|Home | About | Journals | Submit | Contact Us | Français|
In plants and animals, cryptochromes function as either photoreceptors or circadian clock components. We have examined the cryptochrome from the filamentous fungus Neurospora crassa and demonstrate that Neurospora cry encodes a DASH-type cryptochrome that appears capable of binding flavin adenine dinucleotide (FAD) and methenyltetrahydrofolate (MTHF). The cry transcript and CRY protein levels are strongly induced by blue light in a wc-1-dependent manner, and cry transcript is circadianly regulated, with a peak abundance opposite in phase to frq. Neither deletion nor overexpression of cry appears to perturb the free-running circadian clock. However, cry disruption knockout mutants show a small phase delay under circadian entrainment. Using electrophoretic mobility shift assays (EMSA), we show that CRY is capable of binding single- and double-stranded DNA (ssDNA and dsDNA, respectively) and ssRNA and dsRNA. Whole-genome microarray experiments failed to identify substantive transcriptional regulatory activity of cry under our laboratory conditions.
Cryptochromes (CRY) are defined as proteins that share sequence similarity with the DNA-repairing enzyme photolyase but lack conventional photolyase activity (45, 58, 59). Phylogenetic analysis of cryptochrome sequences indicates three general classes: plant cryptochromes, animal cryptochromes, and DASH-type cryptochromes (46). Structurally, most cryptochromes have an amino-terminal photolyase-related region (PHR) and a carboxyl-terminal domain. However, the carboxyl-terminal domain is missing in DASH-type cryptochromes (46). In Arabidopsis thaliana, cry1 and cry2 encode cryptochromes required to mediate the entrainment of the circadian clock, as well as expression of a variety of light-induced genes and developmental processes (31, 48, 62, 72). In contrast, Arabidopsis cry3 is a DASH-type cryptochrome and appears to have a photolyase activity specific to single-stranded or looped duplex DNA (55, 58, 59). Animal cryptochromes serve functions similar to those of their plant counterparts—acting as either photoreceptors or components of circadian clocks (46). For instance, in Drosophila, cry is a blue-light photoreceptor (21, 63). Once light activated, CRY facilitates resetting of the clock by mediating the light-dependent degradation of TIM (7). In contrast, mammalian cry appears to play no role in circadian photobiology but instead is an integral part of the negative feedback loop of the clock (43). Light-dependent and -independent effects of animal cry on the regulation of the circadian clock have been well documented for both Drosophila and mice (30, 40, 60, 68). In addition to its widespread clock functions, animal cry has been shown to mediate light-dependent magnetosensitivity in Drosophila (27).
Neurospora crassa has served as a model organism to study the circadian clock and light responses in eukaryotic cells for several decades (13, 19, 34, 35, 56). Two GATA family zinc finger transcription factors, white collar-1 (WC-1) and white collar-2 (WC-2), form an obligate heterodimer (WCC) via their Per-Arnt-Sim (PAS) domains that acts both as the photoreceptor for light responses/clock resetting and, in the dark, as a transcription factor complex that promotes the expression of FRQ, a core clock component analogous to PER/CRY in mammalian clock systems. WC-1 shares extended sequence and functional similarity with one of the heterodimeric activators in the mammalian feedback loop, BMAL1 (brain and muscle Arnt-like protein-1) (44); in the feedback loop, WCC drives expression of FRQ, which then feeds back to repress the activity of its activator, WCC, resulting in closure of the negative feedback loop (19). WC-1 is also a flavin adenine dinucleotide (FAD) binding protein capable of sensing light. For light responses, WCC activates downstream target genes through recognition of the light-responsive elements (LREs) (9, 23) in the promoters of target genes and activation of another light-responsive transcription factor, sub-1, to activate late light responses (8). In the absence of WC-1 or WC-2, most early and late light responses are lost, suggesting their dominant roles in mediating light signals in Neurospora. Notably, despite extensive efforts in several labs to identify blind strains, only wc-1 and wc-2 mutants have been repeatedly isolated (47). However, unexpectedly, the completion of the Neurospora genome sequencing revealed the presence of an additional putative blue-light photoreceptor, cryptochrome (cry), based on its sequence similarity to known cry genes (25). Given the regulatory similarities in the molecular basis of the circadian feedback loop and light signaling cascades among different model systems, we were interested in characterizing the clock and light functions of this novel cry at the molecular level.
We show that Neurospora CRY shares a high level of sequence similarity and domain structure with other DASH-type members, including FAD and MTHF (methenyltetrahydrofolate) binding sites as well as the residues potentially interacting with cyclobutane pyrimidine dimers (CPDs). Spectral analysis with purified CRY verified the interactions with FAD and MTHF. Like other photoreceptors in Neurospora, the transcript and protein levels of cry are highly induced by light in a wc-1-dependent manner. Meanwhile, we found that the transcript of cry in the dark is under circadian clock control, with peak expression antiphasic to frq, while the protein level becomes quickly dampened after 12 h in darkness. Race tube analysis with the knockout strains suggests that Neurospora cry is not a clock component insofar as its loss does not change the free-running period of the circadian rhythm; however, an alteration of the light-entrained phase is noted in cry mutant strains. Interestingly, Neurospora CRY is capable of binding single- and double-stranded DNA and RNA in vitro, as demonstrated by electrophoretic mobility shift assays (EMSA). However, whole-genome microarray analysis indicates that both early and late transcriptional light responses remain unaltered in the absence of cry. Thus, although the explicit signal transduction pathway that CRY impacts is unclear, we show that its loss affects the phase of entrainment of the circadian clock by light.
