Adaptive changes in BMAL2 with increased locomotion associated with the evolution of unihemispheric slow-wave sleep in mammals

Abstract

Marine mammals, especially cetaceans, have evolved a very special form of sleep characterized by unihemispheric slow-wave sleep (USWS) and a negligible amount or complete absence of rapid-eye-movement sleep; however, the underlying genetic mechanisms remain unclear. Here, we detected unique, significant selection signatures in basic helix-loop-helix ARNT like 2 (BMAL2; also called ARNTL2), a key circadian regulator, in marine mammal lineages, and identified two nonsynonymous amino acid substitutions (K204E and K346Q) in the important PER-ARNT-SIM domain of cetacean BMAL2 via sequence comparison with other mammals. In vitro assays revealed that these cetacean-specific mutations specifically enhanced the response to E-box-like enhancer and consequently promoted the transcriptional activation of PER2, which is closely linked to sleep regulation. The increased PER2 expression, which was further confirmed both in vitro and in vivo, is beneficial for allowing cetaceans to maintain continuous movement and alertness during sleep. Concordantly, the locomotor activities of zebrafish overexpressing the cetacean-specific mutant bmal2 were significantly higher than the zebrafish overexpressing the wild-type gene. Subsequently, transcriptome analyses revealed that cetacean-specific mutations caused the upregulation of arousal-related genes and the downregulation of several sleep-promoting genes, which is consistent with the need to maintain hemispheric arousal during USWS. Our findings suggest a potential close relationship between adaptive changes in BMAL2 and the remarkable adaptation of USWS and may provide novel insights into the genetic basis of the evolution of animal sleep.

Graphical Abstract

Statement of Significance.

Sleep is an indispensable part of animal life; however, the molecular mechanisms underlying different sleep features among animals remain largely unknown. Marine mammals, especially cetaceans, have been considered as an exciting opportunity to understand sleep evolution, because they have evolved a very special form of sleep characterized by unihemispheric slow-wave sleep (USWS), which is remarkably different from bihemispheric slow-wave sleep observed in all terrestrial mammals with both brain hemispheres sleeping simultaneously. Here, we described the unique and significant evolutionary signatures of the basic helix-loop-helix ARNT-like 2 (BMAL2) gene in marine mammal lineages, and discovered two unique amino acid substitutions (K204E and K346Q) in cetaceans. Furthermore, cetacean-specific mutations are shown to cause adaptive modifications in BMAL2, which specifically enhance the PER2 expression and the related wakefulness-promoting effect to help cetaceans to depress whole-brain sleep demand and maintain continuous alertness during USWS. Our result is the first to link mutations in key sleep-regulated genes to mammalian sleep adaptations, and shed novel insights into the understanding of the genetic basis underlying sleep evolution.

Introduction

Sleep, a fundamentally important biological process, is conserved from jellyfish to humans, and is essential for optimal organism function [1]. Typically, mammalian sleep consists of two highly different stages: rapid eye movement (REM) sleep and slow-wave sleep (SWS, also called NREM sleep) [2]. REM sleep is characterized by muscle atonia, episodic bursts of rapid eye movement, and a high level of brain activity. SWS, by contrast, is characterized by the maintenance of muscle tone, electrocortical synchronization, and reduced neuronal activity [3]. Notably, all terrestrial mammals have relatively high-voltage and slow neocortical activity bilaterally during SWS (i.e. bihemispheric SWS); however, cetaceans (whales, dolphins, and porpoises), sirenians (dugongs and manatees), and non-phocid pinnipeds exhibit a unique sleep pattern known as unihemispheric SWS (USWS), in which one cerebral hemisphere exhibits SWS while the other remains awake [4, 5]. In addition, it is reported that cetaceans have a negligible amount or complete absence of REM sleep [6]. This unusual form of sleep allows cetaceans to benefit from sleep while accommodating their need for motion to come to the surface for breathing and enables more efficient environmental monitoring, which helps alleviate the contradiction between sleep and drowning and confers better adaptation to aquatic habitats [4, 7]. Although cetacean sleep has been considered one of the most interesting topics in understanding the evolution of sleep [8], the underlying molecular mechanisms are still enigmatic.

There is a wealth of evidence suggesting that mammalian sleep is modulated by the circadian rhythm system [9, 10]. At the molecular level, the circadian system is based on autoregulatory transcription- and translation-based negative feedback loops [11]. Briefly, four members of the basic helix-loop-helix (bHLH) proteins containing two PER-ARNT-SIM (PAS) domains (PAS A and B)—namely, CLOCK/NAPS2 and BMAL1/BMAL2 (also known as ARNTL1/ ARNTL2)—act as the positive regulator by heterodimerizing and initiating the transcription of circadian genes through binding to the E-box or E-box-like regulatory elements in their promoters. The period (PER1, PER2, and PER3) and cryptochrome (CRY1 and CRY2) proteins form multimeric complexes that translocate to the nucleus to inhibit their own transcription by directly interacting with the heterodimers of the positive regulators, thereby closing the circadian feedback loop. The deletion or mutation of these circadian genes in mammals has been shown to cause rhythm disturbances and sleep abnormalities [12]. For example, mice lacking BMAL1 exhibit arrhythmic behavior, increased total sleep time, and an attenuated compensatory response to sleep deprivation [13, 14]. Genetic knockout studies subsequently revealed that the disruption of Bmal1 in mice leads to reduced Bmal2 gene expression and that continuously expressed Bmal2 can rescue aberrant circadian phenotypes [15]. BMAL2 was found to be a much more efficient transactivating partner of CLOCK compared to BMAL1 [16]. These findings suggest that the role of BMAL2 in circadian regulation is much more important than previous assumptions.

