Involvement of the Clock Gene period in the Circadian Rhythm of the Silkmoth Bombyx mori

The strain pnd w-1 (pnd, pigmented and non-diapausing egg; w-1, white egg 1), a standard strain for transgenesis of B. mori (Tamura et al. 2000), was used as the wild type strain in the present study. Larvae were reared on an artificial diet (Kuwano-hana, JA Zennoh Gunma, Maebashi, Japan) under a LD 12:12 h cycle at 25.0°C [1.0°C]. The light intensity in the photophase produced by a daylight fluorescent lamp was 870 to 1450 mW/m². Knockout B.mori were generated using TALENs, as described previously (Takasu et al., 2013; Suppl. Fig. S1). The left and right TALEN binding sites are 5’-CATCGCTGACGGCTAC-3’ and 5’-CCCGATCCACATGTCCT-3’, respectively. TALEN mRNAs (400 ng; 200 ng + 200 ng for left and right TALEN mRNA, respectively) were injected into preblastoderm embryos. Established per knockout lines were not outcrossed to other standard strains so that we could compare the phenotypes and gene expression levels in the same genetic background, i.e., pnd w-1.

For genotyping, the heads of moths or whole bodies of first-instar larvae were crushed in alkaline solution (50 mM NaOH) and then heated at 95°C for 10 to 15 min. After neutralization by the addition of an equal volume of 0.2 M Tris-HCl (pH 8.0), the supernatants were used as templates for PCR. PCR was performed using ExTaq (TaKaRa, Kusatsu, Japan) or Paq5000 (Agilent Technologies, Tokyo, Japan). Primers used for genotyping were per-F (5’-ATGACTGCATGACGGCAACT-3’) and per-R (5’-CTCTTCGACA AAGGATACGTAGC-3’). These primers were designed to hybridize to sequences upstream and downstream of the TALEN target site.

To examine the hatching rhythm, each egg mass was divided into batches of approximately 20 eggs and kept under LD conditions until recording. To examine the eclosion rhythm, pupae were removed from cocoons 6 days after pupation and kept under LD conditions until recording. Hatching and eclosion rhythms were examined under LD, constant light (LL) or constant darkness (DD) at 25.0°C [1.0°C]. Recording commenced at lights-on (LL) or lights-off (LD and DD) before the predicted time of hatching. Hatching or eclosion was recorded using a digital camera (EX-ZR3100, CASIO, Tokyo, Japan or D5100, Nikon, Tokyo, Japan) at 30-min intervals. In darkness, red LED light (Kaito Denshi, Hasuda, Japan, 660 nm) was used as the light source for recording. The LED was wrapped with white tape to reduce the light intensity to approximately 0.3 mW/m².

The degree of rhythmicity in eclosion was measured by the parameter R (Winfree, 1970). Eclosion data from several days were pooled to calculate the total number of eclosions for each hour of the day. The 8-h period (gate) of the day with the highest number of eclosions was then determined. The parameter R was calculated by dividing the number of eclosions outside this 8-h period by the number of eclosions within it and multiplying by 100. The theoretical range of R is from 0, if all moths emerge within the gate, to 200, if eclosion is distributed uniformly throughout the day. R values of 150 or greater show statistically uniform eclosion (Winfree, 1970). R values of 60 or less represent rhythmic eclosion, those between 60 and 90 represent weakly rhythmic eclosion, and those greater than 90 represent arrhythmic eclosion (e.g., Saunders, 1979; Smith, 1985; Watari, 2005).

The free-running period was defined as the mean interval between the medians of eclosion/hatching peaks. To examine the arrhythmicity in the knockout strain within the range of the free-running period in the wild type under the same conditions ±4 h (test range), the Rayleigh test was used with the “circular” package (Version 1.1.3) in R (Version 3.3.0) (Lund and Agostinelli, 2013; R Development Core Team 2013). (This mode of analysis was selected because the present data of eclosion and hatching were not quantitatively sufficient for a chi-square periodogram.) Before the Rayleigh test, time series data were converted to angle data using a test period chosen from the test range. Accepting the null hypothesis of uniformity (p > 0.05) indicated arrhythmicity during the chosen test period. The test period was chosen in turn from the test range, and the Rayleigh test was repeatedly carried out.

Quantitative real-time PCR (qPCR) was used to investigate the temporal expression patterns of per and tim mRNAs. Eggs were incubated under LD conditions until sampling. Sampling was started at lights-off one day before hatching, when embryos were at the body pigmentation stage. Sampling was performed every 4 h. Manipulation of the eggs under darkness was conducted under dim red light. Total RNA from an egg was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA, USA), treated with DNase I (Invitrogen), and then used to synthesize cDNA with High-Capacity cDNA Reverse Transcription Kits (Applied Biosystems, Foster City, CA, USA). Quantitative PCR assays were performed using SYBR Green I (Roche, Basel, Switzerland) and a Light Cycler 96 system (Roche, Basel, Switzerland). The reaction conditions were as follows: 95°C preinitiation heating for 1 min, followed by 45 cycles of 95°C for 10 s, 58°C for 10 s and 72°C for 10 s. After PCR, a melting curve analysis was used to confirm the amplification of specific products. The BmRp49 gene was used as a reference. The standard curve method was used to determine the expression levels of samples. The primers used for qPCR were the same as those used previously (Iwai et al. 2006): per-F (GTAATGCTCGGCGGGATATC) and per-R (AGCGGTGTTTCTGTGCTTGA) for per, tim-F (CTTGAAGTTTCTGCAGCTGTTACAC) and tim-R (AAACACTTTCCCGGAATCGA) for tim, and Rp49-F (ATGACGGGTCTCTTTGTTGGAA) and Rp49-R (CAGGCGATTCAAGGGTCAAT) for rp49.

