J Biol Chem, Vol. 274, Issue 39, 27371-27378, September 24, 1999
Use of a Peptide Mimotope to Guide the Humanization of
MRK-16, an Anti-P-glycoprotein Monoclonal Antibody*
Ying
Tang,
Greg
Beuerlein,
Gerlinde
Pecht,
Todd
Chilton,
William D.
Huse, and
Jeffry D.
Watkins
From Ixsys, Inc., San Diego, California 92121
 |
ABSTRACT |
A mimotope-guided strategy for engineering
antibodies directed against orphan targets or antigens that are
difficult to purify was developed and used to humanize the murine
MRK-16 monoclonal antibody (mAb). MRK-16 recognizes a conformational
epitope of a 170-kDa membrane protein, termed P-glycoprotein (P-gp).
Elevated expression of P-gp on tumor cells is associated with
resistance to cytotoxic drugs, a major obstacle in chemotherapy. Murine
MRK-16 was used to enrich and screen a phage-displayed peptide library to identify reactive mimotopes. One peptide, termed ALR1, was enriched
to a greater extent than others and subsequently was expressed as a
fusion protein with glutathione S-transferase. ALR1 fusion
protein bound MRK-16 specifically and inhibited binding of MRK-16 to
cells expressing elevated levels of P-gp. To humanize MRK-16, the
murine complementarity determining regions were grafted onto homologous
human heavy and light chain variable region frameworks. Framework
residues that differed between the murine MRK-16 and the homologous
human templates were analyzed and subsequently, five framework
positions potentially important for maintaining the specificity and
affinity of MRK-16 were identified. A combinatorial library consisting
of 32 variants encoding all possible combinations of murine and human
residues at the five differing framework positions was expressed in a
phage system. In the absence of purified P-gp, ALR1 fusion protein was
used as surrogate antigen to screen the antibody library to identify
the framework combination that most preserved the binding activity of
the mAb. On the basis of the initial screening against the mimotope
four antibody variants were selected for further characterization. The
binding affinity of these variants for the ALR1 fusion protein
correlated with their binding to cells expressing elevated levels of
P-gp. Thus, peptide mimotopes which can be identified for virtually any
antibody including those that recognize conformational or carbohydrate epitopes, can serve as antigen templates for antibody engineering.
| |
INTRODUCTION |
The widespread success of murine hybridoma technology has resulted
in the discovery of numerous well characterized monoclonal antibodies
(mAbs)1 with unique
specificities. Many of these mAbs display tremendous therapeutic
potential both as vehicles for targeting cytotoxic agents (reviewed in
Ref. 1) and as function blocking molecules (2-7). However, murine mAbs
are generally recognized as foreign antigens by the human immune system
preventing the administration of multiple doses (8, 9). As a result,
there has been considerable effort devoted to circumventing the
immunogenicity of murine mAbs, including the development of methods for
discovering human antibodies. For example, human lymphocytes have been
stimulated in vitro (10, 11), phage-expressed human antibody
libraries have been synthesized (reviewed in Ref. 12), and transgenic
mice expressing human Ig genes have been created (13). Although human
mAbs have been discovered by these approaches there remains a need to
fully exploit the potential clinical benefits of murine antibodies.
As an alternative to discovering human mAbs, antibody engineering
approaches have been developed to reduce the potential immunogenicity of murine mAbs. For example, chimeric antibodies consisting of the
murine variable region fused to a human constant region have been
constructed (14-16). Antibody variants containing even fewer murine
residues have been synthesized by grafting murine CDRs onto human
variable region frameworks (17-19). Unfortunately, chimeric antibodies
may still display immunogenic properties in humans (20, 21) and CDR
grafting often diminishes the affinity of mAbs (18, 22). A general and
simple approach for reducing the immunogenicity of murine mAbs without
diminishing their affinity would permit a greater number of the
existing well characterized mAbs to be evaluated as therapeutics.
The diminished affinity often observed with antibodies humanized by CDR
grafting reflects the structural significance of certain framework
residues. The relative importance of specific framework residues varies
between different mAbs and consequently, identifying important residues
and determining the optimal amino acid at those positions has proven
difficult. Previously, we described an approach for the humanization of
antibodies that uses phage expression of combinatorial antibody
framework libraries and CDR grafting coupled with identification of the
most active humanized variant by rapid screening methods (23-25).
Phage expression of combinatorial framework libraries takes advantage
of the efficiency of bacterial cloning systems and, when coupled with
the appropriate assays, increases the likelihood of identifying fully
active humanized variants. Examination of a wide range of framework
structures reduces the requirement for structural modeling to predict
precisely the framework residues critical for binding activity.
Furthermore, the identification of fully active humanized variants from
combinatorial libraries is straightforward, using screening methods
that permit the affinities of numerous (>105) soluble Fabs
to be distinguished rapidly (24, 26). To date, the optimal screening
methods have required soluble labeled antigen. However, therapeutic
mAbs are often directed against cell surface antigens that are integral
membrane proteins and therefore, the purification or expression of
quantities of antigen sufficient to support screening of antibody
libraries has been limiting in certain instances.
Antibodies recognize a restricted portion of protein targets, termed
the antigenic determinant or epitope, which can consist of peptide,
protein, carbohydrate, or other chemical moieties. Regardless of the
precise chemical nature of the epitope, peptides can often be used to
mimic the structure of the antigenic determinant (27). Moreover, the
efficient synthesis of peptides or the display of random peptide
libraries on the surface of phage (28) have greatly facilitated the
identification of peptides reactive with a wide range of antibodies.
For example, peptides reactive with mAbs that recognize linear (29),
conformational (30), and carbohydrate (31) epitopes have been
described. The ability of these mAb-reactive peptides to mimic the
actual antigen epitope has been demonstrated by the use of the peptides
as assay reagents and affinity matrices (32, 33), to provide structural
information about the antigen (34, 35) or antibody (36), and as
immunogens (37-40). The ability of peptides to effectively mimic a
portion of the structure of the antigen suggests that it may also be
possible to use peptide mimotopes as templates for screening libraries of humanized variants to identify clones that retain the conformation of the binding site of the murine mAb. As such, a readily available supply of soluble antigen would be available to direct the humanization of mAbs, regardless of the nature or availability of the actual antigen.
In the present study we wanted to determine if a mimotope could in fact
be used to guide the successful humanization of an antibody. The murine
mAb MRK-16, which is directed against an external domain of
P-glycoprotein (41), was selected as a model system. P-gp is an
integral membrane protein that putatively spans the plasma membrane 12 times (42-44), is expressed at elevated levels on many tumor cells
(45), and acts as a drug efflux pump (reviewed in Ref. 46). The
expression of elevated levels of P-gp on tumor cells is associated with
multidrug resistance, a primary cause of failure of chemotherapy.
