Neisseria meningitidis escape from the bactericidal activity of a monoclonal antibody is mediated by phase variation of lgtG and enhanced by a mutator phenotype

Abstract

Bacteria adapt to environmental changes through high-frequency switches in expression of specific phenotypes. Localized hypermutation mediated by simple sequence repeats is an important mechanism of such phase variation (PV) in Neisseria meningitidis. Loss or gain of nucleotides in a poly(C) tract located in the reading frame results in switches in expression of lgtG and determines whether a glucose or a phosphoethanolamine (PEtn) is added at a specific position in the inner core lipopolysaccharide (LPS). Monoclonal antibody (MAb) B5 is bactericidal for N. meningitidis strain 8047 when PEtn is present in the inner core LPS and lgtG is switched "off." Escape from the bactericidal activity of this antibody was examined by subjecting strain 8047 to multiple cycles of growth in the presence of MAb B5 and human serum. Escape variants with alterations in the lgtG repeat tract rapidly accumulated in bacterial populations during selection with this antibody. Strain 8047 was outcompeted in this assay by the 8047 Delta mutS strain due to the elevated PV rate of this mismatch repair mutant and hence the greater proportion of preexisting phase variants of lgtG in the inoculum. This mutS mutant was also more virulent than strain 8047 during escape from passive protection by MAb B5 in an in vivo infant rat model of bacteremia. These results provide an example of how PV rates can modulate the occurrence and severity of infection and have important implications for understanding the evolution of bacterial fitness in species subject to environmental variations that occur during persistence within and transmission between hosts.

FIG. 1.

Escape of N. meningitidis strain 8047 from MAb B5-mediated serum bactericidal activity in an interrupted selection assay. Strain 8047 or 8047 ΔmutS was incubated in 1 ml of PBSB containing 0.1% glucose and 5% pooled human serum supplemented with 15 μl of a 1:50 dilution of MAb B5 ascites fluid. Three 180-min passages were performed. The inocula for each passage were prepared following overnight growth of bacteria on BHI medium plates (as indicated at the bottom). The inocula for passages 2 and 3 were prepared from plates inoculated with bacterial cells present at the 80-min time point in the previous passage. The data are the numbers of viable cells present in the inoculum and each passage, which were determined using serial dilutions of appropriate samples, and the frequencies of MAb B5 nonreactive variants (mAb B5-ves), determined by probing colony immunoblots with MAb B5. Dashed lines indicate the frequency of MAb B5 nonreactive variants, and solid lines indicate the size of the total population. Squares, strain 8047; triangles, strain 8047 ΔmutS.

FIG. 2.

Escape of N. meningitidis strain 8047 from MAb B5-mediated serum bactericidal activity in a continuous selection assay. Strain 8047 or the 8047 ΔmutS mutant was incubated in 1 ml of PBSB containing 0.1% glucose and 5% human serum either with or without MAb B5. Two 120-min passages and one 40-min passage were performed. The inoculum for the first passage was prepared following overnight growth of bacteria on BHI medium plates, while for subsequent passages a 500-μl sample from the previous passage was mixed with an equal volume of PBSB containing 1% glucose, 5% human serum, and antibody (as indicated at the bottom). The data are the numbers of viable cells present in the inoculum and each passage, which were determined by plating serial dilutions of appropriate samples. Samples for passages were collected after 40 min of incubation. The frequencies of MAb B5 nonreactive variants (Freq. B5-ves) present in the final passage were determined using colony immunoblots probed with MAb B5 and are indicated on the right. Filled squares, strain 8047 and 1.2 μg of purified MAb B5; open squares, strain 8047 and no MAb B5; filled triangles, strain 8047 ΔmutS and 1.2 μg of purified MAb B5; open triangles, strain 8047 ΔmutS and no MAb B5.

FIG. 3.

Influence of high levels of antibody on escape from MAb B5-mediated serum bactericidal activity. Strain 8047 was passaged in a continuous selection assay as described in the legend to Fig. 2 using various amounts of purified MAb B5. The total number of viable cells was plotted for the inoculum and each passage. The frequencies of MAb B5 nonreactive variants (Freq. B5-ves) present at the end of the assay are indicated on the right. Squares, 0.6 μg of MAb B5; diamonds, 1.5 μg of MAb B5; triangles, 6 μg of MAb B5; circles, 12 μg of MAb B5.

FIG. 4.

