B Cell Responses to a Peptide Epitope

IV. Subtle Sequence Changes in Flanking Residues Modulate Immunogenicity

Lalitha Vijayakrishnan,* Surojit Sarkar,* Rajendra P. Roy,' and Kanury V.S. Rao'*
We examined modulation of primary humoral responses to a model synthetic peptide immunogen,peptide PS1 CT3, asa consequence
of single aminoacid substitutions. Two analogues were employed, one
in which the amino-terminal histidine (His', peptide C28CT3)
and another in which an internal proline (Prol4, peptide C41CT3) were replacedwith glycine residues. Peptide C28CT3 displayed
markedly enhanced immunogenicity relative to peptide PS1 CT3 in BALB/c mice, whereas peptide C41CT3 was only poorly immunogenic. Nevertheless, in all three cases the mature polyclonal IgC responsewas predominantly directed against a tetrapeptide
segment of sequence Asp-Pro-Ala-Phe between positions 4 and7 of the sequence. While all three peptides proved equally capable
of priming Ag-specific Thcells, they, however, displayed significant differences in their abilities to recall T cell responses. Regardless
of the priming immunogen, in vitro challenge with either PSlCT3 or its analogues consistently gave a hierarchy of potencies as
C28CT3 > PSlCT3 > C41CT3. Thiscould also becorrelated with B cell recall responses in which an identical hierarchy was obtained
on restimulation of G41CT3-primed B cells in adoptive transfer experiments. Subsequent studies revealed that peptide-mediated
modulation of Th cell recruitment by Ag-primed B cells was probably due to differences in on-rates for engagement of B cell Ag
receptor by these analogues. This was despite
the factthat all three peptides displayed equally randomized conformations
in solution.
These studies indicate that even subtle variations in the flanking sequences can markedly influence the immunogenicity of B cell
epitopes. The journal of Immunology, 1997, 159: 1809-1 819.
Y

n systemic primary humoral responses, an encounter between

Ag and the host preimmune B cell repertoire results in activation of that B cell subset bearing appropriate surface immunoglobulin (sIg)* receptors. The principal outcome of such a
recognition is the induction of Ag-specific IgM Abs. Subsequent
developments include Ab isotype switch (usually predominantly to
IgG) and B cell differentiation into either plasma or sIgGt memory B cells (1, 2). In T-dependent responses, progression of Agactivated B cells through each of these stages is driven by Agspecific activated Th cells (3-5). B cells bind Ag through their sIg
receptors, which is then followed by endocytosis, Ag processing,
and presentation of appropriate fragments in the context of MHC
class I1 molecules (6). Such Ag-presenting B cells engage relevant Th cells in a cognate interaction, one consequence of which
is B cell proliferation and differentiation (4-6). While there appears to be some debate in the literature about the ability of Agpresenting B cells to prime naive T cells (7-16), a considerable
bodyof evidence suggests that the Ag-presenting function of
B cells is critical for sustaining B cell components of humoral
responses ( 17-25).
A variety of studies have demonstrated that Ags that bind specifically to Ig receptors on B cells are presented extremely effi'Immunology Group, lnternatlonal Centre ior Genetic Engineering and Biotech
nology, and 'National Institute of Immunology, New Delhi, India
Received iorpublicationNovember
19, 1997.

4, 1996.AcceptediorpublicationMay

The costs of publication oi this artlcle were defrayed in part by the payment of
page charges. This article must therefore be hereby marked advertisement
in
accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

'

Address correspondence and reprint requests to Dr. Kanury V.S. Rao, Immunology Croup, International Centre ior Genetic Engineering and Biotechnology,
Aruna Asaf Ali Marg. New Delhi-110067, India.

'

Abbrevlations used in thrs paper: slg, surface immunoglobulin: BCR, B cell

antigen receptor; HRPO, horseradish peroxidase; LNC, lymph node cells; CD,
circular dichroism: DPAF, a tetrapeptide segment spanning positions 4 to positions 7 within the antigens used here o i sequence Asp-Pro-Ala-Phe.
Copyright 0 1997 by The Amerlcan Association of Immunologists

ciently (reviewed in Ref. 6). Indeed, under conditions of limiting

Ag supply, B cells are probably the most efficient among the various Ag-presenting cell types. The bulk of available evidence supports the view that the Ig receptor functions primarily as a high
affinity trapping site for Ag for eventual presentation (6, 26, 27).
Thus, the ability of a given B cell to present a specific Ag can be
expected to be governed by a variety of factors. For example,
avidity of sIg for Ag, or more precisely its epitope on Ag, may be
expected to eventually define the ability of the B cell to recruit T
cell help. Other parameters such as stability of the receptor-Ag
complex, endocytic rate, Ag-dependent induction of costimulatory
molecules on the B cell surface, etc., may also be expected to
influence the Ag-presenting efficacy of B cells.
In addition to influencing quantitative aspects of humoral responses, our own recent results suggest that the interplay between
Ag-activated B and Th cells also plays a crucial role in defining the
Ab specificities selected for maturation during a primary response
(28,29). Using both recombinant and synthetic peptide model Ags,
we have demonstrated, in the mouse model, that while the IgM
response initially induced on primary immunization was composed
of heterogeneous specificities, class switch to the IgG isotype of
Abs was accompanied by a strict selection for only a limited subset
of these (28-30). In both of these systems, success or failure of an
Ag-activated B cell of a given fine specificity for retention in the
response was primarily determined by its ability to access T cell
help (28, 29). Thus, it would seem that individual clonotypes that
are activated upon initial encounter withAg are subjected to a
selection pressure in which the critical determinant is the efficacy
of these B cells to function as APCs. The Ag-presenting ability
was in turn found to be regulated by a variety of extraneous parameters that included affinity of B cell Ag receptor (BCR) for Ag
(29) and competition between BCR and secreted Ig for the corresponding epitope on Ag (28). It is the culmination of such processes that eventually defines the spectrum of Ab fine specificities
obtained in the later stages of a humoral response.
0022-1 767/97/$02.00

1810

We have continued these studies to further explore the regulation of epitope-specific B cell responses to polypeptide immunogens. In this report, we show that single amino acid substitutions
outside of the immunodominant domain of a model synthetic peptide immunogen, peptide PSlCT3, can have profound effects on
immunogenicity without altering the fine specificityof the IgG Ab
response invoked. Interestingly, these differences in immunogenicity did not appear to be due to any alteration in the levels of T
cell activation but, rather, are
due to varying abilities of these
analogues to engage the BCR and thereby differentially influence
Ag-specific B cell recruitment of T cell help. It therefore appears
that, in addition to T cell determinants, interactions between B cells
and their corresponding epitopesalso represents an important paraneter for defining immunogenicity of polypeptide immunogens.

Materials and Methods

Materials
Heavy chain-specific horseradish peroxidase (HRPO) labeled secondary
Abs and anti-mouse IgG (7-specific) were purchased from Sigma Chemical
Co. (St. Louis, MO). Magnetic beads for panning of T (Dynabeads Mouse
pan T) and B (Dynabeads Mouse pan B) cells were obtained from Dynal
(Oslo, Norway). ['HITdR (specific activity, 5 Ci/mmol) was purchased
from Amersham (Arlington Heights, IL). For multipin synthesis of peptides (31), both the noncleavable kits and F-moc amino acid derivatives
were purchased from Chiron Mimotopes Inc. (Victoria, Australia).

Peptide synthesis
Peptides were synthesized by the solid phase method (32.33) on a Milligen
9050 automated synthesizer using the F-moc chemistry (34). Crude peptides were purified to at least 95% by reverse phase HPLC on a C,, column
(15-pm 6 Pak; (Waters, Milford, MA) 19 X 300 mm) using an aqueous
gradient of 0 to 70% acetonitrile in 0.1 % trifluoroacetic acid. Identities of
all peptides were ascertained by amino acid analysis.
The overlapping hexapeptide panels were synthesized using a Multipin
noncleavable peptide kit (Chiron Mimotopes). strictly adhering to the protocol recommended by the manufacturer. After completing the synthesis,
all hexapeptides were routinely acetylated at the amino terminus with a
50/5/1 (v/v/v) mixture of dimethylformamide, acetic anhydride, and triethylamine. Side chain deprotection was accomplished over a 2-h period at
room temperature with a 3X/1/1/ (v/v/v) mixture of trifluoracetic acid.
ethanedithiol, and anisole.