However, MRK-16 can block the efflux of certain chemotherapeutic drugs
in vitro (41) and treatment of athymic mice with MRK-16
inhibited the growth of drug-resistant human tumor cells in
vivo (47). Thus, MRK-16 may increase the benefits of certain
chemotherapeutic agents. Although a humanized version of MRK-16 is
preferred for human clinical studies, the difficulty in obtaining
sufficient levels of soluble P-gp has hindered progress. In this study
we describe the humanization of MRK-16 using a peptide mimotope as
surrogate for the P-gp antigen for screening.
| |
EXPERIMENTAL PROCEDURES |
Materials--
HCT-15 human colorectal carcinoma cells were
obtained from American Type Culture Collection (Rockville, MD). The
human adriamycin-resistant K562/ADM cell line was established from the
human myelogenous leukemia K562 cell line and was maintained as
described previously (48). MRK-16 antibody and K562/ADM cells were
obtained from Hoechst Japan.
Identification of MRK-16 Mimotopes--
Polystyrene latex beads
(1 µm diameter, Interfacial Dynamics Corp., Portland, OR) were washed
twice and resuspended in 100 mM MES, pH 5.0, to the
original volume. The beads were incubated at 45 °C for 5 min at
which time 70 µl were combined with 150 µg of MRK-16 in 700 µl of
100 mM MES, pH 5.0, for 1 h at 45 °C. The beads
were collected by centrifugation, rinsed twice with 1 mM
MOPS, pH 7, 150 mM NaCl, 1 mM
MgCl2, 0.1 mM ZnCl2, 1% poly(vinyl alcohol), 0.1% sodium azide (MOPS, 1% poly(vinyl alcohol)) containing 3% BSA followed by two washes with MOPS, 1% poly(vinyl alcohol) alone. The beads were resuspended in 500 µl of MOPS, 1% poly(vinyl alcohol) and were stored at 4 °C until use.
The phage-expressed peptide library was designed to randomly express
cysteine residues at a 5-fold higher frequency and was synthesized as
described previously (49). To amplify the peptide library, 300 ml of
supE amber suppressor strain XL1-Blue (Stratagene, Inc., San
Diego, CA) was grown at 37 °C until the cultures reached a density
of 0.8 at OD600 at which time
isopropyl-thio-
-D-galactoside was added to a final
concentration of 2 mM and the culture was infected with 100 µl of a high titer phage stock (1010-1011
phage/ml). After 6 h of growth, bacteria were removed from 10 ml
of culture media by centrifugation at 2000 × g for 25 min.
For enrichment of phage, 100 µl of MRK-16-coated beads were added to
a microcentrifuge tube which had been blocked with MOPS, 1% poly(vinyl
alcohol) for 15 min at 25 °C. The beads were collected by
centrifugation, resuspended in 1 ml of culture media containing the
amplified peptide library for 16 h at 4 °C, and collected by
centrifugation. The beads were washed six times with 1 ml of Tris-buffered saline, resuspended in 500 µl of Tris-buffered saline, and stored at 4 °C. The bound phage were amplified as described above, using 400 µl of the beads to infect a 50-ml culture of XL1-Blue. The cells were pelleted by centrifugation at 11,000 × g for 30 min and the amplified phage were precipitated from the culture supernatant following a 30-min incubation at 4 °C with
1.25% PEG 8000 in 220 mM ammonium acetate. The phage were collected by centrifugation at 11,000 × g for 30 min
and resuspended in 2 ml of 1% BSA in PBS. Subsequently, 1 ml of the
amplified phage was used for the next round of selection.
Following the second and third rounds of enrichment an aliquot of beads
was screened by plaque lift, as described previously (50). The filters
were incubated with 5 µg/ml MRK-16 in 5% nonfat powdered milk, 0.2%
Tween 20, 0.01% anti-foam A emulsion, and 0.01% thimerosal in PBS and
washed three times with PBS containing 0.1% Tween 20 (PBS-T). Reactive
plaques were identified colorimetrically following incubation with goat
anti-murine IgG-alkaline phosphatase conjugate (25).
Expression and Biotinylation of GST-Peptide Fusion
Proteins--
Three overlapping oligonucleotides encoding the reactive
peptide, a six-proline linker and unique EcoRI and
NotI restriction sites were synthesized. The overlapping
oligonucleotides were annealed and amplified by polymerase chain
reaction. Following digestion with EcoRI and NotI
and purification by agarose gel electrophoresis, the DNA fragment was
ligated into the pGEX-4T-1 vector (Amersham Pharmacia Biotech) and
introduced into DH10B cells by electroporation. Expression and
purification of GST-peptides was performed as described by the
manufacturer. Purified GST-peptides was dialyzed into PBS and incubated
with 10-fold molar excess sulfosuccinimidobiotin for 2 h at
25 °C. The reaction was terminated by addition of 15 mM
ethanolamine and excess biotin was removed by dialysis against PBS.
Design, Synthesis, and Characterization of Phage-expressed
Humanized MRK-16 Fab--
Overlapping oligonucleotides (69-75 bases
in length) encoding the framework regions of the H and L chain variable
domains of the human templates and the murine MRK-16 CDRs, as defined by Kabat et al. (51, 52), were synthesized and used to
construct a CDR-grafted version of MRK-16. Briefly, the H and L chain
variable domains were synthesized separately by combining equimolar
amounts of the overlapping oligonucleotides with Pfu DNA
polymerase in a polymerase chain reaction. Subsequently, the variable
domains were fused to human 
1 and 
constant region sequences in a
modified M13IX104 vector (53), termed M13IX104CS, using hybridization mutagenesis (23, 54, 55). The M13IX104 vector was modified by replacing
cysteine residues at the end of the and 1 constant regions with
serine. Subsequently, all possible combinations of murine and human
residues at five framework positions (Fig. 3, asterisks)
were synthesized using site-directed mutagenesis, resulting in the
expression of 32 unique variants.
ELISA Screening--
The relative affinities of the humanized
MRK-16 variants expressed in small-scale (<1 ml) bacterial cultures
were assessed by ELISA (24). Briefly, microtiter plates were coated
with 5 µg/ml goat anti-human antibody (Southern Biotechnology)
and blocked with 3% BSA in PBS. Next, 50 µl of Fab from
Escherichia coli cell lysates were incubated with the plate
for 1 h at 25 °C, the plate was washed three times with PBS-T,
and 1 µg/ml biotinylated GST-peptide in PBS containing 1% BSA was
added for 2 h at 25 °C. The plate was washed three times with
PBS-T and NeutrAvidin-alkaline phosphatase conjugate diluted 1000-fold
in PBS-T was added for 30 min at 25 °C. The plate was washed three
times with PBS-T and binding was quantitated colorimetrically (24).
Purified Fab and Fab in the periplasmic space fraction from bacterial
lysates was quantitated as described (24).
