Histone H1 Is Required for Proper Regulation of Pyruvate Decarboxylase Gene Expression in Neurosporacrassa

We show that Neurospora crassa has a single histone H1 gene,hH1, which encodes a typical linker histone with highly basicN- and C-terminal tails and a central globular domain. A greenfluorescent protein-tagged histone H1 chimeric protein was localizedexclusively to nuclei. Mutation of hH1 by repeat-induced pointmutation (RIP) did not result in detectable defects in morphology,DNA methylation, mutagen sensitivity, DNA repair, fertility,RIP, chromosome pairing, or chromosome segregation. Nevertheless,hH1 mutants had mycelial elongation rates that were lower thannormal on all tested carbon sources. This slow linear growthphenotype, however, was less evident on medium containing ethanol.The pyruvate decarboxylase gene, cfp, was abnormally derepressedin hH1 mutants on ethanol-containing medium. This derepressionwas also found when an ectopically integrated fusion of thecfp gene promoter to the reporter gene hph was analyzed. Thus,Neurospora histone H1 is required for the proper regulationof cfp, a gene with a key role in the respiratory-fermentativepathway.

INTRODUCTION

Top

Introduction

Eukaryotic DNA is compacted into chromatin, a dynamic nucleoproteincomplex organized into nucleosomes (31). Each nucleosome containsan octameric core of four conserved histones, H2A, H2B, H3,and H4, around which $~$ 146 bp of DNA are wrapped (35). The N-terminaltails of core histones are thought to interact with nuclearproteins and to be involved in internucleosomal interactionsthat can lead to chromatin condensation (36, 71). Posttranslationalmodifications of histone tails have been shown to be involvedin various aspects of gene regulation, suggesting the existenceof a "histone code" specifying gene silencing, activation, andother processes (6, 26). Chromatin of most eukaryotes includesless conserved "linker histones," most notably, histone H1 (28,29, 63). Histone H1 has been proposed to play a role in creatingand/or maintaining the higher-order structure of chromatin (52).In contrast to core histones, linker histones appear rathermobile and in fact are not permanently associated with chromatin(33, 44). The results of in vitro studies have suggested thatH1 participates in the organization of nucleosomes (62), restrictsthe translational mobility of nucleosomes (48, 66), and caninhibit transcription (70). Interestingly, H1 is enriched innuclease-resistant and inactive chromatin (69) and in regionsof DNA that are methylated (3). Additional clues to the possiblefunctions of H1 have come from in vivo studies. In higher eukaryotes,which have multiple linker histone genes subject to developmentalregulation (30), some H1 variants seem to be dispensable (13,60). In tobacco, H1 plays roles in development and male meiosis(49). An H1 variant is essential for the development of thegerm line in Caenorhabditis elegans (25). Changes in H1 stoichiometrymodulate core histone acetylation in cultured mammalian cells(18).

The functions of H1 have been studied most extensively for Tetrahymenaand fungi. Both Tetrahymena thermophila and Saccharomyces cerevisiaehave single genes for linker histones with unconventional structures(5, 72). The Tetrahymena gene encodes a lysine- and alanine-richprotein that lacks the globular domain, whereas the yeast proteinconsists almost entirely of two globular domains without a longC-terminal tail. Deletion of the linker histone gene did notlead to gross phenotypic changes in either organism. Nevertheless,the Tetrahymena linker histone mutant showed increased nuclearsize (57) as well as abnormal activation of some genes and repressionof others (58). Disruption of the yeast H1-like gene causeddecreases in the steady-state levels of a substantial numberof unrelated mRNAs (22).

Filamentous fungi have canonical H1 linker histones. Ascobolusimmersus contains a single H1 gene encoding a typical tripartiteH1 protein with a charged amino-terminal domain, a globularwinged helix domain, and a charged alanine- and lysine-richcarboxy-terminal domain (4). Ascobolus strains in which theH1 gene was silenced by methylation showed increased sensitivityof chromatin to micrococcal nuclease (MNase), global DNA hypermethylation,and a shortened "life span" (4). These phenotypes are consistentwith important roles of H1 in global chromatin function andgene regulation (2). The dispensability of the equivalent genein Aspergillus nidulans (hhoA) (53), however, challenged theidea that canonical H1 proteins are essential in eukaryotes.We chose Neurospora crassa to further define the functions oflinker histones. This fungus exhibits DNA methylation (32, 61),repeat-induced point mutation (RIP) (55), and a fully characterizedset of core histones (21); in addition, as in A. immersus, nucleosomalrepeats are slightly longer than in Aspergillus and yeast cells(53).

Here we report the cloning and characterization of the hH1 geneof N. crassa. This gene encodes a typical histone H1 protein.We constructed viable hH1 null mutants, demonstrating that hH1is not essential in Neurospora. We found, however, that H1 mutantsgrew more slowly than wild-type strains on the various testedcarbon sources, and we observed misregulation of the cfp (cellularfilament polypeptide) gene (1), which encodes Neurospora pyruvatedecarboxylase. Our results indicate that histone H1 is requiredfor the proper regulation of cfp expression.

