Identification of Dermatophytes by an Oligonucleotide Array

Abstract

Species of dermatophytes are classified into three anamorphic (asexual) genera, Epidermophyton, Microsporum, and Trichophyton. Conventional methods used to identify dermatophytes are often lengthy and may be inconclusive because of atypical microscopic or colony morphology. Based on the internal transcribed spacer 1 (ITS-1) and ITS-2 sequences of the rRNA genes, an oligonucleotide array was developed to identify 17 dermatophyte species. The method consisted of PCR amplification of the ITS regions using universal primers, followed by hybridization of the digoxigenin-labeled PCR products to an array of oligonucleotides (17- to 30-mers) immobilized on a nylon membrane. Of 198 dermatophyte strains and 90 nontarget strains tested, the sensitivity and specificity of the array were 99.5% and 97.8%, respectively. The only strain not identified (Microsporum audouinii LMA 597) was found to have a nucleotide insertion at the ITS-2 region where the probe was designed. Two nontarget strains, Microsporum equinum LMA 40396666 and Trichophyton gourvilii var. intermedium CBS 170.65, were misidentified as Microsporum canis and Trichophyton soudanense, respectively. Sequence analysis of the ITS regions revealed that the two misidentified strains displayed high sequence homology with the probes designed for M. canis and T. soudanense, respectively. The present method can be used as a reliable alternative to conventional identification methods and can be completed with isolated colonies within 24 h.

Routine procedures for dermatophyte identification rely on an examination of strain characteristics and microscopic morphology (43). Many typical isolates of common dermatophytes can be identified directly from primary isolation medium. However, morphological and physiological characteristics may vary; the phenotypic features can be influenced by factors such as temperature variation and medium (43). Frequently, the absence of reproductive structures of some strains on solid medium poses another problem, since these structures are important for conventional identification methods.

Molecular approaches have been developed to provide more rapid and accurate alternatives for dermatophyte identification. These methods include restriction fragment length polymorphism analysis (4, 17, 22), sequencing of the large-submit rRNA gene (34) and protein-encoding genes (19), gene-specific PCR (16, 17), random amplification of polymorphic DNA (20, 33), PCR fingerprinting and amplified fragment length polymorphism analysis (12, 13), and dot blot hybridization (6). Recently, several molecular studies have focused on the internal transcribed spacer (ITS) region of the rRNA gene. Sequence analysis of the ITS regions has proven to be a useful tool for phylogenetic delineation and for the identification of some dermatophytes (10-13, 15, 28, 29, 40, 46).

In recent years, DNA array technology has been used to identify a variety of fungi (14, 45). Wu et al. (45) previously developed an array consisting of 33 probes designed from the 18S rRNA gene to detect airborne fungi. An array was recently developed in our laboratory to identify a wide spectrum of medically relevant molds including several species of dermatophytes (14). However, several important dermatophyte species such as Epidermophyton floccosum, Microsporum canis, and Microsporum audouinii were not included in our previous study (14). In addition, a common probe was used to identify Trichophyton rubrum, Trichophyton soudanense, and Trichophyton violaceum; i.e., the three important species could not be differentiated. Trichophyton mentagrophyte, Trichophyton tonsurans, and Trichophyton schoenleinii also shared a common probe and could not be differentiated (14).

The aim of this study was to develop an oligonucleotide array to identify 17 dermatophyte species of clinical relevance. The species names used in this study are according to current taxonomy used in the lists of cultures of the Centraalbureau voor Schimmelcultures (CBS) (Utrecht, The Netherlands) and the American Type Culture Collection (ATCC) (Manassas, VA).

MATERIALS AND METHODS

Fungal strains.

A total of 67 reference strains and 131 clinical isolates belonging to 17 target species (i.e., species that could be identified by the array) of dermatophytes were analyzed (Table 1). Reference strains were obtained from the Bioresources Collection and Research Center (BCRC, Hsinchu, Taiwan), ATCC, CBS, and the Belgian Coordinated Collections of Microorganisms (IHEM) (Brussels, Belgium). Clinical isolates were isolated from a variety of specimens including skin, nail scrapings, and hair fragments. Clinical isolates with the prefix LMA, LM, or NCKU were obtained from the Laboratory of Parasitology and Mycology, Angers University Hospital (Angers, France); the Mycology Reference Centre, School of Biochemistry and Microbiology, University of Leeds (Leeds, United Kingdom); and the National Cheng Kung University Medical Center (Tainan, Taiwan), respectively. Identification of clinical isolates to the species level was accomplished by using established procedures, including microscopic and macroscopic characteristics (5, 18, 24, 25). In addition, a collection of 90 nontarget strains (i.e., species that could not be identified by the array) representing 84 different fungal species (see Table S1 in the supplemental material) was used for specificity testing of the oligonucleotide array. For DNA extraction, fungal strains were grown on Sabouraud dextrose agar (Difco, Detroit, MI) and incubated at 30°C until there was evidence of hyphal growth.

