the transmission of HIV-1 requires
a deeper understanding of what is needed for an effective host
defense from retroviral invasion of the human immune system.
Of particular relevance, a group of epidemiologically resistant
Kenyan commercial sex workers have been shown to resist
HIV-1 infection despite frequent exposure.(1) These subjects
characteristically have HIV-1 specific cell-mediated immunity2–4 and HIV-1 specific mucosal IgA(5) without detectable
serum IgG reactive with HIV-1. Other exposed seronegative
H
1The
Skaggs Institute for Chemical Biology and the Department of Molecular Biology, The Scripps Research Institute.
of Medical Microbiology, University of Manitoba, Canada.
3University of California San Diego, and the Sam and Rose Stein Institute for Aging, San Diego, CA.
4Department of Microbiology, Oxford, UK.
5National Microbiology Laboratory, Health Canada, Canada.
6Present address: Monoclonal Antibody Section, National Centre for Foreign Animal Disease, 1015 Arlington St., Winnipeg, Manitoba, Canada.
2Department
97
98
HIV vaccines. Neutralizing monoclonal antibodies (MAbs) to
HIV-1 also have potential therapeutic value for the prevention
of infection after accidental exposures to the virus, and for the
interruption of maternal–fetal transmission of HIV-1. Moreover, these antibodies can be used to identify potential protective B-cell epitopes to facilitate vaccine design. The viral
neutralization sites for HIV-1 reside largely in the envelope protein.(11) The Envelope protein, which consists of gp120 and
gp41, has been shown to contain a variety of epitopes capable
of eliciting neutralizing antiviral antibody.
Mucosal antibody has been postulated to play an important
role in the acquisition of resistance to progressive HIV-1 infection. A likely explanation is that these IgA responses are acquired through sexual exposure to HIV-1, and because lymphocytes tend to traffic to sites of initial antigen encounter the
local cervical lymphocyte pool of HEPS should be enriched
with B cells producing antibody to HIV-1.(12) The demonstration of neutralizing IgA in cervical secretions, and the predominance of B cells in the local cervical mononuclear cell population,(13) would therefore make the cervical B-cell pool a
potentially valuable source of antibodies.
Of the human MAbs to HIV-1 that have been reported, the
IgG MAb b12 is among the most potent in virus neutralization
activity.(14,15) IgGb12 was selected from a combinatorial antibody library derived from a long-term asymptomatic HIV-1
seropositive patient, and the activities of b12 variants were subsequently improved with alternating rounds of in vitro mutagenesis and selection. IgGb12 recognizes a highly conserved
but relatively obscure epitope on gp120 shared by diverse isolates, and this antibody can neutralize primary HIV-1 isolates
in vitro.(16) Passive transfer of IgGb12 into hu-PBL-SCID mice
conveys protection from challenge with primary isolates of
HIV-1,(17) and protect macaques from vaginal challenge with
R5 isolates.(15) Based on these findings, we speculate that it
may be feasible to develop a vaccine capable of inducing antibodies with potency akin to IgGb12, that may act as a preformed protective mucosal immune barrier. To attain this goal,
the immunogenicity and immunodominance of protective epitopes on the HIV-1 envelope proteins on the mucosal surface
must be investigated.
Herein, we describe the isolation and characterization of a
panel of novel recombinant single-chain variable fragment
(scFv) antibody against HIV-1. The genes encoding these antibody binding domains were cloned from total RNA of cervical mononuclear cells, collected by routine noninvasive gynecological procedures. Thereby, we tested the feasibility of
selecting new recombinant MAbs from limited numbers
(1000–100,000 per sample) of lymphocytes. The libraries of
cloned cervical VH or VL genes were paired with the VL or
VH genes for a previously described variant of the human IgG
anti-gp120 MAb, b12, respectively. Novel antibodies specific
for gp120 were isolated from these hybrid libraries using a specially tailored phage-selection.(18) In these studies, the guided
selection strategy was designed for the section of hybrid MAbs
with broad reactivity for diverse HIV-1 envelope proteins now
found in low frequency in East-African isolates. Because cervical antibodies may contribute to protection against HIV-1 infection, the ability to study cervical B-cell pools will help to
identify candidate vaccine antigens that may be important for
the development of mucosal vaccines.
BERRY ET AL.
MATERIALS AND METHODS
Study subjects, cervical lymphocyte collection, and
RNA purification
Samples were collected from female commercial sex workers in the Pumwani area of Nairobi, Kenya, as described previously.(5) To collect mucosal lymphocyte samples, cytobrushes
were used during regular scheduled clinic visits. Immediate after collection, the cytobrush was placed into RNA lysis buffer
containing guanidium isothiocyanate (TRIreagent, Molecular
Research Diagnostics, Inc., Cincinnati, OH), and the RNA was
later purified using the Eppendorf 5-Prime phaselock gel tubes,
a procedure ideal for the isolation of small amounts of total
RNA from samples containing extraneous material. Each cytobrush sample contains between 1000–100,000 lymphocytes (R.
Kaul, personal observations). For the single-donor antibody library, a sample from the HEPS study subject, ML 556, was
used. For the pooled 12-person IgA selection experiments, samples from HEPS ML 767, 881, 889, 896, 1192, 1260, 1327,
1441, 1705, 1732, 1747, and 1832 were used. All of the HEPS
were screened by clinical tests for binding to clade B envelope
and were serum negative, and polymerase chain reaction (PCR)
negative (R. Kaul, data not shown). Only purified serum IgA
is measurable to the Env protein from HEPS, as direct serum
analysis for IgA binding from HEPS is negative. In contrast,
HIV-1 infected Kenyans have highly positive IgG binding to
clade B-envelope protein measurable directly from their sera (J.
Berry, data not shown).
Reverse transcriptase (RT)-PCR and Phagemid System
First strand cDNA was reverse transcribed from total RNA
using oligo-dt(15) primers and the Thermoscript cDNA kit
(GIBCO-BRL, Fredrich, MI). Due to the limited amount of
cDNA retrieved from each sample, the previously described oligonucleotide primers were pooled for the separate PCR amplification of human IgA heavy chain genes or light (kappa and
lambda) chain genes, respectively (Table 1). Two additional
primers were also used, that included the kappa forward primer,
HSCK4I-F (5-TGG CCC AGG CGG CCG GCG CGC CTG
AAC TCA CGC AGT CTC CAG GC-3), and the VH primer,
HSCVH31-FL (5-GGT GGT TCC TCT AGA TCT TCC TCC
TCT GGT GGC GGT GGC TCG GGC GGT GGT GGG GAG
CAG CTC GAG CAG TCT GGG GCT GAG GTG-3). In one
reaction, genes were amplified with the heavy chain (IgA1/A2)
back primer HSCA-B (5-CCT GGC CGG CCT GGC CAC
TAG TGA CCT TGG GGC TGG TCG GGG ATG C-3). In
contrast, attempts to amplify rearranged IgG heavy chain Vgenes from cervical cDNAs were unsuccessful, which supports
IgA as being a predominant locally expressed heavy chain class
in the cervix.