The N. crassa strains used were 343-25 (ras-1bd cry::hph), with 328-4 (ras-1bd) as the corresponding ras-1bd cry+ strain, and 378-5 (cry::hph), with 378-6 as the corresponding wild-type (WT) strain. 378-5 and 378-6 are sibling progeny from a cross of 343-25 with N. crassa wild-type FGSC 2489 (OR74A). For circadian entrainment experiments, an independently but almost identically produced knockout strain was used. The 5′ and 3′ untranslated regions (UTRs) of NCU00582.3 (cryptochrome) were amplified with 5′CTTGCCTCTTCCAACTTGAG3′ (forward) and 5′CTTGGACAACAGGTAAGG3′ (reverse) and with 5′CCGCTTTCCTTACCAGCCTGCG3′ (forward) and 5′CCGCGCCTCCGTGCTCAAATAC3′ (reverse), respectively. The hygromycin resistance cassette-encoding hph was amplified from pCSN44 with GAGGTCGACAGAAGATGATATTG (forward) and GAGTCTAGAAAGAAGGATTACCTC (reverse). Blunt-end ligations followed by restriction digests and subsequent PCR amplification yielded fragments consisting of 5′UTR-hph and hph-3′UTR. An additional PCR step using primers within the hygR cassette, hSM-f (5′CGTCGCGGTGAGTTCAGGCTTTTT3′) together with the 3′ UTR reverse primer and hSM-r (5′TCGCCTCGCTCCAGTCAATGACC3′) (both sequences from H. Colot [personal communication]) together with the 5′ UTR forward primer, generated two fragments with a ca. 600-bp overhang within the hph-encoded hygR cassette. These were gel purified and cotransformed into ras-1bdA (FGSC 1489), where they underwent homologous recombination. Strains selected for hygromycin resistance were verified for correct insertion by PCR, restriction digests, and sequencing. Homokaryons were produced by microconidiation and are called ΔcryCM in the text. A post hoc analysis confirmed that the strains created in the two labs performed comparably in all circadian protocols. FGSC 2489 (OR74A) was also used for all microarray expression analyses and reverse transcription-PCR (RT-PCR) follow-up experiments, and for microarray work, the cry and vvd knockout strains came from the Neurospora knockout project (11) and have been deposited in the Fungal Genetics Stock Center, Kansas City, MO (www.fgsc.net). NCU strain numbers are from Neurospora annotation (http://www.broad.mit.edu/annotation/genome/neurospora/Home.html).
Construct pAF67 was generated in which the 5′ and 3′ sequences flanking the CRY open reading frame (ORF) were appended to the hygromycin B phosphotransferase (hph) coding region, whose expression was driven by the Aspergillus nidulans trpC promoter, thus allowing selection by hygromycin B resistance. pAF67 was transformed into N. crassa, and hygromycin-resistant transformants containing single homologous integrations of the DNA were backcrossed to the wild type (WT) to obtain homokaryotic disruption strains (343-25), as judged by DNA gel blot analysis of genomic DNA (data not shown). Because the annotation of the N. crassa genome was incomplete when 343-25 was made, the disruption retained the first exon but spanned the entire second exon, including all chromophore binding sites. The pAF76 construct, similar to those described by Larrondo et al. (42), was used for overexpressing CRY in Neurospora and contains the quinic acid-2 (qa-2) promoter and 5′ UTR followed by a copy of the genomic region spanning the CRY ORF and then the Aspergillus nidulans trpC 3′ UTR in a his-3 targeting vector. The qa-2 promoter enables regulated gene induction by the addition of quinic acid. pAF76 was transformed into N. crassa strain 87-74 (ras-1bd; his-3) following standard transformation protocols so that the overexpression of cry mRNA, confirmed to be approximately 25-fold (as determined by quantitative RT-PCR, n = 3; data not shown), is driven from a transgene at the his-3 locus.
The general conditions for growth and light manipulations are described elsewhere (8, 16). Liquid culture and race tube experiments were performed as previously described (1, 14, 24) except those experiments with the CRY overexpressor strain, which used a modified growth medium containing 1× Vogel's medium, 0.17% arginine, 50 ng/ml biotin, 1.5% Bacto agar, and 0.015 M quinic acid. For the expression analysis of cry and frq following light exposure, cultures were irradiated with ~40 μmol photons/m2/s of white light (GE F20T12-CW cool white fluorescent light bulb) for the indicated durations, following of 20 h of dark incubation. Red-, far-red-, and blue-light experiments were performed using an E-30LED growth chamber equipped with red-, far-red-, and blue-emitting diodes (Percival Scientific, Inc., Perry, IA). For circadian entrainment in race tubes, methods were as described previously (64) using 4 μmol photons/m2/s of light and race tube media lacking glucose (or any other obvious carbon source; in this case, OSRAM cool white fluorescent light was used). Data were analyzed using CHRONO and CHRONOSX (57). For microarray experiments, 1 × 107 conidia were inoculated into a 10-cm petri dish with 20 ml Bird medium (50) containing 2% glucose. After 24 h of incubation in darkness at 25°C, a mycelial plug was cut with a no. 4 cork borer (diameter, 8 mm) and transferred into a 125-ml flask with 50 ml Bird medium containing 2% glucose. All procedures were performed under a low-red-light environment to avoid any possible light-stimulating effects. After another 24 h of culture with constant shaking (125 rpm) in darkness (DD) at 25°C, the flasks were moved to a shaker at 25°C with a continuous white-light stimulus (LL) (GE F20T12-CW, 40 to 50 μmol photons/m2/s), harvested at the indicated time points using vacuum filtration, immediately frozen in liquid nitrogen, and stored at −80°C until RNA extraction.
