
==== Front
Immunohorizons
Immunohorizons
IH
ImmunoHorizons
2573-7732
AAI

39093310
immunohorizons_2400055
10.4049/immunohorizons.2400055
Adaptive Immunity
Antigen-driven Convergent Evolution of Polysaccharide-specific “DH-less” B Cells in Glycoconjugate Immunized Mice
https://orcid.org/0000-0002-9819-4100
Kushwaha Sachin
https://orcid.org/0009-0005-5611-4802
Shome Pratiksha
https://orcid.org/0000-0002-4869-7105
Sehgal Devinder
Molecular Immunology Laboratory, National Institute of Immunology, Aruna Asaf Ali Marg, New Delhi, India
Address correspondence and reprint requests to: Dr. Devinder Sehgal, Molecular Immunology Laboratory, National Institute of Immunology, Aruna Asaf Ali Marg, New Delhi, India. E-mail address: devinder@nii.ac.in
8 2024
02 8 2024
8 8 511526
08 7 2024
10 7 2024
Copyright © 2024 The Authors
2024
https://creativecommons.org/licenses/by-nc-nd/4.0/ This article is distributed under the terms of the CC BY-NC-ND 4.0 Unported license.

Abstract

Glycoconjugate vaccines elicit robust anti-polysaccharide Ab response by recruiting T-cell help. Multiple doses of glycoconjugate vaccine are required to induce long-lasting immunity. The characteristics of anti-polysaccharide Ab response have been reported previously. However, the effect of glycoconjugate booster immunization on anti-polysaccharide and anti–carrier protein Ab repertoire remains poorly understood. In this study, we used clinically relevant pneumococcal capsular polysaccharide type 14 (PCP14) conjugated with cross-reactive material 197 (CRM197) as a model glycoconjugate Ag (PCP14-CRM197). We performed a comprehensive sequence analysis of mouse mAbs generated against PCP14 and CRM197 following immunization with one or three doses of PCP14-CRM197. Analysis of the paired Ig H and L chain transcripts revealed that anti-PCP14 Ab repertoire is extremely restricted. The reoccurrence of five replacement mutations at identical positions in anti-polysaccharide mAbs generated from different mice provided evidence for Ag-driven selection in PCP14-specific B cells. Convergent evolution was observed wherein distinct V(D)J rearrangements resulted in identical or nearly identical CDR3 in anti-PCP14 mAbs. Abs that lacked DH encoded amino acids dominated the anti-PCP14 Ab response. In contrast, anti-CRM197 Ab response was quite diverse, with fewer mutations compared with the anti-PCP14 mAbs, suggesting that conjugation of the polysaccharide to a carrier protein interferes with the development of carrier protein–specific Ab responses. Our findings provide molecular insights into the maturation of Ab responses driven by booster doses of glycoconjugate. This has fundamental implications for the design of glycoconjugate vaccines, especially where the development of Ab response against the carrier protein is also crucial.

10.13039/ BT/COE/34/SP15189/2015
==== Body
pmcIntroduction

The polysaccharide-based capsule is the primary virulence component of encapsulated bacterial pathogens. Abs directed against the bacterial capsular polysaccharide have been shown to confer protection. For this reason, capsule-derived polysaccharides are considered good targets for developing vaccines against encapsulated bacterial pathogens. Glycoconjugate vaccines against encapsulated bacteria have contributed significantly to the decline in the incidence of bacterial pneumonia and meningitis. To generate glycoconjugate vaccine, the bacterial polysaccharide is chemically coupled to a carrier protein. At present, glycoconjugate vaccines are licensed for use against the human bacterial pathogens Streptococcus pneumoniae, Haemophilus influenzae type B, and Neisseria meningitidis. The glycoconjugate vaccines were developed to overcome the shortcomings of the unconjugated polysaccharide vaccines (1, 2). Rubinstein et al. (3) reported that mice primed with meningococcal group C capsular polysaccharide–tetanus toxoid (MCPS-TT) glycoconjugate showed enhanced Ab response following booster shots. Not only were the anti-polysaccharide Ab levels elevated but showed superior serum bactericidal titers during recall response.

Unconjugated polysaccharides are considered T cell–independent Ags (4). Coupling the polysaccharide to a carrier protein imparts to the anti-polysaccharide Ab response a T cell–dependent character. A robust anti-polysaccharide B cell memory compartment and isotype class switching was induced on administering booster doses of the cross-reactive material 197 (CRM197)–conjugated pneumococcal capsular polysaccharide type 14 (PCP14) and PCP19F (5). The carrier protein contributes T-cell epitopes to trigger a robust germinal center reaction and expansion of polysaccharide-specific B cells. Avci et al. (6) proposed a model to explain how glycoconjugate vaccines induce polysaccharide-specific T-cell responses. These authors showed that the glycoconjugate Ag is taken up, processed into small glycan peptides, and presented in the context of MHC class II on APCs. The carrier protein–derived peptide part of the glycan peptide binds to the MHC class II molecule making the glycan component available for recognition by T helper cells (7).

To get a detailed understanding of the anti-polysaccharide Ab response elicited in response to immunization with a glycoconjugate Ag, various groups have resorted to making mAbs. Yano and Pirofski (8) generated and characterized anti-PCP8 mAbs from mice immunized with PCP8 conjugated to TT (PCP8-TT). The authors observed that the mAbs were diverse, and there was no evidence of somatic hypermutation. Tian et al. (9) reported that the gene usage in anti-PCP3 mAbs raised from mice administered PCP3-TT was diverse and mutated. The anti-polysaccharide Ab response to MCPS-TT glycoconjugate was found to be oligoclonal (10). Mukherjee et al. (11) molecularly characterized the Ab response to Cryptococcus neoformans–derived glucuronoxylomannan chemically coupled with TT (GXM-TT) in BALB/c mice. All the anti-GXM mAbs generated from the GXM-TT immunized mice were found to be highly restricted and used the same VH7183 family member. All the mAbs used the same Vκ5.1 and the Jκ1 gene segment and had the same Vκ–Jκ junction. The same VH and VL restriction was found in the anti-GXM Ab response in GXM-TT immunized NZB/W and C3H mice (12).

Glycoconjugate vaccines are effective in all age groups including infants. We do not fully understand why they work and the molecular details of the Ab response elicited against glycoconjugate vaccines. Because a single dose of the glycoconjugate is insufficient, two or more shots are typically required to elicit Abs of high quality and in sufficient quantity to achieve long-term protection. How the anti-polysaccharide and anti–carrier protein Ab response evolves in the spleen following booster shots of glycoconjugate vaccine is not documented. To our knowledge, a direct comparison of the molecular events that are targeted to the rearranged V region in the anti-polysaccharide and anti–carrier protein Ab genes has not been done. In this article, we show that the anti-polysaccharide Ab response is highly restricted. The polysaccharide-specific B cell underwent somatic hypermutation and class switching upon re-exposure to the Ag, hallmarks of T cell–dependent response. Remarkably, a vast majority of the anti-polysaccharide Abs do not have any DH gene segment–encoded amino acids. In contrast to the anti-polysaccharide Ab response, anti–carrier protein Ab response was diverse but with germline or near germline sequences. The glycoconjugate induced clonal expansion of polysaccharide reactive B cells and exhibited a highly conserved CDR3 signature. This study provides molecular insights regarding how the immune system concomitantly responds to the polysaccharide and carrier protein component of the glycoconjugate vaccine. This information will be valuable in guiding better vaccine design and developing novel glycoconjugate vaccines against encapsulated bacterial pathogens.

Materials and Methods

Immunization

Six- to eight-week-old female BALB/c mice (originally sourced from The Jackson Laboratory [USA] and reared at the Small Animal Facility at the National Institute of Immunology) were immunized with PCP14 chemically coupled to CRM197 (PCP14-CRM197) glycoconjugate Ag, i.p. PCP14-CRM197 was custom synthesized (Fina Biosolutions, USA). All animal experimental procedures were approved by the Institutional Animal Ethics Committee and were performed following the guidelines laid out by the Committee for Control and Supervision of Experiments on Animals of the Government of India.

For the primary immunization, the mice (n = 10) received a single bolus of glycoconjugate (5.175 µg per mouse; the PCP14 to CRM197 ratio was 1:1.07) with 25 µg of alum in 100 µl of PBS. For the tertiary immunization, the mice were given three shots of the glycoconjugate with alum on days 0, 14, and 28. Control groups received only alum.