To generate composite genes for scFv antibodies, overlap
PCR was performed as described with human IgA1/2 and kappa
and lambda primer long-linker scFv sets(19,20) with the primers
HSC-F (5-GAG GAG GAG GAG GAG GAG GCG GGG CCC
AGG CGG CCG AGC TC-3) and RSC-B (5-GAG GAG GAG
GAG GAG GAG CCT GGC CGG CCT GGC CAC TAG TG3). Thereby, a scFv antibody library from ML556 was made
with VL genes fused via a synthetic linker to VH genes, which
were linked to the gene-encoding the C-terminal domain of the
OLIGONUCLEOTIDE PRIMERS
FOR
PCR AMPLIFICATION
CCG
CCG
CCG
CCG
CCG
CCG
CCG
CCG
CCG
TTT
TTT
TTT
TTT
TTT
GAT YTC CAC CTT GGT CCC-3
GAT CTC CAG CTT GGT CCC-3
GAT ATC CAC TTT GGT CCC-3
AAT CTC CAG TCG TCT CCC-3
CCT CTC CAG ATT GGT CCC CTG G-3
TGB TGA CGC AGC CGC CCT C-3
TGC TGA CTC AGC CAC CCT C-3
CCC TGA CTC AGC CTC CCT CCG T-3
AGC TGA CTC AGC CAC CCT CAG TGT C-3
TGC TGA CTC AAT CGC CCT C-3
TGC TGA CTC AGC CCC ACT C-3
TGG TGA CYC AGG AGC CMT C-3
TGC TGA CTC AGC CAC CTT C-3
GGC AGA CTC AGC AGC TCT C-3
ACC
ACC
ACC
ACC
ACC
TCG
TCG
TCG
TCG
TCG
TCA
TCG
TCG
TCG
ACC
ACC
ACC
ACC
ACC
AGC
AGC
AGC
AGC
AGC
AGC
AGC
AGC
AGC
GGA
GGA
GGA
GGA
GGA
AND
GGT
GGT
GGT
GGT
GGT
GGT
GGT
TCC
TCC
TCC
TCC
TCC
TCC
TCC
TCT
TCT
TCT
TCT
TCT
TCT
TCT
AGA
AGA
AGA
AGA
AGA
AGA
AGA
TCT
TCT
TCT
TCT
TCT
TCT
TCT
TCC
TCC
TCC
TCC
TCC
TCC
TCC
TCC
TCC
TCC
TCC
TCC
TCC
TCC
TCT
TCT
TCT
TCT
TCT
TCT
TCT
GGT
GGT
GGT
GGT
GGT
GGT
GGT
GGC
GGC
GGC
GGC
GGC
GGC
GGC
GGT
GGT
GGT
GGT
GGT
GGT
GGT
GGC
GGC
GGC
GGC
GGC
GGC
GGC
TCG
TCG
TCG
TCG
TCG
TCG
TCG
GGC
GGC
GGC
GGC
GGC
GGC
GGC
GGT
GGT
GGT
GGT
GGT
GGT
GGT
GGT
GGT
GGT
GGT
GGT
GGT
GGT
*The Sfi I site is underlined; B C/G/T, K G/T, M A/C, R A/G, S C/G, W A/T, and Y C/T.
Overlap Extension Primers
CSC-F (sense)
5-GAG GAG GAG GAG GAG GAG GCG GGG CCC AGG CGG CCG AGC TC-3
RSC-B (anti-sense) 5-GAG GAG GAG GAG GAG GAG CCT GGC CGG CCT GGC CAC TAG TG-3
IgA 3 anti-sense, back primer*
HSCA-B 5-CCT GGC CGG CCT GGC CAC TAG TGA CCT TGG GGC TGG TCG GGG ATG C-3
Heavy Chain forward
HSCVH1-FL 5-GGT
HSCVH2-FL 5-GGT
HSCVH35-FL 5-GGT
HSCVH3a-FL 5-GGT
HSCVH4-FL 5-GGT
HSCVH4a-FL 5-GGT
HSCVH31-FL 5-GGT
GGG
GGG
GGG
GGG
GGG
GGG
GGG
SCFV
ASSEMBLY
CAG GTG CAG CTG GTG CAG TCT GG-3
CAG ATC ACC TTG AAG GAG TCT GG-3
GAG GTG CAG CTG GTG SAG TCT GG-3
GAG GTG CAG CTG KTG GAG TCT G-3
CAG GTG CAG CTG CAG GAG TCG GG-3
CAG GTG CAG CTA CAG CAG TGG GG-3
GAG CAG CTC GAG CAG TCT GGG GCT GAG GTG-3
LONG-LINKER
V Lambda 3, anti-sense back primers
HSCJLam1236 5-GGA AGA TCT AGA GGA ACC ACC GCC TAG GAC GGT CAS CTT GGT SCC-3
HSCJLam4
5-GGA AGA TCT AGA GGA ACC ACC GCC TAA AAT GAT CAG CTG GGT TCC-3
HSCJLam57
5-GGA AGA TCT AGA GGA ACC ACC GCC GAG GAC GGT CAG CTS GGT SCC-3
V Lambda 5 sense, forward primers
HSCLamla 5-GGG CCC AGG CGG
HSCLam1b 5-GGG CCC AGG CGG
HSCLam2
5-GGG CCC AGG CGG
HSCLam3
5-GGG CCC AGG CGG
HSCLam4
5-GGG CCC AGG CGG
HSCLam6
5-GGG CCC AGG CGG
HSCLam78 5-GGG CCC AGG CGG
HSCLam9
5-GGG CCC AGG CGG
HSCLam10 5-GGG CCC AGG CGG
V Kappa 3, anti-sense back primers
HSCJK140-B 5-GGA AGA TCT AGA
HSCJK20-B
5-GGA AGA TCT AGA
HSCJK30-B
5-GGA AGA TCT AGA
HSCJK50-B
5-GGA AGA TCT AGA
HSCJK4L-B 5-GGA AGA TCT AGA
Kappa 5 sense, forward primers*
HSCK1-F 5-GGG CCC AGG CGG CCG AGC TCC AGA TGA CCC AGT CTC C-3
HSCK24-F 5-GGG CCC AGG CGG CCG AGC TCG TGA TGA CYC AGT CTC C-3
HSCK3-F
5-GGG CCC AGG CGG CCG AGC TCG TGW TGA CRC AGT CTC C-3
HSCK5-F
5-GGG CCC AGG CGG CCG AGC TCA CAC TCA CGC AGT CTC C-3
HSCK41-F 5-TGG CCC AGG CGG CCG GCG CGC CTG AAC TCA CGC AGT CTC CAG GC-3
TABLE 1.
