How Cryptochrome and Photolyase Proteins Have Been Gained and Lost Over Evolution

6-4 PLs are characterized by their α helices (α8, α10, α12, α14, α15, α16, α17, α18) and the sulfur loop (sl). In addition, 76% of them also possess the plant N-terminal motif (pn). All motifs found in 6-4 PLs are also abundant in the other CRY/PLs belonging to the 6-4 cluster, especially in MCRY and in CRY4 sequences. The motif analysis revealed that the most common motif among all members of the CRY/PL family is the α helix10 (α10) (Figure 3). This is strongly conserved across all CRY/PL members. Only in CPDI photolyases we found it rarely (38%). In our analysis, we found a novel 6-4 PL specific motif (YIYEPWKAP) in their C-terminal tail. It is present in 81% of 6-4 PL sequences and 73% of DCRY sequences, while only 44% of MCRYs and 21% of CRY4 sequences feature it (not shown in Figure 3 as < 50%). CRY4s are very similar to the 6-4 PLs in terms of their motifs. In addition to the previously described 4c and 4Y motifs, we also identified two new CRY4 characteristic motifs: 4a and 4b. All, except 4a, are almost absent in other subfamilies. Only 4a occurs in some MCRY sequences with a probability of 36% (not plotted).

MCRY is the most conserved subfamily in our analysis as all major motifs, especially the 6-4 PL specific helices, occur very frequently. All α helices and the sulfur loop occur in MCRY subfamily with higher frequencies than in the 6-4 PL subfamily. In addition, we also found the MCRY specific motifs second pocket up (m2pu), second pocket low (m2pl), phosphate motif (mpm), protrusion loop (ml), C-terminal lid (mcl) and C-terminal tail (mct). These MCRY-specific motifs, on the other hand, are not as strongly conserved, but feature characteristic differences between duplications and species. We will discuss these in more detail later.

The α10 motif is the most conserved motif from the Drosophila-type DCRY subfamily (90% frequency). In addition, most DCRY sequences feature further 6-4 PL motifs such as α-helices α14 (78%) and α17 (69%) and the plant N-terminal motif (pn) (76%). Besides the known DCRY typical sequences second pocket up (d2pu), second pocket low (d2dpl) and the C-terminal lid (dcl), we discovered in our analysis a new DCRY specific motif (dm1, HTLWxP), which appears in 70% of the investigated DCRY sequences. Most ACRY sequences are duplicated and, with about 1000 amino acids, form the largest members of the CRY/PL family in terms of protein size. Like MCRY and DCRY, ACRY also feature 6-4 PL motifs (α8, α10, α14, and α16) as well as the plant N-terminal motif (pn). The above motifs occur at least once (very often twice) in the protein sequence. The second plant N-terminal motif and the second α14-motif occur in all ACRY sequences (Figure 3).

Many CPDII sequences also contain the α10 motif (90%), though the subfamily-typical motifs CPDII conserved1 (cc1) and CPDII mitochond. target (mts) occur more frequently. The mts motif is the most common, accounting for 98%. CPDII conserved2 (cc2) and CPDII conserved3 (cc3) are not as highly conserved with a frequency of 85% and 64%. CPDI photolyases are very different from the other members of the CRY/PL family. The two highly conserved motifs α10 and the plant N-terminus occur in only 38% and 10% of all CPDI sequences, respectively. Similarly, the CPDI/III-specific motifs are hardly found in other members of the CRY/PL subfamily, except CPDIII PLs and DASH-CRYs. While CPDIII conserved (c3) and CPDI/III conserved2 (c3-2) are relatively highly conserved in the subfamily (75% and 78%), CPDI/III conserved1 (c1/3-1) is less conserved with a frequency of 60%. In contrast most of the CPDIII sequences possess plant N-terminal (78%) and the α10 (67%). The c3 motif is most conserved as all tested CPDIII sequences possess it, while the c1/3-1 motif is only found in 22% CPDIII sequences (not shown).

The most common motif in the PCRY sequences is DAS motif1 (pd1) with a frequency of 97%. The plant N-terminal motif occurs at 93% in all analyzed sequences. The other PCRY-typical motifs, plant CRY conserved (pc) and plant CRY1 conserved (pc1), occur with 90% frequency. Only plant DAS motif2 (pd2) occurs less frequently at only 55%. None of the PCRY-specific motifs occur in PCRY-like sequences. PCRY-like cryptochromes possess only three known CRY/PL motifs. These are α10 (90%), α14 (85%) and the plant N-terminal (77%) (). Both are generally among the most frequent motifs in the entire CRY/PL family. In our study, we identified three novel PCRY-like specific motifs (pl1, pl2, and pl3). PCRY-like 1 (pl1) consists of only four amino acids VGLR and occurs with a frequency of 74% in the PCRY-like subfamily. PCRY-like 2 (pl2, sequence: WRDLAY) is the second most common with 83%. The third motif PCRY-like 3 (pl3, sequence: MWQNxG) is the most conserved with 97%. All three motifs barely occur in other subfamilies. Supplementary Materials S4

The DASH-CRY typical motifs are D-R salt bridge2 (db2, 93%), CPD-lesion repair (dlr, 73%) and D-R salt bridge1 (db1, 75%). In addition, DASH sequences have both the 6-4 photolyase-specific motifs α10 (73%) and α8 (52%), and c3, which is typical of CPDI/III photolyases, at a frequency of 88%. The DASH CPD lesion repair (dlr) also occurs in the plant photolyases PPL subfamily with a probability of 80%. All PPL sequences possess the α10 motif. Furthermore, we found by alignment and consensus sequence analyses three novel PPL specific motifs ppl1 (85%, ESxSxxPVYCFDPR), ppl2 (75%, BSJYGANFSCKI), ppl3 (85%, FFRFxTxK) (Figure 3, Supplementary Materials S3). Except for ppl1 that also occurs with 2% frequency in the DASH subfamily, the motifs ppl2 and ppl3 are highly subfamily-specific allowing strict differentiation from DASH-CRY sequences clustering very closely in the phylogenetic tree.