Conjugation of peptides to anti-mouse IgG

Peptides were conjugated to anti-mouse IgC using a standard procedure
(35). Briefly, 1 mg each of peptide and anti-mouse IgC were dissolved in
a total volume of 450 p1 of PBS. To this was added, dropwise with mixing,
500 pl of a 0.2% glutaraldehyde solution in PBS. The solution was incubated with gentle mixing at room temperature for 1 h. Then, 240 pI of 1 M
glycine in PBS was added to quench the reaction over a further I -h incubation at room temperature. Any unconjugated peptide was removed by
overnight dialysis against PBS with a minimum of four changes. Concentration of each peptide in the solution of conjugate was estimated by incubating with an equal volume of a 1 % solution of trifluoroacetic acid in
water for 30 min at room temperature. Released peptide was then estimated
by quantitative reverse phase HPLC with dual detection at 254 nm and
215 nm.

Animals and immunization

Female BALB/c mice (6-X wk old) were obtained from the small animal
breeding facility at the National Institute of Nutrition (Hyderabad, India).
Immunizations were generally given i.p. with a dose of SO pglmouse as an
emulsion in CFA (except where indicated). For polyclonal sera, mice were
bled from the retroorbital plexus, and sera within a group were usually
pooled.

ELlSAs
Plates were coated with 2 pg of peptide/well in 100 pl of PBS, pH 7.2, at
37C for 3.5 h. Subsequently, they were blocked with 300 pl/well of a 5%
solution of fat-free dry milk powder in PBS at 37C for 1 h. Then, 100 pI
of the appropriate dilution of mouse antiserum was added and incubated at
37C for 1 h. After washing, bound Ab was detected with HRPO-labeled

MODULATION OF B CELL RESPONSES

secondary Ab (37C 1 h), followed by color development with o-phenylenediamine as chromogen. Absorbance was measured at 490 nm.
For competitive ELISA experiments, antisera were used at dilutions
representing 50% of titer value. Twofold higher concentrations of antiserum and competitor peptide were mixed in equal volume and incubated
for 10 min at room temperature. This was then added to duplicate wells at
100 pl/well. The remaining procedure was as described above.

ELlSAs for pin-bound hexapeptides

Hexapeptide sets synthesized on noncleavable pins were also evaluated for
Ab cross-reactivity by ELISA. For this, the protocol recommended by the
manufacturer was strictly followed. Primary Abs were diluted appropriately in PBS containing 2% BSA, 0.1% (v/v) Tween-20, and 0.1% (vlv)
Tween-20, and 0.1% sodium azide. Pins were incubated in 200 p l each of
Ab solution at 4C overnight with gentle shaking. Subsequently, they were
washed and subjected to a second round of incubation with appropriate
dilution of HRPO-labeled goat anti-mouse IgM or IgG at room temperature
for 1 h, again with gentle shaking. The chromogen used for detecting
bound Ab was 2,2'-azino-bis(3-ethylbenzthiazoline-6-sulfonicacid)diammonium, and the absorbance was measured at 405 nm after subtracting
background at 490 nm.

Enrichment of cell populations

For enrichment of T cells, lymph node cells (LNCs) from either immune or
nonimmune mice were first depleted of RBCs by lysis with ammonium
chloride. and subsequently, adherent cells were removed by panning on
plastic plates at 37C for 1 h. Nonadherent cells were collected and diluted
to a cell concentration of 5 X 10' cells/ml. B cells were depleted from this
population by two rounds of panning with Dynabeads mouse antikB220
(Dynal, 4 X IO* beaddml) following the recommended protocol of the
manufacturer.
For obtaining enriched B cells, mice immunized with peptide G41CT3
were boosted 3 wk later with a 50 pg/mouse doseof G41CT3 in PBS given
i.v. Three days later, the spleen was removed and depleted of RBCs by
ammonium chloride lysis, following which adherent cells were removed by
panning as described above. Nonadherent cells were diluted to 5 X IO7
cells and T cells were depleted by two rounds of panning with Dynabeads
anti-mouse Thy 1.2 (Dynal, 4 X 10* beads/ml) as recommended by the
manufacturer. The resultant enriched B cells were then treated with mitomycin C at a final concentration of 50 pg/ml at 37C for 20 min. Cells were
washed thoroughly in culture medium prior to use. When unprimed B cells
were required, an identical protocol was followed except splenocytes were
taken from nonimmune mice.

Lymphocyte proliferation assays

BALB/c mice were immunized S.C. at the base of the tail with 50 pg/mouse
of the relevant peptide as an emulsion in CFA. Seven days later, the mice
were killed and inguinal lymph nodes removed. LNCs, at 5 X I O 5 cells/
well, were cultured in quadruplicate in 200 p1 of RPMI 1640 containing
10% FCS, gentamycin (2 mg/L)and 2-ME (0.05 mM) in wells of a 96-well
plate, In addition, indicated concentrations of appropriate peptides were
also included as challenge Ag. Cultures were incubated at 37C in a humidified atmosphere of 5% carbon dioxide for 72 h. Theywere then pulsed
with 1 pCilwell of ['HlTdR for an additional 18 h prior to harvesting for
determination of radioactive TdR incorporation by liquid scintillation
counting. Values are presented as stimulation index, which represents the
ratio of mean counts obtained at a given concentration ofAg over that
obtained in the absence of any challenge Ag. Background counts generally
varied between 1000 to 2500 cpm.

Determination of on-rates
Equal volumes of mAb PC286 and relevant peptide in PBS were mixed at
room temperature, and time-dependent Ab binding in terms of quenching
of tryptophan fluorescence was continuously monitored over a 100-min
period in a Shimadzu RF-1501 spectrofluorimeter (Shimadzu Corporation,
Tokyo, Japan). The excitation wavelength used was 280 nm. and emission
was recorded at 330 nm. The final Ab concentration employed was 300
nM, whereas peptide was maintained at > 10-fold in excess over binding
sites (7.5 p,M) to ensure psuedo-first-order conditions. Extent of fluorescence quenching was used to determine unbound Ab concentrations as a
function of time. The log of concentration of unbound Ab was plotted vs
time, and the slope, which was obtained by linear regression
analysis, was used
to determine k,,,. The k,,, value was subsequently calculated by dividing k,,,
with peptide concentration. Values of k,,, presented are the mean ( i S D ) of
three independent determinations.

181 1

The Journal of Immunology

Results
Peptide PS 7 CT3 and its single-amino acid-substituted
analogues

As a follow up to our earlier studies (29), we have continued to use

the model synthetic peptide immunogen PSlCT3. Peptide PSICT3
represents a chimeric molecule containing a B cell epitope (segment PSI) derived from the large protein of the hepatitis B virus
surface Ag (36) and a well-characterized promiscuous T cell
epitope (segment CT3) from the circumsporozoite protein of the
malaria parasite P/asmodium falciparum (37). For reasons described earlier (38), the PS1 and CT3 segments were separated by
a spacerof two glycine residues. Prior studies have established that
the PS 1 segment does not encode a T cell determinant, at least for
the murine system (39, 40).
In the course of a study of immunogenicity with a variety of
amino acid-substituted analogues of peptide PSlCT3, we noticed
that there were two analogues that showed a marked difference in
behavior from the parent peptide. Both of these represented single
amino acid-substituted analogues of peptide PSlCT3 in which either the histidine residue at position 1 (peptide G28CT3) or the
proline residue at position 14 (peptide G41CT3) were substituted
by glycines. It must be emphasized here that both substitutions
were performed within the B epitope segment of the molecule,
whereas the T cell epitope sequence remained unaltered. Figure I A
gives the amino acid sequence of peptide PSICT3 and its two
analogues.
We have shown earlier that peptide PS ICT3 exists in a predominantly random structure in solution (29). To verify whether glycine substitutions in peptides G28CT3 and G41CT3 had any influence on secondary structural preferences, we examined the
circular dichroism (CD) spectra of these peptides. As shown in
Figure lB, all three peptides gave virtually identical CD spectra of
a negative peak with a minima at 199.3 nm. This finding indicates
that the degree of randomness (41) was retained in spite of glycine substitutions, with no new secondary structural constraint
imposed.
Immunogenicity of peptide PS 7 CT3 and its analogues