100
filamentous phage protein pIII, enabling display on the surface
of filamentous phage following VCS M13 helper phage rescue,
as described.(19) The Kenyan VH IgA repertoire from the cervical lymphocytes of 12 HEPS subjects were cloned into the
pComb3X phagemid vector, also as described.(19) The result of
this approach is the display of a IgA/kappa/lambda polypeptides representing the product of scFv cDNA ligated in frame
with the C-terminus of pIII via directional Sfi I sites within the
pComb3X phagemid vector.(20)
Phagemid selection and scFv characterization
The H31/4L MAb is an affinity improved IgGb12 with several point mutations in the VH region and the parental light
chain of IgGb12, 4L. H31/4L neutralizes many TCLA and PI
strains of HIV-1. Briefly, via overlap-PCR with linker
oligonucleotides, the cDNA of VH and VL of the b12 mutant H31/4L Fab were PCR amplified and linked to cDNAderived cervical VL and VH genes, respectively. For the IgA
VH library, the repertoires in PCR-amplified cervical B lymphocyte cDNA from 12 HEPS subjects was pooled, and then
fused to the 4L VI cDNA, prior to cloning into the pComb3X
phagemid vector. Similarly, the HEPS cervical V kappa and
V lambda cDNA repertoire was PCR-amplified and fused to
the H31 VH cDNA and then cloned into pComb3X. A single round of affinity selection was performed with 15 wash
steps in the presence of 5 g/mL of native IgGb12 as a competitor, on wells coated with gp120IIIB protein, generously
provided by Dr. Pat Earl (National Institutes of Health,
Bethesda, MD). The gp120 binding phagemids were eluted
and expanded, as described,(19) and then used to infect F pilus
XL-1 E. coli (Stratagene, Cambridge, UK), and then plated
onto LB/carbenecillin (100 g/mL) plates. To assess the
binding characteristics of the scFvs, bacterial supernatants
from individual clones transformed into Top-10 E. coli (amber nonsuppressor-strain) were expressed without the pIII fusion protein and screened for binding to gp120SF2 in standard
EIA. To enable detection in both enzyme immunoassay (EIA)
and flow cytometric assays was via an engineered epitope tag
in the pComb3X vector an hemagglutin (HA) is fused to the
carboxy-end of the scFv molecules upstream of the amber
stop codon. Binding was detected with a high affinity rat antiHA tag antibody (Roche, Nutley, NJ), followed by a goat
anti-rat fluorescein isothiocyanate (FITC) or horseradish peroxidase (HRP) conjugated commercial antibody (Roche).
The recombinant and wildtype vaccinia strains were generously provided by Dr. Christopher Brodeur (Uniformed
Services University, Bethesda, MD). Log phase 293T cells
were infected at an multiplicity of infection (MOI) of 2 with
recombinant vaccinia expressing Ugandan (A) HIV-1 envelopes. Cells were harvested at 24 h by mechanical scraping, blocked in FACS buffer [1% bovine serum albumin
(BSA), 25 mM HEPES, phosphate-buffered saline (PBS) pH
7.2] and stained with antibodies at room temperature. Cells
were then treated with 2% paraformaldehyde to destroy any
residual infectious vaccinia virus. Data was acquired for
stained cells with a FACScan (Becton-Dickinson, Oxnard,
CA). The in vitro p24 neutralization assay was performed, as
described.(16)
BERRY ET AL.
RESULTS
Characteristics of the cloned human cervical
immunoglobulin repertoire
To assess the diversity of antibody genes cloned into the
phagemid vector from cervical lymphocytes, Alu I restriction
enzyme digestion was performed on individual phagemid
clones from the unselected combinatorial IgA library. This demonstrated diverse restriction fragments patterns indicating that
the library consisted of diverse variable region genes from the
mucosal lymphocytes (data not shown). To further characterize these plasmids, individual clones were then randomly selected for detailed sequence analysis using vector-specific
primers. The sequences show that diverse Ig V-genes regions
can be cloned from small samples of cervical lymphocytes.
Light chains of both kappa (Vk-1, and 3) and lambda (VLambda
1, 2, and 3) isotypes were identified among clones of the unselected mucosal library. Heavy chain IgA variable region
genes have highest identity with germline members of the VH1,
3, 4, and 5 VH gene families. Examination of several random,
unselected cloned cervical V-genes shows each is found only
once in this library, indicating that our primer sets amplify a
representative assortment of heavy and light chain variable
region gene cDNAs from cervical B cells (Fig. 1).
Selection of gp120 specific hybrid MAbs
Initial selection experiments were done with a ScFv
phagemid library produced from the cervical lymphocytes of
individual ML556. This library was biopanned for three rounds
upon a recombinant gp41 (Intracel, Cambridge, MA) and several Chlamydia trachomatis antigens. In all cases the ML556
library failed to show any specific enrichment, and individual
clones were nonbinding. We could not conclude that our library
was functional because there is no precedent for small cervical
antibody libraries based upon this failure. Furthermore, in retrospect, our failure to select gp41 specific MAbs from the scFv
library made from ML556 cervical lymphocyte cDNA could
have been predicted as ML556 is one of those rare resistant
prostitutes without measurable mucosal antibody to gp160 in
her secretions (R. Kaul, observations). Thus, we adopted the
guided selection strategy to test the feasibility of using cervical B cell RNA. Cervical VH and VL cDNA was recombined
with MAb H31-VH and 4L-VL to produce functional libraries.
