Dynamic capsule restructuring by the main pneumococcal autolysin LytA in response to the epithelium

Abstract

Bacterial pathogens produce complex carbohydrate capsules to protect against bactericidal immune molecules. Paradoxically, the pneumococcal capsule sensitizes the bacterium to antimicrobial peptides found on epithelial surfaces. Here we show that upon interaction with antimicrobial peptides, encapsulated pneumococci survive by removing capsule from the cell surface within minutes in a process dependent on the suicidal amidase autolysin LytA. In contrast to classical bacterial autolysis, during capsule shedding, LytA promotes bacterial survival and is dispersed circumferentially around the cell. However, both autolysis and capsule shedding depend on the cell wall hydrolytic activity of LytA. Capsule shedding drastically increases invasion of epithelial cells and is the main pathway by which pneumococci reduce surface bound capsule during early acute lung infection of mice. The previously unrecognized role of LytA in removing capsule to combat antimicrobial peptides may explain why nearly all clinical isolates of pneumococci conserve this enzyme despite the lethal selective pressure of antibiotics.

Figure 1. Effects of LL-37 on viability and capsule shedding.

(a) Representative growth curve of WT strain TIGR4 (left) or a TIGR4-derived mutant in the capsule synthesis genes cps4E-F (right) grown in C+Y. LL-37 was added at t=0 at 16 μg ml⁻¹ (triangles), 32 μg ml⁻¹ (squares) or was omitted from cultures (circles), and growth at 37 °C was monitored by optical density. (b) WT strains TIGR4 or D39 were treated with LL-37 at concentrations ranging from 0.25 to 4 μg ml⁻¹ in a standard capsule shedding assay (see Methods). The resulting supernatants were then processed and analysed via capsule blot (see Methods). n/a: no addition; std: purified capsule. Experiments were repeated at least three times.

Figure 2. Role of LytA autolysin in capsule shedding.

(a) WT TIGR4 or the isogenic deletion mutant in the lytA gene (LytA⁻) were treated with 4 μg ml⁻¹ LL-37 in a capsule-shedding assay for 30 or 60 min as indicated. The supernatants were then analysed by capsule blot. (b) TIGR4 (WT) or the mutant in lytA (LytA⁻) were treated at t=0 with LL-37 at 16 μg ml⁻¹ (squares), 8 μg ml⁻¹ (triangles), 4 μg ml⁻¹ (inverted triangles) or untreated (circles). Cultures were then monitored by optical density and samples collected for enumeration of viable CFUs at the indicated times. (c) TIGR4 (WT) was grown in C+Y to OD 0.4. Cultures were then back diluted to OD 0.2 (1 mass doubling) or OD 0.05 (three mass doublings) and LL-37 was added at the indicated concentrations. At OD 0.4 the cultures were harvested and pellet and supernatant fractions were analysed by capsule blot. Relative capsule amounts were calculated by comparison with a purified capsule standard and were normalized to untreated cells for each fraction. The results are the mean and s.d. from three independent experiments (100% pellet: 55.9 μg ml⁻¹±32.1 n=3; supernatant 2.68 μg ml⁻¹±0.827, n=3). (d) The LytA⁻ strain was grown as in c and treated with 4 μg ml⁻¹ LL-37 where indicated. For reference, the results for WT in the same conditions in c were compared and plotted against the LytA⁻ results. *P<0.05 **P<0.01 ***P=0.001 unpaired t-test with Welch's correction.

Figure 3. Amidase activity of LytA is required for capsule shedding.