To assess the influence of glycine substitutions on immunogenicity, we immunized separate groups of BALB/c mice with a single
dose of either peptide PSICT3, G28CT3, or G41CT3, and serum
Ab responses were monitored periodically. Surprisingly, glycine
substitutions at both positions were found to have a profound but
opposite effect on the immunogenicity of the resultant analogues
(Fig. 2). While primary IgG Ab titers to peptide G28CT3 were at
least sixfold higher than that obtained against peptide PSlCT3,
that against peptide G41CT3 was severely diminished (Fig. 2). In
contrast to IgG Ab titers, early primary IgM responses were relatively unaffected with all three peptides eliciting comparable levels
of Abs (Fig. 2, inset). Finally, immunization of ndnu mice with
either of these peptides did not yield primary IgM or IgG antibodies, suggesting that both IgM and IgG stages of the primary response to these peptides were T dependent.
We also investigated whether Ab responses to these three immunogens were restricted to the B epitope segments only (i.e.,
either PSI, G28. or G41) or whether they were more broadly distributed. For this determination, we performed competitive inhibition experiments in which binding of day 28 sera from each of the
groups represented in Figure 2 to homologous immunogen was
examined in the presence of various inhibitors. The inhibitors used
were synthetic peptides representing the B epitope only (iz., PSI,
G41, or G28), the T epitope only (i.e., CT3), an equimolar mixture
of both, or the homologous immunogen. Results from these ex-

A
SEQUENCE

PEPTIDE

PS l a 3

HQLDPAFGANSTNPDGGDIEKKIAKMEKASSVFNVVNS

G28Cr3

GQLDPAFGANSTNPDEDIEKKIAKMEKASSVFNVVNS

G41CT3

HQLDPAFGANSTNGDGGDIEKKIAKMEKASSVFNVVNS

E
Y

a
0

-3.500E+01

L
\L:vi/ I

190.0
260.0
WAVELENGTH ( n m )

FIGURE 1 . Single amino acid glycine-substituted analogues of peptidePSlCT3andtheir

CD spectra. A, The amino acid sequence of
peptide PSlCT3
and
its analogues, where either His (peptide
G28CT3) or Pro14 (peptide G41CT3)wereindividuallysubstituted
with a glycine residue. The substituted glycine is shown in bold and
the two-glycine residue spacer is underlined. 5, TheCD spectra obtained for thesepeptides.Thesewererecorded
on a JASCOmodel
J710 spectropolarimeter (JASCO Corporation, Tokyo, Japan) over a
wavelength range of 190 to 260 n m , with a step resolution of 0.1 n m
and a scan speed of 50 nm/min. Peptides were dissolved at a final
concentration of 0.25 mdml in PBS before recording the spectra. -,
Indicates peptide PSlCT3;--, peptide G41CT3; ---,peptide G28CT3.

periments are shown in Figure 3. For all three antisera, nearly

identical inhibition profiles were obtained regardless of whether
the competitive inhibitor employed was the B epitope peptide or
the total homologous immunogen (Fig. 3). In contrast, no inhibition was observed when peptide CT3 was used as inhibitor,
whereas a mixture of B epitope peptide and CT3 yielded an inhibition pattern identical to that with B epitope peptide alone (Fig.
3). Collectively, these results indicate that, for all three analogues,
day 28 IgG Ab responses are, at least predominantly, directed
against the B epitope segment only. Also notable from Figure 3 is
that peptide G41CT3 induces a day 28 IgG Ab response thatis
between 5- to IO-fold lower in afinity relative to that induced by
either peptide PSlCT3 or G28CT3 (compare Fig. 3, panel C with
A and B). Thus, substitution of Pro for Gly at position 14 has both
a quantitative and a qualitative influence on the humoral response.
Given the relatively lower affinity of anti-G41CT3 IgG Abs, we
also performed a comparative evaluation of its affinity for the other
two peptide analogues. As shown in Figure 4, nearly identical
competitive inhibition profiles were obtained with all three peptides, indicating that day 28 anti-G41CT3 polyclonal antiserum

1812

MODULATION OF B CELL RESPONSES

pression of immunogenicity seen as a consequence of glycine substitution at either position I or position 14 simply represents quantitative alterations in humoral recognition an
of
identical
determinant within the PS1 segment of peptide PSlCT3.
Modulation of PSlCT3 immunogenicity by glycine
substitution is not due to altered Th cell activation

I
5
10
15
20
25
30
DAYS AFTER PRIMARY IMMUNIZATION

FIGURE 2. Relative immunogenicity of peptides PSlCT3 and its analogues in BALB/c mice. Groups of four BALB/c mice were each immunized with either peptide PS1 CT3 (O),
peptide C41CT3 (0)or peptide C28CT3 (A), and pooled sera obtained at indicated time points
were takenfor determination of peptide-specific IgC titers by ELISA
(see Materials and Methods). /met showspeptide-specific IgM response at day 5 after primary immunization. a, Peptide PSlCT3; 0,
peptide G41 CT3; B,peptide G28CT3.This figure is a representative of
five independent experiments.

does not discriminate significantly between peptide G41CT3 and

its analogues. From this we also infer that at least a predominant
component of the Ab fine specificities that comprise an antiG41CT3 response is directed against determinants common to all
three analogues and does not include the substituted positions.
Polyclonal IgG responses to all three immunogens are
monospecific

In our previous report, we showed that immunization of BALB/c

mice with peptide PS 1CT3 induced an early primary IgM response
that was composed of multiple fine specificities collectively spanning the entire PSI domain (29). However, subsequent class
switch to IgG Abs was accompanied by stringent selection for a
predominantly monospecific population directed against a tetrapeptide segment spanning positions 4 to 7 (sequence: DPAF)
(29). It was of interest, therefore, to ascertain the situation with
respect to peptides G28CT3 and G41CT3. For this we used a panel
of overlapping hexapeptides representing single amino acid displacements over either the PSI or its analogue sequences. Polyclonal Ab cross-reactivities with this hexapeptide panel was determined by ELISA. As shown in Figure 5 , qualitatively similar
results were obtained for primary IgM and IgG responses to all
three immunogens. While the early primary IgM response displayed multiple specificities, cross-reactivities of the mature IgG
response were restricted to within three overlapping hexapeptides
(sequences: QLDPAF, LDPAFG,and DPAFGA) between residues
2 and 9 in all three cases (Fig.5). Thus, similar to the situation with
PSICT3 (29), both peptides G28CT3 and G41CT3 also induce a
primary IgG response in BALB/c mice that is at least predominantly directed against the tetrapeptide sequence DPAF between
positions 4 and 7. It therefore appears that enhancement or sup-

A simple explanation for the observed effects of glycine substitution on immunogenicity could be that the analogues are altered in
their ability to activate Ag-specific Th cells. Although both substitutions performed were not within the Tepitope of peptide
PSlCT3, prior studies have shown that amino acid changes outside
of the T cell determinant nevertheless have dramatic effects on the
ability to activate T cells (42-44). We therefore conducted a series
of experiments to ascertain whether this could account for our
observations. As a first step, we examined the ability of these individual immunogens to prime Th cells specific for the determinant(s) encoded within the CT3 segment of PSlCT3; for this, individual groups of BALB/c mice were primed with either peptide
PSICT3, G28CT3 or G41CT3. For comparative purposes, an additional group was primed with a peptide representing the CT3
segment of peptide PSlCT3 (peptide CT3). Seven days later, inguinal lymph nodes were removed and the resulting cells (LNCs)
challenged with peptide CT3 in vitro. As shown in Figure 6A,
recall responses were generally poorer in LNCs from mice primed
with peptide PSlCT3 or its analogues relative to CT3 priming.
Nevertheless, priming with either PSlCT3 or its analogues yielded
comparable recall responses to peptide CT3 challenge, suggesting
that all three of these immunogens are equally proficient at priming
CT3-specific Th cells (Fig. 6A). The poorer priming ability of
these peptides relative to peptide CT3 may be indicative of a requirement for processing prior to presentation for Th cell
recruitment.
We also evaluated the ability of peptides PSlCT3 and its analogues to elicit recall responses in LNCs from mice primed with
peptide CT3. As shown in Figure 6B, all Ags tested gave nearly
identical proliferative responses, suggesting that both of the glycine substitutions performed in the PSI sequence did not alter the
ability of PS 1CT3 to recall a preprimed population of CT3-specific
T cells. Also, if processing of PSICT3 is indeed required for optimal presentation of the CT3 fragment, then the data shown in
Figure 6B may also be taken to indicate that peptides G28CT3 and
G41CT3 do not differ significantly, in terms of processability, from
peptide PSlCT3. Finally, the data shown in Figure 6 also indicate
that while peptide CT3 is more potent relative to peptide PS 1CT3
and its analogues in terms of priming for a CT3-specific T cell
response, they are all equally proficient at restimulating a CT3primed T cell population.
To further confirm the absence of altered T cell immunogenicity
as a consequence of glycine substitution, we performed an additional experiment in which LNCs from mice primed with each of
these immunogens were challenged in vitro with a common peptide. Figure 7A shows the results of an experiment in which LNCs
from mice primed with either PSlCT3, G28CT3, or G41CT3 were
challenged with peptide PSlCT3. As is evident, very similar lymphocyte proliferative responses were obtained in all three groups.
This was also true in parallel experiments in which the challenge
Ag was either G28CT3 (Fig. 7 B ) or G41CT3 (Fig. 7 0 . We interpret these data to suggest that priming with either of these Ags
activates T cell subsets sharing, at least predominantly, a common
spectrum of epitope specificities. Further, this result also extends
the observation made in Figure 6A that these three immunogens
prime Ag-specific T cells equally well.