Based upon cell counting experiments with other cytobrush
samples we estimate that the chain shuffled libraries from 12
individuals would have a maximum cellular diversity of less
than approximately 12 100,000 total lymphocytes. We rationalize that the use of the functional H31/4L binding domains,
which were optimized through prokaryotic expression and selection, would facilitates maximum sampling of the cervical antibody domains as hybrids and the b12 domains are highly expressed in E. coli. This could help to reduce any potential
deleterious effects of random pairings formed during the randomized VH and VL pairing inherent in this cloning procedure.(21) The use of the b12 variant was intentional as we hoped
to at least select new hybrid scFv with diversified gp120 binding reactivity should the guided selection succeed. The cervi-
CHAIN-SHUFFLING WITH CERVICAL V-GENES
101
FIG. 1. Diversity of unselected immunoglobulin variable genes recovered from the human cervix. Inferred amino acid alignment depicting the sequences of representative (A) light-chain and (B) heavy-chain variable regions recovered from prostitute
ML556. Light-chain variable regions are comprised of kappa and lambda chains of 4 different V1 families, and 4 different
kappa/lambda joining regions. Heavy-chain genes with homology to 4 different VH families, 5 DH, and 4 JH alleles were recovered. Percent homology to the closest germline V gene is shown in brackets. Numbering is according to the Immunogenetics database (http://imgt.cines.fr:8104), which was also used to find the homologues for VH, DH, and JH. The H-CDR3 length
varies from 8 to 21 residues in length indicating that an assortment of V genes can be cloned from B cells retrieved from the ectocervix with a single cytobrush sample. The inferred amino acid sequence of the heavy chain region corresponding to the IgA
antibody class and the back primer is underlined. aIMGT database CDR length determinations. bAlleles with highest similarity
from IMGT. Percent homology shown in parenthesis.
cal light chain shuffled library contained 3.3 107 hybrid
clones comprised of kappa and lambda isotypes with the H31
heavy chain. The cervical IgA heavy chain shuffled library contained 6.4 107 hybrid clones with the 4L light chain. Using
gp 120IIIB in solid phase, we selected phage-borne scFv from
both libraries.
Four novel binding clones were identified using enzymelinked immunosorbant assay (ELISA) (Fig. 2a). The hybrid
clones had comparable gp120 binding reactivity to the parental
H31/4L when expressed as scFv in bacterial supernatant. We
sequenced the VH and VL of these binding clones to confirm
the selection of new V-gene binders. Two cervical VH domains
(IgA6 and IgA9) were identified in combination with the
parental 4L and were expressed well in vitro. Two cervical VL,
clone 5-kappa and 30-lambda domains (signifying the light
chain isotype classes) were identified in scFv paired with the
parental H31 VH-gene. The molecular sequences of the cervical VH and VL domains selected in hybrid ScFv were determined and are shown in Figs. 3 and 4, respectively. Characterization of the variable region genes of the binding clones
revealed that they were encoded by diverse antibody variable
region genes and without evidence of clonal restriction. Of note,
the IgA6-VH was selected independently in several different
phagemid clones suggesting it was enriched in the eluted
102
BERRY ET AL.
FIG. 2. Identification of the selected hybrid ScFvs. (A) Clade B gp120SF2 was used to coat wells of a 96 well plate. Single
colony bacterial supernatants were tested in ELISA to identify binding clones for further analysis. Muck 4 is a previously selected single chain which binds to an irrelevant antigen. IgA6/4L, and IgA9/4L were selected from the cervical heavy chain shuffled library and have novel VH genes; H31/5kappa, and H31/30lambda were selected from the cervical light chain shuffled library and have novel light chain variable regions. The parental H31/4L was similarly expressed for a positive control. (B)
Demonstration of binding of purified hybrid ScFv on the gp120 oligomer expressed on cells in flow cytometry. Human 293T
cells were infected with recombinant vaccinia virus constructs expressing the gp160 molecule. Clone muck 4 is a ScFv to an irrelevant antigen and used to control specificity of binding (green fill). The parental H31/4L (green line) binds to an epitope in
the CD4 binding site which is present on both the clade A and B gp120 molecules. IgA9/4L [blue line] binds with a similar shift
as the H31/4L. IgA6/4L binds with two decade of shift to both gp120 molecules suggesting it binds to a conserved epitope present on these diverse gp120 molecules (red line). (C) In vitro neutralization of HIV-1 with monoclonal antibodies. Serial dilution’s of the mAbs were incubated with the HXB2 virus for 30 minutes at 370c before being transferred to PBLs. Viral growth
was estimated using standard P24 ELISA on the supernatant following 7 days growth (Burton et al., 1994). The OD is proportional to the relative amount of p24 produced at 7 days in human PBL culture. The assays were performed in triplicate and in a
blinded fashion. The whole IgG B12 and ScFvs H31/4L, and Muck 4 were used as controls. IgGb12; H31/41 scfv; muck
4; IgA6/4L; and IgA9/4L.
phagemid population. Because of the general importance of the
VH in determining antigen specificity,(22) the hybrid scFv molecules with new VH domains were further tested for binding to
the gp120 oligomer expressed on the surface of eukaryotic cells.
These hybrid molecules bind to cell-expressed HIV-1 envelope
of a Ugandan clade A strain in flow cytometry compared to a
control antibody (Fig. 2b). IgA6/4L in particular is capable of
good binding to the clade A gp120 molecule, with 2 decades
of binding, suggesting a new epitope is targeted on the envelope molecule.
Hybrid scFv were further assessed in in vitro HIV virus neutralization assay. Purified hybrid scFv from clone IgA6/4L neutralizes HXB2, in contrast to clone IgA9/4L, which did not produce significant neutralization relative to control antibodies (Fig.
CHAIN-SHUFFLING WITH CERVICAL V-GENES
103
FIG. 3. Nucleotide sequences of the selected heavy-chain variable regions. The sequences were compared to the known immunoglobulin sequences in the databanks on the worldwide web. The nucleotide and inferred amino acid translations of IgA6 VH
(upper) and IgA 9 VH (lower) and depicted in alignment with the closest matching germline alleles for human VH, DH, and JH. A
DH germline match for IgA9 was not identified by the ImmunoGenetics program. The nucleotide sequences for IgA6 and IgA9 have
been submitted to GenBank (accession numbers AY044149 and AY044150, respectively). Alignments and CDR delineation were
performed according to LeFranc et al.(8) and the ImmunoGenetics databanks. “” denotes identity, and “” denotes a deletion.
2c). The H31/4L ScFv neutralizes this T-cell line adapted (TCLA)
isolate very well, which is consistent with the improved affinity
of H31/4L for this envelope compared with the original b12 antibody. These data show that a hybrid cervical scFv MAb IgA6/4L
neutralizes HXB2 in vitro, and displays unique reactivity for an
envelope epitope common to isolates from clades A and B.
Characterization of the combined cervical MAbs
The cervical antibody binding domains were tested for gp120
binding activity when cervical VH cDNA is recombined with cer-
vical VI cDNA (Fig. 5a). The gp120 binding ability was retained
by IgA6-VH linked to either of the two selected cervical VL domains. Both the IgA6/2k and the IgA6/30 combinations show
significant binding to the clade B gp120. The binding curve for
IgA6/5k and the IgA6/30I cervical ScFvs were tested with purified scFv protein in a capture EIA assay on a clade B gp120JRFL.