MATERIALS AND METHODS

Top

Materials and Methods

Strains and genetic methods. Standard Neurospora methods were used (10). The N. crassa strainsused in this work are described in Table 1. Strain N228 carriesa duplication of the al-1 gene and was kindly provided by C.Staben (University of Kentucky). Strains N1445 and N1264 weredescribed previously (32, 43). Strains HH3 (hH1^RIP his-3) andHH28 (his-3) were obtained by a cross between DF3 (mat a hH1^RIP)and a his-3 mutant (FGSC 6103). Strains used to measure RIPand linear growth were constructed for this study as detailedin Tables 2 and 3, respectively.

View this table:
[in this window]
[in a new window]
TABLE 1. N. crassa strains used in this study

View this table:
[in this window]
[in a new window]
TABLE 2. Lack of effect of disruption of hH1 on RIP mutations

View this table:
[in this window]
[in a new window]
TABLE 3. Linear growth rates for wild-type and hH1^RIP mutant strains

Nucleic acid manipulations. N. crassa genomic DNA was prepared and analyzed as previouslydescribed (20, 37). For Southern blot analyses, 0.5 to 1.0 µgof DNA was used. Total mycelial RNA was purified by a methodused for plant tissues (68). Northern analyses were performedas described previously (37). The following probes were usedfor Southern and Northern hybridizations: a 1.2-kb EcoRI hH1fragmentfrom pDF2 (see below), a 1.7-kb BamHI-EcoRI hH1 fragment frompMF233 (see below), 0.6-kb HindIII and 0.2-kb HindIII-EcoRIhH1 fragments from pDF1 (see below), a 537-bp PCR fragment (positions+460 to +997) and a 338-bp PCR fragment (positions -348 to -10)from cfp, a 0.9-kb BanII-EcoO109I ${Psi}$ ₆₃ (Fsr-63; 5S RNA pseudogene)fragment from pPG22 (43), an 0.8-kb BamHI ${zeta}$ - ${eta}$ (Fsr-33; 5S RNApseudogene) fragment from pVM152 (43), a 2.7-kb HindIII his-3fragment from pBM60 (38), and a 1.1-kb BamHI-ClaI hph fragmentfrom pCSN44. For methylation analyses, genomic DNA was digestedwith methylation-insensitive DpnII or its methylation-sensitiveisoschizomer Sau3AI. The his-3 probe was used as a control probeto verify that digestions were complete.

Protein purification and electrophoresis. Protein purification and electrophoresis were carried out asdescribed previously (53).

Cloning of the hH1 gene. Genomic DNA (N. crassa FGSC 987) and DNA purified from platestocks of phage ${lambda}$ ZAPII from a mycelial cDNA N. crassa library(47) were used as templates in PCR experiments with oligonucleotidesNC1 (5'-ACCTTTTGCACGCCCCTT-3') and HDF1 (5'-AGTGGTATCGTTCCAAGACGG-3').PCR products of 1.7 kb (genomic DNA) and 1.2 kb (cDNA) wereobtained, cloned into the p-GEM T Easy vector (Promega) (plasmidspDF1 and pDF2, respectively), and sequenced.

Because the hH1 gene is not available in Neurospora cosmid libraries,a genomic library of wild-type N. crassa strain FGSC 987 wasconstructed in the ${lambda}$ DASHII vector by using genomic DNA partiallydigested with Sau3AI (10 to 18 kb) and an in vitro packagingkit (Gigapack II Gold; Stratagene). A positive clone containingthe whole hH1 gene was isolated by screening with the 0.2-kbHindIII-EcoRI genomic fragment from pDF1. Oligonucleotide primerH1rt (5'-GGATGTGGTTGTGGTTTTAGG-3') was used to sequence a 250-bpregion upstream of the ATG start translation codon. 5' Rapidamplification of cDNA ends (RACE) was carried out with a 5'/3'Race kit (Boehringer Mannheim, Mannheim, Germany) and oligonucleotidesNC2 (5'-TTCTTGGCAAGCTTGGTGC-3') and NC4 (5'-CTTGTTGAAGAGCGAGTC-3').The hH1 gene was mapped by restriction fragment length polymorphism(RFLP) analysis (42) with the 0.6-kb HindIII genomic fragmentfrom pDF1 as a probe.

EGFP fusion constructs. To generate histone H1 fusions to enhanced green fluorescentprotein (EGFP), the BamHI-BglII fragment from plasmid pFA6a-GFP(S65T)-kanMX6(34) was inserted into BamHI-digested pBM60 (38) to yield pMF255.The GFP (S65T) gene was replaced by a PCR-amplified EGFP gene(19), and the fragment was inserted into BamHI-EcoRI-digestedpMF255, yielding pMF267. The Neurospora ccg-1 promoter (Pccg-1)and 5' untranslated region of the transcript directly precedingthe ccg-1 ATG (nucleotides 738 to 1659) were PCR amplified andinserted into NotI-XbaI-digested pMF267, yielding pMF272. PCR-amplifiedNeurospora hH1 was inserted into BamHI-PacI-digested pMF272to yield pMF280. Both pMF272 and pMF280 were targeted to thehis-3 locus of N623 (38), yielding strains N2261 and N2276,respectively. EGFP imaging was performed as previously described(15).