The perithecial beak phototropism assays were performed on cultures grown on Westergaard's synthetic crossing medium. In brief, strains were inoculated onto crossing plates and kept at 25°C in DD for 7 days. Fifty microliters of conidial suspension (strain 343-25, 328-4, or wc-1ER53) was pipetted as a thin line along the diameter of the plate, and the plates were returned to the dark. The plates were exposed to a 12-h light–12-h dark cycle (the saturating light was provided by fluorescent lighting), with the plates positioned in a box with a 4-cm-wide opening so that the direction of the light was perpendicular to the line of perithecia. Perithecial beaks were scored 14 days after inoculation, with the orientation of the beaks (toward, neutral, or away) scored relative to the direction of the light. For the dark-grown samples, an arbitrary “direction” was chosen for scoring purposes.
The RT-PCR analysis was performed as previously described (24). The cry-specific primers were ACF160 (5′AAGTGGCATAGATTTCAAGC3′) and ACF161 (5′TTCGGTAGGTGGTGAAGACA3′). Statistical differences between RNA transcript levels were tested by one-factor analysis of variance and a subsequent post hoc Dunnett t test, where significance was set at a P value of <0.05. Northern blotting and Western blotting were performed as previously described (23).
The DNA sequence encoding the full-length CRY cDNA was PCR amplified using primers designed to add an NdeI site proximal to the start codon and an EcoRI site immediately before the stop codon. The product was digested and inserted into pET24a (Novagen), and the resulting C-terminally six-His-tagged expression construct (pAF77) was transformed into BL21-Codon Plus-RIL cells (Stratagene). Transformed cells were grown to logarithmic phase and induced with 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside) for 4 h at 37°C. Following sonication of the cells, a majority of the six-His-tagged CRY polypeptide was in the inclusion body fraction, necessitating the addition of 6 M urea for solubilization. This soluble protein was purified by nickel chelate affinity chromatography using Ni-nitrilotriacetic acid (Ni-NTA) Superflow agarose (Qiagen) according to the manufacturer's directions. The eluate was dialyzed against phosphate-buffered saline (PBS) buffer (137 mM NaCl, 10 mM KH2PO4, 100 mM Na2HPO4, 27 mM KCl, pH 7.4) and used directly as the antigen for injection into rabbits following standard procedures (Pocono Rabbit Farm & Laboratory, Canadensis, PA).
Using BL21-Codon Plus-RIL cells transformed with pAF77, six-HIS-CRY was induced with 0.75 mM IPTG for 4 h at 37°C, purified over a Ni column, and then dialyzed with 1× PBS. To isolate glutathione S-transferase (GST)-tagged CRY, full-length CRY was PCR amplified from pAF77 using primers designed to add an EcoRI site proximal to the start codon and a NotI site immediately before the stop codon. The product was digested and inserted into pET41b (Novagen), and the resulting GST/six-His/CRY/six-His expression construct (pAF88) was transformed into BL21-Codon Plus-RIL cells and then induced with 0.1 M IPTG for 6 h at 30°C before being purified over a Ni column. The CRY-containing fractions were pooled and run over an Amersham GSTrap FF column according to the manufacturer's instructions. For GST protein production, pET41b with no insertion was used under similar induction/purification conditions. Cleaved CRY was generated by the following steps: GST-CRY was bound to a GSTrap FF column (Amersham), and the column was then loaded with thrombin and left at room temperature overnight. An Amersham HiTrap Benzamide FF (high sub) column was placed in line with the GST-CRY-containing column, and cleaved CRY was eluted off the column (thrombin retained in the HiTrap column) using 1 M NaCl and 1× PBS. Eluted fractions were dialyzed against 1× PBS with 20% glycerol. Binding was done in a 20-μl reaction volume for 20 min at room temperature; the reaction mixtures contained 0.5 fmol biotinylated probe, 100 ng bovine serum albumin (BSA), 17.5 mM Tris-HCl, pH 7.5, 52.5 mM NaCl, 0.35 mM EDTA, 5 mM dithiothreitol (DTT), 2.5% glycerol, and the indicated proteins. The following protein amounts were used (see Fig. 6): 2 μg CRY-HIS, 4.5 μg GST-CRY, 3.5 μg GST, and 2 μg cleaved CRY. Reaction mixtures included the following increasing amounts of CRY: 0, 0.0625, 0.125, 0.25, 0.5, 1, and 2 μg. The products of the binding were subjected to electrophoresis on 0.5× Tris-borate-EDTA (TBE)–5% acrylamide gels with 0.5× TBE buffer at 4°C/150V/60 min. The nucleic acids were transferred to nylon membrane (Hybond+), UV cross-linked, and detected using the LightShift chemiluminescent EMSA kit (Pierce) according to the manufacturer's instructions. DNA and RNA probes consisted of 30-mer oligonucleotides that were biotinylated at the 5′ end (IDT). The single-stranded probes were a single oligonucleotide, whereas the double-stranded probes consisted of two complementary oligonucleotides annealed by heating to 100°C and then slowly cooling the mixture to room temperature in 10 mM Tris, pH 8, and 0.04× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate). The sequence of the DNA probe was 5′GGGGGTGGGGTTGGGTTTGGTTTTGTTTTT3′, and that of the RNA probe was 5′GGGGGUGGGGUUGGGUUUGGUUUUGUUUUU3′. The probe sequences were designed to minimize improper annealing or secondary structure.