While there was no detectable binding by the IgA9 heavy chain
when recombined with either of the light chains and expressed as
supernatant, we have yet normalized the expression of the IgA9
clones as there may be inclusion formation or toxicity from these
proteins. Neither IgA6/5k nor IgA6/30I are inhibited by whole
104
BERRY ET AL.
FIG. 4. Nucleotide sequences of the selected light-chain variable regions. The sequences were compared to the known immunoglobulin sequences in the databanks on the worldwide web. The nucleotide and inferred amino acid translations of the 30lambda VL (upper) and 5-kappa VL (lower) and depicted in alignment with the closest matching germline alleles for human
VL, and JL. The 30-lambda VL has a polymorphism in the CDR1 domain with a deletion of codon 35. The nucleotide sequences
for 30-lambda and 5-kappa VLs have been submitted to GenBank (accession numbers AY044151 and AY044152, respectively).
Alignments and CDR delineation were performed according to LeFranc et al.(8) and the ImmunoGenetics databanks. denotes
identity; and denotes a deletion.
IgGb12 antibody for gp120 binding over several logs b12 protein
concentration (Fig. 5b). This shows that the gp120 epitope targeted by the cervical MAbs are different from the parental IgGb12
epitope. This is consistent with a previously identified hepatitis
MAb, where changing the light-chain partner was found to alter
epitope specificity.(23)
DISCUSSION
We used phage display to select human gp120 binding domains from cervical B-lymphocytes of HEPS. Using gp120 in
solid phase we selected hybrid phage-borne ScFv from a chain
shuffled IgA/k/ library generated from the pooled cervical B
lymphocytes of 12 Kenyan HEPS. The HEPS cervical B cells
contain antibody binding domains, which neutralize HIV-1.
This is corroborates previous findings of IgA antibody to HIV1 in HEPS,(5) and the neutralizing capacity of these secreted
antibodies in two different in vitro culture assays.(8,9) The neutralization of HIV-1 by hybrid clone IgA6/4L correlates with
the ability to bind to the gp120 envelope oligomer on cells (Fig.
2b). Consistent with the compartmentalized nature of the IgA
response to HIV, local B cells on the cervix directly produce
antibodies to HIV-1. Further tests on the neutralizing spectrum
CHAIN-SHUFFLING WITH CERVICAL V-GENES
105
FIG. 5. Characteristics of cervical MAbs obtained through guided selection. (A) Graph depicting binding of African cervical
monoclonal ScFv, expressed in bacterial culture supernatants, on clade B gp120JRFL in a solid-phase ELISA. Black bars represent MAbs comprised of recombined cervical heavy and light-chain variable domains derived by guided selections. The MAbs
encoded by IgA6 VH cDNA fused to either the 5-kappa or 30-lambda light chain cDNAs bind to gp120. The mAbs encoded by
IgA9 VH cDNA fused to either of the cervical VL do not bind only 2- to 3-fold over background). The parental H31/4L b12
variant was tested in parallel as a control. Results are representative of three identical assays. (B) Reactivity of cervical MAbs
obtained through guided selection. Purified cervical MAbs IgA6/5-kappa [open triangle] and IgA6/30-lambda [open circle] were
tested in ELISA. Commercial polyclonal sheep anti gp120 C-terminus IgG (International Enzymes Inc.) was used to capture recombinant gp120JRFL on a 96-well polystyrene plate. Purified ScFv were serially diluted in 2% BSA-PBS and reacted upon a
4% BSA preblocked plate. Commercial secondary antibody (anti-HA tag-HRP; Roche) was used to detect the ScFv molecules.
B-(INSET). Whole recombinant IgG B12 molecular was tested in competition against cervical MAbs IgA6/301 (filled circles)
or IgA6/5k (filled triangles) for binding to gp120JRFL. The concentration of the ScFvs were normalized at 7.5 g/mL in 2%
BSA, which was about 50% of maximum signal. The open symbols represent binding without IgG B12 as an inhibitor; the diamond is an irrelevant ScFv control. The whole IgG B12 molecule showed no inhibition of binding to either ScFv over several
logs, which is consistent with the use of whole IgG B12 as an inhibitor during the wash steps in the phage selection.
106
of these new cervical V-domains on additional TCLA and primary isolates are underway. The reason for the absence of in
vitro neutralization seen with MAb IgA9/4L is not clear, but
could be due to unknown parameters for neutralization of HIV1 in vivo on the mucosal surface. Nevertheless, the IgA9/4L
acts as an additional in vitro control for neutralization by new
VH with the parental 4L light chain. This study shows that
MAbs can be selected from cDNA cloned from the ectocervical plasmablast population of HEPS.
This is the first set of recombinant MAbs to be derived from
B cells of the human cervix. While the human cervix is rich in
B cells and plasmablasts,(13) the clonal diversity was uncharacterized. Johansson et al.(24) showed that CD38 plasmablasts
are highly concentrated on the lamina propria of the cervix and
the vagina, which corroborates earlier observations of local populations of antibody producing cells in the human cervix.(25)
Despite the predominance of plasmablasts in the cervix, which
are highly enriched in specific Ig mRNA,(26) we observe an assortment of light and heavy chain V genes in the cervix. We
can conclude that cervical IgA antibody is not restricted, and
that diverse B-cell pool is found in the cervix.
We used guided selection to produce new MAbs to gp120
from cervical B cells libraries of HEPS prostitutes. The new
gp120 specific VH and VL domains were combined to form
completely Kenyan MAbs and these were tested for gp120 reactivity. Purified scFv fragments from MAb IgA6-30 lambda
bound the recombinant gp120JRFL in EIA much more strongly
than the IgA6 VH-5 kappa, which also binds the clade B gp120
(Fig. 5). The absence of competition of whole IgG B12 with
cervical MAbs IgA6/5k and IgA6/30 lambda shows that a new
epitope on gp120 is bound (Fig. 5b). This is consistent with the
unique reactivity of the IgA6/4L hybrid observed initially in
flow cytometry (Fig. 2b) and suggests the IgA6 VH is specific
to a different epitope on gp120, conserved on both the Clade
A and B envelope protein oligomers. The epitope identified by
IgA-6 VH may be immunodominant and/or protective on the
cervix, however, its conservation among HIV-1 strains is not
currently known.
