The Neurospora crassa gene responsible for the cut and ovc phenotypes encodes a protein of the haloacid dehalogenase family

Authors


E-mail avalos@us.es; Tel. (+34) 954 557 110; Fax (+34) 954 557 104.

Summary

Light stimulation of carotenogenesis in Neurospora crassa, mediated by the White Collar proteins, is enhanced in some regulatory mutants, such as vivid and ovc. The gene responsible for the vivid mutation has been identified, but not the one responsible for the ovc phenotype. The ovc mutant is sensitive to high osmotic conditions and allelic with another mutant, cut, also osmosensitive but not affected in carotenogenesis. A phenotypic characterization of both strains is presented. Light induction of mRNA levels of the carotenoid genes al-1 and al-2, the regulatory gene wc-1 or the conidiation-specific gene con-10 is not significantly changed in the ovc mutant when compared with the wild type. We have identified the gene affected in the ovc mutant by complementation of osmosensitivity with a cosmid library. This gene, which we call cut-1, codes for an enzyme of the haloacid dehalogenase family, which includes different classes of phosphatases. cut-1 is able to restore the wild-type phenotype of the ovc and cut strains, confirming that they are affected in the same gene. DNA sequence analysis identified a point mutation in the cut mutant, leading to a truncated protein. The ovc mutant represents a deletion encompassing the entire gene and surrounding sequences. The cut-1 promoter contains putative regulatory elements involved in osmotic or thermal stress. We show that cut-1 transcription is low in illuminated or dark-grown cultures, and is induced by high osmotic conditions or by heat shock.

Introduction

The ascomycete Neurospora crassa is a well-known microbial model extensively used in genetic research (Davis, 2000). A large number of mutants and genetic markers from this fungus have been phenotypically characterized and genetically mapped (Perkins et al., 1982; 2001). The sequence of its genome has been recently determined (Galagan et al., 2003) and the annotation process has provided a valuable tool to assign genes to formerly identified genetic loci.

As with all microorganisms, Neurospora responds to changes in environmental conditions activating or repressing particular sets of genes. Two widely investigated external conditions are light and osmotic pressure. Neurospora exhibits different light responses, such as photoinduction of carotenoid biosynthesis (Harding and Turner, 1981) and conidiation (Lauter et al., 1997), phototropism of perithecial beaks (Harding and Melles, 1983) and entrainment of the circadian clock (Dunlap, 1999). The variety and complexity of its photoresponses have made this fungus a leading research subject in photobiology (Linden, 2002).

Particular attention has been devoted to photoinduction of the carotenoid pathway. Carotenoids are terpenoid pigments synthesized by photosynthetic organisms and many microorganisms, including fungi (Sandmann and Misawa, 2002). The end-product of Neurospora carotenogenesis is neurosporaxanthin, an acidic apocarotenoid (Aasen and Jensen, 1965) produced through the sequential activities of a prenyl transferase (coded by the gene al-3; Sandmann et al., 1993), phytoene synthase (al-2; Schmidhauser et al., 1994), a dehydrogenase (al-1; Schmidhauser et al., 1990), a cyclase (5′domain of al-2; Arrach et al., 2002) and a hypothetical cleaving oxydase (gene not identified). The amounts of carotenoids synthesized by Neurospora mycelia are higher in the light than in the dark. All Neurospora photoresponses, including photoinduction of carotenoid biosynthesis, depend on the transcriptional activation of the al genes by the White Collar (WC) proteins, WC-1 and WC-2 (Talora et al., 1999). The WC regulatory proteins interact to form active heterodimers (Froehlich et al., 2002) in which WC-1 functions as a photoreceptor (Cheng et al., 2003). A role as a photoreceptor has also been demonstrated for VIVID, involved in this case in the adaptation to light (Schwerdtfeger and Linden, 2003). In addition to the photoresponse, carotenoids accumulate in conidia in the dark through a developmental activating mechanism independent of the WC proteins (Li and Schmidhauser, 1995).

Osmotic tolerance is a phenotypic trait widely investigated in Neurospora. The ability to grow under high osmotic conditions is lost in the os (osmosensitive) mutants. At least seven mutants (os-1 to os-6 and cut) have been classified as osmosensitive (Perkins et al., 1982) due to an inability to grow on media supplemented with 0.65 M NaCl or 1 M sorbitol (Perkins et al., 2001). Some of the os genes encode components of osmotic sensing pathways. os-1 codes for a putative osmotic sensing histidine kinase (Miller et al., 2002), os-2 codes for a putative mitogen-activated protein (MAP) kinase (Zhang et al., 2002), and os-4 and os-5 code for further components of a MAP kinase cascade (Fujimura et al., 2003). These regulatory proteins suggest the occurrence in Neurospora of a regulatory mechanism similar to the high-osmolarity glycerol (HOG) response of yeast, controlled by a MAP kinase cascade (Gustin et al., 1998). The os mutants of Neurospora often exhibit morphological alterations, including reduced conidiation, and a tendency of their hyphae to aggregate or develop leaking breaks (Perkins et al., 2001), reflecting alterations in cell-wall composition. The phenotype of the cut mutant is peculiar, exhibiting a characteristic well-delimited upper edge in aerial mycelia (Kuwana, 1953), and no particular aggregating or leaking properties when compared with the wild type. The genes affected in the os-3, os-6 and cut mutants have not been identified.