Determination of serum anti-PCP14 and anti-CRM197 antibody titer

Serum anti-PCP14 and anti-CRM197 Ab titers were measured by indirect ELISA using purified PCP14 (catalog no. 76943, SSI Diagnostica, Denmark) and CRM197 (Fina Biosolutions) as the coating Ags, respectively. Preimmune sera were collected retro-orbitally 3 d prior to primary immunization. The mice were bled retro-orbitally at weekly intervals following immunization. The sera obtained from the group were pooled, and 2-fold serial dilutions (1:200 to 1:2,04,800) were done in PBS containing cell wall polysaccharide (CWPS; 5 μg/ml; 50 μl/well; catalog no. 3459, SSI Diagnostica). The dilutions were incubated at 37°C for 1 h to neutralize potential anti-CWPS Abs. Briefly, a 96-well ELISA plate was coated with PCP14 (5 μg/ml) in PBS (50 μl/well) overnight at 4°C. The plate was washed with PBST (PBS containing 0.05% Tween 20) and blocked with 5% skim milk (in PBST) for 1 h at 37°C. After washing, the CWPS preabsorbed sera dilutions were added to the plate and incubated for 2 h at 37°C in a humidified chamber. The plate was washed thoroughly and incubated for 2 h at room temperature with alkaline phosphatase (AP)–conjugated goat anti-mouse secondary Ab (diluted 1 in 3,000 in 2.5% PBST; 50 μl/well; catalog no. 1010-04, Southern Biotech, USA). The plate was washed, substrate p-nitrophenyl phosphate (1 mg/ml) was added (50 μl/well), and the plate was incubated in the dark at room temperature for 1 h. The reaction was stopped with 3 N NaOH (50 μl/well), and absorbance was recorded at 405 nm. Absorbance of the serially diluted pooled sera samples was plotted along with the control group of the corresponding time point, and their slopes were calculated using linear regression (GraphPad Prism version 8, GraphStats, USA). Intersection point of the control and test group slopes was identified. An OD value twice that of the intersection value was used as the cutoff for determining endpoint titers.

The same procedure was followed for determining anti-CRM197 titers with minor changes. The plate was coated with CRM197 in carbonate buffer pH 9.6 (5 μg/ml; 50 μl/well), overnight at 4°C. Serum samples were diluted in PBS. The plate was blocked with 2% BSA in PBST for 1 h at 37°C. AP-conjugated goat anti-mouse Ab (diluted 1 in 3,000 in 1% BSA in PBST; 50 μl/well) was added. For IgM- and IgG-specific Ab titer determination, the same procedure was followed except that the isotype-specific AP-conjugated goat anti-mouse Ab was used (catalog nos. 1020-04 and 1030-04, Southern Biotech).

Generation of PCP14- and CRM197-specific hybridomas

Monoclonal Abs were generated using hybridoma technology (13, 14). Briefly, splenocytes obtained from mice immunized with either one or three doses of glycoconjugate were fused with SP2/0-Ag14 myeloma cells (catalog no. CRL-1581, ATCC, USA) in a 4:1 ratio using polyethylene glycol (molecular weight 1,450, 50% [w/v]; catalog no. P-7181, Sigma, USA). Hybridomas were selected using hypoxanthine–aminopterin–thymidine medium (catalog no. H0262, Sigma) containing 20% FBS (catalog no. 04-127-1A, Biological Industries, Israel). In all, five fusions were performed, each from a different mouse. The splenocytes from a given mouse were used for performing only one fusion. For the one-dose regimen, three fusions were performed: on day 14 (mouse ID 1, screened for anti-PCP14 mAbs only), day 21 (mouse ID 2, screened for anti-PCP14 and anti-CRM197 mAbs), and day 28 (mouse ID 3, screened for anti-PCP14 mAbs only). Two fusions were done for the three-dose regimen at day 35 from two separate mice (mouse IDs 4 and 5). For both of these fusions, the hybridomas were screened for reactivity against PCP14 and CRM197.

PCP14- and CRM197-reactive mAb-secreting hybridomas were screened using an indirect ELISA following the same procedure described above for Ab titer determination. In the latter case, culture supernatant was used in place of diluted sera samples. Hybridoma cells from the well showing reactivity with the PCP14 or CRM197 Ag were subjected to two further rounds of limiting dilution cloning in the presence of irradiated EL-4-B5 cells (catalog no. EVU301, Kerafast Inc., USA) as a feeder layer to achieve monoclonality.

Isotyping of the anti-PCP14 and anti-CRM197 mAbs was performed using hybridoma culture supernatant by ELISA as described above. Eleven goat anti-mouse secondary Abs specific to different isotypes of the H (IgM, IgG, IgA [catalog no. 1040-04], IgG1 [catalog no. 1070-04], IgG2a [catalog no. 1080-04], IgG2b [catalog no. 1090-04], IgG2c [catalog no. 1079-04], and IgG3 [catalog no. 1100-04], Southern Biotech) and two L chains (Igκ [catalog no. 1050-04] and Igλ [catalog no. 1060-04]) were used including one as control (total Ig; Southern Biotech). The 96-well plate was coated with hybridoma supernatant followed by incubation with isotype-specific secondary Abs conjugated with AP. The color development on addition of the substrate (p-nitrophenyl phosphate) was recorded at 405 nm and used to assign isotype.

Amplification of immunoglobulin H and L chain transcripts from hybridomas by RT-PCR

Ig H and L chain transcripts expressed in PCP14- and CRM197-specific hybridomas were amplified using hot-start touch down nested RT-PCR as described previously (14, 15). Briefly, 10 to 50 hybridoma cells suspended in 20 μl of nuclease-free water were lysed by snap freezing in liquid nitrogen. The cellular content released upon thawing the lysed cells was used as the source of RNA. The cDNA was synthesized using a commercially available cDNA synthesis kit (catalog no. 205313, Qiagen, Germany) following the manufacturer’s instructions. The 20 μl reaction mixture with isotype-specific primers for the Ig H and L chain (0.75 μM each) was incubated at 42°C for 30 min followed by incubation at 95°C for 3 min to inactivate the enzyme. The nested PCR amplification for Ig H and L chains were performed separately. Four microliters of cDNA was used as the template in a 50 μl first-round PCR containing external antisense primer (0.25 μM) and a mixture of VH (or VL as the case may be) family-specific external sense primers, each at a final concentration of 0.1 μM, 1× Q5 DNA polymerase buffer, dNTPs (200 μM), and Q5 DNA polymerase (0.5 U) as recommended by the manufacturer (catalog no. M04901, New England Biolabs, USA). The second round PCR was performed using 2 μl of first-round PCR product as the template in a 50 μl reaction following the protocol described for the first-round PCR. The second round PCR product was column-purified according to the protocol provided by the manufacturer (catalog no. K310002, Invitrogen, USA) and sequenced.

Bioinformatic analysis of immunoglobulin H and L chain sequences

The Ig nucleotide sequences were analyzed using MacVector (with Assembler) software (version 17.0.5; MacVector Inc., USA). Briefly, the sense and antisense sequences were assembled to generate a contig. The primer binding site was edited out from the sequence. The final sequence was analyzed using IMGT/V-QUEST (https://www.imgt.org/IMGT_vquest/input) (16, 17) and IgBlast (https://www.ncbi.nlm.nih.gov/igblast/) (18) using the default parameters. Output from the IMGT/V-QUEST was used to assign germline VH, DH, JH, VL, and JL gene segments and for calculating the mutational load. The 3′ boundary of the VH gene segment and 5′ boundary of the JH gene segment were identified (with a maximum of two mutations allowed, each mutation being at least two nucleotides away from the junction ends). The remaining unassigned nucleotides were matched with the DH gene segments in the database (minimum of four nucleotides should match with the DH gene segment with a maximum of four mutations permitted with no change in two junctional nucleotides at both ends). P- and N-nucleotide assignment was done manually by analyzing the CDR3 sequence. The nucleotide(s) in the CDR3 that were not assigned to VH, DH, JH, VL, or JL gene segments were considered for assigning P- and N-nucleotides. The nucleotides derived from the asymmetric cleavage of the hairpin turns generated during somatic V(D)J recombination that resulted in palindromic sequences were considered as P nucleotides. Nucleotides with nontemplate origin were classified as N nucleotides.

Hybridomas that shared the same VH, DH, JH, VL, JL, HCDR3, and LCDR3 along with N- and P-nucleotide additions and that had undergone identical exonuclease trimming in their expressed Ig transcripts were considered to have been derived from clonally related B cells. The genealogical tree was constructed starting with a hypothetical germline (unmutated) ancestral precursor B cell following the principle of shared mutations and stepwise accumulation of mutations during clonal expansion. The framework region (FR), CDR, and codon numbering was according to the IMGT system (19). The IMGT output provided the number of mutations and was used for comparison of mutational load across groups. To avoid the issue of over-representation, we considered only one member (with the highest mutational load [Ig H and L chains combined]) of the clone for the analysis.

Statistical analysis

Statistical analyses were performed using the GraphPad Prism, and the data are presented as means ± SEM. The mutational load data were analyzed using two-tailed Mann–Whitney U test. A p value of < 0.05 was considered statistically significant.