Generating hH1 mutants by RIP A 1.7-kb BamHI-EcoRI PCR fragment from hH1 DNA cloned into pBM61(pMF233) was targeted to the his-3 locus (38). Transformantswere crossed to strain N1264, and random ascospores were isolated28 days after fertilization. Strain N1264 is mat A and carriesnic-3, which was determined by RFLP analysis (see above) tomap closely to hH1 on LG VIIL. Therefore, selection for growthon minimal medium and screening for mat A increased the likelihoodof isolating strains with a single mutated hH1 copy at the nativelocus. We isolated 43 prototrophs, 27 of which were mat A and22 of which showed evidence of RIP mutations. The mutated allelesof strains N1815 (hH1^RIP1), N1817 (hH1^RIP2), N1816 (hH1^RIP3),and N1819 (hH1^RIP4) were amplified by PCR with Pfu DNA polymerase.Products from at least five reactions per strain were pooledprior to sequencing.

Construction of the ectopic cfp-hph reporter locus. Plasmid pDF12 was generated by ligation of the 2.7-kb ApaI-SacIfragment from pA3, which carries hph under the control of thecfp promoter and the Aspergillus nidulans trpC terminator (E.D. Temporini and A. L. Rosa, unpublished data), into ApaI-SmaI-digestedpBM61 (38). Control plasmid pDF85, with a trpC promoter insteadof the cfp promoter, was constructed by ligating a 2.4-kb ApaI-NotIfragment from pCSN44 into ApaI-NotI-digested pBM61. These constructswere targeted to his-3 of strains HH3 and HH28, and correctintegrations were confirmed by Southern analysis.

MNase digestion of chromatin. Chromatin structure was analyzed by a modification of a methoddescribed for A. nidulans (17). Briefly, N. crassa was grownat 30°C in Vogel's minimal medium with 2% sucrose for 16h. Ethanol (2%) or glucose (2%) was added, and mycelia wereharvested 4 h later and used immediately or freeze-dried. Frozenand powdered mycelia were suspended (100 mg/ml) in a buffercontaining 250 mM sucrose, 60 mM KCl, 15 mM NaCl, 0.05 mM CaCl₂,3.0 mM MgCl₂, 0.5 mM dithiothreitol, and 15 mM Tris-HCl (pH7.5), and aliquots were treated with various amounts of MNasefor 5 min at 30°C. DNA was purified by phenol extractionand ethanol precipitation, digested with either EcoRI or BglII,and subjected to Southern blot analysis and probing with the537-bp PCR fragment (+460 to +997) or the 338-bp PCR fragment(-348 to -10) from cfp, respectively.

Sensitivity to MMS, UV, and DMSO. Neurospora conidia were plated at a density of 500 conidia/plateon Vogel's minimal medium containing 1.5% sorbose, 0.1% glucose,and 0.1% fructose and supplemented with methyl methanesulfonate(MMS) at 0.005, 0.015, or 0.030% (vol/vol). Colonies were countedat 48 and 72 h (24). UV sensitivity was tested as describedpreviously (23). The UV dose was 25 J min^-1, and exposure timeswere 0, 8, 16, 20, and 24 min. For dimethyl sulfoxide (DMSO)assays, equal numbers of conidia ( $~$ 500) were plated on petridishes containing Vogel's minimal medium with 4.5% DMSO. Colonieswere counted after 72 h at 30°C, and morphology was noted.

Measurement of linear growth rates with Race tubes Linear growth rates were determined at 30°C with glass Racetubes (40 cm long, 12 mm in diameter) (10). Ten microlitersof a conidial suspension ( $~$ 10⁴/ml) was inoculated at one end,and growth was monitored for 4 days at intervals of 8 to 12h.

Measurement of RIP frequencies. Crosses were carried out on Westergaard-Mitchell medium (10)after the "female" strain had grown for 5 days at 25°C inthe dark. Random ascospores were harvested, heat activated,and plated on Vogel's minimal medium containing 2% sucrose and0.0075% Tergitol (Sigma) (59). White (RIP-mutated) and orange(wild-type) colonies were scored by visual inspection.

Microscopic examination of vegetative and sexual tissues. Asci and chromosomes were visualized by staining with hematoxylin-ferricacetate or acriflavine (50). The number, size, and shape ofnuclei (i.e., mycelia and conidia) were determined after staining(Hoechst 33258 or 4',6'-diamidino-2-phenylindole [DAPI] dyes)and visualization under a fluorescence microscope. Imaging ofEGFP was carried out as described previously (15). Strains weregrown overnight at room temperature in a thin layer of Vogel'sminimal medium without a carbon source on microscope slidesto induce the ccg-1 promoter.