Genetic differences in the circadian system are suggested to be associated with the different features of sleep and rhythms found in mammalian species. Indeed, evidence of positive selection in CRY1 and PER3 in several subterranean rodents indicated adaptations to dark habitats [17]. A comparative analysis of the giraffe’s (Giraffa camelopardalis rothschildi) genome revealed adaptive modifications of the circadian network, including the rapid evolution of PER1 and a premature stop mutation in PER2; these may contribute to the short and fragmented sleep patterns of the giraffe [18]. Delving into the underlying circadian basis of special sleep traits in animals, such as cetaceans, provides an exciting opportunity to understand sleep mechanisms. In the present study, we focused on the evolutionary characteristics of the BMAL2 gene across multiple mammalian lineages. Notably, the cetacean BMAL2 gene was found to contain two specific amino acid substitutions that are absent in any other mammals. In vitro functional experiments further determined the functional significance of these cetacean-specific mutations. Finally, we used zebrafish as model organisms to transiently overexpress the cetacean-specific mutant BMAL2 gene to evaluate phenotypic and behavioral consequences and to perform transcriptome analysis. This model has several advantages, including its simplicity, highly conserved sleep mechanisms, and amenability to genetic manipulation and for high-throughput behavioral assay [19]. As a newly emerged model for sleep study, zebrafish exhibit established, behaviorally defined diurnal sleep featuring circadian-regulated periods of reversible immobility associated with an increased arousal threshold [20] as well as increased quietness following rest deprivation [21].

Materials and Methods

Alignment and phylogenetic analysis

The nucleotide coding sequences (CDS) of the BMAL2 gene were collected from the National Center for Biotechnology Information (https://www.ncbi.nlm.nih.gov/) for a selected set of species representative of all major mammalian lineages (Supplementary Table S1). In addition, the BMAL2 sequences of the bowhead whale (Balaena mysticetus) and zebrafish (Danio rerio) were separately downloaded from http://www.bowhead-whale.org/ and Ensembl (https://www.ensembl.org/). Sequences were aligned using MEGA7 with the MUSCLE [22] plugin, and minor adjustments were made by eye. The maximum likelihood tree of BMAL2 was inferred using a supermatrix approach by IQ-TREE [23]. Bootstrap analyses were conducted with 1000 pseudoreplicates. All other options were set to their default values in IQ-TREE.

Molecular evolution analysis

Signatures of selection were first examined based on the nonsynonymous (d_N)/synonymous (d_S) substitution ratios (ω = d_N/d_S) using six methods implemented in the HyPhy package through the Datamonkey webserver (http://www.datamonkey.org/) as previously described [24], including branch-site unrestricted statistical test for episodic diversification (BUSTED), mixed effects model of evolution (MEME), adaptive branch-site random effects likelihood (aBSREL), fixed effects likelihood (FEL), FUBAR, and Contrast-FEL (fixed effects site-level model). For these ML methods, we considered tests with a p-value < 0.1 as statistically significant, whereas for the Bayesian approach implemented in FUBAR, we considered a posterior probability larger than 0.90 (corresponding to a Bayes Factor > 9.0). Selective pressures were then assessed using the branch-site model using the CODEML program in PAML 4.7 [25]. The well-accepted mammalian phylogeny obtained from multiple references was used as an input tree in PAML analysis (Figure 1). The likelihood ratio test with an χ² distribution was applied to evaluate which models were statistically different from the null model at a threshold of p-value < 0.05. Bayes Empirical Bayes (BEB) analysis was used to identify positively selected sites with posterior probabilities ≥ 0.80 [25].

Figure 1.

Phylogenetic tree of mammals. The topology of the tree is based on previous studies (Yu et al., 2021). Shaded boxes (A: non-phocid pinnipeds, B: cetaceans, C: sirenians) represent USWS-specific taxa, while branches D–F represent sister taxa.

Plasmid construction and mutations

The full-length cDNAs of mouse Bmal2 and Clock were amplified using PCR and cloned into pcDNA3.1(+) with HindIII and XhoI sites, respectively. These two proteins were expressed without any tag (mBMAL2 and mCLOCK) for transcription activation assays using a previously described method [26]. The mouse Per1 promoter region and mouse Per2 promoter region were isolated as described previously [27, 28] and cloned into the pGL4.17 luciferase reporter vector. For fluorescence microscopy, mouse Bmal2 cDNA was PCR-amplified and subcloned in a pCS2-mCherry vector with BamHI and EcoRI sites to express fusion proteins with a C-terminal mCherry red fluorescent protein tag (mCherry-mBMAL2). For FLAG-mBMAL2, an oligonucleotide encoding the FLAG epitope sequence was fused to the 5ʹ-end of the mouse Bmal2 for western blotting analysis. Zebrafish bmal2 cDNA was PCR-amplified and cloned into pCS2-mCherry with BamHI and EcoRI sites (zBmal2). Mutated plasmids were generated using QuickMutation™Plus Site-Directed Mutagenesis Kit (Beyotime Biotechnology). The constructs were verified using Sanger sequencing. All primers are listed in Supplementary Table S8.

