How do lycophytes differ from bryophytes

Phytochrome diversity in green plants and origin of canonical plant phytochromes


  • Phylogenetics
  • Plant genetics
  • Plant physiology


Phytochromes are red / far-red photoreceptors that play an essential role in various plant morphogenetic and physiological reactions to light. Despite their functional importance, the phytochrome diversity and evolution in photosynthetic eukaryotes are poorly understood. Using newly available transcriptomic and genomic data, we show that canonical plant phytochromes are derived from a common ancestor of streptophytes (charophyte algae and land plants). Phytochromes in charophyte algae are structurally diverse, including canonical and non-canonical forms, while in land plants the phytochrome structure is highly conserved. Liverwort, hornwort and selaginella appear to have a single phytochrome, whereas independent gene duplications occurred in mosses, lycopods, ferns and seed plants, which led to different phytochrome families in these clades. Surprisingly, the phytochrome parts of algae and land plant neochromes, a chimera of phytochrome and phototropin, seem to have a common origin. Our results uncover new phytochrome clades and form the basis for understanding the functional evolution of phytochrome in land plants and their algae relatives.


Plants use a range of photoreceptors to measure the quality, quantity, and direction of light in order to respond to ever-changing light environments 1 . Four families of photoreceptors - phytochromes, phototropins, slow motion and cryptochromes - together with UVR8 regulate most of the developmental and physiological processes mediated by ultraviolet B and visible light 1, 2 .

Phytochromes are red / far-red light sensors, which are especially used for the control of the germination of seeds, the photomorphogenesis of seedlings, the avoidance of shadows, the resting phases, the daily rhythm, the phototropism and the flowering 1, 3, 4 are important . Due to their biological importance, phytochromes are of great importance in plant research. The photochemistry, function and related signal transduction mechanisms of phytochromes have been extensively studied, mainly using the Model flower Arabidopsis thaliana 1, 3, 4, 5 .

Canonical plant phytochromes comprise an N-terminal core photosensory module (PCM) and a C-terminal regulatory module 3, 4 . The PCM contains three conserved domains in the linear sequence Per / Arnt / Sim (PAS), cGMP phosphodiesterase / adenylate cyclase / FhlA (GAF) and phytochrome (PHY). This is essential for light reception and photoconversion between reversible conformations that absorb maximally in the red (650–670 nm) or distant red (705–740 nm) regions of the spectrum, which are referred to as Pr and Pfr, respectively. The C-terminal module consists of a PAS-PAS repeat followed by a histidine kinase domain. The domain related to the histidine kinase domain resembles a histidine kinase domain, but lacks the conserved histidine phosphorylation site, which instead has a serine / threonine kinase activity of 6, 7 has .

Plant phytochromes occur as a small nuclear-encoded gene family and fall into three different classes in seed plants: PHYA, PHYB / E and PHYC 8 . The phylogenetic relationships between these groups are well understood, leading to the formulation of functional hypotheses for seed plant phytochromes based on their orthology with Arabidopsis phytochromes 8 allows . However, phytochrome diversity in non-seed plants is very poorly understood as the limited data available from the Physcomitrella (moss) and Selaginella (Lycophyte) -Genome projects 9, 10 and some cloning studies 11, 12, 13 originate. 14, 15 . The lack of a comprehensive phytochrome evolutionary framework for all terrestrial plants is an obstacle to understanding the evolution of the functional diversity of phytochrome and makes it difficult, for example, to correctly interpret results from functional comparisons in A. thaliana and Physcomitrella patens.

A particularly notable plant phytochrome derivative is neochrome, a chimeric photoreceptor that includes a phytochrome PCM and a blue light sensing phototropin 16 combined. Neochromes have only been detected in zygne metal algae, ferns and ruffle horns 17, 18 . While the phototropin component of neochromes has been shown to have two independent origins (one in zygne metal algae and the other in hornwort) 18, the origin of the phytochrome part remains unclear.

In addition to plants, phytochromes are found in prokaryotes, fungi and various protist and Algae lines before 19, 20 . These phytochromes share the PCM domain architecture at the N-terminal with canonical plant phytochromes, but differ in their C-terminal regulation modules. For example, in prokaryotic and fungal phytochromes, the PAS-PAS repeat is absent and there is a functional histidine kinase domain with the conserved histidine residue. Recently, Rockwell et al. 20 and Duanmu et al. 21 examined the phytochromes in different algae lines (brown algae, cryptophytes, glaucophytes and prasinophytes) and found that some of them not only have a great spectral diversity, but also have novel domain combinations within the C-terminal module. Despite these important discoveries, phytochromes remain unaffected by the majority of algal lines. Duanmu et al. 21 suggested that the canonical plant phytochrome may have come from charophyte algae, but failed to confirm this.

