- Email: [email protected]
Effect of in vivo growth conditions upon expression of surface protein antigens in Enterococcus faecalis
FEMS MicrobiologyImmunology64 (1990) 5t-54 Publishedby Elsevier
51
FEM$IM 00100
Effect of in vivo growth conditions upon expression of surface protein antigens in Enterococcus faecalis P.A. L a m b e r t 1, p.j. Shorrock 1, E.J. Aitchison 1, P.A.G. D o m i n g u e 2, M.E. Power 2 a n d J.W. Costerton 2 / MicrobiologF . . Research . Group, . Pharmaceutical . . Sciences . lnstztute, . . Unwer~zty,Birrmngham, U.K, Aston and z Department of BiologicalSciences, The University of Calgary, Calga~. Alberta, Canada Received16 November1989 Accepted 5 December1989 Key words: Enterococcus faecalis; Surface protein; Antigen; Growth conditions; Biofilm
1. SUMMARY Western blotting of whole-cell preparations of Enterococcus faecalis showed the protein-antigen profiles to be markedly influenced by growth conditions. The E. faevalis-speCtfic antigens of 40 and 37 kDa, which are prominent in endocarditis, were strongly expressed following growth in serum or brain heart infusion, but not after growth in a chemically defined medium. The expression of these antigens in vivo was demonstrated in cells grown as a biofilm on silastic discs in the peritoneum of rabbits. These in vivo culture conditions may be useful in studying the pathogenesis of E. faecalis infections and the effectiveness of antibiotic therapy. 2. INTRODUCTION E. faecalis produces a number of surface protein anti~,ens which can be exploited in the serodiagnosis of E. faecalis endocarditis [1]. In developing a serodiagnostic test, cells were grown in
serum to simulate the in vivo nutritional conditions of the transient bacteremia presumed to occur before colonisation of the cardiac vegetations [2]. To study more closely the mechanisms of pathogenesis of E. faecalis in endocarditis, in particular the attachment of the organism to components of the cardiac vegetation, we require a model system in which cells are grown in vivo as an adherent biofilm or microcolonies on a surface [3]. This mode of growth permits interaction with the environment of the host, e.g. binding of host proteins [4] and phenotypic response to it, which could significantly influence the surface properties of the organism [5]. We have developed and evaluated one such model using silastic (silicone rubber) discs implanted in specially-designed chambers in the peritoneum of rabbits. These provide a surface for the organism to colonise and grow under in vivo nutritional conditions, yet are easily recovered for immunochemical analysis. We report here on the antigenic composition of the in vivo grown ceils. 3. MATERIALS A N D METHODS
Correspondence to: P.A. Lambert, Microbiology Research Group, PharmaceuticalSciences Institute, Aston University, AstonTriangle,BirminshamB4 7ET, U.K.
3.1. Organism and in vitro culture conditions E. faecalis strain EBH1 was obtained from a
0920-8534/90/$03.50© 1990 Federationof EuropeanMicrobiologicalSocieties
~ ~
discs with a scalpel blade were checked for purity, washed and resuspended as above.
~ C A P
FLANGE
(~)~RUBBER
WASHER
~
RUBBERWASHER ~MEMBRANE FILTER (~RUBBER WASHER ~FLANGE ~CAP
Fig. 1. Exploded view of the 'Calgary Chamber' with disc assembly.
