Comparison of Serologic and Genetic porB-Based Typing of Neisseria gonorrhoeae: Consequences for Future Characterization

Bacterial isolates and culture conditions.A total of 108 N. gonorrhoeae isolates were examined in the present study (see Fig. 1A and 2A). The clinical isolates (n = 74) were cultured from patients in Sweden between 1998 and 2001. The N. gonorrhoeae reference strains (n = 34) originated from different geographic localities between 1973 and 1997. The selection of the clinical isolates and reference strains was based on the results of the initial serovar determination and on results obtained in a previous study (30). One to three isolates that differed only in the reactivity with a single MAb (of all the 12 used MAbs) were included, as well as all isolates comprising a unique porB gene sequence. The isolates comprised 28 different serovars, as well as two nonserotypeable isolates, according to the initial serovar determination performed on N. gonorrhoeae isolates at the Swedish Reference Laboratory for Pathogenic Neisseria (see Fig. 1A and 2A). All isolates were cultured as previously described (30) and preserved at −70°C.

• Open in new tab
• Download powerpoint

FIG. 1.

(A) Neighbor-joining phylogenetic tree of the entire porB1a alleles (924 unambiguously aligned nucleotides) coding for the mature PorB1a proteins of N. gonorrhoeae isolates (n = 35), comprising nine different serovars and two nonserotypeable (NT) serovars according to the initial serovar determination. The serovars of the isolates are included; discrepancies between initial and repeated (one to three times) serovar determination are indicated by boldface italics. *, the isolates were reactive with MAbs 2F12, 5G9, 5D1, and assigned serovar IA-25 in the present study. (B) Neighbor-joining phylogenetic tree of the loop sequences 1, 2, 4, and 8 (174 unambiguously aligned nucleotides) of the porB1a alleles in the isolates described above. The entire and partial porB1a sequences, respectively, of the isolates 34/98 and 114/98, which represent an outgroup, were used to root the trees. The lengths of the branches leading to the outgroups have been reduced by a factor of three. All bootstrap values (as a percentage of 1,000 resamplings) are shown.

• Open in new tab
• Download powerpoint

FIG. 2.

(A) Neighbor-joining phylogenetic tree of the entire porB1b alleles (999 unambiguously aligned nucleotides) coding for the mature PorB1b proteins of N. gonorrhoeae isolates (n = 73), comprising 19 different serovars according to the initial serovar determination. The serovars of the isolates are included; discrepancies between initial and repeated (one or two times) serovar determination are indicated by boldface italics. (B) Neighbor-joining phylogenetic tree of the loop sequences 1, 3, 5, 6, 7, and 8 (369 unambiguously aligned nucleotides) of the porB1b alleles in the isolates described above. The entire and partial porB1b sequence, respectively, of the strain CCUG 42287, which represents an outgroup, was used to root the trees. All bootstrap values (as a percentage of 1,000 resamplings) are shown. The isolates that were indistinguishable by using only the loop sequences are shaded.

Serological characterization.Serovar determination was performed by a coagglutination technique (23) with the Genetic Systems (GS) panel of MAbs (14), which comprises six different MAbs against PorB1a and six different MAbs against PorB1b.

Isolation of genomic DNA.Isolation of bacterial DNA was performed as previously described (30). Briefly, bacterial suspensions (approximately 3 × 10⁸ cells/ml) were prepared in sterile 0.15 M NaCl. DNA was isolated from 500 μl of each suspension using the Dynabeads DNA DIRECT Universal kit (Dynal ASA, Oslo, Norway) according to the manufacturer's instructions. The DNA preparations were stored at 4°C prior to PCR.

porB PCR.The entire porB gene was amplified by using the primers PorBU and PorBL, yielding a PCR product of approximately 1.0 to 1.1 kbp, as previously described (30). Briefly, a PCR mixture (50 μl) was prepared by using 1.0 U of AmpliTaq Gold DNA polymerase (Applied Biosystems, Foster City, Calif.), 1× PCR Gold buffer (Applied Biosystems), 2.5 mM MgCl₂ (Applied Biosystems), 0.1 mM deoxynucleoside triphosphates (Applied Biosystems), 0.5 μM concentrations of each primer, and 1 μl of the DNA template. The cycling parameters of the amplification were as follows: an enzyme activation step at 94°C for 10 min, followed by 30 sequential cycles of 94°C for 30 s, 60°C for 30 s, and 72°C for 90 s. At the end of the final cycle, an extension phase at 72°C for 4 min was included before storage at 4°C.