Results

Booster shots of glycoconjugate enhance serum anti-polysaccharide and anti–carrier protein antibody response

To determine the kinetics of Ab response, we followed anti-PCP14 and anti-CRM197 serum Ab titers in mice who received either one or three doses of PCP14-CRM197 glycoconjugate. As expected, we observed a characteristic heightened total Ab response with booster immunization (Fig. 1A, 1B). We did not observe a boosting effect in the anti-PCP14 IgM Ab response apart from the peak following third shot, after which the response waned to baseline level (Fig. 1C). Surprisingly, we did not observe any detectable IgM Ab response against CRM197 (Fig. 1D). IgG titers against PCP14 (Fig. 1E) and CRM197 (Fig. 1F) showed heightened and sustained titers upon receiving booster dose.

FIGURE 1. Serum anti-PCP14 and anti-CRM197 Ab response kinetics following immunization with PCP14-CRM197.

Groups of female BALB/c mice (n = 10) were immunized with PCP14-CRM197 with alum, i.p. In the case of the single-dose regimen, the mice received a single dose at day 0 (shown as filled triangle below the y-axis), whereas in the case of the three-dose regimen, the mice were immunized on days 0, 14, and 28 (shown as open triangles). Control groups received only alum. The mice were bled at weekly intervals until week 14. The sera samples were pooled, and anti-PCP14 (A, C, and E) and anti-CRM197 (B, D, and F) endpoint Ab titers were determined by ELISA. The one- and three-dose regimens are indicated by filled and open circles, respectively. The curves were nudged in (D) for the ease of visualization.

Unlike for CRM197, anti-PCP14 antibody response is very restricted

We obtained 17 anti-PCP14 hybridomas from 3 fusions (at days 14, 21, and 28; Mouse IDs 1, 2, and 3) for which the mice received a single dose of PCP14-CRM197 (Table I). The majority (13 of 17; 76.47%) of these anti-PCP14 mAbs were of the IgM isotype. The remaining 4 were of IgG isotype. In case of CRM197-specific hybridomas, all 7 mAbs obtained in a single fusion (at day 21; Mouse ID 2) from mouse that received only a single dose of PCP14-CRM197 were of IgM isotype (Table II). Two fusions were performed from mice (Mouse IDs 4 and 5) that were administered three doses of PCP14-CRM197, at day 35. Anti-PCP14 hybridomas had 6 mAbs of IgG isotype and only 3 of IgM isotype. In contrast, 11 of 15 (73.33%) CRM197-specific mAbs obtained from these fusions were of IgM isotype, and the remaining 4 were of IgG isotype.

Table I. V, D, and J gene usage in PCP14-specific hybridomas

Mice ID.	Pri/Tera	Dayb	Name	Isotype	VH	DH	JH	VL	JL	
1	Pri	14	B07B02	IgM/κ	VH1-26	Absente	JH2	Vκ1-117	Jκ2	
			B12G10	IgM/κ	VH1-4 c	Absent e	JH2	Vκ1-117	Jκ1	
			C01G01	IgM/κ	VH1-4c	DH4-1	JH3	Vκ1-117	Jκ1	
			D03E02	IgM/κ	VH1-4 c	Absent e	JH2	Vκ1-117	Jκ1	
			D10H02	IgM/κ	VH5-4	DH4-1	JH3	Vκ1-117	Jκ2	
			E07A04	IgM/λ	VH5-17	DH1-1	JH2	Vλ1	Jλ1	
			F03A11	IgM/κ	VH1-26	Absente	JH2	Vκ1-117	Jκ2	
			F03E04	IgG1/κ	VH1-26	Absente	JH2	Vκ1-117	Jκ1	
			F04F09	IgM/κ	VH1-26	Absente	JH2	Vκ1-117	Jκ2	
			G06C07	IgM/κ	VH1-S29	Absente	JH2	Vκ1-117	Jκ2	
2	Pri	21	B08F09	IgG1/κ	VH1-4c	DH4-1	JH3	Vκ1-117	Jκ1	
			E03A02	IgG3/κ	VH1-4d	Absente	JH4	Vκ1-117	Jκ1	
			F08B01	IgG1/κ	VH1-26	Absente	JH2	Vκ1-117	Jκ1	
3	Pri	28	A12A03	IgM/κ	VH1-S29	Absente	JH2	Vκ1-117	Jκ1	
			D08H04	IgM/κ	VH1-S29	Absente	JH2	Vκ1-117	Jκ1	
			G11A06	IgM/κ	VH1-S29	Absente	JH2	Vκ1-117	Jκ1	
			H05C03	IgM/κ	VH1-26	DH4-1	JH3	Vκ1-117	Jκ1	
4	Ter	35	A04A09	IgM/κ	VH2-9	DH2-10	JH4	Vκ8-24	Jκ2	
			C02H08	IgM/κ	VH1-26	Absente	JH2	Vκ1-117	Jκ2	
			D05B02	IgG1/κ	VH1-26	Absente	JH2	Vκ1-117	Jκ1	
			D05C03	IgG1/κ	VH2-6	Absente	JH4	Vκ1-117	Jκ1	
			E08B03	IgG3/κ	VH3-1	DH6-1	JH3	Vκ1-117	Jκ1	
5	Ter	35	B05E02	IgM/κ	VH1-26	DH4-1	JH3	Vκ1-117	Jκ1	
			B07D06	IgG3/κ	VH1-26	Absente	JH2	Vκ1-117	Jκ1	
			C06D11	IgG1/κ	VH1-26	Absente	JH4	Vκ1-117	Jκ1	
			D10C04	IgG3/κ	VH1-26	Absente	JH2	Vκ1-117	Jκ1	
Clonally related hybridomas B07B02, F03A11, and F04F09 are highlighted with bold type. The second clone comprising of hybridomas B12G10 and D03E02 are highlighted with underlining. For the subsequent comparative analyses, we have considered only one representative member of each clone.

Assignment for the VH, DH, JH, VL, and JL gene segments was done using IMGT/V-QUEST. Alleles assigned by IMGT/V-QUEST for the VH, DH, JH, VL, and JL gene segments are not shown here.

a Monoclonal Abs generated under the one- and three-dose regimens.

b The day on which the splenocytes were harvested from mice immunized with PCP14-CRM197 (one or three doses) for generating the hybridomas.

c The germline VH gene segment used in the expressed H chain could be either VH1-4 or VH1-7.

d The germline VH gene segment used in the expressed H chain could be VH1-4, VH1-7 or VH1-26.

e The expressed Ig H chain lacks DH gene–encoded amino acids.

Pri, one-dose regimen; Ter, three-dose regimen.

Table II. V, D, and J gene usage in CRM197-specific hybridomas

Mice ID.	Pri/Tera	Dayb	Name	Isotype	VH	DH	JH	VL	JL	
2	Pri	21	B02A07	IgM/κ	VH1-9	DH1-2	JH2	Vκ1-135	Jκ1	
			C07C02	IgM/κ	VH1-S81	DH2-1d	JH4	Vκ12-44	Jκ5	
			C07H04	IgM/κ	VH3-2	DH1-1	JH2	Vκ8-24	Jκ5	
			E07D09	IgM/κ	VH14-3	DH1-1	JH3	Vκ8-21	Jκ1	
			F08D01	IgM/κ	VH2-6-7	DH6-2	JH4	Vκ10-94	Jκ1	
			F10B03	IgM/κ	VH1-S135	DH1-2	JH2	Vκ16-104	Jκ2	
			G09H08	IgM/κ	VH5-6-2	DH2-1	JH4	Vκ10-96	Jκ1	
4	Ter	35	A06C04	IgM/κ	VH2-6-7	DH1-1	JH3	Vκ8-21	Jκ1	
			A07B05	IgM/κ	VH5-9	DH1-1	JH2	Vκ6-32	Jκ1	
			A07D08	IgM/κ	VH1-S81	DH2-3	JH4	Vκ6-15	Jκ2	
			C09G10	IgM/κ	VH1-S53	Absente	JH3	Vκ8-30	Jκ4	
			D09C04	IgM/κ	VH2-6	DH3-1	JH4	Vκ8-21	Jκ4	
			D10G02	IgG1/κ	VH1-55	Absente	JH2	Vκ3-12	Jκ5	
			G05F03	IgM/κ	VH1-S136	DH2-1f	JH1	Vκ10-96	Jκ1	
			G05G05	IgG2b/κ	VH1-64c	DH1-1	JH3	Vκ6-32	Jκ4	
5	Ter	35	B01E02	IgM/κ	VH1-5	DH1-1	JH4	Vκ4-57	Jκ1	
			B02F09	IgM/κ	VH1-S12	DH2-4	JH3	Vκ14-111	Jκ5	
			B04D10	IgM/κ	VH5-6	DH1-2	JH3	Vκ12-46	Jκ1	
			C04B04	IgG1/κ	VH1-18	DH2-4	JH4	Vκ12-41	Jκ5	
			E07F03	IgG1/κ	VH1-9	DH2-12	JH2	Vκ8-27	Jκ2	
			G06F07	IgM/κ	VH1-5	DH2-4	JH3	Vκ12-41	Jκ5	
			G08E08	IgM/κ	VH5-12-1	DH3-3	JH2	Vκ14-111	Jκ1	
Assignment for the VH, DH, JH, VL, and JL gene segments was done using IMGT/V-QUEST. Alleles assigned by IMGT/V-QUEST for the VH, DH, JH, VL, and JL gene segments are not shown.

a Monoclonal Abs generated under the one- and three-dose regimens.

b The day the splenocytes were harvested from mice immunized with PCP14-CRM197 (one or three doses) for generating the hybridomas.

c The germline VH gene segment used in the expressed H chain could be either VH1-64 or VH1-S81.

d The germline DH gene segment used in the H chain could be either DH2-1 or DH2-10.

e The expressed Ig H chain lacks DH gene–encoded amino acids.

f The germline DH gene segment used in the H chain could be DH2-1, DH2-10, or DH2-11.