A Neurospora mutant called ovc, isolated as a carotenoid overproducer in the light (Harding et al., 1984), is osmosensitive and allelic to cut (Banks et al., 1997). The ovc mutant shows normal carotenoid accumulation upon incubation of illuminated mycelia at low temperature (6°C), culture conditions under which Neurospora photocarotenogenesis is particularly efficient (Harding, 1974). We have carried out a detailed characterization of the ovc and cut phenotypes, including the effect of light on mRNA levels of their carotenoid genes, and we have cloned a gene able to restore osmotolerance in both strains. The gene that we call cut-1 complements other phenotypic traits of the ovc and cut mutants, and we have found significant differences in the molecular basis of their mutations.

Results

Carotenoid accumulation and other phenotypic traits of the cut and ovc mutants

The ovc and cut mutants differ in their abilities to accumulate carotenoids. As expected, illuminated cultures of the ovc mutant show a deeper pigmentation and produce more carotenoids than those of the wild type (Fig. 1A). Carotenoid concentration increases at approximately constant rates in the wild type and the ovc mutant for at least 2 days after the onset of illumination. Carotenoids accumulate faster in the ovc mutant, and the difference broadens over time. The increased response to light is similar in mycelia grown on agar (232 and 258 µg per gram dry weight in the ovc mutant against 97 and 105 in the wild type in two independent determinations). In contrast, the cut mutant exhibits a normal carotenoid photoinduction (Fig. 1B).

Figure 1.

Phenotype of the ovc and cut mutants.
A. Time-course of carotenoid accumulation in the wild type (wt) and in the ovc mutant. The cultures were incubated for 2 days in the dark and transferred to light for the times indicated.
B. Carotenoids accumulated by the wild type and the mutants ovc, cut and vvd in the same conditions after 24 h illumination at either 30°C or 8°C.
C. Phytoene, neutral and polar carotenoids accumulated by the wild type, and the ovc and cut mutants in the conditions described in (B).

The increase in carotenoid biosynthesis in the ovc strain, at least twice that of the wild type, is more modest than the increase exhibited by a vivid mutant, a strain also characterized by a deep pigmentation in the light (Shrode et al., 2001). Chemical analyses of the carotenoids accumulated by these strains show no significant difference in the amounts of the colourless precursor phytoene and the coloured neutral intermediates of the pathway. The increased pigmentation results only from a higher accumulation of polar carotenoids (Fig. 1C), interpreted as an increase in the levels of neurosporaxanthin, the only polar carotenoid of the pathway. As expected (Harding et al., 1984), there is no quantitative or qualitative difference in carotenoid accumulation between these strains when illumination and subsequent incubation are performed at 8°C (Fig. 1B and C). Only traces of phytoene were found in the samples at 8°C.

Apart from carotenoid biosynthesis, the ovc and cut mutants share phenotypic similarities consistent with genetic data that attribute their phenotypes to a single gene (Banks et al., 1997). Both strains exhibit a comparable growth inhibition if the medium is supplemented with 4% (0.65 M) NaCl (Fig. 2A). Similar results are observed for growth in the presence of KCl, NaNO3 or KNO3 at the same molar concentration. Growth of both strains is also affected if the medium is supplemented with 1.3 M sorbitol yielding the same osmotic pressure (0.65 M) as for the tested salts, confirming the sensitivity of these strains to hyperosmotic conditions. Additionally, surface cultures of both strains exhibit a similar developmental alteration, characterized by a sudden cut-off of aerial hypha formation in slant cultures, ending as a velvet-like surface that contrasts with the powdery look of the wild-type cultures (compare wild type and cut slant cultures in Fig. 5B).

Figure 2.

Growth inhibition produced by high osmotic conditions or by the antibiotic fludioxonil on the wild type (black bars), cut (grey bars) or ovc (white bars) strains.
A. Inhibition produced by the presence of 0.68 M of the indicated salts or the equivalent osmomolarity (1.36 M) of sorbitol.
B. Inhibition produced by the presence of 10 mg l−1 or 50 mg l−1 fludioxonil.
Relative growth is a comparison of mycelial dry weight of strains grown for 3 days in the dark on VMM medium without addition. Data are the average and standard deviation from at least two independent experiments.