TABLE 1.

Dermatophyte strains used in this study and results of identification by array hybridization

Species (anamorph/teleomorph)	Reference strains	Clinical isolate(s)a	No. of discrepant strains/total no. of strains
E. floccosum	ATCC 9646, ATCC 52066	NCKU 1862, NCKU 1973, NCKU 2827	0/5
M. audouinii	ATCC 10008, ATCC 10216, ATCC 42558	LMA 597, LMA 854, LMA 971, LMA 941276, LMA 50002007, LMA 50500271, LMA 50501240, LMA 50600031, LMA 59401341, LMA 59700523	1/13
M. canis/Arthroderma otae	ATCC 10214, CBS 282.63, CBS 643.74	LM 40, LMA 720, LMA 921, LMA 922, LMA 94595, LMA 94604, LMA 5990586, LMA 50000068, LMA 50000070, LMA 50000071, LMA 50000091, LMA 50200087, LMA 50401206, NCKU 1200, NCKU 3401, NCKU 3404, NCKU 3486	1/20
M. cookeib/Arthroderma cajetanum	CBS 129.67, CBS 202.66, CBS 227.58, CBS 228.58, IHEM 14178, IHEM 14180	LMA 94548	0/7
M. ferrugineum	ATCC 11992, ATCC 42560, ATCC 52060	NCKU 1582, NCKU 1627, NCKU 2513, NCKU 3385	0/7
M. gallinae/Arthroderma grubyi	CBS 133.89, CBS 215.38, CBS 221.55		0/3
M. gypseumc/Arthroderma gypseum	BCRC 30542, CBS 161.69, CBS 286.63, CBS 311.61	LMA 961, LMA 90603, LMA 93629, LMA 94446, LMA 911379, LMA 981320, LMA 9801190, LMA 50401101, NCKU 2001, NCKU 3386	1/14
M. nanum/Arthroderma obtusum	CBS 314.54, CBS 321.61, CBS 569.80, CBS 632.82, IHEM 19092, IHEM 21111		0/6
M. persicolor/Arthroderma persicolord	CBS 421.74, CBS 468.74, CBS 871.70	LMA 9076, LMA 92136, LMA 95974, LMA 951741, LMA 40401293, LMA 50401174, LMA 59200806, LMA 59500974, LMA 59900598	0/12
T. mentagrophytes	BCRC 32066, CBS 160.66, CBS 361.62	LM 48, LM 49, LMA 604, LMA 617, LMA 90791, LMA 94677, LMA 951932, LMA 5002007, LMA 50300859, LMA 50300894, LMA 50400833, LMA 50401331, LMA 50401377, LMA 59700858, LMA 59801326, NCKU 3361, NCKU 3389, NCKU 3534	0/21
T. mentagrophytes var. erinaceie	CBS 511.73, IHEM 15931, IHEM 20118		0/3
T. mentagrophytes var. interdigitale	ATCC 200099	LMA 214, LMA 627, LMA 90592, LMA 91309, LMA 951682, LMA 50100370	1/7
T. rubrum	ATCC 28188, BCRC 32805, BCRC 32806	LMA 559, LMA 569, LMA 585, LMA 590, LMA 606, LMA 610, LMA 679, LMA 685, LMA 9025, LMA 90412, LMA 96432, LMA 97858, LMA 9500311, NCKU 1213, NCKU 3399, NCKU 3591	0/19
T. schoenleinii	ATCC 22775, ATCC 22776, CBS 174.40		0/3
T. soudanensef	ATCC 24583, ATCC 52936, ATCC 64654	LMA 831, LMA 835, LMA 966, LMA 979, LMA 9532, LMA 96248, LMA 951336, LMA 50401328, LMA 50500572, LMA 50500769, LMA 50500944, LMA 50600067, LMA 59900985	6/16
T. terrestre/Arthroderma quadrifidumg	CBS 117.61, CBS 464.62, CBS 465.62, CBS 567.94	LMA 92347, LMA 93293, LMA 7971445, LMA 50200212, LMA 59901091	0/9
T. tonsurans	ATCC 10217, ATCC 10220, ATCC 56186	LMA 856, LMA 857, LMA 50500106, LMA 50500764, LMA 50600039, NCKU 3059, NCKU 3156, NCKU 3166, NCKU 3364, NCKU 3383, NCKU 3549	1/14
T. verrucosum	ATCC 28203, ATCC 28943, ATCC 38485 IHEM 5480, IHEM 17755	LMA 40204990, LMA 50200673	2/7
T. violaceum	ATCC 24787, ATCC 28944, ATCC 62858	NCKU 3164, NCKU 3264, NCKU 3372, NCKU 3395, NCKU 3415	0/8
Total no.	67	131	13/197