HCT-15 cells were plated in 96-well cell culture dishes at 7.5 × 104 cells/well, a cell density sufficient to produce
90-95% confluent monolayers 48 h. The culture media was removed
and 200 µl of Z-FIX (Anatech Ltd., Battle Creek, MI) diluted 5-fold
in water was added for 30 min at 25 °C. The cells were washed once
with PBS and blocked with 1% BSA in PBS for 1 h at 25 °C.
Fixed cell monolayers were incubated with 50 µl of Fab diluted in 1%
BSA in PBS for 2 h at 25 °C, washed three times with PBS, and
Fab binding was detected by incubating with goat anti-human alkaline phosphatase conjugate (Southern Biotech). The cells were
washed three times with PBS and binding was quantitated
colorimetrically (24).
DNA Sequencing--
Single-stranded DNA was isolated and the H
and L chain variable region genes were sequenced by the fluorescent
dideoxynucleotide termination method (Perkin-Elmer, Foster City, CA).
Expression and Purification of Fab--
Fabs were cloned into a
different expression vector under control of the arabinose-regulated
BAD promoter. In addition, a six-histidine tag was fused to the
carboxyl terminus of the H chain to permit purification with
nickel-chelating resins. Purified Fab was quantitated as described
(24).
| |
RESULTS |
Mimotope-guided Humanization of mAbs--
The goal of the present
study was to humanize the murine MRK-16 mAb by screening a
combinatorial framework library for antigen binding. Unfortunately, the
target antigen P-gp, a large integral membrane protein with multiple
transmembrane domains, was not readily available as a soluble protein.
Consequently, we evaluated an alternative strategy for humanizing
MRK-16 and other antibodies in the absence of labeled soluble antigen.
The approach, outlined in Fig. 1, uses a
peptide mimotope to serve as surrogate antigen. Initially, the murine
antibody is used to identify reactive peptides that mimic the antigen
(step 1). Next, the mimotope is expressed as a fusion protein with GST
to ensure solubility and permit efficient purification and
biotinylation (steps 2 and 3). The biotinylated GST mimotope is then
used to screen the phage-expressed combinatorial antibody libraries,
leading to the identification of the most active humanized construct
(step 4). Finally, the murine and humanized antibodies are
characterized for binding to cells expressing the antigen to verify
that the mimotope-guided process does not alter the specificity of the
antibody (step 5).
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 1.
Mimotope-guided humanization of MRK-16.
The murine MRK-16 mAb was used to identify reactive peptides from a
phage-expressed peptide library (step 1). The most highly
enriched mimotope was sequenced, expressed as a GST fusion protein
(step 2), purified, and labeled with biotin (step
3). Thirty-two phage-expressed humanized variants of MRK-16 were
screened with the biotinylated GST-mimotope (step 4). The
most reactive humanized variant containing the fewest murine residues
was characterized further by assessing reactivity with cells expressing
elevated levels of P-gp (step 5).
|
|
Identification and Characterization of MRK-16-reactive
Peptides--
A phage-expressed peptide library was synthesized using
codon-based mutagenesis (49, 56) such that peptides 20 amino acids in
length were fused to the pVIII major coat protein of M13 filamentous phage. Subsequently, the library was incubated with immobilized MRK-16
to enrich phage displaying reactive peptides. After three rounds of
selection, MRK-16-reactive phage were identified by a plaque lift
assay. DNA sequencing of randomly selected reactive phage identified
multiple unique peptides of which one, termed ALR1, was the predominant
species. MRK-16 is not reactive with P-gp under denaturing conditions,
suggesting the mAb recognizes a conformational epitope. Nonetheless,
ALR1 displayed a small degree of homology with a region of P-gp
predicted to be the fourth extracellular loop (Fig.
2A). MRK-16 reactivity has
previously been mapped to this region of P-gp using overlapping
synthetic peptides (57). Interestingly, a peptide corresponding
directly to the linear sequence 737-756 of the extracellular P-gp loop expressed as a GST fusion protein was not reactive with MRK-16 either
by ELISA or Western blot analysis (data not shown). The reason(s) for
this apparent discrepancy are not clear but may reflect differences in
conformation adopted by the free peptide used in the previous study
(57) versus the fusion peptide described herein.
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 2.
Identification and characterization of
peptide reactive with murine MRK-16. A, phage enriched
with immobilized MRK-16 were screened by plaque lift and reactive
peptides were sequenced. The 20-mer peptide identified most frequently
was aligned with P-gp residues 737-756, an extracellular loop reactive
with MRK-16 (57). Underlined amino acids are thought to be
critical for MRK-16 binding (57) and identical/similar amino acids are
indicated with vertical lines. B, the
MRK-16-reactive peptide, termed ALR1, was expressed as a fusion protein
with GST and characterized further. Varying quantities of MRK-16
antibody were incubated with GST-ALR1 (filled circles),
biotinylated GST-ALR1 (open triangles), and an unrelated GST
fusion peptide (open circles) in an ELISA format. MRK-16
binding was quantitated colorimetrically following incubation with
anti-murine IgG-alkaline phosphatase conjugate, as described under
"Experimental Procedures." C, HCT-15 cells expressing
elevated levels of P-gp were incubated with 5 µg/ml MRK-16 antibody
alone (open triangles) or in the presence of varying
quantities of GST-ALR1 (open circles) or an unrelated
GST-peptide (filled circles). Subsequently, MRK-16 binding
was quantitated colorimetrically following incubation with anti-murine
IgG-alkaline phosphatase conjugate, as described under "Experimental
Procedures." The assays were performed two times and the data are the
average of duplicate determinations from a single representative
experiment. The range of the values for the duplicate determinations
was <10% (B) or <5% (C).
|
|
To permit further characterization of ALR1 the peptide was expressed as
a fusion protein with GST. Bacterial lysates containing GST-ALR1 or an
unrelated GST-peptide were resolved by SDS-polyacrylamide gel
electrophoresis and transferred to nitrocellulose for Western blot
analysis. Although MRK-16 is unreactive with Western blots of P-gp, it
bound GST-ALR1. MRK-16 was unreactive with other bacterial lysate
proteins as well as the unrelated GST-peptide, providing evidence for
the specificity of the MRK-16/peptide interaction (data not shown).
ALR1 contains a single cysteine residue and therefore, is unlikely to
be constrained by disulfide bonds. Nonetheless, purified GST-ALR1
protein was reduced with 10 mM dithiothreitol in the
presence of 0.75% SDS and subsequently, the sample was heated to
85 °C for 10 min, cooled, and treated with freshly prepared 0.04 M N-ethylmaleimide for 60 min on ice. GST-ALR1,
reduced GST-ALR1, and reduced and alkylated GST-ALR1 were resolved by
SDS-polyacrylamide gel electrophoresis and transferred to
nitrocellulose. Little or no higher molecular weight bands were
observed under nonreducing conditions, demonstrating the lack of
disulfide bridges between molecules. Furthermore, MRK-16 reacted with
non-reduced, reduced, and reduced and alkylated GST-ALR1 in monomer
form, as determined by Western blot. Although these studies do not rule
out dimerization of the peptide through noncovalent interactions under
nondenaturing conditions, the Western blot data suggest that
dimerization is not essential for its activity.