In this study, we examined newly available genomic and transcriptomic resources to discover phytochrome homologs outside of seed plants. We examined a total of 300 genomes and transcriptomes from seed plants, ferns, lycophytes, bryophytes, charophytes, chlorophytes and prasinophytes (all in Viridiplantae) as well as from other algal lines with plastids, the glaucophytes, cryptophytes, rhodophytes, haptophytes and stramenophytes. We used these data to reconstruct the first detailed phytochrome phylogenesis for the eukaryotic branches of the tree of life and to map all important events of gene duplication and domain architecture transitions onto this evolutionary tree. We uncover new phytochrome lines and show that the canonical plant phytochromes come from an ancestor of streptophytes (charophyte algae and land plants).


Phytochrome phylogenetic reconstructions

We discovered a total of 350 phytochrome homologs in 148 transcriptome assemblies and 12 whole genome sequences (supplementary tables 1 and 2), which extend over the existing plant and algae diversity. No phytochrome homologues could be detected in the remaining 140 assemblies and genome sequences. We derived phytochrome phylogenesis from an amino acid matrix that contained the sequences we discovered along with previously published sequences from GenBank. To improve understanding of phytochrome and neochrome evolution, especially in ferns and bryophytes, we have also put together three nucleotide matrices. The fern and bryophyte matrices contained 113 and 97 phytochrome sequences, respectively. The neochrome matrix comprised 16 neochromes and 95 phytochromes from selected bryophytes and charophytes.

The topologies of our phytochrome Gene trees are right good with the published organizational relationships 22, 23, 24, 25, 26, 27, 28, 29, 30, 31 matchwhereby we can locate the phylogenetic positions of gene duplication events and delineate new phytochrome clades. In the following we report on the results of the phytochrome diversity, the phylogenetic structure and the domain architecture in the stramenopiles, cryptophytes and archaeplastida (or 'plantae': red algae + glaucophytes + viridiplantae) 32 .

Phytochrome gene line names

The high diversity of phytochromes that we discovered in charophytes, mosses and ferns - resulting from multiple, independent gene duplications - required a sensible system for naming the gene lines. Within each major organismal group of Archaeplastida (with the exception of seed plants where a system for naming PHY is already well established) we used numerical markings for the phytochrome clades that arose from major gene duplication events (e.g. fern PHY1-4 and charophyte) PHY1-2). Subclasses resulting from more local duplications were then named alphabetically within the classes (e.g. Polypodiales PHY4A-B and Desmidiales PHY2A-C). It should be emphasized that this alphanumeric system does not imply orthology across groups of organisms. For example, Fern PHY1 is less related to Charophyte PHY1 than to Fern PHY2. Charophyt PHYX1 and PHYX2 were named here because they are not canonical plant phytochromes like Charophyt PHY1-2 and their evolutionary origin is less clear. For phytochromes in glaucophytes, we used the term glaucophyte phytochrome sensors (GPS) from Rockwell et al. 20 and for the cryptophyte phytochromes with C-terminal serine / threonine kinase we followed Duanmu et al. 21 and named them phytochrome eukaryotic kinase hybrids (PEK).

Stramenopiles and haptophytes

Stramenopiles are a large eukaryotic group that includes brown algae (e.g. kelps), golden algae, and diatoms. The latter are an important part of the plankton. Within this group, phytochromes are only known from brown algae, some of their viruses and diatoms. Their sequences form a clade that is the sister of fungal phytochromes (1 and supplementary 2). Interestingly, the phytochrome of the brown algae virus EsV-1 (Ref. 33) is not grouped with brown algae phytochromes, but is more closely related to those of the diatoms. This relationship was not supported in a bootstrapping analysis (Supplementary Fig. 2); however, it was also used by Duanmu et al. 21 (but without support). Additional phytochrome data from stramenopils will be required to elucidate the origin of these viral phytochromes. We also examined haptophytes, a predominantly marine line of phytoplankton (their relationships with stramenopils and other protists are unclear 27, 28 ). No phytochrome could be found in the haptophyte transcriptomes.

Terminal classes are grouped into higher taxonomic units (usually orders or classes) for display purposes. Orange circles indicate duplicate genes. Italic capital letters in each circle correspond to the duplication events mentioned in the text, and the numbers / letters next to each orange circle are the names of the gene duplicates. Canonical plant phytochromes are from an ancestor of streptophytes (green star), and some charophyte algae retain non-canonical phytochromes (PHYX1 and PHYX2). Phytochrome domain architectures are shown on the right. Domains that are not always present are indicated by dashed outlines. Domain names: GAF (cGMP phosphodiesterase / adenylate cyclase / FhlA); H / KD (Histidine phosphorylation site (H) in the histidine kinase domain (KD)); PAS (Per / Arnt / Sim); PHY (phytochrome); PKC (Protein kinase C); REC (Response Regulator); and RING (Really interesting new gene). * Traditional archaeplastida do not contain cryptophytes 32 .

Full size image

Red seaweed

Red algae are mostly multicellular marine species that include many coral reef algae. No phytochromes were found in the 28 red algae transcriptomes examined or in the published genomes of Porphyridium purpureum, Chondrus crispus, Cyanidioschyzon merolae, Galdieria sulphuraria and Pyropia yezoensis (Supplementary Table 1). This result, based on data from all Rhodophyta classes 34 based, provides convincing evidence for the absence of phytochromes in red algae (supplementary Figure 1).