Photolyases are ancient DNA repair enzymes thought to have evolved before atmospheric oxygenation and the formation of the ozone layer [19,20,45,46]. They were crucial at those times as intense UV radiation could reach the surface of the Earth without attenuation by the ozone layer [45,47]. Indeed, we found all four photolyases in Archaea and Bacteria with CPDII being the most abundant in Archaea and CPDI the most abundant in Bacteria. Most procaryote species possess only one photolyase, but E. archaeon has the 6-4 and CPDIII photolyases together, and several bacteria possess the 6-4 and the CPDII photolyases together with the CPDI photolyase. Interestingly, the CPDI and CPDIII photolyases never occur simultaneously, and this also applies to all eukaryotes examined in this study. Zhang et al. (2013), assumed that a prokaryotic 6-4 photolyase with an iron-sulfur cluster (FeS-BCP) is the first common ancestor of photolyases [48]. In contrast, the recent study by Xu et al. (2021), considers this unlikely, as functional changes from 6-4 to CPD repair mechanisms must then have occurred multiple times during evolution [49]. Furthermore, they emphasize that CPDs are the major UV-induced DNA lesions and therefore it is more likely that CPD photolyases are the most ancient. They consider a new protein class of short photolyase-like (SPL) proteins similar to CPD photolyases, as possible origin of photolyases and thus of the CRY/PL family [49]. Our study confirms the special role of CPD photolyases as they are the most abundant across taxa. We also confirm the common hypothesis that photolyases are the origin of the CRY/PL family as we find all photolyases in all three domains (Table 1, Figure 7).

DASH-CRYs are probably the first cryptochromes within the CRY/PL family [25]. Among archaea, only the extremely halophilic archaeon Halorhabdus utahensis (not included in our study) appears to have a DASH-CRY [19], but DASH-CRYs are widely distributed in the domains of Bacteria (especially Cyanobacteria) and Eukaryota. In addition to the DASH-typical motifs, DASH-CRYs possess the c3 motif, which is found only in CPDI/III photolyases and their relatives. Because of this structural similarity and because of their proximity to the CPDI&III branch in the phylogenetic tree (Figure 2), we hypothesize that DASH-CRYs are closely related to CPDI/III photolyases and have evolved from a common precursor, forming sister clades. The fact that at least some DASH-CRYs are also able to repair CPD lesions supports our hypothesis [25]. Furthermore, DASH-CRYs form a bridge between CPDI/CPDIII photolyases and other members of the CRY/PL family, especially the 6-4 PLs since most DASH-CRYs additionally possess the major motif of the CRY/PL family, α10, and the α8 motif, which we find mainly in 6-4 PLs and in MCRYs. Whether the presence of different CRY/PL subfamily motifs has arisen through convergent evolution or rather through homologous recombination between a CPDI or CPDIII photolyase and a 6-4 PL in an early bacterium is still unclear. DASH-CRYs are versatile in their function and, particularly in Fungi, they appear to be involved in gene regulation and development [27,50,51]. In the green algae C. reinhardtii, DASH-CRYs seem to play a role in the photosynthetic machinery [52], while biochemical studies show that many DASH-CRYs are capable of CPD repair in ssDNA [25,53,54,55]. Apparently, DASH-CRYs partially retained the DNA repair activity of their photolyase precursor, but simultaneously acquired new photoreceptive functions, particularly in light-dependent organisms.

PCRY-likes (plant CRY-likes) belong to the next CRY/PL member that appeared in early evolutionary history, initially in simple eukaryotes (Figure 7, Supplementary Materials S8). A PCRY-like was first described in the diatom P. tricornutum and in other marine organisms [28,29,30]. Although named PCRY-likes, they bear little resemblance to PCRYs in terms of motifs (Figure 3). The plant N-terminal (pn) and the α-helices α10 and α14 are the only common motifs, but since these are anyway most common in all CRY/PLs they are not helpful in PCRY-like assignment. However, the pl1, pl2 and pl3 motifs identified in our study allow a precise assignment. In particular, the pl3 motif (MWQNxG) occurs with 97% in all PCRY-like sequences (100 out of 103 sequences), but in none of the other 2144 sequences belonging to other CRY/PL subfamilies. Interestingly, nearly all organisms that possess a PCRY-like additionally possess a DASH-CRY (Supplementary Materials S8). Only five organisms do not possess both PCRY-like and DASH-CRY simultaneously. Apparently, PCRY-likes and DASH-CRYs are linked in an interplay, which has not been investigated so far. Moreover, PCRY-likes seem to have a unique position in the CRY/PL family, as their occurrence has little to do with phylogenetic relationship between species, but rather with their natural habitat. While we find PCRY-likes in isolated species of fish and frogs as well as in water-associated crustaceans, we do not find it in their close relatives living in terrestrial habitats. Also, in plants and in other unicellular eukaryotes they occur exclusively in aquatic species, regardless of their phylogenetic relationship. Horizontal gene transfer (HGT) seems to play an important role in PCRY-like distribution. HGT is already a known factor in the CRY/PL family evolution occurring more frequently in aquatic organisms [20,56,57].

Continuing the evolutionary history, we find that kingdom-specific CRY/PL members have evolved in multicellular organisms. Thus, plants additionally possess the plant CRYs (PCRYs) as well as the plant photolyases (PPLs). Both are mainly found in higher plants and some red algae (Figure 7). Already Charles Darwin described the ability of plants to perceive blue light [58], and thus a “cryptic” blue-light photoreceptor (namesake for cryptochrome) was suspected in them. Indeed, the first cryptochrome ever, encoded by the gene HY4, was discovered in the thale cress A. thaliana [59]. After characterizing and finding its similarity to photolyases, its involvement in the plant circadian clock was revealed [11,60,61,62]. PCRYs are thought to be closely related and derived from CPDIII photolyases [19,30,44,63]. Although the PCRYs are also very close to the CPDIII photolyase sequences in our phylogenetic tree (Figure 2), we do not find a direct link between CPDIII and PCRY sequences in terms of protein motifs. Moreover, a duplication of the PCRY gene led to the emergence of two gene products PCRY1 and PCRY2 with specialized functions in higher plants. A detailed analysis of these PCRY subgroups can be found in the reviews by Liu et al. (2016) and Wang et al. (2018) [64,65].

Plants as sessile and photoautotrophic organisms are at high risk of UV damage. To protect themselves they express besides the known photolyases (especially 6-4 and CPDII photolyase) also the plant specific photolyases PPL. Although PPLs have no sequence homologies to CPD photolyases, studies in C. reinhardtii suggest a CPD photolyase repair activity for chloroplast and nuclear DNA [34]. In contrast to PCRYs, which are related to CPDIII photolyases, PPLs are related to and likely evolved from DASH-CRYs [19]. Our study further supports this hypothesis, as PPLs particularly cluster very close to plant DASH-CRYs (Figure 2). Most likely, PPLs have evolved from a gene duplication of a DASH-CRY gene within a plant ancestor. A characteristic of the PPLs is that their FAD-binding domains are truncated and less conserved than in the other subfamilies [19]. Other characteristic motifs were previously not reported for PPLs. We identified for the first time three PPL-specific motifs, PPL1, PPL2, and PPL3, which allow differentiation from other members of the CRY/PL family, in particular from DASH-CRYs (Figure 3, Supplementary Materials S3).

