Emerging, Non-PCV13 Serotypes 11A and 35B of Streptococcus pneumoniae Show High Potential for Biofilm Formation In Vitro

Background Since the use of pneumococcal conjugate vaccines PCV7 and PCV13 in children became widespread, invasive pneumococcal disease (IPD) has dramatically decreased. Nevertheless, there has been a rise in incidence of Streptococcus pneumoniae non-vaccine serotypes (NVT) colonising the human nasopharynx. Nasopharyngeal colonisation, an essential step in the development of S. pneumoniae-induced IPD, is associated with biofilm formation. Although the capsule is the main pneumococcal virulence factor, the formation of pneumococcal biofilms might, in fact, be limited by the presence of capsular polysaccharide (CPS). Methodology/Principal Findings We used clinical isolates of 16 emerging, non-PCV13 serotypes as well as isogenic transformants of the same serotypes. The biofilm formation capacity of isogenic transformants expressing CPSs from NVT was evaluated in vitro to ascertain whether this trait can be used to predict the emergence of NVT. Fourteen out of 16 NVT analysed were not good biofilm formers, presumably because of the presence of CPS. In contrast, serotypes 11A and 35B formed ≥45% of the biofilm produced by the non-encapsulated M11 strain. Conclusions/Significance This study suggest that emerging, NVT serotypes 11A and 35B deserve a close surveillance.


Introduction
In the present study, 16 isogenic pneumococcal transformants expressing CPSs of non-PCV13 serotypes were constructed to determine whether biofilm formation can reliably predict the emergence of NVT of S. pneumoniae that colonise the human nasopharynx.

Materials and Methods
Media, growth conditions, DNA purification, and genetic transformation Table 1 shows the pneumococcal strains examined. Clinical isolates of S. pneumoniae were from the collection of the Spanish Pneumococcal Reference Laboratory (SPRL; Centro Nacional de Microbiología ISCIII, Majadahonda, Madrid, Spain). All strains were grown in liquid CpH8 medium supplemented (or not) with 0.08% yeast extract (C+Y medium), or in solid medium involving trypticase soy agar or D agar supplemented with 5% defibrinated sheep blood (Thermo Scientific; Hampshire, England) [11]. Growth was controlled by measuring the optical density at 550 nm (OD 550 ).
S. pneumoniae chromosomal DNA was purified from clinical strains and encapsulated, isogenic transformants constructed by transformation of strain M11, as described elsewhere [9]. The selection of transformants was made by enriching those encapsulated by successive transfer to C+A medium (CpH8 containing 0.08% of bovine seroalbumin) supplemented with 0.5 μl/ml anti-R antiserum (which agglutinates only non-encapsulated pneumococci) [9]. Serotyping of S. pneumoniae was kindly performed by D. Vicioso at the SPRL.

Biofilm formation assay, quantification, and statistical analysis
The optimal conditions for biofilm formation by pneumococcal cells on polystyrene microtitre plates have been previously described [11,14]. In short, pneumococcal cells were grown in C+Y medium at 37°C to an OD 550 of 0.5-0.6 and diluted (1:100) in fresh C+Y medium. Aliquots (200 μl) were dispensed into each well in triplicate, and the plate incubated at 34°C for 6 h. Growth (OD 595 ) was measured and the biofilm formed was stained with 1% crystal violet [11]. In the present study, a strain was considered as "good" or "intermediate" biofilm producer when it formed 45% or between 10% and 30% of the biofilm formed by the control, non-encapsulated M11 strain respectively. The data for biofilm formation include the mean ± standard error of at least three independent experiments, each performed in triplicate. Statistical significance was examined using the Student t test. For multiples comparisons, one-way analyses of variance (ANOVA) were performed, followed by Tukey's post hoc test when the ANOVA rejected the null hypothesis. The SAS 9.3 statistical packge (SAS Institute, Cary, NC) was used for all analyses. Differences were considered statistically significant when P <0.05.  (Table 1). Two clinical isolates of each serotype were analysed in detail.

