OUP user menu

PclA, a pneumococcal collagen-like protein with selected strain distribution, contributes to adherence and invasion of host cells

Gavin K. Paterson, Leena Nieminen, Johanna M.C. Jefferies, Tim J. Mitchell
DOI: http://dx.doi.org/10.1111/j.1574-6968.2008.01217.x 170-176 First published online: 1 August 2008

Abstract

Analysis of Streptococcus pneumoniae sequenced genomes revealed a region present only in selected strains consisting of two ORFs: a putative cell wall anchored protein and a putative transcriptional regulator. The cell wall anchored protein contains large regions of collagen-like repeats, the number of which varies between strains. We have therefore named this protein PclA for pneumococcal collagen-like protein A. The second gene, spr1404, encodes a putative transcriptional regulator. We examined the strain distribution of these two genes among a collection of clinical isolates from invasive pneumococcal disease and found them to be present in 39% of the strains examined. Strains were either positive for both genes or lacked both, with the two genes always present together in the same location of the genome. RT-PCR analysis revealed that pclA is transcribed in vitro, even in the absence of spr1404. Single deletion mutants lacking either gene were not attenuated in a mouse model of invasive pneumonia. However, the pclA mutant was defective in adherence and invasion of host cells in vitro.

Keywords
  • Streptococcus pneumoniae
  • collagen-like protein
  • adherence
  • virulence

Introduction

The Gram-positive bacterium Streptococcus pneumoniae (the pneumococcus) is an important human pathogen responsible for considerable disease worldwide. In particular, it is an important cause of pneumonia, meningitis, bacteraemia and otitis media. Drawbacks with the current vaccines and the spread of antibiotic resistance drive the need to better understand pneumococcal–host interactions. An important aspect of these interactions is the high degree of genetic diversity between pneumococcal strains and how this variation influences the behaviour of different strains and serotypes (Hakenbeck et al., 2001; Bruckner et al., 2004; Obert et al., 2006; Silva et al., 2006).

Sortase-processed LPXTG motif proteins are important virulence factors for numerous Gram-positive pathogens, including the pneumococcus (Paterson & Mitchell, 2004; Marraffini et al., 2006). For example, the signature-tagged mutagenesis screen of TIGR4 identified 10 such proteins as virulence factors in a pneumonia model (Hava & Camilli, 2002). A well-characterized example is neuramidase A (NanA), which through desialylation of host molecules is proposed to contribute to adherence (Tong et al., 2001, 2002; King et al., 2006), immune evasion (King et al., 2004, 2006) and nutrient acquisition (Burnaugh et al., 2008). In addition, the desialylation of lipopolysaccharides of other microorganisms by NanA may provide a competitive advantage for the pneumococcus in nasopharyngeal colonization (Shakhnovich et al., 2002). Another notable example is the rlrA islet-encoded pilus, which acts to promote adherence and inflammation (Barocchi et al., 2006; Nelson et al., 2007).

Here we describe initial characterization of a pneumococcal genomic region found only in select strains, encoding an LPXTG motif containing a collagen-like repeat protein and a putative transcriptional regulator. The region's selected strain distribution makes it a potential contributor to the observed differences in the behaviour of strains and serotypes.

Materials and methods

Bacterial strains and growth conditions

The strains used were S. pneumoniae D39, serotype 2 (NCTC 7466), R6 (an unencapsulated derivative of D39) and a panel of clinical isolates from invasive pneumococcal disease in Scotland, deposited at the Scottish Meningococcus and Pneumococcus Reference Laboratory (SMPRL), Stobhill Hospital, Glasgow, UK (Table 1).