Glaucophytes are a small group of unicellular freshwater algae with unusual plastids called cyanelles and, unlike plastids in rhodophytes and green plants, have a peptidoglycan layer 35 have . Phytochromes are present in glaucophytes (GPS 20 ), and if our tree on the branch is rooted to prokaryotes / fungi / stramenopil phytochromes, GPS will be resolved as the sister of cryptophytes + viridiplantae phytochromes (1 and supplementary 2). GPS, unlike canonical plant phytochromes, has a single PAS domain in the C-terminal module, and the conserved histidine residue is present in the kinase domain, suggesting that histidine kinase activity is retained 21 .


The phylogenetic position of cryptophytes remains controversial. They were previously thought to be related to stramenopils and haptophytes (belonging to the kingdom of Chromalveolata), but some recent phylogenomic studies place them either as in Archaeplastida 26, 27, 28 nested or as a sister of Archaeplastida 26, 27, 28 on . In our analyzes, Cryptophyte + Viridiplantae-Phytochromes form a clade that is the sister of glaucophyte phytochromes (Fig. 1 and supplementary Fig. 2). Phytochromes from Viridiplantae and from some cryptophytes also share the characteristic PAS-PAS repetition in the C-terminus (Fig. 2). These cryptophyte phytochromes differ from the canonical phytochromes in that the conserved histidine phosphorylation site is retained in the kinase domain (1 and 2). Some cryptophyte phytochromes do not have a PAS-PAS repeat in the C-terminus, but rather have a single PAS, followed by a serine / threonine kinase domain (“PKC” in Figs. 1 and 2). Despite this variation of the C-terminus, the N-terminal photosensitive modules of all cryptophyte phytochromes are monophyletic (Fig. 1 and supplementary Fig. 2).

The tree shows the organismal phylogeny of all phytochrome lines. The domain architecture of the C-terminal regulatory module, which is characteristic of each line, is connected by dashed lines on the right. The N-terminal photosensory module has a largely conserved domain sequence of PAS-GAF-PHY and is not shown here. The substitution of the histidine phosphorylation site (H) in the histidine kinase domain (KD) took place after the divergence of prasinophytes. The canonical plant phytochrome is restricted to streptophytes (in gray box); Zygnematales and Coleochaetales also have non-canonical plant phytochromes. Domain names: PAS (Per / Arnt / Sim); PKC (Protein kinase C); REC (Response Regulator); and RING (Really interesting new gene). * Traditional archaeplastida do not contain cryptophytes 32 . † Full-length phytochrome was not available from Charales and its domain structure was inferred.

Full size image


Viridiplantae comprise two lines, Chlorophyta and Streptophyta. Chlorophyta include chlorophytes (Trebouxiophyceae + Ulvophyceae + Chlorophyceae + Pedinophyceae) and prasinophytes (Supplementary Fig. 1). Chlorophytes appear to be completely absent from phytochromes; No homologues were found in any of the chlorophyte transcriptomes examined, including 14 Trebouxiophyceae, 21 Ulvophyceae, 59 Chlorophyceae and 2 Pedinophyceae. This result agrees with the available data on the whole genome sequence. the genomes of Chlamydomonas reinhardtii, Volvox carteri and Chlorella variabilis (Chlorophyceae) lack phytochromes. Prasinophytes, on the other hand, have phytochromes. Most of these have a PAS-PAS repeat, a histidine kinase domain, and a response regulator domain at the C-terminus 21 . Prasinophyte phytochromes are monophyletic and form the sister group of the streptophyte phytochromes (Fig. 1 and supplementary Fig. 2).

Streptophyta (or streptophytes) are a collection of charophytes (a paraphyletic quality of algae) and terrestrial plants 22 (supplementary 1). We found phytochrome homologues in all terrestrial plant cages and in all charophyte lines: Mesostigmatales (including Chlorokybales), Klebsormidiales, Coleochaetales, Charales, Zygnematales and Desmidiales (Fig. 1 and Supplementary Fig. 1). The Charales phytochromes were not included in our final phylogenetic analyzes because the transcriptome contigs (and also the data currently available on GenBank) are too short to reveal their relationships. All streptophytes have canonical plant phytochromes, including mesostigmatales, the earliest divergent charophyte line (Fig. 1 and 2, supplementary Fig. 1). This result suggests that the canonical plant phytochrome originated in the ancestor of existing streptophytes.

We identified several gene duplication events within charophyte algae. We conclude that duplication occurred after the divergence of mesostigmatales ('A' in Fig. 1), resulting in two clades: one is small and charophyte-specific (Charophyte PHY1) while the other is large and Charophyte PHY2 and the Land plant contains phytochromes. Members of the Charophyte PHY1 clade are not frequent in our algae transcriptomes and were only found in Desmidiales and in Entransia of the early divergent Klebsormidiales (Supplementary Fig. 2). On the other hand, the Charophyte PHY2 homolog is consistently found across algae transcriptomes. Additional duplications ('B' and 'C' in Fig. 1) resulted in three phytochrome subclades within Desmidiales (Desmidiales PHY2A-C). The relationships restored within each of these phytochrome subclasses correspond well to the species phylogenies for Desmidiales 25 .