Cell culture and transfections

HEK293T and NIH 3T3 cells were cultured and passaged under 5% CO₂ in DMEM (HyClone) containing 1.8 mg/mL NaHCO₃, 4.5 mg/mL glucose, 100 U/mL penicillin, 100 mg/mL streptomycin, and 10% fetal bovine serum (SenBeiJia Biological Technology Co., Ltd). NIH 3T3 cells that contain self-sustained circadian oscillators and express circadian proteins [29] were used to estimate the effect on the expression of circadian genes. HEK293T and NIH 3T3 cells were transiently transfected using LipoD293^TM DNA In Vitro Transfection Reagent (Signagene) in accordance with the manufacturer’s protocols.

Luciferase assays

Dual-luciferase assays were performed in a 24-well plate by transiently transfecting HEK293T with a total of 525 ng of DNA (350 ng of luciferase reporter plasmids, 62.5 ng of mCLOCK, 50 ng of Renilla internal control pRL-SV40, and 62.5 ng of test plasmids) and 1.5 μL of LipoD293^TM. At 48 hours after transfection, the cells were prepared using the Dual-Luciferase Reporter Assay Kit (Vazyme Biotech Co., Ltd.) and read by a Synergy H1 Multimode microplate reader (BioTek). Light output from the transcriptional activity was divided by the output from Renilla luciferase activity to normalize the samples.

Degradation experiment and western blotting

To determine protein stability, the FLag-mBMAL2 and FLag-mBMAL2-mut plasmids were transfected into HEK293T cells. After 36 hours, the cells were treated with 100 μg/mL CHX (MedChemExpress). Collection of the transfected cells was performed from 0 to 6 hours after CHX treatment. RIPA buffer (Beyotime Biotechnology) containing 1 mM phenylmethanesulfonyl fluoride (PMSF) was used to make whole-cell extracts.

The protein samples were subjected to denaturing 8% or 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred onto polyvinylidene difluoride (PVDF) membranes. For the block buffer, 5% nonfat dry milk in Tris-buffered saline containing 0.05% Tween 20 was used. Anti-FLAG (Sigma, 1:1000) and anti-PER2 (Abcam, 1:1000) antibodies were used to detect mBMAL2 (or mBMAL2-mut), and PER2, respectively.

Subcellular localization assay

HEK293T cells were transfected with mCherry-mBMAL2 or mCherry-mBMAL2-mut plasmids. At 48 hours after transfection, the cells were stained with DAPI staining solution (Yeasen Biotechnology Co., Ltd., Shanghai) and fixed with 4% paraformaldehyde in PBS. Samples were observed using a Zeiss Imager A2-M2 microscope (Carl Zeiss, Göttingen), and the data were analyzed using ImageJ software.

RNA isolation and real-time PCR

Total RNA was prepared using an RNA isolater (Vazyme Biotech Co., Ltd.) in accordance with the manufacturer’s protocol. The total RNA was reverse transcribed using the HiScript III 1st Strand cDNA Synthesis Kit (+gDNA wiper) (Vazyme Biotech Co., Ltd.). The reverse transcripts were subjected to real-time PCR (Roche LightCycler^® 480) using SYBR qPCR Master Mix (Vazyme Biotech Co., Ltd.) with gene-specific primers. Analysis was performed as described previously [30]. The primer sequences are presented in Supplementary Table S9.

Generation of bmal2-overexpressing zebrafish

The zBmal2 and zBmal2-mut plasmids were linearized by ApaI and separately transcribed into mRNAs using the mMESSAGE mMACHINE™ SP6 Transcription Kit (Thermo Fisher Scientific). The mRNA products were finally diluted to 400 ng/μL for microinjection. wild-type (WT) (AB-type) zebrafish were housed at the Nanjing Normal University following standard procedures. Embryos were obtained by natural crosses and fertilized eggs were raised at 28.5°C. All experiments were conducted in accordance with guidelines approved by the Animal Ethical and Welfare Committee of Nanjing Normal University. The mRNA products from the two treatment groups (zbmal2 and zbmal2-mut) were microinjected into the single-cell-stage WT zebrafish embryos using an ASI pressure injection system under a stereomicroscope (Leica). The injected zebrafish larvae at 72 hours postfertilization (hpf) were used for behavioral assays, as previously reported [31], or were collected and stored in an RNA isolator for transcriptome sequencing and real-time PCR analysis [32]. Embryos injected with an empty plasmid (pCS2-mCherry) were sampled at equivalent developmental stages and used as the negative control.

Zebrafish behavioral assays

A single zebrafish larva was placed in each well of a 96-well plate at 72 hpf (each group contained 18 zebrafish larvae) for behavioral assays. All behavioral tracking was performed in the DanioVision chamber, and the locomotor activity was monitored for 24 hours using the EthoVision XT Version 14 software (Noldus Information Technology). The tracking area inside the chamber was maintained at 28.5°C using a temperature control unit.

Transcriptome sequencing

Total RNA was extracted and sequenced on Illumina NovaSeq platform (Novogene Co., Ltd., Beijing, China). Paired-end reads were mapped to the zebrafish reference genome (GRCz11) by HISAT2 [33]. Alignments were sorted by reading name using Samtools [34]. Preprocessing, normalization, and differential expression analysis were performed in R using the DESeq2 package [35]. differentially expressed genes (DEGs) were selected with criteria of p < 0.05 and fold change ≥ 1. Log-transformed gene expression was used to generate the heatmap for DEGs.

