Sequence Analyses Suggest Fungal CP2 Proteins are More Functionally Similar to Animal GRH Proteins than to Animal LSF Proteins

The CP2 superfamily is composed of the GRH and LSF families of transcription factors. A comprehensive review of the functions of LSF-like proteins is beyond the scope of this paper, but there appears to be little overlap between the biological roles of the GRH and LSF families in animals [29], and the two families have diverged greatly in their modes of DNA binding [30], [31]. It is clear that the last common ancestor of Metazoa and Fungi possessed at least one CP2 superfamily protein, although phylogenetic analysis indicates that fungal CP2 superfamily proteins form a separate outgroup with respect to metazoan GRH and LSF family proteins [28]. With few exceptions, all sequenced metazoan genomes possess one or more copies of both GRH and LSF family proteins. Among the Fungi, only ascomycete and zygomycete genomes encode a CP2 superfamily protein (or multiple paralogs), while known basidiomycete genomes do not. Some ascomycetes (e.g., Saccharomyces cerevisiae) appear to have lost the CP2 superfamily. The unicellular sister-group organisms M. brevicollis (a choanoflagellate) and C. owczarzaki (a filasterean) both contain single CP2 superfamily proteins [32], [33].

Although a recently published phylogenetic analysis using gap-free alignments of near full-length protein sequences showed that fungal CP2 superfamily proteins are roughly equally related to both the GRH and LSF protein families [28], we decided to look more closely at the DNA-binding domain sequences of extant opisthokont CP2 superfamily proteins to identify specific residues that might be characteristic of GRH or LSF proteins. An alignment between the DNA-binding domains of two GRH family proteins (D. melanogaster GRH and H. sapiens Grhl1), two LSF family proteins (D. melanogaster GEM and H. sapiens LSF), and a representative fungal CP2 superfamily protein (referred to as Neurospora Grainy head-like, or GRHL, for reasons described below) highlights the extensive sequence conservation throughout this domain (Figures 1A and B). It has been predicted that part of the region containing the DNA-binding domain of CP2 superfamily proteins adopts a similar tertiary structure to the DNA-binding domain of p53 [34], which has a well-characterized three-dimensional structure. Strikingly, the identity of eight amino acid residues at and around positions predicted to be crucial for DNA binding, based on mapping to the p53 structure (i.e., major- and minor-groove contacts, zinc-binding residues, and residues involved in dimerization) suggest that the DNA-binding properties of fungal CP2 proteins might be more similar to GRH than to LSF family proteins. For example, relative to positions 194–198 of the Neurospora GRHL DNA-binding domain (a region predicted to be involved in major-groove interaction) the same amino acid sequence GAERK is found in nearly all available metazoan GRH and fungal GRHL ortholog sequences, while the sequence GADRK is found in all available metazoan LSF sequences (Figure 1B). Similarly, in three other regions predicted to be important for DNA binding (positions 81–84, 142–147 and 150–153), we find that some Neurospora GRHL amino acid residues are identical to those found in nearly all fungal CP2 proteins and metazoan GRH proteins, but they differ from those found in LSF proteins (indicated with asterisks in Figure 1B). Based on these observations, we hypothesized that the DNA-binding characteristics of the Neurospora GRHL protein might be more similar to those of animal GRH family proteins than animal LSF family proteins. Another indication that fungal CP2 superfamily proteins might be more functionally similar to animal GRH proteins is that fungal CP2 superfamily proteins all lack SAM oligomerization domains – animal GRH proteins also lack SAM domains, but all known animal LSF family proteins possess SAM domains [28].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Neurospora GRHL has a similar DNA binding specificity as Drosophila GRH.

(A) The Neurospora GRHL protein shares sequence similarity with both Drosophila GRH and mammalian Grhl proteins, as well as with LSF family proteins [28]. The areas of highest similarity include the region containing the GRH DNA-binding domain, and a region near the C-terminus containing the GRH dimerization domain [30]. (B) A comparison of the DNA-binding domain of a representative fungal CP2 superfamily protein (N. cra GRHL) with those of the Drosophila and human GRH family proteins (D. mel GRH and H. sap Grhl1) and LSF family proteins (D. mel GEM and H. sap LSF). Amino acid residues predicted to be important for DNA interactions based on comparisons with p53 transcription factors [34] are marked above the alignment as follows: “D” – dimerization; “Z” – zinc-binding; “m” – minor-groove interaction; and “M” – major-groove interaction. Residues that distinguish GRH family from LSF family proteins are indicated below the alignment with asterisks. The indicated amino acids in nearly all known fungal CP2 superfamily proteins are identical to the Neurospora residues. (C) Oligonucleotides used in the gel-shifts. Bases that include the LSF or GRH optimal consensus binding sites are indicated with asterisks. (D) Gel-shift assays testing Drosophila GRH and Neurospora GRHL binding to the oligonucleotides in (C). The bottom panels were exposed for 16.5 h, and the top panels were exposed for 75 h. Specific bands are indicated with black arrowheads, and weak specific bands are also highlighted with asterisks in the top panels. Nonspecific (NS) bands were also detected in the no-protein-template negative controls (data not shown), and they are indicated with white arrowheads in the top panels and with a bar in the bottom panels.

https://doi.org/10.1371/journal.pone.0036254.g001

The Neurospora CP2 Protein GRHL and Drosophila GRH have Similar DNA-binding Specificities

To test whether a fungal CP2 proteins binds DNA similarly to animal GRH family proteins, we decided to study the Neurospora crassa CP2 superfamily protein. We chose Neurospora as our model organism because it is a fairly typical representative of a filamentous ascomycete fungus, there exist a number of molecular tools to work with (including gene-knockout technologies), and it has a fully sequenced genome. The Neurospora crassa genome possesses a single CP2 superfamily gene (designated as NCU06095), which has been called grainy-head homolog (ghh) [35], but which we will hereafter refer to as grainy head-like (grhl) or csp-2, for the reasons described in the preceding and following sections. Using RT-PCR and primers specific to the predicted start and stop sites, we cloned and sequenced the full-length grhl coding region and found the sequence and exon structure to be identical to that in the Broad Institute Neurospora database. No splice variants were detected, although we cannot rule out the possibility of grhl transcripts that include additional upstream exons or alternate 3’UTRs.

We synthesized full-length Neurospora GRHL protein in order to characterize its DNA-binding properties using gel-shift analyses. It has been shown that Drosophila and C. elegans GRH family proteins can both bind with high affinity as homodimers to the palindromic DNA sequence ACCGGTT from the Ddc promoter [14], [30] and that the optimal consensus binding site for murine GRH family proteins contains the palindromic DNA sequence AACCGGTT [19], [27]. Mammalian LSF has been shown to bind as a tetramer to DNA fragments containing the sequence CTGG-N₆-CTGG; LSF does not bind to DNA fragments containing a GRH site from the Ubx promoter [31]. On the other hand, Drosophila GRH can weakly bind to both full-length LSF sites and CTGG half-sites [31]. Therefore, we tested the ability of the Neurospora GRHL protein to bind DNA oligonucleotides containing one of the following sites: the endogenous GRH binding site (GRH-Ddc), a mutated GRH binding site (GRH-mut), or the consensus GRH binding site (GRH-con) (Figure 1C). We also tested the ability of the GRHL protein to bind DNA oligonucleotides containing one of the following sites: the endogenous LSF consensus site (LSF-con), an LSF half-site (LSF-1/2), or a mutated LSF site (LSF-mut) (Figure 1C). The binding of full-length Drosophila GRH protein to these oligonucleotides was tested as a comparison.