We found that Zygnematales and Coleochaetales (Charophytes) also have two non-canonical phytochrome clades (Charophytes PHYX1 and PHYX2, Fig. 1). Some PHYX1 have a response regulator domain at the C-terminus, similar to the prasinophyte, cryptophyte, and glaucophyte phytochromes (1 and 2). Interestingly, all known conserved cysteine ​​residues (CysA-D 20 ) in the PAS-GAF region of the N-terminus that bind bilin chromophores, indicating that this protein may not bind bilin or that an unpreserved binding site is needed.


Our data suggest that the phytochrome module of neochrome had a single origin (Fig. 1 and supplementary Fig. 2). Published data suggest that the phototropin module of neochromes, in contrast, has independent origins in algae and hornwort 18 had, suggesting two separate fusion events involving phytochromes that shared a common ancestor. To further investigate this finding, we analyzed the neochrome nucleotide data set (see above) using several nucleotide, codon and amino acid models and performed a topology test. We obtained the monophysis of the phytochrome module of neochromes consistently and usually with high support from analyzes with all models (Fig. 3). Although Anthoceros (a hornwort) neochrome was resolved as the sister of a Zygnematales (algae) neochrome, this relationship was not supported (except in the MrBayes analysis of the nucleotide data set). We then used the Swofford-Olsen-Waddell-Hillis test (SOWH) to compare the topology with all neochromes (the phytochrome module) forming a single clade and an alternative where neochromes were forced by Zygnematales not to group with hornwort + ferns. The alternative hypothesis was rejected (P <0.00001) and the monophysis of the phytochrome module of neochromes was preferred.

The support values ​​are only displayed for the neochrome branches in the following order: Maximum likelihood bootstrap support (BS) from the GTR nucleotide model / Bayesian posterior probabilities (PP) from the GTR nucleotide model / aLRT support from the codon model / Bootstrap maximum likelihood values ​​from the Jones-Taylor-Thornton (JTT) amino acid model / Bayesian posterior probabilities from the JTT amino acid model. '*' Displays all support values ​​= 100 or 1, 0. '-' denotes BS <70, aLRT <70 or PP70, aLRT> 70 and PP> 0, 95.

Full size image


Phytochromes from mosses, liverwort and hornwort each form a monophyletic group (Fig. 4). We discovered individual phytochrome homologs in horn and liverwort transcriptomes. The Gene phylogenies agree with the Species Relations 30, 31 matchwhich is consistent with the presence of single orthologous genes in these taxa. Indeed, a single phytochrome was found in the liverwort Marchantia paleacea var. Identified. diptera 15 . We also have the weakly opaque design genome of the horn word Anthoceros punctatus (20x; Li et al. 18 ) searched and found only one phytochrome. To further evaluate the gene copy number, we hybridized the genomic DNA of A. punctatus with phytochrome RNA probes and sequenced the captured DNA fragments with Illumina MiSeq. The same phytochrome contig (and only this contig) was recovered, suggesting that this hornwort does not contain any additional, divergent phytochrome copies.

Phytochromes previously identified are shown in bold. Support values ​​for branches are Bootstrap Maximum Probabilities (BS) / Bayesian Posterior Probabilities (PP); These are only displayed (together with thickened branches) if BS> 70 and PP> 0.95. Thickened branches without numbers are 100/1, 0. The position of the orange circles indicates suspected gene duplication. Italicized capital letters in each circle correspond to the duplication event mentioned in the text, and the numbers / letters next to each circle indicate the names of the gene duplicates.

Full size image

In contrast, phytochromes are different in mosses, with at least four different clades resulting from three gene duplications (4). The phylogeny shows those moosphytochromes which are orthologous to the aforementioned P. patens phytochromes PpPHY1–5. The Physcomitrella phytochromes and their orthologues form the following classes: moss PHY1_3 (including PpPHY1 and PpPHY3), moss PHY2_4 (including PpPHY2 and PpPHY4), and moss PHY5 (including PpPHY5A-C). An old replication ('D' in Fig. 4) gave moss-PHY1_3 and moss-PHY2_4 + PHY5 clades. The timing of this duplication depends on the phylogenetic position of the Takakia phytochrome, which was resolved here as a sister of the moss PHY2_4 + PHY5 clade, but without support (Fig. 4). Da Takakia (Takakiopsida) the earliest divergent lineage in moss species phylogeny 36 represents, the first phytochrome replication probably preceded the origin of all existing mosses. In the moss PHY2_4 + PHY5 clade, after the splitting of Andreaea (Andreaeopsida), a further doubling occurred ('E' in Fig. 5), but before atrichum (Polytrichopsida) diverged and moss separated PHY2_4 and PHY5. The moss PHY5 clade had an additional duplication ('F' in Figure 4), probably after Physcomitrella had diverged, resulting in moss PHY5D and PHY5E subclasses.