Statistical analysis

Statistical calculations were performed using the GraphPad Prism 8 software. The D’Agostino & Pearson normality and Shapiro–Wilk tests were used to check the data distribution. One-way repeated-measures ANOVA was used to evaluate group differences, followed by the Holm–Sidak multiple posthoc test for multiple comparisons. All experiments were independently repeated at least three times. All data were expressed as the mean ± SEM. The labels “ns,” “*,” “**,” and “***” indicate not significant, p < 0.05, p < 0.01, and p < 0.001, respectively.

Results

Unique selection signature in the BMAL2 gene of marine mammals

Complete BMAL2 coding sequences for 85 representative mammal species were recovered in this study, including 24 cetaceans, 2 sirenians, and 11 pinnipeds (six phocids, four otariids, and the walrus Odobenus rosmarus, Figure 1 and Supplementary Table S1). A phylogenetic tree of the BMAL2 gene was constructed, with 80% of nodes having bootstrap values of > 90 (Supplementary Figure S1). Almost all orders and superorders (Laurasiatheria, Euarchontoglires, Xenarthra, and Afrotheria) were consistent with the widely recognized phylogenetic relationships. To understand the evolutionary pattern of the BMAL2 gene, signatures of natural selection were tested using several statistical tests based on the ω ratio. A pair of site models (M8a vs. M8) was first used to identify whether some codons were the targets of positive selection in marine mammals (except phocids). LRTs revealed that the model incorporating selection (M8) fitted significantly better than the neutral model (M8a) for the BMAL2 gene (ω = 2.056, p = 0.040), and 10 specific codons were detected under positive selection using the BEB method with a posterior probability of ≥ 0.8 (Supplementary Table S2). Further evidence of positive selection was presented by three other site-level analyses implemented in FEL, MEME, and FUBAR (Supplementary Table S2). The FEL method identified four positively selected codons and MEME identified nine positively selected codons, with a significance level of 0.1. Additionally, FUBAR found 10 sites under diversifying selection with a posterior probability of > 0.9. Of these putative positively selected codons, nine (258, 259, 301, 334, 500, 521, 545, 555, and 564) were detected by at least two methods and four (259, 301, 334, and 564) by all methods.

Further tests were performed to determine whether a similar selection pattern also occurred within cetaceans. The results showed that the BMAL2 gene had undergone positive selection in the cetaceans-only dataset, where the LRTs of the site model were statistically significant (ω = 8.042, p = 0.002). Signatures of positive selection were found in five codons via the BEB approach with posterior probabilities ≥ 0.8 (Supplementary Table S3). Furthermore, three, six, and two positively selected codons were individually identified using FEL, MEME, and FUBAR, respectively (Supplementary Table S3). Among them, five codons (258, 259, 334, 500, and 564) were detected using at least three methods and two (334 and 564) were detected using all methods. In contrast, none of the four methods identified a no specific site under positive selection within the sirenians-only dataset, the phocids-only dataset, and non-phocid pinnipeds-only dataset.

A more stringent branch-site model was then used to investigate whether positive selection was acting on specific sites in mammalian lineages exhibiting USWS across the phylogeny of all mammals (Figure 1). Significant signs of positive selection were detected in cetaceans (ω = 11.552, p < 0.001) and sirenians (ω = 21.127, p = 0.017), with four and two codons identified as possible positively selected targets by the BEB procedure, respectively. Furthermore, there was no evidence of positive selection found in their respective sister taxa (Supplementary Table S4). Additionally, positive selection was identified in sirenians using the gene-wide test implemented in BUSTED (Supplementary Table S5).

Cetacean-specific mutations caused functional alteration in BMAL2

The inferred protein sequences were then examined further to identify unique amino acid residue substitutions in species with USWS. Notably, the cetacean BMAL2 proteins had two amino acid replacements that differed from all other species: the first residue was found in the PAS A domain (a K-to-E replacement at position 204 in the human exon) and the second within the PAS B domain (a K-to-Q replacement at position 346 in the human exon) (Figure 2A). These two cetacean-specific mutations were predicted by the Polyphen-2 and PROVEAN algorithms to influence protein function (Supplementary Table S6).

Figure 2.

Altered molecular function of BMAL2 by cetacean-specific mutations. (A) The K204E and K346Q mutations are located in the PAS A domain and PAS B domains, respectively, and are highly conserved among mammalian species and zebrafish. (B) Fluorescence microscopy of HEK293T cells expressing mCherry-tagged mBMAL2 and mBMAL2-mut. Nuclei stained with DAPI are shown in blue. (C) Western blot and time course of mBMAL2 and mBMAL2-mut abundance after cycloheximide (CHX) was supplied to transfected 293T cells. (D) Activation activity of mBMAL2 or mBMAL2-mut was examined using the mouse Per2 promoter-driven luciferase reporter system in HEK293T cells. (E) Activation activity of mBMAL2 or mBMAL2-mut was examined using the mouse Per1 promoter-driven luciferase reporter system in HEK293T cells. ***p < 0.001; ns, not significant. Error bars represent ± SEM.