Conidia were harvested, treated, and diluted with 1 M sorbitol. Conidial suspensions were irradiated with various levels of UV light using a UV Stratalinker. Appropriate dilutions of the conidia were plated in duplicate in standard transformation media (1× Vogel's, 1× fructose-inositol-glucose-sorbose medium, 1.5% agar cooled to ~45°C) (11), one plate from each pair was exposed to fluorescent light at 40 to 60 μmol photons/m2/s for 60 min (photoreactivation), and then all plates were kept in constant darkness until colonies formed. All subsequent work was done under red safe lights. The colonies on each plate were counted, and the numbers were plotted as percent survival relative to non-UV-treated conidia. The strains used for the photoreactivation assay were 378-6 (WT), 378-5 (cry::hph), and FGSC 8929 (phrRIP).
Bacterially expressed and purified His-CRY (375 μg) was mixed 1:1 with chloroform-isoamyl alcohol (24:1), vortexed for 5 min, and centrifuged for 10 min at 16,000 × g at 4°C. The upper aqueous phase, containing the released chromophore, was transferred to a new tube, and the extraction was repeated. Fluorescence emission between 470 nm and 600 nm was measured for the aqueous chromophore-containing fraction using an ISS PC1 photon-counting spectrofluorometer and excitation at 450 nm/4.0-mm slit. Fluorescence measurements were also performed on FAD and flavin mononucleotide (FMN) standards under the same conditions at pH 2 and pH 7.4. Protein absorbance spectra were measured between 300 and 600 nm using a Beckman DU 600 spectrophotometer.
Thin-layer chromatography (TLC) was performed essentially as described previously (10). His-tagged CRY was purified as described above. Then, 200 μg of CRY was combined with ethanol to a final concentration of 70% ethanol, boiled for 2 min, chilled on ice, and centrifuged at 16,000 × g for 10 min at 4°C. The supernatant was snap-frozen, lyophilized in a speed vacuum, and then resuspended in 10 μl of 35% ethanol. FAD, FMN, and riboflavin standards were dissolved in 35% ethanol. The TLC plate was prewarmed, and 10 μl of the standards and CRY chromophore was loaded. The plate was placed in a TLC chamber saturated with 150 ml n-butanol–acetic acid–H2O (3:1:1) and incubated for 3 h.
The detailed protocols for microarray sample preparation, hybridization, image, and data analysis have been described by Chen et al. (8).
Microarray data from the wild-type strain were used as class I (n = 3) in contrast to class II, the data from the cryptochrome knockout strain. A two-class unpaired test was performed with a t-statistic method. Missing data were imputed via a k-nearest neighbor algorithm (k = 10) before permutations (n = 100) for estimating the false discovery rate (FDR).
Complete data sets are available on Gene Expression Omnibus (data accession numbers GSE8932 and GSE14909; http://www.ncbi.nlm.nih.gov/projects/geo).
CRY (encoded by gene NCU00582.4) is predicted to be a 722-amino-acid protein expressed from a 2,956-nucleotide (nt) transcript having one intron (Fig. 1A). Based on the phylogenetic analysis of the cryptochrome family across different species, CRY is categorized as a DASH-type cryptochrome (46), most similar to those found in bacteria, plants, and animals (15, 46). Given that Arabidopsis thaliana cry3 (Atcry3) is one of the best-characterized members in this family, we aligned the amino acid sequence of Neurospora CRY (XP_965722) with that of AtCRY3 (NP_568461), together with those of two other DASH-type members from Xenopus (XlCRY-DASH, NP_001084438) and zebrafish (DrCRY-DASH, NP_991249), to check the integrity of its functional regions. Overall amino acid similarity between Neurospora CRY and AtCRY3 is about 44%, with a very low E value (4e-53, NCBI BLASTP). Of 16 amino acids which act in binding FAD in AtCRY3 (marked by black asterisks in Fig. 1) (5), 15 are conserved in CRY. In addition, four of the five residues (E129, E130, E459, and Y465, marked by green asterisks in Fig. 1) essential for hydrogen bonds with methenyltetrahydrofolate (MTHF) (37) are identical across the four species, suggesting that Neurospora CRY might bind FAD and MTHF in a manner similar to that of other members in the DASH-type family. The crystal structure study of AtCRY3 predicts six residues to directly interact with CPD-containing DNA (37); all six residues (R281, E342, W345, N433, R434 and Q437, marked by red asterisks in Fig. 1) are conserved in CRY, suggesting that Neurospora CRY might also interact with DNA (55, 59) in a manner similar to that of AtCRY3. In addition to the DNA/FAD binding domains which constitute the N-terminal two-thirds of the protein, CRY contains an ~200-amino-acid C-terminal region rich in glycines (Fig. 1A) not found in any other predicted members of the photolyase/cryptochrome family.
To find out if the Neurospora CRY could function as a photolyase in vivo, a disruption mutant strain of cry (343-25) was tested for photoreactivation. In the WT strain, increasing exposure to UV irradiation results in decreased survival, while light treatment following UV exposure increases survival of the WT due to the activation of DNA-repairing photolyase (Fig. 1C). With light treatment following UV irradiation, the survival of 343-25 was restored to the same level as in the wild-type strain, suggesting that the photolyase activity is undamaged in the absence of CRY. In contrast, as an internal control for the assay, a loss-of-function mutant strain of the previously characterized Neurospora CPD photolyase (61), phrRIP, showed no sign of increased survival with light treatment, suggesting a total lack of photoreactivation in this strain. The phr gene appears to encode the only photolyase in Neurospora, and CRY, lacking any detectable photoreactivation activity at least under these assay conditions, would be by definition a cryptochrome.