Pri, one-dose regimen; Ter, three-dose regimen.

The V, D, and J gene segment usage was analyzed for the PCP14- and CRM197-specific hybridomas (Tables I and II). Our analysis indicated the presence of two sets of clonally related mAbs in the anti-PCP14 set. To avoid over-representation, we considered only a single representative member from each clone ([B07B02, F03A11, and F04F09], and [B12G10 and D03E02]; Table I) in our subsequent analysis. We observed a preferential use of variable (VH and VL) and joining (JH and JL) gene segments in the Ig H and L chain transcripts expressed in the PCP14-specific hybridomas. A majority (18 of 23; 78.26%) of anti-PCP14 mAbs used IGHV1 subgroup members (Fig. 2A, Table I). The remaining anti-PCP14 hybridomas used IGHV2 (n = 2), IGHV3 (n = 1), and IGHV5 (n = 2). The most frequently used VH gene segment was IGHV1-26 (10 of 23; 43.47%). One additional PCP14-specific hybridoma used a VH gene segment that could be IGHV1-4, IGHV1-7, or IGHV1-26. Of 22 anti-CRM197 mAbs, 13 (59.09%) used IGHV1 subgroup members (Table II). We did not observe preference for any VH gene segments in the anti-CRM197 hybridomas. For the JH gene usage (Fig. 2B), CRM197-specific hybridomas had representation from all four JH gene segments, whereas the PCP14-specific hybridomas used only JH2, JH3, and JH4 gene segments.

FIGURE 2. PCP14- and CRM197-specific mAbs exhibit distinct Ig gene segment usage.

(A–D) Variable (VH and VL; A and C, respectively) and joining (JH and JL; B and D, respectively) gene segments used in the Ig H and L chain transcripts expressed in the PCP14-specific (blue) and CRM197-specific (red) hybridomas are shown. Only the gene segments that were present in our hybridoma set are shown. $The VH gene segment used could be either VH1-4 or VH1-7. #The VH gene segment used could be either VH1-4, VH1-7, or VH1-26. &The VH gene segment used could be either VH1-64 or VH1-S81. Dark and light shades represent the single- and triple-dose regimens, respectively. (E) Ig H and L chain pairing for the PCP14- and CRM197-specific mAbs is shown. The size of the circle is proportionate to the relative number of times a given VH and VL pair was observed in our set of mAbs. The Vλ and Jλ gene segment usage and VH–VL pairing for Igλ chain is not shown.

In the case of L chain, 21 of 23 (91.30%) anti-PCP14 hybridomas (Fig. 2C) used the IGκV1-117 gene segment. By contrast, the VL gene segments used in CRM197-specific hybridomas were quite diverse. VL gene segments used in CRM197-specific hybridomas were different from the VL gene segments used in the PCP14-specific hybridomas with the exception of one hybridoma. We observed only one PCP14-specific hybridoma that used the λ L chain. In the case of JL gene, we did not find any hybridoma that used Jκ3 gene segment (Fig. 2D). The Jκ gene usage was restricted to Jκ1 and Jκ2 for PCP14-specific hybridomas (Table I). The lone λ L chain expressing anti-PCP14 hybridoma used Jλ1 gene segment. Anti-CRM197 mAbs used all four Jκ gene segments except Jκ3 (Table II). We observed a strong selection in favor of IGHV1-26 and IGκV1-117 gene segment pairing in the anti-PCP14 hybridomas (Fig. 2E). No such VH–VL pairing preference was seen in the case of anti-CRM197 hybridomas, and the VH–VL pairs obtained were unique. A bias in the V and J gene pairing was observed (Supplemental Fig. 1). IGκV1-117 preferentially paired with Jκ1 in 17 of 23 (73.91%) anti-PCP14 mAbs. This analysis suggests that the anti-PCP14 B cell repertoire is highly restricted, whereas it is diverse in the case of CRM197.

CDR3 shows extreme restriction in anti-PCP14 response

The HCDRs and LCDRs together make up the paratope that binds the epitope on the Ag. Among the six CDRs, HCDR3 is the most diverse and is considered to be the major contributor toward Ag recognition. We analyzed the CDR3 length distribution in our mAb set. The HCDR3 length in anti-PCP14 mAbs ranged from 5 to 10 codons (mean ± SEM = 5.82 ± 0.35) with the HCDR3 length of 5 amino acids being observed in 18 of 23 (78.26%) mAbs (Fig. 3A, Table III). HCDR3 length in anti-CRM197 mAbs ranged from 5 to 16 codons (mean ± SEM = 11.63 ± 0.59) with 19 of 22 (86.36%) mAbs preferring codon lengths equal to or greater than 10. Only a single CRM197-specific mAb had a HCDR3 length of five codons (Fig. 3A, Table IV). None of the anti-PCP14 and anti-CRM197 mAbs had a HCDR3 length of six amino acids. In addition to short HCDR3, anti-PCP14 mAbs preferentially used the conserved motif ARWDY (Table III). In contrast, the HCDR3 sequences of anti-CRM197 mAbs were very diverse, in both length and amino acid sequence (Table IV). The LCDR3 length in the case of anti-PCP14 mAbs was nine codons (Fig. 3B). Of 23 anti-PCP14 mAbs, 21 (91.30%) used the motif FQGSHVP(R/W/Y)T (Table III), whereas the LCDR3 length in anti-CRM197 mAbs was limited to 8 or 9 codons (mean ± SEM = 8.81 ± 0.08), with 4 mAbs having a codon length of 8, and 18 mAbs having a length of 9 codons.

FIGURE 3. CDR3 length distribution in anti-PCP14 and anti-CRM197 mAbs.

HCDR3 (A) and LCDR3 (B) lengths (in codons) for the PCP14- and CRM197-specific mAbs are plotted. CDR3 lengths for the anti-PCP14 and anti-CRM197 mAbs are represented as black and gray bars, respectively.

Table III. HCDR3 and LCDR3 sequences in anti-PCP14 mAbs

Mice ID.	Pri/Tera	Dayb	Name	HCDR3	LCDR3	
1	Pri	14	B07B02	ARWDY	FQGSHVPYT	
			B12G10	ARWDY	FQGSHVPRT	
			C01G01	AΚWDY	FQGSHVPRT	
			D03E02	ARWDC	FQGSHVPRT	
			D10H02	ARWDEWFAY	FQGSHVPYT	
			E07A04	AREWTTVGDY	ALWYSNHWV	
			F03A11	ARWDY	FQGSHVPYT	
			F03E04	TRWDY	FQGSHVPWT	
			F04F09	ARWDY	FQGSHVPYT	
			G06C07	ARWDY	FQGSHVPYT	
2	Pri	21	B08F09	ARWDY	FQGSHVPRT	
			E03A02	ARWDY	FQGSHVPRT	
			F08B01	ARWDY	FQGSHVPWT	
3	Pri	28	A12A03	ARWDY	FQGSHVPRT	
			D08H04	ARWDY	FQGSHVPWT	
			G11A06	ARWDY	FQGSHVPWT	
			H05C03	ATWAY	FQGSHVPWT	
4	Ter	35	A04A09	ARAYYDAMDY	QQHYSTPYT	
			C02H08	LRWDY	FQGSHVPYT	
			D05B02	ARWDY	FQGSHVPWT	
			D05C03	ARAPMDY	FQGSHVPWT	
			E08B03	ASLWPFAY	FQGSHVPWT	
5	Ter	35	B05E02	ARWDR	FQGSHVPWT	
			B07D06	ARWDC	FQGSHVPWT	
			C06D11	ARWDY	FQGSHVPWT	
			D10C04	ARWDC	FQGSHVPWT	
Clonally related hybridomas B12G10 and D03E02 are highlighted with underlining, and clonally related hybridomas B07B02, F03A11, and F04F09 are highlighted in bold type.

a Monoclonal Abs generated under the one- and three-dose regimens.

b The day on which the splenocytes were harvested from mice immunized with PCP14-CRM197 for generating the hybridomas.