MRK-16 binding to GST-ALR1 was also evaluated by ELISA under
nondenaturing conditions and was concentration-dependent
and saturable (Fig. 2B, filled circles) while binding to an
unrelated GST-peptide fusion was undetectable (Fig. 2B, open
circles). In order to further demonstrate the specificity of the
mimotope for MRK-16, GST-ALR1 was titrated with 5 µg/ml MRK-16 and
incubated with fixed monolayers of P-gp-expressing HCT-15 cells.
GST-ALR1 inhibited the binding of MRK-16 to the HCT-15 cells in a
concentration-dependent manner (Fig. 2C, open
circles) while an unrelated GST-peptide did not affect MRK-16
binding (Fig. 2C, filled circles). Collectively, these
results are consistent with ALR1 binding at or near the combining site
of MRK-16.
Next, GST-ALR1 was labeled with biotin to permit the screening of
phage-expressed MRK-16 by a previously described rapid approach (24).
MRK-16 and an unrelated mAb were immobilized on a microtiter plate,
incubated with biotinylated GST-ALR1, and binding was detected colorimetrically following incubation with NeutrAvidin-alkaline phosphatase conjugate. As expected, biotinylated GST-ALR1 bound MRK-16
specifically (data not shown). Because biotinylation of GST-ALR1 could
potentially diminish binding to MRK-16 the immunoreactivity of
biotinylated GST-ALR1 was compared with the unmodified fusion protein
(Fig. 2B). Both unmodified (Fig. 2B, filled
circles) and biotinylated GST-ALR1 (Fig. 2B, open
triangles) were immobilized on a microtiter plate and subsequently
incubated with varying concentrations of MRK-16. No detectable
difference in binding of MRK-16 to either form was detected, consistent
with biotinylation occurring at sites distal to the site of
antibody/peptide interaction.
Synthesis of Humanized MRK-16 Variants--
CDR grafting, or the
transfer of murine CDRs onto a human framework, is the most
straightforward approach to humanizing murine mAbs (17-19). However,
the resulting humanized mAbs often display diminished affinity because
certain murine framework residues serve a critical role either in
maintaining the conformation of the CDRs or in antigen interactions.
Previously, we have described a phage expression combinatorial approach
to humanizing mAbs (23). Briefly, the murine variable region sequence
is used to identify the most homologous human framework to serve as a
template. Framework positions where the amino acids differ between the
murine mAb and the human template are assessed individually and
residues that are potentially important for maintaining the full
binding activity of the mAb are characterized by synthesizing a
combinatorial antibody library that examines all possible combinations
of amino acids found at these locations in the murine parent mAb and
the human template. Thus, the complexity of the combinatorial library is 2x, where x is the number of differing framework
positions that potentially contribute to the binding activity of the
mAb.
The amino acid sequences of murine MRK-16 H and L chain variable
regions were used to identify homologous human variable region sequences to serve as templates for humanization (Fig.
3). CDR residues, as defined by Kabat
et al. (51, 52), are underlined and were excluded from
subsequent analysis. The H chain template LJ38 was missing the first
three amino acids of framework 1 and was 88% identical to MRK-16 at
the remaining 84 framework residues, differing at 10 positions (Fig. 3,
differences indicated by vertical lines). The L chain
template KV2D was 84% identical to the MRK-16 framework, differing at
13 of 80 framework residues (Fig. 3).
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 3.
Alignment of MRK-16 variable region and human
template sequences. The amino acid sequences of the murine MRK-16
H and L chain variable region were used to identify homologous human
variable regions. The numbering of residues and the definition of CDRs
(underlined) were based on Kabat et al. (51, 52).
Differences in sequence are indicated by vertical lines and
framework positions characterized in the combinatorial expression
library are marked with an asterisk.
|
|
Multiple parameters were considered in identifying framework positions
to include in the combinatorial library. For example, surface residues
not normally found on human antibodies are likely to contribute to the
immunogenicity of the humanized mAb. Thus, the majority of murine
MRK-16 framework residues predicted to be located on the surface based
on solvent exposure (58) were changed to the corresponding human
template amino acid. In addition, differing residues were analyzed for
potential to contact the opposite variable region domain in the
VH-VL interface, for predicted importance in
modulating CDR activity as defined by Studnicka et al. (59),
and for similarity of amino acids based on the charge and size of side chain.
Assessment of the potential importance of the residues at all framework
positions differing between murine MRK-16 and the homologous human
templates resulted in five positions being selected for further
characterization; L chain residues 12 and 100 and H chain residues 3, 44, and 77 (Fig. 3, asterisks). L chain residue 12 and H
chain residue 77 are predicted to be buried and the amino acids found
on the murine MRK-16 and human templates are dissimilar while L chain
residue 100 and H chain residue 44 are solvent exposed but potentially
involved in contacting the variable region of the other chain. Finally,
H chain residue 3 was included in the combinatorial library because the
sequence of the first three amino acids of the human H chain template
was unknown and residue 3 is typically glutamine in human germline
sequences. The remaining H and L chain framework positions were
identical to the human templates. Because all murine/human residue
combinations at five sites were tested the combinatorial framework
library contained 32 unique variants.
DNA from 20 randomly selected clones from the combinatorial framework
library was sequenced to characterize the mutagenesis efficiency at
each of the framework library sites. As expected, sequencing of random
clones demonstrated representation of both the murine and human
residues at all five framework positions (Table
I), although a modest overrepresentation
of variants expressing the human proline residue at L chain position 12 (75%) and the murine asparagine residue at H chain position 77 (75%)
was observed.
View this table:
[in this window]
[in a new window]
|
Table I
DNA sequencing of combinatorial framework residues of variants most
reactive with the ALR1 mimotope
Randomly selected variants were screened against biotinylated GST-ALR1
and the DNA from the most active clones was sequenced. Two variants,
clones 4 and 12, were identified multiple times based on reactivity.
|
|
Screening of Humanized MRK-16 Variants with Mimotope Fusion
Protein--
The precise combination of murine/human framework
residues that most closely preserves the conformation of the CDR
residues involved in antigen binding was identified based on affinity
screening of the combinatorial framework library. Randomly selected
MRK-16 variants from the library were expressed as soluble Fabs in the periplasmic space of small scale (<1 ml) bacteria cultures grown in a
96-well format. Although variable amounts of MRK-16 Fab were released
from the periplasmic space, uniform quantities were captured on a
microtiter plate coated with a limiting (saturable) quantity of goat
anti-human chain antibody, as described previously (24). Subsequently, the Fabs were incubated with biotinylated GST-ALR1 and
binding was detected with NeutrAvidin-alkaline phosphatase. The
relative affinities of the humanized variants were compared with
bacterially expressed chimeric MRK-16 and an irrelevant Fab based on
the colorimetric signal generated in the ELISA.
