Practical identification of eight medically important Trichosporon species by reverse line blot hybridization (RLB) assay and rolling circle amplification (RCA)

Abstract

We developed a reverse line blot (RLB) hybridization-, and rolling circle amplification (RCA)-based assays for the identification of Trichoporon species and evaluated them with 48 isolates that had been previously recognized as belonging to eight species (Trichosporon asahii, T. cutaneum, T. dermatis, T. domesticum, T. inkin, T. japonicum, T. jirovecii, and T. laibachii). Results were compared to those obtained with DNA sequencing of three rRNA gene loci, i.e., the internal transcribed spacer (ITS) region, D1/D2 domain of the 28S rRNA gene and intergenic spacer 1 (IGS1) region. Using species-specific, or group-specific probes targeted at the ITS region and the D1/D2 domain, the RLB assay permitted accurate species identification of all 48 isolates with 100% specificity. Species-specific RLB probes correctly assigned 45/48 (94%) of the isolates (six species) with the exception of T. dermatis and T. japonicum isolates which were not targeted by the assay. Identification of T. dermatis relied on a positive hybridization result with the group-specific probe hybridizing with T. dermatis and T. jirovecii and the absence of a signal with the T. jirovecii-specific probe. T. japonicum strains were first assigned to the T. asahii-T. japonicum group by hybridization with the two species group-specific probe and then as T. japonicum by the absence of signal with a T. asahii-specific probe. Twelve species-specific RCA probes targeting the eight species studied detected templates of all 48 Trichosporon isolates and an artificial template of T. asteroides, all with good specificity. Both RLB and RCA are potential alternatives to DNA sequencing for the identification of Trichosporon species. The RLB approach is suited for the batched simultaneous analysis of large numbers of isolates, while RCA is more appropriate for the immediate study of single isolates. Comparative costs are US$7 and US$2 per assay for the RLB and RCA methods, respectively.

Introduction

Trichosporon species are important emerging fungal pathogens associated with high mortality (42–80%). Typically, infection occurs in severely immunocompromised patients, including those receiving chemotherapy for underlying hematological malignancies and organ transplantations [1–4].

Species identification of Trichosporon clinical isolates is important to help to define species-specific clinical associations and geographic variation in their epidemiology. In addition, certain Trichosporon species may be resistant or less susceptible to antifungal drugs such as echinocandin and/or amphotericin B [1,4–7]. Since the introduction of molecular identification techniques for fungi, the nomenclature within this genus has been significantly revised. While Trichosporon beigelii was considered to be the most important etiologic agent [8], at least 16 species have now been reported to be associated with human disease [1,5,6,9–11]. DNA sequencing-based methods have primarily been required to delineate the taxonomy and epidemiology of Trichosporon species [1,6,9,12,13]. We previously characterized 48 Trichosporon strains from China belonging to eight species (Trichosporon asahii, T. cutaneum, T. dermatis, T. domesticum, T. inkin, T. japonicum, T. jirovecii and T. laibachii) by sequencing three fungal rRNA gene loci, i.e., the internal transcribed spacer (ITS), D1/D2 domain of the 28S rRNA gene, and intergenic spacer 1 (IGS1), regions. The study of nucleotide polymorphisms within the ITS and D1/D2 regions have especially proved to be sensitive techniques in species identification [12].

However, DNA sequencing may be slow (turnaround time of 2–3 days), expensive, and not easily adaptable to use in the routine clinical laboratory. Two PCR-based approaches, the PCR-reverse line blot (RLB) and rolling cycle amplification (RCA) methods, have been shown to be accurate and practical alternatives to DNA sequencing in the identification/detection of medically important fungal species [14–20]. RCA methodology is particularly well suited for the sensitive detection of single nucleotide polymorphisms (SNPs) and can discriminate between genetically closely-related species, and within species [14,20]. In the present study, by using species-, and group-specific probes targeted against members of the genus Trichosporon, we designed and evaluated a (i) RLB-, and (ii) RCA-based assay for species identification of the same 48 strains previously-characterized by combined ITS/D1-D2 sequencing [12].

