Photoperiodically driven transcriptome-wide changes in the hypothalamus reveal transcriptional differences between physiologically contrasting seasonal life-history states in migratory songbirds

A single long day induces the photoperiodic molecular response

Figure 1c shows results from the experiment 1, as evidenced from the qPCR measurement of mRNA expression of genes of known biological functions in the blood and hypothalamus. Clearly, the exposure to extended light period induced a molecular response by hour 18 of the first long day, as shown by change in mRNA levels of candidate genes in both central (hypothalamus) peripheral (blood) tissues of photosensitive buntings. Blood mRNA levels of peroxiredoxin 4 (prdx4) were significantly lower at hour 18 mimicking a long 18 h photoperiod than those at hour 10 mimicking a short 10 h photoperiod (p = 0.002, t = 5.18, n = 4/time point). Paradoxically, this indicated a reduced cellular response against oxidative stress in the otherwise photo stimulated birds on the first long day. We speculate that prdx4 expression pattern would be inversed (i.e. increased prdx4 mRNA levels) after several long days when birds show photoperiodically stimulated hyperphagia (increased food intake) and lipogenesis (fat accumulation). Intriguingly, however, blood mRNA levels of gpx1 (p = 0.399, t = 0.91, n = 4/time point) and sod1 (p = 0.845, t = 0.20, n = 4/time point) genes were not different between hours 10 and 18 (Student’s t-test, Fig. 1c(a–c)). Taken together differences in the expression pattern of these enzymes, we speculate differential activation of the enzymatic pathways that are probably involved in the oxidative cellular response when migratory birds are exposed to an acute change in their photoperiodic environment.

On the other hand, blood il1β mRNA levels were significantly higher at hour 18 than the hour 10 (p = 0.041, t = 2.58, n = 4/time point; Student’s t-test, Fig. 1c(d)). It is consistent with the known role of il1β-encoded interleukin 1β, as a crucial mediator of the inflammation and a marker of the innate immune system^22,23. Increased il1β mRNA expression on the first long day is consistent with the idea of parallel photoperiodic induction of multiple biological processes, including those associated with the innate immune response, body fattening and gonadal maturation in migratory songbirds²⁸; however, the possibility that an upregulated interleukin was an indicative a stress response cannot be excluded at this time.

Changes in hypothalamic gene expressions further confirm a rapid molecular response to the extended light period when it surpasses the threshold photoperiod, i.e. acts as the stimulatory long day. Reciprocal switching of genes involved in the thyroid hormone responsive pathway at hour 18 particularly evidences this. Hypothalamic mRNA levels of tshβ (p = 0.033, t = 2.75, n = 4/time point) and dio2 (p = 0.0004, t = 7.14, n = 4/time point) genes were higher, and that of dio3 gene expression was lower at hour 18 than the hour 10 (p = 0.036, t = 2.68, n = 4/time point). This is also in agreement with the rapid photoperiodic response found on the first long day in plasma LH secretion, and in hypothalamic expressions of Fos-immunoreactivity and thyroid hormone responsive genes in blackheaded buntings^14,33 and other photoperiodic birds^{15,17,19,32,34,35,36,37,38}. However, gnrh mRNA levels were not found significantly different between hours 10 and 18 of the first long day (p = 0.324, t = 1.07, n = 4/time point; Student’s t-test, Fig. 1c(e–h) indicating that hour 18 was probably too early a time for an upregulated gnrh expression on the first long day^37,38,39.

RNA-Seq reveals differences in time course of the photoperiodic response

Table S2 summarizes the primary statistics used for RNA-Seq results. Using only transcripts with non-zero abundance, we compared the time course of transcriptome-wide response in the hypothalamus both as the function of time (within photosensitive or photorefractory state) and LHS (photosensitive vs. photorefractory state; n = 2/time point/state except at hour 22 in photorefractory state which had n = 1 sample size). Further, to show a functional linkage of differentially expressed genes (DEGs), we performed STRING analysis that predicts the protein–protein interaction (see methods for details).

Results on hypothalamic gene expressions suggest that buntings react to the acute photoperiodic change in photorefractory state almost as they do in the photosensitive state. However, the comparison of the overall RNASeq data from both states revealed LHS-dependent pattern in the time course of transcriptional response, with differences in the number and functions of DEGs and associated physiological pathways.