The selection of new hybrid MAbs with altered epitope
recognition is an expected feature of guided selection.(27) For
example, light chain shuffling facilitates the selection of antibodies with a shift in epitope recognition.(23,28) Herein we show
that heavy chain shuffling and whole IgGb12 antibody competiton during the selection allows the recovery of heavy chains
with novel gp120 epitope reactivity. Cervical antibody libraries
from HEPS commercial sex workers are a useful source of lymphocytes for deriving human MAbs to STD pathogens.
The combinatorial antibody selection strategy works despite
the scrambling of heavy and light chain pairs inherent to library
construction. Guided-selection in combination with immune
mRNA sources seems to work by generating a reasonable proportion of productive heavy–light chain combinations.(22)
While it is possible that the representation of dominant antigp120 transcripts in vivo is simply too low for the usual library
approach, there are several potential limitations to this study.
For example, the number of HEPS individuals and whether
there is a requirement for exposure of HEPS to have this specificity are both unknown. While our findings are compelling, especially in light of the in vivo resistance phenotype, we cannot
rule out the possibility that the specificity may have been created by the chain-shuffling strategy. Chain shuffling is a con-
BERRY ET AL.
structive approach for obtaining binding clones as not all H/L
combinations may be present in the original randomized library.(21) The advantages reside in the generation of large numbers of functional antibody combinations beginning with a single antibody template.(28,29) For example, Kang et al.(29) showed
that chain shuffling of hapten specific VH and VL pairs into an
immune, but not a naive cDNA lambda phage library, produced
new binding clones. Thus “immune” libraries are enriched in
antigen-specific VH and VL partners compared with libraries
from lymphocytes from unexposed mRNA sources. The small
diversity in our libraries suggests it is unlikely that natural pairs
of gp120 specific MAbs could be formed in vitro unless they
are over-represented, for example, by plasmablasts, in the
cervix.
The chain-shuffling strategy selected novel immunoglobulin
binding domains to gp120. At the nucleotide level, clone IgA9
has greatest homology to a VH1-like gene (V1 4603; 91%)
and clone IgA6 has highest identity to a VH3 related gene (V3
6602; 85%) (Fig. 3). According to the NCBI databanks, clone
IgA9 has a maximum identity of 85% at the inferred amino acid
level with HV1B. IgA9 VH also has 74% identity a VH-domain from systemic B-cell cDNA selected from a phage Fab
library on herpes simplex virus.(30) The IgA6 VH3 like gene
has a maximum identify of 75% at the inferred amino acid level
with previously identified V genes in the databank [HV3F
(P01767)]. Planar-type binding sites, which are in general
elicited against protein-antigens, tend to be formed by short
CDR loops.(31) Analysis of the cDNA sequences of the selected
cervical heavy chains IgA6 and IgA9 reveals relatively short
but unique CDR3 regions of 9 inferred amino acid residues
each. The selection of gp120-specific VH genes from the cervical B-cell pool of these Kenyan prostitutes supports the notion that V-genes with short CDR3 lengths may comprise high
afinity antibodies elicited against commonly encountered antigens of the mucosal immune system.(32) The reduced length in
the CDR3 region supports the binding of the cervical ScFv to
different epitopes from the parental b12 molecule. The IgGb12
is proposed to utilize an extended H3 to access critical contacts
of the CD4 binding site on gp120,(33) although b12 has not been
crystallized while bound to any form of gp120.
The selection of new light chain partners for H31 VH from
cervical B cells, while novel, is not surprising as light-chain
promiscuity has been observed with gp120 specific VH domains.(34) Cervical light chains 5-kappa and 30-lambda have
identity to previously published sequences of 85% and 74%,
respectively. IGKV3 2001 and has a IGKJ201 joining region.
The 30-lambda light chain has 74% identity with IGLV2-1403
and has a lambda IGLJ302 joining region (Fig. 4). Of note,
the 30-lambda light chain is clearly polymorphic compared to
the databank sequences. Codon 35 is deleted and an insertion
in the CDR3 region, which could be due to true polymorphism
or from somatic mutation during antigen-specific immune responses(35) or acquired during rearrangement and junctional diversification.
MAbs comprised of new VH and VL combinations are expected to target a different spectrum of HIV-1 antigenic variants for neutralization. For example, in vivo replacement of VL,
termed receptor editing,(36) produces antibodies to the same epitope, as directed by the heavy chain, but utilize a different set
of light chain contacts.(37) Indeed, designed insertions into the
variable regions of MAbs can have variable effects upon anti-
107
CHAIN-SHUFFLING WITH CERVICAL V-GENES
gen binding depending upon the type of antigen they are raised
against.(38) While the cervical V genes we selected are somatically mutated and/or polymorphic, the role of polymorphic V
genes in HIV-1 resistance remains to be determined. V genes
encoding unselected human parotid IgA is also somatically mutated,(32) which likely reflects antigen-specific affinity maturation. Collectively, the relative importance of the heavy chain
domain in determining antibody specificity(22) and the presence
of somatically mutated gp120 specific IgA antibodies in the
genital tract of HEPS individuals(8,9) show that a localized
gp120 specific B-cell response has occurred in Kenyan HEPS.
Pools of potent MAbs to diverse HIV types will produce broad
neutralizing coverage and help to delineate new epitopes for active vaccine discovery.
The molecular basis for immune exclusion of HIV-1 at the
mucosal surface is not clear. Mucosal IgA has traditionally been
attributed as being a major component of protection against
STD pathogens despite limited molecular evidence. Limitations
of our work include the indirect method of selection of these
cervical MAbs to and the lack of available Kenyan gp120 to
reflect the predominant clades of the HIV-1 virus at the outset
of these experiments. The relevance of host co-receptors or even
in vitro neutralization assays to sexual transmission is unclear.
This is important as most HIV transmission is sexual and few
broadly neutralizing gp120 MAbs have been derived from systemic B cells of HIV-positive individuals. There are several antibody-virus parameters that may differ in the cervix. First, for
local protection, a high concentration of HIV-1-specific antibodies at the site of HIV-1 entry may be more important than
high avidity in vivo.(39) Second, it is unclear if immunodominant B-cell epitopes targeted by systemic immune response are
also immunodominant in immune responses, or to B cells on
mucosal surfaces. Third, it is unclear if CD4-based cellular inhibition assays are relevant to the mode of entry of HIV-1 from
mucosal surfaces.(40) Therefore, it is critical to adapt tools, like
immune libraries, to be able to study the B-cell pools that first
encounter HIV-1. Moreover, it will be important to test MAbs
derived from the cervix on HIV-1 strains relevant to sexual exposures in that host.(41)
In conclusion, we have successfully isolated and characterized the first functionally active anti-HIV 1 virus scFv
MAbs from the human cervix. The cervix is a potentially
valuable source of B-cell diversity for the generation of human MAbs to STD pathogens. The mucosal antibody pool
should be well suited for in vitro modification strategies, like
humanization,(42) for antibody against microbes as this pool
is sensitized against a wide variety of pathogens. Clearly,
these data lead the way for more studies aimed at the role of
the cervical antibody repertoire in resistance to HIV, which
appears to be broadly neutralizing.(43) Determination of the
epitopes to which these antibodies were raised may lead to
an effective vaccine for HIV-1 in humans. The use of antibody libraries to sample local mucosal B-cell responses of
the urogenital tract has broad implications for STD vaccine
research.
