Enzymatic differentiation of Candida parapsilosis from other Candida spp. in a membrane filtration test

Transcription

1 Journal of Microbiological Methods 53 (2003) Enzymatic differentiation of Candida parapsilosis from other Candida spp. in a membrane filtration test Tiene G.M. Bauters a, Renaat Peleman b, Marc Dhont c, Piet Vanhaesebrouck d, Hans J. Nelis a, * a Laboratory for Pharmaceutical Microbiology, Department of Pharmaceutical Analysis, Ghent University, Harelbekestraat 72, B-9000 Ghent, Belgium b Department of Internal Medicine, Division of Infectious Diseases, Ghent University Hospital, De Pintelaan 185, B-9000 Ghent, Belgium c Department of Obstetrics and Gynecology, Ghent University Hospital, De Pintelaan 185, B-9000 Ghent, Belgium d Neonatal Intensive Care Unit, Department of Paediatrics, Ghent University Hospital, De Pintelaan 185, B-9000 Ghent, Belgium Received 21 May 2002; received in revised form 24 July 2002; accepted 30 September 2002 Abstract A previously reported enzyme assay on a membrane filter using 4-methylumbelliferyl (4-MU)-N-acetyl-h-D-galactosaminide, -phosphate and -pyrophosphate as substrates for the differentiation of four Candida spp. has been extended to Candida parapsilosis. The substrate 4-MU-h-D-glucoside was hydrolyzed by 28 test strains of this species but to a variable extent by seven other yeasts also. For a full enzymatic differentiation of C. parapsilosis from other medical yeasts, a battery of six reactions was required. Of 71 C. parapsilosis positive clinical samples, 4.2% gave a false negative result due to overgrowth by Candida albicans. The present assay is more rapid than a described spectrofluorometric determination of h-d-glucosidase in a broth, i.e., 9 11 h versus up to >48 h. D 2002 Elsevier Science B.V. All rights reserved. Keywords: Candida parapsilosis; h-d-glucosidase; Membrane filtration 1. Introduction In recent years, non-albicans Candida species such as Candida parapsilosis are increasingly implicated in nosocomial infections, notably in cardiac surgery patients and neonates (Moran et al., 2002; Weems, 1992). Candida species contain a range of hydrolytic enzymes that are useful for their identification in clinical samples. Bobey and Ederer (1981) used a * Corresponding author. Tel.: ; fax: address: Hans.Nelis@rug.ac.be (H.J. Nelis). limited number of key enzymatic activities for the differentiation of Candida and Cryptococcus spp. As substrates, they used fluorogenic 4-methylumbelliferyl (4-MU) derivatives. Based on the differential reactions of yeast isolates with 4-MU-phosphate, 4- MU-pyrophosphate and 4-MU-h-D-glucoside at acidic ph, they were able to distinguish between Cryptococcus spp., Candida tropicalis, Candida glabrata and Candida albicans/candida krusei. At acidic ph, C. parapsilosis was 4-MU-h-D-glucoside negative. However, at neutral ph, Smitka and Jackson (1989) found C. parapsilosis to be strongly 4-MUh-D-glucoside positive. Some other Candida spp. also /02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved. PII: S (02)

2 12 T.G.M. Bauters et al. / Journal of Microbiological Methods 53 (2003) cleaved this substrate but at a much lower rate. To avoid false positives, the inoculum was standardized and fluorescence was read in a spectrofluorometer as a function of time. A reading above a defined threshold value indicated C. parapsilosis with allegedly 100% specificity. However, only a limited number of other Candida spp. was tested for possible interference. Although the test itself is rapid ( min), it requires preliminary culturing of the yeasts for h to prepare the heavy inoculum required. Hence, the real analysis time is >1 2 days. To reduce the test time of the detection of enzymes in Candida spp., the approach of Bobey and Ederer has been previously adapted to a membrane filter format (Bauters et al., 1999). The sample is filtered over a nylon membrane filter, which is subsequently incubated on a modified Sabouraud dextrose agar to yield