Nucleotide sequence accession number. The GenBank accession number for the sequence reported in thisarticle is AY124883.

RESULTS

Top

Results

Isolation of the hH1 gene from N. crassa. We identified a histone H1 cDNA clone in a Neurospora expressedsequence tag (EST) database (b7 h12ne; http://www.genome.ou.edu)and used it to design specific primers to isolate the hH1 genefrom Neurospora genomic DNA and from a mycelial cDNA library(47) by PCR. We isolated and sequenced a 1.7-kb genomic fragmentas well as a 1.2-kb cDNA fragment. Because of uncertaintiesregarding the N-terminal region of the predicted protein andthe promoter region, the entire hH1 gene was cloned from a ${lambda}$ DASHN. crassa genomic library (Fig. 1A). We mapped hH1 to the leftarm of LG VII between 5:5A and 00003 by RFLP analysis (42) andconfirmed by additional Southern analyses that hH1 is the onlyhistone H1 gene in N. crassa (data not shown). This conclusionwas further supported by BLAST searches of the >98% completedNeurospora genome sequence (http://www-genome.wi.mit.edu/annotation/fungi/neurospora/;release 3: 02.12.2002).

View larger version (108K):
[in this window]
[in a new window]
FIG. 1. hH1 gene of N. crassa. (A) Nucleotide and predicted amino acid sequences for the N. crassa hH1 gene. Coding regions are shown in uppercase letters, and noncoding sequences are shown in lowercase letters. Transcription start sites (asterisks) and the 3' end of the longest available EST sequence (black diamond) are indicated. The putative TATA box is shown in bold type. The sequence around the most likely ATG start codon is doubly underlined. The central globular domain of the protein is underlined. Genomic sequences flanking the hH1 gene, from contig 3.396 (http://www-genome.wi.mit.edu/annotation/fungi/neurospora), are shown in italic type. The cytosine at +577 and Q60, mutated in hH1^RIP strains, are indicated in bold type. (B) Comparison of the highly conserved globular domains of histone H1 proteins from N. crassa (Nc); A. fumigatus (Af; sequence deduced from data available at http://tigrblast.tigr.org/ufmg/); A. nidulans (An; CAB72936); Fusarium sporotrichioides (Fs), Magnaporthe grisea (Mg), and Botrytis fuckeliana (Bf) (these three sequences deduced from ESTs availables at http://www.cogeme.ex.ac.uk); A. immersus (Ai; AAF16011); S. cerevisiae (ScD1 and ScD2; NP_015198); Volvox carterii (Vc; H1-1; Q08864); tobacco (Nt; S53502); human (Hs; H1⁰; XP_009973); and Drosophila melanogaster (Dm; P02255). Yeast H1 is unconventional because it does not contain N- or C-terminal tails but rather contains two fused globular domains (ScD1 and ScD2). The globular domains were aligned with CLUSTAL W (64). Identical residues (black) are indicated by asterisks, conservative substitutions of major groups (dark gray) are indicated by colons, and conservative substitutions of minor groups (light gray) are indicated by periods. Regions identified as ${alpha}$ helices or ß sheets in the histone H5 crystal structure are indicated.

Alignment of genomic and cDNA sequences allowed us to determinethe positiond of four introns. The G+C content of the hH1 openreading frame (61.2%) is significantly higher than that of theintrons (46.4%), consistent with the previously described codonbias for constitutively expressed genes of N. crassa (12). Wedetermined the major transcription start sites by RACE experiments(Fig. 1A) and identified a consensus TATA box (8). Northernblot analyses of total RNA extracted from vegetative myceliumrevealed a single 1.4-kb hH1 transcript. RNA levels were similarwhen mycelia were grown on sucrose, glucose, ethanol, or ethanol-glucosemedia (data not shown).

The sequence surrounding the most likely initiation codon ofhH1, CCATCACCATGCC (initiation codon in italic type), fits theconsensus translation start site sequence for Neurospora (7,12) (Fig. 1A). Conceptual translation of the hH1 open readingframe yields a protein of 236 amino acids with a predicted molecularmass of 24.2 kDa and high contents of alanine (26.3%) and lysine(23.3%). This predicted protein shows strong similarity to H1from various organisms, suggesting that hH1 encodes a bona fidehistone H1. Sequence alignments revealed that residues 36 to112 of Neurospora H1 correspond to the globular winged helixdomain, the most conserved region of linker histones (Fig. 1B).Moreover, the Neurospora H1 protein shows the characteristicthree-domain structure of metazoan histone H1 proteins: (i)an N-terminal region of 35 amino acids, (ii) a globular regionof 77 amino acids, and (iii) a positively charged C-terminalregion of 124 amino acids. The predicted isoelectric point (10.2)may account for our observation that H1 migrates as a proteinof 32 kDa (see below), a behavior shared by its homologs fromA. nidulans (53) and A. immersus (4).

