- Email: [email protected]
Importance of medium and atmosphere type to both slime production and adherence by coagulase-negative staphylococci
Journal
of Hospital
Infection
(19’92) 20, 173-184
Importance of medium and atmosphere both slime production and adherence coagulase-negative staphylococci M. Hussain”f,
M. H. Wilcox*, P. J. White?, R. C. Spencer*
M. K.‘Faulknerf
Departments of “Experimental and Clinical Microbiology, and Biotechnology, and f Pathology, University Acceptedjfoor publication
type to by
and
tMolecular Biology of She&Geld
3 December 1991
Summary:
Marked differences in both the production of slime and adherence by Staphylococcus epidermidis were observed when comparing four culture media. Slime isolated from a strain cultured in a chemically defined medium (HHW) in air was chemically indistinguishable from that formed in both HHW and synthetic dialysis fluid (SDF) in air with 5% CO,. The presence of a physiological level of CO, during culture in tryptone soya broth (TSB) prevented production of slime. It was not possible to separate the constituents of slime from those of the culture medium in bacteria grown in TSB in air using DEAE cellulose. Slime production was notably poor in used peritoneal dialysis fluid (PUD). Adherent growth was marked in HHW and SDF but was poor in TSIB and PUD when air with 5% CO, was used. These findings emphasize the adlvantages in using chemically defined and biological fluids when studying slime production and adherence by S. epidermidis. Keywords:
Staphylococcus
epidermidis;
slime production;
adherence.
Introduction Controversy surrounds the importance of slime production by coagulase-negative staphyl.ococci (CNS) as a virulence determinant in medical device-associated infections. Although the terms adherence and slime production are frequ.ently used as if they are the same process it is apparent that the two processes are not the same.’ It has been noted that CNS slime production is media-dependent,2 and more recently that the atmospheric CO, content may significantly affect adherence to polymer surfaces.3r4 A chemically defined culture medium (HHW)’ has been successfully employed to determine the composition of CNS slime.6 However, the in-vivo behaviour of CNS is likely to be significantly different from that observed in both chemically defined and conventional complex Correspondence to: Dr M. H. Wilcox Department of Experimental and University of Sheffield Medical School, Beech Hill Road, Sheffield SlO 2RX 0195-6701/92/030173+
12 SO3,OOjO
Clinical
0 1992 The Hospml
173
Microbiology,
Infection
Society
174
Hussain
et al.
media in vitro. We have, therefore, compared both production of a well-characterized CNS strain7,’ in a biological fluid, represented by used peritoneal addition, the effect of atmospheric CO, content production of slime has been examined.
the adherence and slime such media, with that in dialysis fluid (PUD). In during growth on the
Methods
Media Four media were used: a chemically defined medium (HHW)-described previously5 consisting primarily of glucose, 18 amino acids, two purines and six vitamins; tryptone soya broth (TSB, Oxoid); a synthetic dialysis fluid (SDF),9 and sterile pooled used dialysis fluid (PUD),” collected from ten patients undergoing continuous ambulatory peritoneal dialysis (CAPD) at the Northern General Hospital, Sheffield. . Bacterium A Staphylococcus
epidermidis
strain
RP62A
(ATCC
35984) was used.7Ts
Isolation and characterization of slime The methods employed to identify the composition of slime have been described in detail elsewhere.6 The strain was cultured in 500 ml of medium in a 1 1 glass flask in either air, or air containing 5% CO, for 24 h at 37°C. Following incubation, the contents of the flask were blended in an M.S.E. Atomax at half maximum speed for 1.5 min, and then centrifuged at 6000 g for 30 min at 4°C. Slime was separated from the supernatant by ultrafiltration (Filtron membrane cells, Gallenkamp) using a molecular mass cut-off of 10 kDa; 1.50 ml of supernatant was concentrated to 30 ml. The concentrate was washed twice by adding 120 ml distilled water and reconcentrating, and was then freeze-dried. Crude slime was purified by ion-exchange and affinity chromatography, and characterized by gas liquid chromatography, enzymic and chemical assays.6 Electron microscopy Bacterial adherence and slime production by planktonic cells were visualized by scanning and transmission electron microscopy (SEM and TEM), respectively. Bacteria were cultured in glass universal containers with loose-fitting tops in 10 ml of medium for 24 h at 37°C in either air, or air with 5 % CO,. Each vessel contained a flat polystyrene square (1 cm2) cut from weighing boats (L.I.P. Equipment and Services Ltd, UK), suspended vertically beneath the surface of the medium by wire. Following incubation, the polystyrene squares were carefully removed, washed three times by immersion in sterile saline, and the remaining adherent bacteria was then fixed in 3% glutaraldehyde. Planktonic bacteria were collected by centrifugation (1500 g for 5 min) and then fixed in 3 % glutaraldehyde
CNS slime
production
and
175
adherence
containing 0.1% ruthenium red. The procedures for sample preparation for SEM and TEM were similar to those described by Mayberry-Carson et al.” Following fixation in glutaraldehyde, samples were washed twice in 200 mM cacodylate buffer, fixed for 1 h in osmium tetroxide (1% w/v in distilled water) and then dehydrated in alcohol (TEM). Critical point drying (SEM) was carried out in alcohol and liquid CO,. TEM samples were viewed on a Philips 400 machine at an acceleration voltage of 60 kV. Gold-coated SEM samples were examined on a Philips 501 apparatus at an acceleration voltage of 15kV. Results
Table I indicates the relative amounts of slime that were recovered by ultrafiltration, according to the media and atmosphere employed. Slime produced by RP62A cultured in HHW in air, or in air enriched with CO, were chemically similar. Both specimens fractionated into one major and one minor peak by stepwise elution on DEAE-cellulose, and consisted primarily of glucose, glucosamine, alanine, glycerol phosphate and protein (15%). Following culture in TSB in air, isolated slime was fractionated into four peaks on DEAE-cellulose (Figure la). Peaks two and three, eluted with 0.25 M NaCl in buffer, were better separated in material obtained from the culture heated to 100°C prior to the isolation of crude slime (Figure lb). Uninoculated TSB subjected to the same isolation and fractionation procedures gave two peaks (Figure lc). The fractionated material was analysed by gas chromatography: the first two slime fractions contained appreciable amounts of glucose, mannose, glucuronic and galacturonic
Table I. Yields of extracellular material (mg of freezedried material I-’ culture media), isolated by ultrafiltration,from RP62A cultured in air and air with 5% CO, in four different media Incubation Medium
Air
atmosphere Air with
type 5% CO,
HHW
89
88
TSB
163
130*
SDF
ND
54
PUD
ND
350*
ND: not done (see text). * The same quantities of material were recovered from uninoculated media.