All tested Porifera species possess a species-specific cryptochrome, only the demosponge A. queenslandica possess a CPDII photolyase in addition. The species-specific CRY was previously described as eye type or type 0 CRY [1,44,76]. In these studies, as well as in our study, the sponge CRYs cluster near to the animal CRYs, especially near DCRYs (Figure 2). But in our study, they do not form a distinct group, even when spatially grouped together. This could be due to the early separation of single species of this evolutionarily primitive phylum. Moreover, we could not find any type-specific motifs in the sponge CRYs. Some of them are only characterized by the presence of sporadic α-helices of the 6-4 photolyase (α8, α10, sl & α18, see also Supplementary Materials S5). Studies on the sponge Suberites domuncula indicate that these CRYs are photosensitive and involved in various processes such as phototactic behavior, phototransduction, and circadian rhythmicity [76,77]. This suggests that the common ancestor of the Porifera and Bilateria already possessed an animal CRY. T. adhaerens is one of the three recognized species of Placozoa and the only one with a sequenced and annotated genome. T. adhaerens possesses the smallest animal genome and is the simplest free-living multicellular animal [78]. This could be the reason why it has only one CPDIII photolyase, which can be considered as all-rounder featuring PCRY-, MCRY-, 6-4- as well as CPDI/III-specific motifs (Supplementary Materials S4). In contrast, all four animals belonging to the phylum Ctenophora are specialized to only 6-4-PLs. Although we find no evidence for animal CRYs in both phyla, their exclusion should be considered with reservation due to the few animals sequenced.

Cnidarians being relatively close to bilaterians are not considered to be among the most primitive animals [69,70,79]. We found more than 12 sequenced cnidarians allowing us to study individual classes and subclasses in more detail. In cnidarians, which are known for their remarkable plasticity of morphology and life we find a kind of “CRY/PL expansion” (Supplementary Materials S8). While we found only the DASH-CRY in the freshwater polyp Hydra vulgaris (class: Hydrozoa) and the CPDII photolyase in other cnidarians of the class Myxozoa, we detected several CRY/PL family members in the class Anthozoa, specifically in the subclass Hexacorallia. Animals of this subclass possess many CRY/PL family members, including the already known CPDII and 6-4 photolyases as well as the DASH-CRYs. In them, we also discovered for the first time the major animal cryptochrome MCRY (Figure 4b and Figure 7, Supplementary Materials S8). Previous studies have already reported the presence of MCRY in cnidarians [40,80]. Their MCRY sequences differ from the other sequences and barely possess any MCRY-typical motifs, except the MCRY2 specific motif II (Figure 3, Supplementary Materials S3 and S4).

Hexacorallia also possess ACRY, which is not present in any of the other organisms studied (Figure 4b and Figure 7). ACRY sequences are characterized by duplications and the basic sequence is appended up to five times. For example, the sequences of NematostellaXP_032237794.1↗ and XP_032237792.1↗, which consist of 1985 and 2482 amino acids, have 4-5-fold repeats. In our analysis, we only considered CRY/PL homologues with single or double repeats. While in former studies ACRY was classified to be a type I cryptochrome [80], which is a synonym for DCRY, in our study it forms a separate clade close to DCRY (Figure 2). Moreover, except the ACRY of the mountain star coral Orbicella faveolata, having the DCRY-specific motif d2pl (Figure 3, Supplementary Materials S4), all other ACRY lack any DCRY specific motifs. ACRY and DCRY though forming separate subfamilies, possibly share a common ancestor (Figure 2, Supplementary Materials S4). Whether ACRY, like the other animal CRYs, is also involved in the circadian clock is uncertain. Gornik et al. (2021), ruled this out in their studies as they observed a light-induced rather than rhythmic expression in ACRY [40]. This conclusion should be taken with caution because for example one major clock gene of D. melanogaster, cycle, is also not rhythmically transcribed but is kept in sync by its partner CLOCK, which is rhythmically transcribed [81]. Thus, rhythmic transcription is no prerequisite for a clock gene.

Besides cnidarians, also animals belonging to the subkingdom Bilateria possess the MCRY. This suggests that the origin of the MCRY occurred before the separation of the Cnidaria from the Bilateria, i.e., in the common ancestors of the Planulozoa [82]. DCRY, on the other hand, seems to have evolved later. So far, we have detected DCRY only in animals of the subkingdom Bilateria (Figure 4b and Figure 7). Here it occurs in both protostomes and deuterostomes, suggesting its emergence before the split of the two clades. In general, MCRY subfamily members are highly conserved across phyla. In our study, the abbreviation MCRY derives from mammalian-type CRY, but we may even call it the metazoan-type CRY. Indeed, MCRY is the most abundant cryptochrome in the animal kingdom, with a frequency of 86%. While CPDII photolyase is the second most abundant with 76%, the other animal cryptochrome DCRY is found in only 18% of animals (Table 1).

Some chordates, among them bony fishes, amphibians, turtles, lizards, and birds, possess an additional cryptochrome, CRY4. The phylogenetic tree clearly reveals that CRY4 is closely related to, and probably derived from, the 6-4 PL clade (Figure 2). It shares strong similarities with 6-4 PLs, though minimal amino acid changes, particularly the conserved Y319 in CRY4 proteins (4Y motif, Figure 3), seem to alter its photochemical properties and its mode of action [83]. In addition to the two known CRY4-specific motifs 4c and 4Y, we also identified two novel protein motifs enabling strict differentiation between 6-4 PLs and CRY4 (4a & 4b, Figure 3, Supplementary Materials S3). Based on further analysis, we found that the 4a motif is also very abundant in MCRYs, especially in MCRY1 sequences. Some animals examined possess both CRY4 and 6-4 PLs (bony fishes, amphibians, lizards, and turtles), while other possess only the 6-4 PL or only the CRY4 (Supplementary Materials S8). Whereas only single fishes of the Infraclass Neopterygii have CRY4, almost all have the 6-4 PL. During chordate evolution, the ratio reverses, as most birds have CRY4 and less 6-4 photolyase. The lancelets Leptocardii and birds not belonging to the infraclass Neognathae retain CRY4 only (Figure 5A). Besides birds, insects such as the monarch butterfly Danaus plexippus, migrating many hundreds of kilometers annually, are also renowned for magnetoreception [84]. CRY4 is however restricted to chordates. Since CRY4 is very similar to the 6-4 PL, but also to DCRY and MCRY, these might also function as magnetoreceptors by minimal modifications. Studies in cockroaches [85,86], fruit flies [87,88,89] and butterflies [90,91] already indicate their involvement in insect magnetoreception.