An H1-EGFP fusion protein is localized in nuclei. To confirm that Neurospora histone H1 is localized in nuclei,we constructed translational fusions of the H1 and EGFP genes(19) and targeted these fusion constructs to the his-3 locus.Fusion constructs with the native hH1 promoter or the inducibleqa-2 promoter did not yield sufficient H1-EGFP for visualization.We therefore made use of the Neurospora ccg-1 (grg-1) promoter(Pccg-1) (41) to drive the overexpression of EGFP fusion genes(see Materials and Methods). Expression patterns observed weretypical for normal ccg-1 induction; i.e., on sucrose medium,little or no expression was observed in hyphae growing on orthrough agar. Fusions of Pccg-1-H1-EGFP genes produced highlevels of H1-EGFP fusion protein, localized exclusively in thenucleus (Fig. 2a to d). Control fusions of Pccg-1-EGFP genesproduced cytoplasmic EGFP (Fig. 2e and f).

View larger version (77K):
[in this window]
[in a new window]
FIG. 2. EGFP-tagged H1 is localized in nuclei. Overexpressed H1-EGFP fusion protein is exclusively localized in nuclei of fully developed hyphae (a and b) and germinating conidia (c and d) of strain N2276. The nuclear localization of H1-EGFP was confirmed by staining with DAPI (data not shown). Nonfused EGFP expressed under the control of the Pccg-1 promoter is localized in the cytoplasm (e and f) of a germinating conidium of strain N2261. a, c, and e, phase-contrast microscopy; b, d, and f, fluorescence microscopy. EGFP imaging was carried out as described previously (15).

Construction of hH1 mutants. We made use of the RIP process (55) to create hH1 mutants. A1.7-kb fragment of hH1 was introduced at the his-3 locus (LGIR) by gene replacement (38). Single-copy transformants of matingtype mat a (LG IL) were crossed with a mat A strain marked onLG VII (nic-3 wc-1 arg-10). Progeny carrying RIP-mutated allelesat the endogenous hH1 locus were recovered as mat A prototrophs.Southern analysis of genomic DNAs from these strains showedevidence of DNA methylation and C:G to T:A transition mutations,both indicators of RIP (Fig. 3). Four heavily mutated hH1^RIPstrains were characterized in detail. A 700-bp genomic fragmentfrom the 5' region of the hH1 gene was amplified by PCR andsequenced. Strains N1815 (hH1^RIP1) and N1817 (hH1^RIP2) showedthe most mutations (85 and 92, respectively) and were thereforeselected for further characterization. In addition to missensemutations, both strains have a CAA-to-TAA nonsense mutationat residue Q60 (Fig. 1A). Histone H1 was absent from perchloricacid (PCA) extracts obtained from vegetative tissue of strainsN1815 and N1817, confirming that the hH1 gene was disrupted(Fig. 4). No hH1 transcript was detectable by Northern blotanalyses of total RNAs isolated from these strains (data notshown). These observations support the conclusion that N. crassahas a single active histone H1 gene.

View larger version (75K):
[in this window]
[in a new window]
FIG. 3. Generation of hH1 mutants by RIP. RFLPs and DNA methylation at the hH1 locus of four RIP-mutated strains are revealed by Southern analysis. Genomic DNA from the parental strain (lanes P) with the hH1 duplication and four prototrophic mat A progeny (lanes 1 to 4) were digested with either DpnII (D; 5-methylcytosine insensitive) or Sau3AI (S; 5-methylcytosine sensitive). Isolates 1, 2, and 4 carry mutated hH1 alleles, as indicated by RFLPs in the DpnII lanes (white dots), and exhibit methylation, as indicated by high-molecular-weight fragments (bracket on the left) in the Sau3AI lanes. Isolate 3 carries two copies of hH1 and may be free of mutations or only sparsely mutated. A 1.7-kb hH1 fragment was used as a probe. Molecular size markers (in kilobases) are shown on the right.

View larger version (77K):
[in this window]
[in a new window]
FIG. 4. Histone H1 is absent from hH1^RIP mutants. PCA-soluble protein extracts from N. crassa wild-type and hH1^RIP1 and hH1^RIP2 mutant strains were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Coomassie blue staining. N. crassa H1 (arrow) is readily identifiable in protein extracts from total mycelium of the wild-type strain. H1 is absent from extracts from total mycelium of the hH1^RIP mutant strains. Molecular weight (MW) markers (in thousands) are shown on the left.

Absence of histone H1 does not affect DNA methylation but results in subtle changes in chromatin accessibility to MNase. Conflicting reports on the association of linker histones withmethylated DNA in vivo (3, 27) and in vitro (9, 40) led us toexamine effects of the hH1^RIP mutants on DNA methylation. Grossglobal DNA methylation was not affected, as indicated by digestionof total DNA with a methylation-sensitive restriction enzymeand staining of gels with ethidium bromide (data not shown).No localized changes in the DNA methylation levels of severalknown methylated regions tested ( ${psi}$ 63, ${zeta}$ - ${eta}$ , and ribosomal DNA) (39,56) were detected in hH1^RIP mutants (Fig. 5). These findingsare in stark contrast to those of studies with silenced H1 inA. immersus, in which the lack of H1 resulted in global hypermethylation(4).