OD 488
176
2.5 D
2.5
OD 488
c
nm
nm
Tube
number
CNS slime
production
and
177
adherence
acids and traces of ribose and galactose (Table II). These same monosaccharides were also present in the two fractions of uninoculated TSB. The third and fourth fractions contained mainly glucose and galacturonic acid and, additionally, the third obtained from the material isolated after heating, contained about 25% galactosamine. Slime was not detected by these methods after culture in TSB in air with 5% CO,; the fractionation pattern was the same as that obtained from uninoculated TSB. Precipitation due to alterations in pH occurred in SDF and PUD” incubated in air, and for this reason RP62A was only cultured in these fluids in an atmosphere enriched with carbon dioxide. A small amount of slime material was isolated from the strain cultured in SDF. Chemically, it was indistinguishable from that obtained from culture in HHW. Ultrafiltration of PUD following culture was extremely slow, probably due to the Table
II.
Gas-liquid chromatograph analysis of high molecular weight material RP62A cultured in TSB (and in TSB heated after incubation) DEAE
Constituent
cellulose
obtained from
fraction
I
II
2 (1)
3 (4) ND
1 (1)
1 (ND)
26#0)
2 i4)
1 (1)
1 (1)
Galactose
5 (7) 3
2 (2)
2 (2)
Glucose
12 (3) 2
19 (20)
4 (11)
Galactosamine
4 (6) ND
2 (26)
3 (2)
19 (16)
22 (12)
3 (2)
5 (ND)
Ribose
1
Mannose
Galacturonic Glucuronic
acid acid
Glycerol Figures isolated
4 (5) 13 ND
represent % (w/w) of freeze-dried from uninoculated TSB.
(ND) ND
III
3 (ND) 2 ND
(ND) ND
ND
material. ND: not detected. Figures
IV
(2)
4 (2)
in bold refer to material
Figure 1 (left). Fractionation of slime from S. epidermidis RP62A isolated after growth in tryptic soy broth on DEAE-cellul~ose by stepwise elution. The first peak (left to right) eluted with loading buffer, peaks 2 and 3 with 0.25 M NaCl in buffer and peak 4 with 0.5 M NaCl in buffer. Peaks were detected by assay with phenol/H,SO,. Tryptic soy broth containing slime was concentrated in ultrafiltration cell and freeze-dried material was loaded on DEAEcellulose (a). Before ultrafiltration the tryptic soy broth containing slime was heated to 100°C in a water bath for 15 min (b). Uninoculated tryptic soy broth concentrated in an ultrafiltration cell and freeze-dried material fractionated on DEAE-cellulose (c).
Hussain
et al.
b
Figure 2. Adherence and PUD (4.
of RP62A
cultured
in air with
5% CO, in HHW
(a), SDF
(b) TSB
(c)
CNS slime
d
production
and
adherence
179
180
Hussain
et al.
precipitation of albumin and salts.” The material isolated by this procedure material) with no was mostly protein (8.5% w/ w of freeze-dried carbohydrate detected by either phenol/sulphuric acid assay or paper chromatography. This protein is most probably derived not from RP62A, but from PUD. A similar amount of protein (81% w/w) was isolated from uninoculated PUD. Adherence of RP62A to polystrene varied markedly in the four media studied, and was greatest in HHW and poorest in PUD (in air with 5% CO,) (Figure 2u-4. Growth yield in these media varied significantly ( lo9 cfu ml-’ in HHW and TSB, compared with lo8 and lo7 cfu ml-’ in SDF and PUD, respectively). TEM was generally a more sensitive technique for identifying slime production compared with the isolation of material by ultrafiltration described above. Hence, small amounts of extracellular material were seen by TEM of RP62A cultured in PUD (Figure 3) and TSB in air with 5% CO, (not shown). As viewed by TEM, slime production was dramatically greater in TSB incubated in air compared with air with 5% CO, (not shown). Discussion
CNS are the most common cause of infection associated with the use of a wide range of indwelling devices. I2 How they adhere to and then persist on the surface of such implants remains uncertain. CNS slime was first described by Bayston and Penny,13 and the initial inclination to ascribe the process of adherence to this extracellular material has now been modified.
