A New Pneumococcal Capsule Type, 10D, is the 100th Serotype and Has a Large cps Fragment from an Oral Streptococcus

The polysaccharide capsule is essential for the pathogenicity of pneumococcus, which is responsible for millions of deaths worldwide each year. Currently available pneumococcal vaccines are designed to elicit antibodies to the capsule polysaccharides of the pneumococcal isolates commonly causing diseases, and the antibodies provide protection only against the pneumococcus expressing the vaccine-targeted capsules. Since pneumococci can produce different capsule polysaccharides and therefore reduce vaccine effectiveness, it is important to track the appearance of novel pneumococcal capsule types and how these new capsules are created. Herein, we describe a new and the 100th pneumococcal capsule type with unique chemical and serological properties. The capsule type was named 10D for its serologic similarity to 10A. Genetic studies provide strong evidence that pneumococcus created 10D capsule polysaccharide by capturing a large genetic fragment from an oral streptococcus. Such interspecies genetic exchanges could greatly increase diversity of pneumococcal capsules and complicate serotype shifts.

S treptococcus pneumoniae (pneumococcus) has a storied history. It is one of the first pathogens ever discovered and is a major cause of pneumonia and mortality (1,2). Its virulence is largely due to its thick polysaccharide (PS) capsule (3), and studies of its capsule elucidated that the genetic material was DNA (4). Currently, the capsule is used to produce pneumococcal conjugate vaccines (PCVs), a major public health tool worldwide. Since PCV protection is serotype specific, PCVs provide protection only against the serotypes included in the vaccine (5). Pneumococcus displays one of many structurally diverse capsules, with 99 identified so far (Table 1) (6). The widespread use of PCVs has altered the natural distribution of pneumococcal serotypes by increasing invasive pneumococcal diseases (IPDs) and nasopharyngeal carriage of nonvaccine serotypes (7). This change in serotype distribution is called serotype replacement (8)(9)(10). Diversity in pneumococcal capsule types and serotype replacement are major challenges in making and using PCVs.
Serotype replacement can occur in several ways. It may occur when a minor preexisting serotype increases its prevalence following vaccination. For instance, PCV vaccination could have increased the prevalence of serotypes 6C and 19A, which previously existed as minor populations (11)(12)(13)(14). Alternatively, replacement may occur if a pneumococcus strain expressing a new capsule type appears by capturing a gene fragment from another pneumococcal strain or related species like oral streptococcus (15). The genetic transfer is facilitated since almost all the genes necessary for capsule biosynthesis are in a genetic cassette called cps in these species. Knowledge of capsular PS diversity, and understanding the origin of capsule diversity, is of fundamental importance in using PCVs and controlling pneumococcal infections (16).
To investigate capsule diversity, capsule loci of 21,853 pneumococcal genomes were studied as a part of the Global Pneumococcal Sequencing (GPS) project (17) and a mother-infant carriage study in the Maela refugee camp, Thailand (18). The genome studies revealed that several nasopharyngeal isolates from the children of Ethiopia (19), Cambodia (20), and Thailand (21) express a novel type of cps, labeled 39X. Their cps loci have parts of serotypes 6C and 39, separated by a gene of unknown origin, wcrO. The isolates were previously serotyped as either 39 or 10A by the Quellung and latex agglutination tests (17,(19)(20)(21). These isolates were provisionally designated as expressing serotype 39X, pending biochemical characterization, since a large genetic difference in cps may not confer a difference in the PS (22). Here, we determined the chemical structure of the 39X capsule to be unique, named the new type serotype 10D, and show that its cps has captured a large genetic fragment from oral streptococci.

RESULTS
Serotype 10D (39X) displays unique serologic properties. Following the confirmation that a pneumococcal strain, Cam853, has the genetic marker described for a novel capsule type "39X" (21) (see Fig. S1 in the supplemental material), we investigated its serologic properties along with five strains expressing related serotypes (serotypes 39, 10A, 10B, 10C, and 10F) (23) using a panel of commercially available antisera and agglutination reaction ( Table 2). Antisera used include factor serum, type serum, group serum, and pool serum, which recognize an epitope, serotype, serogroup, and multiple serotypes, respectively. Strain Cam853 agglutinated with factor serum 10d (FS10d) but failed to react with other antisera, including type 39 antiserum, even though the antisera reacted with the related strains as expected (23). In addition, Cam853 reproducibly failed to agglutinate with any of the three pool antisera (pool E, S, and T) as well as with serogroup 6 reagents. Therefore, Cam853 appeared to have serologic properties not described so far for pneumococci.
