Identification of putative circadian clock genes in the American horseshoe crab, L. polyphemus

Abstract

While the American horseshoe crab, Limulus polyphemus, has robust circadian and circatidal rhythms, virtually nothing is known about the molecular basis of these rhythms in this species or any other chelicerate. In this study, next generation sequencing was used to assemble transcriptomic reads and then putative homologs of known core and accessory circadian genes were identified in these databases. Homologous transcripts were discovered for one circadian clock input gene, five core genes, 22 accessory genes, and two possible output pathways. Alignments and functional domain analyses showed generally high conservation between the putative L. polyphemus clock genes and homologs from Drosophila melanogaster and Daphnia pulex. The presence of both cry1 and cry2 in the L. polyphemus transcriptome would classify its system as an “ancestral”, type 2 clock system. In addition, a novel duplication of CYCLE, and a novel triplication of PERIOD were found. Investigations are currently underway to determine if any of these “circadian” genes also participate in the molecular processes that drive the Limulus circatidal clock.

1. Introduction.

Endogenously driven biological rhythms have been observed in virtually all organisms in which a concerted effort has been made to look for them (Dunlap, 1999). These rhythmic changes in physiology and behavior are driven by internal biological clocks and help organisms to both anticipate, and synchronize to, predictable rhythmic changes in their environment. These clocks are composed of three essential elements: 1) the input proteins that allow the clock to synchronize to environmental rhythms, such as light:dark cycles; 2) the clock itself, made up of the proteins that allow the clock to keep time; and 3) the output proteins that help to mediate changes in physiology and behavior (Allada and Chung, 2010; Dunlap, 1999). The molecular basis of these three components of the clock have been elucidated in a few model systems such as the fruit fly, Drosophila melanogaster and the mouse, Mus musculus (Zhang and Kay, 2010) and these findings have provided the framework for studies in other organisms. Although the specific molecular mechanisms of the circadian clock vary greatly between distantly related animal models (i.e. M. musculus and D. melanogaster), the general architecture appears to be well preserved across nearly all living organisms (Dunlap, 1999). In plants, fungi, and animals, circadian clocks appear to be composed of interlocking transcription-translation cycles that feedback to control their own transcription.

These circadian clocks involve two sets of proteins: the core proteins (which are part of a negative feedback loop) and the accessory proteins (which modulate the core proteins and are necessary for maintaining the ~24hr periodicity of the core clock). In D. melanogaster the core clock is composed of four proteins: PERIOD (PER), TIMELESS (TIM), CLOCK (CLK), and CYCLE (CYC) (Allada and Chung, 2010). When the per and tim genes are transcribed and translated, they form a heterodimer that acts at the clk promoter to up-regulate CLK expression (Allada and Chung, 2010; Chang and Reppert, 2003). In other insect species Non-photoreceptive(np) CRYPTOCHROME 2 (npCRY2) acts in addition to, or in place of, TIM (Rubin et al., 2006; Yuan et al., 2007). As PER and TIM build up over time, they dimerize and become phosphorylated by accessory genes such as CASEIN KINASE Ie (CKIe). This phosphorylation allows this heterodimer to enter the nucleus and to remove CLK-CYC from the E-box, turning off per and tim (Allada and Chung, 2010). The proteins CLK and CYC (which appear to be constitutively expressed) are, in turn, capable of repressing the activity of per and tim, completing the negative feedback loop.

This core negative feedback loop is also modulated by a host of accessory proteins, such as VRILLE (VRI), CLOCKWORK ORANGE (CWO), and PAR DOMAIN PROTEIN 1e (PDP1e), which form a secondary feedback loop with CLK (Allada and Chung, 2010). Additionally, the proteins SUPERNUMERARY LIMBS (SLIMB) and ARYL HYDROCARBON RECEPTOR NUCLEAR TRANSPORTER (ARNT) also help to modulate the transcription of the core clock components. The protein kinases CASEIN KINASE IIa (CKIIa), and CASEIN KINASE IIß (CKIIß) serve to phosphorylate PER, while TIM is phosphorylated by CKIIa, CKIIß, CASEIN KINASE Ia (CKIa), SHAGGY (SGG), and JETLAG (JET). Along with these kinases, the protein phosphatases PROTEIN PHOSPHATASE 1a (PP1a), PROTEIN PHOSPHATASE 1ß (PP1ß), PROTEIN PHOSPHATASE 2a-MTS (PP2a-MTS), PROTEIN PHOSPHATASE 2a-WBT (PP2a-WBT), PROTEIN PHOSPHATASE 2a-WS (PP2a-TWS) are also crucial for proper clock function (Allada and Chung, 2010).

The input portion of the circadian clock system is provided by the photoreactive(p) protein CRYPTOCHROME 1 (pCRY1). In D. melanogaster, pCRY1 undergoes a conformational change in the presence of blue light (such as in sunlight) and ubiquinates TIM, tagging it for degradation via the proteasome (Emery et al., 1998; Tauber et al., 2004; Yuan et al., 2007). This degradation acts to “reset” the clock, because when TIM is degraded, it can no longer form a heterodimer with PER and the transcriptional regulation of clk is terminated.

The neuropeptides Pigment Dispersing Hormone (PDH) and NEUROPEPTIDE F (NPF) (Lee et al., 2006; Strauss et al., 2011; Taghert and Shafer, 2006) are crucial for the output of the clock in D. melanogaster. Ablation of PDH cells, or mutations of the pdh gene, disrupts the expression of circadian rhythms in this species D. melanogaster (Renn et al., 1999), and these effects are exacerbated by the ablation of NPF containing cells (Hermann et al., 2012). While these peptides are thought to induce numerous physiological and behavioral changes (Lee et al., 2006; Strauss et al., 2011), the mechanism of the control of either NPF or PDH release by the core circadian clock is not known (Depetris-Chauvin et al., 2011).

While much is known about the molecular mechanisms of the circadian clock in D. melanogaster and a few other insect species (Dunlap, 1999; Tomioka and Matsumoto, 2010), much less is known about the circadian clocks of non-insect arthropods, with the exception of Daphnia pulex (Tilden et al., 2011) and Calanus finmarchicus (Christie et al., 2013). One entire sub-phylum of Arthropoda that has been ignored is the chelicerata, which encompasses a range of organisms including spiders, ticks, mites, scorpions, and horseshoe crabs (Giribet and Edgecombe, 2012). The American horseshoe crab, Limulus polyphemus, has long been known to exhibit robust circadian rhythms of lateral eye sensitivity to light and has served as a model species for studies in photophysiology and its circadian control (Barlow, 1983; Barlow et al., 1980). Bob Barlow, his colleagues, and other investigators have shown that more than twenty independent changes occur in the eye of L. polyphemus to achieve an approximately 100,000 fold change in eye sensitivity between night and day (Barlow et al., 1980; Battelle, 2013). Yet, despite decades of study of this system, the molecular mechanisms that drive this clock in L. polyphemus are completely unknown.

Horseshoe crabs also have an additional timing system that serves to synchronize its locomotor activity to the ~ 12.4 hour tidal cycles (Chabot and Watson, 2010; Watson et al., 2008). While circatidal rhythms such as this have been documented in several species, the molecular basis of these clocks is unknown (de la Iglesia and Hsu, 2010; Takekata et al., 2014; Tessmar-Raible et al., 2011). Furthermore, there is some controversy surrounding the nature of the timing system that drives these rhythms. The two primary competing theories that have been put forth to explain the underlying mechanisms that give rise to these 12.4 h rhythms are: 1) The circatidal oscillator theory, which states that two bouts of activity per day are controlled by one, ~ 12.4 h, circatidal clock (which can also be modulated by a circadian clock (Naylor, 1996)) and; 2) the circalunidian theory, which states that each bout of activity is controlled by a separate, ~24.8 hr (the time between successive moonrises), circalunidian oscillator (Palmer and Williams, 1986). This circalunidian model has been well supported by behavioral evidence in L. polyphemus (Chabot et al., 2016; Chabot and Watson, 2010) and our working hypothesis is that each bout of activity is controlled by a modified circadian (Palmer and Williams, 1986) clock that has a slightly longer period than 24 hours. This would be similar to the situation in D. melanogaster, where dawn and dusk bouts of activity are controlled by separate, coordinated, circadian clocks (Stoleru et al., 2004). The goal of this study was to determine if horseshoe crabs possess some of the circadian genes found in other model species. The identification of these genes in L. polyphemus would not only provide the scientific community with a resource to begin studying circadian clock in chelicerates, but may also allow us to begin to understand the molecular basis of circatidal/circalunidian clocks.