Figure 3. Evidence 5% co,.
of limited
production
of slime by RP62A
cultured
in PUD
in air with
CNS slime
production
and
adherence
181
Many believe that slime may have a role in consolidating adherent bacteria, so allowing the formation of a biofilm. Progress in understanding the pathogenesis of CNS device-associated infection has been hindered by a failure to appreciate that slime production and adherence are two distinct processes. ’ Furthermore, the in-vitro study of these phenomena has almost exclusively been carried out using complex culture media, and the relevance of results to the in-vivo situation is doubtful. Hussain et aZ.6and Drewry et aZ.14have recently demonstrated, for example, that the high galactose content of slime isolated from CNS cultured on agar is, in fact, agar derived. Four distinct, chromatographic fractions of slime are obtained after growth on agar, whereas only one major peak is seen in material isolated from organisms cultured on a dialysis membrane or in liquid.6 A chemically defined medium has been developed (HHW)’ consisting of low molecular weight constituents, that support the production of CNS slime in vitro; high molecular weight substances recovered from slime isolated after growth in HHW must be of bacterial origin. In this study we have demonstrated marked differences in both the adherence and production of slime by a CNS strain cultured in four media: HHW,’ TSB, which has often been used previously for in-vitro studies with CNS ‘,* SDF which was used elsewhere’ to simulate the conditions found in the’ peritoneal cavity of patients on CAPD, and PUD, which most closely represents the in-vivo situation. A previous study5 determined that HHW containing glucose supported the growth of eight CNS strains (including RP62A); none of the six other carbohydrates tested allowed growth of all the strains. The glucose content of the media used in the present study was lOgl-‘, 4.5 gl-‘, 3.5 gl-’ and 2.5 gl-’ f or HHW, PUD, SDF and TSB, respectively. Bacteria causing CAPD peritonitis, or indeed those causing venous catheter infection, are exposed to elevated tensions of CO, which are simulated in vitro by employing an atmosphere of air with 5% CO, during culture.” When cultured in TSB in air with 5% CO, compared with air alone, the adherent growth of the majority of CNS is grossly reduced.3 In the present study we have shown that the production of slime in TSB is also dramatically reduced in CO,-enriched conditions. Such a phenomenon was not observed in RP62A cultured in HHW. The reasons for this are uncertain but cannot be explained by growth yield. The influence of incubation atmosphere on the function of bacterial enzyme systems in these media is not clear.” The slime isolated from RP62A cultured in HHW in air’ was chemically indistinguishable from that formed in both HHW and SDF, in air enriched with CO,. It is assumed that the constituents of slime isolated from these media are primarily of microbial origin. Glycerol phosphate, alanine, glucose and glucosamine were consistently identified in the slime material, and these are also present in s,taphylococcal teichoic acid.16 It is feasible that slime consists mainly of cell wall derived factors. It was not possible on
182
Hussain
et al.
DEAE cellulose, however, to separate the constituents of slime from those of the culture medium in TSB grown cells. Heating of the bacterial culture to 100°C prior to isolation of slime l7 allowed a better fractionation of the crude material, but this was still less than ideal. Brock and Reiter” noted that mannose-containing sugars in slime isolated from S. aUYeUS also appeared to be derived from the medium (modified 110 broth) rather than from the bacteria. These observations reinforce the argument that complex conventional culture media are best avoided when studying CNS slime in vitro, and particularly if attempting to purify this material. Clearly, the behaviour of RP62A in PUD bore little resemblance to that in the other three media examined. SDF, which contains 0.4% nutrient broth, failed to simulate the conditions existing in human peritoneal dialysis fluid. Although bacterial growth was relatively poor in PUD, the difference in yield compared with culture in SDF was only small (c. lo-fold difference in cfu ml-‘). It is unlikely, therefore, that growth yield alone can explain the grossly reduced slime production and adherence that was observed in PUD. High molecular weight extracellular material could not be isolated from PUD-grown cells, but extracellular material was observed by TEM using planktonic bacteria. It was interesting to note that the latter material had a more delicate and closely cell-associated appearance than that observed in bacteria grown in CDM, TSB or SDF. Peritoneal dialysate is nutritionally poor with particularly low levels of free iron.‘9,20 There is a marked shut-down of cell surface proteins in CNS cultured in PUD compared with nutrient broth.‘i,** The composition of slime produced in PUD can still only be summarized, but is likely to be altered from that observed in the non-biological media examined. One explanation for the grossly reduced adherence of RP62A cultured in PUD compared with SDF, is the absence of albumin from the latter medium. Coating of polymer surfaces with albumin has previously been shown to reduce CNS adherence. 23 Media such as SDF, which are designed to simulate conditions in Z&JO, should be tested alongside a biological medium where possible, before assuming that representative bacterial behaviour will be obtained. In the conditions examined, where there was slime production by RP62A, there was adherence (and vice versa). Results using a different strain, however, do not confirm such an association.’ Kotilainen et al. have also recently demonstrated that both adherent and non-adherent S. epidermidis strains produce an immunogenic extracellular material.24 It was evident that RP62A adhered only to some areas of the polystyrene (Figure 3), rather than forming a complete monolayer. It is not clear whether subsequent microbial accumulation results primarily from multiplication of adherent cells, or from further bacterial deposition. We are presently studying the role of slime in such processes. Coagulase-negative staphylococci grown in nutrient broth adhere differently, and have altered cell surface protein profiles and carbohydrate content, when cultured in air compared