ACKNOWLEDGMENTS
The authors acknowledge the assistance of the women of the
Pumwani sex cohort and support from Dr. Dennis Burton and
the Scripps Research Institute’s level-three containment facility. Work on this article was supported by grants from the National Institute of Health 5R01AI37470-06 (C.F.B. III),
1R01AI48373-01 (F.P.). J.D.B. was supported by a CIHR
HIV/AIDS fellowship. J.R. is supported by a CANVAC grant
to F.A.P.
REFERENCES
1. Fowke KR, Nagelkerke NJ, Kimani J, Simonsen JN, Anzala AO,
Bwayo JJ, MacDonald KS, Ngugi EN, and Plummer FA: Resistance to HIV-1 infection among persistently seronegative prostitutes in Nairobi, Kenya. Lancet 1996;348:1347–1350.
2. Plummer F, Ball TB, Kimani J, and Fowke KR: Resistance to HIV1 infection among highly exposed sex workers in Nairobi: what
mediates protection and why does it develop? Immunol Lett
1999;66:27–34.
3. Kaul R, Plummer FA, Kimani J, Dong T, Kiama P, Rostron T,
Njagi E, MacDonald KS, Bwayo JJ, McMichael AJ, and RowlandJones SL: HIV-1 specific mucosal CD8 lymphocyte responses in
the cervix of HIV-1 resistant prostitutes in Nairobi. J Immunol
2000;164:1602–1611.
4. Fowke KR, Kaul R, Rosenthal KL, Oyugi J, Kimani J, Rutherford
WJ, Nagelkerke NJ, Ball TB, Bwayo JJ, Simonsen JN, Shearer
GM, and Plummer FA: HIV-1-specific cellular immune responses
among HIV-1-resistant sex workers. Immunol Cell Biol 2000;
78:586–595.
5. Kaul R, Trabattoni D, Bwayo JJ, Arienti D, Zagliani A, Mwangi
FM, Kariuki C, Ngugi EN, MacDonald KS, Ball TB, Clerici M,
and Plummer FA: HIV-specific mucosal IgA in a cohort of HIV1 resistant Kenyan sex workers. AIDS 1997;12:23–27.
6. Beyrer C, Artenstein AW, Rugpao S, Stephens H, VanCott TC,
Robb ML, Rinkaew M, Birx DL, Khamboonruang C, Zimmerman
PA, Nelson KE, and Natpratan C: Epidemiologic and biologic characterization of a cohort of human immunodeficiency virus type 1
highly exposed, persistently seronegative female sex workers in
northern Thailand. Chiang Mai HEPS Working Group. J Infect Dis
1999;179:59–67.
7. Mazzoli S, Lopalco L, Salvi A, Trabattoni D, Lo Caputo S, Semplici F, Biasin M, BI C, Cosma A, Pastori C, Meacci F, Mazzotta
F, Villa ML, Siccardi AG, and Clerici M: Human immunodeficiency virus-specific IgA and HIV neutralizing activity in the serum
of exposed seronegative partners of HIV-seropositive persons. J Infect Dis 1999;180:871–875.
8. Devito C, Hinkula J, Kaul R, Lopalco L, Bwayo JJ, Plummer F,
Clerici M, and Broliden K: Mucosal and plasma IgA from HIVexposed seronegative individuals neutralize a primary HIV-1 isolate. AIDS 2000a;14:1917–1920.
9. Devito C, Broliden K, Kaul R, Svensson L, Johansen K, Kiama P,
Kimani J, Lopalco L, Piconi S, Bwayo JJ, Plummer F, Clerici M,
and Hinkula J: Mucosal and Plasma IgA from HIV-1-Exposed Uninfected Individuals Inhibit HIV-1 Transcytosis Across Human Epithelial Cells. J Immunol 2000b;165:5170–5176.
10. Ada GL, and Ramsay AJ: Vaccines: vaccination and the immune
response. Lippincott-Raven, New York, 1996.
11. Burton DR: Antibodies, viruses and vaccines. Nat Rev Immunol
2002;2:706–713.
12. Brandtzaeg, P, Farstad, IN, and Haraldsen G: Regional specialization in the mucosal immune system: primed cells do not always
home along the same track. Immunol Today 1999;20:267–277.
13. Crowley-Nowick PA, Bell M, Edwards RP, McCallister D, Gore
H, and Kanbour-Shakir A: Normal uterine cervix: Characteristics
of isolated lymphocyte phenotypes and immunoglobulin secretion.
Am J Reprod Immunol 1995;34:241–247.
14. Barbas CF, 3rd, Collet TA, Amberg W, Roben P, Binley JM, Hoek-
108
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
BERRY ET AL.
stra D, and Burton DR: Molecular profile of an antibody response
to HIV-1 as probed by combinatorial libraries. J Mol Biol
1993;230:812–823.
Parren PW, Marx PA, Hessell AJ, Luckay A, Harouse J, ChengMayer C, Moore JP, and Burton DR: Antibody protects macaques
against vaginal challenge with a pathogenic R5 simian/human immunodeficiency virus at serum levels giving complete neutralization in vitro. J Virol 2001;75:8340–8347.
Burton DR, Pyati J, Koduri R, Sharp SJ, Thornton GB, Parren PW,
and Barbas CF, III: Efficient neutralization of primary isolates of
HIV-1 by a recombinant human monoclonal antibody. Science
1994;266:1024–1027.
Gauduin MC, Parren PW, Weir R, Barbas CF, Burton DR, and
Koup RA: Passive immunization with a human monoclonal antibody protects hu-PBL-SCID mice against challenge by primary isolates of HIV-1. Nature Med 1997;3:1389–1393.
Beiboer SH, Reurs A, Roovers RC, Arends JW, Whitelegg NR,
Rees AR, and Hoogenboom HR: Guided selection of a pan carcinoma specific antibody reveals similar binding characteristics yet
structural divergence between the original murine antibody and its
human equivalent. J Mol Biol 2000;296:833–849.