microcolonies. In a second step, the membrane filter is cut in four quarters which are incubated for 30 min on a glass fiber pad impregnated with enzyme substrates, i.e., 4-MU-N-acetyl-h-D-galactosaminide, 4-MU-phosphate and 4-MU-pyrophosphate or no substrate, and digitonin acting as a membrane permeabilizer. C. albicans, C. tropicalis and C. krusei are differentiated based on the presence/absence of blue fluorescence on the first three parts of the filter. C. glabrata does not hydrolyze any of the substrates but was found to exhibit orange fluorescence in the absence of a substrate, on the fourth part of the filter. As C. parapsilosis was not included in the original test, the purpose of the present work was to investigate whether the addition of 4-MU-h-D-glucoside to the battery of enzyme substrates would be capable of differentiating C. parapsilosis, C. albicans, C. glabrata, C. krusei and C. tropicalis under the conditions of the membrane filtration procedure. In addition, in view of a recent report about the use of h-d-glucosidase activity to distinguish between C. albicans and C. dubliniensis (Sullivan and Coleman, 1998), an isolate of C. dubliniensis was also subjected to the current test conditions. 2. Materials and methods 2.1. Test organisms Yeasts used to test sensitivity and specificity included C. parapsilosis (n = 28), C. albicans (n = 129), C. glabrata (n = 39), C. krusei (n = 25), C. tropicalis (n = 26), C. lusitaniae (n = 4), C. dubliniensis (n = 1), C. kefyr (n = 2), C. guilliermondii (n = 2), Cryptococcus neoformans (n = 17), Saccharomyces cerevisiae, Geotrichum candidum and Trichosporon cutaneum. C. parapsilosis and C. dubliniensis strains were obtained from the Institute of Hygiene and Epidemiology, Mycology (IHEM) (Brussels, Belgium) (C. parapsilosis IHEM 2305, 4224, 2052, 1716, 4223, 4024, 4606 and 6395, and C. dubliniensis IHEM 14280). Other Candida and non-candida yeast isolates were from our own collection, as previously described (Bauters et al., 1999) Clinical samples Clinical C. parapsilosis positive samples included whole blood (n = 14), rectal (n = 26), oropharyngeal (n = 15), vaginal (n = 11) and finger swabs (n = 5). They were obtained from the Departments of Internal Medicine, Division of Infectious Diseases; Obstetrics and Gynecology, and from the Neonatal Intensive Care Unit from the Ghent University Hospital Filters, growth media, chemicals and reagents Nylon membrane filters (47 mm, 0.45-Am pore size) used for filtration of the samples and absorbent glass fiber pads were purchased from Gelman Sciences (Ann Arbor, MI, USA). Sabouraud glucose agar (SGA) (Difco Laboratories, Detroit, MI, USA) was supplemented with 5000 Ag/ml of a combination of ticarcillin (4688 Ag/ml) and clavulanic acid (312 Ag/ml) (TimentinR) (SmithKline Beecham Pharma, Genval, Belgium) (SGA-T). Other growth media were CHROMagar CandidaR (CHRO- Magar, CHROMagar Company, Paris, France) and cornmeal agar (Difco Laboratories) supplemented with 0.5% Tween 80. For the differentiation of Candida spp., five substrates were used, i.e., 4-MU-h-D-glucoside, 4-MUphosphate, 4-MU-h-D-galactoside (all from Sigma, St. Louis, MO, USA), 4-MU-N-acetyl-h-D-galactosaminide and 4-MU-pyrophosphate (both from Melford Laboratories, Ipswich, UK). Of each substrate, 1 mg was dissolved in 1 ml of dimethylsulfoxide (DMSO) (Fluka, Buchs, Switzerland). This stock solution was further diluted in 0.1 M phosphate buffer, ph 7.0 (4-