1813

The journal of I m m u n o l o g y

810

20

6 810

20

C O M P E T I T O RC O N C E N T R A T I O N

810

20

OJM)

FIGURE 3. Anti-peptide polyclonal IgG responses are exclusively directed against the B epitope segment. Polyclonal day 28 antisera from each
of the groups listed in Figure 2 were diluted to50% of titer value and incubated with indicated final concentrationsof either a peptiderepresenting
only the homologous B cell epitope
(i.e., either PS1, C41, or G28(0),only the T cell epitope(CT3, A),a mixture ofthe two (A),
or the homologous
immunogen (i.e., either PS1CT3, G41 CT3, or G28CT3 (0))
as described in Materials and Methods. Subsequently, 100-pl aliquots were added in
duplicate wells coated with homologous Ag and bound Abs determined by ELlSA (Materials and Methods). A, Anti-PSlCT3 antiserum; B, antiG28CT.3 antiserum; c, anti-C41CT3 antiserum. Dilutions of individualsera used were 1:lo00 for anti-PS1 CT3,l :3000 for antiLG28CT3, and 1 : l o 0
for anti-G41 CT3. Results are representative of three separate experiments

A
1.21

PSlCT

s%Il
u
PSICT 3

0.4
0.4

10

20

30

40

COMPETITORCONCENTRATION(lJM1
FIGURE 4. Relativeaffinity of antiLC41CT3 antiserum for peptide
PSlCT3and its analogues. Diluted day 28 antiLG41CT3 polyclonal
serum was incubated with indicated concentrations of either peptide
PSlCT3 (O),
G28CT3 (A),
or G41 CT3 (0)as described in the legend of
Figure 3. Residual Ab binding to wells coated with peptide G41CT3
was determined byELSA (Materials and Methods). Data are presented
in terms of bound Ab obtainedat each competitor peptide concentration expressed as a percentage of that obtained in the absence of any
competitor.

Thus, the data given in Figures 6 and 7 collectively rule out the
possibility that the differences in humoral responses to peptides
PSICT3, G28CT3,andG41CT3 seen in Figure 2 arisedue to
altered immunogenicity at the T cell level as a consequence of
glycine substitution.
Peptides PS1 CT3, G28CT3, and G4 I CT3 differ in their
abilities to restimulate I cells primed with the whole
immunogen

An interesting feature of the data given in Figure 7 was that, while

T cell proliferative responses to acommon challenge Ag were
independent of priming immunogen, there were nevertheless quantitative differences in the ability of a given peptide to induce recall
responses (compare panels A , B, and C in Fig. 7). This, again, was
true regardless of the priming Ag. To confirm this, LNCs from
mice primed with either peptide PS1CT3, G28CT3, or G41CT3
were individually challenged with each of these immunogens and
lymphocyte proliferation measured. These results are depicted in

0.8

I
GIlCT3

0.6
0.4

0.3
H Q

L D P A F G A N

HEXAPEPTIDE SEQUENCE
FIGURE 5. Hexapeptide mapping of IgM and IgG responses elicited
by peptide PSlCT3 andits analogues in BALB/c mice. Day 5 (A)or day
PSlCT3,
28 ( 6 ) sera frommiceimmunizedwitheitherpeptide
G41CT3, or G28CT3 werescreened for either IgM (A)or IgG ( 6 )crossreactivity against a panel of overlapping
hexapeptides derived from the
homologous B cell epitope segment (i.e., either PS1, C41, or C28), as
described in Materials and Methods. The x-axis denotes each
hexapeptide as its N-terminal residue; only the PS1 sequence is given
for the sake of convenience. Dilutions of individual sera used were:
1/1000 for PSlCT3; 1/4000 for G28CT3; and 1/100 for G41CT3.

Figure 8. It is notable that, regardless of the Ag with which the

LNCs were primed, proliferation was always maximal with peptide G28CT3as recall Ag. The next most potent was peptide
PSICT3, followed by peptide G41CT3, in which responses were
barely significant. This hierarchy was common to LNCs primed
with all three Ags. For example, G41CT3-primed LNCs were
more responsive to in vitro challenge with G28CT3 relative to
challenge with peptide G41CT3, a situation identical to that observed with PSICT3- and G28CT3-primed LNCs.
Thus, combining these results with those given in Figure 7, it
would appear that while peptide PSICT3 and its analogues are

1814

MODULATION OF B CELL RESPONSES

FIGURE 6. Peptides PSlCT3, G41CT3, and

G28CT3 are equipotent at both priming CT3-specific T cells and eliciting recall responses from CT3
primed LNCs. A, Priming of CT3-specific T cells.
LNCs from miceimmunizedwith
either peptide
PSlCT3 (GI, G4lCT3 (01,C28CT3 [A),or CT3 (A)
were cultured in vitro in the presence of indicated
concentrations of peptide CT3 and lymphocyte proliferation measured as described in Materials and
Methods. Values are presented as stimulation index
(Materials and Methods) and are representative of
three independent experiments. Mean background
counts obtained in this experiment were 1680 cpm.
B, Recall of a CT3-primed T cell
response.LNCs
from mice immunized with peptide CT3 were CUItured in vitro in the presence of indicated concentrations of either peptide PS1CT3 [G),G41 CT3 (Oj,
or G28CT3 (A)as challenge Ag. Measurement of
lymphocyte proliferation and data presentation are
as described in Maferials and Methods. Mean background counts obtained were 1460 cpm; data are
from one of three separate experiments.

7t

.-

7
6
x

lJJ5
0

f
2 4

c
I-

d 5
1
a

A 3
3

z
L

25

50

75

100 125
25
50
75
50
A N T I G E NC O N C E N T R A T I O N f p M )
'

100

125

I-L
6
-I

z3

b-

ul

7 55 02 5

100

125 7 55 02 5

100

125

25

50

75

100

125

ANTIGEN CONCENTRATION (uM)

FIGURE 7. T cell recall response to either peptide PSlCT3 or its analogues is independent of the priming immunogen. LNCs from mice primed
with either PSI CT3 (O),
C41CT3 (O),or C28CT3(A)
were challenged in vitro with indicated concentrations
of either peptide PSI CT3 (A),G28CT3
(B), or G41CT3 (0.
Further protocol and data presentation is as described in Figure 6. Mean background counts obtained were201 0 cpm. Results
are representative of four independent experiments.

equally proficient at priming Ag-specific T cells, they nevertheless

differ in their ability to elicit T cell recall responses, at least in this
system. Also of interest is the fact that the relative efficacies of
these peptides to recall Ag-specific T cell responses parallels the
trend of the relative immunogenicity of these Ags seen in Figure 2.
Differential recall responses to peptide PS7 CT3 and its
analogues are due to Ag presentation by activated B cells

Figure 8 demonstrates that when T cells are primed with either

PS 1CT3, G28CT3, or G41CT3, maximal proliferative responses
are obtained on in vitro challenge with peptide G28CT3, followed
by peptide PSlCT3, and finally by peptide G41CT3, which was
only weakly stimulatory. However, this was in contrast to the earlier experiment with CT3-primed LNCs in which comparable proliferative responses were obtained with all three peptides (Fig. 6B).
Since the principal difference in these two experiments was either
the inclusion or exclusion of the B cell epitope segment in the
priming Ag, we reasoned that the hierarchy seen in Figure 8 could
represent a consequence of Ag-activated B cells functioning
as APCs.
To examine this possibility, we performed an experiment in
which G41CT3-primed LNCs were cultured in vitro with peptide