2. Methods

2.1. Animals and RNA extraction.

L. polyphemus individuals were collected from Great Bay in Durham NH. Animals were treated in accordance with NIH guide for the care and use of animals (NIH publications No. 8023, revised 1978). The entire central nervous system (brain and ventral nerve cord) was dissected from 4 individuals, immediately snap frozen on dry ice, and stored in 1.5ml Eppendorf tubes at −80°C. Samples were then shipped on dry ice to University of Vermont Cancer Center DNA Analysis Facility (Burlington, VT), where RNA was extracted using Trizol RNA extraction (Invitrogen, Carlsbad, CA) and cleaned using the Qiagen RNAeasy Mini column (Qiagen, Valencia, CA). Quality and concentration of cleaned RNA was determined using Nanodrop spectrophotometer (Thermo Scientific, Madison WI), Qubit Spectrofluorometer (Life Technologies, Carlsbad, CA) and Agilent Bioanalyzer 2100 (Agilent technologies, Santa Clara, CA).

2.1. Library construction and sequencing.

A transcriptome library was constructed using Illumina TruSeq RNA Sample Prep LT version 2 (RS-122–2001/2002). 1 ug of each sample was PolyA enriched using magnetic beads and reverse transcribed into cDNA via Superscript II [Invitrogen]. The cDNA was then fragmented, end-repaired, adenylated, and tagged with sequence adaptors for source identification. The cDNA was sequenced using 12pM/flow cell lane with Illumina CBOT and HiSeq1000/1500.

2.2. Transcriptome de novo assembly.

Transcriptome reads were de novo assembled using the CLC genomics workbench proprietary algorithm (CLC version 5.1.2, CLCBio, Boston, MA), as well as SOAP de novo assembly software (Luo et al., 2012) (http://soap.genomics.org.cn/soapdenovo.html, version 1.02) with the default settings.

2.2. Transcriptome mining and gene identification.

The transcriptome was mined with the CLC genomics workbench tBLASTn program (version 5.1.2), using D. melanogaster and other arthropod query sequences. The top hits from the L. polyphemus transcriptome were translated and subjected to a reciprocal BLASTp against the entire UniProtKB protein database to verify sequence homology. A second reciprocal BLASTp was also performed against D. melanogaster and D. pulex circadian proteins exclusively to compare scores to these model organisms.

2.3. Sequence analysis.

Transcripts of interest were translated into protein sequences using CLC genomics workbench (version 5.1.2). These sequences were then aligned to respective orthologs of D. melanogaster and D. pulex, along with an ortholog of a chelicerate species when one was available using CLUSTAL Omega program (http://www.ebi.ac.uk/Tools/msa/clustalo/). Functional domains of all proteins of interest in L. polyphemus, D. melanogaster, Daphnia, and chelicerate species when present were identified using SMART Genome program (http://smart.embl-heidelberg.de/ (Letunic et al., 2015; Schultz et al., 1998)). The %Identity/%Similarity of the amino acid sequences these proteins and their respective functional domains were calculated using default settings of the SIAS program (http://imed.med.ucm.es/Tools/sias.html). Sequence alignments from CLUSTAL Omega were pasted into a word document and functional domains were color coded as follows: Light blue:bHLH, Green:PAS, Red:PAC, Grey:Period C, Dark green:Photolyase, Gold:FAD Binding 7, Dark blue:Timeless. Phylogenetic trees were constructed using Mega 5.2 (http://www.megasoftware.net/). Sequences, other than L. polyphemus proteins, were extracted from UniProt database and aligned using MUSCLE (http://www.ebi.ac.uk/Tools/msa/muscle/). After alignment a Maximum likelihood tree was generated using default setting and 500 Bootstraps. Tree branches were color coded to specify different clades on the trees: Red:Chelicerata, Blue:Non-chelicerate arthropods, Black:Vertebrates, Violet:Fungi, Green:Nematode. Diamonds were added at gene duplication events.

3. Results.

3.1. Top 5 BLAST hits from UniProtKB

In our analysis of the L. polyphemus (Lp) transcriptome we identified orthologs of all five core clock components (PER, TIM, CLK, CYC, npCRY2), including 3 copies of the per gene [perioda (pera), periodb (perb), and periodc (perc)] and 2 copies of the cyc gene [cycle1 (cyc1) and cycle2 (cyc2)]. Top BLAST hits for all core clock genes are reported in Table 1. A reciprocal BLASTp of all three LpPER paralogs against the UniProt KB Protein database showed closest homology to arthropod PER proteins, including orthologs from three insect species and one chelicerate species (Ixodes scapularis, Black legged tick). BLASTp of LpTIM, LpCLK, LpCYC1, LpCYC2 and LpnpCRY2 showed closest homology to insect orthologs. Top BLAST hits for all core clock components were significant, with e-values ranging from 10^−114 to less than 10^−180, and BLAST scores ranging from 1004 to 2199.

Table 1:

Top 5 BLASTp hits from UniProt.