Drosophila GRH bound DNA sequences as previously reported [30], [31], interacting strongly with the GRH-Ddc and GRH-con oligonucleotides, but not with the GRH-mut oligonucleotide (Figure 1D, right panels). Drosophila GRH also bound very weakly to both the LSF-con and LSF-1/2 oligonucleotides, but not to the mutated LSF-mut oligonucleotide (Figure 1D, top right panel). Neurospora GRHL bound with a similar specificity as Drosophila GRH, albeit with apparent lower affinity. GRHL bound strongly to the GRH-con oligonucleotide, weakly to the GRH-Ddc oligonucleotide, and very weakly to the LSF-con oligonucleotide (Figure 1D, left panels). Considering these results, along with the similarities in their DNA-binding domain sequences, we conclude that the last common ancestor of opisthokonts possessed a CP2 superfamily protein with a similar DNA-binding specificity to existing metazoan GRH family proteins.

Phenotypes of the Neurospora grhl Knockout Mutants

The fungus Neurospora crassa has a simple cellular organization and life cycle compared with most animals and plants (for an in-depth treatment on the subject, see [36]). The most visually obvious phase of the Neurospora life cycle is asexual proliferation – single spores (conidia) germinate on a food source and form a densely interwoven mat of thread-like mycelia, which spreads quickly to form a colony. Neurospora colonies exist as syncytial collections of “cells” which share a common extracellular barrier – the cell wall. While there are regularly spaced septa along the length of the mycelial and hyphal axes, these divisions are not complete, and the “cells” use vigorous cytoplasmic streaming to move nutrients and other molecules throughout the colony. After about a day (and every day after that, according to a circadian rhythm) aerial hyphae grow up and away from the food source and bud off chains of new conidia. These conidial chains become quite delicate as they mature, as the thick cross-walls between individual conidia are remodeled into thin, easily broken connectives – this allows mature conidia to readily detach and disperse to found new colonies.

Neurospora strains containing precise deletions of the entire grhl coding region were obtained from the Fungal Genetics Stock Center (FGSC) for both mating type (mat) backgrounds: FGSC13563 (mat A) and FGSC13564 (mat a). In addition, we created multiple independently derived grhl knockout strains using targeted homologous recombination to replace the grhl locus with a hygromycin cassette. The phenotypes of these mutant strains were indistinguishable from those of the deletion mutant stocks obtained from the FGSC, indicating that the phenotypes described below are indeed due to the loss of grhl function. PCR amplification of a region within the grhl locus verified that all strains were indeed lacking the grhl gene (Figure 2A). Furthermore, RT-PCR amplification of a region of the grhl mRNA yielded no product when RNA from grhl mutants was used as template, compared with robust detection of grhl transcripts using wild-type RNA as template (Figure 2B). Transcripts from the grhl gene were readily detectable by RT-PCR using wild-type RNA templates from either pure mycelial samples or samples of aerial hyphae and conidia (the latter yielding slightly stronger amplifications; data not shown), which suggests the GRHL transcription factor is expressed in most Neurospora cell types during asexual proliferation.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Neurospora grhl mutants display a conidial-separation phenotype.

(A) PCR verification that the grhl locus is deleted in two independently generated grhl knockout strains (grhl-1 and grhl-2) and in a knockout obtained from the FGSC (grhl). The actin locus was amplified as a control. (B) RT-PCR demonstrates that grhl transcripts are not being produced in the FGSC grhl knockout strains (mating types a or A). Similar results were obtained using total RNA from conidia or mycelia. Transcripts from the actin gene were amplified as a control. (C) Mutant grhl strains reach full pigmentation more quickly than wild type. Colonies were grown on Petri dishes for 48 h at room temperature in a 12 h light/dark cycle. Approximately 4 cm² of each mature conidiating colony are shown here. (D–F) Neurospora grhl and wild-type strains have similar growth rates and eventually reach equal pigmentation levels. Shown are Vogel’s agar slants with colony growth after 5 days at 30°C in constant light (D), a 12 h light/dark cycle (E), or constant dark (F). (G–I) Neurospora grhl strains display a pronounced conidial-separation phenotype. (G) Wild-type conidial chains readily disperse in glycerol to yield individual spores. (H) Mutant grhl conidial chains remain intact in glycerol, even after vigorous agitation. (I) A close-up view of a grhl conidial chain demonstrating that the conidia remain connected by thick connectives (arrows).

https://doi.org/10.1371/journal.pone.0036254.g002

Mutant grhl strains are viable and can be propagated asexually as homokaryonic colonies (i.e., all nuclei in the colony are clonal) on minimal media. Both grhl mating-type strains can serve as males or females in sexual crosses to wild-type or grhl strains of the opposite mating type, indicating that grhl function is dispensable for sexual reproduction (data not shown). The grhl strains appear quite healthy and in many ways are indistinguishable from wild type, at least under laboratory conditions (Figures 2D–F).

The grhl mutant strains display a slightly altered circadian rhythm [35], develop orange pigmentation slightly more quickly than wild type (Figure 2C), and sometimes have paler mycelia than wild type (data not shown). However, the most striking phenotype of grhl mutants is a pronounced conidial-separation defect. In grhl strains, conidial chains fail to completely separate, even upon physical stress or immersion in liquid (Figures 2G–I). This phenotype is identical to that observed in the conidial separation mutants csp-1 and csp-2, whose phenotypes have been investigated in some detail. It was shown that the csp mutant strains begin conidial development normally; however, the chitinous cross-walls between adjacent conidia do not become remodeled into thin connectives, precluding conidial separation [37]. This phenotype was correlated with a decrease in the autocatalytic activity of the Neurospora cell wall, which was hypothesized to be due to the loss of secreted enzymes such as chitinase [38].

While csp-1 and csp-2 have long been popular background strains for Neurospora researchers (they help prevent the cross-contamination of stocks), the nature of the mutant genes responsible for these phenotypes remained unknown for many years. Recently, it was shown that csp-1 (NCU02713) encodes a zinc-finger transcription factor on chromosome 1 [39]. However, nature of the gene underlying the csp-2 phenotype remained unclear, except that it mapped to chromosome 7 between the genes thi-3 and ace-8 [40] – precisely the region where grhl is located. Therefore, we believed that a lesion in the grhl gene might be responsible for the csp-2 phenotype.

We first carried out genetic complementation tests with the recessive alleles to test whether csp-2 and grhl are allelic. Different Neurospora strains can fuse to form syncytial heterokaryonic colonies containing nuclei from both parental strains; these fused colonies are often able to grow in conditions that their parents cannot, as each type of nucleus will complement the requirements of the other. For instance, fused colonies from different nutritional-auxotroph parents can survive on minimal media, which can be taken advantage of to test for genetic complementation at another non-selectable locus. Using standard sexual-crossing procedures, csp-2 and grhl mutations were placed into different auxotrophic backgrounds (inos and his-3, respectively), and conidia from each strain were combined on minimal media. We found that all viable heterokaryonic fusings resulted in colonies that still displayed the conidial-separation phenotype, demonstrating that csp-2 and grhl mutant alleles fail to complement (see Materials and Methods for details). To assay for the basis of the non-complementation, we sequenced the grhl open reading frame of the csp-2^[FS590] allele, and found a one base pair deletion in codon S509. This mutation would be predicted to result in a premature stop codon after 14 out-of-frame codons, leading to the removal of the proper 286 C-terminal amino acids of the GRHL protein (Figure S1). Therefore, we conclude that grhl and csp-2 are allelic and that the conidial-separation phenotype observed in grhl strains is due to a reduction in the autocatalytic activity of the cell wall [38], which in turn precludes remodeling of the cross-walls between adjacent conidia [37].