View larger version (67K):
[in this window]
[in a new window]
FIG. 5. Absence of histone H1 does not affect DNA methylation. For Southern blot analysis, genomic DNAs from a wild-type strain (wt) and two strains with hH1^RIP alleles were digested with DpnII (D) or Sau3AI (S) and probed with a ³²P-labeled ribosomal DNA repeat. Methylation patterns and relative intensities of bands were identical when wild-type and mutant strains were compared. Molecular size standards (in kilobases) are shown on the right.

Global chromatin structure was investigated with MNase (17).A subtle alteration in the pattern of nucleosomal DNA digestionwas reproducibly found with chromatin from the Neurospora hH1^RIPmutants (Fig. 6A). Although the size of the nucleosomal repeatwas similar to that in the wild type, a sharper banding patternwas observed in hH1 chromatin. Densitometry confirmed that wild-typechromatin yielded more diffuse bands after nuclease digestion(Fig. 6B). Similar differences were observed with chromatinpurified from nuclei (data not shown). These observations areconsistent with the idea that in H1-depleted chromatin, linkerDNA is abnormally accessible to MNase.

View larger version (41K):
[in this window]
[in a new window]
FIG. 6. Chromatin of hH1^RIP mutants is more accessible to MNase. (A) Chromatin from the wild-type strain and a strain carrying an hH1^RIP null allele (hH1) was obtained from mycelia and treated with various amounts of MNase (250, 100, 50, 25, 10, and 0 U/g of mycelium, indicated by the ramps). Molecular weight markers are indicated in lane MW. (B) Densitometric analysis of a gel similar to that shown in panel A. Plots were made with the program Scion Image (http://www.scioncorp.com/). A sharper profile of peaks and valleys was observed for hH1 chromatin than for wild-type chromatin. Amounts of MNase in units per gram are indicated between the plots.

Histone H1 is not required during the sexual cycle. In both heterozygous and homozygous crosses, hH1^RIP mutantswere fertile when used as either male or female. The developmentof asci and ascospores was normal, and ascospores exhibitednormal morphology, viability, and germination rates. Meioticchromosomes were stained with hematoxylin-ferric acetate andacriflavine and observed from 3 days postfertilization untilthe end of ascus development. Compared to controls examinedin parallel and also in comparison to the results of previousstudies (51), no changes in chromatin condensation or chromosomebehavior were observed (data not shown).

To study the effect of histone H1 on RIP mutations, we generatedhH1^RIP strains carrying a duplication of the al-1 gene (54)by crossing the hH1^RIP1 strain (N1815) to an al-1 duplicationstrain (N228). RIP frequencies in either homozygous or heterozygouscrosses were equivalent to those in wild-type control crosses(Table 2).

H1 mutants exhibit a slow linear growth phenotype. Both hH1^RIP null mutants studied (N1815 and N1817) retainednormal morphology in liquid or on solid minimal media (Vogel'sminimal medium, PPC, and Westergaard-Mitchell medium) (10) atall temperatures tested (10 to 39°C) (data not shown). Thesize, shape, and distribution of nuclei in mycelia were normal.Mutants produced normal quantities of conidia (asexual spores),with the expected number of nuclei per conidium, and conidiashowed normal viability. The mutants also showed normal sensitivityto MMS, UV irradiation, and DMSO. Measurements of linear growthrates, however, revealed that hH1^RIP mutants had lower mycelialelongation rates on all carbon sources tested (P < 0.005)(Table 3). This slow linear growth phenotype segregated withthe hH1^RIP allele (data not shown). Interestingly, hH1 mutantsshowed better linear growth rates on ethanol (91%) than on sucrose(85%), glucose (85%), or agar (80%) (P < 0.001). Based onthese selective responses to different carbon sources, we hypothesizedthat H1 is required for the expression of one or more specificgenes involved in carbon metabolism.

cfp is misregulated in hH1 mutants. The possibility that H1 plays a regulatory role in carbon metabolismwas explored by studying cfp, the Neurospora gene encoding pyruvatedecarboxylase (1, 20). Pyruvate decarboxylase is a key postglycolyticenzyme and converts pyruvate to acetaldehyde (65). The expressionof the cfp gene is strongly induced by glucose and repressedby ethanol (1). Northern analyses showed that cfp mRNA levelswere similar when wild-type and hH1^RIP mutants were grown onmedia containing either sucrose or glucose (Fig. 7). In contrast,no cfp mRNA was detected when the wild-type strain was grownunder repressing conditions on media containing ethanol (1)or ethanol-glucose, but cfp mRNA was readily detected when hH1^RIPmutants were grown under these conditions (Fig. 7). These resultssuggested that histone H1 is required for the proper regulationof cfp.