CNS slime
production
and
adherence
183
with air enriched with carbon dioxide.4 It remains to be elucidated whether surface proteins or carbohydrates are the more important determinants in CNS adherence. References MK, White PJ, Spencer RC. Slime production and 1. Wilcox MH, Hussain M, Faulkner adherence by coagulase-negative staphylococci. J Hasp Infect 1991; 18: 327-332. 2. Ludwicka A, Uhlenbruck G, Peters G et al. Investigation on extracellular slime substance produced by Staphylococcus epidermidis. Zentralbl Bakt Hyg A 1984; 258: 256-267. P. Effects of carbon dioxide 3. Wilcox MH, Denyer SF’, Finch RG, Smith DGE, Williams and sub-lethal levels of antibiotics on adherence of coagulase-negative staphylococci to polystyrene and silicone rubber. J Antimicrob Chemother 1991; 27: 577-587. 4. Denyer SP, Davies MC, Evans JA et al. Influence of carbon dioxide on the surface characteristics and adherence potential of coagulase-negative staphylococci. J Clin Microbial 1990; 28: 1813-1817. 5. Hussain M, Hastings JGM, White PJ. A chemically defined medium for slime production by coagulase-negative staphylococci. J Med Microbial 1991; 34: 143-147. and composition of the extracellular 6. Hussain M, Hastings JGM, White PJ. Isolation slime made by coagulase-negative staphylococci in a chemically defined medium. J Infect Dis 1991; 163: 534-541. GD, Simpson WA, Bisno AL, Beachy EH. Adherence of slime-producing 7. Christensen strains of Staphylococcus epidermidis to smooth surfaces. Infect Immun 1982; 37: 3 18-326. 8. Christensen Gfi, Baddou; LM, Simpson WA. The role ofadherence in the pathogenesis of coaeulase-negative stanhvlococcal infections. In: Pulverer G. Ouie PG. Peters G. Eds. Pathogenicity aid Clinical ‘Signi$cance of Coagulase-negative ‘Staphylococci. Stuttgart: Gustav Fischer Verlag 1987; 103-l 11. 9. Evans RC, Holmes CJ. Effect of vancomycin hydrochloride on Staphylococcus epidermidis biofilm associated with silicone elastomer. Antimicrob Agents Chemother 1987; 31: 889-894. IO. Wilcox MH, Denver SP, Fincih RG, Smith DGE, Williams P. Influence of carbon dioxide on the growth and antibiotic sensitivity of coagulase-negative staphylococci cultured in human peritoneal dialvsate. 7 Clin Microbial 1990: 28: 2183-2186. I?J, Tober-Me$er B,>mith JK, Lambe D\iv, Costerton JW. Bacter11. Mayberry-Carson ial adherence and glycocalyx formation in osteomyelitis experimentally induced with Staphylococcus aureus. Infect Immun 1984; 43: 825-833. LA. Laboratory, clinical and epidemiological aspects of coagu12. Pfaller MA, Herwaldt lase-negative staphylococci. Clin! Microbial Rev 1988; 1: 281-299. production of mucoid substance in Staphylococcus 13. Bayston R, Penny SR. Excessive SIIA: a possible factor in colonisation of Holter shunts. Develop Med Child Neural 1972; 14 (Suppl. 27): 25-28. DT,’ Galbraith L, Wilkinson BJ, Wilkinson SG. Staphylococcal slime: a 14. Drewr;cautionarv tale. ?’ Clin Microbial 1990: 28: 1292-1296. CO, of the growth and metabolism of micro15. Dixon N&l, Keil DB. The inhibition’by organisms. J Appl Bacterial 1989; 67: 109-36. and structure of cell 16. End1 J, Seidl HP, Fielder F, SchLleifer KM. Chemical composition wall teichoic acids of staphylococci. Arch Microbial 1983; 135: 215-233. N, Goldm,an DA, Pier GB. Isolation and characterisation of a 17. Tojo M, Yamashita capsular polysaccharide adhesin from Staphylococcus epidermidis. J Infect Dis 1988; 157: 713-722. 18. Brock JH, Reiter B. Chemical and biological properties of extracellular slime produced by Staphylococcus aureus grown in high carbohydrate, high salt medium. Infect Immun 1976; 13: 653-660. R, Lamiere N. The trace elements Br, Co, Cr, Cu, Fe, Mn, Rb, 19. Wallaeys B, Cornelius Se and Zn in serum, packed cells and dialysate of CAPD patients. In: Maher JF, Winchester JF, Eds. Frontiers in Peritoneal Dialysis. New York: Field, Rich and Assoc. Inc. 1986; 478-481.
184
Hussain
et al.
20. Williams P, Denyer SP, Finch RG. Protein antigens of Staphylococcus epidermidis grown under iron-restricted conditions in human peritoneal dialysate. FEMS Microbial Lett 1988; 50: 29-33. 21. Smith DGE, Wilcox MH, Finch RG, Denyer SP, Williams P. Characterisation of the cell envelope proteins of Staphylococcus epidermidis cultured in human peritoneal dialysate. Infect Immun 1991; 59: 617-624. 22. Wilcox MH, Williams P, Smith DGE, Modun B, Finch RG, Denyer SP. Variation in the expression of cell envelope proteins of coagulase-negative staphylococci cultured under iron-restricted conditions in human peritoneal dialysate. J Gen Microbial 1991; 137: 2561-2570. 23. Peters G, Schumacker-Pedreau F, Jansen B, Bey M, Pulverer G. Biology of S. epidermidis extracellular slime. In: Pulverer G, Quie PG, Peters G, Eds. Pathogenicity and Clinical SigniJicance of Coagulase-Negative Staphylococci. Stuttgart: Gustav Fischer \~ Verlag 1987; 137: 2561-2570. 24. Kotilainen P, Maki J, Oksman P, Viljanen MK, Nikoskelainen J, Huovinen P. Immunochemical analysis of the extracellular slime substance of Staphylococcus epidermidis. Eur J Cl% Microbial Infect Dis 1990; 9: 262-270.