Approximately 200 randomly selected clones were screened individually
for mimotope binding. Eleven clones that displayed the strongest
binding based on the colorimetric signal were selected for further
characterization. DNA sequencing of the clones identified seven unique
framework sequences and two clones, 4 and 12, were identified multiple
times (n = 2 and 4, respectively), as summarized in
Table I.
Although DNA sequencing of 20 clones randomly selected from the library
demonstrated the diversity of the framework library, the frequency at
which certain clones were identified (clones 4 and 12, for example)
relative to other variants might reflect an abundance of these variants
in the library. Therefore, in order to more fully characterize the
mimotope binding, additional Fab was isolated from larger cultures of
clones 4 and 12, as well as from a clone which did not display
appreciable mimotope binding. Of the 32 distinct members of the
combinatorial framework library, one contains an entirely human
framework. Presumably, this CDR-grafted variant was not identified
during the screening of the mimotope because it binds the mimotope
weakly or because it is present in the library at an abnormally low
frequency. Nonetheless, because the simplest approach to antibody
humanization consists of CDR grafting, phage-expressed Fab consisting
of the murine MRK-16 CDRs grafted to an entirely human framework was
characterized in addition to clones 4 and 12. The variants were
titrated against immobilized GST-ALR1 fusion protein (Fig.
4A). Clones 4 (open circles), 12 (filled circles), and the CDR-grafted
variant (filled squares) all bound the immobilized mimotope
in a concentration-dependent fashion while the non-reactive
clone (open squares) and irrelevant clone (open
triangles) did not bind appreciably. Based on the titration
profile, each clone displayed different affinity for the mimotope with
clone 4 binding with the highest affinity, followed by clone 12, and
the CDR-grafted variant.
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 4.
Characterization of humanized MRK-16
variants. Phage-expressed Fab isolated from bacterial cultures of
clones 4 (open circles), 12 (filled circles),
CDR-grafted MRK-16 (filled squares), a non-reactive clone
(open squares), and an irrelevant antibody (open
triangles) was screened for reactivity with immobilized GST-ALR1
(A) or with fixed monolayers of cells overexpressing P-gp
(B). Fab binding was detected as described under
"Experimental Procedures." The assays were performed three times
and the data are the average of duplicate determinations from a single
representative experiment. The range of the values for the duplicate
determinations was <9%.
|
|
To determine if the strength of the binding interaction between
GST-ALR1 and MRK-16 variants correlated with binding strength to the
actual antigen, the same variants were titrated on cells expressing
elevated levels of P-gp. Human HCT-15 colorectal carcinoma cell
monolayers were fixed in 96-well microtiter plates, incubated with
varying amounts of phage-expressed MRK-16 variants, and binding was
quantitated colorimetrically following incubation with goat anti-human
-alkaline phosphatase conjugate (Fig. 4B). All of the
MRK-16 variants bound the fixed monolayers in a
concentration-dependent and saturable manner (Fig.
4B), while an irrelevant Fab control did not (Fig. 4B,
open triangles). Importantly, the relative affinity of these
clones for the antigen expressed on cells correlated with the titration
profiles observed with mimotope. Consistent with previous studies,
simple grafting of MRK-16 CDRs to the homologous human template
resulted in a humanized mAb with diminished binding, as demonstrated by
clone 4, which had two murine framework residues (L chain residues 12 and 100) and displayed stronger binding to the cell monolayers.
| |
DISCUSSION |
The premise of this study was that optimization of MRK-16 binding
to a peptide mimotope would lead to optimization of MRK-16 binding
toward the antigen, P-gp. The success of mimotope-guided humanization
relies on the ability of peptides to mimic the epitope recognized by
mAbs and consequently, peptide mimicry of the antigen must be reflected
by the specificity of peptide binding to the target mAb. Several lines
of evidence demonstrate the specificity of ALR1 binding to murine
MRK-16. First, using immobilized murine MRK-16 mAb ALR1 was
specifically enriched from a phage-expressed peptide library containing
as many as 109 unique peptides. In addition, the binding of
GST-ALR1 fusion peptide to MRK-16 was concentration-dependent
and saturable while binding of unrelated GST-peptide fusions was
undetectable. Finally, GST-ALR1 inhibited the binding of MRK-16 to
fixed and live cells expressing elevated levels of P-gp. Collectively,
these results are consistent with ALR1 binding at or near the CDR of
MRK-16.
Additional evidence that ALR1 mimicked the P-gp epitope closely was
provided by the screening of the library of humanized variants. The
primary amino acid sequences of the 32 combinatorial framework variants
are highly homologous, differing by five or fewer amino acids.
Nonetheless, the five framework positions examined in the combinatorial
humanization library were selected based on their potential to affect
the CDR structure and thus, the antigen binding of MRK-16. Despite the
high degree of primary sequence similarity between the 32 variants,
ALR1 binding to the humanized MRK-16 clones was variable. Differences
in mimotope binding were observed both in the primary screening of
random clones as well as in the titrations performed on select
variants. Thus, similar to antigen-antibody interactions the affinity
of the mimotope-antibody interaction was dependent on the framework
structure of the variable region.
Screening of 200 randomly selected clones identified seven distinct
framework sequences that appeared to bind the mimotope with higher
affinity. Of these, two variants, clones 4 and 12, were identified on
multiple occasions providing further evidence that the enhanced signal
observed during the initial screening of these particular clones was
not an artifact. Sequence analysis of the 11 most active variants
suggests that L chain residue 12 is a framework residue important for
maintaining the affinity of humanized MRK-16. DNA sequencing of 20 randomly selected clones revealed a potential bias at L chain position
12 in the library synthesis with 15/20 (75%) clones expressing the
human proline residue. However, DNA sequencing of the 12 clones
selected on the basis of mimotope reactivity revealed that 10/11 (91%)
clones expressed the murine serine residue. Because the method
described herein does not use affinity selection/enrichment methods the results suggest that serine is favored over proline at L chain position
12. Preferences for either murine or human residues at the other four
framework library positions were not observed.
Several lines of evidence suggest that certain combinations of
framework residues other than serine at L chain position 12 serve
critical roles in preserving the binding activity of humanized MRK-16.
First, the combinatorial library contains 16 distinct members that
express a serine residue at L chain position 12 yet only six of these
were identified based on binding assays. In addition, one of the active
variants, clone 21, expressed proline at L chain position 12. Finally,
the clone non-reactive with the mimotope expressed serine at L chain 12 and bound P-gp less effectively than other clones.