Pri, one-dose regimen; Ter, three-dose regimen.

Table IV. HCDR3 and LCDR3 sequences in anti-CRM197 mAbs

Mice ID.	Pri/Tera	Dayb	Name	HCDR3	LCDR3	
2	Pri	21	B02A07	AIHYYFDY	WQGTHFPQT	
			C07C02	ARΚGGNYFYAMDY	QHHYGTPLT	
			C07H04	ARERPYYGSSFLFDY	QQHYSTPLT	
			E07D09	ARGGVYYYGSSGAY	ΚQSYNLWT	
			F08D01	ARAYSLYAMDY	QQYSΚLPWT	
			F10B03	ARSPRALRLYYFDY	QQHNEYPYT	
			G09H08	ARQGDNYGDYVMDY	QQGNTLPWT	
4	Ter	35	A06C04	AREGGPWFAY	ΚQSYNLRT	
			A07B05	ARRDYYYYGSSYFDY	QQDYSSPWT	
			A07D08	TIGWLLRVRHYYAMDY	QQYNSYPYT	
			C09G10	KSPRY	QQYYSYPFT	
			D09C04	ARGLPYAMDY	KQSYNLFT	
			D10G02	ARGDFDY	QHSRELLTF	
			G05F03	ARRGYYGNYGGYFDV	QQGNTLPPT	
			G05G05	TRRGSSAFAY	HQDYSSPPT	
5	Ter	35	B01E02	TLYYGYAMDY	QQYSGYPWT	
			B02F09	ARSNYDYDWFAY	LQYDEFPLT	
			B04D10	ARHGVHYFAWFAY	QHFWGTPWT	
			C04B04	ALHDFDYAMDY	QHFWSTPLT	
			E07F03	TRCNYDAASDY	HQYLSSYT	
			G06F07	TRYDYVAWFAY	QHFWSTPLT	
			G08E08	ARHGGΚGWPDY	LQYDEFPWT	
a Monoclonal Abs generated under the one- and three-dose regimens.

b The day on which the splenocytes were harvested from mice immunized with PCP14-CRM197 for generating the hybridomas.

Pri, one-dose regimen; Ter, three-dose regimen.

Clonal expansion of PCP14-specific B cells in response to PCP14-CRM197

In our analysis, we found two sets of anti-PCP14 hybridomas that were derived from clonally related B cells. The first clone had three members, namely B07B02, F03A11, and F04F09. The hypothetical ancestral B cell of this clone rearranged IGHV1-26 and IGHJ2 in the H chain and IGκV1-117 and IGκJ2 in the L chain (Fig. 4A, 4C). It may be noted that this H chain lacks DH gene–encoded amino acids. The ancestral B cell acquired eight shared mutations in the H chain to give rise to a hypothetical intermediate precursor 1 (HP1). This precursor accumulated one and two additional mutations in the H and L chains, respectively, to form B cell B07B02, whereas an additional mutation at a distinct position in the HP1 gave rise to a hypothetical intermediate precursor 2 (HP2), which further gained two mutations in the H chain and one mutation in the L chain to form B cell F04F09. The other daughter B cell acquired two further mutations in the H chain, and three mutations in the L chain to give rise to B cell F03A11. These clonally related Abs were of the IgM/κ isotype.

FIGURE 4. Clonal expansion of PCP14-specific B cells in mice immunized with PCP14-CRM197.

Sequence alignment of the Ig H and L chain genes expressed in hybridomas derived from clonally related B cells with their germline counterpart are shown. Clone 1 comprising B07B02, F03A11, and F04F09 and clone 2 with members D03E02 and B12G10 are presented in (A) and (B), respectively. The nucleotide sequence of germline VH, DH, JH, VL, and JL gene segments used are shown on top. The suffixes “-H” and “-L” denote H and L chain genes expressed in the anti-PCP14 hybridomas, respectively. The FR, CDR, and codon positions are according to the IMGT numbering system. The amino acid translation (in single-letter code; in italics) is given below the codons. Nucleotide identity with the corresponding nucleotide in the germline gene is indicated by a dash (-). The figure shows only those codon positions where mutation(s) have occurred. Replacement and silent mutations are shown in upper and lowercase, respectively. CDR3 sequence is, however, presented in full. P- and N-nucleotide addition(s) are indicated, and the nucleotides lost due to exonuclease activity in the CDR3s during the V(D)J recombination are shown as missing. Deduced genealogical trees for clones 1 (C) and 2 (D) are shown as schematic diagrams. The number of nucleotide changes observed in the expressed H chain gene is written first, followed by the number of changes in the expressed L chain gene (italicized). HP denotes the hypothetical intermediate precursor. V(D)J rearrangements used in the hypothetical germline precursor and the isotypes of the expressed H and L chains are also indicated.

The second clone was initiated by a B cell that rearranged IGHV1-4 to IGHJ2 and IGκV1-117 to IGκJ1 for H and L chains, respectively (Fig. 4B, 4D). Like the clone described above, this clone also shared a H chain that was devoid of DH gene–encoded amino acids. The hypothetical ancestral B cell acquired four mutations in the H chain and one mutation in the L chain to generate a hypothetical intermediate precursor (HP). The three distinct additional mutations in the H chain resulted in the formation of B cell B12G10. HP accumulated two mutations each in the H and L chains to form B cell D03E02. This clonal pair was of IgM/κ isotype.

Booster doses drive mutational load in PCP14-specific B cells

We observed a significant increase in the mutational load in the H chains of anti-PCP14 hybridomas that were generated under the three-dose regimen (mean ± SEM = 14.33 ± 1.52; range, 5 to 20) in comparison with anti-PCP14 hybridomas generated after single dose of PCP14-CRM197 (mean ± SEM = 10.14 ± 1.23; range, 2 to 18; Fig. 5A). Similar increase in the number of mutations were found in the L chains of anti-PCP14 hybridomas obtained from mice immunized with three doses (mean ± SEM = 5.33 ± 0.83; range, 0 to 8) as compared with single-dose immunization (mean ± SEM = 2.64 ± 0.40; range, 0 to 5) of PCP14-CRM197. Contrary to expectations, we did not observe a significant increase in the mutational load in the H (mean ± SEM = 5.66 ± 1.82 versus 2.85 ± 1.42; range, 0 to 26 versus 1 to 10) and L (mean ± SEM = 1.13 ± 0.58 versus 1.00 ± 0.84; range, 0 to 7 versus 1 to 6) chains in the CRM197-specific hybridomas after booster shots (Fig. 5B). Although the number of IgG CRM197-specific hybridomas was low, the trend indicated that carrier-specific IgG mAbs (H chain, 14.00 ± 4.26; L chain, 3.50 ± 1.75) have higher numbers of mutations as compared with the number of mutations in anti-CRM197 IgM hybridomas (H chain, 2.72 ± 0.77; L chain, 0.55 ± 0.34). The magnitude of the difference was much lower in the case of PCP14-specific IgM (H chain, 10.46 ± 1.40; L chain, 2.84 ± 0.54) and IgG (H chain, 12.70 ± 1.75; L chain, 4.70 ± 0.79) mAbs (Fig. 5A).

FIGURE 5. Booster shots of PCP14-CRM197 drives mutational load in PCP14 specific B cells.

Mutational load in the rearranged Ig H and L chain genes in the PCP14- and CRM197-specific hybridomas. Representative mAb from each clone was considered for the analysis. The number of mutations in PCP14-specific (A) and CRM197-specific (B) hybridomas obtained from the one-dose (filled circle) and three-dose (open circle) regimens are shown. The mutational loads in mAbs of IgM and IgG isotype are represented in red and blue, respectively. Error bars represent means ± SEM. The p values were calculated using two-tailed Mann–Whitney U test.

Specific amino acids at specific codon positions in the mAb are required for binding PCP14

We analyzed the anti-PCP14 mAbs that used IGHV1-26 gene in detail. Remarkably, we found a set of five replacement mutations (40:N > K; 55:G > D; 66:S > F; 90:E > Q; 92:R > N) that co-occurred at the specific codon positions in five anti-PCP14 hybridomas generated from different animals at different time points. In a sixth anti-PCP14 hybridoma, four of these five amino acid replacement mutations co-occurred. Unexpectedly, of the five replacement mutations, only one occurred in the CDR of the H chain, whereas four occurred in the framework regions (Fig. 6). Analysis of PCP14-specific hybridomas that used the IGκV1-117 gene segment indicated that the majority (71.42%; 15 of 21) of the L chains had an N to D replacement mutation at codon position 34 in the CDR1 (Supplemental Fig. 2). Additionally, we observed that hybridomas raised following booster shots accumulated higher numbers of mutations (Fig. 5A). These data suggest that a specific set of amino acids (i.e., K40, D55, F66, Q90, and N92) are necessary at the indicated positions for the recognition of the epitope on PCP14. These data provide evidence for Ag-driven selection.