To test the effect of the cetacean-specific mutations, E and Q amino acids were mutated to the corresponding positions in mouse WT BMAL2 (K179 K319), and the functional differences between the mBMAL2 and mBMAL2-mut proteins were compared in cultured HEK293T cells. The results revealed that both mBMAL2 and mBMAL2-mut were predominantly localized in the nucleus (Figure 2B), with no statistically significant differences. In addition, the mutant protein showed similar protein stability to mBMAL2 when treated with cycloheximide (CHX) (Figure 2C). However, luciferase assays using a mouse Per2 promoter construct showed that mBMAL2-mut possessed significantly stronger activation activity than mBMAL2 (p = 0.0001, Figure 2D). In contrast, the ability of mBMAL2-mut to activate a mouse Per1 promoter-driven luciferase reporter was indistinguishable from that of mBMAL2 (p = 0.9937, Figure 2E).

Cetacean-specific mutations enhance Per2 gene expression by interacting with an E-box-like transcriptional element

Because the circadian transcription of the PER2 gene is driven by an E-box-like enhancer (CACGTT) and the expression of PER1 is dependent on a classical E-box enhancer (CACGTG), we examined whether the altered transcriptional activation of mBMAL2-mut was related to the E-box-like element. Luciferase assays showed that mBMAL2-mut exhibited a significantly enhanced ability to activate a mutant Per1 promoter-driven luciferase reporter when the E-box sequence was substituted for the Per2 E-box-like sequence (p = 0.0213, Figure 3A). In contrast, mBMAL2-mut had similar transactivation ability to the WT for the mutant Per2 promoter, wherein the E-box-like sequence was mutated into the Per1 E-box sequence (p > 0.9999, Figure 3B). These indicated that mBMAL2-mut was more responsive to the E-box-like enhancer in the Per2 promoter compared to the E-box enhancer in the Per1 promoter.

Figure 3.

The response to the Per2 E-box-like regulatory element and the expression of the PER2 gene are changed by cetacean-specific mutation. (A) mBMAL2-mut shows increased transcriptional activation of the Per1 promoter after replacing the E-box sequence with the Per2 E-box-like sequence. (B) mBMAL2-mut shows decreased transcriptional activation of the Per2 promoter after replacing the E-box-like sequence with the Per1 E-box sequence. (C) Comparison of circadian gene mRNA expression levels between mBmal2-mut transfected, mBmal2-mut transfected, and control cells. (D) qRT-PCR shows upregulation of Per2 expression in mBmal2-mut transfected cells. (E) Western blot shows increased PER2 protein expression in mBmal2-mut transfected cells. Relative abundance of PER2 was quantified and is shown on the right (n = 5). *p < 0.05; **p < 0.01; ****p < 0.0001; ns, not significant. Error bars represent ± SEM.

Further expression assays were performed in NIH 3T3 cells to determine whether the cetacean-specific mutations had any influence on circadian regulation. The quantitative real-time PCR (qRT-PCR) analysis showed that the changes in the mRNA expression of Clock in the three transfected groups were statistically insignificant. However, the mRNA expression of Per2 in mBMAL2-mut-transfected cells was significantly higher than in the mBMAL2-transfected (p = 0.0422) or empty vector control cells (p = 0.0253) (Figure 3C). Furthermore, the mRNA expression of Bmal1 and Per1 in mBMAL2-mut-transfected cells was obviously lower than in control cells (Bmal1: p = 0.0007; Per1: p = 0.0081) but not significantly different from that in mBMAL2-transfected (Bmal1: p = 0.1023; Per1: p > 0.9999) cells, whereas Cry1 mRNA expression in mBMAL2-mut-transfected cells was significantly higher than in control cells (p < 0.0001) but almost identical to that in mBMAL2-transfected cells (p = 0.4780) (Figure 3C). The temporal profiles of gene expression revealed that the Per2 levels were higher at most time points in mBMAL2-mut-transfected cells than in mBMAL2-transfected or control cells (Figure 3D). The elevated Per2 protein expression in the mBMAL2-mut-transfected cells was confirmed by western blotting analysis after transfection for 48 hours (Figure 3E). These findings suggested a role for the cetacean-specific mutations of Bmal2 in promoting the expression of Per2.

Zebrafish overexpressing zbmal2-mut have increased per2 expression and locomotor activities

To investigate the effects of the cetacean-specific mutations in vivo, zebrafish WT and mutant bmal2 (K220E and K363Q) mRNAs were overexpressed in zebrafish embryos, and the expression of core circadian clock genes was examined using qRT-PCR. This was consistent with the above findings that the expression of the mammalian PER2 ortholog per2 was significantly higher in zbmal2-mut-overexpressing zebrafish than in the zbmal2 overexpression group (p = 0.0001) or the control group (p < 0.0001) (Figure 4A). In contrast, the expression of bmal1b, clock1a, per1a, per1b, and cry1b in zbmal2-mut-overexpressing zebrafish was lower than in zbmal2-overexpressing zebrafish but higher than in the control zebrafish (Figure 4A).

Figure 4.

Altered circadian gene expressions and locomotor activities in zbmal2-mut overexpressed zebrafish. (A) Comparison of circadian gene mRNA expression levels between zbmal2 overexpressed, zbmal2-mut overexpressed, and control zebrafish. (B) Locomotor activities were monitored and analyzed in zbmal2 overexpressed, zbmal2-mut overexpressed, and control zebrafish larvae 72 hours postfertilization. (C-E) Total mobile distance within 24 hours (C), light phase (D), and dark phase (E) were calculated in zbmal2 overexpressed, zbmal2-mut overexpressed, and control zebrafish. *p < 0.05; **p < 0.01; ****p < 0.0001. Error bars represent ± SEM.