3.3. SDS-PAGE and western blotting The methods were as described [1]. Samples applied to the gels were either whole-cell suspensions or mutanolysin digests of whole cells. The latter were prepared by treating cell suspensions with mutanolysin (Sigma, Poole, U.K.), 50 u n i t s / m l in 2 m M Hepes buffer, p H 7.2, containing 1 mM phenylmethylsulphonyl fluoride for 18 h at 37°C. Reduction in turbidity was at least 80~ as reported by Morris et al. [7]. Whole cells or the mutanolysin digests were heated at 100°C for 10 rain in denaturing buffer containing 3~ SDS and 5~ mercaptoethanol before electrophoresis. Serum for detect~c)n of antigens was obtained from a patient who had suffered repeated episodes of E. faecalis endocarditis. The high IgG titre to E. faecalis antigens was useful in revealing the full range of antigens expressed. 4. RESULTS
blood culture of an endocarditis patient at East Birmingham Hospital, Birmingham, U.K. [1]. Cells were grown for 18 h with shaking at 3 7 ° C in brain heart infusion (Difco. Detroit, U.S.A. or Labm, Bury, U.K.), heat-inactivated horse serum (Gibco Europe, Paisley, U.K.), or a chemically defined medium (CDM) [6]. Cells were washed three times in saline and resuspended in water to an absorbance (470 rim) of 5. 3.2. Implanted chambers for in vivo growth Sterile 8 nun diameter discs, cut from silastic sheeting (1 mm thick, Dow Coming, MI, U.S.A.) and threaded onto niehrome wire (Fig. 1), were colonised by suspension in inoculated serum for 1 h at 37 ° C prior to implantation. Disc assemblies were then drained and transferred to acrylic implant chambers (the 'Calgary Chamber'; made to our design by Tyler Research Instruments, Edmonton, Canada) which were subsequently sealed with sterile 0.22/tm cellulose acetate filters. These were then positioned intraperitoneally in New Zealand White rabbits and left in situ for 3 days. Bacteria recovered by scraping the silastic
The species-specific antigens of 40 and 37 kDa were most strongly expressed following growth in I
2
3
4
$
6
7
S
9
10
|t
12
Fig. 2. We~terublot of £./aecalls wholecells (odd-numbered lanes) and mutanolysindigests (even-numberedlanes) probed with serum (1:50 dilution) from an E./aecalis endocerditis patient with a high lgG titre to E. faecalls and revealed with protein A-peroxidase.Growthconditionswere: 1 & 2, CDM; 3 & 4, CDM+ 1~ hone serum;5 & 6 Difcobrain heart infusion; 7 & 8, Labm brain heart infusion;9 & 10, horse serum; II & 12, cells recovered from silasti¢discs implantedin inoculated chambersin the peritoneumof rabbits for 3 days.
serum (Fig. 2, lanes 9 and 10) and were also prominent in the profdes of brain heart infusiongrown cells (lanes 5-8) and in biofilm-grown cells (lanes 11 and 12). These antigens were conspicuously lacking from the cells grown in CDM or in CDM + 1% serum (lanes 1-4). No antigens were released from the CDM-grown cells by boiling in denaturing buffer. Mutanolysin digestion of these cells was required to give a profile on the western blots. Addition of 1% serum to the CDM restored the sensitivity to denaturing buffer and increased the expression of the 73-kDa antigen.
5. DISCUSSION Growth conditions had a pronounced influence upon the antigenic profiles of E. faecalis as revealed by western blotting. The 40- and 37-kDa protein antigens showed greatest phenotypic variation with maximum expression in serum and virtual absence in CDM. The CDM-grown cells were extremely refractive to treatment with denaturing buffer. Antigens were only released by digestion of the cell wall with mutanolysin. Surprisingly, addition of just 1% horse serum rendered the cells sensitive to denaturing buffer and increased expression of the 73-kDa antigen. The factor in the serum responsible for these changes is unknown. Expression of the 73-, 40- and 37-kDa-antigens in vivo was demonstrated directly by growth of the cells adherent to silasric discs in rabbit peritoneal chambers. Their detection by lgG in human antiserum from an E. faecalis endocarditis patient strongly suggests that these antigens are prominently expressed in endocarditis. Our previous studies have shown that these antigens are particularly associated with endocarditis, and are not generally detected by serum from patients with other E. faecalis infections [1,2]. This could be due to the protracted nature of the infection in endocarditis and prolonged exposure to antigens, or to the growth conditions in the cardiac vegetation, where the cells are exposed to soluble serum factors but not to cellular defence mechanisms. There was a difference in relative levels of these antigens between the in vivo chamber- and
serum-grown cells. This could be due both to differences in the nutritional contents of the chamber fluid and the heat-inactivated serum, and to the different mode of growth, i.e. free suspended cells in serum and biofilm cells on the discs in the chamber. The pore size of the filters used to seal the chambers allowed exposure only to humoral factors and lyric products of degranulation. Thus components likely to influence bacterial growth, such as complement and cationic antibacterial proteins would be present in the chamber fluid but not in the heat-inactivated serum. Both the serum-grown cells and the cells recovered from the chambers expressed broad, diffuse bands in the 80-kDa region which were recognised by the patient's serum. These bands were not detected in cells grown in CDM or brain heart infusion. Serum from the rabbits used for the implanted chambers contained no E. faecalis antibody, due to the short incubation period (3 days). However, immune rabbit serum prepared by sensirising rabbits over 8 weeks with serum-grown ceils gave essentially the same profiles as in Fig. 2 (data not shown). We believe that growth on implanted silasric discs in our chamber is attractive as a model for study of microbial properties related to pathogenicity [3]. Firstly, it provides a rapid, reproducible source of in vivo-grown cells in sufficient quantity for immunochemical study. Secondly, the biofdm mode of growth simulates the conditions in infected tissues or colouised devices and therefore provides a more realistic model for study of the effectiveness of antimierobial therapy than the planktonic (free suspended) cells which conventionally are used [8]. Clearly further studies are required to elucidate the nutritional conditions which influence growth in vivo.
