Isolation, characterization, and disruption of dnr1, the areA/nit-2-like nitrogen regulatory gene of the zoophilic dermatophyte, Microsporum canis

A homolog of the major nitrogen regulatory genes areA from Aspergillus nidulans and nit-2 from Neurospora crassa was isolated from the zoophilic dermatophyte, Microsporum canis . This gene, dnr1 , encodes a polypeptide of 761 amino acid residues containing a single zinc-finger DNA-binding domain, which is almost identical in amino acid sequence to the zinc-finger domains of AREA and NIT-2. The functional equivalence of dnr1 to areA was demonstrated by complementation of an areA loss-of-function mutant of A. nidulans with dnr1 cDNA. To further characterize this gene, dnr1 was disrupted by gene replacement based on homologous recombination. Of 100 transformants analyzed, two showed the results expected for replacement of dnr1 . The growth properties of the two dnr1 (cid:2) mutant strains on various nitrogen sources were examined. Unlike the A. nidulans areA (cid:2) mutant, these dnr1 (cid:2) mutants showed significantly reduced growth on ammonia, a preferred nitrogen source for fungi. These mutant strains were also able to utilize various amino acids for growth. In comparison with wild-type M. canis , the two dnr1 (cid:2) mutants showed reduced growth on medium containing keratin as the sole nitrogen source. This is the first report describing successful production of targeted gene-disrupted mutants by homologous recombination and their phenotypic analysis in dermatophytes.


Introduction
Dermatophytes are a closely related group of keratinophilic fungal pathogens, which are responsible for a superficial cutaneous infection called dermatophytosis (ringworm) in both humans and animals. Microsporum canis is a zoophilic fungus, and is the most common causative agent of dermatophytosis in animals. Cats and dogs are regarded as the natural hosts for this fungus, and may even act as reservoirs [1,2], thus contributing to the high prevalence of zoonotic human infections [3 Á6]. Zoonosis is often caused by direct contact with animals infected with M. canis [7].
Dermatophytes commonly gain access to the host via keratinized structures, such as the hair, skin, or nails, which are cornified tissues that form solid structural barriers against invasion. To overcome complex host defense systems and progress the infection cycle, most pathogenic microorganisms produce a variety of extracellular hydrolytic enzymes with different activities both constitutively and inductively. Similarly, dermatophytes appear to use several extracellular hydrolytic enzymes for penetration through cornified tissues and maintenance of fungal growth on host tissues. Among these enzymes, a great deal of attention has been focused on the keratinolytic proteases (the so-called keratinases) that digest keratins, the major constituents of cornified tissues, as possible virulence-related factors in dermatophytosis.
Although little is currently known about the regulation of expression of extracellular proteases by dermatophytes, extensive studies of protease production in the model filamentous fungi, Aspergillus nidulans and Neurospora crassa , have suggested that extracellular protease expression is regulated by both nitrogen (N) and carbon (C) catabolite repression, i.e., control of nitrogen and carbon availability. In these two fungi, a number of genes have been shown to be expressed in response to restriction of nitrogen and/or carbon source availability. areA from A. nidulans [8,9] and nit-2 from N. crassa [10,11] are well-characterized genes involved in nitrogen catabolite repression in these fungi. areA and nit-2 encode transcription factors containing a DNA-binding domain consisting of a Cys-2/Cys-2 zinc-finger motif followed by an adjacent basic region. The DNA-binding domains of AREA and NIT-2 are characteristic of the GATA family of transcription factors [12]. These molecules activate expression of a large number of genes encoding permeases and enzymes that mediate uptake and utilization of various secondary nitrogen sources, in the absence of preferred nitrogen sources, such as ammonia and glutamine [13]. These genes include those encoding extracellular proteases. Genes encoding products that are structurally and functionally similar to AREA and NIT-2 have been found in both human and plant pathogenic fungi, such as Aspergillus fumigatus [14], Candida albicans [15], and Magnaporthe grisea [16], and inactivation of the corresponding genes could affect the production of extracellular proteases in these fungi.
Here, we report the molecular characterization of the dnr1 gene from M. canis, which is homologous to areA and nit-2 . Complementation of the A. nidulans areA ( mutant with the isolated cDNA and phenotypic analyses of the dnr1 ( mutants produced by homologous recombination indicated that dnr1 is involved in regulation of nitrogen metabolism in M. canis. The reduced growth activity of these dnr1 ( mutants in liquid medium containing keratin as the sole nitrogen source also suggested a relationship between dnr1 and expression of extracellular proteases during the infection cycle on host tissues.