G28CT3 either in the presence or absence of varying concentrations of peptides representing the B cell epitope segments of either
peptide G28CT3 (peptide G28) or peptide G41CT3 (peptide G41).
A parallel set of wells containing a peptide of scrambled G28
sequence was also included as control. Our objective was to assess
T cell proliferation in a situation in which B cell sIg receptor recognition of peptide G28CT3 was competitively inhibited by the
presence of B cell epitope-only peptides. The results of such an
experiment are shown in Figure 9. Peptide G28CT3 induced proliferation of G41CT3-primed LNCs was clearly inhibited, in a
dose-dependent manner, by both peptide G41 and G28 but not by
the scrambled peptide. Maximal inhibition obtained was between
85 and 95% with an eightfold molar excess of the B epitope peptide (Fig. 9). We interpret these results as suggesting that the principal APCs in this system are the Ag-activated B cells. This is also
consistent with the data in Figure 6 showing that PSlCT3 and its
analogues were more efficient as recall Ags when LNCs were
primed with total immunogen rather than with the T cell determinant peptide CT3 only.
If, indeed, Ag-specific T cell recall responses to PS 1 CT3 or its
analogue peptide-primed LNCs were being mediated by Ag-specific B cells functioning as APCs, then it was logical to expect that

1815

The Journal of Immunology

ANTIGEN CONCENTRATION OJM)

FIGURE 8. Peptide PSlCT3 and its analogues differ in their ability to recall immunogen-primed T cell responses. LNCs from mice primed with
were each cultured in vitro with the indicated concentrations of
either peptide PSlCT3 (O),
either PS1CT3 (A), G28CT3 (g), or G41CT3 (0
G41CT3 (Oj, or G28CT3 (A),
and lymphocyte proliferationwas subsequently measured as described in Materials and Methods. Mean background
counts were 1750 cpm. Values presented are mean stimulation indices of quadruplicates (2SEj; results are representative of four independent
experiments.

COMPETlTOR CONCENTRATION ( p M )

FIGURE 9. Differential T cell recall abilities of peptide analogues is

a consequence of B cell-mediated Ag presentation. G41CT3-primed
LNCs were cultured in vitro with a 30 pM concentration of peptide
G28CT3 either in the presence or absence of indicated concentrations
of either peptide G28 (A),
G41 (O),or a control peptide of scrambled
C28 sequence (x). Lymphocyte proliferation obtained in these three
groups was determined as described in Materials and Mefhods, and
data were presented as stimulation index obtained at each competitor
concentration expressed as a percentage of that obtained in the absence of any competitor peptide. A stimulation index of 6.5 was obtained in the absence of competitor peptides (100%) against a backgroundof1400cpm.
Results presented are representative ofthree
separate experiments.

differences in Ag recognition by B cells may account for the differential T cell recall abilities of these peptides. We examined this
possibility in a relatively more defined experiment in which purified CT3-primed T cellswere cocultured with mitomycin C-treated
enriched splenic B cells from mice primed with peptide G41CT3.
Thisculture was subjected to antigenic challenge with optimal
concentrations of either peptide PSlCT3, G28CT3 or G41CT3 and
CT3-specific T cell proliferation measured as counts of [3H]TdR
incorporated. As shown in Figure 10, restimulation of CT3-primed
T cells was by far the most efficient when peptide G28CT3 was
present. This was followed by peptide PSlCT3, whereas the homologous immunogen, peptide G41CT3, was the least effective.

CHALLENGE ANTIGEN
FIGURE 10. Differentialrestimulation
of CT3-primedTcellsby
G41CT3-activated B cells In the presence the peptide analogues. MItomycin C-treated enriched splenic B cells
derived
from
either
mice were cocultured at 2.5 X
G41CT3-immunized (0)or naive (0)
10' cells/welI along with 2.5 X 10' enriched T cells from
LNCs of
mice primed with peptide CT3. In addition, a 50 pM concentration of
either peptidePSlCT3,C41CT3,
or G28CT3 was also includedin
quadruplicate wells in a total culture volume
of 200 PI.After 72 h, the
cells were pulsed with ['HITdR and radioactivity incorporated determined as described in Materials and Methods.Meanbackground
counts obtained were 1a40 cprn for primed and 1630 for unprimedB
cells. This figure is representative of twoseparate experiments; data are
the mean stimulation indices (+SE) of quadruplicate sets.

Such differences were not observed when unprimed B cells were

used (Fig. lo), suggesting that Ag priming of B cells was a necessary prerequisite for differential T cell recall.

1816

MODULATION OF B CELL RESPONSES

10

e
e
"a"
e

CHALLENGE ANTIGEN

Conjugation of peptidestoanti-mouse
IgG abolishes
their differential abilities to restimulate CT3-specific
T cells. Mitomycin
C-treated enriched splenic B cells derived from G41CT3 immune mice
were cocultured with enriched T cells from CT3-primed mice as described in Figure 10. In addition, 50 FM concentrations of the indioras conjugates
cated peptides were added either in free form (O),
with anti-mouse IgC (0).For the purposes of control, a 32-amino acid
residue irrelevant peptide derived from the ORF-2Agof
hepatitis E
virus (sequence: AGAGPRVRQPARPLCSAWRDQAQRPAVASRRR)
wasalso included. Further protocoland data presentation is as described in Figure 10. Mean background counts obtained were 2250
cpm. Data are representative of three separate experiments.
FIGURE 11.

Differential recall abilities of peptide PSlCT3 and its

analogues are a consequence of differential BCR-Ag
interactions

The data shown in Figure 10 indicate that CT3-primed T cells are

optimally restimulated by G41CT3-primed B cells in the presence
of peptide G28CT3 as challenge Ag, whereas the homologous Ag,
G41CT3, was the least effective. Given that these differences seem
to arrive as a consequence of Ag-primed B cells functioning as
APCs, we thought it likely that they may reflect differences at
the level of BCR-Ag interactions. To verify this, we prepared
individual conjugates of the peptides with anti-mouse IgG. It
was anticipated that such a conjugation would provide for an
alternate targeting pathway to B cells, bypassing the need for
peptide-specificrecognition toenableuptakeandsubsequent
presentation to Th cells.
Mitomycin C-treated enriched splenic B cells and enriched
lymph node-derived Tcells, both from G41CT3-primed mice,
were cocultured in vitro in the presence of either free peptides or
peptides conjugated to anti-mouse IgG, and lymphocyte proliferation was measured subsequently. As shown in Figure 11, proliferative responses to the free peptides were essentially similar to
those obtained earlier. However, conjugation of either peptide
PSlCT3 or its analogues to anti-mouse IgG abolished these differences in their B cell-dependent abilities to recall T cell responses. All three analogues were now equipotent (Fig. 11).
The data shown in Figure 11 suggest to us that the observed
hierarchy of G28CT3 > PSlCT3 > G41CT3 in recruiting T cell
help is probably a consequenceof differential engagement of either
PSlCT3 or its analogue peptides by the DPAF-specific Ig receptor
on Ag-primed B cells. Where DPAF is a tetrapeptide segment
spanning position 4 to position 7 within the antigene used here of
sequence Asp-Pro-Ala-Phe.

FIGURE 12. Peptide PS1 CT3 and its analogues also induce differential E cell recall responses from G41CT3-primed splenocytes. Day 21
splenocytes (5 X 107/mouse) from C41CT3-primedmicewere
injected i.v. in 0.5 ml of PES into the tailvein of irradiated (550 rad)
BALBlc recipients. At 16 h after transfer, the hosts werechallenged i.v.
with soluble forms of either peptide PSlCT3, C41CT3, or C28CT3 in
sterile PES. A n additional control Ag,V3MNCT3,was also included;
V3MN represents a 15-amino acid residue sequence derived from the
V3 loop of the M N isolate ofHIV-1 (sequence: RKRIHICPGRAFYTT).
Bloodwas collected 6 daysafterantigenic challenge and peptidespecific IgC Abs determined by ELlSA (Materials and Methods). A parallel set of experiments in which naive splenocytes were transferred
did not yield any IgG responses atthis time point. Data from individual
mice are shown and the mean is indicated. Results are representative
of two separate experiments.