REFERENCES [l] Aitchison, E.J., Lambert, P.A., Smith, E.G. and Farrell, I.D. (1987) Serodiagnosisof Streptococcus faecalis endocarditis by immunoblottingof surface proteinantigens.J. Clin. MicrobioL25, 211-215. [2] Shorrock,P.J., Lambert,P.A.,Aitchison,E.J., Smith,E.G., Farrell, I.D. and Gutschik,E. (1990)Serologicalresponse
in Enterococcus faecalis endocarditis determined by an enzyn~linked immunosorbent assay. J. Ciin. Microbiol. 28 (2). [3] Costerton, J.W., Chen8, K-J., Gececy, G.G., Ladd, T.I., Nickel, J.C., Dasgupta, M. and Marrie, TJ. (1987)Bacterial biofilms in nature and disease. Ann. Rev. Microbiol. 41, 435-464. [4] Shorrock, P.J. and Lambert, P.A. (1989) Binding of fibron~fin and albumin to Enterococcus (Streptococcus) faeca//~. Microb. Pathogenesh 6, 61-67. [5] Brown, M.R.W. and Williams, P. (1985) The influence of environm~t on envelope properties affecting survival of bacteria in infections. Ann. Rev. Microbiol. 39, 527-556.
[6] Shcekman, G.D., Conover, MJ., Kolb, JJ., Riley, L.S. and T o n n ~ G. (1961) Nutritional requirements for bacterial cell wall synthesh. J. Ba~eriol. 81, 44--S0. [7] Morris, F..I., ~ , N. and McBride, B.C. (1985) Cell s u r f ~ components of Streptococcus ~m~i~: relationship to a88regation, adherence and hydrophobicity..i. Bacterlol. 164, 255-262. [8] Nickel, J.C., Ruseska, I., Wright, J.B. and Coslerton, J.W. (1985) Tobramycin res/sumce of Pseudomonas aeruSinosa cells growing as a biofilm on urinary catheter material. Antimicrob. Agent Chemother. 27, 619-624.
51
FEM$IM 00100
Effect of in vivo growth conditions upon expression of surface protein antigens in Enterococcus faecalis P.A. L a m b e r t 1, p.j. Shorrock 1, E.J. Aitchison 1, P.A.G. D o m i n g u e 2, M.E. Power 2 a n d J.W. Costerton 2 / MicrobiologF . . Research . Group, . Pharmaceutical . . Sciences . lnstztute, . . Unwer~zty,Birrmngham, U.K, Aston and z Department of BiologicalSciences, The University of Calgary, Calga~. Alberta, Canada Received16 November1989 Accepted 5 December1989 Key words: Enterococcus faecalis; Surface protein; Antigen; Growth conditions; Biofilm
1. SUMMARY Western blotting of whole-cell preparations of Enterococcus faecalis showed the protein-antigen profiles to be markedly influenced by growth conditions. The E. faevalis-speCtfic antigens of 40 and 37 kDa, which are prominent in endocarditis, were strongly expressed following growth in serum or brain heart infusion, but not after growth in a chemically defined medium. The expression of these antigens in vivo was demonstrated in cells grown as a biofilm on silastic discs in the peritoneum of rabbits. These in vivo culture conditions may be useful in studying the pathogenesis of E. faecalis infections and the effectiveness of antibiotic therapy. 2. INTRODUCTION E. faecalis produces a number of surface protein anti~,ens which can be exploited in the serodiagnosis of E. faecalis endocarditis [1]. In developing a serodiagnostic test, cells were grown in
serum to simulate the in vivo nutritional conditions of the transient bacteremia presumed to occur before colonisation of the cardiac vegetations [2]. To study more closely the mechanisms of pathogenesis of E. faecalis in endocarditis, in particular the attachment of the organism to components of the cardiac vegetation, we require a model system in which cells are grown in vivo as an adherent biofilm or microcolonies on a surface [3]. This mode of growth permits interaction with the environment of the host, e.g. binding of host proteins [4] and phenotypic response to it, which could significantly influence the surface properties of the organism [5]. We have developed and evaluated one such model using silastic (silicone rubber) discs implanted in specially-designed chambers in the peritoneum of rabbits. These provide a surface for the organism to colonise and grow under in vivo nutritional conditions, yet are easily recovered for immunochemical analysis. We report here on the antigenic composition of the in vivo grown ceils. 3. MATERIALS A N D METHODS
Correspondence to: P.A. Lambert, Microbiology Research Group, PharmaceuticalSciences Institute, Aston University, AstonTriangle,BirminshamB4 7ET, U.K.