of Hospital
Infection
(19’92) 20, 173-184
Importance of medium and atmosphere both slime production and adherence coagulase-negative staphylococci M. Hussain”f,
M. H. Wilcox*, P. J. White?, R. C. Spencer*
M. K.‘Faulknerf
Departments of “Experimental and Clinical Microbiology, and Biotechnology, and f Pathology, University Acceptedjfoor publication
type to by
and
tMolecular Biology of She&Geld
3 December 1991
Summary:
Marked differences in both the production of slime and adherence by Staphylococcus epidermidis were observed when comparing four culture media. Slime isolated from a strain cultured in a chemically defined medium (HHW) in air was chemically indistinguishable from that formed in both HHW and synthetic dialysis fluid (SDF) in air with 5% CO,. The presence of a physiological level of CO, during culture in tryptone soya broth (TSB) prevented production of slime. It was not possible to separate the constituents of slime from those of the culture medium in bacteria grown in TSB in air using DEAE cellulose. Slime production was notably poor in used peritoneal dialysis fluid (PUD). Adherent growth was marked in HHW and SDF but was poor in TSIB and PUD when air with 5% CO, was used. These findings emphasize the adlvantages in using chemically defined and biological fluids when studying slime production and adherence by S. epidermidis. Keywords:
Staphylococcus
epidermidis;
slime production;
adherence.
Introduction Controversy surrounds the importance of slime production by coagulase-negative staphyl.ococci (CNS) as a virulence determinant in medical device-associated infections. Although the terms adherence and slime production are frequ.ently used as if they are the same process it is apparent that the two processes are not the same.’ It has been noted that CNS slime production is media-dependent,2 and more recently that the atmospheric CO, content may significantly affect adherence to polymer surfaces.3r4 A chemically defined culture medium (HHW)’ has been successfully employed to determine the composition of CNS slime.6 However, the in-vivo behaviour of CNS is likely to be significantly different from that observed in both chemically defined and conventional complex Correspondence to: Dr M. H. Wilcox Department of Experimental and University of Sheffield Medical School, Beech Hill Road, Sheffield SlO 2RX 0195-6701/92/030173+
12 SO3,OOjO
Clinical
0 1992 The Hospml
173
Microbiology,
Infection
Society
174
Hussain
et al.
media in vitro. We have, therefore, compared both production of a well-characterized CNS strain7,’ in a biological fluid, represented by used peritoneal addition, the effect of atmospheric CO, content production of slime has been examined.
the adherence and slime such media, with that in dialysis fluid (PUD). In during growth on the
Methods
Media Four media were used: a chemically defined medium (HHW)-described previously5 consisting primarily of glucose, 18 amino acids, two purines and six vitamins; tryptone soya broth (TSB, Oxoid); a synthetic dialysis fluid (SDF),9 and sterile pooled used dialysis fluid (PUD),” collected from ten patients undergoing continuous ambulatory peritoneal dialysis (CAPD) at the Northern General Hospital, Sheffield. . Bacterium A Staphylococcus
epidermidis
strain
RP62A
(ATCC
35984) was used.7Ts
Isolation and characterization of slime The methods employed to identify the composition of slime have been described in detail elsewhere.6 The strain was cultured in 500 ml of medium in a 1 1 glass flask in either air, or air containing 5% CO, for 24 h at 37°C. Following incubation, the contents of the flask were blended in an M.S.E. Atomax at half maximum speed for 1.5 min, and then centrifuged at 6000 g for 30 min at 4°C. Slime was separated from the supernatant by ultrafiltration (Filtron membrane cells, Gallenkamp) using a molecular mass cut-off of 10 kDa; 1.50 ml of supernatant was concentrated to 30 ml. The concentrate was washed twice by adding 120 ml distilled water and reconcentrating, and was then freeze-dried. Crude slime was purified by ion-exchange and affinity chromatography, and characterized by gas liquid chromatography, enzymic and chemical assays.6 Electron microscopy Bacterial adherence and slime production by planktonic cells were visualized by scanning and transmission electron microscopy (SEM and TEM), respectively. Bacteria were cultured in glass universal containers with loose-fitting tops in 10 ml of medium for 24 h at 37°C in either air, or air with 5 % CO,. Each vessel contained a flat polystyrene square (1 cm2) cut from weighing boats (L.I.P. Equipment and Services Ltd, UK), suspended vertically beneath the surface of the medium by wire. Following incubation, the polystyrene squares were carefully removed, washed three times by immersion in sterile saline, and the remaining adherent bacteria was then fixed in 3% glutaraldehyde. Planktonic bacteria were collected by centrifugation (1500 g for 5 min) and then fixed in 3 % glutaraldehyde
CNS slime
production
and
175
adherence
containing 0.1% ruthenium red. The procedures for sample preparation for SEM and TEM were similar to those described by Mayberry-Carson et al.” Following fixation in glutaraldehyde, samples were washed twice in 200 mM cacodylate buffer, fixed for 1 h in osmium tetroxide (1% w/v in distilled water) and then dehydrated in alcohol (TEM). Critical point drying (SEM) was carried out in alcohol and liquid CO,. TEM samples were viewed on a Philips 400 machine at an acceleration voltage of 60 kV. Gold-coated SEM samples were examined on a Philips 501 apparatus at an acceleration voltage of 15kV. Results
Table I indicates the relative amounts of slime that were recovered by ultrafiltration, according to the media and atmosphere employed. Slime produced by RP62A cultured in HHW in air, or in air enriched with CO, were chemically similar. Both specimens fractionated into one major and one minor peak by stepwise elution on DEAE-cellulose, and consisted primarily of glucose, glucosamine, alanine, glycerol phosphate and protein (15%). Following culture in TSB in air, isolated slime was fractionated into four peaks on DEAE-cellulose (Figure la). Peaks two and three, eluted with 0.25 M NaCl in buffer, were better separated in material obtained from the culture heated to 100°C prior to the isolation of crude slime (Figure lb). Uninoculated TSB subjected to the same isolation and fractionation procedures gave two peaks (Figure lc). The fractionated material was analysed by gas chromatography: the first two slime fractions contained appreciable amounts of glucose, mannose, glucuronic and galacturonic
Table I. Yields of extracellular material (mg of freezedried material I-’ culture media), isolated by ultrafiltration,from RP62A cultured in air and air with 5% CO, in four different media Incubation Medium
Air
atmosphere Air with
type 5% CO,
HHW
89
88
TSB
163
130*
SDF
ND
54
PUD
ND
350*
ND: not done (see text). * The same quantities of material were recovered from uninoculated media.