To empirically test whether Neurospora CRY binds FAD and MTHF, CRY was expressed and purified in Escherichia coli using a six-His tag and Ni-NTA affinity chromatography. A small proportion of the expressed protein was soluble, allowing purification of CRY under nondenaturing conditions with its associated cofactors as shown in Fig. 2A. Fractions containing CRY appeared yellow, supporting that CRY binds a flavin (59). The CRY absorbance spectrum exhibited a major peak at 375 nm, two minor peaks at 445 and 470 to 472 nm, and a tail extending out to 600 nm (Fig. 2B). The dominant peak at 375 nm suggests the presence of MTHF (49); the two smaller peaks indicate the presence of fully oxidized FAD (445 nm) and flavin neutral radical (FADH0, 470 to 472 nm) (3, 37). Notably, the absorption spectrum of CRY is almost identical to the spectra of Atcry3 shown to bind FAD and MTHF (37). To confirm the flavin moiety, the bound cofactors were released from CRY by a chloroform extraction method and the fluorescence spectrum was recorded. Measurement of the excitation at 450 nm resulted in fluorescence emission with a maximum at 522 nm that was 3.5-fold higher at pH 2.0 than at pH 7.4 (Fig. 2C). This emission peak is indicative of a flavin moiety, and the increase in fluorescence intensity with decreased pH identifies FAD as the specific flavin derivative (70). A thin-layer chromatography (TLC) assay further confirmed one of the released cofactors to be FAD (Fig. 2D). Overall, our spectral analysis data support that the purified Neurospora CRY was loaded with FAD and MTHF, as predicted by the sequence analysis.
To determine if the cry gene is regulated by light, Northern blot analysis was performed on RNA isolated from fungal mycelia. A 30-min white-light treatment resulted in a dramatic increase in cry transcript levels relative to a dark-grown sample (Fig. 3A). The light-induced cry transcript ran at approximately the same position as the smallest Neurospora rRNA species, indicating that the cry transcript is approximately 2 kb, in agreement with the cDNA analysis. Light failed to induce cry transcript levels in a WC-1 mutant (wc-1ER53), suggesting that WC-1 is the photoreceptor/transcription factor responsible for the light regulation of cry transcription. Using blue-, red-, or far-red-light-emitting diodes (LEDs) as the light source, only blue-light treatment resulted in an increase in Neurospora cry transcript levels (Fig. 3B). WC-1 is responsive to blue-light wavelengths but not to red or far-red light, consistent with WC-1 being the photoreceptor directly responsible for the light induction of cry. Further quantification was performed using RT-QPCR. A wild-type strain was subjected to increasing doses of white light for 15 min to 24 h. The cry transcript was rapidly and highly induced, reaching levels 200- to 300-fold higher than levels found under dark conditions within 15 to 30 min (Fig. 3C). The cry transcript levels declined following the initial peak due to photoadaptation but remained elevated approximately 50-fold higher than dark levels after 24 h in constant light.
To detect CRY protein levels and study its regulation, we generated antiserum against the full-length CRY protein. Western blot analysis of extracts from the WT and Δcry strains demonstrated the specificity of the antiserum (Fig. 3D). The antibody recognized two proteins: one was nonspecific, and a single specific protein band was detected in the WT extracts, but not in the disruption mutant (343-25) extracts, at the approximate predicted molecular mass of CRY (80 kDa). CRY was present at very low to undetectable levels in the dark but was detectable following as little as 30 min of light treatment (Fig. 3D). The CRY protein levels continued to increase, reaching a peak between 4 and 8 h, and then remained elevated under constant light exposure (Fig. 3D and E). Although both the transcript and protein levels remain elevated under constant light, the profiles differ slightly, with transcript levels rapidly peaking at 15 to 30 min and protein levels lagging by several hours.
The transcript levels of cry genes from organisms, including Arabidopsis, Drosophila, and mouse, are circadianly regulated, with amplitudes of less than 3-fold (21, 33, 41, 51, 66). To test for circadian regulation of the cry transcript, mycelial tissue of approximately the same developmental age was harvested at 4-h intervals following a light-to-dark transfer. A light-to-dark transfer resets the Neurospora clock to subjective dusk, after which the clock continues to run in constant darkness (19). The circadian regulation of the central clock component frequency (frq) was apparent under these conditions, with a peak in mRNA levels occurring ~12 to 16 h in constant darkness (F8,27 = 7.3; P = 0.0001), as shown in Fig. 4A (1). The cry transcript appeared to be circadianly regulated (F8,27 = 2.27; P = 0.05), with levels decreasing following the light-to-dark transfer, reaching a trough after ~12 to 16 h followed by an increase reaching a peak after a total of ~20 to 24 h in the dark. The cry transcript oscillated with an ~2-fold amplitude but at levels well below those seen in the light (i.e., time zero in Fig. 3C is equal to 20 h in Fig. 4A). Notably, in Neurospora, where most clock-regulated genes are morning specific (12), including frq, cry appears to be an evening-specific gene. Meanwhile, CRY protein levels remained elevated for the first 12 h but then decreased to low/undetectable levels by 16 h with no subsequent increase detected, as shown in Fig. 4B. In this regard, CRY is similar to the Neurospora photoreceptor VVD, which regulates photoadaptation and thereby acts to modify the primary biological response elicited by the WCC. As a positive control, FRQ abundance and phosphorylation patterns were seen to oscillate in a circadian manner as previously described (26). Although CRY does not appear to be circadianly regulated, it is possible that CRY is fluctuating at very low levels, similar to what is found in the cry transcript.