In contrast to the deuterostomes, DCRY is widespread among the protostomes. For the first time we see DCRY in the brachyopod L. anatina. Almost all studied animals belonging to the superphylum Spiralia including animals of the phyla Chaetognatha, Mollusca, Rotifera and Annelida contain DCRY. Platyhelminthes seem to be an exception in Spiralia as most of them only possess a CPDII PL (some also a 6-4 PL in addition). In the flatworm Macrostomum lignano we found another CRY/PL family member that is unassigned. Since it shares some common motifs with the 6-4 PL and the animal CRYs and is positioned between them in the phylogenetic tree, we assume that it might be a modified form of the animal CRY. The superphylum Ecdyzosoa consists of the phyla Arthropoda, Nematoda and Tardigrada, among others. While we identified both animal CRYs MCRY and DCRY in Arthropoda, neither the Tardigrada nor the Nematoda possess them. In the water bear R. varieornatus (Tardigrada) and in most nematodes, we found only the CPDII PL. This raises the question why some evolutionarily recent phyla lack the animal CRYs that appeared already in early Bilateria. Interestingly, some evolutionarily young phyla seem to have lost multiple CRY/PLs and retain only a few. We suggest that the era of development and retention of diverse CRY/PL family members was apparently followed by an era of reduction and concentration on single CRY/PLs. These reductions have not always occurred at the level of phyla, but often at the level of classes, families, or genera.

The large number of sequenced animals in both deuterostomes and protostomes allowed us to analyze the CRY/PL restrictions in more detail. Interestingly, many echinoderms (Deuterostomia) possess five metazoan specific CRY/PL family members being MCRY, DCRY, DASH-CRY and both photolyases 6-4- and CPDII (Figure 4). Exceptions are the three sea urchins Mesocentrotus franciscanus, Evechinus chloroticus and Sterechinus neumayeri in which we could only detect the CPDII PL. This probably results of incomplete annotation as these photolyase sequences are derived from a single study [92]. The absence of 6-4 PL in the feather star Anneissia japonica might be true as BLAST-search on the respective assembly with related 6-4 sequences did not yield reliable results.

Recently Kotwica-Rolinska et al. (2021), revealed that Ptychodera flava, a hemichordate, possesses DCRY [43]. Our study lacks hemichordates because most of them are not sequenced and annotated yet. An exception is the acorn worm Sacoglossus kowaleyskii, though having an annotated genome, we could not find any CRY/PL family members in it. The presence of DCRY in echinoderms and hemichordates strongly indicates that DCRY was lost only after the split of chordates. Besides the chordates listed in the phylogenetic tree, we also performed additional BLAST and motif analysis in all chordates sequenced so far and can confirm this loss with certainty. The pelagic tunicate O. dioica and the two lancelets Branchiostoma floridae and Branchiostoma belcheri are the only chordates in our study not belonging to the subphylum Craniata. In the lancelets, we find besides MCRY and CPDII photolyase also the chordate specific CRY4, a putative magnetoreceptor, which is likely derived from 6-4 PLs. Homologues of CRY4 were further identified and characterized in zebrafish [31], birds [93,94], amphibians [95], and in lizards [96]. Our study revealed that turtles possess CRY4 too. Interestingly, all these animals are supposed to have a magnetic sense [97,98,99,100]. Only for lancelets we could not find any studies on magnetic sense so far. However, the presence of CRY4 suggests a magnetic sense in them as well. Furthermore, the occurrence of CRY4 already in Craniata, suggests its early evolution in original chordates (Figure 5A).

The tunicate O. dioica, however, differs from all other Deuterostomia by having two CRY/PLs, one clustering with CPDI PL and the other near DCRY. Since hardly any metazoans and especially no chordates possess CPDI photolyase, we assume that this is either a contamination or a misclassification of the CRY/PL sequence. The former is supported by the fact that it is an aquatic species and contamination is common in these. However, in favor of the second assumption are its motifs, since this one, in addition to its α10, also has an MCRY protrusion loop (ml) motif, which would be more indicative of an MCRY. The second sequence (CAG5112545.1↗), has in addition to the plant N-terminus also the 6-4 PL specific motif, which occurs in 6-4 photolyases, but also in DCRY sequences. Given the proximity to DCRYs in the phylogenetic tree, this could indicate a DCRY relationship (Figure 3). Barring possible contamination and sequencing errors, this would suggest that this more primordial creature has greatly modified its CRY/PLs during evolution.

The sea lamprey P. marinus (class: Hyperoartia) possesses a CPDII photolyase in addition to MCRY. In almost all craniates, MCRY duplicates, MCRY1 and MCRY2, are present (Figure 5A). While lancelets still have an unduplicated standard MCRY, the MCRY in sea lamprey lies between the two duplicates (Supplementary Materials S6). While we define it as MCRY1 based on its clustering (even if as an outgroup), it is still the only one to exhibit the MCRY2-specific motif (II) in the entire MCRY1cluster (Supplementary Materials S4, S6 and S7). Carpet sharks of the order Orectolobiformes (class: Chondrichthyes) have a similar CRY/PL distribution possessing an MCRY1 and a CPDII photolyase. But we find MCRY duplication in other cartilaginous fishes. Interestingly, all species except the great white shark Carcharodon carcharias that have MCRY duplication, have lost CPDII photolyase (Supplementary Materials S8). Still, we assume that ancient chordates probably had many more CRY/PLs than today’s cartilaginous fishes and lancelets, as evolutionarily younger chordates such as bony fishes or frogs also possess many CRY/PLs. Possibly the lamprey and cartilaginous fishes of the early Devonian had more CRY/PLs and only lost them over time.