To investigate its serologic properties in detail, we examined strain Cam853 along with the five strains expressing related serotypes by using flow cytometry, which is more sensitive and reproducible than agglutination reactions. Reactions with type 39 antiserum and a monoclonal antibody (MAb) Hyp10AG1, which was produced as described previously (23), distinguished strain Cam853 and serotypes 39 and 10A from other related serotypes (10B, 10C, 10F, and 6C) ( Fig. 1 and Table 3). Strain Cam853 could be further distinguished from serotypes 10A and 39 by their reactions with pool S and group 10 antisera. Cam853 reacted with pool S, but not with group 10 antiserum, whereas serotype 39 reacted with none and serotype 10A reacted with both ( Fig. 1). Additional serologic properties are summarized in Table 3. The other two 39X strains (PATH4346 and Cam657) showed the same serologic profiles as Cam853 (Fig. S2). Taken together, 39X shares epitopes associated with serotype 10A but has a unique serologic profile.
The provisional name 39X was replaced with a new name, 10D, for the following reasons. Although 39X is genetically most similar to 39, it has substantial similarity to 6C and 10A cps. Serologically, 39X cross-reacts with FS10d and Hyp10AG1, which recognize a structural motif shared by serotypes 39 and 10A (Fig. S3) (24). Indeed, Cam853 was mistyped as 10A (17,20). Further, 10A will be in a future PCV (25) and may elicit antibodies cross-reacting with 39X. Thus, the name 10D would facilitate crossprotection investigations of new PCVs.
10D capsular polysaccharide has a unique chemical structure. Capsular PS of serotype 10D was purified from a culture seeded with one colony of strain Cam853 using anion-exchange chromatography. Each fraction was tested for a 10D capsule (in an inhibition enzyme-linked immunosorbent assay [ELISA]) as well as absorbance at 280 and 260 nm for protein and nucleic acid contamination, respectively ( Fig. 2A). Fraction 15, which had high concentration of 10D PS with low protein or nucleotide contaminants, was selected for nuclear magnetic resonance (NMR) studies. One-dimensional (1D) 1 H NMR spectra were obtained with Cam853 capsular PS as well as capsular PS from serotypes 6C, 10A, and 39. The latter three serotypes were investigated for direct comparison with Cam853 PS. As shown in Fig. 2B, the 1D 1 H NMR spectra of the anomeric carbons (chemical shifts of 4.6 to 5.6 ppm) are distinct for each serotype, clearly indicating that Cam853 capsule is chemically distinct from the other three. The  Ϫ Cam853 is serologically unique from other serogroup 10 members and serotype 39. Pneumococcal strains SSISP10A, SSISP10B, SSISP10C, SSISP10F, and SSISP39 are from Statens Serum Institut (SSI) (23) and represent serotypes 10A, 10B, 10C, 10F, and 39, respectively. All group-specific, type-specific, and factorspecific (FS) antisera are also from SSI. Cam853 reacted only with FS10d in the agglutination reaction. Negative reactions (Ϫ) and positive reactions (ϩ) are shown.
NMR spectrum of 10D showed a minor signal from teichoic acid (3.17 ppm), suggesting that the 10D capsular PS is sufficiently pure for detailed structural studies. Next, we employed additional NMR techniques to determine the chemical structure of 10D capsular PS (Fig. 3). Shifts in the NMR signals of 1 H and 13 C (i.e., chemical shifts) reflect the chemical milieu of each carbon atom, and they can be used to determine the structure of a PS. The 1 H and 13 C chemical shifts of 10D PS ( Fig. 3A and B and Table 4) were assigned using a set of standard two-dimensional NMR experiments as described in Material and Methods. To facilitate assignments of the chemical shifts for 10D PS, similar NMR experiments were also performed with the 10A, 6C, and 39 PS. Chemical shifts for 10A, 39, and 6C PS were identical to those reported in previous studies (24,26). The phosphate glycosidic linkage of 10D PS was confirmed by 1 H-31 P correlation (Fig. 3C). Table 4 shows the full assignments for 10D PS, and the deduced chemical structure of 10D PS is shown in Fig. 4 along with those of related serotypes. Taken together, the NMR data unambiguously confirmed that 10D produces a chemically distinct capsule.