^aStrains producing different identifications by array hybridization are underlined, and strains not identified by array hybridization are in boldface type. Strains with the prefix LMA, LM, or NCKU were obtained from the Laboratory of Parasitology and Mycology, Angers University Hospital (Angers, France); the Mycology Reference Centre, School of Biochemistry and Microbiology, University of Leeds (Leeds, United Kingdom); and the National Cheng Kung University Medical Center (Tainan, Taiwan), respectively.

^bStrains CBS 129.67, CBS 202.66, and CBS 227.58 were received as Microsporum cookei strains; until 2001, they were preserved as Nannizzia cajetani strains.

^cStrain CBS 161.69, until 1989, was preserved as a Microsporum gypseum var. gypseum strain; until 2001, it was preserved as a Nannizzia gypseum strain.

^dOnly the teleomorph name (Arthroderma persicolor) was given for strains CBS 421.74, CBS 468.74, and CBS 871.70 on the CBS website (http://www.cbs.knaw.nl).

^eT. mentagrophytes var. erinacei CBS 511.73 is a synonym of Trichophyton erinacei on the CBS website.

^fT. soudanense is considered a synonym of T. violaceum on the CBS website, but T. soudanense and T. violaceum are two different species on the ATCC website (http://www.atcc.org/).

^gOnly the teleomorph name (Arthroderma quadrifidum) was given for strains CBS 117.61, CBS 464.62, CBS 465.62, and CBS 567.94 on the CBS website.

DNA extraction.

Mycelial cells were disrupted by vigorous shaking with 0.5-mm-diameter zirconium-silica beads (Azygen Sientific, Union City, CA) in a mechanical cell disrupter (Mini-Beadbeater; Biospec Products, Bartlesville, OK) as previously described (14). The disrupted cell suspension was centrifuged at 8,000 × g for 10 min in a microcentrifuge. Fungal DNA in the supernatant was extracted by a genomic DNA extraction kit (Viogene, Taipei, Taiwan) according to the manufacturer's instructions.

ITS amplification.

PCR conditions used to amplify the ITS regions for array hybridization were described previously (14). A negative control was performed with each test run by replacing the template DNA with sterilized water in the PCR mixture.

Design of oligonucleotide probes.

Species-specific oligonucleotide probes (17- to 31-mers) used for the identification of 17 dermatophyte species are listed in Table 2. Probe design was based on sequence data from the ITS-1 and ITS-2 regions, with some sequences being determined in our previous study and submitted to GenBank (14). The designed probes were checked for internal repeats, secondary structure, melting temperature (T_m), and GC content using Vector NTI software (Invitrogen Corporation, Carlsbad, CA). A total of 21 probes (Table 2) including one positive control (designed from a conserved sequence of the 5.8S rRNA gene) were used to prepare the oligonucleotide array (data not shown). Six to 18 additional bases of thymine were added to the 3′ ends of probes that displayed low hybridization signals after preliminary testing (2). In addition, the digoxigenin-labeled fungal universal primer ITS4 (44) was spotted on the array and used as a position marker of hybridization (data not shown).

TABLE 2.