Materials and methods

Trichosporon strains and DNA extraction

The identification of the 48 Trichosporon isolates (Supplementary Table 1, available online at http://informahealthcare.com/doi/abs/10.3109/13693786.2012.723223) employed in this study had been previously determined by combined ITS and D1-D2 sequencing [12]. DNA extraction was performed as described before [12] and stored at −20°C prior to use.

Primers for PCR, RLB and RCA

For all isolates, primer pairs ITS1/ITS4 and F63/R635 [12] were used to amplify the ITS region and D1/D2 domain, respectively. For PCR in preparation for the RLB assay, 5′-end biotin-labelled primers, ITS1b/R635b, (Table 1) were employed. For the RCA assay, all four primers (ITS1/ITS4 and F63/R635) were non-labeled. Two additional primers (RCA primer 1 and RCA primer 2; Table 1) were specifically designed to bind the linker region of the padlock probes [14,21]. All primers were synthesized at Beijing AuGCT Biotechnology Co. Ltd. (Beijing, P. R. China).

Table 1.

Primers, probes and artificial template used in the present study.

Oligonucleotide	Specificity	Target gene	GenBank accession no.	Sequence (5′–3′)*
Primers
ITS1/ITS1b†	Panfungal	18S rRNA gene	HM802135	4 TCCGTAGGTGAACCTGCG 21
ITS4	Panfungal	28S rRNA gene	HM802135	534 TCCTCCGCTTATTGATATGC 515
F63	Panfungal	28S rRNA gene	EU882103	1 GCATATCAATAAGCGGAGGAAAAG 24
R635/R635b†	Panfungal	28S rRNA gene	EU882103	640 GGTCCGTGTTTCAAGACG 623
RCA primer 1‡	–	Padlock probe linker region	–	ATGGGCACCGAAGAAGCA
RCA primer 2‡	–	Padlock probe linker region	–	CGCGCAGACACGATA
RLB probes†
12AP	Panfungal	ITS region	AF455524	211 CCAAGAGATCCGTTGTTGAAAG 190
23SP	Panfungal	ITS region	AF455524	271 GTGAATCATCGARTCTTTGAACG 293
Tri-ITS-1	T. asahii, T. domesticum and T. japonicum	ITS region	AF444473	31 GTGATTGCCTTAATTGGCTTATAAC 55
Tri-ITS-2	T. asahii, T. inkin and T. japonicum	ITS region	AF444473	93 ACGCAAGTCGAGTATTTTTACAAAC 117
Tri-ITS-3	T. asahii and T. japonicum	ITS region	AJ864867	10 GTGATTGCCTTTATAGGCTTATAAC 34
Tri-ITS-4	T. cutaneum	ITS region	AF444325	89 TTCGGTCAATTGATTTTACAAAC 111
Tri-ITS-5	T. dermatis and T. jirovecii	ITS region	AF444437	88 CTCCGGTCAATTACTTTACAAAC 110
Tri-ITS-6	T. jirovecii	ITS region	AF444437	373 GAGTTAGCGTGTTTAACTTGTCGAT 397
Tri-ITS-7	T. dermatis and T. jirovecii	ITS region	AF444437	385 TTAACTTGTCGATCTGGCGTA 405
Tri-ITS-8	T. domesticum	ITS region	AF444422	85 TTGAATCTTCGGATTCGATTTTATACAAA 113
Tri-ITS-9	T. domesticum and T. laibachii	ITS region	AF444422	377 AAAGAGTTAGCAAGTTGAACTATTGCTAT 405
Tri-ITS-10	T. inkin	ITS region	AF444420	31 GTGATTGCCTTTACAGGCTTAACTA 55
