Candida albicans - Wiley Online Library

2849

Electrophoresis 2010, 31, 2849–2853

Andrew W. Lantz1 Bledar Bisha2 Man-Yung Tong3 Ryan E. Nelson1 Byron F. Brehm-Stecher2 Daniel W. Armstrong3 1

Department of Chemistry, Grand Valley State University, Allendale, MI, USA 2 Rapid Microbial Detection and Control Laboratory, Department of Food Science and Human Nutrition, Iowa State University, Ames, IA, USA 3 Department of Chemistry and Biochemistry, University of Texas, Arlington, TX, USA

Received March 16, 2010 Revised April 23, 2010 Accepted May 10, 2010

Short Communication

Rapid identification of Candida albicans in blood by combined capillary electrophoresis and fluorescence in situ hybridization A CE method based on whole-cell molecular labeling via fluorescence in situ hybridization was developed for the detection of Candida albicans in whole blood. Removal of potentially interfering red blood cells (RBC) with a simple hypotonic/detergent lysis step enabled us to detect and quantitate contaminating C. albicans cells at concentrations that were orders of magnitude lower than background RBC counts (7.0 109 RBC/mL). In the presence of the lysed blood matrix, yeast cells aggregated without the use of a blocking plug to stack the cells. Short (15 min) hybridizations yielded bright Candidaspecific fluorescence in situ hybridization signals, enabling us to detect as few as a single injected cell. The peak area response of the stacked Candida cells showed a strong linear correlation with cell concentrations determined by plate counts, up to 107 CFU/mL (or 1 104 injected cells). This rapid and quantitative method for detecting Candida in blood may have advantageous applications in both human and veterinary diagnostics. Keywords: Blood / Candida albicans / CE / Fluorescence in situ hybridization DOI 10.1002/elps.201000159

Several industries, including the food and pharmaceutical industries, rely on rapid microbial testing to ensure consumer or patient safety, to maintain product quality and to meet regulatory requirements. However, current culture-based methods for microbial testing are not sufficient to meet industrial demands, mainly due to their slow speed of analysis [1]. Furthermore, microorganisms require specific nutrients, cofactors and environmental conditions for growth. As a result, culture-based methods may only be able to detect a small portion of the target cells that are present in a sample [2]. Depending on sample type, high backgrounds of host cells (red blood cells (RBC), for example) or fast-growing non-target organisms may also be present that can outnumber or outcompete target cells, making pathogen detection more difficult. In this case, growth-based microbial testing may not be able to yield conclusive determinations for pathogen presence [3, 4]. The inherent limitations of growth-based microbial testing are driving the development of rapid alternatives to

Correspondence: Dr. Andrew W. Lantz, Department of Chemistry, Grand Valley State University, 351 Padnos Hall, Allendale, MI, USA E-mail: [email protected] Fax: 11-616-331-3230

Abbreviations: CFU, colony forming unit; FAM, 6-carboxyfluorescein; FISH, fluorescence in situ hybridization; RBC, red blood cell; SB3-10, caprylyl sulfobetaine (n-decyl-N,Ndimethyl-c-ammonio-1-porpanesulfonate)

& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

traditional cultural methods [5–10]. These include impedance measurements, enzyme immunoassays, molecular techniques such as the polymerase chain reaction and flow cytometry. Although these methods are promising, drawbacks may include high cost, complex methodology or low sensitivity (>103 cells), depending on the assay used [10]. CE for analysis of whole, intact microorganisms has emerged as another promising rapid approach and, with further development, may ultimately have advantages over other alternative methods in terms of speed, simplicity and cost. Recent study in CE analysis of cells has focused on using polymeric and/or surfactant additives to control the electrophoretic mobility of microorganisms within the applied electric field as well as capillary coatings to suppress cell interactions with the capillary wall [11–22]. Isoelectric focusing has also been explored for the separation of bacteria, fungi and cellular organelles [11, 19, 23, 24]. Previously, we developed a CE technique for the analysis of microbial contamination that is capable of single cell detection [25]. This method combines a cationic surfactant buffer additive (CTAB) to coat the surface of the microorganisms and effectively sweep them out of their initial sample zone with a short injection plug of zwitterionic surfactant blocking agent (caprylyl sulfobetaine (n-decylN,N-dimethyl-c-ammonio-1-porpanesulfonate), SB3-10) to Additional corresponding authors: Dr. Byron F. Brehm-Stecher