Fungal strains and culture conditions
The wild-type M. canis strain, TIMM4092 [17 Á19], the wild-type Aspergillus nidulans strain, 2373 (inoB, glrIA1 ), and the A. nidulans areA ( loss-of-function mutant strain, 2207 (wA3 , inoB, areA r 5), were used in this study. areA r 5 has the molecular nature of the mutation that the Cys673 of the zinc finger domain was substituted by a tyrosine [20]. M. canis strains were maintained at 288C on solid Sabouraud dextrose (SD) medium with or without an appropriate concentration of hygromycin B, whereas A. nidulans strains were maintained at 288C on solid SD medium, solid Aspergillus complete medium (ACM), or minimal medium (AMM) [21] with 10 mM of sodium nitrate (NaNO 3 ) or ammonium chloride (NH 4 Cl).

Genomic DNA isolation and Southern blotting analysis
Genomic DNA of each M. canis and A. nidulans strain was isolated from the growing mycelia, according to the method of Girardin and Latge [22].
Aliquots of 3 mg of genomic DNA were digested with appropriate restriction enzymes, separated by electrophoresis on 0.7% (w/v) agarose gels, and transferred onto Hybond N ' membranes (Amersham Biosciences, Piscataway, NJ, USA). Southern blot hybridization was performed using the ECL TM Direct Nucleic Acid Labeling and Detection System (Amersham Biosciences).
Isolation of cDNA and genomic clones encoding the M. canis AREA/NIT-2 ortholog Total RNA of the M. canis strain, TIMM4092, was isolated from the growing mycelia cultured for 4 days in 50 ml of modified dermatophyte minimal medium (DMM) [23] with 250 mg of dog hair, by using an RNeasy Plant Mini Kit (Qiagen, Hilden, Germany).
A cDNA clone encoding the M. canis AREA/NIT-2 ortholog was isolated by RT-PCR and RACE-PCR with the 5? and 3? Rapid Amplification of cDNA Ends method with Super Script II reverse transcriptase (Invitrogen, Carlsbad, CA, USA). A genomic clone of the M. canis AREA/NIT-2 ortholog was also amplified by PCR. Nucleotide sequences of primers (SP1, ASP1, 3?RACE1, ASP2, 5?RACE1, 5?RACE2, SP2 and ASP3) used for amplification of the cDNA and genomic clones were shown in Table 1.
Nucleotide sequencing reactions of all the amplified fragments were performed using a BigDye Terminator Cycle Sequencing FS Ready Reaction Kit (Applied Biosystems, Foster City, CA, USA).

Protoplast preparation and genetic transformation
Protoplast preparation and polyethylene glycol (PEG)mediated transformation of the M. canis strain, TIMM4092, and the A. nidulans areA ( mutant strain, 2207, were performed, according to the method of Yamada et al . [25], except that the selection of A. nidulans transformants.