Proteins	UniProt KB Protein Database
Core Proteins	UniProt ID	Species	Protein description	E-value	BLAST score
PERIODA (PERA)
LpPERA	H7C8F3	Apteronemobius asahinai	PERIOD, isoform 1	1e-121	1,058
	H7C8F4	Apteronemobius asahinai	PERIOD, isoform 2	2e-121	1,054
	B7PDL3	Ixodes scapularis	Period circadian protein, putative	4e-118	1,013
	Q8MMG2	Blattella germanica	Circadian clock protein PERIOD	1e-114	1,007
	L0CAI9	Rhyparobia maderae	Period	2e-114	1,009
PERIODB (PERB)
LpPERB	B7PDL3	Ixodes scapularis	Period circadian protein, putative	6e-119	1,020
	Q8MMG2	Blattella germanica	Circadian clock protein PERIOD	8e-118	1,031
	A1EA98	Blattella bisignata	Circadian clock protein period	2e-115	1,008
	H7C8F3	Apteronemobius asahinai	PERIOD, isoform 1	2e-114	1,008
	H7C8F4	Apteronemobius asahinai	PERIOD, isoform 2	4e-114	1,004
PERIODC (PERC)
LpPERc	H7C8F3	Apteronemobius asahinai	PERIOD, isoform 1	4e-133	1,142
	H7C8F4	Apteronemobius asahinai	PERIOD, isoform 2	3e-132	1,134
	B7PDL3	Ixodes scapularis	Period circadian protein, putative	2e-129	1,095
	L0CAI9	Rhyparobia maderae	Period	9e-129	1,115
	A1EA98	Blattella bisignata	Circadian clock protein period	5e-124	1,071
TIMELESS (TIM)
LpTIM	J9Y3V3	Clunio marinus	Timeless	2e-180	1,464
	G9M9V8	Thermobia domestica	TIMELESS	4e-178	1,458
	H2D5T9	Aedes albopictus	Timeless	6e-165	1,370
	E0D5C1	Gryllus bimaculatus	TIMELESS	5e-164	1,359
	Q05E95	Aedes aegypti	TIMELESS	2e-161	1,349
CLOCK (CLK)
LpCLK	E2C765	Harpegnathos saltator	Circadian locomoter output cycles protein kaput	4e-164	1,276
	H9KJ84	Apis mellifera	Uncharacterized protein	3e-161	1,264
	E2AY34	Camponotus floridanus	Circadian locomoter output cycles protein kaput	3e-159	1,251
	F4WUR0	Acromyrmex echinatior	Circadian locomoter output cycles protein kaput	1e-156	1,233
	H9HL46	Atta cephalotes	Uncharacterized protein	4e-156	1,223
CYCLE 1 (CYC1)
LpCYC1	E0D6T3	Thermobia domestica	CYCLE	<1e180	1,786
	A0MH07	Lutzomyia longipalpis	Cycle	<1e180	1,622
	E2A3F0	Camponotus floridanus	Aryl hydrocarbon receptor nuclear translocator	<1e180	1,643
	A9XCF1	Tribolium castaneum	CYCLE	<1e180	1,621
	E2B7K2	Harpegnathos saltator	Aryl hydrocarbon receptor nuclear translocator	<1e180	1,620
CYCLE 2 (CYC2)
LpCYC2	E0D6T3	Thermobia domestica	CYCLE	<1e180	1,674
	E2A3F0	Camponotus floridanus	Aryl hydrocarbon receptor nuclear translocator	<1e180	1,606
	H9K2D4	Apis mellifera	CYC	<1e180	1,562
	A0MH07	Lutzomyia longipalpis	Cycle	<1e180	1,555
	A9XCF1	Tribolium castaneum	CYCLE	<1e180	1,560
CRYPTOCHROME 2 (npCRY2)
LpnpCRY2	E2BGC2	Harpegnathos saltator	Cryptochrome-1	<1e180	2,213
	L7YAB3	Solenopsis invicta	Cryptochrome	<1e180	2,210
	A8QSC1	Bombus impatiens	Cryptochrome 2 protein	<1e180	2,202
	L0C8K6	Rhyparobia maderae	Cryptochrome 2	<1e180	2,199
	F4WVC9	Acromyrmex echinatior	Cryptochrome-1	<1e180	2,199
Accessory Proteins
VRILLE (VRI)
LpVRI	E0VHE1	Pediculus humanus	Transcriptional factor nfil3/e4bp4, putative	1e-43	424
	B7PEG8	Ixodes scapularis	Transcriptional factor nfil3/e4bp4, putative	9e-43	417
	Q699T4	Antheraea pernyi	Vrille	6e-42	410
	Q1XD36	Danaus plexippus	Vrille	1e-41	408
	J9K7G7	Acyrthosiphon pisum	Uncharacterized protein	4e-40	403
CLOCK WORK ORANGE (CWO)
LpCWO	B7PGM4	Ixodes scapularis	Putative uncharacterized protein	1e-57	528
	B4LW47	Drosophila virilis	GJ23560	2e-42	430
	Q7QFU3	Anopheles gambiae	AGAP003844-PA	2e-41	422
	E9GEU4	Daphnia pulex	Putative uncharacterized protein	7e-41	395
	B4JIN1	Drosophila grimshawi	GH19128	2e-40	415
SUPERNUMERARY LIMBS1 (SLIMB1)
LpSLIMB1	L7M0T0	Rhipicephalus pulchellus	Uncharacterized protein	<1e180	2,159
	E9IHL0	Solenopsis invicta	Putative uncharacterized protein	<1e180	2,159
	D2XMQ7	Saccoglossus kowalevskii	Beta-TCRP E3 ligase	<1e180	2,145
	H9HC45	Atta cephalotes	Uncharacterized protein	<1e180	2,142
	D6WA15	Tribolium castaneum	Supernumerary limbs	<1e180	2,138
SUPERNUMERARY LIMBS2 (SLIMB2)
LpSLIMB2	L7M0T0	Rhipicephalus pulchellus	Uncharacterized protein	<1e180	2,153
	E9HMX3	Daphnia pulex	Putative uncharacterized protein	<1e180	2,150
	E9IHL0	Solenopsis invicta	Putative uncharacterized protein	<1e180	2,147
	H3JK44	Strongylocentrotus purpuratus	Uncharacterized protein	<1e180	2,135
	G3I574	Cricetulus griseus	F-box/WD repeat-containing protein 1A	<1e180	2,133
ARYL HYDROCARBON RECEPTOR NUCLEAR TRANSPORTER1 (ARNT1)
LpARNT1	C7B7E8	Litopenaeus vannamei	Hypoxia inducible factor 1 beta	<1e180	1,799
	E0VXW6	Pediculus humanus	Putative uncharacterized protein	<1e180	1,805
	L7MH74	Rhipicephalus pulchellus	Putative tango	<1e180	1,770
	D6W6H5	Tribolium castaneum	Putative uncharacterized protein	<1e180	1,739
	Q1JUI4	Daphnia magna	Aryl hydrocarbone receptor nuclear translocator	<1e180	1,723
ARYL HYDROCARBON RECEPTOR NUCLEAR TRANSPORTER2 (ARNT2)
LpARNT2	E0VXW6	Pediculus humanus	Putative uncharacterized protein	<1e180	1,900
	C7B7E8	Litopenaeus vannamei	Hypoxia inducible factor 1 beta	<1e180	1,858
	Q1JUI4	Daphnia magna	Aryl hydrocarbone receptor nuclear translocator	<1e180	1,850
	Q1JUI5	Daphnia magna	Aryl hydrocarbone receptor nuclear translocator	<1e180	1,838
	E9FQM5	Daphnia pulex	Putative aryl hydrocarbon receptor nuclear translocator	<1e180	1,838
CASIEN KINASE Iα (CKIα)
LpCKIα	Q8JG73	Danio rerio	Casein kinase 1alpha S	<1e180	1,570
	Q75WS8	Carassius auratus	Casein kinase I alpha S	<1e180	1,570
	Q8JGT0	Danio rerio	Casein kinase 1, alpha 1	<1e180	1,568
	Q75WS9	Carassius auratus	Casein kinase I alpha	<1e180	1,568
	B5THN0	Saccoglossus kowalevskii	Casein kinase 1 protein catalytic subunit	<1e180	1,567
CASIEN KINASE Iε (CKIε)
LpXKIε	F6YKD5	Equus caballus	Uncharacterized protein	<1e180	1,422
	G1SGL5	Oryctolagus cuniculus	Uncharacterized protein	<1e180	1,418
	G1U4F8	Oryctolagus cuniculus	Uncharacterized protein	<1e180	1,418
	M7AVG9	Chelonia mydas	Casein kinase I isoform epsilon	<1e180	1,416
	H0V5Y3	Cavia porcellus	Uncharacterized protein	<1e180	1,416
CASIEN KINASE IIα (CKIIα)
LpCKIIα	E2AW17	Camponotus floridanus	Casein kinase II subunit alpha	<1e180	1,674
	F4WU82	Acromyrmex echinatior	Casein kinase II subunit alpha	<1e180	1,666
	B7PVH2	Ixodes scapularis	Mitogen-activated protein kinase, putative	<1e180	1,671
	E2BH98	Harpegnathos saltator	Casein kinase II subunit alpha	<1e180	1,669
	K7GIG0	Pelodiscus sinensis)	Uncharacterized protein	<1e180	1,665
CASIEN KINASE IIβ (CKIIβ)
LpCKIIβ	Q71U52	Cyprinus carpio	CK2 beta subunit	5e-144	1,057
	Q6DEU1	Xenopus tropicalis	Casein kinase 2, beta polypeptide	5e-144	1,057
	Q1LXD2	Danio rerio	Casein kinase 2 beta	5e-144	1,057
	M4AI87	Xiphophorus maculatus	Uncharacterized protein	5e-144	1,057
	J3SBX9	Crotalus adamanteus	Casein kinase II subunit beta	5e-144	1,057
JETLAG (JET)
LpJET	K1P9I2	Crassostrea gigas	F-box only protein 37	1e-37	354
	C3Y7V1	Branchiostoma floridae	Putative uncharacterized protein	8e-37	356
	C3ZZX8	Branchiostoma floridae	Putative uncharacterized protein	6e-35	342
	R7VDT7	Capitella teleta	Uncharacterized protein	1e-32	326
	A7SQL5	Nematostella vectensis	Predicted protein	1e-31	320
SHAGGY (SGG)
LpSGG	E2C155	Harpegnathos saltator	Protein kinase shaggy	<1e180	1,478
	E9H6Q6	Daphnia pulex	Glycogen synthase kinase 3 beta	<1e180	1,475
	E9I8X0	Solenopsis invicta	Putative uncharacterized protein	<1e180	1,469
	G3MII6	Amblyomma maculatum	Putative uncharacterized protein	<1e180	1,453
	F4X3V9	Acromyrmex echinatior	Protein kinase shaggy	<1e180	1,467
PROTEIN PHOSPHOTASE 1α (PP1α)
LpPP1α	K1PXG6	Crassostrea gigas	Serine/threonine-protein phosphatase	<1e180	1,631
	A7RVJ0	Nematostella vectensis	Serine/threonine-protein phosphatase	<1e180	1,627
	E9HKA6	Daphnia pulex	Serine/threonine-protein phosphatase	<1e180	1,624
	H2L9G2	Oryzias latipes	Serine/threonine-protein phosphatase	<1e180	1,620
	H9KP59	Apis mellifera	Serine/threonine-protein phosphatase	<1e180	1,618
PROTEIN PHOSPHOTASE 1β (PP1β)
LpPP1β	L7MB61	Rhipicephalus pulchellus	Serine/threonine-protein phosphatase	<1e180	1,470
	H2UK23	Takifugu rubripes	Serine/threonine-protein phosphatase	<1e180	1,458
	H3DIM0	Tetraodon nigroviridis	Serine/threonine-protein phosphatase	<1e180	1,457
	H2MDU5	Oryzias latipes	Serine/threonine-protein phosphatase	<1e180	1,457
	G3P7R9	Gasterosteus aculeatus	Serine/threonine-protein phosphatase	<1e180	1,457
PROTEIN PHOSPHOTASE 2α-mts (PP2α-mts)
LpPP2α-mts	L7M975	Rhipicephalus pulchellus	Serine/threonine-protein phosphatase	<1e180	1,336
	B7QGY5	Ixodes scapularis	Serine/threonine-protein phosphatase	<1e180	1,335
	E0VKT2	Pediculus humanus	Serine/threonine-protein phosphatase	<1e180	1,333
	R4G8Q4	Rhodnius prolixus	Putative serine/threonine protein phosphatase	<1e180	1,330
	E9IAY8	Solenopsis invicta	Serine/threonine-protein phosphatase	<1e180	1,334
PROTEIN PHOSPHOTASE 2α-WBT (PP2α-WBT)
LpPP2α-WBT	L7M3M4	Rhipicephalus pulchellus	Putative serine/threonine protein phosphatase…	<1e180	2,137
	H9J639	Bombyx mori	Uncharacterized protein	<1e180	2,073
	E2AI01	Camponotus floridanus	Serine/threonine-protein phosphatase 2A 56 kDa regulatory subunit	<1e180	2,045
	H9K2X9	Apis mellifera	Uncharacterized protein	<1e180	2,044
	E0VNG0	Pediculus humanus	Serine/threonine-protein phosphatase 2A 56 kDa regulatory subunit	<1e180	2,041
PROTEIN PHOSPHOTASE 2α-TWS (PP2α-TWS)
LpPP2α-TWS	H9KM79	Apis mellifera	Uncharacterized protein	<1e180	2,081
	I3RJJ3	Scylla paramamosain	Protein phosphatase 2A regulatory subunit B	<1e180	2,074
	G3MLH6	Amblyomma maculatum	Putative uncharacterized protein	<1e180	2,070
	L7M886	Rhipicephalus pulchellus	Putative serine/threonine protein phosphatase	<1e180	2,063
	F4WRB8	Acromyrmex echinatior	Protein phosphatase PP2A 55 kDa regulatory subunit	<1e180	2,055
Input Pathway Protein
CRYPTOCHROME 1 (CRY1)†
LpCRY1	A2A264	Dianemobius nigrofasciatus	Cryptochrome	4e-132	1,022
	K1PRK6	Crassostrea gigas	Cryptochrome-1	3e-120	941
	R7V1J3	Capitella teleta	Uncharacterized protein	5e-119	937
	Q2TJN5	Danaus plexippus	Cryptochrome	5e-115	908
	J9T2S0	Mythimna separata	Cryptochrome 1	4e-112	888
Output Pathway Proteins
NEUROPEPTIDE F RECEPTOR (NPFR)†
LpNPFR	T1KR44	Tetranychus urticae	Uncharacterized protein	6e-96	759
	T1IW42	Strigamia maritima	Uncharacterized protein	7e-88	707
	B0WJ64	Culex quinquefasciatus	G-protein coupled receptor	3e78	642
	Q16SC4	Aedes aegypti	AAEL010626-PA	5e-78	636
	T1JTP7	Tetranychus urticae	Uncharacterized protein	1e-77	642
NEUROPEPTIDE F (NPF)
LpNPF	T1KF83	Tetranychus urticae	Uncharacterized protein	23e-12	156
	A0A087UNM1	Stegodyphus mimosarum	Uncharacterized protein	160e-9	128
	T1JM02	Strigamia maritima	Uncharacterized protein	670e-9	125
	F6KM62	Litopenaeus vannamei	Preproneuropeptide F I	0.0000013	123
	F6KM64	Melicertus marginatus	Preproneuropeptide F I	0.0000013	123
PIGMENT DISPERSING HORMONE RECEPTOR (PDHR)†
LpPDHR	K7JRN8	Nasonia vitripennis	Uncharacterized protein	2.4e-63	548
	U3U904	Nilaparvata lugens	Neuropeptide GPCR B2	95e-63	545
	Q7PQE3	Anopheles gambiae	AGAP003654-PA	830e-63	537
	A0A0P5AQT3	Daphnia magna	Class b secretin g-protein coupled receptor	590e-63	533
	A0A0P5CML9	Daphnia magna	Class b secretin g-protein coupled receptor	1.4e-60	533