Oligonucleotide probes used in this study

Microorganism	Probe	Sequence (5′-3′)a	Length (mer)	Tm (°C)	Locationb	GenBank accession no.c
E. floccosum	Epflo2c	CCATAGGTGGTTCAGTCTGAGCGT	24	61.0	280-303 (1)	AY213646
M. audouinii	Miaud3b	CGACCGTCCCCCCCCAATAAC(tttttttttttttt)d	21	62.8	566-586 (2)	AJ252333
M. canis	Mican2g	CCAACTCCCCAGTAACCACCCACC(ttttttttttt)	24	63.1	517-540 (2)	AY213657
M. cookei	Micok4	GCCTTGACTGGACTCCTTTGTCCGTTAAATe	30	65.0	152-181 (2)	AY830116
M. ferrugineum	Mifer1c	CCCCTCCCCAACAACCACCCAC(tttttttt)	22	64.9	677-698 (2)	AJ252338
M. gallinae	Migal3	GGCCTCGTTTCAATAATTGTCGTTAGAGAATe	31	62.6	536-566 (2)	AJ000620
M. gypseum	Migyp3b	CGGTTTTCTGGCCTAGTTTTAGTTAGG	27	58.2	572-598 (2)	AJ970141
	Migyp5b	CCTAGTTTCCGTCAGAGATGTATTT	25	52.2	544-568 (2)	AJ000621
M. nanum	Minan1	AGGAACGATCAAAACACGCGAACACe	25	62.1	157-181 (1)	AJ970149
M. persicolor	Miper2	TGGCCTAGTTTTGGTCAAGGAC(tttttttt)	22	54.8	590-611 (2)	AJ000615
T. mentagrophytes	Trmen2d	GCCCCCGTCTTTGGGGGTGCG(ttttttttttttttttt)	21	69.6	507-527 (2)	Z97999
	Trmen3g	ATAGGGACCAACGTTCCGTCAGGG	24	62.9	82-105 (1)	Z98016
T. rubrum	Trrub3	GCAGCCAATYCAGCGCCCT(ttttttttttt)	19	60.8	698-716 (2)	AF170472
	Trrs1c	CAGACACCAAGAAAAAATTCTCTG(tttttt)	24	51.9	198-221 (1)	AJ222626
T. schoenleinii	Trsch2b	CGAATGGGCGCAACAAACCA(ttttttt)	20	61.1	575-594 (2)	Z98011
T. soudanense	Trrs1c	CAGACACCAAGAAAAAATTCTCTG(tttttt)	24	51.9	198-221 (1)	AJ222626
T. terrestre	Trter1	AAACACGGTCTTAACTGACCATCTAGG	27	57.2	598-624 (2)	AJ877214
	Trter6b	TTAACCAAAGTCCCCCAGG(tttttttttttttttt)	19	51.4	597-615 (2)	AJ877213
T. tonsurans	Trton1e	TCCGGCTTTCTAGGCGA(tttttttttttttttttt)	17	51.6	545-561 (2)	AY213690
T. verrucosum	Trver2e	TAGGGATCAGCGTTCCATCA(ttttttttttttttt)	20	52.9	82-101 (1)	Z98003
T. violaceum	Trvio1c	CAGACACCAAGGAAAATTCTCTG(ttttttt)	23	52.3	277-299 (1)	AJ270811
Positive control	PC	GCATCGATGAAGAACGCAGC(ttttttttt)	20	55.7	5.8S rRNA gene	EF134625

^aY stands for C or T.

^bThe location of probe is shown by the nucleotide number of either ITS-1 or ITS-2; the number (1 or 2) in parentheses indicates the ITS region from where the probe was designed.

^cGenBank accession numbers with a prefix of AY were determined in our laboratory and submitted to GenBank.

^dIndicates that 6 to 18 bases of thymine were added to the 3′ end of the probe.

^eProbes were described previously by Hsiao et al. (14).

Fabrication of oligonucleotide arrays.

Oligonucleotide probes were diluted 1:1 (final concentration, 10 μM) with a tracking dye solution (30% [vol/vol] glycerol, 40% [vol/vol] dimethyl sulfoxide, 1 mM disodium EDTA, 0.15% [wt/vol] bromophenol blue, and 10 mM Tris-HCl [pH 7.5]) (3). Probes were spotted onto a positively charged nylon membrane (Roche, Mannheim, Germany) by an Ezspot arrayer (SR-A300; EZlife Technology, Taipei, Taiwan) using a 400-μm-diameter solid pin. The array (0.5 by 0.5 cm) contained 36 dots (6 dots by 6 dots), including 20 dots for dermatophyte identification, 1 dot for a positive control, 5 dots for negative controls, and 10 dots for position markers (data not shown).

ITS amplification and hybridization procedures.