Microarray Profiling of Neurospora grhl Knockout Mutants

To determine the genes directly and indirectly under the control of GRHL in Neurospora, we carried out microarray-based transcriptome profiling of three different sample types: 1) MYC – actively growing pure mycelial samples; 2) AHC – aerial hyphae and conidia from 48 h old colonies; and 3) ALL – all cell types from 48 h old colonies. We only describe the results from the AHC samples, as these are the cell types that displayed the conidial-separation phenotype (see the Materials and Methods section for the accession numbers of the MYC and ALL microarray datasets).

Of the 10,526 genes that were probed on the microarray, 167 were seen to be misregulated in the grhl AHC samples at a False Discovery Rate (FDR) threshold of less than 0.01 (meaning 1%, or about 2 of these genes, are expected to be false positives). This threshold roughly corresponds to a greater than twofold change in expression up or down relative to wild-type levels. Nearly equal numbers of genes were seen to be up- or down-regulated (84 and 83 genes, respectively), and verification of microarray fold-change directionality for ten genes using quantitative RT-PCR is shown in Figure S2. Up-regulated genes on the grhl AHC microarrays are shown in Figure S3, the largest classes of which include genes involved in nitrogen, sulfur, and selenium metabolism, as well as genes involved in membrane transport and cellular import. As we were interested in finding commonalities between the gene products activated by GRH-like transcription factors in animals and fungi, we focused on the down-regulated genes on the grhl AHC microarrays.

Highly Enriched FunCat Categories of the Down-regulated Genes from the Neurospora grhl AHC Microarrays

In order to parse microarray results, researchers often use the Gene Ontology (GO) functional annotation system (www.geneontology.org) to look for highly enriched classes of genes. As a comprehensive GO annotation of the Neurospora crassa genome did not exist at the time of these analyses, we used an alternative classification system – The Functional Catalogue (FunCat) – for which there did exist a high-quality annotation for Neurospora genes [41]. For the 83 genes that were seen to be significantly down-regulated, there were highly significant enrichments in five FunCat categories (Table 1). Three of these categories are composed of genes involved in amino acid metabolism – specifically that of cysteine, phenylalanine, and tryptophan. A fourth category, “C-compound and carbohydrate transport”, is composed of membrane transport proteins. The fifth highly significant category found was “disease, virulence, and defense”, which is composed of genes predicted to be involved in fungal pathogenicity, defense against other organisms, and certain stress responses.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Enriched functional categories for the down-regulated Neurospora grhl genes and the misregulated Drosophila grh genes.

https://doi.org/10.1371/journal.pone.0036254.t001

We could find no direct connections in the literature between cysteine metabolism and barrier formation in animals. However, most amino-acid-metabolism networks are interlinked, and three of these genes (NCU05499, NCU09183, and NCU01402) are also part of the significantly enriched phenylalanine- and tryptophan-metabolism FunCat categories, for which there are some intriguing connections to barrier formation in animals. Melanization reactions in Drosophila are used to harden and cross-link cuticular structures, and are known (at least in the epidermis) to rely on GRH for activation [3]. The reactive quinone molecules used to carry out these processes are derivatives of dopamine, which is itself a derivative of the amino acids tyrosine and phenylalanine (for a review see [42]). It is possible that an ancestral role in phenylalanine regulation by GRH-like transcription factors could have been co-opted by cuticle-forming animals for use in cross-linking apical extracellular barriers. As for the last amino-acid-related FunCat category, “degradation of tryptophan”, there is some evidence that it is a general mechanism of all cells to degrade tryptophan in response to infection, which is used as a means to slow microbial growth through tryptophan deprivation [43]. If true, this function would link tryptophan degradation to the fifth FunCat category – “disease, virulence, and defense” – the presence of which we found especially intriguing, due to the numerous documented connections in animals between physical epidermal barriers and chemical defense against pathogens [44], [45].

Down-regulated Genes from the Neurospora grhl AHC Microarrays

To investigate the down-regulated genes from the Neurospora grhl AHC samples in more detail, we undertook a manual classification of these genes based on database and literature searches. We were especially interested in finding studies carried out directly on the Neurospora crassa genes or on their close homologs in other fungal species. Of the 83 significantly down-regulated genes, 54 had known functions, or predicted functions based on homology to genes in other fungi (Figure 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Down-regulated genes from the Neurospora grhl Aerial Hyphae and Conidia microarray samples.

A manual classification of the significantly down-regulated genes from the Neurospora grhl AHC microarrays. “Broad ID” entries correspond to the gene IDs found in the Broad Institute Neurospora crassa database. The two italicized entries in this column refer to probes that do not correspond to genes in the Broad database, but which correspond to genes in the MIPS database. “Gene name or Description” and “Function” entries were based on the annotations found in the Broad and MIPS databases, as well literature and homology searches. Numbers in curly brackets indicate genes that belong to one of the five highly enriched FunCat categories: {1} metabolism of the cysteine - aromatic group, {2} metabolism of phenylalanine, {3} C-compound and carbohydrate transport, {4} degradation of tryptophan, and {5} disease, virulence and defense. Entries with asterisks encode experimentally verified components of the Neurospora cell wall [46]. “Fold (wt value)” entries indicate the fold changes observed in grhl mutant aerial hyphae and conidia (relative to wild type); wild-type microarray fluorescence values are shown in parentheses (the background level was ∼100 units). “FDR” entries indicate the False Discovery Rate values calculated for each gene; only genes with FDR values less than 0.01 are shown. Columns 1–9 of the grid represent a simplification of the FunCat classification system; solid-colored blocks indicate those genes are classified in the corresponding FunCat categories; dashes indicate that we found evidence in the literature to suggest these genes belong in the corresponding categories. Column 10 of the grid indicates whether the encoded proteins are predicted to be secreted, according to the SignalP (S) or TargetP (T) prediction algorithms. Significantly down-regulated genes that could not be assigned a function are not shown in this figure.

https://doi.org/10.1371/journal.pone.0036254.g003

Strikingly, the most strongly down-regulated gene on the entire microarray (other than grhl itself) was chitinase 1 (NCU04883) (Figure 3), the lack of which is likely to contribute to the conidial-separation phenotype observed in grhl/csp-2 mutants. The chitinase 1 gene of Neurospora has two consensus GRHL DNA binding sites (AAACGGTT & CACCGGTT) within 875 bp of the ATG codon for the chitinase 1 gene. This suggests that the microarrays had identified at least some biologically relevant genes that are directly regulated by GRHL. There were at least six other GRHL-dependent down-regulated genes that encode proteins predicted (from research on other fungi) or known to be involved in normal “Cell Wall Structure” (Figure 3). All of these proteins contain predicted secretion signals, and four are experimentally verified components of the Neurospora cell wall: gel1 (NCU07253, which has an AACCGGTT sequence ∼130 bp upstream of the transcription start), Mwg1 (NCU05974, which has an AACCGGTT sequence ∼400 bp upstream of the 5′ end of the open reading frame), non-anchored cell wall protein-5 (NCU00716, which has an AACAGGTT sequence ∼1.8 kb upstream of the transcription start), and BYS1 domain-containing protein (NCU08907) [46]. Three of the down-regulated cell wall genes – Mwg1 and NCU04431 (both belonging to glycoside hydrolase family 16) as well as gel1 (belonging to glycoside hydrolase family 72) – encode beta-1,3-glucanases. Beta-1,3-glucans are the major biopolymer constituent of the cell wall in filamentous fungi, and it has been shown in many fungal species that beta-1,3-glucanase enzymes are very abundant components of the cell wall, where they play an active role in cell wall biosynthesis and remodeling, as well as in processes such as biofilm formation [46]–[49]. For instance, mutations in enzymes from glycoside hydrolase family 72 cause cell wall defects in S. cerevisiae [48] and also affect morphogenesis and virulence in Aspergillus fumigatus [50]. We also found that seven of the 29 down-regulated genes that could not be assigned a function encode proteins with predicted secretion signals, and therefore might be components of the cell wall (Figure 3).