^†indicates partial sequence.

We also identified orthologs of 15 accessory clock components (VRI, CWO, SLIMB, ARNT, CK Iα, CK Iε, CK IIα, CK IrlIβ, JET, SGG, PP1α, PP1β, PP2αMTS, PP2αWBT, PP2αTWS), including duplications of both the LpSLIMB and LpARNT genes, for a total of 17 accessory genes. Only one known accessory clock gene (PDP1) was not found in the L. polyphemus transcriptome. Top BLAST hits for all accessory clock genes are reported in Table 1. A reciprocal BLAST of the L. polyphemus orthologs revealed nearly all clock genes (VRI, CWO, SLIMB1, SLIMB2, ARNT1, ARNT2, CKIα, JET, SGG, PP1α, PP1β, PP2α-MTS, PP2α-MTS, PP2α-TWS) are most closely related to arthropod orthologs. However, CKIα, CKIε, and CKIIβ all show highest levels of homology to vertebrate orthologs. The top 5 BLAST hits of 10 of the17 accessory proteins included at least one known chelicerate ortholog, primarily from the tick family. Top BLAST hits for nearly all accessory clock components were significant, with e-values ranging from 10^−31 to less than 10^−180, and BLAST scores ranging from 320 to 2138 (only LpVRI, LpCWO, and LpJET score below 1000).

We were able to identify partial sequences for the input protein pCRY1, the NPF receptor (NPFR) and the PDH receptor (PDHR), as well as a complete sequence for the output neuropeptide NPF. PDH was not identified in the transcriptome of L. polyphemus. Top BLAST hits for all input and output clock genes are reported in Table 1. The top BLAST hits for LppCRY1 showed that LppCRY1 is most closely related to insect pCRY1 sequences. Despite only being a partial sequence this protein is highly conserved and had an e-value of 10^−132, with a BLAST score of 1022. BLASTp of LpNPF, LpNPFR, and LpPDHR showed closest homology to arthorpod orthologs, including several chelicerates. Top BLAST hits for most output clock components (excluding LpNPF) were significant, with e-values ranging from 10^−60 to 10^−150, and BLAST scores ranging from 533 to 1022. Top BLAST hits for LpNPF showed e-values of 10^−3 and BLAST scores of 84 and 86. The high e-values and low BLAST scores of the LpNPF BLAST are likely because NPF is a short peptide (96 amino acids long).

3.2. Comparisons with D. melanogaster and D. pulex clock genes

Top BLAST hits for all core clock genes against the D. melanogaster and D. pulex databases are reported in Table 2. All putative core clock components identified in this study, with the exception of npCRY2, (PERA, PERB, PERC, TIM, CLK, CYC1, and CYC2,) were found to have the highest homology to the orthologous proteins of D. melanogaster and D. pulex. LpnpCRY2 most closely matched a 6–4 photolyase in D. melanogaster, which is the closest relative to npCRY2 on the cryptochrome gene family tree. This mis-match is the product of the fact that the genus D. melanogaster lost its npCRY2 ortholog during its evolution. However, Apis mellifera retained its npCRY2 ortholog, and the top BLAST hit for LpCRY2 in Apis is npCRY2 (data not shown). Similarly, the closest match to LpnpCRY2 in the D. pulex protein database was “CRY-M”, which is another name for npCRY2. The top BLAST hits for D. melanogaster and D. pulex also showed high levels of homology to our genes of interest, with e-values ranging from 10^−68 to less than 10^−180 and BLAST scores ranging from 653 to 2139.

Table 2:

Top BLASTp hits for Drosophila melanogaster and Daphnia pulex.