E-mail: [email protected] Professor Daniel W. Armstrong E-mail: [email protected]

www.electrophoresis-journal.com

2850

A. W. Lantz et al.

stack the cells into a single, sharp peak. The presence or absence of this peak indicates the presence or absence of cells in the sample. Additional information can be obtained by combining electrophoretic analysis of cells with a suitable method for organism-specific cell labeling. Fluorescence in situ hybridization (FISH) is a whole-cell fluorescent labeling method that enables molecular detection of specific cell types based on hybridization of short nucleic acid probes to ribosomal RNA within target cells. Briefly, cells in the sample are first permeabilized with the addition of a suitable aldehyde or alcohol fixative. Apart from permeabilization, the fixation step also kills microbial cells, inhibits DNase activity and preserves cell morphology. After fixation, cells are heated in buffer containing salt, detergent and a DNA probe specific for the organism or group of organisms to be tested for. The probe, which is end-labeled with a fluorescent reporter, penetrates the cell and binds with a 1:1 stoichiometry to its complementary sequence on the ribosome. Because most microbial cells contain between 10 000 and 100 000 ribosomes, ribosomal RNA is considered a naturally amplified target molecule. The aggregate signal from multiple probe-ribosome binding events results in bright fluorescent labeling of target cells, whereas nontarget cells are not labeled. Target cells containing thicker cell walls or fewer ribosomes may have lower staining intensities. To improve staining of these cells, the cell fixation step can be optimized for maximum probe uptake and ‘‘helper’’ probes – unlabeled probes that bind to sites flanking the target region can also be used. The helper probes bind to and modify the local structure of the ribosome, increasing the physical accessibility of the target region to the labeled probe, enhancing labeling intensity. Recently, we coupled FISH labeling with our CE-based method, facilitating rapid and specific identification of Salmonella spp. in mixed cultures via CE [26]. As a continuation of this study, we have applied this approach for the detection of Candida albicans in blood samples. Candida albicans is the most widespread cause of nosocomial fungal infections in the US hospitals, with a mortality rate of over 25%. Apart from an existing FISH-based approach using peptide nucleic acid probes [27], definitive culture-independent microbiological tests for C. albicans and related Candida spp. in blood are not currently available, hampering diagnosis and effective treatment [28, 29]. As a result, there is significant interest in the medical industry for a rapid test for C. albicans in patients’ blood. The sample preparation and instrumental setup used was similar to previously reported for CE-FISH of Salmonella [26]. Briefly, fungal cultures were grown overnight (20–22 h) at 241C in Yeast and Mold broth (Difco Laboratories, Detroit, MI). One-milliliter portions of fungal culture were pelleted for 5 min at 2000 g. The supernatant was discarded and the cell pellet was washed once and resuspended in PBS. This suspension was serially diluted (1:10) into fresh PBS to yield a series of tubes having yeast concentrations ranging from 7.9 100 to 7.9 107 CFU (CFU, colony forming units)/mL. These were pelleted, the & 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim


