Chandler et al. 10.1073/pnas.0505545102. |
Fig. 5. Affinity chromatography of plasma using an iCF10 matrix indicates an interaction between iCF10 and albumin. The iCF10 peptide was coupled with the carboxyl terminus to Sepharose 6B. The matrix material was incubated with human plasma and eluted with acetonitrile (see Supporting Text, main text; and Fig. 3B). Eluate was separated on an SDS/7.5% PAGE gel and subjected to Western blot analysis with anti-human albumin antibodies. Lane 1, human plasma 1:1,000; lane 2, acetonitrile eluate from the affinity column; lane 3, purified (commercially obtained) human serum albumin (100 ng).
Fig. 6. Induction of Asc10 by OBSA complexes and Western blot of cell extracts of OG1RF(pCF10) cells with anti-AS antibody. Lane 1, induced with cCF10 (10 ng/ml); lane 2, uninduced; lane 3, OBSA 5 mg/ml; lane 4, OBSA 1 mg/ml; lane 5, OBSA 500 mg/ml; lane 6, OBSA 100 mg/ml; lane 7, OBSA 50 mg/ml; lane 8, OBSA 10 mg/ml; lane 9, OBSA 1 mg/ml. The same concentration of total protein was loaded in every lane.
Fig. 7. Shift of the peptide balance in favor of cCF10 and whole-cell ELISA with antibody against AS. All cultures were contained in 1 ng/ml exogenously added synthetic cCF10 and synthetic iCF10 at either 20:1 (1 and 2) or 40:1 (3 and 4) molar ratio relative to the cCF10 and were inoculated with the indicator strain OG1RF(pCF10). In assays 2 and 4, 100 ng/ml OBSA complexes were also added. Induced Asc10 expression after 2.5 hr of growth was detected by the antibody (see Materials and Methods in Supporting Text).
Supporting Text
Hirt et al. (1) demonstrated induction of expression of pCF10-encoded conjugation proteins, including Asc10, by growth of the bacteria in rabbits or in human or rabbit plasma. Incubation of plasma with various mixtures of iCF10 and cCF10 caused a shift in the biological activities of the mixtures such that cCF10 activity was increased. In conjunction with the data presented in the main paper of the present study, these results suggest the possibility that one or more plasma components can degrade or sequester iCF10 more efficiently than cCF10, and that this component is albumin. This interaction could cause Enterococcus faecalis strains carrying cCF10 to self-induce when growing in the bloodstream of a mammalian host. This supporting information describes initial attempts to characterize the interaction of albumin from human plasma with the iCF10 and cCF10 peptides. We show that purified albumin/lipid complexes had effects on the activities of iCF10/cCF10 mixtures similar to those observed with whole plasma. These data are consistent with the possibility that in vivo induction may result from an interaction between iCF10 and albumin complexes that reduce iCF10 activity and allow endogenous cCF10 to escape iCF10 neutralization, leading to self-induction.
Materials and Methods
SDS/PAGE, Silver Stain, Cell Extracts, and Western Blotting.
SDS/PAGE and Western blot analysis were performed as described in ref. 2 or with use of the ECL system (Pierce) according to manufacturer recommendations. Cells grown in 5-ml cultures in THB (Difco) were treated with 50 ml of extraction buffer as described (3) for 1 h at room temperature on a shaker. The antibody QSD 3.7, which was raised against whole aggregation substance Asa1, was kindly provided by Albrecht Muscholl (Universität Regensburg, Regensburg, Germany).Heat, Lipase, and Protease Treatment.
A 500-ml lyophilized aliquot of isolated plasma albumin was resuspended in 20 ml of H2O. Heat inactivation was performed by 30-min boiling in a water bath. One microliter of lipase (60 units/ml, type VII from Candida rugosa, Sigma) or 1 ml of protease K (20 mg/ml, Sigma) was added for lipase or protease treatment, respectively. Incubation was performed at 37°C for 30 min. Five microliters was used untreated as positive control and was incubated at 37°C.Whole-Cell ELISA.