Of note, the chemical structure of 10D allowed for the identification of biosynthetic roles of all genes in 10D cps locus as summarized in Fig. 4. 10D and 6C share the structure of the two glucose residues at the reducing end as well as two glycosyltransferases, wchA and wciN␤ in their cps (11,26). Therefore, both genes are involved in forming the glucose dimer. Ribitol is linked to the glucose dimer, and wcrO 10D is the only gene encoding a ribitol transferase in 10D cps. The remaining structure is formed by the products of genes (wcrC, wcrD, wciF, and wcrG), which are shared between serotypes 10A and 39 as described previously (24). . The x axes show the log fluorescence intensity, and the y axes show the number of events (cell counts). Hyp10AG1 is the serotype 10A-specific monoclonal antibody (MAb). All other reagents were obtained from Statens Serum Institut (SSI) (FS10d, factor serum 10d). Each strain was tested three times, and a representative result was shown.
Immunization with 10A polysaccharide elicits antibodies opsonizing 10D. Since serologic studies suggested that 10A PS may elicit antibodies cross-reacting with 10D, we investigated whether the 23-valent pneumococcal PS vaccine (PPV23), which contains the 10A serotype, induces antibodies cross-opsonizing serotype 10D. An opsonophagocytosis assay (OPA) with six pairs of pre-and postvaccinated immune sera a Pneumococcal strains SSISP10A, SSISP10B, SSISP10C, SSISP10F, and SSISP39 representing serotypes 10A, 10B, 10C, 10F, and 39, respectively, were procured from Statens Serum Institut (SSI), Copenhagen, Denmark. MJC705 strain represents the serotype 6C and was obtained from the Nahm laboratory bacterial strain collection. Cam853 is the 10D cps-containing strain and was obtained from during a post-PCV13 introduction colonization study at Angkor Hospital for Children/Cambodia Oxford Medical Research Unit. Pool E and pool S were designed to react with multiple serotypes (58). Group serum reacts to all the serotypes within one serogroup. Type serum reacts only with a single serotype. Factor serum (FS) reacts with a specific epitope in a serotype. All the antisera are polyclonal rabbit antisera and were obtained from SSI. Negative reactions (Ϫ) and positive reactions (ϩ) are shown. Cam853 reacted only with FS10d in agglutination reaction and flow cytometry. Cam853 also reacted with pool E and S antisera in flow cytometry, which otherwise showed negative results in agglutination reaction. Characterization of the 100th Pneumococcal Capsule Type ® was performed using serotype 10D (Cam853) and 10A (SSISP10A) strains as targets.
Vaccination greatly increased opsonic capacity to both serotypes 10A and 10D (P Ͻ 0.005 for both), and the mean OPA titer values for 10A and 10D increased about 10-fold following vaccination (Fig. 5A). Furthermore, the increase in OPA titer is specific to 10A and 10D, since preabsorption with homologous 10A PS reduced the 10A or 10D OPA FIG 3 Two-dimensional nuclear magnetic resonance (NMR) spectra of serotype 10D polysaccharide. 1 H-13 C HMQC spectra of anomeric carbons and the other carbons are shown in panels A and B, respectively. (C) 1 H-31 P HMBC spectrum of the ribitol phosphate region. NMR spectra were obtained at 45°C. In the peak labels, the letter indicates the sugar residue (see Table 4), and the number denotes the carbon position on each unit. Peak assignments are shown in Table 4. The asterisk denotes an unassigned peak. Horizontal dashed lines connect the two proton signals associated with the sugar residue at the indicated carbon position. activities almost completely (P Ͻ 0.0001 for both), but preabsorption with 24F PS (negative control) did not reduce the 10A (P Ͻ 0.0001) or 10D (P Ͻ 0.001) OPA activities (Fig. 5B). It is therefore likely that 10A PS would elicit protective antibodies against 10D. 10D cps has a ϳ6 kb fragment that is syntenic and homologous to a Streptococcus mitis cps. To understand how 10D cps arose, we searched for the origin of wcrO 10D , which is absent in the cps of serotypes 6C and 39. Phylogenetic analysis of wcrO 10D identified various genes encoding ribitol phosphate transferases from oral streptococci and pneumococci with a 45 to 60% nucleotide identity (Fig. 6). No genes from pneumococcal cps loci showed high homology to wcrO 10D . However, wcrO 10D showed high homology (90% nucleotide identity, 94% amino acid identity) to one gene (RS00925) from an oral streptococcus strain, SK145 (Fig. 6).