supernatant removed and yeast cells were resuspended in 1 mL portions of whole bovine blood (catalog ]IR1-040N, Innovative Research, Novi, MI), simulating C. albicans bloodstream infection (candidemia) at various concentrations of the pathogen. To prepare samples for CE-FISH, Candida-contaminated blood samples were pelleted and RBCs instantaneously lysed by resuspending Candida-RBC pellets in 1 mL distilled water plus 0.1% v/v Triton X-100. After lysis of the potentially interfering RBC matrix, yeast cells were pelleted, then fixed for 1 h by adding 1 mL of 60% formalin-40% absolute ethanol [26]. Fixed samples were pelleted and resuspended in a 50:50 mixture of PBS and absolute ethanol, then stored at 201C until analyzed. For analysis, 100 mL stored samples were pelleted and resuspended in an equal volume of hybridization buffer (0.9 M NaCl, 0.1 M Tris (pH 8.0), 0.1% sodium dodecyl sulfate, 10 mM EDTA), preheated to 551C and containing a FISH probe cocktail consisting of a 6-carboxyfluorescein (FAM)-labeled probe, Calb-1249, previously reported to react with C. albicans and the closely related bloodstream pathogen Candida tropicalis, 50 -GCC AAG GCT TAT ACT CGC T-30 -FAM [30], plus two unlabeled helper probes targeting either side of the Calb-1249 binding site, 50 -GGC TCC GTC AGT GTA GC-30 and 50 -AGA TTT CCC AGA CCT CTC G-30 . The Calb-1249 probe was labeled at the 30 end to minimize quenching of FAM by the 50 terminal G residue. All probes were used at a level of 10 ng/mL each and were synthesized and purified by Integrated DNA Technologies, Coralville, IA. Samples were hybridized at 551C for 15 min, and then washed for 30 min with the addition of 500 mL preheated hybridization buffer without probe, vortexing frequently. After hybridization, samples were pelleted and suspended in 0.2 mL of running buffer for analysis via CE. All CE runs were performed on a Beckman Coulter P/ACE MDQ equipped with LIF detector a 488 nm laser. The bare fused silica capillaries used were 30 cm in length (20 cm to the detector) with an id of 100 mm. New capillaries were conditioned with 1 M NaOH for 5 min, distilled water for 5 min and then running buffer for 5 min at 25 psi. Between each run, the capillary was washed with 1 M NaOH for 1 min, distilled water for 1 min followed by running buffer for 1 min. The working buffer used consisted of 1 mM Tris/0.33 mM citric acid at a pH of 7.0. The running buffer was prepared by adding 10 mg/mL CTAB to the working buffer. Sulfobetaine solutions were prepared to a concentration of 10 mg/mL with working buffer. Surfactant solutions were prepared fresh daily. All glassware was autoclaved prior to use. Initially, our previously reported 3-injection technique was used to detect the presence of C. albicans in blood samples [25, 26]. The capillary was filled with running buffer containing CTAB, and the following three injections were performed via hydrodynamic pressure at 0.5 psi: contaminated/lysed blood sample (158 nL) for 5 s, followed by a spacer plug of running buffer for 4 s and finally blocking plug of SB3-10 for 2 s. All CE runs were www.electrophoresis-journal.com

CE and CEC


RFU

A

B

45000 40000 35000 30000 25000 20000 15000 10000 5000 0

45000 30000 15000 0

0 1 2 3 4 5 6

0

1

2

3

4

5

6

0

1

2

3

4

5

6

0

1

2

3

4

5

6

4

5

6

35000 30000

RFU

25000 20000 15000 10000 5000 0

C

800 700 600 500

RFU

performed in the reverse polarity mode with an applied voltage of 3 kV. The fluorescence signal from the probe-labeled C. albicans was detected at 516 nm. Sharp, reproducible peaks indicating the presence of C. albicans cells in spiked blood samples were obtained using this method. However, the spacer and blocker plug injections are only necessary if the cells swept out of the sample zone by the cationic surfactant do not aggregate sufficiently to form a discrete, detectable peak. In these cases, the blocker plug can serve to inhibit the cells’ mobility, stacking them into a focused zone. However, we have observed that some cell types, in the presence of a proper sample matrix, may aggregate with CTAB alone, obviating the need for the blocking plug. In a previous study, numerous species of bacteria were observed to spontaneously aggregate with CTAB in the presence of nutrient broth growth media [18]. To test this effect for C. albicans in a blood matrix, the spacer and blocker plug injections were removed from the injection sequence. A single fluorescence peak with a reproducible migration time was obtained for C. albicans cells, demonstrating that the cells sufficiently focus for proper detection without the blocking agent, with the complex lysed blood matrix possibly playing a role similar to that previously observed for the nutrient broth plug [31]. This simplified method was then used for the remainder of this study (Fig. 1). Contaminated blood samples containing various levels (7.9 103–7.9 107 CFU/mL) of C. albicans cells were used to evaluate quantitative detection of this contaminant in blood via CE. Based on the volume of blood injected into the capillary for analysis (158 nL), these concentrations correspond to approximately 1–10 000 injected cells.

2851

400 300 200 100 0

D

300 300 200 100 0

250 RFU

200 150

0 1 2 3 4 5 6

100 50 0 0

1

2

3

Time (min.)

Figure 1. Schematic diagram of sample loading and fungal peak formation from lysed blood samples containing C. albicans. The lysed blood is represented here as a dark matrix, and the Candida cells as spheres. Prior to application of voltage, the sample plug (158 nL, in running buffer) containing FISH-labeled Candida cells is injected via hydrodynamic pressure (0.5 psi). Upon application of voltage, CTAB from the bulk running buffer migrates into the sample zone, coating the cells with cationic charges and inducing cathodic migration. As they travel, the Candida cells are compressed into a tight peak prior to passing in front of the detector at 3.9 min. All CE runs were performed in the reverse polarity mode with an applied voltage of 3 kV. The fluorescence signal from the probe-labeled C. albicans was detected at 516 nm.