3.1. Organism and in vitro culture conditions E. faecalis strain EBH1 was obtained from a
0920-8534/90/$03.50© 1990 Federationof EuropeanMicrobiologicalSocieties
~ ~
discs with a scalpel blade were checked for purity, washed and resuspended as above.
~ C A P
FLANGE
(~)~RUBBER
WASHER
~
RUBBERWASHER ~MEMBRANE FILTER (~RUBBER WASHER ~FLANGE ~CAP
Fig. 1. Exploded view of the 'Calgary Chamber' with disc assembly.
3.3. SDS-PAGE and western blotting The methods were as described [1]. Samples applied to the gels were either whole-cell suspensions or mutanolysin digests of whole cells. The latter were prepared by treating cell suspensions with mutanolysin (Sigma, Poole, U.K.), 50 u n i t s / m l in 2 m M Hepes buffer, p H 7.2, containing 1 mM phenylmethylsulphonyl fluoride for 18 h at 37°C. Reduction in turbidity was at least 80~ as reported by Morris et al. [7]. Whole cells or the mutanolysin digests were heated at 100°C for 10 rain in denaturing buffer containing 3~ SDS and 5~ mercaptoethanol before electrophoresis. Serum for detect~c)n of antigens was obtained from a patient who had suffered repeated episodes of E. faecalis endocarditis. The high IgG titre to E. faecalis antigens was useful in revealing the full range of antigens expressed. 4. RESULTS
blood culture of an endocarditis patient at East Birmingham Hospital, Birmingham, U.K. [1]. Cells were grown for 18 h with shaking at 3 7 ° C in brain heart infusion (Difco. Detroit, U.S.A. or Labm, Bury, U.K.), heat-inactivated horse serum (Gibco Europe, Paisley, U.K.), or a chemically defined medium (CDM) [6]. Cells were washed three times in saline and resuspended in water to an absorbance (470 rim) of 5. 3.2. Implanted chambers for in vivo growth Sterile 8 nun diameter discs, cut from silastic sheeting (1 mm thick, Dow Coming, MI, U.S.A.) and threaded onto niehrome wire (Fig. 1), were colonised by suspension in inoculated serum for 1 h at 37 ° C prior to implantation. Disc assemblies were then drained and transferred to acrylic implant chambers (the 'Calgary Chamber'; made to our design by Tyler Research Instruments, Edmonton, Canada) which were subsequently sealed with sterile 0.22/tm cellulose acetate filters. These were then positioned intraperitoneally in New Zealand White rabbits and left in situ for 3 days. Bacteria recovered by scraping the silastic
The species-specific antigens of 40 and 37 kDa were most strongly expressed following growth in I
2
3
4
$
6
7
S
9
10
|t
12
Fig. 2. We~terublot of £./aecalls wholecells (odd-numbered lanes) and mutanolysindigests (even-numberedlanes) probed with serum (1:50 dilution) from an E./aecalis endocerditis patient with a high lgG titre to E. faecalls and revealed with protein A-peroxidase.Growthconditionswere: 1 & 2, CDM; 3 & 4, CDM+ 1~ hone serum;5 & 6 Difcobrain heart infusion; 7 & 8, Labm brain heart infusion;9 & 10, horse serum; II & 12, cells recovered from silasti¢discs implantedin inoculated chambersin the peritoneumof rabbits for 3 days.
serum (Fig. 2, lanes 9 and 10) and were also prominent in the profdes of brain heart infusiongrown cells (lanes 5-8) and in biofilm-grown cells (lanes 11 and 12). These antigens were conspicuously lacking from the cells grown in CDM or in CDM + 1% serum (lanes 1-4). No antigens were released from the CDM-grown cells by boiling in denaturing buffer. Mutanolysin digestion of these cells was required to give a profile on the western blots. Addition of 1% serum to the CDM restored the sensitivity to denaturing buffer and increased the expression of the 73-kDa antigen.