We classified 15 grhl-dependent down-regulated genes in the category “Virulence/Defense/Detoxification” (Figure 3). Defense/Virulence have not been studied experimentally in Neurospora crassa, as its normal host-pathogen relationships are unknown, so the functions of the following genes are inferred from research in other fungi. Genes potentially involved in defense include the following: kynureninase (NCU09183, which has two ACAGGTT sites ∼ 650 bp upstream of the open reading frame) and indoleamine 2,3-dioxygenase (NCU01402), which are involved in tryptophan catabolism and microbial growth control [43] (see above); NCU05495 encodes a putative anti-viral factor [51], [52]; and exo-beta-1,3-glucanase (NCU04850) is possibly involved in the degradation of foreign polysaccharides. Other Neurospora grhl-dependent genes potentially involved in fungal virulence include the following: the metalloprotease MEP1 (NCU07200), whose homolog in C. posadasii has been shown to be crucial for evasion of host-detection [53]; the p450 monooxygenase lovA (NCU05376), whose homolog in a Fusarium species has been shown to be directly involved in mycotoxin synthesis [54]; cerato-platanin (NCU07787), which is potentially important for phytotoxin synthesis [55]; the integral membrane protein pth11 (NCU06328), whose homolog in another fungal species is important for appressorium formation [56]; and NCU03643, which encodes a cutinase transcription factor that is likely to control plant cuticle digestion during fungal infection [57]. Finally, several genes potentially involved in the detoxification of harmful chemicals and the stress response include the following: the p450 gene pisatin demethylase (NCU06327), whose ortholog is important in fungal pea-pathogens for detoxifying host defensive chemicals [58]; the aldehyde dehydrogenase gene (NCU03415), which encodes a broadly acting detoxification and metabolic enzyme; the genes catalase-3 (NCU00355) (another verified component of the cell wall [46]) and NAD(P) transhydrogenase (NCU01140), which encode enzymes known to be important for oxygen-radical detoxification; and the YBH1 flavohemoglobin gene (NCU10051), which may be involved in the stress response [59].

Taken together, these results suggest that Neurospora GRHL plays an important role in the regulation of genes that form and remodel the cell wall (at least in developing conidia). Additionally, a significant number of GRHL-dependent down-regulated genes in the aerial hyphae and conidia are involved in virulence, defense, and detoxification. It is should be noted that the cell wall and virulence/defense categories are not mutually exclusive, as seven of the 15 “defensive” gene products (e.g., cerato-platanin and MEP1) have secretion signals (Figure 3) and are likely to be deposited into the cell wall or released into the extracellular space.

Microarray Profiling of Late-stage Drosophila grh Embryos

As a comparison to the Neurospora microarray dataset, we also carried out microarray-based transcriptome profiling of Drosophila grh^IM and control wild-type embryos collected during late-stage 16 and early-stage 17 of embryogenesis [60], when cuticle deposition is occurring. We used flies homozygous for the grh^IM allele because it is the strongest grh allele available (with respect to its cuticle and head-skeleton phenotypes) and because homozygous embryos do not produce any detectable GRH protein (assayed using an antibody against the C-terminal half of the protein [6]). By sequencing the grh^IM transcript, we identified the lesion responsible for the grh^IM allele as a TAT to TAA stop-codon introduction in exon seven, in the N-terminal end of the DNA-binding domain and about half-way through the protein (Figure 1B, amino acid Y29 in the D.mel GRH DNA-binding domain; see Figure S4 for details); this mutation is consistent with a functional null phenotype for grh^IM and the lack of GRH protein detection with a C-terminal specific antibody [6]. See the Materials and Methods and Text S1 for more details on the Drosophila microarrays and data analyses. Besides the mutation in grh, the control and mutant embryos differed slightly with respect to their genetic backgrounds, as the wild-type strain contained the yellow¹ allele, which has an adult pigmentation defect, while the grh^IM embryos were yellow⁺. However, we do not believe this significantly influenced our results since our microarray data indicate that yellow is expressed at extremely low levels during the embryonic stages we tested; furthermore, the yellow transcript expression levels were not significantly different between the grh^IM and control embryos on the microarrays.

Highly Enriched Gene Ontology Categories of the Misregulated Genes from the Drosophila grh Embryo Microarrays

Assayed at stages 16–17 of embryogenesis, zygotic loss of GRH function has a huge impact on the Drosophila transcriptome as a whole, as over 1,200 genes (FDR <0.01) were seen to be misregulated (up or down) in grh^IM mutants compared with wild type (see the Materials and Methods section for the accession numbers of the Drosophila microarray datasets). Verification of microarray fold-change directionality for eight genes using quantitative RT-PCR is shown in Figure S5.

A search for enriched GO “Biological Process” (BP) and “Molecular Function” (MF) categories was performed (Text S1), and the top ten and eleven most significant classes, respectively, are shown in Table 1 (see Table S1 for the full lists of the significantly enriched GO-BP, GO-MF, and GO “Cellular Component” categories). As GRH is known to be very important for cuticle development and wound healing in Drosophila, we expected to see numerous genes involved in these processes misregulated on the microarrays. Indeed, four of the most significant GO-BP classes (e.g., “chitin metabolic process” and “aminoglycan metabolic process”) and five of the most significant GO-MF classes (e.g.,“structural constituent of the cuticle” and “chitin binding”) are consistent with the known role of GRH in regulating the formation of chitin-based cuticular barriers.

Surprisingly, the remaining six GO-BP categories (of the top ten) were all composed of genes involved in either innate immunity or the stress response (e.g., “defense response”, “immune response”, and “humoral immune response”). Similarly, the remaining six GO-MF categories (of the top eleven) were all composed of genes that encode products with either serine-protease or serine-protease-inhibitor activity. This was interesting because serine protease cascades are used to trigger the hemolymph melanization-reactions used in response to infection, and serine protease inhibitors (also known as serpins) are used to limit the spread of this reaction (for a review see [44]). This was initially puzzling, as most of these genes were seen to be up-regulated in the grh^IM mutants (Table 2), and GRH has no known function as an inhibitor of the immune response.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Select misregulated genes from the late-stage grh^IM embryo microarrays.

https://doi.org/10.1371/journal.pone.0036254.t002

From these we results, we conclude that in addition to the expected misregulation of genes involved in cuticle formation, late-stage grh^IM embryos are experiencing a massive wound/immune response as well. During the stages they were collected (late-stage 16 or early-stage 17 of embryogenesis) the grh^IM embryos have weaker and more permeable epidermal barriers [1], [6], [12], yet are still motile, which can cause their fragile cuticles to rupture. Consistent with this, the pale gene, which encodes tyrosine hydroxylase, is known to be up-regulated around sterile wound sites in a largely grh-independent manner [3], [4], and in grh^IM embryos, pale transcripts are significantly up-regulated. This is also consistent with the observation that clean puncture wounding of late-stage embryos (in the absence of intentional microbial infection) also induces the expression of large numbers of Drosophila genes involved in innate immunity and the stress response (R. Patterson & W. McGinnis, unpublished).

Misregulated Genes from the Drosophila grh Embryo Microarrays Reflect the Role of GRH in Barrier Formation and Wound Healing

We carried out a manual classification of the genes both and up- and down-regulated on the Drosophila grh^IM embryo microarrays, and select genes are shown in Table 2. We placed 64 genes in the category “Cuticle Formation/Chitin Metabolism”, including genes involved in the generation and degradation of chitin molecules, as well as genes for many cuticle proteins that are deposited into the cuticle to mediate aspects of cuticle-shape and elasticity [61], [62]. The majority of these genes (42 of 64) were down-regulated, consistent with potential direct regulation by GRH, and consistent with the idea that GRH is a crucial regulator of physical-barrier formation in Drosophila. However, a subset of these "cuticle/chitin" genes (22 of 64) were seen to be up-regulated in the grh mutants. It is possible that these up-regulated cuticle genes are normally directly repressed by GRH in wild-type embryos, but we believe it more likely that they are being overexpressed to compensate for the lack of the GRH-activated cuticle proteins, or they are being overexpressed in response to cuticular damage in grh^IM mutants (see above).