Tri-ITS-11	T. laibachii	ITS region	AF444421	84 TTGAATCTCTGATTCAATTTTACAAAC 110
Tri-ITS-12	T. laibachii	ITS region	AF444421	31 GTGATTGCCATCTTGGCTTAAAC 53
Tri-D1D2-1	T. cutaneum, T. dermatis and T. jirovecii	D1/D2 domain	AF105398	181 GCTTGATACGACGACCAGTGCTCT 204
Tri-D1D2-2	T. cutaneum, T. dermatis, T. domesticum and T. jirovecii	D1/D2 domain	AF105398	405 ATTCAGCTGGTTCTTCCAGTCTACT 429
Tri-D1D2-3	T. dermatis and T. domesticum	D1/D2 domain	EU559351	413 GATTCAGCTAGTTCTTCTAGTCTACTTCC 441
Tri-D1D2-4	T. domesticum	D1/D2 domain	AF189874	181 ACTTGACACAACAATCAGTGCTCT 204
Tri-D1D2-5	T. asahii	D1/D2 domain	AF337949	535 GGCCGGCCTTCGGGCACGTT 554
Tri-D1D2-6	T. asahii, T. inkin and T. japonicum	D1/D2 domain	AF308657	407 TCAGCCAGTTCTGCTGGTCTACT 429
Tri-D1D2-7	T. domesticum, T. inkin and T. laibachii	D1/D2 domain	AJ749822	566 GGCCGGGGTTCGCCCACGTT 585
Tri-D1D2-8	T. laibachii	D1/D2 domain	AJ507664	157 ACTTTACACAATCATCAGTGCTCT 180
Tri-D1D2-9	T. laibachii	D1/D2 domain	AJ507664	381 ATTCAGCCGGTCTTCGGTGTACT 403
RCA padlock probes‡
Tri-jap	T. japonicum	ITS region	AF444473	51 TAAGCCAATTAAGGCAATCACTAATG 26 –linker-74 ACAGGTGTAAGTGGATATAGTTA 52
Tri-cut	T. cutaneum	ITS region	AF444325	354 ACTGGCAGCGCCCAA 340 –linker-366 GCGAGCCAGGCT 355
Tri-der-a	T. dermatis	ITS region	EU559351	441 GGAAGTAGACTAGAAGAACTAGCTGAATCC 412 –linker-457 TGTTGACCCGTTCAAT 442
Tri-dom-a	T. domesticum	ITS region	AF444454	393 AACTTGCTAACTCTTTTAAGAGGAGCC 367 –linker-411 GCCAGATAGCAATAGTTC 394
Tri-asa-a	T. asahii	D1/D2 domain	AF337949	540 CCGGCCATAAAGGCGA 525 –linker-552 CGTGCCCGAAGG 541
Oligonucleotide	Specificity	Target gene	GenBank accession no.	Sequence (5′–3′)*
Tri-asa-b	T. asahii	D1/D2 domain	AF337949	547 CCGAAGGCCGGCCA 534 –linker-563 CTAAGCTCGAACGTGC 548
Tri-der-b	T. dermatis	D1/D2 domain	EU559351	494 TACATTCCTACTATCTTTATCCACCGG 468 –linker-505 CCCGGGGAGC 495
Tri-dom-b	T. domesticum	D1/D2 domain	AF189874	510 TAGGCTATAACACTCCCGAGGGA 488 –linker-528 ACCCAGTGTATGTGAACC 511
Tri-ink	T. inkin	D1/D2 domain	AJ749822	519 CAGGCTATAACACTTCCGGAGAAG 496 –linker-537 ACCCAGTGTATGTGACAG 520
Tri-jir-a	T. jirovecii	D1/D2 domain	AF105398	130 TCCAGCACGGAAAACACG 113 –linker-148 CAAGGGACTTAGATACGA 131
Tri-jir-b	T. jirovecii	D1/D2 domain	AF105398	214 CGTGTATCACAGAGCACTGGTC 193 –linker-233 AACAACTCGACTCGTAGAA 215
Tri-lai	T. laibachii	D1/D2 domain	EU559352	431 ACCGAAGACCGGCTGAAT 414 –linker-450 ACCCGTTTAAAGGAAGTAC 432
Tri-ast§	T. asteroides	D1/D2 domain	AF075513	490 GAAGTCACATTCCTACTACCTTTATCCAC 462 –linker-507 GCTATAACACTTCCGGG 491
Artificial template§
T. asteroides-AT§	T. asteroides	D1/D2 domain	AF075513	457 GTCCGGTGGATAAAGGTAGTAGGAATGT GACTTCCCCGGAAGTGTTATAGCCTATT 512