A CRY/PL increase or retention of multiple CRY/PLs is seen in bony fishes of the class Cladista. They have in addition to the two MCRYs also DASH-CRY, 6-4 and CPDII photolyases. In the class Actinopteri, the fishes belonging to the subclass Chondrostei again lack DASH-CRY, while most fishes of the subclass Neopterygii have DASH-CRY. Many of them also possess the PCRY-like, which appears to be closely linked to DASH-CRY. Fishes having the PCRY-like also usually have the DASH-CRY. The only exceptions are the white flower croaker Nibea albiflora and the allis shad Alosa alosa. The fact that the American shad Alosa sapidissima has both PCRY-like and DASH, but its relative A. alosa does not, is rather unexpected. Since A. alosa has a BUSCO score above 95%, we consider sequencing and annotation errors rather unlikely. Whether A. alosa lost PCRY-like and DASH later in evolution or rather the inclusion of them in A. sapidissima occurred subsequently should be further investigated. However, the distribution of PCRY-like in fish points out that although PCRY-like is present only in water-associated organisms, not all water-associated organisms necessarily possess it.

In fishes belonging to the subclass Neopterygii, the CRY4 emerges for the first time. While all three examined fishes of the Infraclass Holostei possess CRY4, only some fishes of the Infraclass Teleostei do. Fishes from the Cohort Euteleosteomorpha generally do not possess CRY4 (Supplementary Materials S8). The only exception is the alpine whitefish Coregonus sp. “balchen”, where we found a CRY/PL sequence clustering with CRY4 sequences in the phylogenetic tree and possessing the CRY4-specific 4b motif (even if it is incomplete with only 182 amino acids). Assuming this is indeed a CRY4 it would mean that Coregonus, which occurs in alpine lakes such as Lake Constance, has a magnetoreceptor, but its close relative, the pink salmon Oncorhynchus gorbuscha, known for its migrations, does not. Since we also did not find MCRY2 in Oncorhynchus (Supplementary Materials S8), we cannot exclude incomplete sequencing as a possible cause. Likewise, we cannot exclude the possibility that these animals have a peculiar CRY/PL mechanism. In the other classes, on the other hand, CRY4 is very often present. Many fishes possess up to seven CRY/PLs (CRY4, the two MCRYs, DASH-CRY, PCRY-like, 6-4 PL and CPDII PL). However, the catfishes of the order Siluriformes again have a reduced CRY/PL repertoire, with most species having only the CPDII photolyase in addition to the two MCRYs. In general, the combination of MCRY and CPDII photolyases seems to be quite common in the animal kingdom. Especially many animals with reduced CRY/PL repertoire have CPDII besides MCRY, if not only MCRY alone.

The coelacanth L. chalumnae was considered an extinct fossil for a very long time. The deep sea hidden fish was discovered to still exist only in 1939 [101]. Due to its habitat in the dark deep sea, the coelacanth seems to have lost all light-sensitive CRY/PLs. Specific changes in opsins were also detected in them as they only receive light around 480 nm in their environment [102]. Although CRY/PLs could also perceive this wavelength, there is probably no need due to the loss of other wavelengths, especially UV light. As a result, we could only find the light insensitive MCRY duplicates in L chalumnae. In contrast, its close relative the lungfish Protopterus annectens (subclass: Dipnomorpha) inhabiting rather small stagnant or slow flowing waters (in dry periods even without any water) is exposed to large amounts of light. Protopterus possesses again more CRY/PLs, besides the two MCRYs, it also has DASH-CRY and the two photolyases 6-4 and CPDII.

The amphibians are subdivided into Anura (frogs) and the Gymnophiona (caecilians). Frogs that stay in water and on land have many (up to seven) CRY/PLs. In contrast, caecilians living hidden in the upper soil and litter layers seem to have lost many CRY/PLs due to their sheltered-from-light lifestyle and have left only both MCRYs and the CPDII photolyase. While the crocodiles possess the two MCRYs, CPDII and the 6-4 photolyases, the turtles have additionally also CRY4 and DASH-CRY. Although lizards and snakes belong to the same order of Squamata, they also differ greatly in their CRY/PL content. While lizards have similar amount of CRY/PLs like turtles, snakes are like caecilians having only MCRYs and CPDII PL. This again shows that for the CRY/PL repertoire not necessarily the degree of relationship, but other factors, especially environmental conditions of animals play a more important role.

Besides the two MCRYs, most birds also have DASH-CRY and CPDII photolyase. Only one-third of the birds belonging to Infraclass Neognathae possess the 6-4 PL. Birds that do not belong to this Infraclass lack 6-4 PL at all. Instead, many birds, whether migratory or stationary, have CRY4. CRY4 was first discovered in zebrafish and was thought to be a circadian photoreceptor due to its high sequence similarity to DCRY and its rhythmic expression [31]. However, CRY4 has been shown to have a constant expression in zebra finch (Taenopygia guttata) retina, which is considered an important requirement for a magnetoreceptor [103]. Most probably, there are two different CRY4 isoforms as was recently discovered in the night-migratory robin Erithacus rubecula. One of them (CRY4a) is highly regulated during the migratory period and thus has a seasonal rhythm, but no diurnal rhythm [104], suggesting that it is a magnetoreceptor. In contrast, the other isoform CRY4b, showed a circadian regulated expression pattern [105], so that another (possibly clock-related) function cannot be excluded.

In the age of the dinosaurs, most mammals were small and nocturnal. Only after the extinction of dinosaurs about 66 Ma ago, some of them became diurnal [106]. A CRY/PL reduction, caused by nocturnal activity persisted beyond dinosaur extinction. This phenomenon is called nocturnal bottleneck [107] and affects not only CRY/PLs but also other light-associated proteins (such as opsins) and eye structures [108,109]. As a result, most mammals, including their famous representatives mouse and human, possess only MCRY (albeit duplicated as MCRY1 and MCRY2). Moreover, the presence of other DNA repair mechanisms, such as nucleotide excision repair (NER), appears to have allowed mammalian survival even in the absence of photolyases [110,111]. Interestingly, there is also a link between NER and MCRY as NER activity is regulated by the circadian clock and thus by MCRY [112,113,114]. However, most mammals belonging to Prototheria (for example the platypus Ornithorhynchus anatinus) as well as mammals belonging to Metatheria (as example the koala Phascolarctos cinereus or the common wombat Vombatus ursinus) still possess the CPDII photolyase in addition to the two MCRYs. In the long-nosed kangaroo Potorous tridactylus, we even found only one CPDII photolyase and no MCRYs. Interestingly, this CPDII is among the first photolyases discovered [115,116]. Experiments already showed that this one is an all-rounder that can not only repair DNA but also replace MCRY in MCRY-deficient mice [117,118]. Although we cannot exclude the possibility that Potorous indeed lacks MCRY due to the presence of this certain CPDII, we rather assume that MCRY1 and MCRY2 are lacking in Potorous because its genome has not yet been fully annotated. The CPDII PL sequence we obtained was derived from the photolyase-focused study of Yasui et al. (1994) [118].