Figure 5.

Complementation of the cut phenotype.
A. Restriction map of the DNA region of N. crassa containing the cut-1 gene. White arrows indicate annotated genes. Restriction sites are indicated for EcoRI (E), KpnI (K) and HindIII (H). Black arrowheads indicate the location of predicted SRE regulatory elements; white arrowheads indicate predicted heat-shock regulatory elements.
B. Left, Southern blot of genomic DNA from the wild type (wt), the cut mutant and the cut-derived transformant T2 digested with either EcoRI or HindIII. Sizes markers (M) are indicated in kb on the left. Right, VGM slant cultures of the same strains grown for 3 days at 30°C in the dark and two further days at 22°C under light.

The Neurospora os mutants exhibit a characteristic resistance to dicarboximide anti-fungals, such as fludioxonil (Zhang et al., 2002; Fujimura et al., 2003), but the cut mutant is sensitive to this antibiotic (Fujimura et al., 2000). We have compared the sensitivity to fludioxonil of the ovc strain in comparison with the cut mutant and the wild type. As expected, the wild-type growth is severely impaired. The ovc strain is at least as sensitive as the wild type, but the cut mutant is only partially sensitive (Fig. 2B).

Despite exhibiting a developmental alteration in aerial growth, the ovc and cut mutants show conidiation levels similar to wild type (values ranging from 1.2 × 1010 to 2.3 × 1010 conidia per flask).

Carotenoid gene expression in the ovc and cut mutants

To assess the molecular basis for carotenoid accumulation in the ovc mutant in the light, we measured mRNA levels for the carotenoid genes, al-1 and al-2, in mycelia incubated for different times after a 10 min light pulse. The induction kinetics of mRNA accumulation is very similar in the wild-type and ovc strains (Fig. 3A), including a decrease resulting from light adaptation, indicating that the ovc mutation does not affect light regulation. The vivid mutant exhibited a similar induction, although the light adaptation is less apparent in this case. The ovc phenotype has no relation with light sensitivity, as shown by the similar accumulation of al-1 and al-2 mRNAs in this strain and in the wild type upon illumination with low-intensity light pulses (Fig. 3B). The amount of al-2 mRNA in both strains is significantly reduced at very low light intensity, implying lack of saturation under these illumination conditions.

Figure 3.

Effect of light on al-1 and al-2 gene expression in the wild-type and in mutant strains.
A. Northern blots of total RNA from the wild type, and the ovc and vvd mutants grown for 2 days in the dark, exposed for 10 min under light and incubated in the dark for the times indicated in min (including the 10 min light pulse).
B. Northern blots of total RNA from the wild type and ovc mutant illuminated for 30 min under three different light intensities (D, no illumination; VL, very low light intensity: 10 mW m−2; L, low light intensity: 0.7 W m−2; H, high light intensity: 10 W m−2). rRNA bands are shown below each panel as load control.

Similar results are obtained with longer light exposures (Fig. 4A), or after 24 h illumination under the conditions shown in Fig. 1 (Fig. 4C). Because of the differences in the development pattern of aerial mycelia from the wild type and the ovc mutant, expression of a conidiation gene, con-10, was also examined. No difference is observed in con-10 mRNA levels. Finally, no significant difference is found in the mRNA levels of the wc-1 gene when comparing wild type with ovc (Fig. 4B).

Figure 4.

Effect of light on the expression of the al-2, con-10 and wc-1 genes in the wild type and in the ovc mutant. Northern blots of total RNA from the wild type and the ovc mutant grown in the dark for 2 days and in the light for the times indicated (minutes in A and B, hours in C). rRNA bands are shown below each panel as load control.

Cloning of a DNA segment that relieves the 4% NaCl sensitivity of the ovc mutant

We used wild-type growth on 4% NaCl-supplemented solid medium as a selective trait to identify the gene responsible for the osmotic sensitivity of the ovc mutant. Spheroplasts of the ovc strain were first transformed with 50 pools of DNA each representing 96 clones from one microtitre dish of a Neurospora cosmid library and grown on hygromycin-supplemented solid medium. Many plates contained hundreds of transformants, but some of them contained less than 100 or in some cases just a few, indicating only a partial screening of the library. Conidia were collected as a pool from each transformation plate and used to inoculate the middle of NaCl-supplemented plates. Slow-growing mycelia expanded from the inoculated surface of every plate except the one corresponding to pool X13 where a vigorous growth extending to the entire surface of the medium was observed. Upon confirmation of this result, transformations were carried out with 20 subpools with combinations of the 96 cosmids of the X13 pool. Using the same detection method, cosmid X13:E12 was identified as the one conferring normal growth to ovc in high NaCl concentrations.