Assay of growth properties of A. nidulans and M. canis on various nitrogen sources
The conidial suspensions of each A. nidulans strain were spotted onto solid AMM supplemented with various nitrogen sources added to the medium as follows: 10 mM of NaNO 3 , sodium nitrite (NaNO 2 ), and NH 4 Cl, 1 mM L-tyrosine, and 5 mM of other amino acids, adenine, urea, and formamide, 1 mM hypoxanthine, 1.0% (w/v) skimmed milk. After incubation at 378C for 2 to 3 days, growth of fungal colonies was compared.
Protoplast suspensions of each M. canis strain were spotted onto solid DMM supplemented with 1.2 M D-sorbitol and the above nitrogen sources plus 0.5% (w/v) BSA. After incubation at 288C for 7 to 8 days, growth of reproduced fungal colonies was compared.
For the growth property assay of M. canis dnr1 ( mutant strains on the medium containing keratin as the sole nitrogen source, protoplast suspensions of each M. canis strain were spotted onto membrane filters (Omnipore JA; pore size, 1.0 mm, 13.0 mm in diameter) (Nihon Millipore, Yonezawa, Japan) on solid RYSDA [25], and incubated at 288C for 3 days. The filters with regenerated mycelia were lifted from the plates, washed once in 20 ml of PBS, and placed onto dog hair (400 mg) immersed in 10 ml of liquid modified DMM. During incubation at 288C for 7 days, growing mycelia were observed by stereomicroscopy.

Nucleotide sequence accession number
The nucleotide sequence data of dnr1 from M. canis reported here have been deposited in the DDBJ/EMBL/ GenBank under Accession No. AB201458.

Results
Isolation of cDNA and genomic clone encoding M. canis AREA/NIT-2 ortholog An isolated cDNA, designated as dnr1 , contained an open reading frame (ORF) of 2283 bp encoding a predicted product of 761 amino acid residues (Fig. 1). Alignment between the dnr1 cDNA and genomic sequences revealed that the ORF of dnr1 was interrupted by 2 introns of 80 bp and 74 bp. The existence of these two introns was similar to nit-2 from N. crassa [10,11], and nut1 and clnr1 from the plant fungal pathogens Magnaporthe grisea [16] and Colletotrichum lindemuthianum [28]. In contrast, the areA genes from A. nidulans [9] and other Aspergillus species [29,30] have only a single intron in their coding regions. Southern blotting analysis using a Cys-2/Cys-2 zincfinger domain (Fig. 1, shaded box) as a probe suggested the existence of the dnr1 gene as a single copy in the  5?-ACGCAAACTACTCCCCTC-3? Fig. 1 Comparison of the amino acid sequence of the predicted M. canis dnr1 protein with fungal AREA/NIT-2 orthologs. AREA (X52491) [8,9], NIT-2 (AAB03891) [11], and NUT1 (AAB03415) [16] are from A. nidulans, N. crassa , and the plant pathogenic fungus, Magnaporthe grisea , respectively. The zinc-finger region is boxed. Horizontal arrows indicate annealing sites of the oligonucleotide primers used for isolation of the dnr1 cDNA.
M. canis genome (data not shown). The predicted protein of dnr1 shares extensive regions of high similarity with the known AREA/NIT-2 orthologs (Fig. 1). Of these conserved regions, the carboxylterminal conserved region, together with the zinc-finger domain, was shown to be involved in interaction with another nitrogen regulatory protein in N. crassa [31,32]. As the results of estimation by the unweighted pair group method with arithmetic mean (UPGMA), the predicted dnr1 protein product showed amino acid sequence similarities of 22.0%, 15.0%, and 14.0% to AREA, NIT-2, and NUT1, respectively.

Functional analysis of the M. canis dnr1 gene
To determine whether M. canis dnr1 is capable of complementing the functions of areA , the A. nidulans areA ( mutant strain was transformed with pCHSH75-dnr1 carrying the dnr1 cDNA. Of 19 transformants regenerated on solid AMM with 10 mM NaNO 3 , the growth properties of NT10 and NT11, carrying one or two copies of the dnr1 cDNA, were examined on solid AMM containing different nitrogen sources. As shown in Table 2, the A. nidulans areA ( mutant strain (2207) was able to grow only on medium containing ammonium, glutamine, and urea. In contrast, the wild-type A. nidulans strain (2373) and the two A. nidulans transformants showed similar growth properties and were able to grow on all of the media tested ( Table 2), demonstrating that the dnr1 gene product has functions equivalent to those of AREA. The both A. nidulans transformants were found to also produce extracellular proteases on the skimmed milk-containing medium ( Table 2).