3 T.G.M. Bauters et al. / Journal of Microbiological Methods 53 (2003) MU-h-D-glucoside), in 0.1 M citrate phosphate buffer, ph 4.5 (4-MU-N-acetyl-h-D-galactosaminide, 4- MU-pyrophosphate and 4-MU-galactoside) or in 0.1 M citrate phosphate buffer, ph 3.4 (4-MU-phosphate) to a final concentration of 0.01%. The working solutions were sterilized by filtration over Nalgene disposable units (0.45-Am pore size, 250 ml) (Nalge, Rochester, NY, USA), dispensed in test tubes and frozen at 20 jc until use. Triton X-100 and digitonin were obtained from Merck (Darmstadt, Germany) and Sigma, respectively. API 20C AUX strips came from biomérieux Vitek (Hazelwood, MO, USA) and a latex-agglutination test kit (Krusei-ColorR) from Fumouze (Levallois Perret, France) Procedure All clinical specimens were processed within 30 min after arrival. Swabs were extracted by vortex mixing in 10 ml of buffered peptone (Difco Laboratories). Blood samples were pretreated with 50 Al Triton X-100 per ml of blood for 30 min at 30 jc. Half of each sample was filtered over a nylon membrane filter which was subsequently placed on SGA-T and incubated for 9 11 h at 37 jc to yield microcolonies. After incubation, the filter was exposed to UV light to observe a possible orange fluorescence, indicating the presence of C. glabrata. The filter was subsequently removed from the SGA- T medium and cut in five pieces. These were placed on absorbent glass fiber pads, each one impregnated with 340 Al of a buffered solution of the 4-MU substrate (4-MU-h-D-glucoside, 4-MU-N-acetyl-h-Dgalactosaminide, 4-MU-phosphate, 4-MU-pyrophosphate and 4-MU-h-D-galactoside), containing 0.1% digitonin, acting as a membrane permeabilizer and 1 mm MgCl 2. After incubation for 30 min at 30 jc, the filters were sprayed with 1.2 M sodium hydroxide and inspected under a 366-nm UV lamp. Blue fluorescent microcolonies indicated hydrolysis of a substrate. The remaining part of the sample was filtered over a second nylon membrane filter which was incubated on CHROMagar for h at 37 jc. Confirmation of presumably identified colonies relied on the germ tube test, morphology on cornmeal agar with 0.5% Tween 80, sugar assimilation tests (API 20C AUX) and agglutination with the Krusei- ColorR test. All cultures identified as C. albicans were subcultured on SGA at 45 jc to distinguish them from C. dubliniensis, which, unlike C. albicans, is unable to grow at 45 jc (Pinjon et al., 1998). 3. Results All 28 test strains of C. parapsilosis were 4- MU-h-D-glucoside positive. However, as shown in Table 1, other Candida spp. and non-candida yeasts also cleaved this substrate. To ensure specificity of the test for C. parapsilosis, the conversion of 4-MU-h-D-glucoside had to be combined with that of four additional substrates, i.e., 4-MU-Nacetyl-h-D-galactosaminide, 4-MU-phosphate, 4- MU-pyrophosphate and 4-MU-h-D-galactoside, as well as with one unelucidated reaction leading to orange fluorescence (Bauters et al., 1999). This combination allowed a full differentiation of C. parapsilosis from seven other Candida spp. and four non-candida yeasts (Table 2). The method has been applied to 71 clinical samples containing C. parapsilosis as a single species (n = 68) or in a mixed flora (n = 3). An agreement of 95.8% with conventional identification methods was obtained for the target species. Only in the three samples (4.2%) with mixed flora C. parapsilosis was overlooked due to its overgrowth Table 1 4-MU-h-D-glucoside positivity in yeast species in the enzymatic membrane filtration test Number of strains tested % Reactive strains C. parapsilosis C. albicans C. glabrata 39 8 C. krusei 25 0 C. tropicalis 26 0 C. dubliniensis 1 0 C. lusitianiae 4 25 C. kefyr 2 0 C. guilliermondii 2 0 Cr. neoformans S. cerevisiae 4 50 G. candidum 4 50 T. cutaneum 4 50

4 14 T.G.M. Bauters et al. / Journal of Microbiological Methods 53 (2003) Table 2 Combined positive and negative reactions a for the differentiation of C. parapsilosis from other medical yeasts (A) Differentiation of C. parapsilosis from other Candida spp. Reaction C. parapsilosis (n = 28) C. albicans (n = 129) C. glabrata (n = 39) C. krusei (n = 25) C. tropicalis (n = 26) C. guilliermondii (n =2) C. lusitaniae h-d-glucosidase + +/ +/ +/ N-acetyl-h-Dgalactosaminidase c + +/ b Acid phosphatase b + c +/ +/ + +/ + + Pyrophosphatase b +/ c +/ + Unknown b mechanism b,d h-d-galactosidase b + c (B) Differentiation of C. parapsilosis from other yeasts Reaction C. parapsilosis (n = 28) Cr. neoformans (n = 17) S. cerevisiae