Abbreviations: D1/D2 domain, D1/D2 domain of 28S rRNA gene; ITS region, internal transcribed spacer region; RLB, reverse line blot; RCA, rolling cycle amplification. *Numbers represent the numbered base positions where the oligonucleotide sequences start or finish (commencing at point 1 of the corresponding gene GenBank sequence). †Primers ITS1b and R635b were based on previously published ITS1 and R635, respectively, with 5′ end biotin labeled for PCR in preparation for RLB. All RLB probes were 5′-hexylamine modified. ‡All RCA padlock probes were with 5′-end phosphorylation. 5′ and 3′ binding arms of the probes derived from reference GenBank sequence and their positions are showed. These are joined by the 64 bp linker region including (i) binding site of RCA primer 1 to the padlock probe (reverse compliment sequence of RCA primer 1), generating a long single-stranded DNA; (ii) binding site of RCA primer 2 (with the same sequence of RCA primer 2), primer binds to nascent single-stranded DNAs as their binding sites become available; and (iii) the non-specific linker region. §An additional probe Tri-ast was designed to identify T. asteroides, which has high sequence similarity with T. japonicum studied (1 bp difference each within the ITS region and D1/D2 domain, respectively). As no T. asteroides isolates were available for study, an artificial template ‘T. asteroides-AT’ was used along with the 48 Trichosporon isolates to test the sensitivity and specificity of the probe Tri-ast.

Oligonucleotide probes and PCR for RLB

Oligonucleotide probes for the RLB assay were designed to achieve maximum specificity according to published protocols [22]. Six species- and eight group-specific probes were designed to target the ITS region, including two panfungal oligonucleotide probes, i.e., 12AP and 23SP (Table 1) [18]. In addition, four species- and five group-specific probes were directed at nucleotide polymorphisms in the D1/D2 domain of Trichosporon spp. Thus, a total of 23 probes were employed in the RLB assay (Table 1), all of which were 5′-hexylamine modified and synthesized by Beijing AuGCT Biotechnology Co. Ltd.

In preparation for the RLB assay, each PCR reaction (total volume 30 μl) contained: 5 μl template DNA, 10 × PCR buffer (Qiagen, Doncaster, Australia); 0.2 mM each dNTP (Roche Diagnostics, Mannheim, Germany), 1.5 U HotStar Taq polymerase (Qiagen); 0.5 μM each of forward primer and reverse primers and biology grade water (Eppendorf, North Ryde, Australia). Amplification was performed in a Mastercycler gradient thermocycler (Eppendorf, GmbH, Germany), under cycling conditions previously described [12] and consisted of 10 min at 95°C for initial activation, 30 cycles of 30 sec at 95°C for denaturing, 30 sec at 50°C for annealing, and 30 sec at 72°C for extension, followed by 7-min final extension at 72°C. After amplification, equal volumes of ITS and D1/D2 PCR- amplified products (20 μl each) were added to 160 μl 2 × SSPE (Amresco) in 0.1% SDS (Sigma), then denatured at 100°C for 10 min, and cooled immediately on ice. The RLB procedure was then performed as previously described [18,22].

Design of padlock probes and RCA

Initially 14 species-specific RCA padlock probes to detect and identify Trichosporon species were designed using established methods to target nucleotide polymorphisms (including single nucleotide polymorphisms [SNPs]) within the ITS region and D1/D2 domain (see Table 1) [14,21,23]. However, preliminary experiments showed that the ‘T. cutaneum-specific’ and ‘T. inkin-specific’ ITS-targeted probes failed to detect the species. These two probes were not further used in the study and hence a total of 12 species-specific probes were employed (Table 1). Briefly, the probes were 91–113 bp in length, consisting of (i) a 5′-binding arm region, (ii) a 64-bp linker region, and (iii) a 3′-binding arm region (Table 1). The 5′- and 3′-binding arms of the probes are complementary to the 5′ and 3′ termini of the target sequence in reverse. To optimize the binding to target DNA, probes were designed with a minimum of secondary structure and with a T_m of the 5′-end probe binding arm greater than the temperature used for probe ligation (65 ± 1°C). To increase the specificity, the 3′-end binding arm was designed to have a T_m (51–56°C) below the ligation temperature. Careful attention was paid to the linker region for each point mutation-specific probe to (i) minimize similarity to those mutations closely located to the mutation of interest and (ii) to allow primer binding during RCA and amplification of the probe-specific signal. Binding sites of RCA primers 1 and 2 were designed within 64 nucleotides linker region as previously described [14,21,23].