To determine whether cry plays any role in the Neurospora circadian system, the WT strain, a cry disruption strain (343-25), and a strain overexpressing cry (cryOX; see Materials and Methods) were grown in race tubes as described previously (19). Strains 343-25 and cryOX had periods (22.15 ± 0.07 h and 22.77 ± 0.11 h, respectively) similar to that of the WT (22.44 ± 0.11 h; data not shown) in constant darkness, indicating that the cry gene is not a core clock component. In addition, light induction of frq, a mechanism by which light input resets the clock (14), was also unaltered in strain 343-25, suggesting that cry is not involved in light signaling to the clock (data not shown). Using a temperature step-up from 4°C to 25°C (in constant darkness) to reset the race tube cultures also resulted in a WT period for strains 343-25 and cryOX (data not shown).
We then investigated synchronization of the circadian clock to zeitgeber cycles, a protocol that yields a strain-specific phase relationship between endogenous (circadian) and exogenous cycles, an active process called entrainment. (Zeitgebers [from the German for “time-givers”] are environmental stimuli that alter the phase of a biological clock; light and temperature are common examples.) In temperature cycles of 22°C to 27°C, ΔcryCM resembled the wild-type strain (data not shown). That is, conidiation occurred at the same phase of the temperature cycle for both the mutant and the wild type. However, in light cycles (with incubations in light followed by darkness), a delay in conidiation (interpreted as a delay in entrained phase) was observed in the absence of CRY (Fig. 5). In general, when white light was used, longer photoperiods yielded larger phase delays (e.g., 50 min at a light/darkness cycle [LD] of 12 h of light/12 h of darkness [12:12] versus 2 h at LD 18:6; Fig. 5A and and5B).5B). An additional protocol for entrainment of circadian rhythms is that of skeleton photoperiods, which calls for delivering a light pulse at either end of a dark incubation. This procedure often mimics full photoperiod entrainment, suggesting that the lights-on and lights-off signals play a dominant role in entrainment. When we replaced dawn and dusk by 2-h light pulses (LD 2:10:2:10), the phase of entrainment of the ΔcryCM strain was delayed by 3 h compared to that of the wild-type strain (Fig. 5C).
In addition to the interaction with the CPD-containing DNA, DASH-type members from Arabidopsis thaliana (39) and Synechocystis sp. strain PCC6803 (5) have been shown to bind to double-stranded DNA nonspecifically, although the functional significance of this binding is not yet understood. We tested CRY's ability to bind to DNA using bacterially expressed CRY in electrophoretic mobility shift assays (EMSA).
We purified Neurospora CRY with either a six-His C-terminal (CRY-HIS) or a GST (glutathione S-transferase) N-terminal (GST-CRY, Fig. 6A) tag. The CRY-HIS protein was used in binding reactions with a 30-bp double-stranded DNA probe. CRY-HIS clearly bound to the double-stranded DNA probe, as shown by the significant decrease in migration of the DNA (Fig. 6B, compare lanes 1 and 2). To ensure that the slower migration of the DNA was due specifically to binding by CRY and not by a contaminant from the purification, we repeated binding reactions using CRY purified with an N-terminal GST tag. The use of GST-CRY in a binding reaction also resulted in a decreased migration of the DNA (Fig. 6B, lane 3), and more importantly, the GST-CRY reaction had a greater decrease in mobility of the DNA (compare lanes 2 and 3) due to the mass increase of the GST tag. GST alone did not bind to the DNA (lane 4). As an additional control, we cleaved CRY from GST and used this cleaved CRY in a binding reaction which resulted in a decrease in migration of the DNA (lane 5) similar to the HIS-CRY reaction. Taken together, these results indicate that Neurospora CRY is capable of binding to double-stranded DNA in vitro. Meanwhile, CRY appears to bind to DNA in a non-sequence-specific fashion, because binding reactions using DNA probes of different sequences produced similar results and addition of unlabeled DNA of unrelated sequence to the binding reactions resulted in decreased binding of CRY (data not shown).
To further elucidate the nature of the CRY/DNA interaction, we tested CRY's ability to bind to single-stranded DNA and single-/double-stranded RNA. Our rationale for these additional experiments came from several observations. First, AtCRY3 is capable of binding to single-stranded DNA as efficiently as to double-stranded DNA (39). Second, RNA fragments were found to interact with Vibrio cholerae Cry1, another DASH-type member, when it was expressed and purified from E. coli (71). Third, the C terminus of Neurospora CRY contains multiple arginine-glycine-glycine (RGG) repeats, which usually participate in protein-RNA or protein-protein interactions (29). We found that Neurospora CRY was capable of binding double-stranded RNA (lane 7), single-stranded DNA (lane 11), single-stranded RNA (lane 13), and even an RNA-DNA hybrid molecule (lane 9). To begin to determine the relative affinity that CRY has for each type of nucleic acid, we performed a series of binding reactions using increasing amounts of CRY with a fixed amount of each nucleic acid species (dsDNA, ssDNA, dsRNA, or ssRNA) (Fig. 6C). CRY appears to exhibit similar affinities for dsDNA and dsRNA and a lower affinity for ssDNA. The low intensity of the ssRNA probe (bound or unbound) makes it difficult to quantitatively compare binding relative to that of the other three probes. Notably, multiple slower-migrating bands were seen in the binding reactions with dsDNA and dsRNA (Fig. 6C). The slower-migrating bands become more predominant as the CRY concentration increases, presumably due to multiple CRYs bound to each probe, supporting the notion that Neurospora CRY binds to dsDNA/dsRNA in a sequence-independent manner.