In contrast, the Tasmanian devil Sarcophilus harrisii seems to lack the CPDII photolyase. Since the genome from the Tasmanian devil is fully sequenced and annotated, the absence of CPDII seems to be genuine. Thus, the Tasmanian devil is very similar to placental mammals (Eutheria) possessing only MCRY. Now, the question arises why most marsupials, also small and nocturnal in the Cretaceous, and in some cases still are, retained CPDII photolyase. Their geographic location might provide an explanation as most marsupials now live in Australia and in South America, which were quite close to the South Pole when Eutheria and Metatheria diverged about 120–140 million years ago [119,120,121]. Consequently, this resulted in stronger photoperiodic variation and seasonal effects due to the tilted earth axis [120]. Long summer days and the increased UV radiation may have contributed to the preservation of CPDII photolyase. In addition, lower temperatures at certain times in these areas may also have partly prevented the large poikilothermic dinosaurs from populating such habitats that were less problematic for the small homoiothermic mammals [120]. Thus, the latter may have remained diurnal.

Since CPDII has not been detected so far in Eutheria, it is assumed that its loss also occurred at the time when Eutheria and Metatheria separated [108]. The sperm whale P. catodon as a relatively recent mammal, however, challenges this hypothesis as it still possesses a CPDII photolyase (Supplementary Materials S8). The split between the sperm whale and other terrestrial animals such as the hippopotamus possibly occurred only about 50 million years ago [122]. This would imply CPDII photolyases were conserved in Eutheria much longer than believed so far. However, none of the other animals studied in the Suborder Cetacea had CPDII. We even searched for possible CPDII photolyase homologs in all mammals already sequenced and annotated but were excluded from this study due to our limitations. Still, the sperm whale was the only organism that had a CPDII. We then reciprocally BLASTed the CPDII sequence of the sperm whale to find its closest relative. It was striking that, apart from a bird Colinus virginianus, the closest hits were parasites such as Besnoitia besnoiti, Toxoplasma gondii, Cystoisospora suis and Neospora caninum. This rather suggests that the whale tissue used for sequencing was likely contaminated with a parasite. This might also be true for the bird Colinus virginianus. Similarly, in our phylogenetic tree, the CPDII from the sperm whale clusters with the tapeworm Rodentolepis nana and not with other chordates. Furthermore, it should be emphasized that in contrast to the MCRY sequences, the CPDII sequence is a partial mRNA sequence and could not be assigned to any chromosome. Taken together, we strongly believe that CPDII in the sperm whale is an impurity and that none of the eutherians studied to date possess a CPDII photolyase. Thus, our study provides further confirmation of the bottleneck theory.

The transparent arrow worm S. cephaloptera is the only sequenced species we have found in the phylum Chaetognata. It is a marine carnivore that belongs to the abundant planktonic organisms. Whether the Chaetognata belong to the Protostomia or whether they form a sister group to the Protostomia and Deuterostomia is still controversial [123,124]. According to paleontological studies, Chaetognatha already existed in the Early Cambrian (540-520 Ma ago) [125]. Thus, the Chaetognata might be among the first animals to possess a DCRY. Notably, we could not identify all other CRY/PL members in it. However, the DCRY of S. cephaloptera was identified in a study that focused on photoreceptors in the eyes [126]. Since the genome of this species has not yet been fully sequenced and annotated, it cannot be excluded that the MCRY and photolyases merely have not yet been identified.

Mollusks and annelids are representatives of the Protostomia Superhylum Spiralia. Both have an increased number of CRY/PL family members. Most of the mollusks and annelids possess five members of the CRY/PL family. These are MCRY, DCRY, DASH-CRY and the photolyases 6-4 and CPDII (Figure 4 and Figure 7). However, we did not find CPDII photolyases in the three sequenced marine mollusks belonging to the order Nudibranchia and no 6-4 photolyase in the East Asian common octopus Octopus sinensis (Supplementary Materials S8). Some mollusks also lack the 6-4 photolyase or the DASH-CRY but MCRY and DCRY are very common in them. In most cases, the absence of the one or the other CRY/PL members is likely due to incomplete sequencing/annotation. Some mollusks even possess more than five CRY/PL members. While the freshwater snail Biomphalaria glabrata, for example possesses an additional CPDIII photolyase, six other mollusks including the Pacific oyster Crassostrea gigas possess an additional PCRY-like. In the case of B. glabrata the additional CPDIII photolyase, which occur very rare in metazoans, might stem from bacterial infections carried by its parasite Schistosoma mansoni, a platyhelminth [127,128]. The Crassostrea genome, on the other hand, is known to be highly polymorphic with transposable elements still actively shaping it [10]. This, coupled with the fact that all PCRY-like containing mollusks are marine animals being in constant contact with other microorganisms, makes the inclusion of PCRY-like via horizontal gene transfer more likely. We also find PCRY-like in the marine annelid Dumeril’s clam worm Platynereis dumerilii. The gain of PCRY-like again indicate a possible horizontal gene transfer. Besides their circadian clock, P. dumerilii is famous for its lunar clock, which characterizes their monthly reproductive cycles [129]. Whether the additional CRY/PL family members (besides the circadian MCRY and DCRY) play a role here is still elusive. However, the other two annelids lack PCRY-like. In the segmented worm Capitella teleta, only the MCRY, DCRY and the CPDII photolyase appear to be present. Since C. teleta has a well-sequenced genome [130], we exclude any sequencing errors and assume that the CRY/PL members are indeed missing.

The major CRY/PL member in platyhelminths and nematodes is the CPDII photolyase. All examined plathelminth species possess CPDII, and some additionally possess the 6-4 photolyase. The whipworms Trichuris trichiura and Diploscapter pachys are the only nematodes having a CPDI photolyase instead of CPDII. The typical animal CRYs MCRY and DCRY involved in the circadian clock are absent in platyhelminths and in nematodes studied. Still, they might have an intact clock. Circadian clocks were described in some platyhelminths by behavioral studies [131]. Since the organisms studied by them are not yet fully sequenced and annotated, their CRYs may still be undiscovered. However, incomplete sequencing and annotation can be ruled out for the best-known model organism Caenorhabditis elegans. Although C. elegans lacks all types of CRY/PL members, it still retains a functioning circadian clock [132,133]. Redox clocks based on S-sulfinylation of the peroxiredoxin (PRDX) protein represent potential alternatives for CRY-independent molecular clocks [134]. As nematodes and flatworms often live in darker, light-protected areas (such as the seafloor, in the soil, or as parasites inside a host), they are not highly exposed to light and thus might have reduced or even lost their CRY/PLs.