View this table:
Table 1

Clinical isolates used in this study and distribution of pclA and spr1404

Strain numberStrain codeSpecimenSerotypeSTpclA and spr1404 genotype
N1603-1138Blood149
P1100-4850Blood149
P3317533Eye149
P5202-1309Blood149
1901-1508Blood6A65+
2100-3521CSF6A65+
4701-4176Blood1370+
11896-3310CSF2472+
596-5878Blood274+
1701-5034Nasal6B96
12602-3345Blood18C113+
1001-1315Blood14124+
4800-4181Blood14124+
5001-2329Blood14124+
2701-1752Blood14156
2900-1139Eye9 V156
3000-1153Tissue9 V156
3100-5049Ear9 V156
N1403-1134Blood3180
P200-1847Blood3180
P2140544Ear3180
P4902-1198Blood3180
6900-2569?19A199+
699-1181Blood12F218
4400-1124Blood12F218
4500-1151Blood12F218
4601-3039Blood12F218
201-4291Blood1227+
401-2696Blood1227+
7500-4294Blood20235
7601-1277Blood20235
7301-4363Ear19F285+
8500-1225Eye23F311+
N4715401Blood23F311+
P3116438Ear23F311+
6200-3175Blood17F392
6300-5123Blood17F392
8001-3137Blood22F433
12902-3522Eye35F448
7400-5238Blood31567
9400-2510CST31568
2301-2255Blood29575
5200-3618Blood15B579+
4301-1203Blood19F586
7101-1875Blood19F587
10301-3102Sputum35593+

Bacteria were grown on blood agar base number 2 (Oxoid Ltd, Basingstoke, UK) with 5% (v/v) defibrinated horse blood (E&O Laboratories, Bonnybridge, UK) or in brain heart infusion broth (Oxoid Ltd). All incubations were static at 37 °C. Clinical isolates were serotyped and analysed by multilocus sequence typing (Enright & Spratt, 1998) at SMPRL.

Transcript analysis by RT-PCR

RNA from mid-log phase cultures was prepared using a commercially available kit (RNeasy® Mini Kit, Qiagen) following the manufacturer's instructions. Reverse transcription reactions were carried out using the ThermoScript RT-PCR System (Invitrogen Life Technologies). Before use, RNA was treated with DNase (RQ1, Promega) according to the manufacturer's specifications. cDNA synthesis using random hexamers was performed using the recommended ThermoScript RT-PCR System protocol. The resulting cDNA was treated with 1 μL of RNase H (E. coli, 2 U μL−1, Invitrogen Life Technologies) at 37 °C for 20 min, and stored at −20 °C until required. pclA was amplified with primer pair spr1403F and spr1403R and product abundance was visualized by ethidium bromide staining after 15, 20, 25 and 30 cycles.

Allelic replacement mutants lacking pclA or spr1404

Single gene deletion mutants in wild-type D39 lacking either pclA or spr1404 were generated by allelic replacement with a spectinomycin resistance cassette. Regions for homologous recombination for deletion of pclA were amplified with primer pairs 43N/43O and 43P/43Q. These were fused by SOEing and cloned into TOPO-pCR4 (Invitrogen Life Technologies), and confirmed by sequencing. The spectinomycin cassette was then introduced between the pneumococcal sequences via the AscI site introduced by the primers, and this construct was used for transformation. spr1404 was deleted with the same strategy using primer pairs 44O/44P and 44Q/44R to amplify the regions for homologous recombination. Mutants were confirmed by PCR. The primers used in this study are shown in Table 2.