Two-way ANOVA was used to examine statistically significant differences in gene expression between the 2 strains, whereas one-way ANOVA was used to examine statistically significant temporal differences in each strain.

To establish a per knockout line, we used a TALEN-mediated gene-targeting approach, because TALENs have been shown to induce highly efficient targeted gene disruption in B. mori (Daimon et al., 2014; Takasu et al., 2013). We designed a target site in exon 7 of the per gene, an exon that encodes a PAS domain, which is known to interact with other proteins (Fig. 1). We injected TALEN mRNA into early embryos of the pnd w-1 strain. The hatched larvae were reared, and adults were crossed with the parental strain (pnd w-1). In the next generation (G1), we screened for the induced mutations using the CEL-I assay (Daimon et al., 2014) and identified adults carrying mutant alleles of per. From the recovered per mutant adults with various mutant alleles, we selected an adult with a large (64-bp) deletion within exon 7 that produced a premature stop codon downstream of the target site (Fig. 1). This allele was considered a null allele, as it encoded an extensively truncated form of PER (198 aa in the protein from the 64-bp deletion allele; 1113 aa in the wild-type protein). This mutant adult was crossed to an adult of the parental strain, and the 64-bp deletion allele was genetically fixed in the next generation (G2). The G2 adults were crossed to siblings and, in the next generation, we established a per knockout line. As per is located on the Z chromosome (ZW for females and ZZ for males in B. mori), females are hemizygous and males are homozygous for the per knockout allele. In the present study, we used this knockout strain for further experiments and named it per^Δ64. In our rearing conditions and under the pnd w-1 genetic background, viability and fertility of per^Δ64 appeared to be normal without any apparent external phenotype.

Under LD conditions, eclosion was observed in the middle of scotophase in the wild type, whereas in per^Δ64, a steep peak of eclosion occurred 1 h after lights-on, with eclosion also observed in scotophase (Fig. 2A, B). Under DD conditions, eclosion of the wild type showed a free-running rhythm with a period of approximately 23.5 h (Fig. 2C). The parameter R was 4.76 and the Rayleigh tests showed p < 0.01 at any chosen test period from the test range, indicating strong rhythmicity. In per^Δ64, however, eclosion was continuously observed from 17 h after lights-off (Fig. 2D), and the parameter R of 167 and p > 0.1, as shown by Rayleigh tests at any chosen test period from the test range, indicated arrhythmicity. Under LL conditions also, eclosion of the wild type showed a free-running rhythm but the period was approximately 16 h, which was much shorter than that under DD conditions (Fig. 2E). In per^Δ64 under LL conditions, eclosion was observed immediately after lights-on and continuously from 20 h after lights-on (Fig. 2F). In the Rayleigh test, the immediate peak after lights-on was excluded because this peak was considered a direct response to light. The parameter R of 153 and p > 0.4 shown by the Rayleigh tests at any chosen test period from the test range indicated arrhythmicity. Therefore, we concluded that per^Δ64 has lost circadian rhythm in eclosion.

Wild-type eggs hatched mostly around the time of lights-on, and the hatching peak occurred 1 h after lights-on under LD conditions (Fig. 3A). In per^Δ64, however, hatching was observed even at other times of day, even though a steep peak was observed at 1 h after lights-on (Fig. 3B). Hatching of the wild type showed a free-running rhythm with a period of approximately 24.5 h under DD conditions. The parameter R was 10.78 and the Rayleigh tests showed p < 0.01 at any chosen test period during the test range (Fig. 3C), indicating strong rhythmicity, as seen in eclosion. Hatching of per^Δ64 was observed continuously from 2 h after lights-off (Fig. 3D), and the parameter R of 123 and p > 0.03, as shown by Rayleigh tests at any chosen test period during the test range, indicated arrhythmicity. Under LL conditions also, hatching of the wild type was observed as 3 groups, with median times of 1 h, 23 h, and 37 h after the start of the experiment. Although the third hatching group consisted of only a small number of larvae and its peak is unclear, the free-running period seemed to shorten gradually. The time intervals between the hatching groups were much shorter than those under DD conditions (Fig. 3E). Hatching of per^Δ64 was observed continuously from 6.5 h after the start of the observation (Fig. 3F). The Rayleigh tests indicated rejection of the null hypothesis of uniformity (p < 0.01 during 16-22 h), probably because hatching was restricted mostly to the first half of the observation period. However, the parameter R of 94 indicated arrhythmicity, and apparently hatching occurred sporadically in the first half of the observation period. Therefore, we concluded that per^Δ64 has lost circadian rhythm in hatching.

To examine the effect of per knockout on the molecular clockwork, we performed qPCR of 2 clock genes, per and tim, in embryos at the body pigmentation stage. In the wild type, per expression showed temporal changes (p < 0.05, one-way ANOVA), whereas there was no significant difference in per expression over time in per^Δ64 (p > 0.05) (Fig. 4A). The results of the 2-way ANOVA showed that per expression was significantly lower in per^Δ64 than in the wild type (p <0.01) (Fig. 4A). In the wild type, tim expression also changed temporally (p < 0.05, one-way ANOVA), but did not significantly change in per^Δ64 (p > 0.05) (Fig. 4B). In per^Δ64, tim expression was significantly higher than in the wild type (p <0.05) (Fig. 4B).