In addition, a T. asteroides-specific probe (Tri-ast) was designed, targeting a 55-bp artificial template of the species (T. asteroides-AT) derived from Tri-ast target region of T. asteroides sequences AF075513 accessed from the GenBank database (Table 1). An artificial template for this species was constructed because we were unable to obtain T. asteroides isolates for study. All padlock probes were synthesized by Beijing AuGCT Biotechnology Co. Ltd.

For PCR in preparation for RCA analysis, the ITS regions and D1/D2 domains of the test organisms were amplified as described for preparation for the RLB assay using the primer pairs ITS1/R635. Procedures for ligation of the padlock probe, exonucleolysis and hyperbranched RCA reactions were performed as previously described [14,21,23].

Results

Species identification by RLB

The panfungal primers ITS1b/ITS4 and F63/R635b amplified the ITS region and D1/D2 domain, respectively, of all 48 study isolates. All species-, and group-specific, RLB probes were 100% specific for their target species or species group. The ITS-targeted panfungal probes 12AP and 23SP (Table 1) hybridized with amplified DNA of all isolates (Fig. 1).

Fig. 1.

Reverse line blot assay results of eight Trichosporon species studied. The left column shows the probe list (see Table 1). Lanes 1–28 (T. asahii), PUMCHBY12-PUMCHBY19, PUMCH30401-PUMCH30410, PUMCH5Z6443, PUMCH5Z6527, PUMCH6W5203, PUMCH6Z10766, PUMCH6Z2782, PUMCH6Z6579, PUMCH6Z8369, PUMCH6Z9690, PUMCH7R7615, PUMCH7Z102; Lane 29 (T. cutaneum), PUMCHBY28; Lanes 30–33 (T. domesticum), PUMCHBY20-PUMCHBY22, PUMCHBY24; Lane 34 (T. inkin), PUMCHBY23; Lanes 35–36 (T. japonicum), PUMCHBY11, PUMCHBY27; Lane 37 (T. dermatis), PUMCHBY24; Lanes 38–39 (T. jirovecii), PUMCH6Z2374 PUMCH6Z10950; Lane 40 (T. laibachii), PUMCHMC31.

Species-specific RLB probes correctly identified 45 of 48 (94%) Trichosporon isolates as compared with DNA sequencing [12] (Table 2). The exceptions were the non-target species, i.e., T. dermatis and T. japonicum, and as a result T. dermatis PUMCHBY24 and two T. japonicum isolates were not assigned to species. Since we were unable to design ‘T. dermatis’- or ‘T. japonicum’-specific RLB probes, species identification of ‘T. dermatis’ relied on obtaining a positive hybridization result with the group-specific probe Tri-ITS-5 (specifically hybridizing with T. dermatis and T. jirovecii; Table 1) and the subsequent absence of a positive signal using the T. jirovecii-specific probe, Tri-ITS-6 (Figs. 1 and 2). Similarly, T. japonicum strains PUMCHBY11 and strain PUMCHBY27 were first assigned to the T. asahii-T. japonicum group by hybridization with probe Tri-ITS-3 (T. asahii and T. japonicum group-specific) and then as T. japonicum by absence of a probe signal with the T. asahii-specific probe, Tri-D1D2-5 (Figs. 1 and 2).

Table 2.

Comparison of species ID by three-locus sequencing, RLB and RCA amongst eight Trichosporon species studied.