Results
Although pneumococci of every serotype tested produced less biofilm than the non-encapsulated strain M11, significant differences were noted in the biofilm formation capacity of several pairs of clinical strains of the same serotype: e.g., strains 2901 and 2963 (serotype 11A), 2948 and 2990 (serotype 15B/C), and 1544 and 3004 (serotype 23A). This confirms previous observations indicating that the genetic background, and not only the CPS, modulates pneumococcal attachment to the artificial substrate [9,11,15].
Since in vitro biofilm formation by pneumococcus depends on many different genes and/or biochemical traits, the use of isogenic transformants (i.e., strains differing only by a single trait) is necessary when the influence of CPS in biofilm formation is to be measured. To this end, the non-encapsulated M11 strain was transformed with DNA prepared from the corresponding clinical isolates, and at least two independent transformants of each serotype were analysed.
Since initial experiments revealed no significant differences between them, only one transformed strain of each serotype was studied further. The serotype 35B strains were the best biofilm producers; the P241 transformant strain produced 50% of the biofilm formed by M11.
In addition, the P242 transformant (serotype 11A) was capable of synthesizing up to 45% of the biofilm produced by its non-encapsulated progenitor. Intermediate biofilm formation was noted for the isogenic S. pneumoniae transformants of serotype 23A. In contrast, serotype 33F pneumococci (whether clinical isolates or M11 transformant) were unable to produce any substantial quantity of biofilm in vitro (Fig 1). Some variation was noted in the growth rate and/or biofilm formation capacity of the different strains (Fig 1). Consequently, the biofilm formation values for the encapsulated strains were normalised for the OD of the culture measured as stated in Materials and Methods, and the presented percentages were calculated in relation to the parental strain M11 [11]. Fig 2 shows the relative biofilm formation capacity of the isogenic transformants of the non-PCV13 serotypes compared with that of M11. Every serotype analysed formed significantly less biofilm than the non-encapsulated strain, although serotypes 35B and 11A formed significantly more biofilm than the other serotypes analysed.