To delimit internal DNA segments containing the whole gene, cosmid E12 was digested with different restriction enzymes and used for transformation of ovc spheroplasts. Complementation is achieved upon transformation with the cosmid digested with BamHI, HindIII, KpnI or EcoRI, but not with PstI or SacI. Different DNA segments obtained with the former enzymes were subcloned and used for further transformation experiments. A plasmid with a 10 kb EcoRI segment (plasmid pJA10) complements the ovc phenotype and was chosen for further subcloning. Digestion of pJA10 with KpnI results in three DNA segments of approximately 6 kb, 3.5 kb and 2.8 kb. Transformation experiments with the three fragments found complementation only with a plasmid containing the 6 kb KpnI fragment, called pJA23, indicating the presence of the gene responsible for the complementation.

Spheroplasts of the cut mutant were transformed with plasmids pJA10 and PJA23, and with a negative control for the complementation test. Complementation of 4% NaCl sensitivity is achieved with pJA10 or PJA23, confirming that cut and ovc either are mutated in a single gene or are in very closely linked genes.

The sequence from one of the ends of the 6 kb DNA segment of pJA23 was determined and located through blast analysis in the Neurospora genome (release 3). The sequence is contained in contig 276 (supercontig 17), still unassigned to a linkage group. Former genetic mapping of the cut and ovc mutations in LG IV (Harding et al., 1984; Banks et al., 1997) allows assignation of contig 276 to this linkage group. The restriction map of the sequence obtained from the Neurospora genome database confirms the information available on the cloned DNA segment. The EcoRI insert of plasmid pJA10 is 9438 bp long and includes two KpnI sites that subdivide it into three fragments of 539, 2640 and 6259 bp (Fig. 5A).

Identification of the cut-1 gene

The 6259 bp KpnI DNA segment restores the ability of the ovc and cut mutants to grow robustly on 4% NaCl and contains two annotated genes that are transcribed divergently (Fig. 5A). One of them codes for the hypothetical protein NCU04923.1, of 331 amino acids, with sequence similarity to enzymes of the aldo/keto reductase family. This family includes different structurally related NADPH-dependent oxidoreductases, such as aldehyde reductase, aldose reductase, prostaglandin F synthase and xylose reductase. The second gene codes for the hypothetical protein NCU04924.1, with sequence similarity to proteins of the haloacid dehalogenase (HAD) superfamily. DNA segments containing one or the other of the two genes were obtained using a unique HindIII site (Fig. 5A), subcloned into different plasmids and used to transform the ovc and cut mutants. Only the DNA fragment encoding the HAD-like protein is able to restore normal growth in the presence of high NaCl concentrations. Because both mutants coincide in the ‘cut’ morphological phenotype, we have named this gene cut-1.

Transformation of the ovc and cut mutants with the plasmid harbouring the cut-1 gene restores normal growth on 4% NaCl and complements all other ovc/cut phenotypic traits. The alteration of aerial development (the ‘cut’ phenotype) disappeared to a different extent in different transformants, with some of them exhibiting a wild-type appearance (example T2, Fig. 5B). Although selective pressure was kept by incubating the transformants in the presence of hygromycin, no efforts were made to obtain homokaryotic derivatives from the original heterokaryons, suggesting a dominant effect of the wild-type cut-1 allele introduced in different copy numbers and locations in the different transformants (Figs 5B and 6A). Introduction of cut-1 into the ovc mutant results in the recovery of normal photoinduction of carotenoid biosynthesis (Fig. 6B).

Figure 6.

Complementation of the ovc phenotype.
A. Southern blot of genomic DNA from the wild type (wt), the ovc mutant and ovc-derived transformants T1, T3, T6 and T23 digested with either EcoRI or HindIII. Sizes markers (M) are indicated in kb on the right.
B. Carotenoids accumulated by the wild type (wt), the ovc mutant and transformants T1, T3, T6 and T23. The cultures were incubated for 2 days in the dark and 1 day in the light.

The cut-1 alleles of the cut and ovc mutants

Southern blot analyses carried out to investigate the presence of additional copies of cut-1 in the transformants reveal the presence of the endogenous copy in the cut mutant (Fig. 5B), but not in ovc (Fig. 6A). Additional Southern blot analyses confirmed the absence of the cut-1 gene in the ovc mutant, even when the 9.438 pb EcoRI DNA segment containing cut-1 and NCU04923.1 was used as a probe (results not shown), indicating that this strain has a deletion of at least this size. In agreement with this result, we were able to obtain the cut-1 gene by polymerase chain reaction (PCR) from genomic DNA samples of the wild type, but not from the ovc mutant. Preliminary PCR experiments on surrounding DNA indicate that the deleted region extends over a wide region covering additional hypothetical genes. The precise limits of the deletion are under investigation.