Strain ID by three-locus sequencing*	No. of isolates (IGS1 genotype*)	Strain ID by RLB	Strain ID by RCA
Trichosporon asahii	36 (4)	T. asahii	T. asahii
Trichosporon cutaneum	1 (1)	T. cutaneum	T. cutaneum
Trichosporon dermatis	1 (1)	T. dermatis†	T. dermatis
Trichosporon domesticum	4 (1)	T. domesticum	T. domesticum
Trichosporon inkin	1 (1)	T. inkin	T. inkin
Trichosporon japonicum	2 (2)	T. japonicum†	T. japonicum
Trichosporon jirovecii	2 (2)	T. jirovecii	T. jirovecii
Trichosporon laibachii	1 (1)	T. laibachii	T. laibachii

Abbreviations: ID, identification; IGS1, intergenic spacer 1 region; RLB, reverse line blot hybridization assay; RCA, rolling circle amplification. *Genotypes within species were further assigned based on IGS1 sequences as previously described [12]. †Identification of T. dermatis relied on obtaining a positive hybridization result with the group-specific probe which hybridizes with T. dermatis and T. jirovecii and subsequent absence of a RLB signal using a T. jirovecii-specific probe. Similarly, T. japonicum strains were first assigned to the T. asahii-T. japonicum group by hybridization with a combined T. asahii and T. japonicum group-specific probe and then as T. japonicum by absence of a probe signal with a T. asahii-specific probe (Figs. 1 and 2).

Fig. 2.

The in silico results obtained by reverse line blot (RLB) assay for identification of Trichosporon species studied. RLB probes proposed for the species identification of eight medically important Trichosporon species. The black squares represent positive signals on the RLB membrane. The left column shows the probe list (see Table 1) and the top row shows the species identity.

For the eight Trichosporon species tested in this study, we proposed a set of nine RLB probes (eight species- specific probes and one panfungal probe as an internal ‘positive’ control) for species identification (Fig. 2).

Species identification by RCA

The accumulation of double-stranded DNA was detected by staining with Sybr Green I. RCA signals obtained for the target species are shown as exponential increases in fluorescence (Fig. 3). The 12 RCA probes correctly detected their respective DNA templates with 100% specificity and were 100% concordant with DNA sequencing (Table 1 and Supplementary Table 1, available online http://informahealthcare.com/doi/abs/10.3109/13693786.2012.723223). Probe signals were obtained only with matched template-probe mixtures, with no RCA signal observed when templates from isolates of non-targeted species were present. In contrast to results from the RLB assay, T. dermatis and T. japonicum were identified to species level by species-specific RCA probes (Fig. 3).

Fig. 3.

Rolling circle amplification (RCA) results of isolates PUMCHBY11, PUMCHBY27 and of the artificial template T. asteroides-AT. RCA results monitored by the RotorGene 6000 real-time PCR machine (Corbett research). The experiment was conducted using the T. japonicum-specific RCA probe Tri-jap in (a) and T. asteroides-specific probe Tri-ast in (b) tested on all 48 Trichosporon study isolates and the artificial template T. asteroides-AT. The accumulation of double-stranded DNA was detected by staining with Sybr Green I. RCA signals indicating the species identified are shown as exponential increases in fluorescence. Ligation-mediated RCA with matched templates [DNA from T. japonicum isolates PUMCHBY11, PUMCHBY27 in (a) and artificial template T. asteroides-AT in (b)] produced ‘positive signals’. Other templates showed an absence of signal (labeled as ‘Negative Signals’).

In experiments using the artificial template T. asteroides- AT and probe Tri-ast, RCA analysis yielded a signal but produced no signal with all of the 48 Trichosporon isolates studied as illustrated by the negative signals in Fig. 3b.

Discussion

In this study, we evaluated two non-DNA sequencing-based approaches, the RLB and RCA methods, for the identification of 48 Trichosporon isolates representing eight clinically-relevant species in China.