Disruption of dnr1 by homologous recombination
To confirm that the dnr1 is involved in the regulation of nitrogen metabolism in M. canis, we attempted to disrupt the dnr1 based on a gene replacement strategy. Transformation of the wild-type M. canis strain, TIMM4092, with the dnr1 -targeting vector, pR-dnr1 (Fig. 2), yielded more than 100 putative transformants. Of these transformants, 100 were chosen at random Aliquots of 10 ml of each conidial suspension containing about 5)/10 4 conidia were spotted on media and growth of fungal colonies was compared after incubation of 2 to 3 days at 378C. The ability to grow on different media, and protein degradation by extracellular proteases are indicated as '-' (no growth) to ''/'/'/' (vigorous growth) and 'h' (halo), respectively. a) The two A. nidulans strains complemented with the M. canis dnr1 cDNA, NT10 and NT11, exhibited the same morphological characteristics. and screened for replacement of dnr1 by the hph cassette. Southern blotting analysis (Fig. 3)

Growth properties of the M. canis dnr1 ( mutant strains on various nitrogen sources
The growth properties of the M. canis dnr1 ( mutant strains, M47 and M58, on solid DMM containing different nitrogen sources were compared with those of the wild-type strain, TIMM4092, and the ectopic transformant, M55. As TIMM4092 showed absence of conidia productivity within a few passages in subculture, none of the derivatives of TIMM4092 examined here, i.e., M47, M58, and M55, was capable of producing conidia. Therefore, we used protoplasts prepared from growing mycelia. As shown in Table 3, TIMM4092 and M55 were able to grow on all of the nitrogen sources tested, while both M47 and M58 showed reduced or complete loss of ability to utilize nitrate, nitrite, and several other nitrogen compounds. However, unlike the A. nidulans areA ( mutant, M47 and M58 maintained their growth activity on glutamine and many other amino acids. Furthermore, both strains showed difficulties in growth on ammonia, one of the preferred nitrogen sources for fungi (Table 3), and addition of much higher levels of ammonia did not improve their reduced growth activity. TIMM4092 and M55, but not M47 and M58, showed production of extracellular proteases on skimmed milk-containing medium (Table 3). M47 and M58 also showed different growth properties on the medium containing keratin, an important nutrient source for dermatophytes, as compared with TIMM4092 and M55. Mycelia from TIMM4092 and   2 Restriction map of the dnr1 gene-targeting vector, pR-dnr1 . The pR-dnr1 was constructed from pCHSH75 [25]. Abbreviations: Pch, C. heterostrophus promoter 1; hph, E. coli hph gene; TtrpC, terminator sequence of A. nidulans tryptophan C gene; Bg, Bgl II; C, Cla I; E, Eco RI; P, Pst I; Spe, Spe I; Xb, Xba I; X, Xho I. M55 had already begun to grow after 2 days of incubation, whereas those from M47 and M58 showed visible growth only after 7 days of incubation (Fig. 4). Subsequently, their mycelia were incubated for a further 4 days, the extent of which was reduced as compared with TIMM4092 and M55 (data not shown).