(a) TIGR4 (WT) with empty vector, the lytA mutant (LytA⁻) with empty vector or vector containing a WT lytA allele (pLytA) or an allele with mutations in active-site residues (pLytA⁻) were subjected to capsule shedding assay. The supernatant fractions were then analysed by capsule blot. (b) Cultures from a were harvested, mechanically lysed, and lysates, normalized for total protein, were analysed by western blot using LytA antiserum (see full blot in Supplementary Fig. 6a). (c) Representative growth curve of strains from a grown in C+Y. At the arrow LL-37 was added to indicated cultures (open symbols) at 8 μg ml⁻¹. (d) The lytA mutant was subjected to a capsule-shedding assay, and recombinant LytA protein was added at the indicated concentration to the assay. LL-37 was added to the indicated cultures at 4 μg ml⁻¹. The supernatant fraction was then analysed by capsule blot.

Figure 4. Distinction between LytA activation during capsule shedding versus by antibiotics.

(a) Schematic representation of localization of cell wall synthesis and autolysin (red=capsule, green=old cell wall, black=new cell wall containing choline, white=new cell wall containing ethanolamine, LytA=black diamonds). (b) TIGR4 was grown as described for ethanolamine pulse experiments (Methods). Penicillin was added as indicated at 40 ng ml⁻¹. For an autolytic negative reference the mutant in lytA (LytA⁻) grown in choline was included. (c) Cultures from b after ethanolamine or choline pulse were harvested and used in a standard capsule-shedding assay and the supernatant fractions were analysed by capsule blot. Std: purified capsule polysaccharides. (d) TIGR4 was subjected to a capsule-shedding assay with either LL-37 at 4 μg ml⁻¹, pencillin (10–80 ng ml⁻¹) or vancomycin (100–800 ng ml⁻¹). Supernatants from these assays were then analysed by capsule blot. Std: purified capsule polysaccharides. (e) At the arrows, cultures of TIGR4 were exposed to either penicillin (Pen) or vancomycin (Vanc) at the indicated concentrations and growth was monitored by optical density.

Figure 5. Capsule shedding as a function of antimicrobial peptides or capsule serotype.

(a) TIGR4 was treated with LL-37 (4 μg ml⁻¹), human β-defensin 3 (hβD-3) at the indicated concentration, nisin (64 μg ml⁻¹) or polymyxin B (32 μg ml⁻¹). The supernatants from these assays were then analysed by capsule blot. n/a: no addition. (b) Serotypes 6A and 3 strains were subjected to the shedding assay with LL-37 (4 μg ml⁻¹) and the supernatants were analysed by capsule blot.

Figure 6. Analysis of capsule shedding in vivo.

(a) Schematic representation of fluorescent microscopic measurement of encapsulation. Briefly, the diameter of the bacterial membrane (stained with Nile Red) is subtracted from the diameter of the capsule (green). The resulting distance (green arrows) is directly proportional to the amount of surface bound capsule. Scale bar, 1 μm. (b) Encapsulation of TIGR4 (WT) and the lytA mutant (LytA⁻) was measured on bacteria recovered from BAL fluid and compared with encapsulation in C+Y (medium) or after one mass doubling in 8 μg ml⁻¹ LL-37 (WT+LL-37). Either Balb/c mice (WT, n=3 per experiment), or the cathelicidin-deficient mutant (mCRAMP⁻, n=3) and its WT parental control (Bl/6, n=3) mouse strains were infected intratracheally. After 3 h, BALs were performed and samples were subjected to analysis by fluorescent microscopy. The results are expressed as per cent encapsulation of the TIGR4 strain in medium (mean+s.e.m. from three independent experiments). *P>0.001; **P=0.0278. Unpaired t-test with Welch's correction. (c) WT (TIGR4) pneumococci grown in with or without 8 μg ml⁻¹ LL-37 for one mass doubling (LL-37) were incubated with A549 epithelial cells and assayed for either adherence (adh) or invasion (inv) (see Methods). The results are mean+s.e.m. from three independent experiments. *P=0.0307. unpaired t-test with Welch's correction. (d) TIGR4 was utilized in a capsule shedding assay with 4 μg ml⁻¹ LL-37. FBS added at the same time as LL-37 at the indicated concentration. The supernatant fractions were then analysed by capsule blot.