Peptide PSI CT3 and its analogues also vary in their ability to
recall B cell responses

The results described thus far indicatethat recruitment of activated

Th cells by Ag-primed B cells is most efficient in the presence of
peptide G28CT3 and least efficient in the presence of peptide
G41CT3. Furthermore, this trend was maintained regardless of
whether the B cells were primed with peptide PSlCT3, G28CT3,
or G41CT3, implying differential B cell recognition of a common
determinant within these peptides. In addition, we also demonstrated earlier that murine humoral responses to all three peptides
are T dependent. It was therefore highly probable that Ag-dependent differences in T cell recruitment by B cells might account for
the differences in IgG Ab titers elicited by each of these three
analogous immunogens. To test this probability, we immunized
BALB/c mice with a single dose of peptide G41CT3, and the
spleens were collected 3 wk later. Primed splenocytes were adoptively transferred into irradiated BALB/c recipients, which were
then challenged with soluble forms of either peptide PSlCT3,
G28CT3, or G41CT3. Anamnestic IgG responses were monitored
6 days later, and the results are presented in Figure 12. In accordance with our results on T cell recall, anamnestic Ab titers were
maximal with peptide G28CT3 challenge, followed by PSlCT3,
whereas G41CT3, although it is the homologous immunogen, was
extremely inefficient. Thus, there is a direct correlation between
individual T and B cell recall propensities of peptide PS 1CT3 and
its two analogues.

?817

The journal of Immunology

2.4

2.2
2

Time (seconds)

FIGURE 14. On-rates of peptide binding to mAb PC286. The exper-

COMPETITOR CONCENTRATION (pM)

FIGURE 13. Relative binding ofmAb PC286 to peptide PS1CT3 and

its analogues. mAb PC286 at a final concentration of 120 ndml was
incubated with indicated concentrations of either peptide PSlCT3 (O),
G41CT3 (O),or G28CT3 (A) as described in Materials and Methods.
Subsequently, 1OO-pl aliquots were added in duplicate wells coated
with peptide PSlCT3, and bound Ab detected was byELlSA as described in the legend to Figure3. Results presented are a representative
of three separate experiments.

imental protocol employed for determination of on-rates is described

in Materials and Methods; data analysis was as previously described
(54). This figure shows linear regression plotsoflogof unbound Ab vs
time for peptides PSlCT3 (O), C41CT3 (O), and C28CT3 (A).Data
presented are from one of three independent runs.

any Ag can be recognized (45, 46). However, for generation of

effective Ab responses, Ag recognition must necessarily be followed by clonal amplification of responder B cells,somatic
mutations in Ig variable region genes, with subsequent selection
for higher affinity clonotypes and eventual differentiation into
Ab-secreting plasma or memory B cells (1, 2). In T-dependent
Peptides PSlCT3 and its analogues differ in their on-rates for
responses,thedriveforeach
of these stagesisprovided by
Ab binding
Ag-activated Th cells that are recruited by the relevant B cells
The cumulative evidence described thus far indicates that the hiin an Ag-presenting mode (17-25).
erarchical ability of peptides PSlCT3, G28CT3, and G41CT3 to
In the case of multideterminant Ags such as polypeptides, the
induce Ag-primed recruitment of T cell help may account for the
situation is somewhat more complex, with activation of B cells
differential immunogenicities of these molecules. However, thebadirected against the range of epitopes presented by Ag. This, howsis for differences in recognition of these Ag by primed B cells
ever, is probably true only for the early stages of a primary huremained unclear. This was particularly true given that there was
moral response. Subsequent progression appears to be accompano evidence for sequence-dependent structural alterations in these
nied by a combination of selection processes that serve to restrict
peptides as revealed by their CD spectra. Furthermore, we were
both heterogeneity and fine specificity of the B cells that are evenalso unable to detect any heteroclitic binding of these peptides by
tually retained (28, 29). While reducing levels of Ag supply has
anti-G41CT3 antisera. We therefore decided to investigate whether
been suggested as the driving force for competitiveB cell selection
the single amino acid substitutions performed on peptide PSlCT3
(47), our own results have identified at least two additional mechhad any influence on the kinetics of Ag-Ab interactions.
anisms. One of these is the competition between secreted Ig and B
Ourprior
observation that the hierarchy of efficacies as
cell mIg receptor for Ag binding (28). This, in turn, limits the
G28CT3 > PSICT3 G41CT3 was independent of priming imamount of available Ag for B cell binding and eventual presentamunogen suggested to us that the relative potencies of these anation for T help (28). Additionally, we have also recently demonlogues was an inherent property. For purposes of the present study,
strated that the limiting pool of activated Ag-specific Th cells in
therefore, we used a mAb that had previously been generated
the early stages of a primary response enforces a competitive inagainst peptide PSlCT3 and found to be directed against the
teraction between Ag-activated B cells, resulting in an Ag affinityDPAF determinant (mAb PC286 (29)) for determination of ondriven selection (29). Indeed, the latter mechanism appears to be
rates for binding to either peptide PSICT3 or its analogues. Prior
the principal one accounting for dominance of anti-DPAF Abs
to this, however, we first established that this mAb bound all three
over alternate specificities during maturation of a murine primary
peptides with comparable affinities (shown in Fig. 13). Subsehumoral response to PSlCT3 (29).
quently, on-rates for peptide binding were determined in a fluoThis report extends our previous observations by demonstrating
rescence quenching assay under pseudo-first-order conditions, as
that subtle sequence variations outside of the immunodominant B
described in Materials and Methods. The results of such an excell domain can dramatically influence immunogenicity, without
periment are presented in Figure 14 and the calculated k,, values
affecting fine specificities of the IgG Abs invoked. An obvious
were 2.7 2 0.1 X lo7 "Is"
for peptide G41CT3,4.2 ? 0.2 X
explanation, which could account for such a finding, is that these
I O ~ M - for
~ ~peptide
- ~ PSlCT3, and 9.0 ? 0.4X lo7 "'s"
for
changes produced an alteration in the ability of the resulting peppeptide G28CT3. As is evident, the on-rate for individual peptide
tides to stimulate/activate Ag-specific T cells. Indeed, a number of
binding by mAbPC286follows
the trend of G28CT3 >
prior studies have documented such influences in which sequence
PSICT3 > G41CT3. This is consistent with the observed efficachanges distal to T cell determinants within polypeptide Ags have
cies of these peptides in terms of B cell-mediated Th cell recruitbeen shown to modulate T cell activation potencies (42-44). Such
ment and relative immunogenicities, implying a direct causal
effects are thought to be mediated by an alteration in either peptide
relationship.
processing or in the affinityof the resulting fragment for MHC
class I1 binding (48). This explanation, however, did not appear to
Discussion
apply in our case, since all three peptides were found to be equally
The preimmuneB cell repertoire is generally thought to be com- proficient at both priming for a CT3-specific response and restimposed of such a diverse array of Ab specificities that virtually
ulating CT3-primed T cells. Furthermore, if processing of either