Even more surprising is the fact that SK145 and 10D cps displayed very high homology (ϳ94% nucleotide identity) for a very large (ϳ6-kb) cps region beginning from the end of the 6C-like region (wcrO) to wzy (Fig. 7). SK145 cps even contained the unique 3= end of wciN␤ 10D and the genetic fragment of wciN␣ 10D , both of which came from wciN␣ SK145 , including the 7-base overlap that both wciN␣ and wciN␤ have (bases 6073 to 6079 of 10D cps). As a result, wciN␤ 10D is shorter than wciN␤ 6C by 6 nucleotides at the 3= end, and 10D WciN␤ protein should be shorter than 6C WciN␤ (373 versus 375 amino acids) (Fig. 7). 5= to the 7-base overlap, 10D cps (from cpsA to wciN␤) is highly homologous (98% nucleotide identity) to the 6C cps locus. The 3= end of 10D cps (from wcrG to glf) is more homologous (88% nucleotide identity) to serotype 39 cps than to SK145 cps. Thus, serotype 10D cps might have evolved by double recombination events between serotype 6C, strain SK145, and serotype 39 cps loci. However, given that upper respiratory streptococci are genetically very diverse and their cps loci have not been extensively studied, it is more likely that there was a single recombination event between serotype 6C and an SK145-like strain harboring all the genes from the wciN␣ genetic fragment to glf (ϳ9 kb). The seven bases shared between 6C and SK145 could have facilitated the crossover between serotype 6C cps and SK145-like strain (Fig. 7).

DISCUSSION
Pneumococcus is a highly successful pathogen in part due to its high level of capsule diversity, resulting in a plethora of unique serotypes (6). Recent studies show its capsule diversity to be greater than previously appreciated. Although almost a century of studies based on Quellung reactions identified 90 different capsule types (27), the use of MAbs and genetic screens enabled the discovery of nine new capsule types starting with the discovery of serotype 6C in 2007 (11). Herein, we describe the 100th pneumococcal capsule type, named serotype 10D, by providing the complete chemical structure along with identification of biosynthetic roles for all the cps genes. In addition, the chemical structure explains its serologic properties: its reaction with polyclonal serum (FS10d) and MAb Hyp10AG1, which targets ␤Gal(1-6) created by wcrG (24).
Phylogenetic analysis of 10D cps suggests that 10D cps captured the wcrO 10D gene from an oral streptococcus species, which often contains pneumococcus-like cps loci and actually produces capsular PSs (28)(29)(30)(31)(32). Specifically, wcrO 10D sequence has low homology with the four S. pneumoniae wcrO genes in serotype 33C, 34, 35F, and 36 (40 to 50% amino acid identity) even though pneumococcal cps loci have been extensively studied (more than 20,000 pneumococcal isolates [17]). In contrast, wcrO 10D shows surprisingly high homology to a gene (RS00925) in the cps gene of S. mitis strain SK145. In fact, the wcrO 10D gene may be common among oral streptococci, since oral streptococcal cps loci have not been studied as extensively as pneumococci. Taking all the evidence together, the wcrO 10D gene has likely originated from oral streptococci, not from pneumococci.