Unbound oligonucleotides on arrays were removed by two washes (2 min each) at room temperature in 0.5× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate)-0.1% sodium dodecyl sulfate. Most reagents used for array hybridization, except buffers, were included in the DIG Nucleic Acid Detection kit (Roche). The procedures for prehybridization, hybridization (55°C for 90 min), and color development by using alkaline phosphatase-conjugated anti-digoxigenin antibodies were described previously (3). The images of the hybridized chips were captured and processed by a scanner (Powerlook 3000; Umax, Taipei, Taiwan).

Analysis of discrepant identification.

For strains producing a discrepant identification between the conventional methods and array hybridization, the regions of ITS-1, ITS-2, and the D1-D2 domain of the large-subunit RNA gene were amplified by PCR, sequenced, and then compared with sequences in public databases using the BLAST algorithm for species clarification. The fungus-specific primers ITS1 and ITS2 were used to amplify the ITS-1 region, while the primer pair ITS3 and ITS4 was used for ITS-2 amplification (44). The D1-D2 region was amplified by primers NL1 and NL4 (23). The conditions used to amplify the ITS-1, ITS-2, and D1-D2 regions were described previously (14). PCR products were purified and sequenced using a model 377 sequencing system (Applied Biosystems, Taipei, Taiwan).

Definition of sensitivity and specificity.

A strain was identified as being one of the 17 dermatophyte species listed in Table 1 when the probe (or one of two probes) designed for the species and the positive control probe were hybridized (Table 2 and data not shown). For T. rubrum, the PCR product hybridized to its specific probe (probe Trrub3) and an additional probe (probe Trrslc) derived from the ITS sequence of T. soudanense. Sensitivity was defined as the number of target strains correctly identified (true positives) divided by total number of target strains tested (31). Specificity was defined as the number of nontarget strains producing negative hybridization reactions (true negatives) divided by the total number of nontarget strains tested (31). Strains were identified to the species level only; identification to the subspecies or variant level was not considered.

RESULTS

Probe development.

Initially, 65 probes (data not shown) were designed to identify the 17 dermatophyte species listed in Table 1. Through extensive screening, most probes cross-reacted with heterologous species or produced weak hybridization signals with homologous species. Finally, 20 probes were selected for preparation of the array (Table 2). Among the 20 probes, 3 probes (probes Micok4, Migal3, and Minan1) used to identify Microsporum cookei, Microsporum gallinae, and Microsporum nanum, respectively, were reported in our previous study (14), with the remaining 17 probes being developed in this study. For most dermatophytes, a single probe was designed for each species. However, two probes each were used to identify Microsporum gypseum, Trichophyton mentagrophytes, and Trichophyton terrestre (Table 2), since a consensus sequence could not be found in the ITS regions among different strains in each of the three species. T. rubrum and T. soudanense shared a common probe (probe Trrslc); i.e., both species hybridized to the probe (Table 2 and data not shown). However, T. rubrum was differentiated from T. soudanense by its hybridization to an additional probe (probe Trrub3). This design was due to the difficulty in obtaining a T. soudanense-specific probe that had no cross-hybridization with T. rubrum.

Identification of dermatophytes by the array.

A total of 198 strains of dermatophytes including 67 reference strains and 131 clinical isolates were tested by the array (Table 1). As expected, different strains of M. gypseum, T. mentagrophytes, and T. terrestre hybridized to one of the two probes designed for each of the three species (data not shown). In addition, T. rubrum hybridized with two probes (probes Trrub3 and Trrs1c), while T. soudanense hybridized only to probe Trrs1c (data not shown). All 67 reference strains were correctly identified by the oligonucleotide array (Table 1).

Among the 131 clinical isolates of dermatophytes, 13 produced discrepant identifications between phenotypic characteristics and array hybridization (Table 3). Four (M. audouinii LMA 597, M. canis LMA 922, M. gypseum LMA 90603, and T. mentagrophytes var. interdigitale LMA 951682) of the 13 isolates were not identified by the array because no hybridization signal, except the positive control, was observed. Species recognition of isolate LMA 597 as M. audouinii was confirmed by sequence analyses of the ITS-1, ITS-2, and D1-D2 regions (Table 3). Sequence analysis of the ITS-2 region revealed that M. audouinii LMA 597 has a nucleotide insertion at the region where the oligonucleotide probe (probe Miaud3b) was designed (Table 2), resulting in false-negative hybridization. The identification of isolate LMA 922 as being M. canis and isolate LMA 951682 as being T. mentagrophytes was not confirmed by sequencing of the ITS and D1-D2 regions because there were no matching sequences in public databases (Table 3). Therefore, strains LMA 922 and LMA 951682 were considered to be misidentified by conventional methods. For M. gypseum LMA 90603, sequence analyses of both ITS regions demonstrated that it was a misidentification of Microsporum fulvum, although the D1-D2 sequence revealed that the strain could be M. fulvum or M. gypseum (Table 3). There was no probe for M. fulvum on the array (Table 2), and therefore, M. fulvum could not be identified.