FIGURE 6. Recurrent linked replacement mutations observed in PCP14-specific mAbs.

The amino acid sequence of the Ig H chain expressed in PCP14-specific mAbs was aligned with the corresponding germline VH gene segment VH1-26. The codon position of recurrent amino acid replacement mutation is shown at the bottom of the multiple sequence alignment. The FRs and CDRs were assigned according to the IMGT numbering system. The time point following immunization at which the hybridomas were generated is indicated on the right. Amino acid identity with the corresponding residue in the germline gene is indicated by a dash (-).

DH-less PCP14 reactive B cells exhibit convergent evolution

Upon closer examination, we found that the HCDR3 sequence (ARWDY) was identical in 11 of 23 (47.82%) anti-PCP14 mAbs. Based on the differences in the VH, DH, and JH gene segments used, these 11 mAbs fell into 6 groups (Fig. 7A). Surprisingly, despite differences in the VH, DH, and JH genes used, experiencing different extent of exonuclease activity, N- and P-nucleotide addition(s) at the VH–DH and DH–JH junctions, the sequence of the HCDR3 was identical in this set of anti-PCP14 mAbs. Of these 11 anti-PCP14 mAbs, 10 lacked DH-encoded amino acids. Some of them were obtained from more than one mouse and in both the single- and three-dose immunization regimens. This convergence of HCDR3 amino acid sequences in these Abs was seen in all five mice used to generate these mAbs. The observed convergent evolution of Ab response is indicative of the strong pressure exerted by the Ag (i.e., PCP14) on selection of the Ag-specific B cells.

FIGURE 7. Convergent evolution observed in DH-less PCP14 reactive B cells.

(A) The HCDR3 sequence ARWDY was identical in a subset of PCP14-specific hybridomas despite (i) using different VH, DH, and JH gene segments, (ii) difference in the N- and P-nucleotides, (iii) extent of exonuclease chewing at the V–D and V–J junctions in PCP14-specific hybridomas, and (iv) the fact they were generated from different mice. *The VH gene segment used could be either VH1-4 or VH1-7. #The VH gene segment used could be either VH1-4, VH1-7, or VH1-26. P- and N-nucleotides are shown in red and green, respectively. Additional DH-less anti-PCP14 mAbs are shown in (B). Two anti-CRM197 mAbs that lacked DH derived amino acids are presented in (C). The codons between the conserved cysteine (codon position 104 in VH) and conserved tryptophan (codon position 118 in JH) are shown. The nucleotide sequence of the HCDR3 portion of the IgH transcript expressed in the hybridoma and the corresponding germline VH, DH, and JH gene segments is shown in bold type. P- and N-nucleotide addition(s) are indicated on top of the alignment. Identity with the corresponding nucleotide in the germline gene segment is indicated with a dash (-). Silent mutations are shown in lowercase letters. The amino acid translation is shown in single-letter code below the codons (in italics). It may be noted that some nucleotides are missing at the V–D and D–J junctions due to exonuclease activity. The frequency (Fq.) of occurrence, dose regimen (Rg.), and the mouse ID (Mo. ID.) from which the hybridoma was derived are shown on the right.

In addition to anti-PCP14 mAbs that shared the highly conserved HCDR3 ARWDY signature, five more anti-PCP14 mAbs lacked any codon that can be attributed to the DH gene segment (Fig. 7B). These mAbs had HCDR3 that differed from the ARWDY at any one of the five codon positions. In total, 15 of 23 (65.21%) anti-PCP14 mAbs lacked DH gene–derived amino acids. Incidentally, two CRM197-specific mAbs that used VH1-S53 and VH1-55 gene segments were devoid of DH gene–encoded amino acids (Fig. 7C). This unexpected finding suggests Abs that lack DH gene–derived codons are indeed present in the B-cell repertoire, albeit at a very low frequency, and are perfectly capable of binding the cognate Ag.

Discussion

In the current study, we performed a detailed analysis of the anti-polysaccharide and anti-carrier Ab response in mice immunized with a clinically relevant model glycoconjugate Ag PCP14-CRM197. The serum anti-PCP14 and anti-CRM197 titers were elevated following booster shots of the PCP14-CRM197 (Fig. 1). Similar observations were made with PCP4-TT (20) and PCP19F-CRM197 (5). Ab response to MCPS was boosted when mice primed with the MCPS-TT were administered a second dose of the glycoconjugate (3). In our study, we observed that the anti-CRM197 IgM response was low (Fig. 1D). Mawas et al. (5) found that the secondary anti-CRM197 was subdued, unlike the anti-PCP14 Ab response in mice that received two doses of PCP14-CRM197 (5).

The nature of the polysaccharide is an important determinant of how diverse or restricted the anti-polysaccharide Ab response will be. We observed that a limited set of VH and VL genes were used in the primary and tertiary anti-PCP14 Ab response in mice immunized with PCP14-CRM197. Vκ1-117 gene was used in 21 of 23 (91.30%) anti-PCP14 mAbs. By comparison, the anti-CRM197 mAbs used diverse set of V genes. A globular protein like CRM197 is highly likely to have several unique epitopes each with distinct topologies. As a consequence, the paratopes that would recognize these unique epitopes are expected to be equally diverse. This diversity is also reflected in the V, D, and J gene segment utilization in the anti-CRM197 mAbs. In contrast, PCP14 with its tetrasaccharide repeating unit is likely to have one or two epitopes. This can explain why the anti-PCP14 Ab is severely restricted (21, 22). The VH–VL pairs used in the anti-PCP14 mAbs were also very restricted with a VH1-26–Vκ1-117 pair comprising 43.47% (10 of 23). This is in contrast to anti-CRM197 mAbs in which each VH–VL pair was unique. Using a Qβ-PCP14 tetrasaccharide conjugate, Polonskaya et al. (23) observed that 90.00% (9 in 10) of anti-PCP14 mAbs used Vκ1-117 gene. Based on the crystal structure, these authors concluded that the Vκ1-117 L chain was dominant in recognition of PCP14 tetrasaccharide as 12 of the 16 contact residues were from the L chain. This is consistent with our observation. The anti-polysaccharide Ab repertoire elicited against PCP8-TT (8), PCP3-TT (9), and MCPS-TT (10) was found to be oligoclonal. In C. neoformans, in contrast, the anti-GXM Ab response to GXM-TT was very restricted, with one particular VH (i.e., VH7183) and VL (i.e., Vκ5.1) gene pair occurring in all the mAbs derived from three strains of mice (11, 12). In their study, Mukherjee et al. (11) suggested the L chain to be a major contributor to the paratope that binds GXM in GXM-TT immunized mice.

PCP14-CRM197 elicits convergent anti-PCP14 Ab response with a highly conserved CDR3 signature

The HCDRs and LCDRs of the Ab molecules are responsible for Ag recognition. The HCDR3s are the most diverse because they are generated at the junction of recombining V, D, and J gene segments. CDR3s are believed to be the major contributors toward Ag recognition, with some contribution from the other CDRs. Given the immense diversity observed in the HCDR3s, the probability of finding two individuals having identical HCDR3 sequences during Ab response to a given Ag is vanishingly small. In our study, we observed that the HCDR3 in 18 of 23 (78.26%) of the anti-PCP14 mAbs was 5 amino acids long. Of 23 anti-PCP14 mAbs, 11 (47.82%) had identical HCDR3 sequences; i.e., ARWDY (Table III). In the case of the L chain, 91.30% (21 of 23) of the PCP14 mAbs have FQGSHVP(W/Y/R)T as the LCDR3. Occurrence of such a high proportion of anti-PCP14 with identical HCDR3 and nearly identical LCDR3 is very striking. The recurrent occurrence of these sequence at HCDR3 and LCDR3 indicates that they are required for recognition of a particular epitope on PCP14. The HCDR3 sequence of ARWDY and LCDR3 of FQGSHVWT were observed at different immunization regimens, time points, and in five different animals. Of the 10 anti-PCP14 mAbs, 6 obtained from Qβ-PCP14 tetrasaccharide conjugate immunized mice had ARWD(S/Y) as the HCDR3, and 6 of 10 mAbs had FQGSHVP(Y/R)T as the LCDR3 (23). The HCDR3 sequence in anti-PCP8 mAbs generated from PCP8-TT immunized mice were quite diverse (8). The two anti-PCP3 mAbs reported by Tian et al. (9) showed very distinct HCDRs. A restricted HCDR3 was also observed during the Ab response to GXM in mice that received GXM-TT (11).

Examination of the sequence of the anti-PCP14 mAbs provided evidence for convergent Ab response (Fig. 7A). The HCDR3s of six representative anti-PCP14 mAbs share identical amino acid sequences (i.e., ARWDY). Each of these mAbs has a different configuration resulting from one or more of the following processes: 1) the V, D, and J gene segments; 2) N- and/or P-nucleotide addition; and 3) exonuclease dependent chewing of nucleotides at the V to D junction and the D to J junction. This is suggestive of the criticality of these HCDR3 residues in Ag binding. Convergent Ab response has been observed in the case of another bacterial pathogen (24) and several viruses (25–29).