Discussion
In the present study, we report the isolation of the areA /nit-2 -like nitrogen regulatory gene (dnr1 ) from M. canis. Comparison of the growth properties between M. canis wild-type and dnr1 ( mutant strains on different nitrogen sources (Table 3) indicated that the dnr1 is involved in the control of nitrogen metabolism in this species. The dnr1 ( mutant strains showed significantly reduced growth on ammonia (Table 3). Similar results were reported previously in an areA ( mutant strain of A. oryzae [29], except that growth activity could be restored in this mutant by increasing the amount of ammonia present in the medium. Based on the observation that high levels of ammonia could overcome the poor growth activity of the A. oryzae areA ( mutant strain, Christensen et al . [29] concluded that the uptake of ammonia, rather than its subsequent assimilation cycle, should be the main factor responsible for the reduced growth of the A. oryzae areA ( mutant strain on this nitrogen source. In contrast, growth activity of the dnr1 ( mutant strains was not improved even by the presence of a very high level of ammonia (100 mM) ( Table 3), suggesting that the intracellular ammonia assimilation cycle, together with its uptake, may be related to ammonia availability in dnr1 ( mutant strains. Inactivation of the areA/nit-2 -like genes in many fungi results in loss or a significant reduction in growth activity on most of the amino acids, with the exception of glutamine, but the M. canis dnr1 ( mutants were still able to use glutamine and many other amino acids ( Table 3). The availability of a variety of amino acids in the dnr1 ( mutant strains may be related to nutritional environments encountered by the fungi on the host tissues. In general, dermatophytes grow exclusively in the superficial keratinized tissues, such as the hair, skin, nails, or stratum corneum, which appear to be poor in available nutrient sources. Thus, available exogenous small compounds, such as amino acids and short peptides, which are produced by digestion of keratinbased nutrients, appear to become indispensable nutrients for the fungi for the maintenance of growth during infection. The uptake of exogenous amino acids is usually carried out by specific membrane proteins (the so-called permeases). Of the amino acid permeases identified in yeasts and filamentous fungi, GAP1 , the general amino acid permease from Saccharomyces cerevisiae [34], and prnB, the proline transporter from A. nidulans [35], are known to be under control of the areA/nit-2 -like genes (S. cerevisiae, GAT1/NIL1 and GLN3 [36]; A. nidulans, areA ). The loss or significant reduction of the availability of several amino acids (glycine, proline, etc.) in the dnr1 ( mutants suggests that M. canis should have a common exogenous amino acid uptake system under control of the areA/nit-2 -like gene. In addition to this specific system, the fungus may also develop different exogenous amino acid uptake systems that contribute cooperatively to achieve efficient utilization of exogenous nutrients in poor nutritional environments. The production of extracellular proteases by M. canis was affected by dnr1 ( Table 3). The dnr1 ( mutant strains showed significantly retarded growth on medium containing keratin as the sole nitrogen source, as compared with the wild-type strain (Fig. 4). Several extracellular keratinolytic proteases from M. canis were produced during culture in medium containing keratin as the sole nitrogen source [23,37]. This growth property of these mutants may be related to the production of extracellular keratinolytic proteases. Furthermore, it has been reported that areA/nit-2 -like genes in some human and plant pathogenic fungi play roles in host infection [14,15,28]. Experiments are currently underway in our laboratory to identify the extracellular proteases of M. canis under control of the dnr1 and in vivo studies in an animal model of tinea pedis, involving disruption of a dnr1 -like gene (tnr ) of Trichophyton mentagrophytes. These studies will provide more detailed information regarding the roles of dnr1 and extracellular proteases in dermatophytosis. Targeted gene disruption mediated by genetic transformation is indispensable for determination of the functions and roles of genes isolated from pathogenic fungi. Disruption of the dnr1 was confirmed in only 2 of the 100 transformants analyzed, a lower frequency of homologous recombination than reported previously in other filamentous fungi, such as A. nidulans. The low efficiency of homologous recombination in dermatophytes is attributed mainly to their low transformation frequency [24,38]. To increase the transformation frequency of dermatophytes, several aspects of the transformation procedure must be altered. One candidate is alteration of the transformation vector conformation (i.e., circular vs. linear). Linearization of transformation vectors by restriction enzymes, the socalled restriction enzyme-mediated integration (REMI) procedure [39], has been shown to improve transformation frequency in several filamentous fungi [40,41]. Kaufman et al . [42] recently reported enhanced transformation of T. mentagrophytes by the REMI procedure. In addition, Bird and Bradshaw [43] reported that nucleotide length of homologous fragments within the transformation vectors, together with their conformation, was correlated with the efficiency of homologous recombination in A. nidulans ; an increase in the size of the homologous fragments resulted in an increase in the efficiency of homologous recombination (1.0% for 0.6 kb, 14.0% for 0.9 kb, 27.0% for 1.2 kb). One of the homologous fragments inserted in the pR-dnr1 was only about 0.7 kb in length (Fig. 2). Thus, use of larger homologous fragments may make a contribution to the increase in the efficiency of homologous recombination in dermatophytes.