TABLE 3.

Strains that produced discrepant identifications by conventional methods and array hybridization

Strain	Species identification when received	Species identification(s) by:			Result of discrepant analysis
		Array	ITS-1/ITS-2 sequence (% ITS-1/% ITS-2)a	D1-D2 sequence (%)a
Target strains
LMA 597	M. audouinii	NIb	M. audouinii (100/99)	M. audouinii (100)	M. audouinii
LMA 922	M. canis	NI	NI	NI	NI
			M. canis (42/46)
LMA 90603	M. gypseum	NI	M. fulvum (99/99)	M. fulvum (99)	M. fulvum
			M. gypseum (87/94)	M. gypseum (99)
LMA 951682	T. mentagrophytes var. interdigitale	NI	NI	NI	NI
			T. mentagrophytes (56/57)
LMA 831	T. soudanense	T. violaceum	T. violaceum (97/100)	T. violaceum (99)	T. violaceum
			T. soudanense (96/100)
LMA 835	T. soudanense	T. violaceum	T. violaceum (97/100)	T. violaceum (99)	T. violaceum
			T. soudanense (97/100)
LMA 979	T. soudanense	T. violaceum	T. violaceum (100/99)	T. violaceum (99)	T. violaceum
			T. soudanense (98/99)
LMA 951336	T. soudanense	T. tonsurans	T. tonsurans (99/100)	T. tonsurans (100)	T. tonsurans
			T. soudanense (77/81)
LMA 50500944	T. soudanense	T. violaceum	T. violaceum (99/100)	T. violaceum (99)	T. violaceum
			T. soudanense (96/100)
LMA 50600067	T. soudanense	T. violaceum	T. violaceum (100/99)	T. violaceum (99)	T. violaceum
			T. soudanense (98/100)
NCKU 3156	T. tonsurans	T. rubrum	T. rubrum (100/100)	T. rubrum (100)	T. rubrum
			T. tonsurans (79/80)
LMA 40204990	T. verrucosum	T. rubrum	T. rubrum (100/100)	T. rubrum (100)	T. rubrum
			T. verrucosum (87/93)
LMA 50200673	T. verrucosum	T. rubrum	T. rubrum (100/100)	T. rubrum (100)	T. rubrum
			T. verrucosum (87/93)
Nontarget strains
CBS 170.65	T. gourvilii var. intermedium	T. soudanense	T. gourvilii (97/99)	T. gourvilii (100)	T. gourvilii
			T. soudanense (97/98)
LMA 40396666	M. equinum	M. canis	M. canis (100/100)	M. canis (99.8)	M. canis/M. equinum
			M. equinum (100/100)a

^aResults of BLAST search in GenBank.

^bNI, not identified.

Of the remaining nine discrepant clinical isolates, five T. soudanense isolates (LMA 831, LMA 835, LMA 979, LMA 5050094, and LMA 50600067) were misidentifications of T. violaceum, and T. soudanense LMA 951336 was a misidentification of T. tonsurans, as demonstrated by sequencing of the ITS and D1-D2 regions (Table 3). Likewise, T. tonsurans NCKU 3156 and Trichophyton verrucosum (LMA 40204990 and LMA 50200673) were misidentifications of T. rubrum (Table 3). In summary, 128 out of the 131 clinical isolates were proved to be target dermatophyte strains, and 127 isolates were correctly identified to the species level by array hybridization. If reference strains (67 strains) and clinical isolates (128 isolates) were taken together, a test sensitivity of 99.5% (194/195) was obtained by the method describe here. The array described here was able to detect genomic DNA at levels of 1 to 10 pg per assay by testing serial 10-fold dilutions of fungal DNA (data not shown). If the DNA content of a fungal cell was equivalent to 37 fg as a yeast cell (39), then the detection limit of the array was equivalent to 27 to 270 cells per assay.

Specificity of the array.