A whole-cell ELISA was performed as previously described (2) with a few modifications. In brief, bacteria (OG1RF:pCF10) were inoculated into microtiter plates (type 3590, Costar) with the appropriate peptide mix with or without albumin. The plates were shaken at 37°C for 2.5 hr, then 50 ml of 0.25 M EDTA was added to each well to dissolve eventual clumps resulting from cell aggregation. The plates were then dried overnight at 60°C. After extensive washing with PBS-T (PBS with 0.5% Tween 20), the plates were blocked with PBS-T containing 0.5% albumin for 30 min at room temperature, washed, and incubated with the antibody specific for Asc10 at a dilution of 1:1,000 for 2 hr at room temperature. After washing, the plate was incubated with 1:3,000-diluted anti-mouse periodated conjugate (Jackson ImmunoResearch) for 1 hr at 37°C. Color detection was performed with the substrate o-phenylenediamine (Zymed Laboratories), and the absorbance at 490 nm was determined.Results
As presented in the main paper (Fig. 3B), a protein identical in size to serum albumin was specifically retained on an iCF10 affinity column. To confirm identity of this protein, the fractionated material was subjected to Western blot analysis and shown to be albumin (Fig. 5). To further investigate the properties of the inducing species, the albumin from the iCF10 affinity column was subjected to heat, proteinase, and lipase treatment. As expected, the inducing activity of this material was sensitive to heat and protease treatment. Lipase treatment also significantly reduced the inducing activity, suggesting that an additional lipid component was bound to the albumin retained on the column.
A key function of albumin is the transport of long-chain fatty acids, and the albumin fraction of plasma contains complexes of this protein in association with various other molecules (4), complicating the determination of the precise identity of the inducing species. To provide more direct support that specific albumin complexes could be inducers, commercially obtained oleic acid/BSA (OBSA) and linoleic acid/BSA (LBSA) were added to cultures of OG1RF(pCF10) growing in THB (no exogenously added iCF10 or cCF10). Western blot analysis of cultures exposed to a variety of concentrations of OBSA showed induction of Asc10 expression. An optimal concentration of OBSA for Asc10 expression was observed (100 mg/ml OBSA; Fig. 6, lane 6); higher or lower concentrations were not as efficient. LBSA showed similar activity. We speculate that at the highest concentrations of the albumin complexes, all of the iCF10 and cCF10 produced by the bacteria was bound and therefore no induction was observed. As the concentration of the complexes was reduced to a range of 500 mg/ml to 10 mg/ml, iCF10 was preferentially bound, shifting the balance of free peptides toward cCF10. The lowest concentrations did not have sufficient iCF10-binding activity to affect the peptide balance. No inhibition of growth was observed even with the highest concentrations of the complexes.
In the experiment described in the previous paragraph, the albumin complexes were simply incubated with pCF10-containing bacteria growing in liquid medium; no plasma or peptides were added exogenously, so any pheromone or inhibitor activity present was synthesized by the organisms in the culture. As previously shown (1), plasma also influences the balance of cCF10 and iCF10 activities in solutions containing mixtures of the two peptides. We investigated whether or not purified fatty acid/albumin complexes could also shift this balance in favor of cCF10. In these experiments the pheromone response was measured with a whole-cell ELISA for Asc10 expression by donor cells carrying pCF10. These cells were exposed to various combinations of peptides and albumin complexes. In preliminary experiments involving induction (in the absence of plasma or albumin) using various ratios of exogenously-added iCF10 and cCF10, we found that molar ratios of iCF10 and cCF10 >5-fold decrease the level of Asc10 induction relative to that observed in a parallel culture exposed to only cCF10, with the level of induction inversely correlated with the iCF10/cCF10 ratio. At iCF10/cCF10 ratios of >30:1, the iCF10 completely inhibited induction. An experiment illustrating the effect of albumin/oleic acid complexes on the inducing activities of iCF10/cCF10 mixtures is shown in Fig. 7. At a 20:1 molar ratio, the expression of Asc10 was quite high whether or not the complexes were present (lanes 1 and 2), whereas at a 40:1 excess of iCF10 expression is inhibited in the absence of albumin complexes, and the addition of 100 ng/ml complexes to this peptide mixture had a substantial effect on reversing the inhibitory activity of iCF10.
Conclusions
These experiments do not prove that the observed in vitro interactions between iCF10 and albumin complexes is responsible for in vivo induction, nor do they reveal whether a specific subset of albumin complexes is directly involved. However, they do suggest a plausible model for induction (Fig. 1B) that is entirely consistent with all previous and current results and also provides an explanation for the requirement of endogenous pheromone production for plasma induction.
1. Hirt, H., Schlievert, P. M. & Dunny, G. M. (2002) Infect. Immun. 70, 716-723.
2. Tortorello, M. L. & Dunny, G. M. (1985) J. Bacteriol. 162, 131-137.
3. Galli, D., Friesenegger, A. & Wirth, R. (1990) Mol. Microbiol. 4, 895-904.
4. Spector, A. A. (1986) Methods Enzymol. 128, 320-339.