Proteins	Species name	Protein name	UniProt ID	Blast Score	E-value
CORE PROTEINS
PERa	Drosophila melanogaster	Isoform PER-E	P07663	735	2.0×10–78
	Daphnia pulex	PER	E9GW67	822	1.0×10–90
PERb	Drosophila melanogaster	Isoform PER-E	P07663	653	7.0×10–68
	Daphnia pulex	PER	E9GW67	701	4.0×10–75
PERc	Drosophila melanogaster	Isoform PER-E	P07663	765	5.0×10–82
	Daphnia pulex	PER	E9GW67	876	1.0×10–97
TIM	Drosophila melanogaster	Timeless, isoform G	B7Z007	1,290	3.0×10–152
	Daphnia pulex	Putative TIMELESS/TIM-1 protein	E9FZ81	956	2.0×10–109
CLK	Drosophila melanogaster	Clk CLOCK jrk PAS1	O61735	1,048	1.0×10–125
	Daphnia pulex	CLOCK	E9GKD1	1,033	2.0×10–126
CYC1	Drosophila melanogaster	Protein cycle	O61734	1,220	6.0×10–159
	Daphnia pulex	CYCLE	E9FRH8	1,500	<1.0×10–180
CYC2	Drosophila melanogaster	Protein cycle	O61734	1,227	6.0×10–160
	Daphnia pulex	CYCLE	E9FRH8	1,470	<1.0×10–180
npCRY2	Drosophila melanogaster	phr6-4	Q8SXK5	1,296	3.0×10–171
	Daphnia pulex	CRY-M	E9GDJ9	2,139	<1.0×10–180
Accessory Proteins
VRI	Drosophila melanogaster	Vri	Q7KTN9	306	1.0×10–28
	Daphnia pulex	Vri†	E9HB85	280	1.0×10–28
CWO	Drosophila melanogaster	cwo-RA	B7FNP5	312	1.0×10–29
	Daphnia pulex	Putative uncharacterized protein	E9GEU4	395	2.0×10–42
SLIMB1	Drosophila melanogaster	Slimb	Q9VDE3	2,114	<1.0×10–180
	Daphnia pulex	f-box/wd-repeat protein*	E9HMX3	2,136	<1.0×10–180
SLIMB2	Drosophila melanogaster	Slimb	Q9VDE3	2,107	<1.0×10–180
	Daphnia pulex	f-box/wd-repeat protein*	E9HMX3	2,150	<1.0×10–180
ARNT1	Drosophila melanogaster	Aryl hydrocarbon receptor nuclear translocator homolog	O15945	1,568	<1.0×10–180
	Daphnia pulex	Putative aryl hydrocarbon receptor nuclear translocator	E9FQM5	1,723	<1.0×10–180
ARNT2	Drosophila melanogaster	Aryl hydrocarbon receptor nuclear translocator homolog	O15945	1,680	<1.0×10–180
	Daphnia pulex	Putative aryl hydrocarbon receptor nuclear translocator	E9FQM5	1,838	<1.0×10–180
CKIε	Drosophila melanogaster	Dco dbt	O76324	1,282	6.0×10–173
	Daphnia pulex	Casein kinase i alpha*	E9FS31	1,171	3.0×10–157
CKIα	Drosophila melanogaster	Casein kinase I isoform alpha	O76324	1,348	<1.0×10–180
	Daphnia pulex	Casien kinase I alpha*	E9HGM4	1,526	<1.0×10–180
CKIIα	Drosophila melanogaster	Casein kinase II alpha subunit, isoform C	P08181	1,590	<1.0×10–180
	Daphnia pulex	Casein kinase ii subunit alpha*	E9GCV0	1,659	<1.0×10–180
CKIIβ	Drosophila melanogaster	Casein kinase II beta subunit, isoform C	P08182–3	1,010	6.0×10–137
	Daphnia pulex	Casein kinase ii subunit beta*	E9GTE8	1,033	6.0×10–142
JET	Drosophila melanogaster	Jetlag, isoform B	Q0E8T8	188	3.0×10–15
	Daphnia pulex	f-box/leucine rich repeat protein*	E9G1Z9	155	2.0×10–10
SGG	Drosophila melanogaster	Isoform G of Protein kinase shaggy	P18431	1,357	<1.0×10–180
	Daphnia pulex	Glycogen synthase kinase 3 beta	E9H6Q6	1,475	<1.0×10–180
PP1α	Drosophila melanogaster	Pp1alpha-96A	P48461	1,575	<1.0×10–180
	Daphnia pulex	Serine/threonine-protein phosphatase	E9HKA6	1,624	<1.0×10–180
PP1β	Drosophila melanogaster	Serine/threonine-protein phosphatase beta isofrom	H5V895	1,386	<1.0×10–180
	Daphnia pulex	Serine/threonine-protein phosphatase	E9G7U7	1,427	<1.0×10–180
PP2a-MTS	Drosophila melanogaster	mts PP2A	P23696	1,313	<1.0×10–180
	Daphnia pulex	Serine/threonine-protein phosphatase	E9G8K4	842	5.0×10–110
PP2-WBT	Drosophila melanogaster	wdb	Q9VB23	1,986	<1.0×10–180
	Daphnia pulex	Serine/threonine-protein phosphatase kda regulatory subunit*	E9G2F2	1,932	<1.0×10–180
PP2-TWS	Drosophila melanogaster	tws aar Pp2A-85F	P36872	2,003	<1.0×10–180
	Daphnia pulex	TWS	E9GU66	1,974	<1.0×10–180
INPUT PATHWAY PROTEINS
CRY1†	Drosophila melanogaster	Cryptochrome-1	O77059	754	1.0×10–93
	Daphnia pulex	CRY-D	E9GSJ7	754	9.0×10–94
OUTPUT PATHWAY PROTEINS
NPFR†	Drosophila melanogaster	Isoform 6 of Neuropeptide F receptor	Q9VNM1	583	1.0×10–68
	Daphnia pulex	NPFG-protein-coupled receptor	E9GBE9	593	3.0×10–71
NPF	Drosophila melanogaster	Neuropeptide F	Q9VET0	84	5.0×10–3
	Daphnia pulex	Putative uncharacterized protein	E9GJI3	86	2.0×10–3
PDHR†	Drosophila melanogaster	PDF receptor	Q9W4Y2	445	1.4 ×10–48
	Daphnia pulex	Putative PDF receptor variant 2	E9FR28	532	6.3×10–62

^†indicates partial sequence.

^*indicates Protein name which originated from D. pulex genome map on Fleabase.

Top BLAST hits for all accessory clock genes against the D. melanogaster and D. pulex databases are reported in Table 2. All putative accessory clock components identified in this study were found to have the highest homology to the predicted orthologous proteins of D. melanogaster and D. pulex. Although on several occasions we found that the closest match to the L. polyphemus genes was an uncharacterized protein (as seen when LpCWO was blasted against the D. pulex database), or the name of the hit simply referred to the gene family (e.g. when the LpSLIMB1 and LpSLIMB2 genes matched the “f-box/wd-repeat protein” in D. pulex. Generally, these problems were only found in some D. pulex orthologs, and so the D. melanogaster (and many of the D. pulex) matches give reliable gene annotations in the BLAST results. With the exception of LpVRI, LpCWO and LpJET, all accessory proteins show significant levels of homology to their respective BLAST hits, with e-values ranging from 10^−110 to less than 10^−180 and BLAST scores ranging from 842 to 2136.

Top BLAST hits for all input and output clock genes against the D. melanogaster and D. pulex databases are reported in Table 2. All putative input and output clock components identified in this study (pCRY1, NPF, NPFR and PDHR) were found to have the highest homology to the predicted orthologous proteins of D. melanogaster and D. pulex. The top BLAST hits for the LppCRY1 showed pCRY1 and CRY-D (D. melanogaster-like cryptochrome also known as pCRY1) as the closest hits in our model systems. Similarly the top hits for all putative output genes (NPF, NPFR, and PDHR) for both D. melanogaster and D. pulex were the expected orthologs for both of these species. With the exception of LpNPF, all output proteins show significant levels of homology to their respective BLAST hits, with e-values ranging from 10^−48 to 10^−71 and BLAST scores ranging from 445 to 583. The high e-values and low BLAST scores of the LpNPF BLAST are likely because NPF is a short peptide (96 amino acids long), additionally the D. pulex NPF ortholog appears to have not yet been characterized as the top BLAST hit showed “Putative uncharacterized protein”.

3.3. Protein and domain alignment

The majority of the core clock components were more homologous to D. pulex core clock proteins than those of D. melanogaster (with the exception of LpCYC1 and LpCYC2 (Table 3; Figure 1 & 2; Supp Figure 1)). The similarity between Limulus clock proteins and those in D. pulex and D. melanogaster were most pronounced when comparing the functional domains of these clock proteins (Table 3). Moreover, the levels of conservation within the functional domains were fairly comparable between D. melanogaster and D. pulex, suggesting that the difference in sequence similarity lies in the regions outside and between domains (Table 3). The two clock protein domains with lowest levels of conservation were the PeriodC domains of all three LpPER paralogs (Table 3; Figure 3 & 4; Supp Figure 2) and the TIMELESS domain of LpTIM (Table 3; Figure 5; Supp Figure 3) (Table 3). LpnpCRY2 showed the highest level of conservation in the core clock proteins (Table 3; Figure 6). The full protein alignment of LpCYC2 to LpCYC1 shows a %Identity and %Similarity of 86% and 89%, respectively (Table 3; Figure 2). The three copies of the period genes also show high levels of homology to one another, with the alignments of the full LpPERA and LpPERB proteins showing a %identity and %similarity of 35% and 48%, LpPERA and LpPERC showing 37% and 49%, and LpPERC and LpPERB showing 65% and 72% (Figure 4). The duplicate per and cyc genes show a greater amount of similarity to their respective paralogs then they do to the respective orthologs of D. melanogaster and D. pulex (per (Figure 3 & 4; Supp Figure 2); cyc (Figure 1 & 2; Supp Figure 1)).