Phytochromes previously identified are shown in bold. Support values ​​for branches are Bootstrap Maximum Probabilities (BS) / Bayesian Posterior Probabilities (PP); These are only displayed (together with thickened branches) if BS> 70 and PP> 0.95. Thickened branches without numbers are 100/1, 0. The position of the orange circles estimates the origin of the suspected gene duplications. Italicized capital letters in each circle correspond to the duplication event mentioned in the text, and the numbers / letters next to each circle indicate the names of the gene duplicates.

Full size image

Our results show that the phytochrome copies previously cloned from Ceratodon purpureus with the designation CpPHY1–4 (Ref. 37) have the following relationships with the mossytochromes: CpPHY1 and CpPHY2 are each closest relatives and members of the moss PHY1_3 descent; CpPHY3 and CpPHY4 are members of the Moos PHY5 line (Fig. 4). These results suggest that the four known C. purpureus phytochromes "CpPHY1", "CpPHY2", "CpPHY3" and "CpPHY4" (Fig. 4) in "CpPHY1_3A", "CpPHY1_3B", "CpPHY5D" and "CpPHY5E" should be renamed The C. purpureus phytochrome discovered here should be designated as CpPHY2_4.


Lycophyte phytochromes are separated as monophyletic and are sister of the fern and the seed plant phytochromes (Fig. 1 and 5, supplementary Fig. 2). Selaginella and Isoetes (Isoetopsida) each have a single phytochrome, with the exception of Selaginella mollendorffii, in which two almost identical phytochromes can be seen in the data of the entire genome sequence. Their high degree of similarity suggests that they are products of a species-specific gene duplication. In contrast, Lycopodiales have two different phytochrome clades that we call Lycopodiales PHY1 and Lycopodiales PHY2. As all Lycopodiales lines 38 are represented in each phytochrome group, we conclude that the duplication of Lycopodiales PHY1 / 2 ('G' in Fig. 5) predates the common ancestor of all existing Lycopodiales.


Fern phytochromes form a clade that is the sister of the seed plant phytochromes (1 and 5, supplementary 2). Within ferns, we have discovered four classes of phytochrome, which we refer to as ferns PHY1, PHY2, PHY4A, and PHY4B. The name PHY3 was previously used to refer to the chimeric photoreceptor, which is now called neochrome 17, 18 is recognized. The deep evolutionary gap between the fern classes PHY1 and PHY2 / 4 precedes the youngest ancestor of the still existing ferns (“H” in Fig. 5). Fern PHY2 and PHY4 were likely separated after Gleicheniales diverged ('I' in Fig. 5), and the earliest divergent fern lines (Gleicheniales, Osmundales, Psilotales, Ophioglossales, Marattiales and Equisetales) have the pre-doubled PHY2 / 4 copy. It should be noted that our broad-scale amino acid dataset resolved a slightly different topology and placed Gleicheniales PHY2 / 4 closer to PHY4 (supplementary Fig. 2). However, the amino acid dataset contained fewer sequences from ferns, which could reduce phylogenetic accuracy 39 . It is likely that the phylogeny (5) inferred from rigorous analyzes of the nucleotide data more accurately reflects the gene relationships.

We found that Ophioglossales and Osmundales each have two PHY2 / 4 copies, which are likely due to independent gene duplications (Fig. 5). The duplication of Ophioglossales PHY2 / 4A and PHY2 / 4B occurred in either the ancestors of Ophioglossales or of Ophioglossales + Psilotales, but the history of PHY2 / 4 in Osmundales is unclear. The Osmundales PHY2 / 4A and PHY2 / 4B have not been resolved as monophyletic, and the phylogenetic position of Osmundales PHY2 / 4B is inconsistent with the published relationships between fern species 23 .

After the cleavage of ferns PHY2 and PHY4, PHY4 duplicated again and resulted in ferns PHY4A and PHY4B ('J' in Fig. 5). Both appear in Polypodiales. We cannot pinpoint the timing of this duplication event because the relationships between Polypodiales PHY4A-B, Cyatheales PHY4, and Salviniales PHY4 are resolved without assistance. Interestingly enough, PHY4A was previously only known from Adiantum capillus-veneris (as AcPHY4). Its first intron contained an inserted Ty3 / Gypsy retrotransposon and the downstream exon sequence was unknown 12 . We found full-length PHY4A transcripts in a variety of Polypodiales, suggesting that PHY4A is likely to be functional in most other species, if not A. capillus veneris. PHY4B is a novel phytochrome clade that has not yet been documented. it is not common in the fern transcriptomes studied.

Seed plants

Seed plant phytochromes cluster in accordance with previous studies 8 in three classes (supplementary 2) corresponding to PHYA, PHYB / E and PHYC. The organizational relationships within the gene subclades are largely in line with those derived from phylogenetic studies of angiosperms 40 . Remarkably, however, support for the gymnosperm monophysis was poor. We found two divergent transcripts from PHYE in Ranunculales, represented by Aquilegia (Ranunculaceae; from whole genomic data) and Capnoides (Papaveraceae; from transcriptome data) (Supplementary 2), suggesting that a gene duplication event occurred deep in Ranunculales; However, larger samples of Ranunculales are required to determine when this duplication occurred.