View this table:
Table 2

Primers used in this study

Primer nameSequenceDescription/use
spr1403FTCATTCTTAGTTCCGTCTGGGTStrain distribution of pclA
spr1403RGATGGTGCTAAGGGAAAAACTGStrain distribution of pclA
spr1404FACATTTTGGGCTTGAAATGACTStrain distribution of spr1404 gene
spr1404RTTCGGCAAATAAACTTCTTGGTStrain distribution of spr1404 gene
45GCGAGAAGATTTGTCACAACCACAGG3′ location with spr1404F
44MAACAAGCTCAAGACTTGGTCGAAGCG5′ location
44SAAGATGGTAGAGATGGACGTGATGGTCG5′ location
44NAGGTTTCTCATACGGGTATTTCCTCCColocationalization and pclA allelic replacement
43OggcgcgccGTTTAGTAACTTCTTCAGTTTTAACTGGColocationalization and pclA allelic replacement. AscI site in lower case
43PGTTACTAAACggcgcgccTTAAACCTGTGCCAGCGCAACCAACACCpclA allelic replacement. AscI site in lower case
43QCATATAGGCTCAATAGAATACCGCTACCpclA allelic replacement
44OCAATTACGCTTAATGGAAACCCspr1404 allelic replacement
44PggcgcgccGTGACCATTTTTAACCCTGTGGspr1404 allelic replacement. AscI site in lower case
44QGTGACCATTTggcgcgccGAGCATTCCGAATTTCCAGGAGACACTTGCspr1404 allelic replacement. AscI site in lower case
44RAAAGCGGTTATTACAGTAATAGGspr1404 allelic replacement

Experimental murine pneumonia and colonization

Female outbred MF1 mice (Harlan Olac, Bicester, UK) aged 8–10 weeks and weighing 30–35 g were used in all animal work. Before use, a standard inoculum was prepared as described by i.p. injection; 24 h later, blood was recovered into Luria–Bertani medium and bacteria were grown until mid-log (Canvin et al., 1995). For pneumonia infection, the mice were lightly anaesthetized with halothane (Zeneca Pharmaceuticals, Macclesfield, UK) over oxygen using a calibrated vaporizer and 50 μL of bacterial suspension containing the desired dose was administered to the nares. The mice were monitored frequently until they were deemed to have irreversibly succumbed to the infection, after which they were humanely sacrificed. At predetermined points, the bacterial load in the blood was enumerated by viable counting of tail bleeds. For colonization, mice were anaesthetized as above and the bacterial dose was administered in a 10-μL volume. Clinical signs and bacteraemia do not develop in this instance because the smaller inoculum volume restricts bacteria to the upper airways. At the time points indicated, pneumococcus was enumerated in nasopharyngeal washes by plating on blood agar plates supplemented with 4 μg mL-1 gentamicin. All animal experiments were carried out with appropriate licensing and approval from the Home Office and the University of Glasgow.

Adherence to host cells

Adherence to the human nasopharyngeal cell line Detroit 562 and the type II epithelial lung carcinoma cell line A549 was tested essentially as described previously (Kharat & Tomasz, 2003). Briefly, cells were maintained in RPMI 1640 medium without phenol red (Gibco-BRL, Paisley, UK) (RPMI 1640 was supplemented with l-glutamine, 1 mM sodium pyruvate and 10% foetal bovine serum). Frozen bacterial stocks were resuspended in unmodified RPMI 1640 without phenol red, supplemented with 1% foetal bovine serum to give a suspension of 107 CFU mL−1. Monolayers of host cells in 24-well plates were washed with 2 mL phosphate-buffered saline (PBS) and incubated with 1 mL of bacterial suspension for 2 h. The bacterial suspension was decanted and the cells were gently washed with 3 × 1 mL PBS to remove nonadherent bacteria. Human cells were detached by treatment with 200 μL trypsin/EDTA and lysed by the addition of 800 μL ice-cold water. The number of adherent/internalized bacteria was quantified immediately by viable counts of the cell lysates. To determine the number of internalized bacteria, following the 2 h incubation, the medium was replaced with 1 mL of fresh medium containing penicillin (10 μg mL-1) and gentamicin (200 μg mL-1). The cells were incubated for a further 1 h and then sampled as above.

All incubations were at 37 °C in 5% CO2 and 95% air. All bacterial strains grew equally well in the tissue culture medium. Assays were performed in triplicate. Data presented are representative of two experiments giving similar results.

Statistical analysis

Data were analysed by Student's t-test using graphpad prism version 4, with P<0.05 considered significant.