Influence of inoculum size on escape from MAb B5-mediated serum bactericidal activity. Assays were performed and results were analyzed as described in the legend to Fig. 2, using 1.2 μg of purified MAb B5 and different inoculum sizes for strain 8047 for the first passage. The data are the total number of viable cells present in the inoculum or after 40 min of incubation for each passage and the frequencies of MAb B5 nonreactive phase variants (mAb B5-ves). The dotted line indicates the upper limit of detection for samples in which no cells were detected. Solid lines and filled symbols, number of CFU/ml; dashed lines and open symbols, frequency of MAb B5 nonreactive phase variants. Squares, inoculum containing 6.6 ×10⁶ CFU; triangles, inoculum containing 6.6 ×10⁵ CFU; circles, inoculum containing 6.6 ×10⁴ CFU; diamonds, inoculum containing 6.6 ×10³ CFU.

FIG. 5.

Competition between strains 8047 and 8047 ΔmutS for escape from MAb B5-mediated serum bactericidal activity in an assay with continuous selection. Assays were performed and results were analyzed as described in the legend to Fig. 2, using either 3 μg or no purified MAb B5 and an inoculum containing 9 ×10⁶ CFU of strain 8047. The inocula for the first passage contained different mixtures of strains 8047 (mutS⁺) and 8047 ΔmutS (ΔmutS). The ratios for subsequent time points were determined using serial dilutions grown overnight on BHI agar plates with and without kanamycin. Open squares, mutS⁺/ΔmutS ratio of 127:1 without MAb B5; open circles, mutS⁺/ΔmutS ratio of 13:1 without MAb B5; filled squares, mutS⁺/ΔmutS ratio of 127:1 with MAb B5; open circles, mutS⁺/ΔmutS ratio of 13:1 with MAb B5.

FIG. 6.

Competition between strains 8047 and 8047 ΔmutS for escape from MAb B5-mediated serum bactericidal activity in an assay with interrupted selection. Assays were performed and results were analyzed as described in the legend to Fig. 1, using 0.6 μg of purified MAb B5 and an inoculum containing 5 ×10⁶ CFU of strain 8047. The inocula for the first passage contained different mixtures of strains 8047 (mutS⁺) and 8047 ΔmutS (ΔmutS). The inocula for passages 2 and 3 were derived from overnight cultures on plates at the 40- and 80-min time points of the previous passage, respectively. Ratios were determined as described in the legend to Fig. 5. Diamonds, mutS⁺/ΔmutS ratio of 18,000:1; triangles, mutS⁺/ΔmutS ratio of 1,800:1; squares, mutS⁺/ΔmutS ratio of 180:1; circles, mutS⁺/ΔmutS ratio of 18:1.

FIG. 7.

Competition between strains 8047 and 8047 ΔmutS for escape from MAb B5-mediated passive protection in an infant rat model of infection. Infant rats were inoculated by the intraperitoneal route in the presence or absence 2 μg of purified MAb B5 with mixtures of strains 8047 (mutS⁺) and 8047 ΔmutS (ΔmutS). Bacteremia was measured at 18 h postinfection, and the “output” ratio of the two strains was assessed by plating serial dilutions of blood on BHI agar plates with and without kanamycin. A CI was derived by dividing the output ratio by the input ratio. CIs for animals inoculated with strain 8047, strain 8047 ΔmutS, and MAb B5 (mutS⁺:ΔmutS Plus mAb B5) were determined following inoculation of a 1:1 mixture of the two strains using a total inoculum containing either 4,200 or 7,200 CFU (four and three animals, respectively [referred to as groups 1 and 2]). CIs for animals inoculated with strains 8047 and 8047 ΔmutS but no MAb B5 (mutS⁺:ΔmutS No mAb B5) were determined for animals inoculated with 5,900 CFU and either a 1:1 or 10:1 mixture of the two strains (five and four animals, respectively [referred to as groups 3 and 4]). A statistical comparison of the CIs of the two groups was performed using a Mann-Whitney rank sum test.

FIG. 8.

Survival in the presence of an immune response to a phase-variable epitope is facilitated by preexposure to low levels of antibody or an increased PV rate. The model shows a bacterial population containing phase variants for a specific surface epitope that are subjected to selection with a bactericidal antibody specific for the phase-variable epitope. The initial population is generated from a single cell with a reactive phenotype (i.e., “on” variants) (open ellipses). The numbers of nonreactive phase variants (i.e., “off” variants) (filled ellipses) in the first population are determined by the PV rate. The PV rate may be determined by the length of the repeat tract in the phase-variable gene or the activity of trans-acting factors. In the case of mononucleotide repeat tracts (such as those present in the meningococcal lgtG gene) this includes components of the mismatch repair system. The number of phase variants in the initial population is low for a nonmutator strain (a and b) and high for a mutator strain (c). If the initial population is subjected to selection with low levels of specific antibody, there is an increase in the proportion of phase variants (b), while in the absence of antibody there is no change in the proportion of variants (a). If each of the populations is then subjected to selection with a high level of antibody, the surviving population is larger for a preexposed population (b) or for a strain with a higher PV rate (c). Ab, antibody.