Our phylogenetic results refute earlier hypotheses that plants are endosymbiotic gene transfer 41, 42 Phytochrome from cyanobacteria have acquired, since streptophyte and cyanobacteria phytochromes are not closely related in our phytochrome trees (Fig. 1 and Supplementary Fig. 2) by Duanmu et al. 21 . Instead, plant phytochromes evolved from a precursor shared with other archaeplastidae. We have clearly classified the origin of canonical plant phytochromes in a common ancestor of existing streptophytes (Fig. 1 and 2). Our data also show that the creation of this structure required several steps. The increase in internal PAS-PAS repetition occurred first in the ancestors of Viridiplantae or Viridplantae + cryptophytes (Fig. 2). As noted above, the location of cryptophytes is uncertain and its inclusion in archaeplastida is discussed in published studies 26, 27, 28 not strongly supported. The topology of our phytochrome trees is consistent with a sister group relationship between Viridiplantae and Cryptophytes, but the topology could also result from endosymbiotic or horizontal gene transfer. The loss of the histidine phosphorylation site in the histidine kinase domain - hence the attainment of the canonical form - occurred later in the ancestor of the streptophytes and appears to be accompanied by permanent dissociation with the reaction regulator at the C-terminal end (Fig. 2). Some streptophytes have additional, non-canonical phytochromes. Charophyte PHYX1 and PHYX2, both found in Zygnematales and Coleochaetales, have the conserved histidine residue, and some PHYX1 also have a response regulator domain (1 and 2). The fact that the charophytes PHYX1 and PHYX2 were not found in all streptophytes suggests that some orthologues may have been missed in our transcriptomic and genomic scans and / or in a scenario where duplications in streptophytes occurred early and of multiple losses were followed.

Our results highlight the different evolutionary modes of the phytochrome N- and C-terminal modules. The N-terminal photosensory module is deeply conserved across eukaryotes and prokaryotes, and the linear domain sequence of PAS-GAF-PHY is found in most known phytochromes (Fig. 1). In contrast, the development of the C-terminal regulatory module was much more dynamic (Fig. 2). For example, the C-terminal PAS can be absent, occur singly, or occur as a tandem repeat (3). Serine / threonine kinase or tyrosine kinase domains were also independently involved in the regulatory module in the phytochromes 43 of Cryptophytes and C. purpureus (moss) were recruited (2). The successful linking of the photosensitive phytochrome module with a large number of C-terminal modules has promoted the functional diversity of the phytochrome. The most convincing example is certainly that of the neochromes. It combines phytochrome and phototropin modules into a single protein to process blue and red / far-red light signals in controlling phototropism 44 . Neochrome was first found in ferns 16 discovered and as a driver of the modern Fern radiation under dominated by angiosperms Forest roofs in poor lighting conditions suspected 45, 46, 47 . Suetsugu et al. 17 later discovered a similar phytochrome phototropin chimera in Mougeotia scalaris (zygnematic algae) and suggested that neochrome had independently evolved twice. A recent study identified another horn root neochrome and showed that ferns get their horn root neochromes through horizontal gene transfer 18 purchased . By Li et al. Classify the phototropin component of neochrome in a broad phylogeny of phototropins. 18 also showed that phototropin modules of neochromes had two separate origins, one in hornwort and one in Zygnematal algae. In contrast, the phytochrome part of neochrome had a different development history, with Zygnematales, hornwort and ferns forming a single monophyletic group (Fig. 3 and supplementary Fig. 2). This result is robust and is supported by most analyzes and a topology test. Our results therefore suggest that neochromes were formed via two separate fusion events in which two different sources of phototropin but the same phytochrome precursor were involved. This is a fascinating extension of the capacity and tendency of the photosensitive phytochrome module to link with functionally different downstream domains.

The main classes of land plants differ significantly in terms of phytochrome gene diversity. Phytochromes appear to be a single copy in most liver, horn and Isoetopsida species (Isoetaceae and Selaginellaceae), while they have diversified independently in Lycopodiales, mosses, ferns and seed plants (Fig. 1). In ferns, a pattern of early gene duplication followed by gene loss could explain the phylogenetic positions of two Osmundales PHY2 / 4 that are incongruent with known species relationships in ferns (Fig. 5). Interestingly, we observed a relationship between phytochrome copy number and species richness. For example, the polypodial ferns (Polypodiales) make up 90% of the existing fern variety 47 make up four phytochrome copies, while other species of ferns with few species have only two or three (Fig. 5). Likewise, moss species belonging to the hyper-diverse Bryopsida and containing 95% of the existing moss diversity have experienced the highest number of phytochrome duplicates compared to other bryophyte lineages (Fig. 4). It is possible that the evolution of the structural and functional diversity of phytochrome may improve the ability of polypod ferns and bryopsida mosses to adapt to different light environments. In fact, seed plants, ferns, and mosses each have at least one phytochrome duplicate that plays the role of mediating high-level irradiance responses 48, 49, 50, 51 developed or maintained converging. These Property is likely important for deep canopy survival 52 ( see below). This "phytochrome-driven biodiversity" hypothesis, however, requires rigorous testing by phylogenetic comparative methods and functional studies on non-seed plants that identify the genetic basis of phytochrome functions.