Given that Neurospora CRY is a putative photoreceptor capable of interacting with DNA directly, we hypothesized that Neurospora CRY might have some light-related functions regarding gene regulation epistatic to WCC and decided to test it using microarrays with full-genome coverage (18, 65). It has been suggested that the DASH-type CRY in Synechocystis sp. PCC6803 functions as a transcriptional repressor (5). Identification of bona fide light-responsive genes in our previous study (8) provides a list of good targets in which to explore the transcriptional regulatory activity of Neurospora CRY in response to a light stimulus. Five sequential time points, from 0 to 15, 30, 60, and 120 min after the onset of light, were monitored for three strains, the WT, Δcry, and Δvvd strains (data accession number GSE 8932). As shown in Fig. 7A and and7B,7B, the regulation of most early light-responsive genes (ELRGs) and late light-responsive genes (LLRGs) appeared unchanged in the Δcry strain. There was no sign of photoadaptation defects in the Δcry strain in contrast to the Δvvd strain, which is a strain defective in regulating photoadaptation for all light responses (8). To add a statistical verification to our conclusion, we focused on two specific time points (DD and LL60) and repeated the microarray experiments each with three independent biological replicates (data accession number GSE14909). The data were analyzed with the SAM (significance analysis of microarrays) package using a 5% false discovery rate (FDR) as a cutoff (67). SAM identified only four genes, including the cry transcript, which were significantly different between the WT and Δcry strains. However, with the exception of cry (as an intrinsic positive control for our approach), none of the other genes could be further validated by RT-QPCR (data not shown). In separate experiments, quantitative RT-PCR analyses in a ras-1bd Δcry strain failed to find statistically significant increases in light induction of frq, al-1, or con-6 expression (22a). Although small effects on the expression of al-1 or con-6 have been recently reported (53), we expect the differences in results are due to genetic background effects rather than major undiscovered influences of CRY, and indeed the reproducible lack of major effects seen in our microarray analyses, which might well not report the results shown in reference 53 as statistically significant, supports this interpretation. Given cryptochrome's central role in the photobiology of many other organisms and the delay in entrained phase in the cry mutant, it is possible that Neurospora CRY shows more substantive transcriptional regulatory activity under nonlaboratory conditions or at other developmental stages in the Neurospora life cycle.
Given that CRY protein levels remain elevated under constant light exposure (Fig. 3D and E), we tested if there is any detectable long-term light-dependent developmental phenotype in the absence of cry. Constant light would trigger the carotenoid biosynthesis pathway, which can be seen in the accumulation of orange pigmentation. However, no gross defects in light-induced carotenoid biosynthesis can be detected (data not shown). Another light-regulated process, the development of Neurospora's sexual spores in the perithecial organ, is also regulated by light (17). Positive phototropism of the perithecial beaks (beak bending, a maternal effect) is induced by blue light, resulting in the sexual spores being ejected toward the direction of light (32). We tested for defects in perithecial phototropism by inoculating crossing plates with one of the three strains (ras-1bd, 343-25, or ras-1bd wc-1ER53) as the female parent and fertilizing these crosses with a WT strain. The plates were placed in directional lighting and then scored for the direction of the perithecial necks relative to the direction of the light. The WT perithecial necks displayed random growth directions when grown in the dark, whereas 82% of the WT perithecial necks pointed toward the light, as shown in Fig. 7C. The wc-1 mutant displayed a random distribution of perithecial necks both in the dark and with directional lighting. In contrast, 343-25 exhibited WT perithecial phototropism, with 74% of perithecial necks pointing toward the light. Having found no differences in 343-25 in the developmental processes known to be regulated by light nor in general growth rate (data not shown), we looked for novel phototropism during Neurospora's asexual life stage. Different assays for phototropism were tested using a variety of culturing methods in combination with various directional lighting configurations (24). Culturing systems included standard shaking and static liquid cultures, solid media in petri plates, and race tube assays, including modified race tubes which enabled inoculation in the middle of the tubes to allow analysis of fluence responses. In order to study the fine branching structure of mycelia, we used thin vertical gels consisting of standard media poured between two glass plates similar to those used to electrophorese protein samples. These various culturing setups were combined with lighting configurations consisting of standard fluorescent lighting, sunlight, and custom-fabricated light-emitting diode arrays (blue, red, and white). The light sources and intensities were tested in a variety of configurations (e.g., high-intensity blue light above the vertical gels and low-intensity red light below). Under none of the conditions tested did we find any signs of phototropism in the WT or mutant cry strains (data not shown).
DASH-type CRY members have been identified from various organisms, ranging from bacteria and plants to fungi and animals. In Synechocystis sp. PCC 6803, CRY-DASH has been shown to have no photolyase activity in vitro but weak photolyase activity in vivo (36). The same protein has nonspecific DNA binding ability and possibly acts as a transcriptional repressor for a few genes (5). In Vibrio cholerae, two DASH-type members (Vccry1 and Vccry2) have no photolyase activity in vivo or in vitro (71). In Arabidopsis thaliana, the only DASH-type member, Atcry3, is capable of binding to single- and double-stranded DNA (nonspecifically) in vitro but without any conventional photolyase activity in vivo (39); however, biochemical and structural studies suggest that Atcry3 in vitro can act as a CPD photolyase with specificity for either single-stranded (37, 59) or looped double-stranded (55) DNA. Among animals, DASH-type members have been identified only in Xenopus laevis and Danio rerio (15); both have a weak photolyase activity when expressed in E. coli. No nonspecific DNA binding activity could be detected in vitro, and only DrCRY-DASH displays CPD binding activity on double-stranded DNA (15). Overall, although weak photolyase activity and DNA binding ability have been demonstrated for several DASH-type members, the definitive biological function of these DASH-type cryptochromes is still unclear: no strong phenotype could be observed in any case, and they might be involved in other biological processes not necessarily requiring DNA repair functions (5, 69). Because of this, we have focused on exploring clock and light phenotypes in the knockout strain of Neurospora crassa CRY, focusing on characteristics that have not yet been extensively examined in other DASH-type members. To be clear, the CryA recently reported in Aspergillus nidulans is actually a class I CPD photolyase based on both enzymatic and phylogenetic analyses (2).