Of the 90 nontarget strains (84 species) (see Table S1 in the supplemental material) used for specificity tests, 88 did not produce any hybridization signal with probes on the array except the positive control probe. Trichophyton gourvilii var. intermedium CBS 170.65 and Microsporum equinum LMA 40396666 were misidentified as being T. soudanense and M. canis, respectively, by the array (Table 3). The ITS sequences of T. gourvilii var. intermedium CBS 170.65 and M. equinum LMA 40396666 were found to have high homology with the probes designed for T. soudanense and M. canis, respectively. T. gourvilii var. intermedium CBS 170.65 displayed high ITS sequence similarities (97 to 99%) with reference sequences of T. soudanense in GenBank (Table 3). BLAST searches of both ITS-1 and ITS-2 sequences of M. equinum LMA 40396666 hit sequences of M. equinum and M. canis with an identity of 100%. Since 2 of the 90 nontarget strains were misidentified by hybridization, a test specificity of 97.8% (88/90) was obtained by the array.

DISCUSSION

In this study, an oligonucleotide array was developed to identify 17 dermatophyte species, including common and some less common species (Table 1). With a test sensitivity of 99.5% (194/195) and a specificity of 97.8% (88/90), the present approach provides a reliable method for dermatophyte identification. The prominent feature of the present method is that it used a standardized molecular protocol encompassing DNA extraction, ITS amplification, and array hybridization for dermatophyte identification. Another advantage of the method reported here is that microscopic examination of fungal reproductive structures that are essential for classical identification is not necessary.

Ten target and two nontarget strains were found to produce discrepant identifications between the conventional methods and array hybridization (Table 3). Since M. gypseum, T. rubrum, T. soudanense, T. tonsurans, and T. verrucosum are distinct species (5, 25), M. gypseum LMA 90603, T. soudanense LMA 951336, T. tonsurans NCKU 3156, and T. verrucosum LMA 40204990 and LMA 50200673 were misidentified by conventional methods. The accuracy of identification of the five clinical isolates by array hybridization was confirmed by sequencing of both ITS regions and the D1-D2 domain of the 28S rRNA gene (Table 3). It should be noted that the identity of isolate LMA 90603 was not resolved by sequencing of the D1-D2 region, although the strain was unambiguously identified as being M. fulvum by ITS sequencing. This indicates that the ITS region might be a more discriminative target than the D1-D2 region for fungal identification, as found in our previous study (26). It is not easy to differentiate M. gypseum from M. fulvum on the basis of morphological criteria (5). To our knowledge, M. fulvum has not been described as a human pathogen (5), and the misidentification of M. fulvum as M. gypseum raises an important question. Isolate LMA 90603 was recovered from a patient with an inflammatory lesion of the left thigh. Therefore, it would be interesting to verify the morphologically based identification of M. gypseum isolates of human origin in order to determine the prevalence of M. fulvum in human dermatophytosis.

The results of array hybridization revealed that five T. soudanense isolates (LMA 831, LMA 835, LMA 979, LMA 50500944, and LMA 50600067) were misidentifications of T. violaceum (Table 3). Characteristic features of T. soudanense and T. violaceum are presented in standard mycology textbooks (25, 36). However, T. soudanense and T. violaceum are closely related species, composing a single, robust clade in ITS phylogeny (13). T. soudanense was considered to be a variant of the T. violaceum complex (together with T. gourvilii, T. soudanense, and Trichophyton yaoundei), which mainly causes tinea capitis but differs in cultural characteristics and the production of extracellular metabolites. The taxon complex was reduced to synonymy of T. violaceum based on the ITS sequence, PCR fingerprinting, and amplified fragment length polymorphism analysis (13). On the CBS website (http://www.cbs.knaw.nl) and in the Atlas of Clinical Fungi (5), T. soudanense is also recognized as a synonym of T. violaceum. If T. soudanense and T. gourvilii are conspecific with T. violaceum, then the five T. soudanense isolates and T. gourvilii var. intermedium CBS 170.65 could not be considered to be misidentifications caused by conventional methods (Table 3).