Interestingly, one of the most strongly down-regulated genes on the microarray was chitinase 3 (∼13 fold down) (Table 2), which is a Drosophila homolog of Neurospora chitinase 1– the most strongly down-regulated gene in the Neurospora grhl mutants (see above). We identified three high-affinity GRH binding sites within the two kb region upstream of the Drosophila chitinase 3 transcriptional start site, and chitinase 3 is also extensively co-expressed with GRH throughout the Drosophila epidermis and tracheal system during embryogenesis (data not shown), consistent with direct regulation by GRH.

Also consistent with the grh mutant phenotypes, 19 genes known or predicted to be involved in a category we called “Cell Adhesion/Apical-Basal Polarity/Cytoskeleton” were significantly misregulated (Table 2). This category includes three Cadherin-domain-containing protein genes, including Dystroglycan (∼5 fold down), which is required for maintaining the apical-basal polarity of epithelial cells and anchoring the intracellular actin cytoskeleton to the extracellular matrix [63]. Another gene required for apical-basal polarity and adhesion of epidermal cells that was significantly down-regulated in grh^IM mutants (∼2.3 fold down) is bazooka [64]. Previous studies by Narasimha et al. [13] have shown that two genes encoding components of the Drosophila septate junction – coracle and Fasciclin 3– were expressed at reduced levels in grh mutant clones in imaginal disc epithelia, and that coracle and Fasciclin 3 gene expression could be activated by ectopic GRH protein in embryonic amnioserosa cells. As measured by our late-stage embryonic grh mutant microarrays, coracle transcript levels were very slightly lowered, but the difference compared to wild-type levels was not statistically significant. Fasciclin 3, which was probed by eleven different sequences on the microarray chip we used, was reproducibly reduced in grh mutants; we found that the expression of this gene was ∼25% lower than wild type in every experimental and biological replicate, although this difference was never significant enough to reach the stringent FDR threshold we set. Therefore, it is important to note that since we used whole embryos as the RNA source for our microarray experiments there are likely to be many true GRH-regulated genes that were not identified as significant in our analyses, such as coracle and Fasciclin 3. It is likely that this category will include genes whose expression is only quantitatively changed in grh mutant backgrounds, or whose expression is limited to only a subset of the cells that produce GRH protein, and thus would not pass our FDR threshold.

We placed nine genes in the category “Melanization/Wound Healing” (Table 2), four of which were down-regulated. The three most strongly down-regulated genes in this class (alpha methyl dopa-resistant, Pcd, and prophenol oxidase A1) are known or suspected to be directly involved in the cuticular melanization/sclerotization pathway. Stitcher/Cad96Ca, a wound-induced gene known to be directly activated by GRH [5], was significantly down-regulated on our microarrays. Dopa decarboxylase, another gene known to be directly dependent on GRH for its expression during development [1] and the epidermal wound-response [3], was down-regulated (∼1.7 fold down), although it did not pass the stringent FDR threshold we set and is not shown in Table 2.

Misregulated Genes from the Drosophila Microarrays Indicate Mutation of grh Triggers Innate-immune and Stress-response Pathways

While a comprehensive analysis of all the genes involved in innate immunity, stress, and detoxification that were seen to be misregulated in grh^IM mutants is beyond the scope and focus of this paper (and because many are very likely to be misregulated due to indirect effects of the grh^IM phenotype), we will only briefly review the major classes of genes.

Thirty-seven genes in the category “Innate Immunity” (Table 2) were misregulated in grh^IM embryos, and they included genes from nearly every aspect of Drosophila innate immunity [44]. One innate-immune gene, IM1, was strongly down-regulated, suggesting the potential for direct activation of IM1 by GRH in the epidermis. Consistent with this possibility, there is a near perfect palindromic GRH binding site (AACTGGTTT) found less than 600 bp upstream of the IM1 gene. Other down-regulated immunity genes potentially under the direct control of GRH include lectin-28C, Mucin 91C, eater, and two putative lysozymes. However, the majority of innate-immune genes that were misregulated in grh^IM mutants were up-regulated (31 of 37), and they include known antimicrobial peptides (e.g., Attacins, Drosomycins, Diptericin B, and Metchnikowin), lysozymes, Pattern-Recognition Receptors, and the Toll-signaling activator Spaetzle-Processing Enzyme (SPE).

Eighteen genes in the “Defense/Stress Response” category were misregulated in grh^IM embryos (Table 2). The two most strongly down-regulated genes were methuselah-like 8 and 3. Mutations in a paralogous gene (methuselah) have been correlated with longer lifespan and increased resistance to stress in Drosophila [65], so it possible that the observed down-regulation of mthl-8 and -3 was a response to tissue damage and stress in these embryos. Seven heat shock protein (hsp) genes were also misregulated in grh^IM embryos, five of which were up-regulated and have been shown elsewhere to be differentially expressed upon infection of adult flies with microbes (Hsp26, 70Bc, 70Bb, 70Bbb, and 70Aa) [66]. Turandot A, C, and Victoria were also seen to be up-regulated, which are believed to act as extra-cellular chaperones, binding to denatured proteins in the hemolymph that are released upon tissue damage or stress [67].

Forty-four genes in the “Detoxification” category were misregulated in grh^IM embryos (Table 2). These include multiple cytochrome p450 genes [68] as well as glutathione S-transferases and UDP-glucuronosyltransferases, which function by chemically modifying toxic compounds in the cytoplasm or hemolymph in order to render them less active [69]. The misregulation of these genes (which were both up- and down-regulated in nearly equal proportions) is still somewhat unclear, although we propose that their expression levels are altered in response to the release of toxic endogenous compounds during tissue damage in grh^IM mutants.

Taken together, these results indicate that late-stage grh^IM embryos have reduced expression of a wide variety of extracellular cuticular-barrier genes as well as a number of cell-cell adhesion genes [13] (Tables 1, 2, and S1); in addition they are experiencing a massive wound/immune response and are undergoing extreme stress, likely due to global tissue damage in response to cuticular tearing or epidermal-barrier permeability [6].

GRH is Required for Epidermal Integrity During Larval Stages of Drosophila Development

We wished to more fully characterize the phenotypes of grh null mutants at later stages of the Drosophila life cycle. However, as grh null-mutant embryos die at the embryonic/larval transition, this prohibited us from determining the function of GRH in grh^IM larvae and adults. To circumvent this, we produced a Drosophila strain in which GRH is knocked down in the larval epidermis (e13C>GRH^RNAi) by crossing a transgenic UAS-GRH^RNAi line with a strain containing the larval driver e13C-GAL4, which produces GAL4 in the larval epidermis, fat body, gut, imaginal discs, and salivary glands [70]. By immunostaining, we observed that GRH protein was undetectable in the epidermal tissues of late third-instar e13C>GRH^RNAi larvae (Figure 4B) compared with control e13C-GAL4 larvae (Figure 4A), demonstrating that RNAi-mediated knockdown of GRH is very efficient in third instar larvae. In both samples, epidermal cell boundaries are clearly marked by Fasciclin 3 staining, indicating that gross cellular morphology remains intact. It is possible that GRH is only partially knocked down during larval molts, allowing for the deposition of a cuticular barrier sufficient for larval survival, albeit a cuticular barrier that is defective enough to be more easily wounded and water-permeable in wandering third-instar larvae (see below).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 4. Knock-down of GRH in the larval epidermis leads to dramatically increased fluid loss following injury.