Since it was first identified in 2003 (25), Neurospora CRY has been referred to as a cryptochrome ortholog based on sequence similarity. Later, it was categorized as a DASH-type member (5, 15, 19), which is the only type of cryptochrome existing in the fungal kingdom; no animal or plant-type cryptochromes have been detected in the sequenced fungal genomes. However, it was surprising to us that free-running rhythms and a catalog of other light-regulated phenotypes are not perturbed in the absence of cry. Only entrainment in light cycles gave a mutant phenotype. The data clearly suggest that Neurospora CRY does not participate in the transcription-translation feedback loop as does its counterpart in other organisms, but this is in fact consistent with findings for other DASH-type members. For instance, in a transient transfection assay using an E-box-driven luciferase that is responsive to CLOCK: BMAL1, the DASH-type members in Xenopus laevis and Danio rerio effect no inhibition on the luciferase activity. In contrast, in the same assay, animal-type CRYs from the same organism (DrCRY1a) inhibit the luciferase activity completely (15), suggesting that DASH-type and animal-type CRYs are functionally distinct from each other. Surprisingly, although Neurospora CRY is not a clock component and is dispensable for most light responses and various light-regulated developmental processes, the cry transcript is controlled by the circadian clock with peaks antiphasic to frq. Interestingly, the gene expression of another DASH-type member from tomato (Solanum lycopersicum) has been shown to be under the control of the circadian clock as well (22). The physiological significance of the circadian regulation of DASH-type CRYs, however, remains unclear.
Here, we showed that the transcript and protein levels of Neurospora CRY are both strongly and rapidly induced by light in a wc-1-dependent manner. However, whole-genome microarray analyses of mycelium samples collected under different light conditions indicate that there is no significant difference between WT and cry knockout strains at the transcriptional level. Given this, perhaps the major biological function of Neurospora CRY might be related to posttranslational regulation (54), or perhaps it becomes evident only at specific developmental stages or under particular growth conditions. For instance, in Neurospora the regulatory function of the opsin photoreceptor gene nop-1 could be observed only in late developmental stages (4), and a recent report suggests that CRY can modulate the primary response elicited by WC-1 and WC-2 (53); this would perhaps provide an explanation for the small effects we report on the entrained phase of the conidiation rhythm in strains lacking CRY. That report (53) also provides confirmation of the light induction of cry reported here. Additionally, some supporting evidence comes from the study of cry1 from Sclerotinia sclerotiorum (69), which is the only other fungal DASH-type member characterized so far and has orthologs for all the Neurospora crassa light-sensing components (i.e., WC-1, WC-2, and VVD). The transcript level of Sscry1 is strongly induced by UV-A and appears to follow induction kinetics similar to those of the cry transcript in Neurospora. Knockout strains of Sscry1 exhibit a slight decrease in sclerotial mass and increased numbers of pigmented hyphal projections on apothecial stipes under UV-A treatment but are otherwise developmentally normal. The study concludes that Sscry1 may have a function during UV exposure but is not essential for completing the developmental life cycle under laboratory conditions. Therefore, these principles may hold true for revealing additional biological functions of Neurospora CRY and other DASH-type members in the future.
Similar to most DASH-type cryptochromes, Neurospora CRY has no detectable photolyase activity in vivo and appears to bind FAD and MTHF when expressed in E. coli. In addition to its nonspecific DNA binding activity, which has been shown for other DASH-type members, we discovered that Neurospora CRY also has the capability to bind to single- and double-stranded RNA, as well as to a DNA-RNA hybrid. Although these properties might be shared among DASH-type family members, we speculate that this feature might be unique to Neurospora CRY due to the existence of multiple arginine-glycine-glycine (RGG) repeats at the C terminus, which are missing for all other DASH-type cryptochromes, including the cry1 found in S. sclerotiorum. RGG repeats are commonly found in proteins involved in regulation and coupling of RNA maturation events (29). RGG repeats are generally found at the C terminus of the protein and contribute to RNA binding nonspecifically (6, 28, 38, 52), both features of Neurospora CRY. Given the expression of cry RNA and protein late in the circadian cycle relative to most clock-regulated transcription, CRY may act via regulation of RNAs to fine-tune the phase of entrainment. Conversely, it may be worth reevaluating whether the capability of binding to both DNA and RNA is restricted to Neurospora CRY or widespread among other CRY-DASH members and if this function is tied to clock phenotypes in these other model organisms.
This work was supported by grants from the National Institutes of Health to J.J.L. (RO1 GM08336) and to J.C.D. (GM34985 and PO1GM68087), the NWO (Dutch Science Foundation) VICI and the University of Groningen Rosalind Franklin programs to M.M., and the DFG (German Science Foundation) and the EU (EUCLOCK) to M.M. and T.R. and by the core grant to the Norris Cotton Cancer Center at Dartmouth.
We are deeply grateful to the Fungal Genetics Stock Center at the University of Missouri, Kansas City, in supporting our work with Neurospora.
Published ahead of print on 19 March 2010.