Locomotor activity assays revealed that zbmal2-mut-overexpressing zebrafish covered longer total distances than the zbmal2 overexpression group (p = 0.0003) or the control zebrafish (p < 0.0001) (Figure 4, B and C). The mobile distance recorded during the light phase was markedly higher in the zbmal2-mut overexpression group than in the control group (p < 0.0001) but not significantly different from that in the zbmal2 overexpression group (p = 0.1717) (Figure 4D). During the dark phase, the distance covered by the zbmal2-mut overexpression group was nearly equal to the control group (p = 0.5351) but significantly higher than in the zbmal2 overexpression group (p = 0.0002) (Figure 4E). These data suggest that the overexpression of cetacean-specific mutated bmal2 leads to increased locomotor activity in zebrafish.

Gene expression changes in zbmal2-mut-overexpressing zebrafish

High-throughput transcriptome sequencing was performed to analyze gene expression alterations induced by the cetacean-specific mutations. Transcriptome analysis identified 434 DEGs in the zbmal2-mut-overexpressing zebrafish compared with the control group; 251 were upregulated and 183 were downregulated (Figure 5A). Similarly, 686 DEGs were identified in the zbmal2-overexpressing zebrafish compared with the control group; 322 were upregulated and 364 were downregulated (Figure 5A). Among them, 312 DEGs were unique to the zbmal2-mut overexpression group but not to the zbmal2 overexpression group (Figure 5B). Gene Ontology and Kyoto Encyclopedia of Genes and Genomes analyses of these 312 unique DEGs revealed that genes involved in metabolism, oxidative phosphorylation, and small-molecule biosynthetic processes were markedly altered in the zbmal2-mut overexpression group (Figure 5C and Supplementary Figure S2). Interestingly, some of the identified DEGs were known to be involved in the regulation of sleep, cognition, and locomotion activity (Supplementary Table S7). For example, both fkbp5 and akap12b were upregulated in the zbmal2-mut overexpression group. The deletion of Fkbp5 in mice yielded a proresilience sleep phenotype, with more time spent on NREM and REM sleep after SD [36], whereas zebrafish deficient in akap12b exhibited severe defects in locomotor activity [37].

Figure 5.

Altered gene expression in the zbmal2-mut overexpressed zebrafish was revealed by transcriptome analysis. (A) The left shows a volcano of differentially expressed genes in zbmal2 overexpressed group versus the control group; the right shows a volcano of differentially expressed genes in zbmal2-mut overexpressed group versus the control group. (B) Venn diagram of differentially expressed genes in the zbmal2 overexpressed group versus the control group and the zbmal2-mut overexpressed group versus the control group. (C) Heatmap of 312 unique differentially expressed genes in the zbmal2 overexpressed, zbmal2-mut overexpressed, and control groups.

In addition, compared with zbmal2-overexpressing zebrafish, 389 upregulated genes and 255 downregulated genes were identified in zbmal2-mut-overexpressing zebrafish (Figure 6A). Functional characterization showed that these DEGs in zbmal2-mut were associated with cell apoptosis, lipid metabolism, and signal transduction (Supplementary Figure S3). In particular, 18 genes involved in promoting wakefulness were upregulated in zbmal2-mut-overexpressing zebrafish (Figure 6B), including three noradrenergic genes (dbh, adra2a, and adra2b), four dopamine receptor genes (drd1b, drd3, drd4a, and drd7), three histamine-related genes (hdc, hrh2b, and hrh3), four glutamate receptor genes (grm1b, grm3, gria1a, and grik1b), and four serotonin receptor genes (htr2ab, htr2cl1, htr2cl2, and htr3b). In contrast, nine genes related to sleep initiation were downregulated (Figure 6B), namely, three cholinergic receptor genes (chrm3a, chrna2b, and chrnb2b), four γ-aminobutyric acid (GABA)-related genes (gad2, gabra4, gabrb1, and gabrb2), and two adenosine receptor genes (adora1b and adora2ab). Of these genes, the trends in adra2a, drd1b, hdc, grm1b, gabra4, gabrb1, and adora2ab expression were reconfirmed using qRT-PCR (Figure 6C).

Figure 6.

The identified differentially expressed genes in zbmal2-mut overexpressed zebrafish compared with zbmal2 overexpressed zebrafish. (A) A volcano of differentially expressed genes in the zbmal2 overexpressed group versus zbmal2-mut overexpressed group. (B) Histogram of 29 genes related to sleep regulation in the zbmal2 overexpressed, zbmal2-mut overexpressed, and control groups. (C) Seven sleep-related genes revealed by transcriptome analysis are reconfirmed by independent qRT-PCR analysis. *p < 0.05; **p < 0.01; ****p < 0.0001. Error bars represent ± SEM.