The HCDR3 is generated as a consequence of somatic VDJ recombination. Multiple molecular processes are at play during the joining of the VH and DH gene segments and of the DH and JH gene segments. These processes include N- and/or P-nucleotide additions alongside exonuclease activity that leads to nibbling at the 3′ VH, 5′ DH, 3′ DH, and 5′JH ends. None, all, or a combination of these processes generate the enormous diversity of HCDR3. Analysis of the HCDR3 sequence from the mAbs revealed that 65.21% (15 of 23) of the anti-PCP14 and 9.09% (2 of 22) of the anti-CRM197 mAbs lacked amino acids attributable to any DH gene (Fig. 7). Curiously, the HCDR3 in these mAbs was five amino acids long, with the exception one mAb in which the HCDR3 length was seven amino acids. These “DH-less” mAbs used diverse combinations of VH and JH gene segments. The lack of DH-encoded amino acids did not prevent these mAbs from binding their cognate Ag; i.e., PCP14 or CRM197, as the case may be. The observation of DH-less mAbs is very unexpected and rare. A previous study reported that DH-less Abs can be generated, but the underlying mechanism is not known (14). Some plausible mechanisms include excessive exonuclease activity leading to chewing of complete DH gene element or direct VH to JH joining through cryptic RSS (30, 31). These results suggest that Abs can recognize Ag or pathogen in the absence of DH-encoded amino acids.

In contrast to the very restricted HCDR3 length range observed for the anti-PCP14 mAbs, the range was broader in the case of anti-CRM197 Abs (Fig. 3A, Tables III and IV). This finding can be interpreted to mean that the anti-PCP14 mAbs recognize only one or two distinct epitopes present on PCP14, whereas anti-CRM197 mAbs recognize multiple epitopes on CRM197. This is also reflected in the conservation in the CDR3 sequences in anti-PCP14 mAbs and diversity in the CDR3 sequences in the mAbs directed against CRM197. These data offer insights regarding the molecular features shared by Abs directed against capsular polysaccharide from S. pneumoniae type 14.

Mutational load in anti-PCP14 mAbs is higher than in anti-CRM197 mAbs

We observed high numbers of mutations in the H chains of the anti-PCP14 mAbs generated from mice administered a single dose of PCP14-CRM197 (Fig. 5A). This is consistent with the notion that covalent conjugation of polysaccharide to a carrier protein converts the anti-polysaccharide response from a T-independent to a T-dependent Ab response, a transition from germline or near germline to more heavily mutated and class-switched sequences. Consistent with this, we observed an increase in the mean mutational load in the H and L chain sequences upon booster immunization in the case of anti-PCP14 mAbs. To best of our knowledge, this is the first experimental demonstration of an increase in the mutational load in the anti-polysaccharide Ab sequences upon booster shots of a clinically relevant glycoconjugate. The higher mean mutational load in the anti-PCP14 mAbs can be due to the engagement of an unique highly repetitive immunodominant epitope on PCP14 by B cells, with the “second signal” coming in the form of T-cell help originating from the covalently conjugated protein carrier CRM197. We found that the mean mutational load in the H chain was much higher than that observed in the case of L chains (Fig. 5A). This difference could be due to intrinsic difference in the H and L chain loci, which may include differences in the promoter, enhancer, or frequency of mutation hotspots among other cis-acting elements.

Booster doses of a T-dependent Ag is expected to result in an increase in the average number of mutations in the Ab genes in Ag-specific B cells. Contrary to our expectation, we found that the Ab sequences in the anti-CRM197 mAbs were germline or near germline, and there was no statistically significant increase in the mean mutational load with booster shots of PCP14-CRM197. To our surprise, we found that the mean mutational load in the Ig H and L chains of anti-CRM197 mAbs was far less than that observed for the anti-PCP14 mAbs (Fig. 5B, Supplemental Tables I and II). There could be several contributing factors and explanations for this observation. PCP14 by virtue of its larger size can mask and make CRM197 less accessible to the B cells. In the germinal centers (GCs) induced in response to PCP14-CRM197, the number of PCP14-specific B cells are likely to far outnumber and outcompete CRM197-specific B cells due to the highly repetitive nature of the epitope present in PCP14 vis-à-vis multiple unique epitopes present in CRM197. The possibility that epitope(s) present in CRM197 are lost as a result of chemical conjugation cannot be ruled out.

Concurrent replacement mutations during anti-PCP14 antibody response

Sequence analysis of the VH1-26 gene using anti-PCP14 mAbs showed the presence of a five-amino acid motif: K40-D55-F66-Q90-N92. The VH1-26 gene is the most frequently used VH gene in the anti-PCP14 Ab response. The co-occurrence of the five amino acid residues is suggestive of convergent evolution and provides evidence of Ag-driven selection. Interestingly, four of these five residues (with the exception of D55) are present in the framework region. Traditionally, it is believed that the Ag-binding residues in the paratope are contributed by the CDRs. Our data suggest that amino acid residues in framework regions also contribute to Ag recognition indirectly or directly. In the case of the L chain, an N to D mutation at codon position 34 was found in CDR1 of 71.42% (15 of 21) of the anti-PCP14 mAbs that used Vκ1-117 gene segment (Supplemental Fig. 2).

Proposed model for the anti-PCP14 and anti-CRM197 Ab response generated in PCP14-CRM197 glycoconjugate immunized mice

Our model is built around our observation that anti-PCP14 Ab response is very restricted and involved recurrent use of identical V, D, and J gene segments (Table I). Upon immunization with PCP14-CRM197 glycoconjugate, PCP14-specific B cells from the preimmune B-cell repertoire participate in the formation of GCs in the spleen. There, PCP14-specific B cells undergo reiterative cycles of divisions and selection by the Ag–Ab complexes decorated on the surface of follicular dendritic cells (FDCs). This is consistent with the mutated and isotype-switched PCP14-specific B cells observed in our study (Fig. 5, Table I, Supplemental Table I). Booster doses of PCP14-CRM197 promote entry of Ag experienced anti-PCP14 B cells into GCs induced upon booster immunization. This results in further increases in their mutational load and the proportion of class-switched B cells during cycles of proliferation and selection by the immune complexes displayed by FDCs.

In the case of CRM197, adequate Ab response does not develop during primary immunization with PCP14-CRM197 due to insufficient density or inaccessibility of CRM197 epitopes present as Ab–glycoconjugate complexes displayed on the surface of FDCs. As a result, the anti-CRM197 hybridomas are unmutated or near germline (Fig. 5, Table II, Supplemental Table II). Upon booster immunization with glycoconjugate, fresh B cells from the naive repertoire participate in the anti-CRM197 Ab response. Due to booster dose, there is an increase in the epitope density of CRM197 on FDCs in the GCs that allows anti-CRM197 B cells to undergo somatic hypermutation and class switching. IgM anti-CRM197 B cells are unable to do so because of the absence of a sufficient number and/or density of CRM197 epitopes during the primary response.

In conclusion, our study showed that the vast majority of the polysaccharide-specific B cells were devoid of DH gene–encoded amino acids and had a highly conserved CDR3 signature. It is noteworthy that the absence of DH gene did not prevent the anti-polysaccharide mAbs from binding their cognate Ag. Anti-polysaccharide mAbs were highly mutated compared with anti–carrier protein mAbs, which were primarily germline or near germline. Our study offers insights into how covalent coupling a carrier protein to a polysaccharide enhanced the anti-polysaccharide Ab response and its impact on the carrier protein-specific Ab response. Interestingly, it suggests that in glycoconjugate immunized mice, the anti-polysaccharide Ab response is prioritized over the anti–carrier protein Ab response. Our findings have implications for designing and developing glycoconjugate vaccines. It would be interesting to 1) functionally characterize these mAbs and corelate their molecular features with the sequence information and 2) decipher the mechanism involved in the generation of DH-less Abs.

Supplementary Material

Supplemental Material (PDF)

Acknowledgments

We acknowledge the expert guidance provided by Dr. Ayub Qadri and Naresh Kumar for generating B cell hybridomas.

Footnotes

Disclosures

The authors have no financial conflicts of interest.

The online version of this article contains supplemental material.

