Peripheral gene expression rhythms in a diurnal rodent.

While most laboratory rodents are active at night, many other species exhibit higher levels of activity during the daytime. The circadian system of diurnal species is not simply inverted relative to that of nocturnal species (Smale et al., 2003). While most rhythms peak at opposite phases in diurnal and nocturnal species, rhythmic melatonin production peaks at night in all species. Furthermore, studies comparing rhythmicity in the brains of diurnal and nocturnal species reveal that the phase of Per and prokineticin 2 gene expression rhythms in the suprachiasmatic nucleus (SCN) are similar among nocturnal and diurnal species (Mrosovsky et al., 2001; Caldelas et al., 2003; Lambert et al., 2005). Several differences in Fos expression downstream of the SCN have been identified, however (for review, see Smale et al., 2003). Collectively, these studies suggest that multiple switches exist downstream of the SCN along output pathways controlling specific behavioral and physiological rhythms (see Mrosovsky, 2003; Smale et al., 2003). SCN-driven rhythms in locomotor behavior, ingestive behavior, body temperature, corticosteroid, and melatonin secretion participate in entrainment of peripheral oscillators (Schibler et al., 2003; von Gall et al., 2002). As the phase of most of these rhythms differs between nocturnal and diurnal species, one would expect diurnal species to have altered phase of peripheral rhythmicity, relative to nocturnal species. We examined Per1 and Per2 RNA levels in the liver (Fig. 1) and testes (Fig. 2) in a diurnal rodent, the unstriped Nile grass rat (Arvicanthis niloticus), to test this prediction. Per1 and Per2 mRNAs were rhythmically expressed in grass rat liver, with peak levels occurring early in the subjective day, at circadian time (CT) 2, and trough values occurring 12 h later at CT 14 (Fig. 1). This phase is opposite to the phase previously reported for Per gene expression rhythms in mouse and hamster liver (Zylka et al., 1998; Tong et al., 2004) (Fig. 3B). In grass rats, Per gene expression reaches peak levels approximately 4 h earlier in the liver than in the SCN, while in mice, Per gene expression in the liver reaches peak levels 3 to 9 h after peak levels are reached in the SCN (Fig. 3). Testicular Per gene expression is of interest because there appears to be a species difference in the regulation of Per1 in this tissue. In murine testes, Per1 levels are elevated and not rhythmic in most studies (Alvarez et al., 2003; Bittman et al., 2003; Morse et al., 2003; but see Zylka et al., 1998), while in hamsters, testicular Per1 expression has a low-amplitude rhythm (Tong et al., 2004). In grass rat testes, we found that Per1 RNA was expressed at high levels and was not rhythmic, while Per2 was undetectable (Fig. 2 and data not shown). Note that the same blots that gave strong Per1 and actin hybridization signals were used with the Per2 probe, and the same Per2 probe gave a good signal on liver blots processed in parallel, making it unlikely that failure to detect Per2 in testes is due to a technical problem with the probes or blots. Peripheral oscillators are synchronized by the SCN through a variety of mechanisms. Most non-SCN oscillators likely are reversed in their phase of oscillation between nocturnal and diurnal species, as demonstrated here for the liver. Within the brain, however, areas that are directly along outflow pathways from the SCN may have a phase of Per oscillation that is similar between nocturnal and diurnal species. Thus, examination of species differences in Per mRNA and PER protein rhythms may help to identify brain sites important for diurnality. Indeed, Smale, Nunez, and colleagues have used FOS immunoreactivity in a similar manner (for review, see Smale et al., 2003). The high amplitude of Per rhythmicity and its widespread expression may allow Per mapping to serve as a useful complement to work using FOS and other markers. Mrosovsky et al. (2001) reported that Per1 gene expression in the motor cortex of diurnal ground squirrels peaks early in subjective day, preceding the SCN peak. Interpretation of this finding is complicated because gene expression in the motor cortex could be directly related to the locomotor activity of the animal rather than representing the oscillator phase. When considered along with our results, however,