(A and B) Whole-mount preparations of dissected larval epidermal tissue from 5 day old control larvae (e13C-GAL4) (A) or 5 day old larvae expressing a UAS-GRH^RNAi transgene driven by e13C-GAL4 (e13C>GRH^RNAi) (B). Antibody stains for GRH (green) and Fasciclin 3 (red) are shown. (A’ and B’) Close-ups of cuticle preparations of dissected larval epidermises from 5-day-old (at the wandering stage, just prior to puparium formation) control larvae (e13C-GAL4) (A’) or 5-day-old larvae expressing a UAS-GRH^RNAi transgene driven by e13C-GAL4 (e13C>GRH^RNAi) (B’). The large dark spots seen in B’ are localized depositions of melanin that appear at random positions in the epidermis/cuticle when GRH is knocked down during larval stages. (C–F’) Control larvae (e13C-GAL4) and GRH knock-down larvae (e13C>GRH^RNAi) were left untreated (C, D, E, and F) or punctured with a clean needle (injured; C’, D’, E’, and F’) under moist (C, C’, D, and D’) or dry (E, E’, F, and F’) conditions. Sites of injury are indicated with arrows.

https://doi.org/10.1371/journal.pone.0036254.g004

While generally healthy, wandering third-instar e13C>GRH^RNAi larvae all developed melanized clots of diverse size and distribution, which were never observed in control larvae (Figures 4A’ and B’) or in e13C>GRH^RNAi larvae during the earlier first, second, or foraging third-instar stages (data not shown). Upon dissection it was found that these melanotic spots were tightly associated with epidermal cells of the body wall, similar to the phenotypes seen when both Dorsal and Dif transcription factors are knocked down during larval stages [70]. Furthermore, all e13C>GRH^RNAi larvae died at the prepupal stage with noticeably decreased body size compared with wild type, despite the fact that experimental and control larvae appeared similar in size prior to wandering (data not shown).

A lack of GRH function during Drosophila embryogenesis leads to a fragile cuticle [1], [12] and increased epidermal permeability [6]. Therefore, it seemed likely that knocking down GRH during larval stages would have similar effects, which could explain both the presence of randomly localized melanized spots as well as the decreased body size (presumably due to fluid loss). In Drosophila, melanized spots have been associated with hyperactivation of the immune system [71], loss of both Dorsal and Dif transcription factors [70], and such dark spots also appear at wound sites, apparently to strengthen clots and prevent body-fluid loss following physical injury.

We propose that upon leaving the moist food source, wandering e13C>GRH^RNAi larvae develop multiple “micro-wounds” (as evidenced by the ectopic melanotic spots) due to the fragility of their cuticles, which is exacerbated by the dry conditions on the vial walls. Furthermore, we hypothesized that the decreased body size observed in these larvae is a result of fluid loss due to increased epithelial permeability as well as a loss of hemolymph following micro-wounding. These observations suggest that during larval stages, GRH is required for the maintenance of epidermal/cuticular-barrier integrity.

To determine whether epidermal GRH activity is required in larvae to prevent body-fluid loss following wounding, we wounded wandering third-instar larvae with a sterile needle and let them recover in either moist or dry conditions. Under moist conditions, both control and e13C>GRH^RNAi larvae maintained approximately the same body mass, even after clean injury (compare Figures 4C’ and D’). However, when injured e13C>GRH^RNAi larvae were placed in dry conditions, they showed an obviously decreased body size 7 h after injury (Figures 4F’). Injured control, uninjured control, or uninjured e13C>GRH^RNAi larvae did not show any obvious decreases in body size under dry conditions after 7 h (Figures 4E, E’, F). These results suggest that GRH activity in the epidermis is needed for properly repairing wounds and preventing catastrophic body-fluid loss following wounding under dry conditions. These effects could be due to failures or delays in wound healing in the knockout larvae, possibly caused by weakened cuticles due to lower GRH protein levels during larval stages.

We were curious to see if epidermal GRH knockdown could lead to desiccation in the absence of injury after longer periods (>7 h) in dry conditions. We incubated both injured and uninjured control and e13C>GRH^RNAi larvae for ∼24 h under dry conditions. Although most uninjured (96.2±0.6%) and injured (65.6±3.9%) control larvae reached the prepupal stage without any obvious decreases in body size, all uninjured and injured e13C>GRH^RNAi larvae (100%) showed decreases in body size and died before initiating the pupariation process (data not shown). These results indicate that the larval function of GRH is crucial for avoiding excessive body-fluid loss under dry conditions, and it is necessary for viability even in the absence of major injury.

Silencing of GRH in Adult Drosophila Increases their Susceptibility to Septic Injury

Due to the relatively short time-course of larval development, and the fact that e13C>GRH^RNAi larvae do not develop past the prepupal stage, examining the role of epidermal GRH expression in larval microbial resistance is problematic. Therefore, we focused on analysis of clean or septic injury in Drosophila adults in which grh function was knocked down by heat shock induced RNAi. We found that heat-shock driven expression of a UAS-GRH^RNAi (hs>GRH^RNAi) can efficiently eliminate GRH protein in most cells of the adult epidermis compared with similarly treated control flies containing only the hs-GAL4 construct (Figures 5A and B). These hs>GRH^RNAi flies were completely viable and did not develop the drying phenotypes observed in e13C>GRH^RNAi larvae.

The hs>GRH^RNAi flies did not show any reduction in normal life span compared with control flies, indicating that GRH is not required for the homeostatic maintenance of adult viability, at least under laboratory conditions (Figure 5C, and data not shown). Next, we challenged knockdown and control flies with either clean or septic injury using Ecc15 (a gram-negative bacterium) or M. luteus (a gram-positive bacterium), and their survival was monitored over a 10 day period. Although control and hs-GRH^RNAi flies showed similar survival curves after clean injury (Figure 5D), hs>GRH^RNAi flies showed decreased survival after Ecc15 and M. luteus infection than controls. Ten days after septic injury with Ecc15 or M. luteus, 40 and 67% of the GRH knockdown adults survived, respectively, compared with 57 and 80% of control adults that were heat shocked without GRH^RNAi knockdown (Figures 5E and F). These results suggest that GRH plays some protective role in Drosophila adults following microbial infection. However, our embryonic microarrays indicate that grh is not required to activate the standard repertoire of antimicrobial peptide genes, so it seems unlikely that GRH is crucial for activating the canonical genes of the innate immune response in adults [44]. Therefore, the exact nature of the protective effects imparted by GRH during Drosophila adulthood remains to be identified.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 5. Loss of GRH function in adult flies increases their susceptibility to bacterial infection.

(A and B) Whole-mount preparations of dissected adult abdominal epidermal tissue from control flies (hs-GAL4) and GRH knock-down flies expressing a UAS-GRH^RNAi transgene driven by hs-GAL4 (hs>GRH^RNAi). Antibody stains for GRH (green) and Fasciclin 3 (red) are shown. (C–F) Survival curves from hs-GAL4 and hs>GRH^RNAi adults after mock treatment (C), clean injury (D), injury with a needle coated with gram-negative Ecc15 bacteria (E), or injury with a needle coated with gram-positive M. luteus bacteria (F). The survival of adult flies was measured over the 10 day period after injury. The average values of three independent experiments are shown along with the standard errors of the mean.

https://doi.org/10.1371/journal.pone.0036254.g005

Neurospora Stocks, grhl Knockouts, and grhl/csp-2 Complementation Assays

Wild-type strains [FGSC2489 (74-OR23-1V, mat A) and FGSC4200 (ORS-SL6, mat a)], grhl knockout strains [FGSC13563 (ΔNCU06095, mat A) and FGSC13564 (ΔNCU06095, mat a)], and the NHEJ-deficient strain [FGSC9720 (Δmus-52::bar+; his-3, mat A)] were obtained from the Fungal Genetics Stock Center (FGSC) [78]. Stocks were maintained on minimal Vogel’s agar slants with 1.5% sucrose and appropriate supplements [36]. Genomic DNA for PCR analysis was obtained according to [79].