View larger version (53K):
[in this window]
[in a new window]
FIG. 7. The cfp gene is misregulated in N. crassa hH1 mutants. Northern blot analyses of cfp mRNAs were performed with wild-type N. crassa and H1 mutant N1815 (hH1) of N. crassa. Strains were grown for 16 h at 30°C in Vogel's minimal medium containing 2% sucrose. Fresh carbon sources (G, glucose; EG, ethanol plus glucose; E, ethanol; and S, sucrose), each at 2%, were added, and mycelia were harvested 4 h later. The 18S rRNA (stained with methylene blue) is shown as a loading control.

Based on this possibility, we examined the chromatin structureof the cfp promoter region in the wild type and the hH1^RIP mutantsunder inducing and repressing conditions. Chromatin was digestedwith MNase, treated with EcoRI or BglII, and analyzed by Southernblotting with a cfp probe. Identical MNase patterns were foundwith chromatin from wild-type and hH1^RIP mutant strains grownunder inducing conditions (i.e., glucose). The same patternwas observed with hH1^RIP mutant strains grown under repressingconditions, while wild-type chromatin from mycelia grown underrepressing conditions (i.e., ethanol) showed subtle differencesin the MNase pattern (data not shown). These results are consistentwith the possibility that cfp is constitutively expressed inhH1^RIP mutants, as suggested by the analyses described above(Fig. 7).

We investigated whether the misregulation of cfp in hH1^RIP mutantswas dependent on the chromosomal position and/or the integrityof the cfp transcriptional unit. A cfp promoter (Pcfp) (positions-847 to -26) was fused to the coding region of the bacterialhph gene, which confers resistance to hygromycin; the fusionwas integrated at the Neurospora his-3 locus. Wild-type or hH1^RIPmutant strains with a single copy of the reporter Pcfp-hph fusiongene were selected. The expression of the Pcfp-hph gene wasexamined in strains grown on minimal media with various carbonsources and hygromycin. As a control, we tested a fusion ofthe constitutive A. nidulans trpC promoter to hph (PtrpC-hph).Both wild-type and hH1^RIP cells transformed with the Pcfp-hphfusion were highly resistant to hygromycin under inducing conditions.On ethanol-containing media, however, wild-type cells with thePcfp-hph fusion were hygromycin-sensitive, whereas hH1^RIP mutantswith the Pcfp-hph fusion were hygromycin resistant. As expected,wild-type and hH1^RIP mutant strains carrying the control PtrpC-hphgene fusion were resistant to hygromycin on media containingeither glucose or ethanol as a carbon source (data not shown).These results show that the cfp promoter is misregulated inhH1^RIP mutants regardless of whether it is driving the expressionof the endogenous cfp gene or expression of a reporter geneat an ectopic location, indicating that the cfp promoter perse is sensitive to negative regulation by H1.

DISCUSSION

Top

Discussion

The hH1 gene of N. crassa encodes a typical eukaryotic histoneH1 composed of a central globular domain flanked by alanine-and lysine-rich regions. Southern blot studies, as well as analysesof available genome sequence data, indicated that hH1 is theonly gene encoding a linker histone in N. crassa. Linker histonesfrom filamentous fungi are well conserved (Fig. 1B). The N-terminaltails of Neurospora H1 and Ascobolus H1 are longer (35 and 25amino acids, respectively) than that of A. nidulans H1 (17 aminoacids). These data are consistent with the longer nucleosomalrepeats in Neurospora (170 ± 5 bp) (46) and Ascobolus(174 ± 6 bp) (53) than in A. nidulans (159 ± 7bp) (17). In both A. nidulans and N. crassa, histone H1 taggedwith EGFP was localized exclusively in the nucleus, consistentwith an association with chromatin.

Based on intron number and distribution, N. crassa hH1 seemsto be more closely related to A. nidulans hhoA and a putativeH1 gene from Aspergillus fumigatus than to the H1 gene fromA. immersus. The positions of the first four introns of theH1 genes from N. crassa, A. nidulans, and A. fumigatus are identical.The position of the first intron of the A. immersus H1 geneis the same as that of the third intron of the Neurospora andAspergillus H1 genes, suggesting that this is the most ancientintron in the H1 genes of filamentous fungi.

It is now clear that histone H1 and closely related linker histonesare not essential in several model organisms—N. crassa,T. thermophila, S. cerevisiae, A. nidulans, and A. immersus—allorganisms with single genes for unusual or bona fide linkerhistones (4, 53, 57, 67). Interestingly, H1 mutants of theseorganisms exhibited different phenotypes. In T. thermophila,deletion of H1 genes resulted in increases in the volume ofnuclei and in alterations in gene expression (57, 58). The functionof the atypical H1 in Tetrahymena is regulated by phosphorylation(11). In A. immersus, silencing of the single H1 gene by MIPresults in global DNA hypermethylation, a shortened life span,and hypersensitivity of chromatin to digestion with MNase (4).In the yeast S. cerevisiae, deletion of the atypical singleHHO1 gene has no dramatic effect on vegetative or sexual phenotypes.Freidkin and Katcoff (14) determined that S. cerevisiae hasonly approximately one HHO1p molecule per 37 nucleosomes, showingthat HHO1p cannot be associated with linker DNA at all nucleosomes.Interestingly, deletion of HHO1 in S. cerevisiae results ina reduction in the steady-state levels of many mRNAs, as shownby microarray analyses (22).