5. DISCUSSION Growth conditions had a pronounced influence upon the antigenic profiles of E. faecalis as revealed by western blotting. The 40- and 37-kDa protein antigens showed greatest phenotypic variation with maximum expression in serum and virtual absence in CDM. The CDM-grown cells were extremely refractive to treatment with denaturing buffer. Antigens were only released by digestion of the cell wall with mutanolysin. Surprisingly, addition of just 1% horse serum rendered the cells sensitive to denaturing buffer and increased expression of the 73-kDa antigen. The factor in the serum responsible for these changes is unknown. Expression of the 73-, 40- and 37-kDa-antigens in vivo was demonstrated directly by growth of the cells adherent to silasric discs in rabbit peritoneal chambers. Their detection by lgG in human antiserum from an E. faecalis endocarditis patient strongly suggests that these antigens are prominently expressed in endocarditis. Our previous studies have shown that these antigens are particularly associated with endocarditis, and are not generally detected by serum from patients with other E. faecalis infections [1,2]. This could be due to the protracted nature of the infection in endocarditis and prolonged exposure to antigens, or to the growth conditions in the cardiac vegetation, where the cells are exposed to soluble serum factors but not to cellular defence mechanisms. There was a difference in relative levels of these antigens between the in vivo chamber- and
serum-grown cells. This could be due both to differences in the nutritional contents of the chamber fluid and the heat-inactivated serum, and to the different mode of growth, i.e. free suspended cells in serum and biofilm cells on the discs in the chamber. The pore size of the filters used to seal the chambers allowed exposure only to humoral factors and lyric products of degranulation. Thus components likely to influence bacterial growth, such as complement and cationic antibacterial proteins would be present in the chamber fluid but not in the heat-inactivated serum. Both the serum-grown cells and the cells recovered from the chambers expressed broad, diffuse bands in the 80-kDa region which were recognised by the patient's serum. These bands were not detected in cells grown in CDM or brain heart infusion. Serum from the rabbits used for the implanted chambers contained no E. faecalis antibody, due to the short incubation period (3 days). However, immune rabbit serum prepared by sensirising rabbits over 8 weeks with serum-grown ceils gave essentially the same profiles as in Fig. 2 (data not shown). We believe that growth on implanted silasric discs in our chamber is attractive as a model for study of microbial properties related to pathogenicity [3]. Firstly, it provides a rapid, reproducible source of in vivo-grown cells in sufficient quantity for immunochemical study. Secondly, the biofdm mode of growth simulates the conditions in infected tissues or colouised devices and therefore provides a more realistic model for study of the effectiveness of antimierobial therapy than the planktonic (free suspended) cells which conventionally are used [8]. Clearly further studies are required to elucidate the nutritional conditions which influence growth in vivo.
REFERENCES [l] Aitchison, E.J., Lambert, P.A., Smith, E.G. and Farrell, I.D. (1987) Serodiagnosisof Streptococcus faecalis endocarditis by immunoblottingof surface proteinantigens.J. Clin. MicrobioL25, 211-215. [2] Shorrock,P.J., Lambert,P.A.,Aitchison,E.J., Smith,E.G., Farrell, I.D. and Gutschik,E. (1990)Serologicalresponse
in Enterococcus faecalis endocarditis determined by an enzyn~linked immunosorbent assay. J. Ciin. Microbiol. 28 (2). [3] Costerton, J.W., Chen8, K-J., Gececy, G.G., Ladd, T.I., Nickel, J.C., Dasgupta, M. and Marrie, TJ. (1987)Bacterial biofilms in nature and disease. Ann. Rev. Microbiol. 41, 435-464. [4] Shorrock, P.J. and Lambert, P.A. (1989) Binding of fibron~fin and albumin to Enterococcus (Streptococcus) faeca//~. Microb. Pathogenesh 6, 61-67. [5] Brown, M.R.W. and Williams, P. (1985) The influence of environm~t on envelope properties affecting survival of bacteria in infections. Ann. Rev. Microbiol. 39, 527-556.
[6] Shcekman, G.D., Conover, MJ., Kolb, JJ., Riley, L.S. and T o n n ~ G. (1961) Nutritional requirements for bacterial cell wall synthesh. J. Ba~eriol. 81, 44--S0. [7] Morris, F..I., ~ , N. and McBride, B.C. (1985) Cell s u r f ~ components of Streptococcus ~m~i~: relationship to a88regation, adherence and hydrophobicity..i. Bacterlol. 164, 255-262. [8] Nickel, J.C., Ruseska, I., Wright, J.B. and Coslerton, J.W. (1985) Tobramycin res/sumce of Pseudomonas aeruSinosa cells growing as a biofilm on urinary catheter material. Antimicrob. Agent Chemother. 27, 619-624.