The isolation of the independently derived grhl deletion strains was performed by transforming a NCU06095-targeted hygromycin replacement cassette (courtesy of the Dunlap lab, Dartmouth) into FGSC9720, as described elsewhere [80]. Hygromycin-resistant colonies were selected, and homokaryonic grhl knockout strains were tested using PCR to verify loss of the grhl locus. All strains missing the grhl locus displayed the conidial separation phenotype. These new strains (Δmus-52::bar+; Δgrhl::hyg+; his-3, mat A) were also used in the complementation-assay fusings described below. The primer sequences for verifying the grhl knockouts and for detecting grhl transcripts (Figures 2A and B) were as follows: grh-For – CACCAGTCAAGCTGGCATC –and grh-Rev – GGCTTATGTCGCTGCTTTTC. Positive control primers were as follows: actin-For – ATCCGACACTTTTCGTCACC – and actin-Rev – TGCAACAACCACCTCTCAAG.

Genetic complementation assays between grhl and csp-2 were carried out by fusing one of the independently derived grhl deletion strains described above (Δmus-52::bar+; Δgrhl::hyg+; his-3, mat A) to ten different isolates of csp-2; bd; inos. The csp-2; bd; inos strains were created using standard crossing methods (csp-2; bd × inos). Exactly as expected, only half (five) of the fusings were viable on minimal media (due to opposite mating-type incompatibility), all of which displayed the conidial separation phenotype. Using PCR, the csp-2/grhl heterokaryons were verified as positive for the grhl locus, which is consistent with the deletion/stop codon in csp-2 strains (data not shown).

Neurospora GRHL and Drosophila GRH Protein Production and Gel-shifts

The full-length Drosophila grh coding sequence was cloned into the plasmid pcDNA 3.1/myc-His(-) A (Invitrogen) as described elsewhere [4]. The full-length Neurospora grhl coding sequence was amplified using the Phusion polymerase (New England Biolabs) from an oligo-dT-primed cDNA library (RETROscript kit, Ambion). The primers Grhl5’XbaIKozak –GCGTCTAGAGCCACCATGTTCAGTCAACGAACAAG – and Grhl3’HindIII –CGCAAGCTTGTAGAGCAGTCGCAGTTCAT – were used to introduce a Kozak sequence (for efficient translation) and restriction endonuclease sites. The fragment was cloned into pcDNA 3.1/myc-His(-) A using the XbaI and HindIII sites in the multiple cloning site. The insert was fully sequenced and was found to be identical in sequence and exon structure to that predicted by the Broad Institute Neurospora database.

GRH and GRHL proteins were translated using the TNT T7 Quick Coupled Transcription/Translation System (Promega) by adding 1 µg of template to each master mix aliquot, according to the manufacturer’s instructions. Protein expression levels were assayed by Western blotting, using antibodies against the C-terminal Myc tags, as described elsewhere [4]. The translated proteins were directly used in the gel-shift assays, as freezing was found to negatively affect DNA-binding activity. For each oligonucleotide pair, 500 pmol of each were annealed in a final volume of 100 µl in annealing buffer (10 mM Tris-HCl pH 7.5, 20 mM NaCl) by heating to 95°C for 5 min and slowly cooling to 25°C. Then, 5 pmol of double-stranded oligonucleotides was labeled with polynucleotide kinase (New England BioLabs) in the presence of ATP-[³²P] for 30 min at 37°C. The double-stranded probes were purified using the QIAquick Nucleotide Removal Kit (Qiagen). Next, 10–20 fmol of radiolabeled double-stranded oligonucleotides and 1.5 µl of protein from the in vitro transcription/translation reactions were added to 10 µl binding buffer [25 mM Hepes pH 7.9, 100 mM KCl, 1 mM DTT, 1% polyvinylalcohol, 1% Nonidet P-40, 0.1% BSA, 10% glycerol, and 20 µg/ml poly(dI-dC)] and incubated with DNA for 30 min at 4°C. The binding reaction was then electrophoresed through a 4% native polyacrylamide gel in 0.5× TBE at 4°C. Gels were dried and autoradiographed with the use of intensifying screens.

Neurospora and Drosophila Microarray Sample Collection

Neurospora samples for microarray analysis were collected according to the following procedures. Seeder slants of wild-type (FGSC2489) and grhl (FGSC13563) strains were grown for 3 days at 30°C under a 12 h light/dark cycle, and conidia were harvested in 1 ml H₂0. As the csp phenotype makes homogenous resuspension of grhl conidia impossible, accurate conidial counts of the suspensions could not be obtained. Therefore, plates and flasks were innoculated with approximately equivalent masses of conidia suspended in water. As Neurospora comes to confluence very quickly on plates, and growth in liquid culture for short periods of time should not be nutrient limiting, we believe the number of starting conidia was not crucial to these experiments. The following collection procedures were carried out in triplicate for both wild-type and grhl samples. 1) “ALL” samples were collected by densely plating conidia on minimal Vogel’s agar medium +1.5% sucrose in 10 cm Petri dishes that had been overlain with disks of cellophane (VWR, 100357–652). After 48 h at 30°C under a 12 h light/dark cycle, the plates were densely covered with conidiating colonies. Samples were scraped off the cellophane using cell-scrapers, submerged in 5 ml Trizol (Invitrogen), and quickly frozen in liquid nitrogen. 2) “AHC” samples were collected by densely plating conidia on minimal Vogel’s agar medium +1.5% sucrose in a deep 10 cm Petri dishes. Disks of medium gauge wire mesh were suspended ∼0.5 cm above the surface of the agar using a ring of plastic tubing around the periphery of each Petri dish. After 48 h at 30°C under a 12 h light/dark cycle, the aerial hyphae and conidia had grown abundantly through the mesh. The mesh disks were carefully peeled off, and the adherent cells were harvested in H₂0 using cell-scrapers. Samples were dried by vacuum filtration, removed from the filter paper using cell-scrapers, submerged in 5 ml Trizol, and quickly frozen in liquid nitrogen. 3) “MYC” samples were collected by inoculating 25 ml of liquid Vogel’s medium +1.5% sucrose in 125 ml Erlenmeyer flasks with sponge stoppers. After 28 h at 28°C with constant shaking in the dark, log-phase mycelial mats were recovered by vacuum filtration, removed from the filter paper using cell-scrapers, and frozen in 2 ml Trizol with liquid nitrogen.

To obtain total RNA from the Neurospora samples for microarray analysis, we followed a protocol similar to that reported elsewhere [81]. Samples in Trizol were thawed and quickly homogenized by vortexing and passing through a P1000 pipet tip multiple times to break up large clumps. Approximately 100 µl of cells were placed in an eppendorf tube with 1 ml Trizol and 200 mg of 0.5 mm Zirconia/Silica Beads (Biospec). Samples were disrupted twice with a MiniBeadBeater (Biospec) at maximum speed for 30 s each time. RNA was then extracted using standard Trizol procedures and resuspended in 100 µl H₂0. RNA was quantified in 10 mM Tris pH 7.5, and 50 µg of total RNA was cleaned further using the RNeasy miniprep kit (Qiagen). RNA integrity was assayed by gel electrophoresis, RT-PCR against several developmentally dynamic genes [81], and Bioanalyzer (Agilent) analysis (data not shown).