Upon more thorough characterization one of the humanized versions of
the murine mAb MRK-16 identified multiple times by screening with ALR1,
clone 4, displayed higher affinity for P-gp than did the CDR-grafted
construct, demonstrating that peptide mimicry of epitopes can be
exploited to generate reagents that enable humanization of mAbs.
Furthermore, although a very small number (n = 4) of
variants from the combinatorial framework library were characterized
fully by titration on both the mimotope and cells overexpressing P-gp
there was good correlation between the relative affinities of these
four variants for the mimotope as compared with P-gp.
Humanization of MRK-16 used a combinatorial framework library that
examined the contributions of five framework positions to the overall
affinity of the antibody. However, it is often necessary to construct
larger combinatorial libraries to characterize a greater number of
potentially important framework residues. Larger libraries increase the
likelihood of identifying fully active humanized variants because
characterization of a greater number of framework positions diminishes
the requirement for completely accurate structural modeling. However,
larger libraries are screened more efficiently if soluble labeled
antigen is available, which is often not the case. The ease with which
mimotopes are identified and expressed as soluble GST fusion proteins
enables the synthesis of large combinatorial antibody libraries by
removing screening restrictions. Although mimotope screening may not be
capable of identifying the optimal clone from large libraries it is
envisioned that it can be used as an initial screen to permit the
identification of a smaller number of variants for further
characterization. Thus, large combinatorial libraries can be reduced to
a few clones which can subsequently be screened by virtually any assay.
The use of surrogate antigen for antibody humanization is of greatest
utility if the mimotopes can be easily identified, are soluble, and can
be readily purified and labeled for use as a screening reagent. In the
present study, phage display peptide libraries were used for the rapid
identification of a mimotope and subsequently, the peptide was
expressed as a fusion protein with GST to enhance solubility and permit
its straightforward purification and biotinylation. A
priori, it was possible that the peptide identified by phage
display would not be active when fused to GST, due to different
interactions between the peptide and the pVIII or GST protein. However,
to date we have successfully transferred four of five mimotopes
identified from phage display libraries from pVIII to GST. The reasons
for our success are unclear but may reflect the expression of the
peptide libraries on the tip of pVIII, as opposed to other phage
surface proteins. F-specific filamentous phage are composed of five
different phage proteins, of which pVIII is the most abundant, with
thousands of copies per phage. The pVIII protein forms a highly
ordered, tightly packed coat while the remaining four phage proteins
are present at small copy number on the ends of the phage particle
(60). The large number of pVIII protein copies per phage reduces the
likelihood of peptides interacting among themselves. Moreover, the
highly ordered packing of pVIII may reduce the number (diversity) of interactions between peptides and phage surface molecules and thus,
increase the chance of successfully fusing the peptide to an unrelated
protein, such as GST. The ease with which mimotopes identified by
screening phage libraries of pVIII fusion peptides can be transferred
to GST enables the surrogate antigen strategy to potentially be applied
for the engineering of other mAbs.
Previously, peptide mimotopes have been identified for mAbs that bind
epitopes of diverse chemical nature and furthermore, mimotopes have
been used as antigen substitutes in multiple applications including
immunoassays, affinity matrices, elicit structural information, and as
immunogens. All these applications provide evidence that peptides can
in fact mimic epitope structure effectively and support the notion that
mimotope-guided humanization is potentially applicable to a wide range
of mAbs, even when the target antigen is unknown or difficult to
isolate. Mimotope-guided humanization of MRK-16 is one example of
antibody engineering in the absence of antigen. It is also possible
that peptides can be used as reagents for in vitro affinity
maturation of mAbs. However, amino acid substitutions in the CDRs that
improve mAb affinity for the peptide may be directed toward portions of
the peptide that are not structurally related to the antigen.
Consequently, although the resulting mAb may bind peptide with higher
affinity, binding to antigen may be unaltered or diminished. Therefore,
for affinity maturation purposes it might be beneficial to optimize the
mAb by screening against multiple mimotopes in parallel.
| |
ACKNOWLEDGEMENTS |
We gratefully acknowledge Herren Wu for
humanization design and critical review of the manuscript and Franz
Triana for design, expression, and purification of GST-peptides.
| |
FOOTNOTES |
*
The costs of publication of this
article were defrayed in part by the
payment of page charges. The article
must therefore be hereby marked
"advertisement" in
accordance with 18 U.S.C. Section
1734 solely to indicate this fact.
To whom correspondence should be addressed. Tel.: 619-597-4990;
Fax: 619-597-4950; E-mail: jwatkins@ixsysinc.com.
| |
ABBREVIATIONS |
The abbreviations used are:
mAb, monoclonal
antibody;
Fab, antigen-binding fragment;
H, heavy;
L, light;
CDR, complementarity-determining region;
ELISA, enzyme-linked immunsorbent
assay;
P-gp, P-glycoprotein;
GST, glutathione S-transferase;
BSA, bovine serum albumin;
PBS, phosphate-buffered saline;
MES, 4-morpholineethanesulfonic acid;
MOPS, 4-morpholinepropanesulfonic
acid.
| |
REFERENCES |
| 1.
|
Scheinberg, D. A.,
and Chapman, P. B.
(1995)
in
Monoclonal Antibodies: Principles and Applications
(Birch, J. R.
, and Lennox, E. S., eds)
, pp. 45-105, Wiley-Liss, New York
|
| 2.
|
Lefkovits, J.,
Ivanhoe, R. J.,
Califf, R. M.,
Bergelson, B. A.,
Anderson, K. M.,
Stoner, G. L.,
Weisman, H. F.,
and Topol, E. J.
(1996)
Am. J. Cardiol.
77,
1045-1051[CrossRef][Medline]
[Order article via Infotrieve]
|
| 3.
|
Queen, C.,
Schneider, W. P.,
Selick, H. E.,
Payne, P. W.,
Landolfi, N. F.,
Duncan, J. F.,
Avdalovic, N. M.,
Levitt, M.,
Junghans, R. P.,
and Waldmann, T. A.
(1989)
Proc. Natl. Acad. Sci. U. S. A.
86,
10029-10033[Abstract/Free Full Text]
|
| 4.
|
Gorman, S. D.,
Clark, M. R.,
Routledge, E. G.,
Cobbold, S. P.,
and Waldmann, H.
(1991)
Proc. Natl. Acad. Sci. U. S. A.
88,
4181-4185[Abstract/Free Full Text]
|
| 5.
|
Woodle, E. S.,
Thistlethwaite, J. R.,
Jolliffe, L. K.,
Zivin, R. A.,
Collins, A.,
Adair, J. R.,
Bodmer, M.,
Athwal, D.,
Alegre, M.-L.,
and Bluestone, J. A.
(1992)
J. Immunol.
148,
2756-2763[Abstract]
|
| 6.
|
Carter, P.,
Presta, L.,
Gorman, C. M.,
Ridgway, J. B. B.,
Henner, D.,
Wong, W. L. T.,
Rowland, A. M.,
Kotts, C.,
Carver, M. E.,
and Shepard, H. M.