Interspecies capsule gene exchanges have been previously suggested (15). For instance, serotypes 19B and 19C may have emerged when 19F captured an ϳ13-kb fragment with 10 functional genes from SK564, an S. mitis strain (21). Another example is suggested by an ϳ16.5-kb homologous region with 16 genes shared between cps of serotype 2 and strain SK95 (28). Like these previous examples, examination of 10D and SK145 cps also shows a very large (ϳ6-kb-long) syntenic and highly homologous region, which has five functional genes, including wcrO 10D . Unlike previous examples, however, the 10D cps includes at the 5ˊend of the syntenic region a short nonfunctional gene fragment of wciN␣ that lacks evolutionary constraint for retention. Since SK145 has a full-length transcribable wciN␣ gene, the presence of nonfunctional wciN␣ gene fragment in 10D cps suggests recent genetic transfer from strain SK145 to 10D. Moreover, the unique 3= end of wciN␤ 10D may be the interspecies recombination crossover point, which was not reported in previous examples (21,28). Thus, 10D cps provides the most compelling evidence for interspecies cps exchange with clear directionality.
Interspecies transfer of the cps locus would significantly increase the diversity of pneumococcal capsule types since the human upper airway harbors many sources of genes. Oral streptococcus species include not only the mitis group species but also many genetically diverse species that have cps loci resembling pneumococcal cps (28,33). Also, genetic materials useful for increasing pneumococcal capsule diversity can be from unrelated Gram-positive or even Gram-negative species. For instance, Streptococcus thermophilus has cps-like loci (34), and E. coli has a lipopolysaccharide (LPS) synthesis locus which resembles pneumococcal cps loci (35). While a large genetic survey (involving more than 21,000 strains) in selected sites found only several pneumococcal strains with novel genetic features (17,21), surveys of additional sites, including the sites with history of extensive usage in PCVs, may reveal additional novel serotypes.
Discoveries of novel capsule types can have a direct impact on PCV usage and  Table S1 in the supplemental material. SGT, serogroup/type.
Characterization of the 100th Pneumococcal Capsule Type ® design. Prior to the discovery of serotype 6C, it was mistyped as "6A" (11). Discovery of serotype 6C allowed one to show that 7-valent PCV (PCV7) was cross-protective against 6A but not 6C (12), and helped improve PCV design. Like 6C, epidemiologic surveys have not distinguished 10D from 10A so far and reported increased prevalence of "10A" following the use of PCVs in children (36)(37)(38)(39). One should now consider that PCV usage may have increased the prevalence of 10D as well as 10A. New PCVs including 10A PS may provide cross-protection against 10D, since we found 10D-opsonic antibodies in persons vaccinated with PPV23, which contains 10A PS. Since PCV20 under development contains 10A, it would be interesting to examine its effect on 10D IPDs and carriage. The natural prevalence of different pneumococcal capsule types in the upper respiratory tract represents equilibrium of various selection forces in nature. When a  (21) with 6C cps (GenBank accession no. EF538714), SK145 cps (NZ_JYGS01000001.1; contig 1), and 39 cps (CR931711) using dot plot analysis. The x axis shows the 10D cps gene arrangement; 6C, SK145, and 39 cps loci are plotted on the y axis. For strain SK145, reverse complement sequence was used for comparison. The numbers on the x and y axes relate to arbitrary nucleotide base numbers (in kilobases) to show relative size. The solid black lines represent the homologous regions, with the percentage nucleotide identity indicated above the line. The vertical gray-shaded areas indicate overlap regions for the potential recombination events I and II. (B) The nucleotide sequences of cps loci for serotype 6C, serotype 10D, and strain SK145 at the transition region for the potential recombination event I. The sequence in bold type indicates the overlap region. Underlined nucleotides indicate the wciN␤ 6C and wciN␤ 10D translation stop codons. Lowercase letters in 10D cps indicate the nucleotide mismatches. The base numbers indicate the actual nucleotide position in the corresponding cps loci. The gray-shaded area indicates the region of homology "upstream" (cps 6C and cps 10D) and "downstream" (cps 10D and cps SK145) of the potential recombination site. The dashed horizontal lines indicate the genes involved in the potential recombination event I. new selective pressure is introduced, the new force alters the equilibrium demonstrated by serotype replacement following the use of PCVs (7). The selective force could foster interspecies genetic exchange. The appearance of antibiotic resistance genes among pneumococci followed the extensive use of antibiotics (40)(41)(42)(43). An intriguing question is whether the use of PCVs increases capturing of oral streptococcus cps by pneumococcus resulting in "vaccine-escape" strains (31). Another interesting possibility is that PCVs may impose selective pressure onto oral streptococci as well (28)(29)(30) and select for nonvaccine capsule types, since they appear to express capsule types that are similar to PCV serotypes. In summary, the large pneumococcal capsule diversity continues to be a significant threat to human health, and the unusual sources of capsule diversity are contributing to the problem. Thus, the long-term strategy of using PCVs must include improved serotype knowledge of pneumococci as well as oral streptococci residing in our upper respiratory tract.