Within state differences in time course of transcriptional response

We examined the time course of response on the first long day, by comparing gene expressions at the hours 14, 18 and 22 of the extended light period that mimicked 14 h, 18 h and 22 h long photoperiods, respectively, with those at hour 10 that mimicked a 10 h short photoperiod.

Photosensitive state

At hour 14, we found 10 differentially expressed genes (DEGs) with 4 upregulated and 6 downregulated genes (Figs. 2a, 3a, Table S3). Of the 10 DEGs, atp6v1e1, atp6v1b2, uqcrc1 and pgam1 genes enriched the oxidative phosphorylation, metabolic pathways, phagosome and mTOR signalling pathways (Table 1). The oxidative phosphorylation and metabolic pathways were upregulated at hour 10, while the phagosome and mTOR signalling pathways were enriched by two genes that were opposite in the expression trend: atp6v1e1 was upregulated while atp6v1b2 was downregulated at hour 14. The STRING analysis showed a significant interaction of atp6v1e1 and atp6v1b2 encoded proteins (ATP6V1E1 and ATP6V1B2). These proteins are the components of vacuolar ATPase enzyme that mediates the acidification of eukaryotic intracellular organelles necessary for protein sorting and zymogen activation. Further, at hour 14, ttr gene that codes for transthyretin (a preferential T3 binder) and pomc gene that codes for the proopiomelanocortin receptor had significantly lower expressions. Whereas, low ttr gene expression, as in photostimulated redheaded buntings⁴⁰, might indicate a reduced trafficking of thyroid hormones via ttr-encoded transthyretins in the photosensitive state, the low pomc gene expression might suggest the removal of inhibitory effects of the opioids (e.g. β-endorphin, a pomc-encoded proopiomelanocortin product) on hypothalamic GnRH and, in turn, pituitary LH secretion^41,42.

Likewise, there were 8 DEGs with 4 upregulated and 4 downregulated genes at hour 18 (Figs. 2b, 3b, Table S3), of which pcca, eno1 and atp6v1e1 genes enriched the upregulated carbon metabolism and other metabolic pathways (Table 1). Besides, both ttr and pomc genes were downregulated, while apoA-1 gene coding for apolipoprotein-A1 was upregulated in expression at the hour 18. An increased apoa1 gene expression probably indicated the activation of the cholesterol metabolism pathway, as apoa1-encoded apolipoprotein-A1 is a major cholesterol transporter protein in both central and peripheral tissues, and regulates energy homeostasis⁴³. Increased brain apoa1 gene expression has been linked also with the photostimulated development migratory phenotype in migratory northern wheatears, Oenanthe oenanthe⁴⁴.

At hour 22, we found 23 DEGs with 6 upregulated and 17 downregulated genes (Figs. 2c, 3c, Table S3), although these did not enrich a functional pathway. Downregulated cga and pomc genes seem to be the part of a protein network, as revealed by STRING analysis showing the network interaction of cga and pomc encoded proteins (CGA and POMC). It may be recalled that cga is alpha subunit of tsh gene (tshα) couples with its beta subunit (tshβ) and forms the thyroid stimulating hormone in pars tuberalis, which is part of the local thyroid hormone pathway that mediates the induction of a photoperiodic response^19,20,32.

The atp6v1e1 and pomc genes were differentially expressed at hours 14, 18 and 22 during the extended 22 h of the light period (Fig. 2d, Table S4). Whereas, as compared to their expressions at hour 10 mimicking a short day, atp6v1e1 gene was upregulated, the pomc gene was downregulated in expression at all these three time points although they all mimicked a stimulatory long day (Figs. 2d; 3a–c).

Photorefractory state

There were transcriptional changes across the second half of the day, as indicated by the comparison of gene expression patterns at hours 14, 18 and 22 (long day), with that at hour 10 (short day). At hour 14, for example, we found 10 DEGs with 5 each having upregulated and downregulated expressions (Figs. 2e, 3d, Table S3). Differentially expressed eif4g2 and eef1a1 genes significantly enriched the “RNA transport” pathway, suggesting a translational effect of the light exposure at hour 14. Consistent with an upregulated translation activity at hour 14, the expressions of eif4g2 and eef1a1 genes were downregulated and upregulated, respectively (Fig. 3d, Table 1). Whereas eef1a1 is a translational elongation factor gene, eif4g2 gene acts as a general repressor of the translation; hence both these genes possibly interact and regulate translational processes following the photoperiodic induction of a response.