The independent phytochrome diversification events in seed plants, ferns, mosses and Lycopodiales have significant implications for phytochrome function studies.For example, moosphytochromes are more closely related than all seed plant phytochromes (and so are phytochromes from ferns and those from Lycopodiales, of course). Seed plant phytochromes have undergone significant differentiation into two main types. One is represented by phyA from A. thaliana, the main mediator of red light responses in deep shade and below the soil surface. It breaks down quickly in light, mediates reactions at very low fluence and high irradiance and depends on the protein partners FHY1 (far-red elongated hypocoytl 1) and FHL (FHY1 like) for nuclear translocation. The other is represented by phyB-E from A. thaliana, which are the primary mediators of red light responses in open habitats. They have a longer light half-life than PhyA, mediate low fluence reactions, and in the case of PhyB, nuclear translocation does not require FHY1 or FHL 4, 5 . A similar division of function was found in some fern and moss phytochrome duplicates 49, 53, 54 documentedwhich shows a case of convergent differentiation after duplication of independent genes. In future studies it would be of particular interest to infer the traditional properties of phytochrome from land plants: did it have a short or long half-life? What types of physiological responses were mediated? How was the nuclear translocation carried out? Studies of liver, horn and selaginella phytochromes, which are available as single copy genes, could serve as "basic models" for understanding the genetic basis of the functional diversification of phytochrome.

Recent functional studies on a small but diverse group of algal phytochromes have revealed a surprising level of spectral diversity, adapting to a range of marine and marine life Water environments could reflect 20, 21 . For example, photoreversible phytochromes in prasinophyte algae include orange / far red and red / far red receptors, and in algae outside of Viridiplantae there are also blue / far red and red / blue receptors 21 . This is in sharp contrast to the very limited spectral diversity in canonical plant phytochromes, all of which are known to be red / far red receptors. The novel algal phytochrome clades that we have discovered are a potential treasure trove for and for discovering the steps in moving from a spectrally distinct set of reversible photoreceptors to one that focuses on the red to far red regions of the spectrum the characterization of new biochemical variants. Some of these could have implications for understanding the role of phytochrome evolution in the recolonization of marine and freshwater environments by terrestrial plants.

In summary, our study has shown that the diversity of Viridiplantae phytochromes is far greater than previously assumed and shows interesting possibilities to link this structural diversity with function and ecology.


Breakdown of transcriptomes and genomes for phytochrome homologs

The transcriptomes and genomes examined in this study are listed in Supplementary Table 1. We used the Python pipeline BlueDevil according to Li et al. 18to break down transcriptomes. To search the total genome data, we used BLASTp, the one in Phytozome 55 or individual genome project portals is implemented (supplementary table 1). The protein domain composition of each of the phytochrome sequences was obtained by querying the NCBI Conserved Domain Database 56 determined .

Sequence alignment

In addition to the phytochrome homologues obtained from transcriptomes and genomes, we collected selected Genbank accessions and a sequence cloned from Marattia howeana (reference: SWGraham and S. Mathews 15, deposited in NSW; primer: 110f -5'GTNACNGCNTAYYTNCARCGNATG3 ', 788r - 5 "GTMACATCTTGRSCMACAAARCAYAC3").

We compiled four sets of sequence data, one was translated into an amino acid alignment and the others were analyzed as nucleotide templates. The amino acid data set contained the majority of the sequences (a total of 423 sequences; additional 2). The sequences were initially created using MUSCLE 57 alignedfollowed by manual curation of the alignment based on known domain boundaries and protein structures. Regions with uncertain or no homology were not included. This includes all response regulators (REC), really interesting new genes (RING), light-oxygen-voltage sensors (LOV) and PKC domains as well as the single PAS domain in GPS. Non-alignable regions were also excluded and the final alignment included 1,106 amino acid sites. The nucleotide datasets were compiled to provide higher phylogenetic resolution within fern + lycophyte phytochromes (113 sequences; Fig. 5), bryophyte phytochromes (97 sequences; Fig. 4) and neochromes (111 sequences; Fig. 3). The sequences were aligned as amino acids and then back-translated to nucleotides, and the alignment was refined by manual editing. The phytochrome alignments of fern + lycophyte and bryophyte contained 3,366 and 3,429 nucleotide sites, respectively. The neochrome alignment only included the N-terminal photosensor module (PAS-GAF-PHY domains; 1,920 nucleotide sites). All alignments are available from Dryad (//

Phylogenetic reconstruction

For the broad-based amino acid targeting, JTT + I + G was selected by ProTest 3 under Akaike Information Criterion 58 selected as the most appropriate empirical model. We used Garli v2.0 (Ref. 59) to find the maximum likelihood tree with ten independent runs and a genthreshfortopoterm of 100,000. The start tree for Garli comes from a RAxML v8.1.11 run (Ref. 60). To get values ​​for bootstrap branch support, RAxML was run with 1,000 replicas using JTT + G.