Table 3:

%Identity/%Similarity for L. polyphemus proteins against Drosophila melanogaster and Daphnia pulex.

Proteins		Drosophila melanogaster		Daphnia pulex
Core proteins	Length	Domains (%Identity/%Similarity)	Total	Domains (%Identity/%Similarity)	Total
PERA	1,001	PASa(40/53);PASb(44/63);PAC(66/82);PERIODC(N/A)	26/34	PASa(40/50);PASb(45/66);PAC(66/84);PERIODC(21/30)	30/38
PERB	1,021	PASa(37/56);PASb(44/55);PAC(75/86);PERIODC(N/A)	25/31	PASa(38/57);PASb(49/58);PAC(70/80);PERIODC(20/29)	32/39
PERC	1,047	PASa(38/56);PASb(45/65);PAC(68/82);PERIODC(N/A)	25/32	PASa(43/57);PASb(49/61);PAC(73/82);PERIODC(22/31)	31/39
TIM	1,100	TIMELESS(35/45)	27/38	TIMELESS(25/40)	36/45
CLK	654	bHLH(67/82);PASa(52/61);PASb(78/82);PAC(75/93)	32/40	bHLH(71/84);PASa(55/66);PASb(76/85);PAC(84/91)	43/50
CYC1	623	bHLH(75/81);PASa(75/87);PASb(53/73);PAC(57/66)	62/68	bHLH(78/85);PASa(87/94);PASb(57/66);PAC(57/70)	31/37
CYC2	632	bHLH(76/81);PASa(76/88);PASb(53/66);PAC(52/66)	60/66	bHLH(76/85);PASa(85/93);PASb(59/68);PAC(57/70)	31/38
npCRY2	549	6–4 Photolyase(N/A);FAD Binding 7(N/A)	N/A	6–4 Photolyase(64/76);FAD Binding 7(84/91)	70/79
Accessory Proteins
VRI	439	BRLZ(69/80)	20/28	BRLZ(68/75)	42/43
CWO	536	bHLH(82/86);ORANGE(32/39)	16/22	bHLH(88/95);ORANGE(34/46)	23/25
SLIMB1	533	b-TrCP D(55/65);FBOX(85/93);WD40(92/92);WD40(84/98); WD40(97/97);WD40(94/94);WD40(97/100);WD40(92/97); WD40(95/95)	73/78	b-TrCP D(53/65);FBOX(75/85);WD40(92/95);WD40(84/95); WD40(87/87);WD40(95/97);WD40(95/100);WD40(89/97); WD40(100/100)	74/80
SLIMB2	528	b-TrCP D(57/68);FBOX(88/93);WD40(92/92);WD40(84/89); WD40(97/97);WD40(95/95);WD40(97/100);WD40(92/97); WD40(92/95)	74/79	b-TrCP D(57/68);FBOX(78/85);WD40(92/92);WD40(84/95); WD40(87/87);WD40(95/97);WD40(95/100);WD40(89/97); WD40(97/100)	76/82
ARNT1		bHLH(76/81);PASa(74/90);PASb(74/83);PAC(61/75)	49/57	bHLH(76/81);PASa(74/88);PASb(75/78);PAC(84/93)	52/60
ARNT2		bHLH(89/91);PASa(76/90);PASb(75/84);PAC(57/70)	49/57	bHLH(89/91);PASa(78/88);PASb(74/78);PAC(84/91)	54/59
CK Iα	402	S TKc(78/84)	79/84	S TKc(64/71)	46/52
CK Iε		S TKc(75/81)	53/58	S TKc(73/77)	64/67
CK IIα	329	S TKc(92/96)	88/92	S TKc(94/98)	86/90
CK IIβ	198	CKIIβ(90/93)	84/86	CKIIβ(93/95)	84/86
JET	234	LRR(12/35);LRR(15/41);LRR(31/50)	41/48	LRR(15/38);LRR(33/48);LRR(31/46)	14/19
SGG	364	S TKc(77/85)	50/57	S TKc(83/88)	69/75
PP1α	330	PP2Ac(94/97)	88/91	PP2Ac(96/99)	90/93
PP1β	274	PP2Ac(86/88)	76/79	PP2Ac(88/89)	81/82
PP2αMTS	296	PP2Ac(79/82)	80/84	PP2Ac(54/65)	51/61
PP2αWBT	494	B56(84/91)	71/78	B56(81/89)	74/82
PP2αTWS	443	WD40(85/97);WD40(90/93);WD40(93/98);WD40(93/95); WD40(87/90); WD40(93/94);WD40(87/92)	73/79	WD40(79/87);WD40(88/100);WD40(93/98);WD40(80/85); WD40(87/92); WD40(87/89);WD40(82/89)	84/91
Input Pathway Proteins
CRY1	313†	FAD Binding 7(47/55)	N/A	FAD Binding 7(45/60)	N/A
Output Pathway Proteins
NPFR	213†	7tM GPCR Srsx(53/66)	N/A	7tM GPCR Srsx(49/61)	N/A
NPF	95	PAH(36/54)	23/34	PAH(41/54)	29/37
PDHR	231 †	7TM-2(40/57)	N/A	7TM-2(46/63)	N/A

^†indicates partial sequence.

Figure 1:

bHLH, PASa, PASb and PAC domains region of putative L. polyphemus CYCLE1 (CYC1) protein. Alignment of L. polyphemus (Lim) CYC1 to Daphnia pulex (Daph) CYC and Drosophila melanogaster (Dros) CYC using CLUSTAL Omega. Symbols immediately below sequence indicates functional similarity of residues between protein orthologs. “*” Indicates identical residues, “:” Indicates residues with strongly similar properties, “.” Indicates residues with weakly similar properties. Blue shows bHLH domain, light green shows PAS domains, Red shows PAC domain. For full alignment see Supplemental Figure 4.

Figure 2:

Putative L. polyphemus CYCLE 1 (CYC1) protein aligned against putative CYCLE 2 (CYC2) using CLUSTAL Omega. Symbols immediately below sequence indicates functional similarity of residues between protein orthologs. “*” indicates identical residues, “:” indicates residues with strongly similar properties, “.” indicates residues with weakly similar properties. Light green shows bHLH domain, Blue shows PAS domain, Red shows PAC domain.

Figure 3:

PASa, PASb and PAC domains region of putative L. polyphemus PERIOD (PER) protein. Alignment of L. polyphemus (Lim) PER to Daphnia pulex (Daph) PER, D. melanogaster(Dros) PER, Ixodes scapularis (Ix) PER using CLUSTAL Omega. Symbols immediately below sequence indicates functional similarity of residues between protein orthologs. “*” Indicates identical residues, “:” Indicates residues with strongly similar properties, “.” Indicates residues with weakly similar properties. Light green shows PAS domain, Red shows PAC domain. For full alignment see Supplemental Figure 1.

Figure 4:

Putative L. polyphemus PERIOD (PER) proteins. Alignment of L. polyphemus (L. polyphemus) PERa to L. polyphemus PERb and PERc using CLUSTAL Omega. Symbols immediately below sequence indicates functional similarity of residues between protein orthologs. “*” Indicates identical residues, “:” Indicates residues with strongly similar properties, “.” Indicates residues with weakly similar properties. Green shows bHLH domain, Blue shows PAS domain, Red shows PAC domain, Grey shows Period C domain.

Figure 5:

TIMELESS domain region of putative L. polyphemus TIMELESS (TIM) protein. Alignment of L. polyphemus (Lim) TIM to Daphnia pulex (Daph) TIM and D. melanogaster(Dros) TIM using CLUSTAL Omega. Symbols immediately below sequence indicates functional similarity of residues between protein orthologs. “*” indicates identical residues, “:” indicates residues with strongly similar properties, “.” Indicates residues with weakly similar properties. Dark Blue shows TIMELESS domain. For full alignment see Supplemental Figure 2.

Figure 6:

Putative L. polyphemus CRYPTOCHROME 2 (CRY2) protein. Alignment of L. polyphemus (Lim) CRY2 to Daphnia pulex (Daph) CRY-M and Apis mellifera (Apis) CRY2 using CLUSTAL Omega. Symbols immediately below sequence indicates functional similarity of residues between protein orthologs. “*” Indicates identical residues, “:” Indicates residues with strongly similar properties, “.” Indicates residues with weakly similar properties. Dark green shows Photolyase domain, Gold shows FAD Binding 7 domain.