Figure 2. Electropherograms of (A) 7.9 107 CFU/mL, (B) 7.9 106 CFU/mL, (C) 7.9 104 CFU/mL and (D) 7.9 103 CFU/mL of C. albicans recovered from lysed blood and hybridized for 15 min with the Calb-1249/helper probe cocktail. These concentrations correspond to approximately 1 104, 1 103, 10 and 1 injected cells, respectively. Insets in panels (A) and (D) display the signal. from hybridized lysed blood matrix without Candida at the corresponding RFU axis scales. Fluorescence excitation performed at 488 nm and detected at 516 nm.

Representative electropherograms of this cell concentration range show a sharp peak at a reproducible time of 3.9 min for C. albicans cells (Fig. 2). This peak was not observed for hybridized blood samples that were not spiked with C. albicans, highlighting the specificity of this method for detection of Candida in the lysed blood matrix (Fig. 2, panel www.electrophoresis-journal.com

2852


A. W. Lantz et al.

A inset). A broad fluorescent band frequently appeared after the fungi peak at 4.4 min and was also present in the unspiked blood blank (Fig. 2, panel D inset). The presence of this band in the blood-only blank indicates that it is independent of C. albicans contamination and not an artifact of yeast clumping, and possibly results from excess FISH probe bound to blood matrix components released via lysis. This peak was clearly resolved from the sharp C. albicans signal and did not interfere with fungal detection. A calibration curve (Fig. 3) was produced by plotting the fluorescence peak area of C. albicans CE signals against cell counts obtained from contaminated blood, as determined by plating. Yeast-spiked blood samples were run in triplicate to assess the method’s reproducibility. A linear relationship between peak area and cell count was observed up to 107 CFU/mL C. albicans. Concentrations beyond this point consistently produced signals that deviated below the fit line. This trend is likely due to the self-limiting effects of high yeast concentrations within the capillary, which could absorb a significant percentage of the excitation beam and/or the emitted light, interfering with the signal-to-cell number proportionality seen at lower yeast concentrations. Therefore, points for concentrations greater than 107 CFU/mL (1 104 injected cells) were not included in the line fitting. Cell concentrations between 103 and 107 CFU/mL showed good linearity with peak area, yielding an R2 value of 0.9897. From the baseline noise and calibration sensitivity of the method, we determined the detection limit of the analysis to be 2.0 103 (70.2 103) CFU/mL, whereas the quantitation limit was 5.2 103 (70.3 103) CFU/mL, both of which are below the concentration corresponding to a single injected cell. The probability of successfully injecting a Candida cell using a 7.9 103 CFU/mL sample with a 158 nL injection (corresponding to 1 injected cell) was found to be 75% with this method. Increasing the volume of lyszed blood sample injected into the capillary would increase the success rate of injecting a Candida cell, and as a result decrease the detection limit further below 2.0 103 CFU/mL. A larger injection volume can easily be accomplished by increasing the sample

The authors acknowledge the support of the Grand Valley State University S3 Program as well as the National Institutes of Health (GM53825-12 to D. W. A.). The authors also acknowledge the reviewers’ helpful comments on this manuscript. The authors have declared no conflict of interest.

References [1] Industrial Microbiology Market Review – 2nd Edition, Strategic Consulting, March 2004. [2] United States Pharmacopeia, 26th Edn, Webcon, Toronto, ON, Canada 2003, pp. 2011–2016. [3] Pratten, J., Wilson, M., Spratt, D. A., Oral Microbiol. Immunol. 2003, 18, 45–49. [4] Moldenhauer, J., Sutton, S. V. W., PDA J. Pharm. Sci. Technol. 2004, 58, 284–286. [5] Silley, P., Forsythe, S., J. Appl. Bacteriol. 1996, 80, 233–243. [6] Bussey, D. M., Tsuji, K., Appl. Environ. Microbiol. 1986, 51, 349–355.

7.0

[7] Alverez-Barrientos, A., Arroyo, J., Canton, R., Nombela, C., Sanchez-Perez, M., Clin. Microbiol. Rev. 2000, 13, 167–195.

6.0 log(Peak Area)

injection time and/or using a wider diameter capillary. The effect of injection volume on this CE-based microbial detection method has been reported and discussed previously [18, 32]. In conclusion, we have combined CE with whole-cell molecular labeling via FISH for the detection of relatively low levels of C. albicans in blood. All steps of the assay, from blood lysis and cell fixation to hybridization and analysis, could be completed in under 2 h. In situations requiring high sample throughput, the sample preparation steps may be performed in parallel to further decrease the average analysis time per sample. Together, these results indicate the potential for this method to enable rapid, quantitative and definitive determination of Candida (C. albicans and C. tropicalis) infections in blood samples, which is expected to be useful in both human and veterinary diagnostics. Future work will involve evaluation of this method in naturally contaminated blood samples or in blood cultures from suspected cases of candidemia.