MATERIALS AND METHODS
Bacterial strains and cultivation. Strains Cam657 (GPS_KH_COMRU657; SRA accession no. ERR2681221) and Cam853 (GPS_KH_COMRU853; ERR2680924) were collected between 2016 and 2017 during a post-PCV13 introduction colonization study at the Angkor Hospital for Children/Cambodia Oxford Medical Research Unit (20). Strain PATH4346 (GPS_US_PATH4346; ERR1453813) was collected in 2007 as a part of cross-sectional population-based study in Ethiopian children during mass azithromycin distribution for trachoma (19). As a part of the Global Pneumococcal Sequencing project (https://www .pneumogen.net/gps/), the whole genomes of these carriage strains were sequenced and these strains were genetically typed as 39X (17).
All pneumococcal strains were isolated on blood agar plates with 5% sheep blood (Remel Laboratories, Lenexa, KS) after overnight incubation at 37°C with 5% CO 2 . Isolated colonies were inoculated into Todd-Hewitt broth with 5% yeast extract (THY) and grown to mid-log density (optical density at 600 nm [OD 600 ] of 0.5), and bacteria were frozen in THY with 15% glycerol at -80°C for later use. All isolates had colony morphology typical of pneumococci and were identified as pneumococcus by matrix-assisted laser desorption ionizationϪtime of flight mass spectrometry (MALDI-TOF MS) using Vitek MS Knowledge Base V3.2 (bioMérieux) (44).
Genetic confirmation of 39X (serotype 10D) genotype. The 39X genotype of strain Cam853 was confirmed using the primers 5943 (5=-AGTTAGGTAGTGCTGCTTTCC-3=) and 3773 (5=-CTATCAGCAGTAGT TAGTGTAATTATC-3=) to amplify a 1,641-bp unique region (wciN␤-wcrO-wcrC) of the 39X cps locus. The PCR mixture was 25 l of 1ϫ Ex Taq PCR buffer containing 50 to 100 ng of genomic DNA, 0.3 U of Ex Taq polymerase (3 U/l; TaKaRa), 1 l deoxynucleotide triphosphates (dNTPs) (2.5 mM each [Fermentas, USA]), and 1 l of each forward and reverse primer (each 10 M). Thermal cycling was performed in GeneAmp PCR system 9700 (Applied Biosystems, USA) under the following conditions: 94°C for 10 min, followed by 35 amplification cycles, with 1 cycle consisting of 94°C for 30 s, 58°C for 30 s, and 72°C for 1 min, and a final extension step at 72°C for 10 min. The PCR products were separated by electrophoresis on 1.5% agarose gel. The amplified products were purified using a QIA quick PCR purification kit (Qiagen, Germany) following the manufacturer's protocol. Purified PCR products were subjected to bidirectional sequencing which was performed by Heflin Center Genomics Core Lab at the University of Alabama at Birmingham (UAB).