The sequence alignments of the accessory proteins were similar to those of the core proteins. For example, the sequence alignment of the duplicated LpSLIMB paralogs shows an overall %identity/%similarity value of 94/97, and the paralogs of the LpARNT show a value of 85/89. The paralogs of both genes share higher levels of sequence identity to one another than they do to D. melanogaster and D. pulex orthologs. Moreover, the majority of the L. polyphemus orthologs have higher sequence identity to D. pulex then to D. melanogaster, with the higher levels of conservation within the functional domains (Table 3). We found that LpVRI, LpCWO, and LpJET had the lowest levels of conservation in the accessory proteins, but higher levels of conservation in the functional domains of LpVRI and LpCWO (Table 3). Interestingly, we found that the functional domains within the LpJET appear to have less conservation than the non-domain regions (Table 3). Additionally, the kinases and phosphatases tended to be noticeably more conserved than the other accessory proteins, with the most conserved protein (PP1α) having a %Identity/%Similarity of 88/91 for D. melanogaster and 90/93 for D. pulex (Table 3).

The sequence alignments for the input and output proteins showed high levels of homology for both LppCRY1, LpNPFR, and LpPDHR, although the sequence identity for LpNPF was much lower. All of the output proteins had higher levels of homology within the functional domains.

3.4. Phylogenetic trees

All the core clock components for L. polyphemus identified in this study (PERA, PERB, PERC, TIM, CLK, CYC1, CYC2, and npCRY2) clearly fall into the appropriate clades for each gene family (Figure 8, 9, 10 & 11). We also found that whenever the protein sequence for a chelicerate ortholog was available, the L. polyphemus ortholog(s) most closely associated with it. Furthermore, all the chelicerate proteins tended to nest within the clades consisting of arthropod orthologs, but always formed a distinct clade outside the insect and crustacean clades. Within the period gene tree the three paralogs of the L. polyphemus period gene probably originated from two novel duplication events that occurred after the formation of the subphylum chelicerata and after the divergence of Ixodes scapularis (Black legged tick, and a fellow chelicerate) and L. polyphemus (Figure 8). Furthermore, the periodB/periodC duplications probably occurred more recently than the duplication event involving periodA and the common ancestor of periodB and periodC. Similarly, the duplication of the cyc gene also appears to be specific to the subphylum chelicerata (Figure 10). However, without the sequences of any other chelicerate orthologs it is not possible to determine when in the evolution of the chelicerates this duplication occurred. Additionally the phylogenetic tree of the cry gene family showed that LpCRY1 nests within the invertebrate-cry1 clade of the cryptochrome gene family.

Figure 8:

L. polyphemus PERIOD phylograms. L. polyphemus homologs are marked with an “*”, chelicerates shown in red, while all other arthropods are shown in blue, and all vertebrates are labeled in black, the root is a fungi ortholog and is labeled in violet. Major clades are bracketed and labeled. Branch lengths correspond to distance bar on bottom left corner. See Supplemental figure 5 for Cladogram.

Figure 9:

L. polyphemus TIMELESS phylograms. L. polyphemus homologs are marked with an “*”, chelicerates shown in red, while all other arthropods are shown in blue, and all vertebrates are labeled in black, the root is a fungi ortholog and is labeled in violet. Major clades are bracketed and labeled. Branch lengths correspond to distance bar on bottom left corner. See Supplemental figure 6 for Cladogram.

Figure 10:

L. polyphemus CLOCK/CYCLE phylograms. L. polyphemus homologs are marked with an “*”, chelicerates shown in red, while all other arthropods are shown in blue, and all vertebrates are labeled in black, the root is a fungi ortholog and is labeled in violet. Major clades are bracketed and labeled. Branch lengths correspond to distance bar on bottom left corner. See Supplemental figure 7 for Cladogram.

Figure 11:

L. polyphemus CRYPTOCHROME phylograms. L. polyphemus homologs are marked with an “*”, chelicerates shown in red, while all other arthropods are shown in blue, and all vertebrates are labeled in black, the root is a fungi ortholog and is labeled in violet. Major clades are bracketed and labeled. Branch lengths correspond to distance bar on bottom left corner. See Supplemental figure 8 for Cladogram.

In the accessory gene group we found two duplicated genes slimb and arnt. Phylogenetic analysis of the LpSLIMB paralogs show that both copies of LpSLIMB nest within the slimb clade (Figure 12). Similarly phylogenetic analysis of the duplicate LpARNT paralogs show that these protein nest within the invertebrate arnt clade (Figure 10). Both of these duplication events appear to have occurred after the divergence of chelicerates, and have not been identified in any other species.

Figure 12:

L. polyphemus SLIMB phylograms. L. polyphemus homologs are marked with an “*”, chelicerates shown in red, while all other arthropods are shown in blue, and all vertebrates are labeled in black, the root is a fungi ortholog and is labeled in violet. Major clades are bracketed and labeled. Branch lengths correspond to distance bar on bottom left corner. See Supplemental figure 9 for Cladogram.

4. Discussion.

4.1. Identification of putative circadian proteins in the L. polyphemus de novo transcriptome assembly.

Many arthropods have been shown to exhibit robust physiological and behavioral rhythms (Palmer, 1973). However, the molecular basis of many of these rhythms has been elucidated in a few model species, such as D. melanogaster (Allada and Chung, 2010). While some progress has been made in understanding the architecture of these clocks in non-model insect and crustacean species (Rubin et al., 2006; Tomioka and Matsumoto, 2010; Yuan et al., 2007), little effort has been put forth the investigate the clock mechanisms in chelicerates, including L. polyphemus. Based on the data obtained in this study, it appears that horseshoe crabs possess endogenous clocks that may rely on many of the same molecular components as other, better understood, invertebrate circadian clocks. Moreover, given the fact that horseshoe crabs have been extant for at least 450 million years (Rudkin and Young, 2009), these data might also provide some insight into the evolution of endogenous clocks.

In this study we identified 29 orthologs of circadian clock related proteins in the L. polyphemus transcriptome, corresponding to insect core clock proteins (PER, TIM, CLK, CYC and npCRY2), accessory proteins (VRI, CWO, SLIMBS, ARNT, CKIa, CKIe, CKIIa, CKIIß, JETLAG, SHAGGY, PPT1a, PPT1ß, PPT 2a-MTS, PPT 2a-WBT, and PPT2a-TWS), input proteins (pCRY1), and output proteins (NPF, its receptor (NPFR), and PDHR). Two of the top 5 BLAST hits for LpCYC1, and 1 top hit for LpCYC2 were labeled as “ARNT” proteins. BLASTp of the LpnpCRY2 showed closest homology to proteins labeled pCRY1 and 1 protein labeled as generic “CRY” protein. Despite the mislabeling of these genes, phylogenetic evidence suggests that these BLAST hits are, in fact, orthologs of npCRY2 and CYC, and nest within the appropriate clades of the cryptochrome and bHLH gene families, respectively (data not shown). The mislabeling of the CYC and npCRY2 genes appears to be the result of an inappropriate reference genome during genome annotation (using a mammalian reference instead of an insect). Only two proteins found in insect circadian systems were not identified in the L. polyphemus transcriptome (PDP1e, and PDH). Even though we were unable to identify PDH in the L. polyphemus transcriptome, the presence of its receptor (PDHR) may imply that LpPDH is likely expressed in the L. polyphemus CNS. The absence of LpPDH from the transcriptome is possibly due to the relatively small size of the preprohormone (D. melanogaster - 102 amino acids – Renn et al., 1999; Crayfish - deKleijn et al., 1993). Among the proteins that were identified, the full-length protein sequences were able to be determined for all but LppCRY1 and LpNPFR, and LpPDHR.

The identification of both pCRY1 and npCRY2 orthologs allows us to infer some additional information regarding the architecture of the circadian clock in L. polyphemus, based on what is known about insect clocks. Insect clock systems have been categorized into three types: 1) type 1, which contains cry1 but lacks cry2 (such as in D. melanogaster melanogaster (Hardin, 2005)); 2) type 2, which contains both cry1 and cry2 (such as in the monarch butterfly, Danaus plexipus (Zhu et al., 2005)) and; 3) type 3, which contains cry2 but lacks cry1 (such as in the honeybee, Apis mellifera; (Yuan et al., 2007)). The presence of both cry1 and cry2 in the L. polyphemus transcriptome would classify L. polyphemus as having a type 2 clock system which is considered an “ancestral clock system” (Yuan et al., 2007), which is fitting because L. polyphemus is considered to be a “living fossil” (Rudkin and Young, 2009). Thus, the circadian system of L. polyphemus is most likely organized more like the butterfly and mosquito (Yuan et al., 2007; Zhu et al., 2005), than D. melanogaster (Hardin, 2005) or Apis mellifera (Yuan et al., 2007).