The sequence of the cut-1 allele in the cut mutant was determined. When compared with the wild-type sequence, this allele contains a single point mutation, a G-to-A transition at position 884 of the gene. The mutation converts a TGG tryptophan codon (residue 295) into the stop codon TAG, resulting in a predicted truncated protein of 294 amino acids (Fig. 7A), that is, about 56% of the wild-type protein.

Figure 7.

Sequence features of the gene cut-1.
A. Relative positions of the HAD domain (shaded segment) in the predicted proteins O13899 from Schizosaccharomyces pombe (Sp), Q9BXW7 from Homo sapiens (Hs) and CUT-1 from N. crassa. An arrowhead indicates the stop mutation in the cut mutant. CDP-AP indicates the CDP-alcohol phosphatidyltransferase domain of the O13899 protein.
B. Amino acid sequence from the amino segment of the CUT-1 protein preceding the HAD domain. White box, poly glutamine domain; grey box, putative PEST domain; underlined, predicted nuclear localization signals.

cut-1 sequence and expression

The gene cut-1 (NCU04924.1) extends along a non-interrupted open reading frame (ORF) of 1578 bp, coding for a predicted protein of 526 amino acids. Sequence similarity to proteins of the HAD family covers 70% of the CUT-1 polypeptide, extending to the carboxy-terminus (Fig. 7A). The HAD domain is present in a diverse group of proteins, many of them with phosphatase activity. The HAD superfamily includes at least five known subfamilies: haloalkanoate dehalogenases, phosphonoacetaldehyde hydrolases, phosphate monoesterases, ATPases (hydrolysing C-Cl, P-C, P-OC and P-OP bonds respectively) and phosphomutases (cleaving a P-OC bond followed by a intramolecular phosphoryl group transfer). The enzymatic activity of most proteins assigned by sequence similarity to this large family remains to be elucidated. The highest similarity to CUT-1 is found with a subgroup of these proteins, none of them with an identified function. Two of them are schematically compared with CUT-1 in Fig. 7A.

The amino-terminal region of the protein preceding the HAD domain, approximately the first 150 residues, contains short sequences that could play a regulatory role (Fig. 7B), such as a glutamine-rich segment, QQQQEQPQQ, and a putative PEST element. A k-NN prediction from the protein sequence on cell localization attributes a 52% probability of being cytoplasmic and 30% of being nuclear, but a search for nuclear localization signals finds three putative elements located in the same region.

Expression of cut-1 is very low in standard Vogel's liquid medium (Fig. 8A). The presence of NaCl in the medium results in an enhanced expression of the gene over a range of 4% to 8% NaCl, with maximal expression at 6%. A similar transcriptional induction was found if NaCl was replaced by other salts or by an equivalent osmomolar sorbitol concentration. We conclude that cut-1 transcription is regulated by the osmotic pressure of the medium.

Figure 8.

Effect of osmotic pressure, heat shock and light on cut-1 expression.
A. Northern blots of total RNA from mycelia grown in the dark in VGM medium supplemented with different concentrations of NaCl or with other salts and sorbitol at an osmomolar concentration equivalent to 4 W/V per cent NaCl.
B. Northern blots of total RNA from mycelia grown for 2 days in the dark at 30°C and for different times at 50°C (indicated in min) or for 30 min under light at 30°C.

A search for conserved binding sites for fungal transcription factors in the upstream regulatory region of cut-1 reveals the presence of many putative regulatory elements. Three of them are recognized as yeast AGGGG SRE elements (Fig. 5A), which mediate the osmotic HOG induction of Saccharomyces cerevisiae (Schueller et al., 1994). The cut-1 regulatory region is shared with the neighbour divergently transcribed gene, to which one of the SRE elements is closely located. Three predicted regulatory elements for heat-shock factors (Fernández et al., 1994) are also found in the same regulatory sequence, two of them close to the cut-1 coding region (Fig. 5A). Accordingly, a significant induction of the cut-1 mRNA levels is found upon incubation of wild-type mycelia for 30–90 min at 50°C (Fig. 8B). Illumination during 30 min of a parallel wild-type mycelial sample shows no light induction of cut-1 transcription.