OD 488
176
2.5 D
2.5
OD 488
c
nm
nm
Tube
number
CNS slime
production
and
177
adherence
acids and traces of ribose and galactose (Table II). These same monosaccharides were also present in the two fractions of uninoculated TSB. The third and fourth fractions contained mainly glucose and galacturonic acid and, additionally, the third obtained from the material isolated after heating, contained about 25% galactosamine. Slime was not detected by these methods after culture in TSB in air with 5% CO,; the fractionation pattern was the same as that obtained from uninoculated TSB. Precipitation due to alterations in pH occurred in SDF and PUD” incubated in air, and for this reason RP62A was only cultured in these fluids in an atmosphere enriched with carbon dioxide. A small amount of slime material was isolated from the strain cultured in SDF. Chemically, it was indistinguishable from that obtained from culture in HHW. Ultrafiltration of PUD following culture was extremely slow, probably due to the Table
II.
Gas-liquid chromatograph analysis of high molecular weight material RP62A cultured in TSB (and in TSB heated after incubation) DEAE
Constituent
cellulose
obtained from
fraction
I
II
2 (1)
3 (4) ND
1 (1)
1 (ND)
26#0)
2 i4)
1 (1)
1 (1)
Galactose
5 (7) 3
2 (2)
2 (2)
Glucose
12 (3) 2
19 (20)
4 (11)
Galactosamine
4 (6) ND
2 (26)
3 (2)
19 (16)
22 (12)
3 (2)
5 (ND)
Ribose
1
Mannose
Galacturonic Glucuronic
acid acid
Glycerol Figures isolated
4 (5) 13 ND
represent % (w/w) of freeze-dried from uninoculated TSB.
(ND) ND
III
3 (ND) 2 ND
(ND) ND
ND
material. ND: not detected. Figures
IV
(2)
4 (2)
in bold refer to material
Figure 1 (left). Fractionation of slime from S. epidermidis RP62A isolated after growth in tryptic soy broth on DEAE-cellul~ose by stepwise elution. The first peak (left to right) eluted with loading buffer, peaks 2 and 3 with 0.25 M NaCl in buffer and peak 4 with 0.5 M NaCl in buffer. Peaks were detected by assay with phenol/H,SO,. Tryptic soy broth containing slime was concentrated in ultrafiltration cell and freeze-dried material was loaded on DEAEcellulose (a). Before ultrafiltration the tryptic soy broth containing slime was heated to 100°C in a water bath for 15 min (b). Uninoculated tryptic soy broth concentrated in an ultrafiltration cell and freeze-dried material fractionated on DEAE-cellulose (c).
Hussain
et al.
b
Figure 2. Adherence and PUD (4.
of RP62A
cultured
in air with
5% CO, in HHW
(a), SDF
(b) TSB
(c)
CNS slime
d
production
and
adherence
179
180
Hussain
et al.
precipitation of albumin and salts.” The material isolated by this procedure material) with no was mostly protein (8.5% w/ w of freeze-dried carbohydrate detected by either phenol/sulphuric acid assay or paper chromatography. This protein is most probably derived not from RP62A, but from PUD. A similar amount of protein (81% w/w) was isolated from uninoculated PUD. Adherence of RP62A to polystrene varied markedly in the four media studied, and was greatest in HHW and poorest in PUD (in air with 5% CO,) (Figure 2u-4. Growth yield in these media varied significantly ( lo9 cfu ml-’ in HHW and TSB, compared with lo8 and lo7 cfu ml-’ in SDF and PUD, respectively). TEM was generally a more sensitive technique for identifying slime production compared with the isolation of material by ultrafiltration described above. Hence, small amounts of extracellular material were seen by TEM of RP62A cultured in PUD (Figure 3) and TSB in air with 5% CO, (not shown). As viewed by TEM, slime production was dramatically greater in TSB incubated in air compared with air with 5% CO, (not shown). Discussion
CNS are the most common cause of infection associated with the use of a wide range of indwelling devices. I2 How they adhere to and then persist on the surface of such implants remains uncertain. CNS slime was first described by Bayston and Penny,13 and the initial inclination to ascribe the process of adherence to this extracellular material has now been modified.