(1992)
Proc. Natl. Acad. Sci. U. S. A.
89,
4285-4289[Abstract/Free Full Text]
|
| 7.
|
Presta, L. G.,
Chen, H.,
O'Connor, S. J.,
Chisholm, V.,
Meng, Y. G.,
Krummen, L.,
Winkler, M.,
and Ferrara, N.
(1997)
Cancer Res.
57,
4593-4599[Abstract/Free Full Text]
|
| 8.
|
Miller, R. A.,
Oseroff, A. R.,
Stratte, P. T.,
and Levy, R.
(1983)
Blood
62,
988-995[Abstract/Free Full Text]
|
| 9.
|
Schroff, R. W.,
Foon, K. A.,
Beatty, S. M.,
Oldham, R. K.,
and Morgan, A. C., Jr.
(1985)
Cancer Res.
45,
879-885[Abstract/Free Full Text]
|
| 10.
|
Ho, M-K.,
Rand, J.,
Murray, J.,
Kato, K.,
and Rabin, H.
(1985)
J. Immunol.
135,
3831-3838[Abstract]
|
| 11.
|
Borrebaeck, C. A. K.,
Danielsson, L.,
and Moeller, S. A.
(1988)
Proc. Natl. Acad. Sci. U. S. A.
85,
3995-3999[Abstract/Free Full Text]
|
| 12.
|
Winter, G.,
Griffiths, A. D.,
Hawkins, R. E.,
and Hoogenboom, H. R.
(1994)
Annu. Rev. Immunol.
12,
433-455[Medline]
[Order article via Infotrieve]
|
| 13.
|
Mendez, M. J.,
Green, L. L.,
Corvalan, J. R. F.,
Jia, X.-C.,
Maynard-Currie, C. E.,
Yang, X.-D.,
Gallo, M. L.,
Louie, D. M.,
Lee, D. V.,
Erickson, K. L.,
Luna, J.,
Roy, C. M.,
Abderrahim, H.,
Kirschenbaum, F.,
Noguchi, M.,
Smith, D. H.,
Fukushima, A.,
Hales, J. F.,
Klapholz, S.,
Finer, M. H.,
Davis, C. G.,
Zsebo, K. M.,
and Jakobovitz, A.
(1997)
Nat. Genet.
15,
146-156[CrossRef][Medline]
[Order article via Infotrieve]
|
| 14.
|
Morrison, S. L.,
Johnson, M. J.,
Herzenberg, L. A.,
and Oi, V. T.
(1984)
Proc. Natl. Acad. Sci. U. S. A.
81,
6851-6855[Abstract/Free Full Text]
|
| 15.
|
Neuberger, M. S.,
Williams, G. T.,
Mitchell, E. B.,
Jouhal, S. S.,
Flanagan, J. G.,
and Rabbitts, T. H.
(1985)
Nature
314,
268-270[CrossRef][Medline]
[Order article via Infotrieve]
|
| 16.
|
Boulianne, G. L.,
Hozumi, N.,
and Shulman, M. J.
(1984)
Nature
312,
643-646[CrossRef][Medline]
[Order article via Infotrieve]
|
| 17.
|
Jones, P. T.,
Dear, P. H.,
Foote, J.,
Neuberger, M. S.,
and Winter, G.
(1986)
Nature
321,
522-525[CrossRef][Medline]
[Order article via Infotrieve]
|
| 18.
|
Riechmann, L.,
Clark, M.,
Waldmann, H.,
and Winter, G.
(1988)
Nature
332,
323-327[CrossRef][Medline]
[Order article via Infotrieve]
|
| 19.
|
Verhoeyen, M.,
Milstein, C.,
and Winter, G.
(1988)
Science
239,
1534-1536[Abstract/Free Full Text]
|
| 20.
|
Hale, G.,
Clark, M. R.,
Marcus, R.,
Winter, G.,
Dyer, M. J. S.,
Phillips, J. M.,
Riechman, L.,
and Waldmann, H.
(1988)
Lancet
2,
1394-1399[Medline]
[Order article via Infotrieve]
|
| 21.
|
Khazaeli, M. B.,
Conry, R. M.,
and LoBuglio, A. F.
(1994)
J. Immunol.
15,
42-52
|
| 22.
|
Foote, J.,
and Winter, G.
(1992)
J. Mol. Biol.
224,
487-499[CrossRef][Medline]
[Order article via Infotrieve]
|
| 23.
|
Rosok, M. J.,
Yelton, D. E.,
Harris, L. J.,
Bajorath, J.,
Hellström, K.-E.,
Hellström, I.,
Cruz, G. A.,
Kristensson, K.,
Lin, H.,
Huse, W. D.,
and Glaser, S. M.
(1996)
J. Biol. Chem.
271,
22611-22618[Abstract/Free Full Text]
|
| 24.
|
Watkins, J. D.,
Beuerlein, G.,
Pecht, G.,
McFadden, P. R.,
Glaser, S. M.,
and Huse, W. D.
(1997)
Anal. Biochem.
253,
37-45[CrossRef][Medline]
[Order article via Infotrieve]
|
| 25.
|
Watkins, J. D.,
Beuerlein, G.,
Wu, H.,
McFadden, P. R.,
Pancook, J. D.,
and Huse, W. D.
(1998)
Anal. Biochem.
256,
169-177[CrossRef][Medline]
[Order article via Infotrieve]
|
| 26.
|
Wu, H.,
Beuerlein, G.,
Nie, Y.,
Smith, H.,
Lee, B. A.,
Hensler, M.,
Huse, W. D.,
and Watkins, J. D.
(1998)
Proc. Natl. Acad. Sci. U. S. A.
95,
6037-6042[Abstract/Free Full Text]
|
| 27.
|
Geysen, H. M.,
Rodda, S. J.,
and Mason, T. J.
(1986)
Mol. Immunol.
23,
709-715[CrossRef][Medline]
[Order article via Infotrieve]
|
| 28.
|
Scott, J. K.,
and Smith, G. P.
(1990)
Science
249,
386-390[Abstract/Free Full Text]
|
| 29.
|
Stephen, C. W.,
and Lane, D. P.
(1992)
J. Mol. Biol.
225,
577-583[CrossRef][Medline]
[Order article via Infotrieve]
|
| 30.
|
Balass, M.,
Heldman, Y.,
Cabilly, S.,
Givol, D.,
Katchalski-Katzir, E.,
and Fuchs, S.
(1993)
Proc. Natl. Acad. Sci. U. S. A.
90,
10638-10642[Abstract/Free Full Text]
|
| 31.
|
Hoess, R.,
Brinkmann, U.,
Handel, T.,
and Pastan, I.
(1993)
Gene (Amst.)