AP alkaline phosphatase

CRM197 cross-reactive material 197

CWPS cell wall polysaccharide

FDC follicular dendritic cell

FR framework region

GC germinal center

GXM glucuronoxylomannan

MCPS meningococcal group C capsular polysaccharide

PCP pneumococcal capsular polysaccharide

TT tetanus toxoid
==== Refs
References

1 Rappuoli, R. 2018. Glycoconjugate vaccines: principles and mechanisms. Sci. Transl. Med. 10 : eaat4615.30158151
2 Kay, E., J. Cuccui, B. W. Wren. 2019. Recent advances in the production of recombinant glycoconjugate vaccines. NPJ Vaccines 4 : 16.31069118
3 Rubinstein, L. J., P. A. García-Ojeda, F. Michon, H. J. Jennings, K. E. Stein. 1998. Murine immune responses to Neisseria meningitidis group C capsular polysaccharide and a thymus-dependent toxoid conjugate vaccine. Infect. Immun. 66 : 5450–5456.9784556
4 Weintraub, A. 2003. Immunology of bacterial polysaccharide antigens. Carbohydr. Res. 338 : 2539–2547.14670715
5 Mawas, F., I. M. Feavers, M. J. Corbel. 2000. Serotype of Streptococcus pneumoniae capsular polysaccharide can modify the Th1/Th2 cytokine profile and IgG subclass response to pneumococal-CRM(197) conjugate vaccines in a murine model. Vaccine 19 : 1159–1166.11137252
6 Avci, F. Y., X. Li, M. Tsuji, D. L. Kasper. 2011. A mechanism for glycoconjugate vaccine activation of the adaptive immune system and its implications for vaccine design. Nat. Med. 17 : 1602–1609.22101769
7 Sun, L., D. R. Middleton, P. L. Wantuch, A. Ozdilek, F. Y. Avci. 2016. Carbohydrates as T-cell antigens with implications in health and disease. Glycobiology 26 : 1029–1040.27236197
8 Yano, M., L. A. Pirofski. 2011. Characterization of gene use and efficacy of mouse monoclonal antibodies to Streptococcus pneumoniae serotype 8. Clin. Vaccine Immunol. 18 : 59–66.21068211
9 Tian, H., S. Weber, P. Thorkildson, T. R. Kozel, L. A. Pirofski. 2009. Efficacy of opsonic and nonopsonic serotype 3 pneumococcal capsular polysaccharide-specific monoclonal antibodies against intranasal challenge with Streptococcus pneumoniae in mice. Infect. Immun. 77 : 1502–1513.19168739
10 García-Ojeda, P. A., M. E. Monser, L. J. Rubinstein, H. J. Jennings, K. E. Stein. 2000. Murine immune response to Neisseria meningitidis group C capsular polysaccharide: analysis of monoclonal antibodies generated in response to a thymus-independent antigen and a thymus-dependent toxoid conjugate vaccine. Infect. Immun. 68 : 239–246.10603394
11 Mukherjee, J., A. Casadevall, M. D. Scharff. 1993. Molecular characterization of the humoral responses to Cryptococcus neoformans infection and glucuronoxylomannan-tetanus toxoid conjugate immunization. J. Exp. Med. 177 : 1105–1116.8459205
12 Nussbaum, G., S. Anandasabapathy, J. Mukherjee, M. Fan, A. Casadevall, M. D. Scharff. 1999. Molecular and idiotypic analyses of the antibody response to Cryptococcus neoformans glucuronoxylomannan-protein conjugate vaccine in autoimmune and nonautoimmune mice. Infect. Immun. 67 : 4469–4476.10456888
13 Köhler, G., C. Milstein. 1975. Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 256 : 495–497.1172191
14 Rohatgi, S., D. Dutta, S. Tahir, D. Sehgal. 2009. Molecular dissection of antibody responses against pneumococcal surface protein A: evidence for diverse DH-less heavy chain gene usage and avidity maturation. J. Immunol. 182 : 5570–5585.19380805
15 Rohatgi, S., P. Ganju, D. Sehgal. 2008. Systematic design and testing of nested (RT-)PCR primers for specific amplification of mouse rearranged/expressed immunoglobulin variable region genes from small number of B cells. J. Immunol. Methods 339 : 205–219.18926828
16 Brochet, X., M. P. Lefranc, V. Giudicelli. 2008. IMGT/V-QUEST: the highly customized and integrated system for IG and TR standardized V-J and V-D-J sequence analysis. Nucleic Acids Res. 36 : W503–W508.18503082
17 Giudicelli, V., X. Brochet, M. P. Lefranc. 2011. IMGT/V-QUEST: IMGT standardized analysis of the immunoglobulin (IG) and T cell receptor (TR) nucleotide sequences. Cold Spring Harb. Protoc. 2011 : 695–715.21632778
18 Ye, J., N. Ma, T. L. Madden, J. M. Ostell. 2013. IgBLAST: an immunoglobulin variable domain sequence analysis tool. Nucleic Acids Res. 41 : W34–W40.23671333
19 Lefranc, M. P., C. Pommié, M. Ruiz, V. Giudicelli, E. Foulquier, L. Truong, V. Thouvenin-Contet, G. Lefranc. 2003. IMGT unique numbering for immunoglobulin and T cell receptor variable domains and Ig superfamily V-like domains. Dev. Comp. Immunol. 27 : 55–77.12477501
20 Deauvieau, F., S. Dussurgey, D. Rossignol, A. de Montfort, N. Burdin, B. Guy. 2009. Memory B and T cell responses induced by serotype 4 Streptococcus pneumoniae vaccines: longitudinal analysis comparing responses elicited by free polysaccharide, conjugate and carrier. Vaccine 28 : 576–582.19878751
21 Safari, D., H. A. Dekker, J. A. Joosten, D. Michalik, A. C. de Souza, R. Adamo, M. Lahmann, A. Sundgren, S. Oscarson, J. P. Kamerling, H. Snippe. 2008. Identification of the smallest structure capable of evoking opsonophagocytic antibodies against Streptococcus pneumoniae type 14. Infect. Immun. 76 : 4615–4623.18678667
22 Mawas, F., J. Niggemann, C. Jones, M. J. Corbel, J. P. Kamerling, J. F. Vliegenthart. 2002. Immunogenicity in a mouse model of a conjugate vaccine made with a synthetic single repeating unit of type 14 pneumococcal polysaccharide coupled to CRM197. Infect. Immun. 70 : 5107–5114.12183560
23 Polonskaya, Z., S. Deng, A. Sarkar, L. Kain, M. Comellas-Aragones, C. S. McKay, K. Kaczanowska, M. Holt, R. McBride, V. Palomo, . 2017. T cells control the generation of nanomolar-affinity anti-glycan antibodies. J. Clin. Invest. 127 : 1491–1504.28287405
24 Dahora, L. C., M. K. Verheul, K. L. Williams, C. Jin, L. Stockdale, G. Cavet, E. Giladi, J. Hill, D. Kim, Y. Leung, . 2021. Salmonella Typhi Vi capsule prime-boost vaccination induces convergent and functional antibody responses. Sci. Immunol. 6 : eabj1181.34714686
25 Ehrhardt, S. A., M. Zehner, V. Krähling, H. Cohen-Dvashi, C. Kreer, N. Elad, H. Gruell, M. S. Ercanoglu, P. Schommers, L. Gieselmann, . 2019. Polyclonal and convergent antibody response to Ebola virus vaccine rVSV-ZEBOV. Nat. Med. 25 : 1589–1600.31591605
26 Jackson, K. J., Y. Liu, K. M. Roskin, J. Glanville, R. A. Hoh, K. Seo, E. L. Marshall, T. C. Gurley, M. A. Moody, B. F. Haynes, . 2014. Human responses to influenza vaccination show seroconversion signatures and convergent antibody rearrangements. Cell. Host. Microbe 16 : 105–114.24981332
27 Parameswaran, P., Y. Liu, K. M. Roskin, K. K. Jackson, V. P. Dixit, J. Y. Lee, K. L. Artiles, S. Zompi, M. J. Vargas, B. B. Simen, . 2013. Convergent antibody signatures in human dengue. Cell Host Microbe 13 : 691–700.23768493
28 Skinner, N. E., C. O. Ogega, N. Frumento, K. E. Clark, S. Yegnasubramanian, K. Schuebel, J. Meyers, A. Gupta, S. Wheelan, A. L. Cox, . 2023. Convergent antibody responses are associated with broad neutralization of hepatitis C virus. Front. Immunol. 14 : 1201033.37180140
29 Chen, E. C., P. Gilchuk, S. J. Zost, N. Suryadevara, E. S. Winkler, C. R. Cabel, E. Binshtein, R. E. Chen, R. E. Sutton, J. Rodriguez, . 2021. Convergent antibody responses to the SARS-CoV-2 spike protein in convalescent and vaccinated individuals. Cell Rep. 36 : 109604.34411541
30 Koralov, S. B., T. I. Novobrantseva, K. Hochedlinger, R. Jaenisch, K. Rajewsky. 2005. Direct in vivo VH to JH rearrangement violating the 12/23 rule. J. Exp. Med. 201 : 341–348.15699070
31 Koralov, S. B., T. I. Novobrantseva, J. Königsmann, A. Ehlich, K. Rajewsky. 2006. Antibody repertoires generated by VH replacement and direct VH to JH joining. Immunity 25 : 43–53.16860756