Figure 3. Evidence 5% co,.
of limited
production
of slime by RP62A
cultured
in PUD
in air with
CNS slime
production
and
adherence
181
Many believe that slime may have a role in consolidating adherent bacteria, so allowing the formation of a biofilm. Progress in understanding the pathogenesis of CNS device-associated infection has been hindered by a failure to appreciate that slime production and adherence are two distinct processes. ’ Furthermore, the in-vitro study of these phenomena has almost exclusively been carried out using complex culture media, and the relevance of results to the in-vivo situation is doubtful. Hussain et aZ.6and Drewry et aZ.14have recently demonstrated, for example, that the high galactose content of slime isolated from CNS cultured on agar is, in fact, agar derived. Four distinct, chromatographic fractions of slime are obtained after growth on agar, whereas only one major peak is seen in material isolated from organisms cultured on a dialysis membrane or in liquid.6 A chemically defined medium has been developed (HHW)’ consisting of low molecular weight constituents, that support the production of CNS slime in vitro; high molecular weight substances recovered from slime isolated after growth in HHW must be of bacterial origin. In this study we have demonstrated marked differences in both the adherence and production of slime by a CNS strain cultured in four media: HHW,’ TSB, which has often been used previously for in-vitro studies with CNS ‘,* SDF which was used elsewhere’ to simulate the conditions found in the’ peritoneal cavity of patients on CAPD, and PUD, which most closely represents the in-vivo situation. A previous study5 determined that HHW containing glucose supported the growth of eight CNS strains (including RP62A); none of the six other carbohydrates tested allowed growth of all the strains. The glucose content of the media used in the present study was lOgl-‘, 4.5 gl-‘, 3.5 gl-’ and 2.5 gl-’ f or HHW, PUD, SDF and TSB, respectively. Bacteria causing CAPD peritonitis, or indeed those causing venous catheter infection, are exposed to elevated tensions of CO, which are simulated in vitro by employing an atmosphere of air with 5% CO, during culture.” When cultured in TSB in air with 5% CO, compared with air alone, the adherent growth of the majority of CNS is grossly reduced.3 In the present study we have shown that the production of slime in TSB is also dramatically reduced in CO,-enriched conditions. Such a phenomenon was not observed in RP62A cultured in HHW. The reasons for this are uncertain but cannot be explained by growth yield. The influence of incubation atmosphere on the function of bacterial enzyme systems in these media is not clear.” The slime isolated from RP62A cultured in HHW in air’ was chemically indistinguishable from that formed in both HHW and SDF, in air enriched with CO,. It is assumed that the constituents of slime isolated from these media are primarily of microbial origin. Glycerol phosphate, alanine, glucose and glucosamine were consistently identified in the slime material, and these are also present in s,taphylococcal teichoic acid.16 It is feasible that slime consists mainly of cell wall derived factors. It was not possible on
182
Hussain
et al.
DEAE cellulose, however, to separate the constituents of slime from those of the culture medium in TSB grown cells. Heating of the bacterial culture to 100°C prior to isolation of slime l7 allowed a better fractionation of the crude material, but this was still less than ideal. Brock and Reiter” noted that mannose-containing sugars in slime isolated from S. aUYeUS also appeared to be derived from the medium (modified 110 broth) rather than from the bacteria. These observations reinforce the argument that complex conventional culture media are best avoided when studying CNS slime in vitro, and particularly if attempting to purify this material. Clearly, the behaviour of RP62A in PUD bore little resemblance to that in the other three media examined. SDF, which contains 0.4% nutrient broth, failed to simulate the conditions existing in human peritoneal dialysis fluid. Although bacterial growth was relatively poor in PUD, the difference in yield compared with culture in SDF was only small (c. lo-fold difference in cfu ml-‘). It is unlikely, therefore, that growth yield alone can explain the grossly reduced slime production and adherence that was observed in PUD. High molecular weight extracellular material could not be isolated from PUD-grown cells, but extracellular material was observed by TEM using planktonic bacteria. It was interesting to note that the latter material had a more delicate and closely cell-associated appearance than that observed in bacteria grown in CDM, TSB or SDF. Peritoneal dialysate is nutritionally poor with particularly low levels of free iron.‘9,20 There is a marked shut-down of cell surface proteins in CNS cultured in PUD compared with nutrient broth.‘i,** The composition of slime produced in PUD can still only be summarized, but is likely to be altered from that observed in the non-biological media examined. One explanation for the grossly reduced adherence of RP62A cultured in PUD compared with SDF, is the absence of albumin from the latter medium. Coating of polymer surfaces with albumin has previously been shown to reduce CNS adherence. 23 Media such as SDF, which are designed to simulate conditions in Z&JO, should be tested alongside a biological medium where possible, before assuming that representative bacterial behaviour will be obtained. In the conditions examined, where there was slime production by RP62A, there was adherence (and vice versa). Results using a different strain, however, do not confirm such an association.’ Kotilainen et al. have also recently demonstrated that both adherent and non-adherent S. epidermidis strains produce an immunogenic extracellular material.24 It was evident that RP62A adhered only to some areas of the polystyrene (Figure 3), rather than forming a complete monolayer. It is not clear whether subsequent microbial accumulation results primarily from multiplication of adherent cells, or from further bacterial deposition. We are presently studying the role of slime in such processes. Coagulase-negative staphylococci grown in nutrient broth adhere differently, and have altered cell surface protein profiles and carbohydrate content, when cultured in air compared