Likewise, there were a total of 27 DEGs with 11 upregulated and 16 downregulated genes at hour 18, Figs. 2f, 3e, Table S3), of which 8 DEGs enriched 4 functional pathways, namely the glycolysis/gluconeogenesis, metabolism, biosynthesis of amino acids and RNA degradation pathways (Fig. 3e, Table 1). In particular, we found an upregulated expression of genes associated with glycolysis/gluconeogenesis and metabolic pathways. Further, the upregulated eno1 gene and downregulated pfkl gene enriched the amino acids biosynthesis and RNA degradation pathways. Differentially expressed npas2 gene could be indicative of differences in in the circadian timing mechanism, although it remains purely speculative in the absence of differential expression of other circadian genes in the current study.

Similarly, at hour 22, we found 39 DEGs with 19 upregulated and 20 downregulated genes (Figs. 2g, 3f, Table S3). Here, 7 DEGs enriched six pathways of which genes including the biosynthesis of amino acids, pentose phosphate pathway, fructose and mannose metabolism, and glycolysis/gluconeogenesis were highly expressed at hour 22. At the same time, genes that enriched the metabolic pathway were lower in expression at hour 22 than the hour 10. The STRING analysis showed an interaction of pfkl and aldoa encoded glycolytic enzyme proteins (Fig. 3f).

The eef1a1 gene was also found to have an upregulated expression at all three timepoints (hours 14, 18 and 22), compared to the hour 10 of the first long day (Fig. 2h, Table S4). This though suggests daily changes in the translational processes in photorefractory birds, although these were probably were not as robust as found in photosensitive birds in which there was also concurrent downregulated expression of the eif4g2 gene, a general repressor of the translation.

LHS-dependent time course of transcriptional response

Differences in the transcriptome-wide response between physiologically contrasting photosensitive and photorefractory states suggest that the time course of transcriptional activation of molecular processes differed between seasonal LHSs comprising annual cycle of migratory buntings. This is shown by differential gene expressions between two states at four times of the day that we have examined and compared in the current study. To begin with, at hour 10, we found 9 DEGs with 4 upregulated and 5 downregulated genes in photorefractory, compared to the photosensitive state (Figs. 2i, 4a; Table S3). Five DEGs enriched three functional pathways (valine, leucine and isoleucine degradation, metabolic, and phagosome pathways) that were upregulated in the photosensitive state (Fig. 4a, Table 1). The amino acids valine, leucine and isoleucine are involved in the biosynthesis of glutamate, which as a major excitatory neurotransmitter is involved in transmission of the photoperiodic information, and serves as a precursor molecule for the biosynthesis of largely inhibitory gamma aminobutyric acid (GABA) neurotransmitter⁴⁵. The upregulated expression of genes associated with these pathways suggested an enhanced neural activity in photosensitive birds in response to their exposure to a stimulatory long light period. This is consistent with evidences for increased neural activity and neurogenesis in the hippocampus and nidopallium caudolaterale brain regions of photostimulated migratory white crowned sparrows, Zonotrichia leucophrys⁴⁶ and reed warblers, Acrocephalus scirpaceus⁴⁷). An increased neuronal activity, as shown by Fos-like immunoreactivity, has also been found in the mediobasal hypothalamus brain region of Japanese quails^17,34 and blackheaded buntings³³ in response to the first long day.

At hour 14, we found 5 DEGs with 2 upregulated and 3 downregulated genes in photorefractory, compared to the photosensitive state, although these DEGs did not significantly enrich a particular functional pathway (Figs. 2j, 4b, Table S3). Intriguingly, we also found a significantly upregulated expression of orexin precursor hcrt gene in the photorefractory state, which probably indicates an association of hcrt expression with LHS-dependent differences in sleep–wake pattern in buntings as they were exposed for many weeks to the 16 h long photoperiod. It is reported that hcrt-encoded hypocretin produced by lateral hypothalamus is essential for the arousal stability⁴⁸, and activated hypocretin-producing neurons by direct photostimulation have been found associated with transition from sleep to awake state in mice⁴⁹.