However, the unification of T. soudanense and T. violaceum may conceal possible evolutionary diversification. Both species have their unique phenotypes and geographic distributions (25, 27). T. violaceum was found to be the most common cause of tinea capitis in countries of West Asia and North Africa (1, 7), while T. soudanense is the most common cause of tinea capitis in schoolchildren of the Ivory Coast (32). Some mycologists refrain from introducing formal nomenclature changes of these dermatophytes (30) and proposed that sequencing of multiple genes should be used for the classification of these pathogens (9, 40, 41). It is interesting that we found, through multiple sequence alignment, some signature sequences that could differentiate these two taxa. In the ITS-1 region, the nucleotide at position 208 was constantly found to be “A” in strains of T. soudanense, while the nucleotide was “G” in strains of T. violaceum. Also, a single nucleotide deletion was found in strains of T. violaceum at position 214 in the ITS-1 region (our unpublished data). Furthermore, by using a microsatellite marker, Ohst et al. (35) previously demonstrated that T. soudanense is actually more closely related to T. rubrum than to T. violaceum. In this study, the hybridization result showed that T. rubrum cross-hybridized to the probe (Trrslc) used for the identification of T. soudanense (data not shown), which partly supported the finding described previously by Ohst and colleagues (35). Based on phenotypic differences (25, 36), geographic distribution (5, 25), the close relationship of T. soudanense with T. rubrum (35), and the presence of signature sequences in the ITS regions, we support that T. soudanense and T. violaceum should be classified as different species.

Species in the Microsporum canis complex (M. audouinii, M. canis, and M. ferrugineum) are phylogenetically closely related (12, 21). In the ITS-1 region, high levels of sequence homology were found in the following pairs (sequence identity is shown in parentheses): M. audouinii and M. canis (99%), M. audouinii and M. ferrugineum (99%), and M. canis and M. ferrugineum (100%) (our unpublished data). For this reason, oligonucleotide probes used for the differentiation of three Microsporum species were designed from the ITS-2 regions (Table 2). In this study, M. equinum LMA 40396666 was misidentified as being M. canis (Table 3). Based on three independent molecular methods (ITS sequencing, PCR fingerprinting, and amplified fragment length polymorphism), M. canis and M. equinum were proposed to be infraspecific taxa, with M. canis being pathogenic mainly to cats and dogs and with M. equinum being a pathogen of horses (12). If T. gourvilii and M. equinum were proposed to be synonyms of T. violaceum and M. canis, respectively, then the specificity of the array increased from 97.8% (88/90) to 100% (88/88).

Species recognition of dermatophytes has important epidemiological implications in relation to the acquisition and spread of human infection (38). For example, currently, the predominant species causing tinea capitis in the United States is T. tonsurans (8). By contrast, T. violaceum and T. soudanense, which are common causes of tinea capitis in parts of Africa and West Asia (7), are rarely isolated from patients in the United States. Recently, Magill et al. (27) found that the isolation rates of T. violaceum and T. soudanense significantly increased in the Baltimore, MD, metropolitan area. They concluded that the changing epidemiology was due to the introduction of a “new” pathogen by increased immigration to the Baltimore area, where the two species were not endemic. Based on normal habitat, the dermatophytes are termed geophilic, zoophilic (such as M. canis and T. mentagrophytes), or anthropophilic (such as T. rubrum and T. violaceum). To prevent the spread or recurrence of zoophilic tinea capitis, it is important to identify potential sources of infection, and suspected pets should also be treated with antifungals (38). With anthropophilic dermatophyte species, patient-to-patient transmission may occur and should be prevented. In addition, dermatophytosis caused by different species may need different treatment protocols. Treatment of tinea capitis caused by M. canis needs a duration of 12 weeks with griseofulvin, but an 8-week treatment with the same agent would be enough if the disease is caused by T. tonsurans (38). The susceptibilities of different species to some antifungal agents currently used also varied (37).

The T_m values of probes used in this study varied from 51.4 to 69.6°C, and some probes had T_m values lower than the hybridization temperature (55°C) (Table 2). However, clear signals were obtained for all 17 species tested (data not shown). Volokhov et al. (42) also reported the successful use of probes that had T_m values lower than the hybridization temperature for the identification of Listeria species. The addition of several thymine bases to the end of a probe has the benefit of reducing steric hindrance between target DNA and the oligonucleotide probe immobilized on a solid support (2).

The array reported here was able to detect genomic DNA at levels of 1 to 10 pg. The low limit of detection might be due to the presence of high copy numbers (40 to 80 copies) of fungal rRNA operon per haploid genome (44), although the copy number of the rRNA operon in dermatophytes was not reported. Based on the low detection limit, the array may have a potential to directly detect dermatophytes in clinical specimens such as skin, hair, and nail. In conclusion, dermatophyte identification by the array reported here is highly reliable and can be used as an accurate alternative to conventional identification methods. The whole procedure can be finished with isolated colonies within 24 h.