Serological analysis of serotype 10D isolates. The serological properties of the serotype 10D (39X) isolates were assessed by the slide agglutination reaction and flow cytometry as previously described (27,45) using a panel of polyclonal antisera obtained from the SSI (Copenhagen, Denmark) as detailed in Table 3, or using our in-house 10A-and 6C/6D-specific monoclonal antibodies (MAbs) as depicted in Fig. 1 and Fig. S4 in the supplemental material (46). The MAbs were produced as described previously (23). For flow cytometry, frozen bacterial stocks were thawed and washed in fluorescence-activated cell sorting (FACS) buffer (1ϫ phosphate-buffered saline [PBS], 3% fetal bovine serum [FBS], 0.1% NaN 3 ), incubated with 1:1,000 dilutions of polyclonal antisera, or 1:100 and 1:20 dilutions of 10A-specific and 6C/6D-specific MAbs, respectively. After the isolates were washed, they were labeled with a phycoerythrin (PE)-conjugated anti-rabbit Ig secondary antibody (Southern Biotech, Birmingham, AL) at 1:1,000 dilution or a PE-conjugated anti-mouse Ig secondary antibody (BD Biosciences, San Jose, CA) at 1:200 dilution. After the bacteria were washed, their fluorescence was measured with a BD Accuri flow cytometer. Data were analyzed in FCS Express version 3.0. The assay was performed with the indicated bacterial strains and antisera three times.
Inhibition ELISA. To detect the capsule polysaccharide (PS) during purification, an inhibition-type enzyme-linked immunosorbent assay (ELISA) was performed. Briefly, the wells of ELISA plates (Corning Costar Corp., Acton, MA) were coated at 37°C with 1 g/ml of 10A capsular PS (SSI, Copenhagen, Denmark) for 5 h in phosphate-buffered saline. After the plates were washed with PBS containing 0.05% Tween 20, 50-l portions of the samples containing PS were added to the wells along with 50 l of factor serum 10d (FS10d) at 1:10,000 dilutions. After 1 h of incubation at room temperature, the plates were washed and incubated for 1 h with 100 l of alkaline phosphatase-conjugated goat anti-rabbit Ig Characterization of the 100th Pneumococcal Capsule Type ® (Southern Biotech) at 1:3,000 dilutions. The amount of the enzyme immobilized to the wells was then determined by incubating paranitrophenyl phosphate substrate (Sigma) in diethanolamine buffer for 1 to 2 h at room temperature and measuring the OD 405 with a microplate reader (BioTek Instruments Inc., Winooski, VT). Purified 10A PS (SSI, Copenhagen, Denmark) was used as an assay standard (Fig. S5).
Purification of capsular polysaccharide. A single colony of the Cam853 strain was grown in 10 ml of THY broth and then expanded in 1 liter of a chemically defined medium (47) supplemented with choline chloride (1 g/liter), sodium bicarbonate (2.5 g/liter), and cysteine HCl (0.73 g/liter). After overnight incubation at 37°C, bacteria were separated from the supernatant by centrifugation (15,344 ϫ g, 30 min, 4°C), washed, and resuspended in 18 ml of 0.9% NaCl. After the pH was adjusted to 7, the bacterial suspension was added with 100 l of 10% sodium deoxycholate, 200 l of mutanolysin (10 U/l), and 1.7 ml of lysozyme (11.8 mg/ml) and incubated at 37°C for 48 to 72 h. The resulting lysate was centrifuged to remove debris, dialyzed overnight against 4 liters of 5 mM Tris (pH 7.3) (dialysis tubing, 3,500molecular-weight cutoff). The dialyzed product was applied to a DEAE Sepharose (GE Healthcare, Uppsala, Sweden) anion-exchange column (40 ml). After the column was washed with 2.5 volumes of 5 mM Tris (pH 7.4), elution was performed with a linear gradient of NaCl ranging from 0 M to 0.4 M over 37 fractions (2.5 ml/fraction). Each fraction was tested for OD 260 , OD 280 , and 10D PS with an inhibition ELISA as described above. Fraction 15 was selected for PS studies and was lyophilized after dialysis.