G. candidum h-d-glucosidase N-acetyl-h-Dgalactosaminidase c c + b Acid phosphatase b + c + c + + Pyrophosphatase b +/ c c Unknown c c mechanism b,d h-d-galactosidase b + c c + a +/ indicates that only part of the strains are positive. b Bauters et al. (1999). c This study and previous study (Bauters et al., 1999). d Leading to orange fluorescence. T. cutaneum C. kefyr (n =2) by C. albicans. An isolate of C. dubliniensis yielded no fluorescence in the current test. 4. Discussion The results show that h-d-glucosidase is a useful marker enzyme for the differentiation of C. parapsilosis from other medically important yeasts on a membrane filter. Within the genus Candida, the specificity of this reaction for C. parapsilosis is fair. Cross-reaction with roughly 10% of C. albicans and C. glabrata strains appears unacceptable in view of the much higher prevalence of these species in clinical practice relative to that of C. parapsilosis. When other non-candida yeasts are also taken into account, specificity for C. parapsilosis is low, as a significant number of strains of the four species tested also reacted. Smitka and Jackson (1989) already noted that h-d-glucosidase is not 100% specific for C. parapsilosis but that the four other Candida spp. they examined exhibited a lower activity. To avoid false positive results, they proposed a semiquantitative approach using spectrofluorometry in connection with cut-off values for interpretation. In the membrane filtration approach, specificity is achieved by combining five enzymatic reactions and one reaction of unknown nature. The fingerprint resulting from the combined positive and negative reactions for C. parapsilosis is different from that of the 11 other common medical yeasts tested. Unlike other experimental conditions, the present ones are not suitable for the differentiation of C. albicans from C. dubliniensis, as only 9% of C. albicans strains were positive, while no reaction was observed for a strain of C. dubliniensis. The method has the advantage of combining isolation and identification and, hence, avoiding the need for a lengthy preliminary culturing step. In addition, owing to a thorough optimization of the conditions for

5 T.G.M. Bauters et al. / Journal of Microbiological Methods 53 (2003) fluorescence development, with membrane filtration being a key factor, the test can be carried out at the microcolony level (Bauters et al, 1999). As a result, the method only takes 9 11 h versus up to >48 h for that of Smitka and Jackson (1989), which requires preliminary isolation and culturing prior to the actual fluorescence assay as well as instrumental reading as a function of time. Acknowledgements We are grateful to J. Meis (Regional Laboratory for Public Health, Nijmegen) and P. Verweij (University Hospital, Nijmegen, The Netherlands) for providing us with blood samples. References for rapid presumptive differentiation of four Candida species. J. Clin. Microbiol. 37, Bobey, D.G., Ederer, G.M., Rapid detection of yeast enzymes by using 4-methylumbelliferyl substrates. J. Clin. Microbiol. 13, Moran, P., Sullivan, D., Coleman, D., Emergence of non- Candida albicans Candida species as pathogens. In: Calderone, R.A. (Ed.), Candida and Candidiasis. ASM Press, Washington, DC, USA, pp Pinjon, D., Sullivan, D., Salkin, I., Shanley, D., Coleman, D., Simple, inexpensive, reliable method for differentiation of Candida dubliniensis from Candida albicans. J. Clin. Microbiol. 36, Smitka, C.M., Jackson, S.G., Rapid fluorogenic assay for differentiation of the Candida parapsilosis group from other Candida spp. J. Clin. Microbiol. 27, Sullivan, D., Coleman, D., Candida dubliniensis: characteristics and identification. J. Clin. Microbiol. 36, Weems, J., Candida parapsilosis: epidemiology, pathogenicity, clinical manifestations, and antimicrobial susceptibility. Clin. Infect. Dis. 14, Bauters, T.G.M., Peleman, R., Moerman, M., Vermeersch, H., De Looze, D., Noens, L., Nelis, H.J., Membrane filtration test