Mycelia from N. crassa hH1 mutants have normal nuclei, and thegross structure of meiotic chromosomes does not appear to beaffected during meiosis. It remains possible that the loss ofH1 is compensated for by other chromatin proteins, e.g., HMG1-classproteins (45, 73), for which there are at least eight homologsin the Neurospora predicted proteome (M. Freitag and E. U. Selker,unpublished data). Chromatin from hH1 mutants showed somewhatsharper bands in MNase digests, similar to what was seen forA. immersus H1-silenced strains (4). Perhaps linker histonesfrom Ascobolus and Neurospora protect linker DNA from proteinaccess, as expected from in vitro studies with chromatin fromhigher eukaryotes. No change in MNase digestion behavior wasobserved for A. nidulans chromatin depleted of H1 (53), consistentwith the possibility that the protein encoded by hhoA representsan evolutionary variant with a more specific function. RIP wasnot affected in Neurospora hH1 mutants. We did not observe globalor localized changes in DNA methylation, in contrast to thesituation for A. immersus (4).

The slow linear growth phenotype of N. crassa hH1 mutants wasdependent on the carbon source, suggesting that histone H1 mayplay a role in pathways associated with carbohydrate metabolism.Based on the linear growth of hH1 mutants in media containingethanol versus sucrose or glucose, we expected a change in thebalance between respiratory and fermentative pathways. Indeed,in hH1 mutants grown under repressing conditions, the expressionof cfp, the gene encoding pyruvate decarboxylase, was misregulatedand abnormally high cfp mRNA levels were detected. Moreover,the chromatin structures at the cfp promoter appeared to beidentical under either inducing or repressing conditions inhH1 mutants. Perhaps Neurospora chromatin lacking H1 fails togenerate the alternative promoter structures associated withnormal regulation of the cfp gene.

H1 may affect the expression of certain genes according to theirpositions in a specific chromosome "neighborhood," rather thanby acting on specific regulatory sequences (16). Nevertheless,an hph reporter gene fused to the cfp promoter and introducedinto an ectopic position in the Neurospora genome showed normalH1 dependence for gene expression. This observation suggeststhat histone H1 specifically modulates cfp expression in N.crassa and that cfp expression is not position dependent.

Our suggestion that histone H1 participates in the expressionof specific genes in Neurospora is consistent with previousstudies showing that the depletion of H1 resulted in the inductionof some genes but in the repression of others (58, 60, 67).Linker histones traditionally have been considered "stabilizers"of higher-order chromatin structure by rendering a condensedchromatin state, but it is also conceivable that histone H1destabilizes chromatin and thus facilitates interactions ofDNA with transcription factors. A recent global analysis ofS. cerevisiae showed that several mRNAs exhibit abnormally lowsteady-state levels in a mutant linker histone background (22).This result indicates that histone H1 can act as a direct oran indirect "facilitator" of gene expression. In this model,the multiple modifications of core histone tails in single nucleosomes,as predicted by the "histone code" (26), may represent variousstates of a chromatin stabilizer. We propose that histone H1plays a role in fine-tuning of the expression of a subset ofgenes involved in primary metabolism. Further studies with NeurosporahH1 mutants and DNA microarrays will explore this hypothesis.

ACKNOWLEDGMENTS

H. Diego Folco and Michael Freitag contributed equally to thiswork. Claudio Scazzocchio is affiliated with Institut Universitairede France.

We thank Chuck Staben for providing strain N228. We are gratefulto D. Perkins, N. Raju, and R. Metzenberg for support, discussions,and suggestions during the study of vegetative and sexual phenotypesof hH1^RIP mutants.

This work was supported by grants from CONICET, SETCIP, FundacionAntorchas (Argentina), and the International Program of theHoward Hughes Medical Institute (to A.L.R.); SETCIP-ECOS SudAction (to A.L.R. and C.S.); the NSF (grant INT-9820195 to E.U.S.);the National Institutes of Health (grant GM35690 to E.U.S. andgrant CA73123 to M.F.); and Université Paris-Sud, CNRS,Institut Universitaire de France, and Association pour la Recherchesur la Cancer (to C.S.). Irene García was supported bya Marie Curie Fellowship of the European Union. Ana Ramónwas supported by the Fondation de la Recherche Medicale.

FOOTNOTES

* Corresponding author. Mailing address: Instituto de Investigación Médica Mercedes y Martín Ferreyra, INIMEC-CONICET, Friuli 2434, 5016 Córdoba, Argentina. Phone: (54)(351)468-1465. Fax: (54)(351)469-5163. E-mail: arosa{at}immf.uncor.edu.

Present address: Department of Plant Pathology, University ofArizona, Tucson, AZ 85721.

REFERENCES

Top