Although the RLB assay correctly assigned all 48 Trichosporon isolates to species, with 100% specificity, species-specific RLB probes only correctly assigned 94% of isolates belonging to six species. Of note, we were not able to design ‘T. japonicum-specific’ and ‘T. dermatis-specific’ RLB probes. Instead, the identification of these two species required the use of group-specific probes – Tri-ITS-5 (T. dermatis and T. jirovecii group-specific) and Tri-ITS-3 (T. asahii and T. japonicum group-specific), followed by the absence of a positive signal using the T. jirovecii-specific probe (Tri-ITS-6) and T. asahii-specific probe (Tri-D1D2-5), respectively. The inability to successfully identify T. dermatis or T. japonicum with species-specific RLB probe is unexplained. Approximately 2% sequence dissimilarity was observed between the ITS and D1/D2 domain sequences of T. dermatis and T. jirovecii. Yet some of these polymorphic sites contained only a SNP and thus may have been unsuitable as RLB-based probe hybridization sites [22]. In addition, preliminary experiments showed cross-hybridization of the ‘T. dermatis- specific’ probe with other Trichosporon species (including the most common species T. asahii, data not shown). Sequence similarity within the ITS region, as well as in the D1/D2 domain between T. asahii and T. japonicum was also high with only 4-bp differences observed. Only the T. asahii-specific probe, but not the ‘T. japonicum-specific’ probe, was successfully designed.

As an alternative approach, RCA has been established as being highly sensitive for detecting SNPs in fungi [14,19,20]. In the present study, RCA assay not only successfully identified all 48 Trichosporon isolates to species, but also correctly identified T. dermatis and T. japonicum and distinguished these species from T. jirovecii and T. asahii, respectively. Although species outside of those tested were not available for analysis in the present study, the return of a positive RCA signal using the artificial template T. asteroides-AT and the T. asteroides-specific probe, Tri-ast, was a proof of the concept that such probe-template combinations may be used with good results to establish working molecular models of probe-based RCA assays [14,23].

In the present study, the RLB assay had a turnaround time of ˜1.5 days [17,18,22] and the RCA method, ˜0.5 days [19–21] (the time needed for the isolation and preliminary identification of yeast isolates was not included). Both approaches were less time-consuming than DNA sequencing methodology which requires ˜2–3 days for a definitive result. DNA sequencing following PCR has been used successfully to detect and identify fungal pathogens directly from clinical samples [18,24]. The PCR/RLB hybridization approach has also been reported to accurately detect pathogenic fungi in clinical specimens [16,18], yet the ability of either RLB or RCA assay to detect Trichosporon species from clinical specimens has not been studied.

Test costs are likely to influence decisions as to their implementation in clinical laboratories. The set-up costs of the RCA assay are relatively high (US$300 per probe), but a typical commercial batch of probes provides sufficient material for up to 5000 assays. Thus, running costs of RCA are no more than US$2 per assay (for consumables) [19,21]. The costs of the RLB method are approximately US$7 per isolate (for consumables), compared to US$14–35 for DNA sequencing [17]. Hence, either RLB or RCA may be employed as a potential alternative tool to DNA sequencing for the identification of Trichosporon species in the clinical laboratory. The RLB approach is suited for the simultaneous analysis of a large number of isolates (up to 43) whereas RCA assays can be used to identify isolates, either as the need arises or in batches. RCA is better suited to detect SNPs or polymorphisms which may differ between species by only a few bp. In the present study, although only 23 RLB probes and 12 RCA padlock probes were utilized for identification of eight Trichosporon species, incorporating more probes will extend the target range of species. This flexibility allows laboratories to customize the assays to meet their specific requirements according to local epidemiology. RLB and/or RCA assays have been successfully applied in the identification of various fungal pathogens including Aspergillus, Candida, Cryptococcus and Fusarium [15–20]. The RLB assay may also be adapted for intra-species genotyping by designing and incorporating ‘genotype-specific’ probes into the blot as has been demonstrated for Fusarium spp. [17].

Limitations of the present study include the relative small number of species tested since only isolates representing eight species were available. Other pathogenic Trichosporon species such as T. dohaense, T. faecale and T. mycotoxinivorans [1,6,10] may not able to be identified to species level using the probes in this study. T. mycotoxinivorans has been reported as an emerging pathogen in cystic fibrosis [10]. Hence assays that target fungal pathogens should take into consideration the context in which they will be applied.

In conclusion, two simple and practical molecular assays, RLB and RCA were evaluated and both assays obtained accurate results for species identification of Trichosporon fungi. The methods can be extended to detect other Trichosporon species not included in the study.

Supplementary material available online

Supplementary Table 1.