For the nucleotide alignments, we used PartitionFinder v1.1.1 (Lit. 61) to derive the best partition schemes and substitution models from Akaike Information Criterion. The search and bootstrapping of the maximum likelihood tree (1,000 replicas) were carried out in RAxML. The Bayesian inference was carried out in MrBayes v3.2.3 (Ref. 62) with two independent Markov chain Monte Carlo runs (MCMC) and four chains each. We decoupled the substitution parameters and set the rate before varying between partitions. The MCMC output was made using the tracer 63 checked to ensure convergence and mixing (effective sample sizes all> 200); 25% of the entire generations were discarded as burn-in prior to the analysis of the posterior distribution.

Additional analyzes were applied to the neochrome data set. First we used CodonPhyML v1.0 (Lit. 64) to derive the tree topology and to evaluate the support (SH-like aLRT Branch Support) using a codon substitution model. Four categories of non-synonym / synonym substitution rate ratios were drawn from a discrete gamma distribution, and the codon frequencies were estimated from the nucleotide frequencies at each codon position (F3 × 4). Second, we translated the nucleotides into amino acids and performed a maximum likelihood tree search and bootstrapping (in RAxML) as well as Bayesian inference (in MrBayes) according to the JTT + I + G model. Eventually we used the one in SOWHAT 65 implemented SOWH test to investigate whether the derived tree topology (phytochrome portion of the neochrome that forms a clade) is significantly better than the alternative topology (neochrome not monophyletic). In SOWHAT we used the standard stop criterion and applied a topological constraint that forces land plants and zygnematal neochrome not to be monophyletic.

In order to derive the organizational relationships shown in Fig. 2 and Fig. 1, we have merged the topologies from three recent phylogenetic studies. The relationships between streptophytes and chlorophyte algae are from Wickett et al. 22 or Marin 29 . For the broader kingdom-level relationship, the topology of Grant and Katz was used 28 referenced.

Confirmation of the gene copy number through target enrichment

We used a target enrichment strategy to test whether the hornwort has a single phytochrome locus. In this approach, specific RNA probes are hybridized to genomic DNA in order to enrich the representation of certain gene fragments. Target enrichment has several advantages over the traditional Southern blotting approach. In particular, thousands of different hybridization probes (and not just a few) are used and the end products are not DNA bands but actual sequence data.

We have developed a total of 7,502 120-mer RNA probes targeting phytochrome, phototropin, and neochrome genes, with an emphasis on hornwort and fern genes (probe sequences available at Dryad // dryad.5rs50). . The probes overlap every 60 bp (a 2X tile strategy) and were synthesized by Mycroarray and labeled with biotin. Hornwort A. punctatus genomic DNA was extracted using a modified hexadecyltrimethylammonium bromide (CTAB) protocol and sheared from Covaris with fragments peaking at 300 bp. Library preparation for Illumina sequencing was done using a KAPA Biosystem Kit in combination with NEBNext Multiplex Oligos. To enrich for potentially divergent homologs, we used the touchdown method of Li et al. 66, in which the genomic DNA library and probes were hybridized for 11 h at 65 ° C, followed by 60 ° C (11 h), 55 ° C (11 h) and 50 ° C (11 h). The hybridized DNA fragments were captured with streptavidin beads and washed according to the Mycroarray protocol. The final product was pooled with nine other libraries in equimolar form and sequenced on Illumina MiSeq (250 bp paired end). To process the reads, we used Scythe v0.994 (Ref. 67) to remove the adapter sequences and Sickle v1.33 (Ref. 68) to trim low quality bases. The resulting measured values ​​were then compiled by SOAPdenovo2 (Ref. 69) and the phytochrome contig was compiled by BLASTn 70 identified. The raw data was stored in NCBI SRA (SRP055877).

Data availability

All relevant data in this publication can be found at //

Additional information

Access codes: The read processes generated in this study from the target enrichment were stored in the NCBI Sequence Read Archive under the access code SRP055877. The phytochrome DNA sequences generated in this study were stored in the GenBank database under the access codes KP872700, KT003814 to KT003816, KT013280 to KT013293 and KT071819 to KT072044.

How to cite this article: Li, F.-W. et al. Phytochrome diversity in green plants and origin of canonical plant phytochromes. Nat. Commun. 6: 7852 doi: 10, 1038 / ncomms8852 (2015).


GenBank / EMBL / DDBJ

  • KP872700
  • KT003814
  • KT003816
  • KT013280
  • KT013293
  • KT071819
  • KT072044

Read sequence archive

Further information

PDF files

  1. 1.

    Further information

    Supplementary figures 1-2, supplementary tables 1-2 and supplementary references


By submitting a comment, you agree to our terms of use and community guidelines. If you find something abusive or that does not comply with our terms or guidelines, please mark it as inappropriate.