1818

peptide PS 1CT3 or its analogues is indeed a prerequisite for CT3

presentation, then the latter result would also suggest that there is
no significant difference in processing of these Ags, at least in the
context of presentation via nonspecific uptake of Ag.
Interestingly, however, peptide PSlCT3 and its analogues did
reveal differences in their abilities to induce ex vivo recall responses in LNCs primed with the whole immunogen. This hierarchy of relative potencies was independent of whether the priming
immunogen was peptide PSlCT3, G28CT3, or G41CT3; suggesting that the observed relative potencies of G28CT3 > PS 1CT3 >
G41CT3 were inherent properties of the peptides. These results
were in stark contrast to those obtained with CT3-primed LNCs,
implicating a role for the B cell epitope in Ag for differential behavior. Subsequent experiments verified this assumption by indicating that these differences in recall abilities were a consequence
of Ag-primed B cells functioning as APCs. Particularly notable
here was our finding that enriched G41CT3-primed B cells were
most efficient at restimulating CT3-primed T cells in the presence
of peptide G28CT3 as Ag. The observed hierarchy in T cell recall
experiments with G41CT3-primed LNCs was also reproducible in
B cell recall experiments in irradiated mice adoptively transferred
with G41 CT3-primed splenocytes, indicating a direct correlation
between the two findings. Collectively, we take these data to suggest that Ag-primed B cells are most efficient at recruiting T cell
help in the presence of peptide G28CT3 and least efficient with
peptide G41CT3. This, in turn, may explain the differences in immunogenicity obtained between these three analogue peptides in
Figure 2.
Since it was the APC activity of primed B cells that was being
differentially influenced depending on challenge Ag, we also characterized serum Ab responses to these three peptides. We hoped
that such studies might provide insights that could be extended to
B cell sIg receptor interactions with Ag. One interesting finding
was that, regardless of immunogenicity, immunization with either
peptide PSICT3, G28CT3, or G41CT3 induced, in all cases,a
primary IgG response that was at least predominantly directed
against the tetrapeptide segment between positions 4 and 7 (sequence DPAF). This could be taken to indicate that, for all three
immunogens, the dominant responder B cell populations during
the IgG phase of the response were those with sIg receptors for the
DPAF sequence within Ag. In other words, regardless of priming
or recall Ag used, the dominant B cell APC activity would be that
resulting from DPAF-mediated Ag capture by sIg receptors on
Ag-activated B cells. Consequently, it was reasonable to infer that
Ag-dependent modulation of the efficiency of primed B cells to
recruit T help may actually reflect differences in interaction of sIg
receptor with the DPAF segment in the context of the various
substitutions performed. Evidence in support of our inference was
obtained in experiments with anti-mouse IgG conjugates of the
peptides, which seemed to indicate differences in interaction of
G41CT3-primed B cells with the three analogues. It is possible that
such differences in BCR engagement may eventually influence the
density of the ligand presented and, consequently, Th cell recruitment. In this connection, it was also of interest that peptide
G41CT3 elicited a lower affinity anti-DPAF IgG response in comparison with that against either peptide PSlCT3 or G28CT3, implying a slower rate of affinity maturation. Thus, substitution of a
distal proline residue at position 14 for glycine does exert a qualitative influence, in addition to a quantitative one, on anti-DPAF B
cell responses.
How such minor changes in primary amino acid sequence could
so markedly influence cognate interactions between immunocompetent cells was an intriguing issue that remained to be resolved.
Our CD studies revealed that all three peptides were equally ran-

MODULATION OF B CELL RESPONSES

domized with respect to conformation in solution. This finding

ruled out induction/disruption of a secondary structural feature as
the causative factor. However, since it is well known that Ag-Ab
binding involves conformational accomodation at the level of both
entities (49, 50), it was possible that the substitutions performed
somehow altered the inducability of peptide (or DPAF) fit within
the paratope of Ab. Even if this were true, it was not however
reflected at the level of affinity of Ab for the three analogues. For
instance, anti-G41CT3 Abs were found to bind all three peptides
equally well, implying the absence of a heteroclitic effect.
In contrast to equilibrium binding, the kinetics of peptide binding to a representative Ab proved to be markedly different. The
obtained k,, values were in the order of G28CT3 > PS 1CT3 >
G41CT3. This hierarchy was consistent with that observed for
quantitative differences in Th cell recruitment by Ag-primed B
cells, recall of G41CT3-primed B cell responses, and relative immunogenicities at the level of a humoral response. We, therefore,
infer that these differences in binding on-rates directly influence
the ability of Ag-primed B cells to recruit T cell help, which in turn
determines the extent of B cell proliferation and levels of secreted
Ig. It is expected that kinetic differences in Ag binding would be
reflected at the level of internalized Ag within B cells, with its
consequences on density ofAg presented for TCR recognition.
The density of surface MHC class 11-associated ligand on an APC
is likely to be important in defining the rate at which that threshold
number of TCRs is triggered as a prerequisite to T cell activation
(51, 52). Prior studies have already established that on-rates represent a regulatory parameter for B cell selection during affinity
maturation of an Ab response (53, 54). Results presented here
suggest that on-rates ofAg recognition may also quantitatively
influence amplification of the responder B cells during the course
of an immune response.
In summary, the results presented in this report indicate that
minor changes in sequences flanking an immunodominant epitope
can markedly influence immunogenicity even in the absence of
any obvious structural alterations. Such changes seem to operate
by modulating the kinetics of B cell recognition of its corresponding epitope and thereby regulating its ability to access T cell help.
In addition, studies of the kind described here may also help provide further insights into the basis the underlying generation of
autoimmune B cell responses as a consequence of antigenic
mimicry.

Acknowledgments
Wearegrateful to Dr. V. Manivel and Mr. B. Ganesan for peptide synthesis and amino acid analysis. We also thank Mr. S. Kumaran (National
Institute of Immunology) for assistance with the on-rate measurements.

References
I . Nossal, G. J. V. 1992. The molecular and cellular basis for affinity maturation in
the antibody response. Cell 68:l.
2. Nossal, G. J. V. 1994. Differentiation of the secondary B lymphocyte repertoire:
the germinal center reaction, Immunol. Rev. 137:173.
3. Lanzdvecchia, A. 198.5. Antigen-speclfic interactlons between T and Bcells.
Nature 314:537.
4. Kishimoto. T., and T. Hirano. 1988. Molecular regulation of B lymphocyte response. Annu. Rev. Imrnunol. 648s.

5. Noelle, R. J.. and E. C . Snow. 1991. T helper cell-dependent B cell activation.

FASEB J. 5:2770.
6 . Lanzavecchia, A. 1990. Receptor-mediated antigen uptake and its effect on antigen presentation toclass 11-restricted Tlymphocytes. Annu. Rev. Immunol.
x: 773.
7. Kneger, J. I.. S. F. Crammer, H. M. Grey. and R. W. Chestnut. 1985. Antigen
presentation by splenic B cells: resting B cells are ineffective whereas activated
B cells are effective accessory cells for T cell responses. J. Immunol. 135:2937.
8. Hayglass, K. T., S. J. Naides, C. F. Scott, B. Benacerraf, and M-S. Sy. 1986. T
cell development in B cell-deficient mice. IV. The role of B cells as antigenpresenting cells in vivo. J. Imntunol. 136:823.