CNS slime
production
and
adherence
183
with air enriched with carbon dioxide.4 It remains to be elucidated whether surface proteins or carbohydrates are the more important determinants in CNS adherence. References MK, White PJ, Spencer RC. Slime production and 1. Wilcox MH, Hussain M, Faulkner adherence by coagulase-negative staphylococci. J Hasp Infect 1991; 18: 327-332. 2. Ludwicka A, Uhlenbruck G, Peters G et al. Investigation on extracellular slime substance produced by Staphylococcus epidermidis. Zentralbl Bakt Hyg A 1984; 258: 256-267. P. Effects of carbon dioxide 3. Wilcox MH, Denyer SF’, Finch RG, Smith DGE, Williams and sub-lethal levels of antibiotics on adherence of coagulase-negative staphylococci to polystyrene and silicone rubber. J Antimicrob Chemother 1991; 27: 577-587. 4. Denyer SP, Davies MC, Evans JA et al. Influence of carbon dioxide on the surface characteristics and adherence potential of coagulase-negative staphylococci. J Clin Microbial 1990; 28: 1813-1817. 5. Hussain M, Hastings JGM, White PJ. A chemically defined medium for slime production by coagulase-negative staphylococci. J Med Microbial 1991; 34: 143-147. and composition of the extracellular 6. Hussain M, Hastings JGM, White PJ. Isolation slime made by coagulase-negative staphylococci in a chemically defined medium. J Infect Dis 1991; 163: 534-541. GD, Simpson WA, Bisno AL, Beachy EH. Adherence of slime-producing 7. Christensen strains of Staphylococcus epidermidis to smooth surfaces. Infect Immun 1982; 37: 3 18-326. 8. Christensen Gfi, Baddou; LM, Simpson WA. The role ofadherence in the pathogenesis of coaeulase-negative stanhvlococcal infections. In: Pulverer G. Ouie PG. Peters G. Eds. Pathogenicity aid Clinical ‘Signi$cance of Coagulase-negative ‘Staphylococci. Stuttgart: Gustav Fischer Verlag 1987; 103-l 11. 9. Evans RC, Holmes CJ. Effect of vancomycin hydrochloride on Staphylococcus epidermidis biofilm associated with silicone elastomer. Antimicrob Agents Chemother 1987; 31: 889-894. IO. Wilcox MH, Denver SP, Fincih RG, Smith DGE, Williams P. Influence of carbon dioxide on the growth and antibiotic sensitivity of coagulase-negative staphylococci cultured in human peritoneal dialvsate. 7 Clin Microbial 1990: 28: 2183-2186. I?J, Tober-Me$er B,>mith JK, Lambe D\iv, Costerton JW. Bacter11. Mayberry-Carson ial adherence and glycocalyx formation in osteomyelitis experimentally induced with Staphylococcus aureus. Infect Immun 1984; 43: 825-833. LA. Laboratory, clinical and epidemiological aspects of coagu12. Pfaller MA, Herwaldt lase-negative staphylococci. Clin! Microbial Rev 1988; 1: 281-299. production of mucoid substance in Staphylococcus 13. Bayston R, Penny SR. Excessive SIIA: a possible factor in colonisation of Holter shunts. Develop Med Child Neural 1972; 14 (Suppl. 27): 25-28. DT,’ Galbraith L, Wilkinson BJ, Wilkinson SG. Staphylococcal slime: a 14. Drewr;cautionarv tale. ?’ Clin Microbial 1990: 28: 1292-1296. CO, of the growth and metabolism of micro15. Dixon N&l, Keil DB. The inhibition’by organisms. J Appl Bacterial 1989; 67: 109-36. and structure of cell 16. End1 J, Seidl HP, Fielder F, SchLleifer KM. Chemical composition wall teichoic acids of staphylococci. Arch Microbial 1983; 135: 215-233. N, Goldm,an DA, Pier GB. Isolation and characterisation of a 17. Tojo M, Yamashita capsular polysaccharide adhesin from Staphylococcus epidermidis. J Infect Dis 1988; 157: 713-722. 18. Brock JH, Reiter B. Chemical and biological properties of extracellular slime produced by Staphylococcus aureus grown in high carbohydrate, high salt medium. Infect Immun 1976; 13: 653-660. R, Lamiere N. The trace elements Br, Co, Cr, Cu, Fe, Mn, Rb, 19. Wallaeys B, Cornelius Se and Zn in serum, packed cells and dialysate of CAPD patients. In: Maher JF, Winchester JF, Eds. Frontiers in Peritoneal Dialysis. New York: Field, Rich and Assoc. Inc. 1986; 478-481.
184
Hussain
et al.
20. Williams P, Denyer SP, Finch RG. Protein antigens of Staphylococcus epidermidis grown under iron-restricted conditions in human peritoneal dialysate. FEMS Microbial Lett 1988; 50: 29-33. 21. Smith DGE, Wilcox MH, Finch RG, Denyer SP, Williams P. Characterisation of the cell envelope proteins of Staphylococcus epidermidis cultured in human peritoneal dialysate. Infect Immun 1991; 59: 617-624. 22. Wilcox MH, Williams P, Smith DGE, Modun B, Finch RG, Denyer SP. Variation in the expression of cell envelope proteins of coagulase-negative staphylococci cultured under iron-restricted conditions in human peritoneal dialysate. J Gen Microbial 1991; 137: 2561-2570. 23. Peters G, Schumacker-Pedreau F, Jansen B, Bey M, Pulverer G. Biology of S. epidermidis extracellular slime. In: Pulverer G, Quie PG, Peters G, Eds. Pathogenicity and Clinical SigniJicance of Coagulase-Negative Staphylococci. Stuttgart: Gustav Fischer \~ Verlag 1987; 137: 2561-2570. 24. Kotilainen P, Maki J, Oksman P, Viljanen MK, Nikoskelainen J, Huovinen P. Immunochemical analysis of the extracellular slime substance of Staphylococcus epidermidis. Eur J Cl% Microbial Infect Dis 1990; 9: 262-270.