Discussion
Recent years have seen continuous interest in the impact of pneumococcal conjugate vaccines on nasopharyngeal colonisation [16]. In fact, surveillance of colonisation has become an important component of the vaccination monitoring process in the post-licensure setting. Nasopharyngeal colonisation, which is associated with biofilm formation, is an essential step in the development of S. pneumoniae disease and prevention of colonisation may reduce host-to-host transmission [17,18]. Moreover, colonisation by multiple serotypes (co-colonisation), which appears to be much more prevalent than previously envisaged [19], is an important factor to consider as it facilitates horizontal gene transfer [20]. Thus, the ability of pneumococcal serotypes to form biofilms-which a previous publication [9] and the present study show can be rapidly tested in vitro-could be used to help predict the expansion of NVT. This method has recently shown isogenic transformants and clinical isolates of serotype 19A and serogroup 6, which were emerging during the PCV7 era, to be good biofilm-formers [9], and many studies have shown that the non-PCV13 serotypes analysed in the present study underwent a rapid increase in prevalence after PCV13 implementation. Nevertheless, no previous studies have assessed the ability of these serotypes to form biofilms in vitro.
The present results clearly show that most (14 out of 16) of the clinical isolates or isogenic transformants expressing the CPSs of NVT were unable to form substantial amounts of biofilm in vitro. Although an inverse relationship between CPS and biofilm formation exists [14], it has previously been shown that a minimum amount of CPS is necessary for efficient nasopharyngeal colonisation in mice [15]. The situation, however, is far from clear. For example, S. pneumoniae isolates of serotypes 8, 22F and 33F, which were found both in carriage and causing IPD, and which are potentially highly invasive [21], failed to produce substantial amounts of biofilms in vitro in the present study. This finding strongly suggests that in vitro biofilm formation is mostly unrelated to IPD-inducing capacity. It should be underlined that the CPSs of serotypes 22F and 33F are included in a 15-valent pneumococcal conjugate vaccine that is currently under pre-clinical evaluation [22]. The present results also show that the biofilm formation capacity of the corresponding isogenic transformants almost paralleled (and in some cases exceeded) that of the corresponding clinical isolates. One exception was seen, however: the clinical isolate 2948 (serotype 15B/C) (Fig 1). This strain showed significantly more biofilm formation capacity than strain P013, an M11 transformant of serotype 15B obtained using DNA from the reference strain SSISP15B/1 ( Table 1). It may be that strain 2948 has a genetic background more prone to allow biofilm formation than strains M11 and 2990. It should be reminded that the serotypes of strains 2948 and 2990 were not determined; they might not, therefore, belong to the same serotype (Table 1, see footnote c).
In sharp contrast to the other serotypes tested here, isogenic transformants expressing the CPS of serotype 35B or 11A were capable of producing substantial amounts of biofilm in vitro, forming 45% of the amount of biofilm produced by the non-encapsulated M11 strain (Fig 2). This may explain the current prevalence of these serotypes in the human nasopharynx [23,24].
Many factors appear to influence the colonisation capacity of pneumococcal serotypes. For example, a previous study showed CPS (the major determinant of surface charge in S. pneumoniae) to be associated with colonisation capacity [25]. With the notable exceptions of serotypes 19A and 11A, higher net negative surface charge was associated with higher resistance to nonopsonic, neutrophil-mediated killing as well as higher carriage prevalence. It has also been suggested that a direct relationship exists between the success (in terms of relative numbers) of a serotype during carriage and the biochemistry of its CPS [26]. In fact, it has been recently shown that the CPSs of clinical isolates and isogenic transformants expressing the CPS of serotypes 19F and 19A, and all members of serogroup 6 (which are good biofilm producers), all include the disaccharides α-D-Glcp-(1!2)-α-L-Rhap-(1! and α-D-Glcp-(1!3)-α-L-Rhap-(1! [9]. However, neither of these disaccharides is present in the capsules of serotypes 11A or Relative biofilm formation capacity of M11 isogenic transformants with non-PCV13 serotypes compared to that of their parental non-encapsulated strain (M11). The percentages shown are the mean ± standard error of at least three independent experiments, each performed in triplicate. Relative biofilm formation was significantly different between serotypes (*, P <0.0001 in a one-way ANOVA). The Tukey's post hoc test showed significant differences in biofilm formation between serotypes 35B and 11A, and the rest of serotypes. No significant differences were observed between the biofilm-forming capacity of serotypes 35B and 11A.
doi:10.1371/journal.pone.0125636.g002 35B (Fig 3). In addition, no chemical similarities between serotypes 11A and 35B CPSs are obvious-although both contain O-acetylated galactose residues. Recently, Calix et al. [27] showed that the newly discovered, invasive serotype 11E contains loss-of-function mutations in the capsule O-acetyltransferase gene wcjE, and does not express β-D-Galp6Ac at all. It is remarkably that, although frequently found in carriers, serotype 11A shows low invasiveness due to ficolin-2 recognition of O-acetylated capsule epitopes and the consequent activation of the lectin complement pathway activation [28]. It it well known that some serotypes are rarely found in carriage although they are known to cause disease, e.g., serotypes/serogroups 1, 3, 5, and 7 [29], but the reasons underlying how and when pneumococcal colonisation will result in dissemination and disease remain poorly understood [16,[30][31][32].
There is increasing interest in the interactions between the different species of the nasopharyngeal microbiota, because of the possibility that PCV vaccination might affect the ecology of the nasopharynx, which could have clinical consequences. This topic is subjected to debate; for example, an inverse relation between carriage of pneumococcus and Staphylococcus aureus has been found in some studies [33,34], but not in others [35][36][37].
Evaluating vaccine efficacy for protection against colonisation with S. pneumoniae and other bacterial pathogens is an area of growing interest [38]. Conjugate vaccines lead to a reduction in the carriage prevalence of vaccine-serotype pneumococci in both vaccinated and unvaccinated individuals and a reduction in hospital admissions for invasive and non-invasive pneumococcal pneumonia in children younger than 5 years, as well as in some adult age groups, indicating herd protection [39]. Interestingly, it has been shown that estimating the adult burden of pneumococcal disease from bacteraemic pneumococcal pneumonia data alone significantly underestimates the true burden of disease in adults. For every case of bacteraemic pneumococcal pneumonia, it has been estimated that there are at least 3 additional cases of non-bacteraemic pneumococcal pneumonia [40]. Management of these infections is potentially being compromised by the increasing resistance of the pathogen to antibiotics commonly used to treat these infections [41]. Of note, increasing antibiotic resistance has been recently detected both in serotype 35B [42,43] and serotype 11A S. pneumoniae isolates [44]. However, since pneumococcal vaccines have been dessigned for the active prevention of IPD and pneumococci of serotypes 11A and 35B only seldom cause IPD [45][46][47], it is questionable whether these serotypes should be included in future higher-valency PCVs. Yet, further studies of surveillance are needed in order to detect the emergence of non PCV-13 serotypes and allow rational vaccine design, implementation and continued effective control of pneumococcal disease.