Results and discussion

Strain distribution of pclA and spr1404

Comparison of the R6 and TIGR4 sequenced genomes revealed a region of c. 9.6 kb in R6 that was absent in TIGR4 and referred to as R6-specific cluster 1 by Bruckner et al. (2004). This region encodes two ORFs, spr1403 and spr1404 (Fig. 1). The upstream and downstream genes encoding a hypothetical (Spr1402) and conserved hypothetical protein (Spr1405), respectively, are conserved in both strains. The reason for this insertion/deletion is not obvious as no insertion sequences or phage elements are present adjacent to the region.

Figure 1

PCR mapping primers of the pclAspr1404 region. Primer direction and approximate location are indicated. Not to scale.

spr1403 is predicted to encode a 265 kDa sortase-anchored cell wall protein with several regions of G–X–Y collagen-like repeats. The largest of these contains 103 repeats of this tripartite sequence. Other regions contain 37, 16, 15 and 10 repeats with numerous other repeats varying in length from two to eight G–X–Y repeats. Similar collagen-like repeats are seen in the streptococcal collagen-like proteins Scl1 and 2, and hence we refer to Spr1403 as PclA for pneumococcal collagen-like protein A. The genome sequence of the 23F strain Spanish 23F-1 shows that it too encodes pclA but with reduced numbers of collagen-like repeats (http://www.sanger.ac.uk/Projects/S_pneumoniae/23F/). Spr1404 encodes a putative transcriptional regulator with homology (46% similarity) to Mga of Streptococcus pyogenes. It shows 61.1% similarity to the pneumococcal protein MgrA, a previously described pneumococcal orthologue of Mga found in both TIGR4 and R6 (Hemsley et al., 2003).

The absence of the pclAspr1404 region in TIGR4 prompted the examination of its distribution in a panel of clinical isolates from invasive disease. Strains were chosen to represent a broad range of serotypes (22 serotypes) and multilocus sequence types (26 STs).

Eighteen out of the 46 strains (39.1%) examined by PCR were positive for pclA and spr1404 (Table 1). The strains were either positive for both or negative for both genes. The ST distribution of these genes showed that they were present in diverse STs and did not cluster with related clones (Fig. 2). Likewise, when a similar phylogenetic tree is drawn based on capsular locus sequence, positive strains do not cluster to related serotypes (data not shown). However, too few strains are included here to make conclusions on the association of this region with particular serotypes and STs.

Figure 2

Distribution of pclA and spr1404 by pneumococcal ST. Alignment was carried out using clustalw algorithm in megalign. STs positive for pclA and spr1404 are underlined.

PCR mapping of the region in positive strains (Fig. 1), showed that it was always present in the same location of the genome (primer pair 44M and 44S and primer pair spr1404F and 45G) relative to the flanking genes spr1402 and spr1405. The two genes pclA and spr1404 were always localized together (primer pair 43O and 44N). PCR between the conserved flanking genes (primer pair 44M and 45G) yielded the expected PCR product in the negative strain TIGR4 of c. 350 bp with a similar-sized product in all the other pclAspr1404 negative strains. This confirmed that they had a genomic organization similar to TIGR4 in this region and were indeed negative, as opposed to the false positive results caused by sequence polymorphisms in the primer-binding sites or poor template quality.

Role in virulence and nasopharyngeal colonization

To test the role in virulence of these genes, allelic replacement mutants lacking either pclA or spr1404 were constructed in the virulent parent of strain R6, D39. Following intranasal infection of mice with doses of 105, 106 and 107 CFU, no significant difference in host survival rates or time was seen between infections with wild-type D39 vs. the two mutant strains (Table 3). Likewise, no significant difference between the strains was seen when bacterial blood counts were enumerated at 24 and 36 h postinfection with any of the three doses tested (Table 3).