The phylum of arthropods, consisting of Crustacea, Chelicerata, and Hexapoda, is one of the largest phyla studied and, accordingly, exhibits the greatest CRY/PL variability among species (Supplementary Materials S8). It is questionable whether we would detect such differences in other phyla, with even similar numbers of animals. Finally, arthropods are not only the most species-rich phylum, but also include animals with diverse habitats and lifestyles. The ancient horseshoe crab L. polyphemus and the many spiders (order: Araneae) still possess MCRY, DCRY, and photolyases 6-4 and CDPII, which are very frequent and typical for arthropods. But the Arizona bark scorpion Centruroides sculpturatus (order: Scorpiones) and all ticks tested (subclass: Acari, order: Ixodida) lack nearly all CRY/PLs except MCRY. CPDI and CPDIII photolyases are also common in Chelicerata. While all studied mites from the order Sarcoptiformes possess only the CPDI photolyase, the sea spider N. striatum possesses a CPDIII photolyase. Moreover, all chelicerates tested lack DASH-CRY.

In general, we find the greatest differences in the subclass of mites Acari. The anatomical, nutritional, and behavioral differences in them are reflected in their differently reduced CRY repertoire. While all ticks have only MCRY and Mesostigmata only CPDII photolyase, Sarcoptiformes have only CPDI photolyase. Since several animals within the orders possess their characteristic CRYs, we cannot assume any sequencing or annotation errors here. The three Trombidiformes examined are the only mites that differ from each other even within their order. The fact that L. polyphemus has four CRY/PLs suggests that the original Chelicerata also possessed multiple CRY/PLs. The large variation and reduction within the subphylum suggest specialization to specific CRY/PLs quite early in evolution (Supplementary Materials S8, Figure 6). The hardened exoskeleton of some chelicerae could protect against UV radiation and thus allow a CRY/PL reduction. However, since crustaceans with hardened exoskeletons (e.g., the American lobster Homarus americanus) continue to have multiple CRY/PLs, this may not be the only reason. The light-protected way of life of some animals in dark places or even parasitic in a host could be another explanation. The parasitic lifestyle makes not only photolyases redundant, but also a separate circadian clock.

In contrast to chelicerates and hexapods, many crustaceans possess DASH-CRY in addition to MCRY, DCRY, and photolyases 6-4 and CPDII. These five are present in almost all Thecostraca, Branchiopoda, and Eucarida. Many crustaceans of the class Hexanauplia even have additional CRY/PLs that are rarely found in arthropods. While the aquatic copepod Eurytemora affinis has an additional PCRY-like, the salmon louse Lepeophtheirus salmonis possesses a CPDIII photolyase. At least for the aquatic copepod we believe a subsequent uptake via HGT. DCRY and DASH-CRY are completely absent in the six Peracarida. Many of them only possess MCRY, and sometimes CPDII photolyase too. Though this CRY/PL reduction could indicate a subsequent loss, we cannot exclude incomplete sequencing as a cause, too. We assume at least that the H. azteca data is complete, as this species has a BUSCO score of >95% in its annotation report. But we could only identify single mRNAs from clock-related studies for Eurydice pulchra and Talitrus saltator [135,136] and no whole genomes. In Eucarida, where we found all five CRY/PLs, we also find very good BUSCO scores. The only exception is the Antarctic krill Euphausia superba, in which we found only MCRY and DCRY. Since both sequences are from a clock-focused study [137], we believe that their photolyases are not sequenced and annotated yet (Supplementary Materials S8).

The springtails Orchesella cincta and Allacma fusca are the only sequenced hexapods that belong to the class Collembola rather than Insecta. The fact that both possess a DASH-CRY suggests that the common ancestor of Crustacea and Hexapoda also possessed one. The silverleaf fly Bemisia tabaci (Hemiptera) and the fungus gnat Bradysia coprophila (Diptera), both pest species exposed to strong sunlight, still possess DASH-CRY. Since DASH-CRY is absent in all other insects studied, this must be a posterior acquisition, e.g., by horizontal gene transfer (HGT). HGT has already been described in several studies with Bemisia [138,139,140]. MCRY, DCRY, and photolyases 6-4 and CPDII are the most abundant CRY/PLs in insects, although they do not always occur together. While some orders possess all four, a reduction is observed in others. Such a reduction may be due to the fact that many published CRY/PL sequences are from CRY-focused studies (e.g., [141,142]) and thus at least some CRY/PLs may not have been discovered. This is supported by the fact that Kotwica-Rolinska et al. (2021), recently reported that the two-spotted cricket Gryllus bimaculatus, in which we found only MCRY and DCRY, also has both photolyases [43]. However, for insects that are generally nocturnal and hide in protected, dark places during the day (termites and cockroaches) or live parasitically hidden and protected (Phthiraptera), we assume that this loss is accurate.

The CRY/PL distribution already differs within insects of individual orders, as exemplified by the Hemiptera. While the Sternorrhyncha (including aphids and whiteflies) have at least four CRY/PL, Laodelphax striatella and Apolygus lucorum lack 6-4 PL, and the fire bug Pyrrhocoris apterus additionally lacks 6-4 PL and DCRY. Similar reduction is seen in the Hymenoptera, which always lack DCRY and 6-4 PL, and sometime CPDII PL. The situation is similar in Coleoptera (Supplementary Materials S8), the only exception being the jewel beetle Agrilus planipennis, which has all four insect CRY/PLs. This could be due to the lifestyle of the invasive beetle, which lays its rather unpigmented eggs in bark cracks of ash trees on the sunny side, where they need to be protected from sun damage. DCRY is the best known CRY of the Diptera. However, our analyses show that it is absent in the fungus gnat B. coprophila, which is curious given the recovery of DASH-CRY mentioned above. Perhaps the absence of DCRY favored the presence of DASH-CRY or vice versa. Another exception is the black soldier fly Hermetia illucens, which has photolyases 6-4 and CPDII (the typical CRY/PLs of dipterans) and MCRY in addition to DCRY. It is also the only Brachycera in our study that does not belong to the suborder Muscomorpha but to Stratiomyomorpha. Recently, other flies belonging to the suborder Muscomorpha were also found to possess MCRY [43]. Since the genomes of these flies have not yet been annotated, they are missing from our study. This finding clearly indicates that MCRY is lost only in certain superfamilies of Muscomorpha.