Similarly, at hour 18, 33 DEGs with 25 upregulated and 8 downregulated genes in photorefractory, compared to the photosensitive state, did not enrich a functional pathway (Figs. 2k, 4c, Table S3). Particularly, we found an upregulated ttr, and downregulated cart (coding for cocaine- and amphetamine-regulated transcript) and grp (coding for gastrin releasing peptide) expressions in photosensitive, compared to photorefractory birds, consistent with photoperiodic initiation of physiological processes. Whereas increased ttr gene expression might indicate enhanced activity of the photostimulated thyroid hormone responsive pathway, the decreased cart and grp gene expressions probably suggest the activation of processes associated with the hyperphagia and consequently the body fattening and weight gain in photosensitive birds exposed to stimulatory long days^50,51,52. Further, there seems to be changes in genes associated with metabolic processes, as suggested by downregulated expression of coa6 and uqcrc1 genes in the photosensitive state. These genes are part of the respiratory chain complexes, and seem to interact as suggested from significant interaction of proteins that they encode for (Fig. 4c; STRING analysis).

At hour 22, we found 15 DEGs with 7 upregulated and 8 downregulated genes in photorefractory, as compared to the photosensitive state (Figs. 2l, 4d, Table S3). Of these DEGs, rpl12 and rpl37 genes that significantly enrich the ribosome pathway (Fig. 4d, Table 1) were upregulated in the photosensitive, as compared to expression in the photorefractory state (Table 1). This suggests the initiation of translational processes later in the night (hour 22) when birds were exposed to light. At this time, an upregulated sod1 expression also suggests the activation of oxidative stress pathways pathway in photosensitive birds, as has been reported recently in photostimulated redheaded buntings²⁸. We did not find a common gene that was differentially expressed at all the four time points between the photosensitive and photorefractory states (Fig. 2m; Table S4).

The exposure to a single long day activated a series of hypothalamic molecular pathways that are crucial for the initiation and maintenance of key biological processes during a seasonal LHS. We interpret that within and between state differences in the time course of transcriptional responses were due to differential light-sensitivity of the underlying circadian photoperiod-responsive rhythm, which is shown to be involved in regulation of the lipogenesis (body fattening and weight gain) and gametogenesis (gonadal maturation) in migratory blackheaded buntings and other birds in response to the long photoperiod^1,3,38,53. Clearly, the photoperiodic induction depends on the extension of the light into the photoinducible phase lying in the second half (about 12 h after the light onset) of the endogenous circadian photoperiodic rhythm^1,12,54. Present experiments also evidence this, by a much smaller transcriptional response at hour 10. There can also be LHS-dependent alteration in the 24-h waveform of the circadian rhythm governing the photoperiod-induced response in blackheaded buntings^31,55. Differential responsiveness of the circadian photoperiodic rhythm seems to be the part of the overall adaptive strategy in avian migrants, which need to begin and end their spring migration to restrict their reproduction during most profitable time of the year. This, we believe, is achieved by the interaction of circadian rhythm and photoperiodism, which are mutually inclusive physiological processes, in a long day species like the migratory blackheaded bunting.

Present results should be viewed with a caution, however. This is because we have not shown changes in known marker genes of the photoperiodic response (e.g. tshβ, dio2 and dio3) in the present RNA-seq study. This is intriguing, but we speculate that such an unexpected discrepancy was because of two possibly. First, the quality filtration used for hypothalamus RNA-Seq led to the loss from the list of data on ‘photoperiodic genes’, which are expressed in low amounts and site-specific manner^19,20. Secondly, in the absence of full gene sequence of our study species (Emberiza melanocephala; family – Emberizidae), we used the reference genome of migratory Ficedula albicollis (family—Muscicapidae), and this probably has filtered out many functionally relevant genes from the annotation of present transcriptome. Despite these limitations of the current study, the striking differences in time course of the transcriptome-wide response both within the state and between states provide a strong evidence for LHS-dependence of the photoperiodically driven activation of hypothalamic molecular pathways during the annual cycle in migratory songbirds.

Source: Ecology - nature.com