DNA sequencing.The amplified porB1a or porB1b allele was sequenced using the primers PorBU, PorBL, PorB1bU, PorB1bL, PorB1aU, and PorB1aL as previously documented (30). Briefly, the PCR products were purified with the High Pure PCR product purification kit (Roche Diagnostics, Mannheim, Germany) and cycle sequenced by using the ABI BigDye terminator cycle sequencing ready reaction kit (Applied Biosystems, Warrington, United Kingdom). The sequence extension products were purified with the DyeEx Spin kit (Qiagen, Hilden, Germany), and the nucleotide sequences were determined by using an ABI PRISM 310 genetic analyzer (Applied Biosystems). The sequence of each strand of each compiled sequence was determined.

Sequence alignments and phylogenetic analysis.Multiple-sequence alignments of the porB gene segments encoding the mature PorB, according to the model of van der Ley et al. (32), as well as of the deduced amino acid sequences of the mature proteins, were performed with BioEdit (version 5.0.9) software and by manual adjustment. Phylogenetic trees were constructed with TREECON (version 1.3b) software by using the Jin and Nei substitution model, the Kimura evolutionary model, an α-value of 0.5, and the neighbor-joining method (31). Bootstrap analysis (6) with 1,000 resamplings was performed by using TREECON software with the same settings described above. The porB1a sequences of the isolates 34/98 and 114/98 (see Fig. 1) and the porB1b sequence of the reference strain CCUG 42287 (see Fig. 2) were used as outgroups to root the trees. The alignments were not stripped of gaps before phylogenetic analysis because most indels occurred in the variable loops. Thus, important phylogeny would have been lost if gap stripping had been used.

The numbers of synonymous and nonsynonymous substitutions, the numbers of each type of site, and the p distances (pS and pN [for synonymous and nonsynonymous, respectively]) were calculated as overall means by the method of Nei and Gojobori (17) using the MEGA (version 2.1) software (15) with pairwise deletions of indels.

This study was performed in the National Reference Laboratory for Pathogenic Neisseria, Department of Clinical Microbiology, Örebro University Hospital, Örebro, Sweden.

Sequencing of porB1a genes from PorB1a strains (n = 35).In the alignment of the porB1a gene sequences encoding the mature PorB1a proteins (924 unambiguously aligned nucleotides), a total of 67 polymorphic nucleotide sites (a ratio of 7.2 per 100 sites) were identified. The regions of the gene encoding the amino acid loops of the porin exhibited 50 polymorphic sites (a ratio of 13.9) compared with 17 (a ratio of 3.0) in the regions between the surface-exposed loops. The p distances of nonsynonymous substitutions (substitutions that result in amino acid replacements per 100 nonsynonymous sites) were 4.6 and 1.0 in loops and interspacing regions, respectively. Synonymous (silent) p distances were 0.2 and 1.2 in the loops and interspacing regions, respectively (Table 1). Thus, the pS/pN ratio was 0.043 in the loops and 1.2 in the interspacing regions. The regions of the gene encoding loops 1, 2, 4, and 8 of the proteins were the most heterogeneous ones, whereas regions encoding loops 3, 5, 6, and 7 were more conserved (Table 2). The alignment of the deduced amino acid sequences of the mature PorB1a proteins contained 36 (30 per 100 sites) heterogeneous sites located in the predicted surface-exposed loops and 12 (6.4 per 100 sites) in the interspacing sequences.

The phylogenetic analysis of the entire porB1a alleles (a total of 924 bp) identified 22 unique gene sequences in the 35 PorB1a isolates. Identical discrimination between the isolates was obtained when the analysis was based only on the four most variable gene regions (a total of 174 bp) (Fig. 1). However, when only the four loop sequences were used, the phylogenetic association was somewhat weaker, as indicated by lower bootstrap values (Fig. 1).

Sequencing of porB1b genes from PorB1b strains (n = 73).The individual porB1b alleles were 57 to 69 nucleotides larger than the porB1a alleles, mostly due to a longer loop 5 in the mature protein (Tables 1 and 2). The alignment of the porB1b sequences encoding the mature PorB1b proteins (999 unambiguously aligned nucleotides) exhibited a total of 145 polymorphic nucleotide sites (a ratio of 14.5). The regions of the gene encoding the loops of the porin exhibited 103 polymorphic sites (a ratio of 23.7) in comparison to 42 (a ratio of 7.4) in the interspacing regions. Nonsynonymous substitutions existed in p distances of 5.9 and 1.1 in loops and interspacing sequences, respectively. Synonymous p distances, however, were similar in loops (4.2) and interspacing regions (4.3) (Table 1). Thus, the pS/pN ratio was 0.7 in the loops and 3.9 in the interspacing regions. The regions of the porB1b gene encoding loops 1, 3, 5, 6, 7, and 8 of the proteins exhibited the most heterogeneity, whereas regions encoding loops 2 and 4 were more homologous (Table 2). The alignment of the deduced amino acid sequences of the mature PorB1b proteins contained 50 (34.5 per 100 sites) heterogeneous sites located in the predicted surface-exposed loops and 15 (8.0 per 100 sites) in the interspacing sequences.