Discussion

We have identified the gene responsible for the ovc and cut phenotypes of Neurospora that we have called cut-1. Our results confirm previous phenotypic traits of the ovc and cut mutants and genetic data showing a close linkage between the mutations (Banks et al., 1997), suggesting allelism. The cut and ovc mutants coincide in their hyperosmotic sensitivity and in a characteristic developmental alteration known as ‘cut’ phenotype. Complementation analyses demonstrate that both phenotypic traits are attributed to lack of a functional cut-1 gene and molecular analysis of their sequences has found different alterations in the gene sequence, a large deletion in the case of ovc and a premature stop mutation in the case of cut. The deletion covers the cut-1 allele and precludes allelic recombination.

The ovc mutant also exhibits an increased carotenoid accumulation in the light, actually the reason for its identification and different denomination (Harding et al., 1984). This trait is not manifested by the cut mutant. An obvious explanation is that the carotenoid overaccumulation is a side-effect of the deletion, caused by the lack of a different unrelated gene. Experiments to determine the extension of the deletion show it covers a large DNA segment,  affecting  at  least  several  neighbouring  genes (L. Youssar and J. Avalos, unpubl. results). However, the introduction of the wild-type cut-1 allele to the ovc mutant restores a normal carotenoid photoinduction, associating unequivocally this gene to the ovc phenotype. Lack of the carotenoid overproduction trait in the cut mutant could be explained by a leaky activity of its abnormal CUT-1 protein, but its severe truncation makes this hypothesis unlikely. Moreover, at least four cut mutants have been described (Mays, 1969), but no mention was made of any colour alteration. An alternative explanation is the involvement of a second gene, whose simultaneous loss with cut-1 would result in the carotenoid overaccumulation phenotype. Such a synergistic effect would be lost by recovery of either gene. Experiments to test this hypothesis are under way.

Carotenoid overaccumulation by the ovc mutant is similar to the vivid phenotype. In contrast to what was found in this mutant (Schwerdtfeger and Linden, 2001), we did not notice any significant change in the transcriptional regulation by light of the genes responsible for carotenoid biosynthesis in the ovc strain. mRNA levels of al-1 or al-2 exhibit the same light induction and adaptation as in the wild type. No relevant change was found in the mRNA levels either of a photoregulated conidiation gene, con-10, or of the wc-1 photoreceptor gene. The higher carotenoid accumulation of the ovc mutant in the light could have other causes, such as an enhanced enzyme activity or stability or a higher capacity for carotenoid storage of the mutant compared with the wild type.

The CUT-1 protein has sequence similarity with a large group of enzymes known as the HAD superfamily (Koonin and Tatusov, 1994; Thaller et al., 1998). The name derives from the HADs, a class of bacterial detoxifying enzymes that hydrolyse carbon–halogen bonds in many chlorinated compounds (Hardman, 1991). Sequence comparison of these proteins in sequence databases reveals conserved motifs in a large family of enzymes, including some epoxide hydrolases, a long series of different phosphatases and many other proteins of unknown function (Koonin and Tatusov, 1994). All the known enzymatic reactions have in common the hydrolysis of different substrates, most of them involving the break of a phosphate bond. The highest similarity of CUT-1 with a protein with a presumed function was found with the HAD domain from a hypothetical CDP-alcohol phosphatidyltransferase of Schizosaccharomyces pombe.

The osmosensitive phenotype of the ovc and cut strains, resulting in a defective ability to grow on media containing high salt concentrations, closely resembles that of the os mutants. The four Neurospora os genes identified at the molecular level, os-1, os-2, os-4 and os-5, are recognized as components of a single transduction chain known in S. cerevisiae as the HOG response. Resistance to phenylpyrrole antibiotics, used as fungicides in agriculture, is a characteristic trait of the mutants of this regulatory pathway. These chemicals induce a glycerol biosynthesis response in the wild type, and its impairment in the os mutants explains their resistance. Interestingly, the cut mutant was found to be sensitive to the phenylpyrrole antibiotic fludioxonil, although this chemical does not induce glycerol accumulation in this strain (Fujimura et al., 2000). We have found that the cut mutant is partially resistant to lower concentrations of this antibiotic, reinforcing the connection of cut-1 with the HOG response of Neurospora. Unexpectedly, the ovc strain exhibits a higher sensitivity to fludioxonil than the cut strain.

The HOG transduction pathway includes a two-component histidine kinase and a series of MAP kinases (MAPK, MAPKK and MAPKKK). cut-1 is the first gene of this phenotypic class whose product does not have an obvious role as a member of this transduction chain. Predominance of phosphatases among the identified enzyme activities of the HAD superfamily proteins evokes a possible regulatory role restoring dephosphorylated states of chain components. Transcription of cut-1 is induced by high-osmolarity conditions, suggesting a connection with the osmolarity signalling pathway. No information is available on the transcriptional regulation of the Neurospora os genes. Preliminary experiments with os-1 and os-4 gene probes show little or no induction of their mRNA levels on media supplemented with 4% NaCl (L. Youssar, unpubl.), suggesting that these genes are regulated by a different mechanism from cut-1. Furthermore, heat-shock activation of cut-1 transcription suggests a more general role of this gene in the Neurospora response to different stress conditions.