Barbas CF, Burton D, Scott JK, and Silverman GJ (eds.): Phage
display: a laboratory manual. Cold Spring Harbour laboratory
Press, New York, 2001.
Andris-Widhopf J, Rader C, Steinberger P, Fuller R, and Barbas
CF: Methods for the generation of chicken monoclonal antibody
fragments by phage display. J Immunol Methods 2000;242:
159–181.
Burton D, and Barbas CF: Human antibodies to HIV-1 by recombinant DNA methods. In: Norrby E (Ed.). Immunochemistry of
AIDS. Karger, Basel, 1193;112–126.
Xu JL, and Davis MM: Diversity in the CDR3 region of VH is sufficient for most antibody specificites. Immunity 2000;13:37–45.
Zebedee SL, Barbas CF 3rd, Hom YL, Caothien RH, Graff R, DeGraw J, Pyati J, LaPolla R, Burton DR, and Lerner RA: Human
combinatorial antibody libraries to hepatitis B surface antigen. Proc
Natl Acad Sci USA 1992;89:3175–3179.
Johansson E, Rudin A, Wassen L, and Holmgren J: Distribution of
lymphocytes and adhesion molecules in human cervix and vagina.
Immunology 1999;96:272–277.
Kutteh W, Hatch K, Blackwell RE, and Mestecky J: Secretory immune system of the female reproductive tract: I. Immunoglobulin
and secretory component-containing cells. Obstet Gynecol 1988;
71:56–60.
Yuan D, and Tucker P: Regulation of IgM and IgD synthesis in B
lymphocytes I. Changes in biosynthesis of mRNA for mu and delta
chains. J Immunol 1984;132:1561–1565.
Rader C, Ritter G, Nathan S, Elia M, Gout I, Jungbluth AA, and
Barbas CF, III: The rabbit antibody repertoire as a novel source for
the generation of therapeutic human antibodies. J Biol Chem
2000;275:13668–13676.
Ohlin M, Owman H, Mach M, and Borrebaeck CAK: Light chain
shuffling of a high affinity antibody results in a drift in epitope
recognition. Mol Immunol 1996;33:47–56.
Kang A, Jones T, and Burton D: Antibody redesign by chain shuffling from random combinatorial immunoglobulin libraries. Proc
Natl Acad Sci USA 1991;88:11120–11123.
Sanna PP, Williamson RA, De Logu A, Bloom FE, and Burton
DR: Directed selection of recombinant human monoclonal antibodies to herpes simplex virus glycoproteins from phage display
libraries. Proc Natl Acad Sci USA 1995;92:6439–6443.
Vargas-Madrazo E, Lara-Ochoa F, and Almagro JC: Canonical
structure repertoire of the antigen-binding site of immunoglobulins
suggests strong geometrical restrictions associated to the mechanism of immune recognition. J Mol Biol 1995;254:497–504.
Dunn-Walters DK, Hackett M, Boursier L, Ciclitira PJ, Morgan P,
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
Challacombe SJ, and Spencer J: Characteristics of human IgA and
IgM genes used by plasma cells in the salivary gland resemble
those used in the duodenum but not those used in the spleen. J Immunol 2000;164:1595–1601.
Saphire EO, Parren PW, Pantophlet R, Zwick MB, Morris GM,
Rudd PM, Dwek RA, Stanfield RL, Burton DR, and Wilson IA:
Crystal structure of a neutralizing human IgG against HIV-1: A
template for vaccine design. Science 2001;293:1155–1159.
Barbas CF 3rd, Hu D, Dunlop N, Sawyer L, Cababa D, Hendry
RM, Nara PL, and Burton DR: In vitro evolution of a neutralizing
human antibody to human immunodeficiency virus type 1 to enhance affinity and broaden strain cross-reactivity. Proc Natl Acad
Sci USA 1994;91:3809–3813.
Ohlin M, and Borrebaeck CA: Insertions and deletions in hypervariable loops of antibody heavy chains contribute to molecular diversity. Mol Immunol 1998;35:233–238.
Casellas R, Shih TA, Kleinewietfeld M, Rakonjac J, Nemazee D,
Rajewsky K, and Nussenzweig MC: Contribution of receptor editing to the antibody repertoire. Science 2001;291:1541–1544.
Fignini M, Marks J, Winter G, and Griffiths A: In vitro assembly
of repertoires of antibody chains on the surface of phage by re-naturation. J Mol Biol 1994;239:68–78.
Lantto J, and Ohlin M: Functional Consequences of Insertions and
Deletions in the Complementarity-determining Regions of Human
Antibodies. J Biol Chem 2002;277:45108–45114.
Bachmann MF, Kalinke U, Althage A, Freer G, Burkhart C, Roost
H, Aguet M, Hengartner H, and Zinkernagel RM: The role of antibody concentration and avidity in antiviral protection. Science
1997;276:2024–2027.
Geijtenbeek TB, van Vliet SJ, van Duijnhoven GC, Figdor CG,
and van Kooyk Y: DC-SIGN, a dendritic cell-specific HIV-1 receptor present in placenta that infects T cells in trans-a review. Placenta 2001;(Suppl A):S19–23.
Ellerbrock TV, Lennox JL, Clancy KA, Schinazi RF, Wright TC,
Pratt-Palmore M, Evans-Strickfaden T, Schnell C, Pai R, Conley
LJ, Parrish-Kohler EE, Bush TJ, Tatti K, and Hart CE: Cellular
replication of human immunodeficiency virus type 1 occurs in vaginal secretions. J Infect Dis 2001;184:28–36.
Rader C, Cheresh DA, and Barbas CF 3rd: A phage display approach for rapid antibody humanization: designed combinatorial V
gene libraries. Proc Natl Acad Sci USA. 1998;95:8910–8915.
Devito C, Hinkula J, Kaul R, Kimani J, Kiama P, Lopalco L, Barass
C, Piconi S, Trabattoni D, Bwayo JJ, Plummer F, Clerici M, and
Broliden K: Cross-clade HIV-1-specific neutralizing IgA in mucosal and systemic compartments of HIV-1-exposed, persistently
seronegative subjects. J Acquir Immune Defic Syndr 2002;30:
413–420.
Address reprint requests to:
Carlos F. Barbas, III, Ph.D.
10550 North Torrey Pines Road
La Jolla, CA 92037
E-mail: carlos@scripps.edu
or
Jody D. Berry, Ph.D.
1015 Arlington Street
Winnepeg, MB
Canada R3E3M4
E-mail: berryjd@inspection.gc.ca
Received for publication January 29, 2003. Accepted for publication February 19, 2003.