128,
43-49[CrossRef][Medline]
[Order article via Infotrieve]
|
| 32.
|
Hale, G.
(1995)
Immunotechnology
1,
175-187[CrossRef][Medline]
[Order article via Infotrieve]
|
| 33.
|
Murray, A.,
Sekowski, M.,
Spencer, D. I.,
Denton, G.,
and Price, M. R.
(1997)
J. Chromatogr.
782,
49-54[CrossRef]
|
| 34.
|
D'Mello, F.,
Partidos, C. D.,
Steward, M. W.,
and Howard, C. R.
(1997)
Virology
237,
319-326[CrossRef][Medline]
[Order article via Infotrieve]
|
| 35.
|
Miller, J. L.,
and Lyle, V. A.
(1996)
Proc. Natl. Acad. Sci. U. S. A.
93,
3565-3569[Abstract/Free Full Text]
|
| 36.
|
Stephen, C. W.,
Helminen, P.,
and Lane, D. P.
(1995)
J. Mol. Biol.
248,
58-78[CrossRef][Medline]
[Order article via Infotrieve]
|
| 37.
|
Rudolf, M. P.,
Vogel, M.,
Kricek, F.,
Ruf, C.,
Zurcher, A. W.,
Reuschel, R.,
Auer, M.,
Miescher, S.,
and Stadler, B. M.
(1998)
J. Immunol.
160,
3315-3321[Abstract/Free Full Text]
|
| 38.
|
Chargelegue, D.,
Obeid, O. E.,
Hsu, S. C.,
Shaw, M. D.,
Denbury, A. N.,
Taylor, G.,
and Steward, M. W.
(1998)
J. Virol.
72,
2040-2046[Abstract/Free Full Text]
|
| 39.
|
Steward, M. W.,
Stanley, C. M.,
and Obeid, O. E.
(1995)
J. Virol.
69,
7668-7673[Abstract]
|
| 40.
|
Stoute, J. A.,
Ballou, W. R.,
Kolodny, N.,
Deal, C. D.,
Wirtz, R. A.,
and Lindler, L. E.
(1995)
Infect. Immun.
63,
934-939[Abstract]
|
| 41.
|
Hamada, H.,
and Tsuruo, T.
(1986)
Proc. Natl. Acad. Sci. U. S. A.
83,
7785-7789[Abstract/Free Full Text]
|
| 42.
|
Gros, P.,
Croop, J.,
and Housman, D.
(1986)
Cell
47,
371-380[CrossRef][Medline]
[Order article via Infotrieve]
|
| 43.
|
Chen, C.,
Chin, J. E.,
Ueda, K.,
Clark, D. P.,
Pastan, I.,
Gottesman, M. M.,
and Roninson, I. B.
(1986)
Cell
47,
381-389[CrossRef][Medline]
[Order article via Infotrieve]
|
| 44.
|
Gerlach, J. H.,
Endicott, J. A.,
Juranka, P. F.,
Henderson, G.,
Sarangi, F.,
Deuchars, K. L.,
and Ling, V.
(1986)
Nature
324,
485-489[CrossRef][Medline]
[Order article via Infotrieve]
|
| 45.
|
Kartner, N.,
Riordan, J. R.,
and Ling, V.
(1983)
Science
221,
1285-1288[Abstract/Free Full Text]
|
| 46.
|
Gottesman, M. M.,
and Pastan, I.
(1993)
Annu. Rev. Biochem.
62,
385-427[CrossRef][Medline]
[Order article via Infotrieve]
|
| 47.
|
Tsuruo, T.,
Hamada, H.,
Sato, S.,
and Heike, Y.
(1989)
Jpn. J. Cancer Res.
80,
627-631[CrossRef][Medline]
[Order article via Infotrieve]
|
| 48.
|
Hamada, H.,
Hagiwara, K.,
Nakajima, T.,
and Tsuruo, T.
(1987)
Cancer Res.
47,
2860-2865[Abstract/Free Full Text]
|
| 49.
|
Haaparanta, T.,
and Huse, W. D.
(1995)
Mol. Div.
1,
39-52[CrossRef][Medline]
[Order article via Infotrieve]
|
| 50.
|
Huse, W. D.,
Stinchcombe, T. J.,
Glaser, S. M.,
Starr, L.,
MacLean, M.,
Hellström, K. E.,
Hellström, I.,
and Yelton, D. E.
(1992)
J. Immunol.
149,
3914-3920[Abstract]
|
| 51.
|
Kabat, E. A.,
Wu, T. T.,
Perry, H. M.,
Gottesham, K. S.,
and Foeller, C.
(1991)
Sequences of Proteins of Immunological Interest
, 5th Ed.
, United States Department of Health and Human Services, Washington, D. C.
|
| 52.
|
Kabat, E. A.,
Wu, T. T.,
and Bilofsky, H.
(1977)
J. Biol. Chem.
252,
6609-6616[Free Full Text]
|
| 53.
|
Kristensson, K.,
Glaser, S. M.,
Zebedee, S. L.,
Huse, W. D.,
Casadevall, A.,
and Scharff, M. D.
(1995)
Vaccines 95
, pp. 39-43, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY
|
| 54.
|
Kunkel, T. A.
(1985)
Proc. Natl. Acad. Sci. U. S. A.
82,
488-492[Abstract/Free Full Text]
|
| 55.
|
Kunkel, T. A.,
Roberts, J. D.,
and Zakour, R. A.
(1987)
Methods Enzymol.
154,
367-382[Medline]
[Order article via Infotrieve]
|
| 56.
|
Glaser, S. M.,
Yelton, D. E.,
and Huse, W. D.
(1992)
J. Immunol.
149,
3903-3913[Abstract]
|
| 57.
|
Georges, E.,
Tsuruo, T.,
and Ling, V.
(1993)
J. Biol. Chem.
268,
1792-1798[Abstract/Free Full Text]
|
| 58.
|
Padlan, E. A.
(1991)
Mol. Immunol.
28,
489-498[CrossRef][Medline]
[Order article via Infotrieve]
|
| 59.
|
Studnicka, G. M.,
Soares, S.,
Better, M.,
Williams, R. E.,
Nadell, R.,
and Horwitz, A. H.
(1994)
Protein Eng.
7,
805-814[Abstract/Free Full Text]
|
| 60.
|
Russel, M.
(1991)
Mol. Microbiol.
5,
1607-1613[CrossRef][Medline]
[Order article via Infotrieve]
|
Copyright © 1999 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
|
|
 
P. Steinberger, J. K. Sutton, C. Rader, M. Elia, and C. F. Barbas III
Generation and Characterization of a Recombinant Human CCR5-specific Antibody. A PHAGE DISPLAY APPROACH FOR RABBIT ANTIBODY HUMANIZATION
J. Biol. Chem.,
November 10, 2000;
275(46):
36073 - 36078.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