Discussion

The evolutionary pressures of more efficient monitoring, thermogenesis, and the need to breathe out of water have driven cetaceans to adopt USWS, a type of sleep that allows one brain hemisphere to sleep while the awake hemisphere coordinates movement for surfacing and heat generation [4]. Although the functions and several physiological foundations of unihemispheric sleep have been proposed [38, 39], the underlying mechanisms remain poorly understood. Sleep is regulated by the circadian system, and genetic changes in major components of the circadian feedback loop have been shown to affect various sleep parameters, such as sleep duration and timing [40]. For example, the clock mutation significantly decreases sleep duration in mice [41], and Npas2 knockout in mice reduces the time spent in SWS [42]. Investigations into the molecular evolution of circadian genes will help clarify the genetic mechanisms of sleep alterations in animals. In this study, we have explored the evolutionary characteristics of BMAL2, one of the important positive regulatory components in the circadian system, in mammals and identified two cetacean-specific amino acid substitutions (K204E and K346Q) that were potential causes of the functional changes. The in vitro transfection of WT and mutant mBMAL2 showed that these mutations facilitated PER2 gene expression by specifically enhancing the response to the E-box-like enhancer, which might be related to the unusual circadian regulation and arousal-promoting needs in cetaceans during USWS. In addition, transgenic zebrafish overexpressing cetacean-specific mutated zbmal2 exhibited significantly increased locomotor activity and altered the expression of sleep-regulating genes. These findings are strong evidence for the role of the BMAL2 gene in the evolution of cetacean USWS.

One intriguing finding is that the cetacean-specific mutations affected the activating effect of BMAL2 on the PER2 gene. It is well established that BMAL2 heterodimerizes with either CLOCK or NPAS2 and that these heterodimers stimulate the transcription of downstream circadian genes, such as PER1 and CRY1, through site-specific binding to the E-box regulator element [16, 43]. However, the identified cetacean-specific mutations in the key PAS domains of BMAL2 were found to specifically enhance its transcriptional activation activity in an E-box-like-dependent manner. The circadian transcription of PER2 was shown to be driven by E-box-like element rather than by E-box element [44]. Accordingly, both in vitro and in vivo expression analyses showed that the altered transcriptional activation caused by cetacean-specific mutations led to the increased mRNA and protein expression of PER2. In addition to circadian modulation, previous studies have shown that PER2 is closely tied to sleep regulation; for example, genetic variation in the PER2 gene was found to be associated with familial advanced sleep phase syndrome [45]. The familial advanced sleep phase syndrome S662G mutation in the human PER2 gene results in increased PER2 expression and a lengthened circadian period in transgenic mice [46]; sustained high levels of mouse Per2 expression have a negative effect on recovery sleep after SD [47]; and per2-null mutant zebrafish exhibit decreased locomotor activities under light–dark and 2-hour phase delay under constant darkness [48]. In addition, the knockout of Bmal1 and accordingly downregulation of Bmal2 resulted in increased total sleep time, sleep fragmentation, and EEG delta power in mice [13, 15], suggesting that BMAL2 has a potential inhibitory effect on sleep regulation. Therefore, the benefits of adaptive changes in cetacean BMAL2 are likely twofold. First, the specifically increased activation activity of BMAL2 may contribute to the adjusted sleep rhythms observed in cetaceans through its influence on the circadian feedback process, which allows constant activity and alertness throughout the day. Consistently, motor activity in several cetacean species is found to be essentially continuous from birth to death, that is, they never float, sink to the bottom, or remain still [49, 50]. Second, the altered activation activity of BMAL2 and the consequent increase in PER2 expression are conducive to reducing the need for sleep and facilitating motion increases. Strong support for this is provided by the observation that cetacean-specific mutated zbmal2-overexpressing zebrafish have significantly greater locomotor activities than the zbmal2 overexpression group or the control group. For cetaceans, the ability to keep the brain active is important to adopt USWS and achieve continuous movement and vigilance.