View this table:
Table 3

Virulence characteristics of D39 and its pclA and spr1404 mutants following intranasal infection of MF-1 mice as described in Materials and methods. n=5

DoseStrainPercentage survival24 h blood count (log10 CFU mL−1 mean ± SE)36 h count (log10 CFU mL−1 mean ± SE)
105 CFUD39403.10 ± 1.195.23 ± 1.52
pclA603.68 ± 1.114.36 ± 1.44
spr1404404.60 ± 1.145.37 ± 1.58
106 CFUD39205.73 ± 1.096.20 ± 1.37
pclA05.58 ± 0.567.53 ± 0.70
spr140405.59 ± 1.065.92 ± 0.72
107 CFUD3906.01 ± 1.077.14 ± 1.20
pclA06.39 ± 0.497.68 ± 0.64
spr140405.56 ± 0.9478.08 ± 0.68

Together, these data show that neither pclA nor spr1404 plays an essential role in pneumococcal virulence in this mouse model.

The two mutants were also assessed in a mouse model of nasopharyngeal colonization. Bacterial counts in nasopharyngeal washes were enumerated at 2, 4, 8 and 12 days post infection, n=3–5. No significant difference was seen between wild-type D39 and the mutants (data not shown).

Transcription of pclA

The distribution linkage of pclA with a putative transcriptional regulator suggested that the two might interact, with the regulator Spr1404 involved in pclA transcription. Semi-quantitative RT-PCR showed that pclA was transcribed in vitro and that deletion of spr1404 had no significant effect on pclA transcript levels (data not shown). It is possible that Spr1404 has subtle effects on pclA transcription, undetected by the assay, or that Spr1404 contributes in other environments. Nonetheless, at the very least, it is clear that Spr1404 is not required for pclA transcription under the in vitro conditions used here.

Adherence to host cells

Deletion of the collagen-like repeat protein gene scl1 in S. pyogenes reduces bacterial binding to host cells (Lukomski et al., 2000). To test whether pclA also contributes to adherence, the adherence and invasion of wild-type and pclA mutant D39 to Detroit 562 and A549 cell lines were tested. Deletion of pclA significantly reduced binding and invasion to both cell lines (Fig. 3). The receptor used by PclA is unknown as yet and merits further investigation. The role of pneumococcal adhensins is often masked by the polysaccharide capsule in vitro assays (King et al., 2006; Paterson & Mitchell, 2006). Therefore, to further explore the role of pclA in pneumococcal adherence and invasion, we tested the effect of its deletion in the unencapsulated strain R6, a derivative of D39 (Fig. 3). As expected for an unencapsulated strain, R6 showed enhanced adherence and invasion compared with D39 with both cell lines. Interestingly, deletion of pclA had no effect on the adherence and invasion properties of R6 on either cell line (Fig. 3). It is possible that the enhanced influence of other adhensins in the unencapsulated background masks and compensates for the deletion of pclA. The influence of PclA, unlike other adhesins, may be apparent in the presence of a capsule, because its larger size (265 kDa) allows it to protrude and function far from the surface beyond the capsule.

Figure 3

pclA contributes to pneumococcal adherence and invasion to Detroit 562 cells and A549 cell in D39 but not R6. D39 wild-type and pclA adherence (a) and invasion (b) of Detroit 562 and A549 cells. R6 wild-type and pclA adherence (c) and invasion (d). Adherence and invasion are expressed as a percentage of inoculum. Mean ± SEM, n=3. *P<0.05 compared with wild-type.

Conclusion

We describe here a variable collagen-like repeat protein in the pneumococcus, referred to as PclA. It is found in a subset of diverse strains linked to a putative transcriptional regulator. PclA is involved in the pneumococcal adherence and invasion of host cells. However, neither gene contributed significantly to virulence or nasopharyngeal colonization in a mouse model and work is required to elucidate further the function(s) and importance of this region in pneumococcal–host interactions. The selected strain distribution of this region indicates that it may contribute to the phenotypic variation seen between pneumococcal strains.

Acknowledgements

This work was supported by a Wellcome Trust project grant to T.J.M.

Footnotes

  • Editor: Ross Fitzgerald

References

View Abstract