Gene duplication is an effective method to obtain genes with new functions without losing the original’s function. Thus, incomplete duplication of plant DASH-CRY likely resulted in the creation of an additional plant photolyase (PPL) without losing DASH-CRY. Gene duplicates with specialized functions are also known and already discussed for PCRY in land plants [143] and for MCRY in mammals [144,145]. Even MCRY itself certainly arose from gene duplications of 6-4 photolyases [38,66]. Our study suggests that ACRY, which occurs in anthozoans and poriferan-specific cryptochromes, has also arisen from incorrect 6-4 duplications and subsequent progression. The same is true for DCRY. But whether DCRY arose from a duplication of 6-4 photolyase or from MCRY we are unable to say at this time.

The emergence of various subfamilies is followed by specialization and the loss of individual subfamilies. The absence of various CRY/PL members during evolution is probably caused by gene loss. Single point mutations that affected the functionality of CRY/PLs may have ultimately led to complete loss of the gene. This was likely corroborated by additional factors such as altered living conditions (parasitic lifestyle, nocturnal activity, motility) and gain/loss of other genes (e.g., other repair mechanisms). As an example, many parasitic animals as well as those that live very closely with humans (like the body louse or bedbug), have generally fewer CRY/PLs than their relatives have. One reason for this could be that the clothing and dwellings that provide protection for humans also protect these animals. In addition, the nocturnal activity of small mammals in the mesozoic era, led to the loss of several CRY/PL members. Light activated enzymes have become expendable, as these animals were not exposed to light. Repairing UV-induced DNA damage may be one strategy, avoiding UV light another.

One reason for the sudden appearance of several CRY/PL family members in certain species may result from the events of horizontal gene transfer (HGT). As an example, the sea anemone Nematostella vectensis contains a CPDI sequence that is not present in the other Anthozoa and is very common in microorganisms. The transfer of certain bacterial genes into this anemone has been described previously [146,147].

Moreover, a recent study attempted to trace HGT events particularly in the CRY/PL family by examining inconsistencies in the phylogenetic tree [20]. By identifying a functional Trp tetrad in Gloeobacterium, the study suggested that this cyanobacterium might have taken up a 6-4 photolyase precursor from green algae. Indeed, Gloeobacterium kilauensis, included in our analysis, is one of the few bacterial species that has such a 6-4 homolog, further supporting this hypothesis. This kind of gene transfer ensures perfect adaptation to host species, which is especially crucial for symbiotic or parasymbiotic organisms. The DASH-CRYs of the whitefly B. tabaci or the fungus gnat Bradysia, both plant pests, could also indicate such an adaptation. That the uptake of a plant gene by the whitefly allows its detoxification from plant defenses has been already shown [148].

PCRY-like is most remarkable with respect to HGT. Its presence does not depend on the phylogeny, but rather on the living environment of the organisms. All organisms with PCRY-like live mainly in aqueous environments. First, we find PCRY-like in various algae, probably this is also their place of origin. Then we find PCRY-like sporadically in other marine organisms such as in the pacific oyster C. gigas, in the annelid P. dumerilii as well as in the marine copepod E. affinis. We also find PCRY-like in vertebrates, in this case mainly in fishes belonging to the class Actinopteri and in amphibians. Since no phylogenetic lineage between organisms is apparent, we assume that the more complex organisms possessing PCRY-like have acquired them subsequently by horizontal gene transfer. Furthermore, aquatic habitats have already been shown to promote HGT [56,57,149]. Remarkably, nearly all animals examined that have a PCRY-like also possess a DASH-CRY (Supplementary Materials S8). However, the interplay between the two cryptochromes has not yet been explored.

Perhaps the inclusion of certain genes in one’s own genome is not even necessary if their function can be outsourced to the microbiome. This might not be possible for the DNA-repairing photolyases, but it would be possible for the cryptochromes involved in the circadian clock. After all, microbiomes can interact with their host by means of their metabolites [150,151] and thus also influence their host’s rhythmicity. Indeed, several studies suggest that a proper clock of the microbiome could complement or even replace that of the host organism [152,153,154,155].

Nowadays, organisms with few and with many CRY/PL family members coexist. Our results suggest that the more primitive animals (as example springtails) tend to have more CRY/PLs than their recent relatives (like flies, bees, or ants). Because organisms are constantly evolving, species with few or more CRY/PLs may have coexisted in earlier periods of evolutionary history as well. Morphological and habitat changes could have occurred that reduced the need for CRY/PLs. But why do most ancestral species present today tend to have more CRY/PLs? One reason is certainly the increased UV radiation at the beginning of life (3500 Mya), which was about 500 times higher than today and consequently caused DNA damages making photolyases indispensable [47,156]. A further reason includes the severe collapses occurring later in Earth’s history resulted in mass extinctions. At least the extinction events 250 Ma and 359 Ma ago were associated with the destruction of the ozone layer and a consequently increased UV radiation [157,158,159]. The destruction of the ozone layer may also have contributed to the extinction of the dinosaurs 66 million years ago, albeit due to an asteroid impact and associated air pollution. Under such conditions, organisms with more CRY/PL members, especially photolyases, certainly have a survival advantage, whereas for organisms without photolyases this could be fatal. An increased number of DNA mutations in organisms around these events has been described previously [159]. Organisms with fewer CRY/PLs may have died out due to increased DNA damage. Thus, we suspect that daylight-exposed species that survived these mass extinctions have relatively high numbers of CRY/PLs. This is certainly true for plants being highly dependent on sunlight.

For completeness, we must also mention that the sudden appearance and disappearance of several genes may also be due to methodological errors. Not all species in this study are fully sequenced and annotated. Some of the annotated sequences we used were obtained from clock-related studies in which only cryptochromes were examined and sequenced. Therefore, we cannot exclude the presence of photolyases in these species that have not yet been fully sequenced and annotated. Additionally, DNA contamination due to microbial infection could also be responsible for the sudden appearance of certain genes. CPDI and CPDIII photolyase are rarely found in metazoans but are very abundant in bacteria and fungi. In some metazoans, we find sporadic CPDI and CPDIII occurrences that often do not result in an obvious pattern. Another example of such contamination is the sperm whale P. catodon, which is probably infested with a parasite that may have had a CPDII photolyase (see 4.3.5).