In addition to the circadian system, the alternation between sleep and wakefulness involves the integration of the influences of several major neurotransmitters responsible for sleep homeostasis [51, 52]. For example, noradrenaline, dopamine, histamine, glutamate, and serotonin are important waking factors that act together for the generation and maintenance of wakefulness. Conversely, acetylcholine, GABA, and several peptide factors such as adenosine are involved in initiating and maintaining sleep. Transcriptome analysis revealed that cetacean-specific mutations led to the significant upregulation of 16 genes coding for arousal neurotransmitters, implying enhancement of the wake-activation effect. Furthermore, dbh and hdc, enzymes coding for genes that catalyze the production of norepinephrine and histamine, respectively, were found to be upregulated in zbmal2-mut-overexpressing zebrafish, whereas the two genes hnmt and comt, which encode enzymes with a role in metabolizing histamine and dopamine, respectively, were downregulated. The dbh-knockout zebrafish lack noradrenaline and have dramatically increased sleep [53]. The conditional knockout of the Hdc gene in mice caused a reduction in brain histamine content and a decrease in wakefulness [54], whereas mice lacking Hnmt had dramatically enhanced histamine content and a disrupted sleep–wake cycle [55]. Furthermore, the V158M polymorphism in the COMT gene was demonstrated to impact normal sleep–wake regulation and even sleep pathologies [56]. These findings provide robust evidence that the functional alterations of cetacean BMAL2 may contribute to prolonged wakefulness on one side of the brain during USWS by promoting the accumulation of arousal neurotransmitters and the subsequent increased activation. In addition, four GABAergic genes (gad2, gabra4, gabrb1, and gabrb2) and two adenosine receptor genes (adora1b and adora2ab) exhibited a significant downregulation in zbmal2-mut-overexpressing zebrafish. Functionally, GAD2 encodes glutamate decarboxylase, which catalyzes the production of GABA from glutamate, and the three GABA receptor genes all encode subunits of GABA_A receptors. The activation of GABA_A receptors favors sleep [57], as GABA_A agonists are able to increase sleep continuity and promote NREM sleep [58]. Additionally, these agonists were found to prevent the homeostatic sleep rebound normally seen after sleep deprivation [59]. Furthermore, adora1b and adora2ab encoded the A1 and A2A subtypes of the adenosine receptor, respectively. Adenosine is an important homeostatic sleep factor [60]. Prolonged wakefulness stimulates the release of adenosine, which, in turn, acts in the basal forebrain and preoptic areas via the A1 and A2A receptors, contributing to sleep recovery [61]. Decreased expression of sleep-promoting genes could be beneficial in suppressing global sleep-inducing effects and maintaining single-hemisphere awakening in cetaceans. More importantly, the potential decline of sleep-homeostatic regulation is consistent with the ability of cetaceans to maintain continuous vigilance for 5 days, and no detectable decrease in activity or signs of sleep rebound was found at the end of this period [62, 63], a finding worthy of further attention. Besides the observed expression changes in these neurotransmitters, sleep regulation is influenced by various pathways. For instance, the neuropeptide orexin and its G-protein–coupled receptors play crucial roles in maintaining wakefulness. A deficiency of orexin signaling leads to dramatic increases in sleep and even narcoleptic phenotypes [64, 65]. Overexpression of the orexin gene promoted wakefulness and locomotor activity in zebrafish [21]. Our present results revealed an increase in locomotor activity in the zbmal2-mut-overexpressing zebrafish. However, no significant difference in orexin signaling was detected among the groups, which could be attributed to differences in timing and the influences of circadian genes on sleep mechanisms. Indeed, previous research revealed that DEC2, a transcription factor of the circadian system, regulates the sleep–wake duration by suppressing orexin expression by binding to E-box elements in the promoter region of the orexin gene [66].

Previous research has shown that the double deletion of Chrm1 and Chrm3 in mice could diminish REM sleep to an extremely low level [67], indicating the essential role of the cholinergic pathway in REM sleep. Interestingly, downregulation of three cholinergic receptor genes (chrm3a, chrna2b, and chrnb2b) was found in zbmal2-mut-overexpressing zebrafish, which may suggest a negative influence on the generation of REM sleep in cetaceans. During REM sleep, muscle tone becomes atonic and the temperature-regulating machinery remains suspended [68]. Therefore, cetaceans have either radically reduced or completely eliminated REM sleep to adapt to their thermally challenging environment [6].

Overall, our results revealed the mechanistic correlation between adaptive changes in the circadian positive regulator BMAL2 and the evolution of unihemispheric sleep in cetaceans. Double-luciferase assays revealed that the identified cetacean-specific mutations significantly increased the transcriptional activation of PER2. This gene is involved in sleep regulation, specifically by promoting the response to an E-box-like enhancer. The consequent increase in PER2 expression and the altered circadian feedback may exert a positive role in adjusting irregular sleep rhythms and facilitating arousal activity, which is strongly supported by the enhanced locomotor activities in zbmal2-mut-overexpressing zebrafish. Furthermore, the significant upregulation of neurotransmitter genes associated with arousal and the corresponding downregulation of sleep-promoting genes in zbmal2-mut-overexpressing zebrafish suggested adaptation to the cetacean need to impair global sleep-inducing effects and improve the capacity to maintain wakefulness during USWS. Additionally, we created transient transgenic zebrafish to investigate the in vivo effects of cetacean-specific mutations in BMAL2 efficiently and reproducibly while reducing the impact of position effects and genetic backgrounds [69]. However, the present approach had several limitations, including transient overexpression effects, nonspecific expression patterns, and difficulty in accurately observing and assessing sleep alterations over the long term. Additional research using stable transgenic zebrafish is required to determine the sleep and behavioral changes, especially homeostatic sleep rebound [70], and further reveal the role of mutant BMAL2 in sleep regulation.

Supplementary Material

Supplementary material is available at SLEEP online.

Funding

This work was supported by the National Natural Science Foundation of China (grant no. 32200348), the China Postdoctoral Science Foundation (2022M710879) to D.Y., the Key Project of the National Natural Science Foundation of China (grant no. 32030011) and the National Key Research and Development Program of China (grant no. 2022YFF1301600) to G. Y., the National Natural Science Foundation of China (grant nos. 32270453, 31772448, and 32270442), the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD), the Qinglan Project of Jiangsu Province to S.X., the PI Project of Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou) (GML2021GD0805) to G.Y., and the Guangzhou Science and Technology Plan Project (2023A04J0769) to D.Y..

Author Contributions

G.Y. and S.X. designed and managed the study. D.Y. and Y.C. collected the data and conducted the evolutionary analyses. B.Z. performed in vitro and in vivo experiments. D.Y. and B.Z. prepared the original draft. G.Y. and S.X. revised the manuscript. All authors read and approved the final manuscript.

Disclosure Statement

Financial disclosure: There are no financial conflicts of interest. Nonfinancial disclosure: None.

Data Availability

The raw transcriptome data have been deposited at NCBI under the project accession number PRJNA956154. All other data are included in the article and/or supporting information.