Additional clues on cut-1 function may be deduced from the presence in its amino region of a potential PEST domain, a glutamine-rich segment, and several nuclear localization signals. PEST domains are polypeptide sequences rich in proline (P), glutamate (E), serine (S) and threonine (T), which serve as signals for controlled protein degradation (Rechsteiner and Rogers, 1996). A different role is attributed to glutamine rich tracts, usually associated to transcriptional regulation involving protein–protein interactions (Gerber et al., 1994; Tanaka et al., 1994). Several regulatory proteins of Neurospora contain functional glutamine-rich domains. One of them is WC-1, with an amino polyglutamine domain playing a role in clock function and photocarotenogenesis (Toyota et al., 2002). The length of these regulatory domains frequently exceeds 20 residues, more than 40 in the case of WC-1, much longer than the modest 9 residues of CUT-1.

Taken together, sequence predictions suggest a role for CUT-1 as a nuclear phosphatase able to interact with target proteins, and subject to a rapid turnover mechanism. This is just a tentative hypothesis, first because the relevance of the regulatory amino domains remains to be elucidated, and second because the similarity of CUT-1 with any of the different protein subgroups of the HAD superfamily is not high enough for a solid functional hypothesis. Actually, many members of this family have been identified by sequence similarity and their biochemical activities have not been investigated. CUT-1 might represent a novel type of protein in this superfamily whose function, including a possible role in the Neurospora HOG response, remains to be determined. Experiments on targeted mutagenesis of the predicted regulatory domains will allow to investigate their role, and heterologous CUT-1 expression and purification will provide the tool to gain information on its cellular location and enzymatic activity.

Experimental procedures

Strains and culture conditions

The wild-type strain of N. crassa is Oak Ridge 74-OR23-1A. Mutant strains were ovc a (FGSC 4503), cut A (allele LLM1, FGSC 2385) and vvd a (vivid SS-692, FGSC 7853).

Unless otherwise stated, incubations were for 3 days as submerged cultures at 30°C in Petri dishes with 25 ml of liquid Vogel's medium (VMS) supplemented with 0.2% tween80 to avoid aerial development, inoculated with 105 conidia. The dicarboximide anti-fungal fludioxonil, obtained from Sigma, was added directly to the medium after autoclaving at final concentrations of either 10 mg l−1 or 50 mg l−1. When required, solid VMS medium was supplemented with 250 mg l−1 hygromycin.

For analysis of carotenoid production of agar cultures, the wild type and the ovc mutant were grown in a 3 l plastic beaker with 100 ml of solid VMS medium, covered with polyvinylidene chloride (PVDC) plastic film. Conidia were removed by washing before separating mycelia from the agar.

Incubations in the light were performed at either 8°C or 30°C under white fluorescents with a light intensity of 10 W m−2. Dark manipulations were performed under safe red light. Heat-shock incubations were performed placing the Petri dishes partially submerged in a water bath at 50°C. For RNA isolation, the mycelia were separated from the medium and frozen in liquid nitrogen immediately.

Neurospora transformation, cloning of cut-1 and complementation analysis

Spheroplasts prepared following Royer and Yamashiro (1992) were transformed according to Vollmer and Yanofsky (1986). Protoplasts from the ovc mutant were transformed with the 50 96-clone pools of the Orbach-Sachs genomic cosmid library (Orbach, 1994). Transformations were performed as described except that the 10xFGS solution contained 5% sorbose, 2% fructose and 2% glucose. Under these conditions, transformants grow semi-colonially, allowing calculation of their approximate number, but the colonies mix and conidiate abundantly. Conidia were collected from each plate and diluted to an approximate concentration of 5 × 106 conidia per millilitre. For complementation tests, a 2 µl sample of the conidia mixture from each transformation was placed in the middle of a VMS plate supplemented with 4% NaCl, incubated at 34°C and checked for growth in the following days. In most cases, only a restricted growth was observed which did not reach the edges of the plate after 4–5 days of incubation. In case of complementation, vigorous growth was observed covering the surface of the plate in less than 3 days.

Fragments from cosmid X13/E12 were subcloned in the polylinker sequence of Bluescript KS- (Stratagene). Complementation tests with subclones from this cosmid were performed by cotransformation with plasmid pMOcosX (Orbach, 1994). Serial subcloning led to plasmid pLY2, containing a 3.6 kb HindIII–EcoRI DNA segment with the entire cut-1 coding sequence (Fig. 5A).