Opsonophagocytosis assay. To investigate whether the 10A serotype elicits cross-opsonizing antibodies to serotype 10D, we adapted a well-characterized UAB opsonophagocytosis assay (OPA) (48) (and described in detail at https://www.vaccine.uab.edu). OPA was performed with six pairs of pre-and post-PPV23-vaccinated immune sera using serogroup/type (SGT) 10A and 10D as targets. Briefly, 30 l of bacteria suspended in OBB (Hanks' buffer supplemented with 0·1% gelatin and 5% fetal calf serum) was mixed with 10 l of baby rabbit serum (BRS) of specified concentration, and 40 l of differentiated HL60 cells (10 7 cells/ml) in OBB. The mixture was incubated with shaking (700 rpm) for 45 min at 37°C with 5% CO 2 . Ten microliters from each well was spotted onto THY agar plates, and the bacterial colonies were counted after overnight incubation. To examine the specificity of the 10A-induced immune response, postvaccinated sera were preabsorbed with the homologous 10A capsule PS (200 g/ml in OBB) or heterologous 24F (negative-control) capsule PS (200 g/ml in OBB), incubated for 30 min at room temperature, followed by the typical OPA procedure. As a control, each sample was tested without the preabsorption of any external capsule PS. NMR spectroscopy. 1 H-1 H and 1 H-13 C nuclear magnetic resonance (NMR) data were collected at 45°C on a Bruker Avance II (700 MHz 1 H) or Avance III (600 MHz 1 H) spectrometers equipped with cryogenic triple-resonance probes, processed with NMRPIPE (49) and analyzed with NMRVIEW (50). NMR samples were prepared by dissolving ϳ5 mg of lyophilized PS in 0.5 ml of 99.99% D 2 O (Cambridge Isotope Laboratories). Complete assignments of 1 H and 13 C signals were achieved by two-dimensional nuclear Overhauser spectroscopy ( 1 H-1 H NOESY), correlation spectroscopy ( 1 H-1 H COSY), total correlation spectroscopy ( 1 H-1 H TOCSY), heteronuclear multiple quantum coherence ( 1 H-13 C HMQC), and heteronuclear multiple bond correlation ( 1 H-13 C HMBC) spectra. HDO signal was used as a reference. For the 1 H-31 P HMBC experiment, ϳ5 mg lyophilized PS was dissolved in 200 l of 99.99% D 2 O (Cambridge Isotope Laboratories) and transferred into a 3-mm NMR tube. NMR spectra were recorded at 45°C on a Varian Inova DD2 NMR spectrometer ( 1 H, 599.7 MHz) equipped with a room-temperature 3-mm inverse broadband probe (Complex Carbohydrate Research Center, University of Georgia, Athens, GA). The 1 H-31 P HMBC spectrum was collected with 2,000 points, 64 increments, four scans per increment, and spectral widths of 7,184 Hz ( 1 H) and 4,856 Hz ( 31 P). The HMBC long-range transfer delay corresponded to a coupling of 8 Hz. 31 P chemical shifts were referenced to 85% H 3 PO 4 at 0 ppm. Genomic analysis. To predict the origin and function of the wcrO 10D gene, we searched for the homologous genes in oral streptococcal and pneumococcal sequence data in GenBank. Phylogenetic analyses were conducted based on genetic distance using Geneious prime v2019.2 (Biomatters). Fulllength nucleotide and translated amino acid sequences above 45% and 30% identity, respectively (multiple alignments by MAFFT plug-in) were used to construct the phylogenetic tree based on the neighbor-joining method using Tamura-Nei genetic distance model with 1,000 bootstrap replicates. Display and manipulation of the phylogenetic tree (Newick format) were performed using the online tool Interactive Tree of Life (51). Gene name and function were predicted based on the already published functions and linkage specificities of the pneumococcal cps genes (52).
To infer the probable evolutionary origin of serotype 10D, we examined the 10D cps locus (SRA accession no. ERR051587) (21) for synteny and homology against serotype 6C (GenBank accession no. EF538714), strain SK145 (NZ_JYGS01000001.1; contig 1), and serotype 39 cps loci (CR931711) by dot plot analysis using YASS (53). The SK145 GenBank sequence was in the opposite orientation, and its reverse complement sequence was used for comparison. Forward strands of each cps locus were analyzed with the following parameters: scoring matrix match, ϩ1; transversion, Ϫ5 (no correction); gap costs (opening Ϫ16 and extension Ϫ4), E-value threshold ϭ 10, and X-drop threshold ϭ 30. Recombination crossover regions were identified from a dot plot using Geneious prime.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.