4.2. Conservation and evolution of circadian proteins.

The conservation of the various circadian clock proteins ranged from poor to highly conserved. Nearly all genes showed higher levels of conservation between D. melanogaster and D. pulex than orthologs of either species did to L. polyphemus orthologs. These findings are consistent with the evolutionary relationships of these three subphyla, with insects and crustaceans being more closely related to one another than to chelicerata (Giribet and Edgecombe, 2012). Conservation within functional domains tended to be much higher, in general, indicating that the functions of these proteins within the circadian clock may be well conserved between insects and chelicerata, as well as Crustacea and chelicerata. As such, the proteins which makeup the circadian clock in D. melanogaster may be relatively well conserved across the entire phylum Arthropoda. Interestingly, several domains (PeriodC domain of all three LpPER paralogs, the Timeless domain of LpTIM, and the LRR domains of LpJET) did show surprisingly low levels of conservation. The lack of conservation in these functional domains may confer novel functions of these proteins in the circadian clock of L. polyphemus. For example, the PeriodC domain is a known DNA-binding domain, and the overabundance of mutations in this domain may facilitate new transcriptional targets in the clock cells. Similarly the Timeless and LRR domains are found to be involved in protein-protein interactions, and the high level of mutations in this domain could allow for previously unknown interactions, thus allowing for new clock mechanisms.

Most (72%) L. polyphemus proteins (19 out of the 26, [pCRY1, NPFR, and PDHR were excluded due to incomplete sequences]) showed higher homology to D. pulex orthologs than to D. melanogaster orthologs. Six proteins (CYC1, CYC2, CKIα, CKIIα, JET, and PP2αMTS) show higher levels of sequence identity to D. melanogaster orthologs, and one gene (CKIIβ) showed equal levels of sequence identity to both D. melanogaster and D. pulex. The finding that of 72% of genes share higher sequence identity with D. pulex may be due to the fact that D. pulex and L. polyphemus evolved in similar marine habitats and thus experienced similar environmental pressures. There were six L. polyphemus genes, which show higher homology to D. melanogaster orthologs than to D. pulex. Most of these genes show low levels of conservation in the D. pulex orthologs, which may suggest that these genes may have undergone rapid evolution in the crustacean lineage

Several gene duplications were identified from the L. polyphemus transcriptome. The two duplications of the period gene in L. polyphemus appear to be a novel discovery in invertebrates, and while there have been several duplications of the period gene in vertebrates (Bae et al., 2001), phylogenetic evidence suggests that the duplications this study revealed are unique to the chelicerate lineage, and independent of the vertebrate duplication events. These three paralogs raise an interesting question regarding the possibility of functional divergence following the duplication. Such differences have been observed between two period alleles in D. melanogaster, where each allele has different thermokinetics and varies by latitude (Sawyer et al., 1997). More extreme functional divergences have also been observed following the duplications of the period genes during vertebrate evolution (Bae et al., 2001). The single duplication events of the cycle, arnt, and slimb genes also appear to be unique and previously unknown duplications. These results present intriguing opportunities for potential novel mechanisms and functions within the clock systems of L. polyphemus. Are these different protein isoforms differentially expressed in different cell-types? Could these duplicated genes have been co-opted to create a clock controlling circatidal rhythms in L. polyphemus? Since these duplications do not appear in other chelicerates, such as scorpions and ticks, it is tempting to consider that they might be part of a clock system controlling these rhythms.

The various gene duplications also support the recent discovery of a whole genome duplication event based on genetic mapping of the L. polyphemus genome (Nossa et al., 2014). This genome duplication event appears to have taken place roughly 300 million years ago (Nossa et al., 2014), well after the divergence of Limulidae from the other chelicerate families (~450 million years ago (Rudkin and Young, 2009). This is consistent with our findings, as none of the observed duplications have been found in other species outside the family Limulidae. Given the evidence of at least one genome duplication event, what is even more intriguing than the presence of these duplications is the absence of duplications of the other circadian clock genes. The absence of these duplicate genes following a genome duplication event can have one of two explanations: 1) these genes underwent a loss of function, resulting in the duplicate genes being deleted from the genome or rendered inert, or 2) these genes underwent a functional divergence, resulting in these genes not being expressed in the CNS. In either case we would not expect to find transcripts of these duplicates in our transcriptome assembly.

4.3. Implications and future research.

This study provides the first insight into the molecular mechanisms underlying the biological clocks of L. polyphemus, a member of the subphylum chelicerata. The conservation of these proteins, particularly within functional domains, provides evidence that some or all of these proteins may be involved in the circadian clock, and possibly the circatidal clock, of L. polyphemus. Immunolabelling of putative circadian proteins has been instrumental in localizing circadian clock cells (Siwicki et al., 1988) and their output cells in a variety of insects (Sehadova et al., 2004; Shao et al., 2006) and this study provides the foundation for similar studies in L. polyphemus. This study also provides a resource to aid in the development of qPCR primers and probes for Southern blotting and in situ hybridization. Moreover, the identification of these genes may also allow us to begin to tease apart the mechanisms underlying circatidal/circalunidian clocks (Chabot et al., 2004; Chabot and Watson, 2010), the molecular bases of which are unknown even in model systems (Tessmar-Raible et al., 2011).

Biological clocks act as master regulators of the physiology and behavior of most organisms. The high level of conservation of circadian clock related genes in Limulus indicates that they may play a role in the biological clock system(s) of this species, thus it is likely that the basic architecture of the circadian clock found in insects dates back prior to the divergence of chelicerata on the arthropod phylogenetic tree.

Supplementary Material

Supplementary Figure 4

Supplemental Figure 4: Putative L. polyphemus CLOCK (CLK) protein. Alignment of L. polyphemus (Lim) CLK to Daphnia pulex (Daph) CLK, D. melanogaster(Dros) CLK, Ixodes scapularis (Ix) CLK using CLUSTAL Omega. Symbols immediately below sequence indicates functional similarity of residues between protein orthologs. “*” Indicates identical residues, “:” Indicates residues with strongly similar properties, “.” Indicates residues with weakly similar properties. Blue shows bHLH domain, Light green shows PAS domain, Red shows PAC domain, Magenta shows Coiled Coil domain.

Supplementary Figure 1

Supplemental Figure 1: Putative L. polyphemus CYCLE1 (CYC1) protein. Alignment of L. polyphemus (Lim) CYC1 to Daphnia pulex (Daph) CYC and Apis mellifera (Apis) CYC using CLUSTAL Omega. Symbols immediately below sequence indicates functional similarity of residues between protein orthologs. “*” Indicates identical residues, “:” Indicates residues with strongly similar properties, “.” Indicates residues with weakly similar properties. Blue shows bHLH domain, light green shows PAS domains, Red shows PAC domain.

Supplementary Figure 2

Supplemental Figure 2: Putative L. polyphemus PERIOD (PER) protein. Alignment of L. polyphemus (Lim) PER to Daphnia pulex (Daph) PER, D. melanogaster(Dros) PER, Ixodes scapularis (Ix) PER using CLUSTAL Omega. Symbols immediately below sequence indicates functional similarity of residues between protein orthologs. “*” Indicates identical residues, “:” Indicates residues with strongly similar properties, “.” Indicates residues with weakly similar properties. Light green shows PAS domain, Red shows PAC domain, Grey shows Period C domain.

Supplementary Figure 3

Supplemental Figure 3: Putative L. polyphemus TIMELESS (TIM) protein. Alignment of L. polyphemus (Lim) TIM to Daphnia pulex (Daph) TIM and D. melanogaster(Dros) TIM using CLUSTAL Omega. Symbols immediately below sequence indicates functional similarity of residues between protein orthologs. “*” indicates identical residues, “:” indicates residues with strongly similar properties, “.” Indicates residues with weakly similar properties. Dark Blue shows TIMELESS domain.

Figure 7:

bHLH, PASa, PASb and PAC domains region of putative L. polyphemus CLOCK (CLK) protein. Alignment of L. polyphemus (Lim) CLK to Daphnia pulex (Daph) CLK, D. melanogaster(Dros) CLK, Ixodes scapularis (Ix) CLK using CLUSTAL Omega. Symbols immediately below sequence indicates functional similarity of residues between protein orthologs. “*” Indicates identical residues, “:” Indicates residues with strongly similar properties, “.” Indicates residues with weakly similar properties. Blue shows bHLH domain, Light green shows PAS domain, Red shows PAC domain. For full alignment see Supplemental Figure 3.