The Journal of Immunology

9. Janeway, C. J., Y. Ron, and M. E. Katz. 1987. The B cell is the initiating antigenpresenting cell in peripheral lymph nodes. J. lmmunol. 138:1051.
10. Kurt-Jones, E. A.,D.
Liano, K.A. Hayglass, B. Benacerraf, M-S. Sy, and
A. K. Abbas. 1988. The role of antigen-presenting B cells in T cell priming in
vivo: studies of B cell-deficient mice. J. Immunol. 140:3773.
11, Jenkins, M. K., E. Burrell, and J. D. Ashwell. 1990. Antigen presentation by
rest~ng B cells: effectiveness at inducing T cell proliferation is determined by
costlmulatory signals, not T cell receptor occupancy. J. Immunol. 144:1585.
12. Roncheae, F., and B. Hausmann. 1993. B lymphocytes in vivo fail to prime naive
Tcells but can stimulate antigen-experienced T lymphocytes. J. Exp. Med.
177.679.
13. Cassell,D. J.. and R. H. Schwartz.1994.A quantitative analysis of antigenpresentmg cell function: activated B cells stimulate naive CD4 T cells but are
inferior to dendritic cells in providing costimulation. J. Exp. Med. 180:1829.
14 Moms, S. C., A. Lees, and F. D. Finkelman. 1994. In vivo activation of naive T
cells by antigen-presenting B cells. J. Immunol. 152:3777.
15. Epstein, M. M., F. Di Rosa, D. Jankovic. A. Sher, and P. Matzinger. 1995.
Successful T cell priming in B cell-deficient mice. J. Exp. Med. 182:915.
16. Constant. S., N. Schweitzer, J. West, P. Ranney. and K. Bottomly. 1995. B lymphocytes can he competent antigen-presenting cells for priming CD4+ T cellsto
protein antigens in vivo. J. lmmunol. 155:3734.
17. Snow, E. C..R. J. Noelle. I. W. Uhr, and E. S. Vitetta. 1983. Activation of
antigen-enriched B cells. 11. Role of linked recognition in B cell proliferation to
thymus-dependent antigens. J. Immunol. 130:614.
18. Noelle. R. J.. E. C.Snow, J. W. Uhr, and E. S. Vitetta. 1983. Activation of
antigen-specific-B cells: role of T cells, cytokines. and antigen in induction of
growth and differentiation. Proc. Natl. Acud. Sci. USA 80:6628.
19. Noelle. R.J., J. Daum, W. C. Bartlett, J. McCann, and D. M. Shepherd. 1991.
Cognate Interactions between helper T cells and B cells. V. Reconstitution of T
helper function using purified plasma membranes from activated Thl and Th2
helper cells and lymphokines. J. Immunol. 146:1118.
20. Croft, M.. and S. L. Swain. 1991. B cell response to T helper subsets. 11. Both the
stage of T cell differentiation and the cytokines secreted determine the extent and
nature of helper activity. J. Im?nuno/. 147:3679.
21. McHeyzer-Williams, M. G., M. J. McLean, P. A. Lalor, and G. J. V. Nossal.
1993. Antigen-driven B cell differentiation in vivo. J. Exp. Med. 178:295.
22. Van den Eertwegh, F., R. J. Noelle, M. Roy, D. M. Shepherd, A. Aruffo,
J. A. Ledbetter, W. J. A. Boersma, and E. Claassen. 1993. In vivo CD40-gp39
interactions are essential for thymus-dependent humoral immunity. I. In vitro
expression of CD40 ligand. cytokines, and antibody production delineates sites of
cognate T-B cell interactlons. J. Exp. Med. 178:lSSS.
23. Foy. T. M., D. M. Shepherd, F.H. Dune, A. Aruffo, J. A. Ledhetter, and
R. J. Noelle. 1993. In vivo CD40-gp39 interactions are essential for thymusdependent humoral immunity. 11. Prolonged suppression of the humoral immune
response by an antibody to the ligand for CD40, gp39. J. Exp. Med. 178:1567.
24. Foy. T. M., J. D. Laman, J. A. Ledhetter, A. Aruffo, E. Claassen, and R. J.Noelle.
1994. gp39-CD40 interactions are essential for germinal center formation and the
development of B cell memory. J. Exp. Med. 180:157.
25. Han, S., K. Hathcock, B. Zhang. T. B. Kepler, R. Hodes, and G. Kelsoe. 1995.
Cellular interaction in germinal centers: roles of CD40 ligand and B7-2 in established germinal centers. J. Immunul. 155:556.
26. Rock, K. L., B. Benacerraf, and A. K. Abhas. 1984. Antigen presentation by
hapten-specific B lymphocytes. 1. Role of surface immunoglobulin receptors.
J. Exp.Med.160:1102.
27. Lichtman, A. H.. H-P. Tony,D.C.Parker,
and A. K. Abbas. 1987. Antigen
presentation by hapten-specific B lymphocytes. IV. Comparative ability of B cells
to present specific antigen and anti-immunoglobulin antibody. J. lmmunol.
138:2822.
28. Vijayakxishnan. L., V. Kumar. J. N. Agrewala. G. C. Mishra, and K. V. S. Rao.
1994. Antigen-specific early primary humoral responses modulate immunodominance of B cell epitopes. J. Immunol. 153:1613.
29. Agarwal, A., S. Sarkar, C. Nazabal, G. Balasundaram, and K. V. S. Rao. 1996.
B cell responses to a peptide epitope. I. The cellular basis for restricted recognition. J. Immunol. 157:2779.
30. Tuteja, R., A. Aganval, L. Vijayakrishnan, B. P. Nayak, S. K. Gupta, V. Kumar,
and K. V. S. Rao. 1997. B cell responses to a peptide epitope. 11. Multiple levels
of selection during maturation of primary responses. Immunol. CeliBiul. In press.

1819
31. Geysen. H. M., S. J. Rodda. T. J. Mason, G. Tribbick, and P. G. Schoofs. 1987.
Strategies for epitope analysis using peptide synthesis. J. Immunol. Methods
302:259.
32. Memfield, R. B. 1986. Solid phase synthesis. Science 232:34l.
33. Stewart, J. M., and J. D. Young. 1984. Solid Phase Peptide Synthesis, 2nd Ed.
Pierce Chemical Co., Rockford, IL.
34. Atherton, E., and R. C. Sheppard. 1989. Solid Phase Peptide Synthesis: A Practicul Approach. IRL Press, Oxford, U.K.
35. Harlow, E., and D. Lane. 1988. Antibodies: A LuboratoT Manual. Cold Spring
Harbor Laboratory, Cold Spring Harbor, NY, p. 78.
36. Neurath, A. R., and S . B . H. Kent. 1988. The pre-S region of hepadna virus
envelope protein. Adv. Virus Res. 34:64.
37. Sinigaglia, F., M. Guttinger, J. Kilgus, D.M.Doran, H. Matile, H. Etlinger,
A. Trezciak, D. Gillessen, and J. R. L. Pink. 1988. A malaria T cell epitope
recognized in association with most mouse and human MHC class I1 molecules.
Nature 336:778.
38. Rao. K. V. S., and A. R. Nayak. 1990. Enhanced immunogenicity of a sequence
derived from hepatitis B virus surface antigen in a composite peptide that includes the immunostimulatory region from human interleukin-I. Pruc.Natl.
Acad. Sci. USA 87:5519.
39. Milich, D.R., A. McLachlan, M. K. McNamara, F. V. Chisari, and G. B.
Thornton. 1986. T-cell and B-cell recognition of native and synthetic pre-S region
determinants on hepatitis B surface antigen. In Vaccines 86. Cold Spring Harbor
Laboratory, Cold Spring Harbor, NY, p. 377.
40. Milich, D. R. 1988. T- and B-cell recognition of hepatitis B viral antigens. lmmunol. Today 9:380.
41. Dyson, J. H., and P. E. Wright. 1991. Defining solution conformations of small
linear peptides. Annu. Rev. Biophys. Biophys. Chem. 20:519.
42. Bhayani, H., F.R. Carhone, and Y. Paterson. 1988. The activation of pigeon
cytochrome c-specific T cell hybridomas by antigenic peptides is influenced by
non-native sequences at the amino terminus of the determinant. J. Immunol.
141:377.
43. Vacchio. M. S., J. A. Berzofsky, U. Krzych, J. A. Smith, R. 1. Hodes, and
A. Finnegan. 1989. Sequences outside a minimal immunodominant site exert
negative effects on recognition by staphylococcal nuclease-specific T cell clones.
J. lmmunol. 143:2814.
44. Liu, Z.. K. P. Williams, Y-H. Chang. and J. A. Smith. 1991. Single amino acid
substitution alters T cell determinant antigen processing of Staphylococcus uureus nuclease. J. Immunol. 146:438.
45. Paige, C. J., and G. E. Wu. 1989. The B cell repertoire. FASEB J. 3 : 8 / 8 .
46. Bona, C. A. 1992, Expression of V gene families during ontogeny and establishment of the B cell repertolre. lnr. Rev. lmmunol. R:83.
47. Jacob, J., K. Ramtin. and G. Kelsoe. 1991. In situ studies of the primary immune
response to (4-hydroxy-3-nitropheny1)acetyl.I. The architecture and dynamics of
responding cell populations. J. Exp. Med. 173:1165.
48. Sercarz, E. E., P. V. Lehman, A. Ametani, G. Benichou. A. Miller, and
K. Moudgil. 1993. Dominance and crypticity of T cell antigenic determinants.
Annu. Rev. lmmunol. 11:729.
49. Geysen, H. M.. J. A. Tainer. S. J. Rodda,T.
J. Mason, H. Alexander.
E. D. Getzoff, and R. A. Lerner. 1987. Chemistry of antibody binding to a protein.
Scwnce 235:I 1 84.
50. Getzoff, E., H.M. Geysen, S. J. Rodda, H. Alexander, J. A. Tainer, and
R. A. Lerner. 1987. Mechanisms of antibody binding to a protein. Science
235:1191.
51. Valitutti, S., S. Muller, M. Cella, E. Padovan. and A. Lanzavecchia. 1995. Serial
triggering of many T cell receptors by a few peptide-MHC complexes. Nature
375:148.
52. Viola, A,, and A. Lanzavecchia. 1996. T cell activation determined by T cell
receptor number and tunable thresholds. Science 273:104.
53. Foote,J., and C. Milstein. 1991. Kinetic maturation ofan immune response.
Nature 352:530.
54. Roost, H.-P., M. E. Bachmann, A. Hang, U. Kalinke, V. Pliska. H. Hengartner,
and R. M. Zinkemdgel. 1995. Early high aitinity neutralizing anti-viral IgG responses without further overall improvements of affinity. Proc. Natl. Acad. Sei.
USA 92:1257.