5.0

[8] Johnson, R., Eur. Pharm. Rev. 1999, 4, 55–62.

4.0

[9] McCarthy, J., Genetic techniques: PCR, NASBA, hybridization and microarrays, in: Detecting Pathogens in Food, Woodhead, Cambridge 2003, pp. 241–384.

3.0 2.0

[10] Fung, D. Y. C., Overview of rapid method of microbial analysis, in: Food Microbiological Analysis New Technologies Series no. 12, Marcel Dekker, New York 1997, 1–23.

1.0 0.0 3.0

4.0

5.0

6.0

7.0

8.0

9.0

log(CFU/mL)

Figure 3. Linear relationship between C. albicans peak area signal and CFU concentration. Points with concentrations 4 107 CFU/mL were excluded from line fitting – see text for details.


[11] Armstrong, D. W., Schulte, G., Schneiderheinze, J. M., Westenberg, D. J., Anal. Chem. 1999, 71, 5465–5469. [12] Armstrong, D. W., Girod, M., He, L., Rodriguez, M. A., Wei, W., Zheng, J., Yeung, E. S., Anal. Chem. 2002, 210, 245–249.



CE and CEC

2853

[13] Armstrong, D. W., Schneiderheinze, J. M., Kullman, J. P., He, L., FEMS Microbiol. Lett. 2001, 194, 33–37.

[24] Horka, M., Horky, J., Matouskova, H., Slais, K., Anal. Chem. 2007, 79, 9539–9546.

[14] Armstrong, D. W., He, L., Anal. Chem. 2001, 73, 4551–4557.

[25] Lantz, A. W., Bao, Y., Armstrong, D. W., Anal. Chem. 2007, 79, 1720–1724.

[15] Armstrong, D. W., Schneiderheinze, J. M., Anal. Chem. 2000, 72, 4474–4476.

[26] Lantz, A. W., Brehm-Stecher, B. F., Armstrong, D. W., Electrophoresis 2008, 29, 2477–2484.

[16] Buszewki, B., Szumski, M., Klodzinska, E., Dahm, H., J. Sep. Sci. 2003, 26, 1045–1049. [17] Tsibakhashvilli, N. Y., Biomed. Chromatogr. 2002, 16, 327–331.

[27] Rigby, S., Procop, G. W., Haase, G., Wilson, D., Hall, G., Kurtzman, C., Oliveira, K., Von Oy, S., Hyldig-Nielsen, J. J., Coull, J., Stender, H., J. Clin. Microbiol. 2002, 40, 2182–2186.

[18] Rodriguez, M. A., Lantz, A. W., Armstrong, D. W., Anal. Chem. 2006, 78, 4759–4767.

[28] Gagne, J. J., Goldfarb, N. I., Expert Opin. Pharmacother. 2007, 8, 1643–1650.

[19] Klodzinska, E., Buszewski, B., Anal. Chem. 2009, 81, 8–15.

[29] Sheppard, D. C., Locas, M. C., Restieri, C., Laverdiere, M., J. Clin. Microbiol. 2008, 46, 3508–3509.

[20] Buszewski, B., Szumski, M., Klodzinska, E., Dahm, H., J. Sep. Sci. 2003, 26, 1045–1049.

[30] Kempf, V. A., Trebesisus, K., Autenrieth, I. B., J. Clin. Microbiol. 2000, 38, 830–838.

[21] Szumski, M., Klodzinska, E., Buszewski, B., J. Chromatogr. A 2005, 1084, 186–193.

[31] Tong, M. Y., Jiang, C., Armstrong, D. W., J. Pharm. Biomed. Anal. 2010, 53, 75–80.

[22] Klodzinska, E., Dahm, H., Rozycki, H., Szeliga, J., Jackowski, M., Buszewski, B., J. Sep. Sci. 2006, 29, 1180–1187.

[32] Petr, J., Jiang, C., Sevcik, J., Tesarova, E., Armstrong, D. W., Electrophoresis 2009, 30, 3870–3876.

[23] Shen, Y., Berger, S. J., Smith, R. D., Anal. Chem. 2000, 72, 4603–4607.