According to the phylogenetic analysis of the entire porB1b alleles (a total of 981 to 993 bp), the 73 PorB1b isolates comprised 65 unique gene sequences. Nearly identical phylogenies were obtained when the analysis was based only on the six most heterogeneous regions, encoding six of the loops, of the gene (a total of 351 to 363 bp); 64 out of the 65 genetic variants were identified (Fig. 2). The entire porB1b alleles of these two indistinguishable isolates differed by a single A→T transversion at nucleotide site 826, located in the region encoding the amino acid sequence between loops 6 and 7. However, when only the six loop sequences were used, the phylogenetic association was somewhat weaker, as indicated by lower bootstrap values (Fig. 2).

Comparison of PorB-based serovar determination and porB gene sequencing.The 35 PorB1a isolates (representing 9 different serovars and 2 nonserotypeable isolates according to the initial serovar determination) and the 73 PorB1b isolates (initially representing 19 different serovars) comprised 22 unique porB1a gene sequences and 65 unique porB1b sequences, respectively. Overall, a substantial genetic heterogeneity was identified in the porB sequences of isolates determined to be different serovars as well as identical serovars. However, a few isolates designated as different serovars in the same assay in most cases differing in reactivity to single MAbs, displayed identical porB gene sequences (Fig. 1A and 2A).

In the present study, differences in the multiple-sequence alignments of the deduced PorB amino acid sequences of strains that differed only by the recognition of a single MAb in the serovar determination were also examined. This strategy was used in an attempt to identify amino acid residues critical for single-MAb reactivity. The results of the present study as well as those of previous publications, which in most cases are in agreement, are summarized in Table 3. Discrepancies and new findings were identified, however (Table 3). Consequently, in contrast to the findings of Mee et al. (16), the sequence ¹⁸VAHHG²², in combination with ²⁹¹GTEKF²⁹⁵, was identified solely in isolates reactive with the MAb 5D1. The sequence ²⁹¹GT(E/Q)KF²⁹⁵ was identified in all 5G9-reactive isolates but also in nonreactive isolates that did not show any sequence disparities in any other amino acid loops. The sequence ²¹¹DAKLTWRDD²¹⁹ was identified only in 6D9-reactive isolates. The amino acid residues ¹¹⁸IAQPEE¹²³, critical for recognition by the MAb 4G5, probably as part of a conformational epitope (3), were identified also in nonreactive isolates that did not differ in the amino acid sequences of any other loop. In addition, the sequence ¹¹⁸IARPEE¹²³ was found only in 4G5-reactive isolates. In the present study, the amino acid sequences ²¹⁵TWRD²¹⁸, ²⁵⁵DA(D/G)H²⁵⁸, and ²⁸⁸KGKGA²⁹² were merely found in isolates recognized by the MAb 4A12. The amino acid sequence ²³³KLYQNQIVRD²⁴², in contrast to the findings of Poh et al. (20), and ²³³KLYQNQIVRG²⁴² were also identified solely in 2D4-reactive isolates. According to previous studies, the amino acid residues ¹⁹⁷YSIPS²⁰¹ were critical for the reactivity with MAb 3C8 and the two residues ¹⁹⁸SI¹⁹⁹ comprised the minimum epitope (3, 4). In the present study, the amino acid residues ¹⁹⁷YSIPS²⁰¹ but also ¹⁹⁷YS(M/T/V)PS²⁰¹ or ¹⁹⁷YNIPS²⁰¹ were exclusively found in 3C8-reactive isolates (Table 3). No amino acid residues critical for the reactivity with MAb 2F12 (PorB1a) and the MAbs 2H1, 1F5, 2D6, and 2G2 (PorB1b) could be identified, due to the absence of distinct amino acid residue disparities, at least in the extracellular amino acid loops, between reactive and nonreactive strains.

Notably, repeated serovar determination for the 35 PorB1a isolates and the 73 PorB1b isolates determined 11 PorB1a isolates and 23 PorB1b isolates, respectively, to be different serovars than in the initial testing. This was mostly due, however, to nonreproducible reactivity with single MAbs (Fig. 1 and 2). One PorB1a isolate (isolate 57/01) was repeatedly nonserotypeable despite a porB1a gene sequence that was identical to that of two other isolates that occasionally expressed serovar-specific epitopes (isolates 51/01 and 260/00) (Fig. 1A).