Nucleic acids manipulations and hybridization analyses

Genomic DNA was isolated following the method described by Lee and Taylor (1990). For Southern blots, Neurospora genomic DNA (1–4 µg), digested to completion with a restriction enzyme as required, was separated by electrophoresis in a 0.8% agarose gel, transferred to a nylon membrane (Hybond-N, Amersham Biosciences), hybridized in 50% formamide with a probe labelled with digoxigenin-11-dUTP at 42°C, and detected with Lumigen PPS following the manufacturer's recommendations (Boehringer Mannheim). Unless indicated, the probe used was the 3.6 kb HindIII–EcoRI segment containing the cut-1 gene from plasmid pLY2.

Northern blot hybridizations were carried out according to Sambrook and Russell (2001). RNA samples were isolated with the Perfect RNA eukaryotic mini kit (Eppendorf). rDNA bands transferred to the membranes were stained with methylene blue as described (Di Pietro and Roncero, 1998) and used as load controls. 32P-labelled probes were prepared by standard random oligomer priming. Probes were obtained as follows: for al-1, a 1990 bp segment containing the whole coding sequence was obtained by PCR with the oligonucleotide pair (5′-ACTTACAGACAAAATGGCTG-3′, 5′-AACCCTACCTCACAAATAGC-3′). Similarly, the oligonucleotide pairs (5′-CCCAAGATGTACGACTATGC-3′, 5′-CCGTC TACTGCTCATACAAC-3′), (5′-GTTCTGATACCCCATCCATC -3′, 5′-AGCCAAGCCAATCAATGCCC-3′), and (5′-CACGAT TCACGCTTCAGATG-3′, 5′-CAAACTCGGCTCTGGATTCC-3′) were used to obtain DNA segments containing coding sequences for genes al-2 (1947 bp), cut-1 (1689 bp) and wc-1 (2021 bp) respectively. For con-10, a 250 bp HindIII–XbaI fragment from plasmid pBW100 (Madi et al., 1994) was used as a probe.

Escherichia coli DH5α was used for the multiplication of plasmids. DNA was extracted from agarose gels after electrophoresis with the Qiaex II kit (Qiagen). For other DNA manipulations and technical details see Sambrook and Russell (2001).

PCR and sequence analyses

Polymerase chain reactions were performed in a final volume of 50 µl with 5 ng of genomic DNA, 0.2 mM dNTPs, 1 µM each primer, and 0.5 µl of either the DNA polymerase Expand High Fidelity (Roche) or the Triplemaster PCR system (Eppendorf AG). After a first incubation at 94°C for 2 min, the mixtures were subject of 35 cycles of incubations at 94°C for 0.5 min, 55°C for 0.5 min and 72°C for 1 min, and a final incubation at 72°C for 5 min in a Techne ftgene2d programmable thermocycler.

The cut-1 allele of the cut mutant was cloned by PCR amplification with the oligonucleotide pair (5′-GTTCTGAT ACCCCATCCATC-3′, 5′-AGCCAAGCCAATCAATGCCC-3′), giving a DNA segment of 1689 bp. The sequence of the central region of this DNA segment was obtained from the PCR product resulting from the oligonucleotide pair (5′-GGTCTTGGACATGCTCAACG-3′, 5′-AGCCAAGCCAATCA ATGCCC-3′). To determine the mutation, the sequences from three independent PCR products were compared with the wild-type sequence.

Sequence alignments were performed with the clustal x 1.63b computer program (National Center for Biotechnology Information). Potential fungal transcription factor binding sites were determined with the MatInspector tool (Genomatix Software GMBH; Quandt et al., 1995). The PEST domain was identified with the algorithm PESTFind through the EmbNet Server (Vienna, Austria, http://www.at.embnet.org/embnet/tools/bio/PESTfind/). Prediction of protein subcellular localization was performed with the psort ii program (Nakai and Horton, 1999) through the psort WWW Server (Tokyo, Japan, http://psort.ims.u-tokyo.ac.jp/).

Carotenoid analysis

Mycelial samples were dried on filter paper, frozen and lyophilized. Carotenoids were extracted from 0.1 to 0.2 g dry weight samples as described (Arrach et al., 2002). Total carotenoids were estimated  from maximal absorption spectra in hexane assuming an average maximal E (1 mg l−1, 1 cm) = 200. Phytoene, neutral and polar carotenoids were determined as described by Avalos and Cerdá-Olmedo (1986).

Acknowledgements

We thank C. Vallejo and L. Pérez de Camino for technical assistance. This work was supported by the European Union (Project QLK1-CT-2001-00780) and the Spanish Government (INIA, Project PB96-1336, Ministerio de Ciencia y Tecnología, Project BIO2003-01548; and Ministerio de Educación y Cultura, short-term fellowship).

Ancillary