The following Drosophila embryo collection procedures were carried out in duplicate. To aid in the collection of homozygous grh-deficient embryos, the cn, grh^IM, bw, sp chromosome [2] was placed over the fluorescent balancer CyO, Kruppel-GFP (CKG) [82]. Heterozygous collections of embryos (+; cn, grh^IM, bw, sp/CKG; +) were allowed to develop at 25°C until ∼15–18 h of age. Embryos were aligned on a thin agar slab on a slide, and GFP-negative embryos (grh^IM homozygotes) were selected using epifluorescent microscopy. Gut autofluorescence in the GFP channel allowed for the selection of viable and properly aged embryos (late stage 16 and early stage 17 [60]) using gut morphology as a guide. Correspondingly staged wild-type (y; cn, bw, sp; +) embryos were similarly raised and selected using gut autofluorescence as a guide. Approximately 500 mutant and wild-type embryos were collected and stored frozen in Trizol. Embryos were ground in Trizol using a pestle, RNA was purified using standard Trizol procedures, and RNA integrity was assayed as described above.

Microarray Design and Analysis

Neurospora microarrays were custom synthesized by Agilent using the sequences from the Neurospora crassa arrays available from the FGSC [81], [83]. All probe sequences were shortened from 70 mers to 60 mers by removing the first ten 5′ nucleotides. A total of 10,526 unique spots were printed on each chip, corresponding to predicted genes from several databases. Once a finalized list of significant genes was obtained, probe sequences were BLASTed against the Neurospora crassa genome to verify the Broad (or MIPS) gene ID annotations.

Predesigned Drosophila melanogaster arrays were ordered from Agilent (Design ID # 18972). A total of 43,603 spots were printed on each chip, which mapped to ∼13,000 unique FlyBase genes. Fluorescence values from redundant probes (or unique probes targeting the same gene) were grouped, and only the highest fold-change values were used in these analyses. Once a finalized list of significant genes was obtained, probe sequences were BLASTed against the Drosophila melanogaster genome to verify the FlyBase CG # annotations.

RNA labeling, hybridizations, fluorescent quantification, data normalization, FDR calculations [84], and GO annotations (Tables 1 and S1) were carried out by the Biogem Core facility (UC San Diego); see Text S1 for an in-depth description of the microarray analyses. Manual Neurospora gene classifications (Figure 3) were carried out by consulting the Broad Institute Neurospora crassa database (www.broadinstitute.org/annotation/genome/neurospora/MultiHome.html), the MIPS Neurospora crassa database (mips.helmholtz-muenchen.de/genre/proj/ncrassa), and the Functional Catalog (FunCat) [41] classifications found in the MIPS database (mips.helmholtz-muenchen.de/genre/proj/ncrassa/Search/Catalogs/searchCatfirstFun.html), as well as with literature and homology searches. Both the SignalP (www.cbs.dtu.dk/services/SignalP) and TargetP (www.cbs.dtu.dk/services/TargetP) servers were used to look for secretion signals in the down-regulated Neurospora proteins. Manual Drosophila gene classifications (Table 2) were carried out by consulting Flybase (flybase.org) and The Gene Ontology (www.geneontology.org), as well as with literature and homology searches. The NCBI BLAST (blast.ncbi.nlm.nih.gov/Blast.cgi) and JGI (genome.jgi.doe.gov/genome-projects) search tools were used extensively in these analyses. The normalized microarray results have been deposited in the NCBI Gene Expression Omnibus (www.ncbi.nlm.nih.gov/geo), and the accession numbers for the Neurospora and Drosophila datasets are GSE35017 and GSE34997, respectively.

Drosophila RNAi Strains and Conditions

Flies were raised on the standard Drosophila medium at 25 or 29°C. The hs-GAL4 driver was located on the third chromosome, and the flies were obtained from the Bloomington Stock Center. The e13C-GAL4 [70] flies were a gift from Dr. Norbert Perrimon at Harvard University. The UAS-GRH^RNAi flies were obtained from the Vienna Drosophila RNAi Center (http://stockcenter.vdrc.at/control/main).

Epidermal Tissue Preparation and Immunostaining

Wandering third-instar larvae were rinsed in phosphate-buffered saline (PBS), dissected in Brower Fixative (0.15 M PIPES pH 6.9, 3 mM MgSO₄, 1.5 mM EGTA, and 1.5% NP40) with 4% formaldehyde (ultrapure, methanol-free from Polysciences Inc.), and fixed for 2 h at 4°C. To visualize melantoic spots, fixed larval tissues were washed in PBS with 0.1% Tween 20 and 0.1% Triton X-100 (PBTwx), mounted in Prolong Gold anti-fade reagent (Invitrogen), and imaged using phase-contrast microscopy on a Leica DM 2500 microscope. Epidermal tissues from adult abdomens were also dissected in Brower Fixative with 4% formaldehyde, fixed for 2 h at 4°C, and excess abdominal fat was removed by gentle aspiration.

For immunostaining, fixed epidermal tissues were washed in PBTwx, then incubated in a blocking solution of PBTwx with Western blocking reagent (WBR; Roche) for 1 h at room temperature. Incubations with primary antibodies were performed in PBT + WBR at 4°C overnight, and incubations with fluorescently labeled secondary antibodies were performed in PBT + WBR at room temperature for 2 hours. Primary antibodies utilized in this paper were as follows: guinea pig anti-GRH at a 1∶300 dilution [6] and mouse anti-Fasciclin 3 (7G10 concentrate, from the Developmental Studies Hybridoma Bank) at a 1∶400 dilution. Fluorescently labeled secondary antibodies from Invitrogen (Alexa Fluor 555 goat anti-guinea pig IgG and Alexa Fluor 488 donkey anti-mouse IgG) were used at 1∶400 dilutions. Tissues were mounted in Prolong Gold.

All fluorescent images were collected using a Leica SP2 laser-scanning confocal microscope, with identical instrument settings (at non-saturated gain levels) for both experimental and control samples. Optical sections were scanned at 0.6–0.8 µm thicknesses, and maximum-projection images are shown in Figures 4 and 5.

Drying Assays for Wandering Third-instar Drosophila Larvae

To cause clean injury, larvae were first rinsed in PBS, placed in a small drop of PBS on a black rubber block, and were punctured with a sterile 0.125 mm tungsten needle (Fine Science Tools) through their posterior-lateral epidermis, as described elsewhere [85], [86]. After epidermal injury, larvae were placed into either a Petri dish (60×15 mm) containing Whatman 3 MM chromatography paper moistened with PBS (moist conditions), or an empty Petri dish (dry conditions), and raised at 25°C. Mock-treated larvae were treated as above, except without any epidermal injury. Individual drying assays were performed using at least 40 larvae for each genotype, and each assay was repeated at least three times. Images of wandering third-instar larvae were obtained using a SteREO Discovery.V12 stereomicroscope (Zeiss), and images of representative larvae are presented in Figure 4.

Clean and Septic Epidermal Injury Experiments in Adult Drosophila

Approximately 24 h old male flies were heat shocked for 3 h at 37°C once a day for four consecutive days; flies were raised at 29°C between and after heat shock treatments. Clean or septic injuries were performed on the fifth day after eclosion using a 0.25 mm tungsten needle to puncture their dorsal abdomens, as described elsewhere [85], [86]. For septic injury, needles were dipped in gram-negative Erwinnia carotovora carotovora 15 (Ecc15) or gram-positive Microccocus luteus (M. luteus) bacterial solutions prior to injury. Mock-treated control flies were heat shocked and raised as described above, except without epidermal injury. Individual survival assays were carried out with at least 50 adult male flies for each genotype, and each assay was repeated at least three times. Injured males were placed into clean vials to monitor survival, with eight to ten flies per vial to avoid crowding. Surviving adult flies were counted and transferred to fresh vials every two days for ten total days. The averages of three experimental replicates are shown with the standard errors of the mean.