Insights into the evolution and diversification of theAT-hook Motif Nuclear Localizedgene family in land plants

Background

Members of the ancient land-plant-specific transcription factorAT-Hook Motif Nuclear Localized(AHL) gene family regulate various biological processes. However, the relationships among theAHLgenes, as well as their evolutionary history, still remain unexplored.

Results

We analyzed over 500AHLgenes from 19 land plant species, ranging from the early divergingPhyscomitrella patensand年代elaginellato a variety of monocot and dicot flowering plants. We classified the AHL proteins into three types (Type-I/-II/-III) based on the number and composition of their functional domains, the AT-hook motif(s) and PPC domain. We further inferred their phylogenies via Bayesian inference analysis and predicted gene gain/loss events throughout their diversification. Our analyses suggested that theAHLgene family emerged in embryophytes and further evolved into two distinct clades, with Type-IAHLs forming one clade (Clade-A), and the other two types together diversifying in another (Clade-B). The twoAHLclades likely diverged before the separation ofPhyscomitrella patensfrom the vascular plant lineage. In angiosperms, Clade-AAHLs expanded into 5 subfamilies; while, the ones in Clade-B expanded into 4 subfamilies. Examination of their expression patterns suggests that theAHLs within each clade share similar expression patterns with each other; however,AHLs in one monophyletic clade exhibit distinct expression patterns from the ones in the other clade. Over-expression of aGlycine maxAHL PPC domain inArabidopsis thalianarecapitulates the phenotype observed when over-expressing itsArabidopsis thalianacounterpart. This result suggests that theAHLgenes from different land plant species may share conserved functions in regulating plant growth and development. Our study further suggests that such functional conservation may be due to conserved physical interactions among the PPC domains of AHL proteins.

Conclusions

Our analyses reveal a possible evolutionary scenario for theAHLgene family in land plants, which will facilitate the design of new studies probing their biological functions. Manipulating theAHLgenes has been suggested to have tremendous effects in agriculture through increased seedling establishment, enhanced plant biomass and improved plant immunity. The information gleaned from this study, in turn, has the potential to be utilized to further improve crop production.

Genes that regulated essential biological processes in ancient plant species constituted a conserved "gene tool kit", which tended to be preserved throughout evolution [1]-[4]. Most of the members in this "tool kit" have generally duplicated and expanded into multi-member-containing gene families with divergent functions in modern land plants [1],[5],[6]. Understanding their functions as well as evolutionary histories have greatly enhanced our knowledge of plant growth and development, such as the cases of the cytochrome P450s [7], MADS-box transcription factors [8]-[12],AP2/EREBPgenes [13]-[16], theTALEhomeobox gene family [17]-[19], NAC transcription factors [20]-[22],HD-ZIPgenes [23]-[25],Basic/Helix-Loop-Helixgenes [26]-[28] and theTCPgene family [29]-[31].

However, there are also many gene families that are important to land plant evolution whose functions and evolutionary histories are not well understood. The ancient transcription factorAT-Hook Motif Nuclear Localized(AHL) gene family has been found in all sequenced plant species, ranging from the mossPhyscomitrella patens, to flowering plants, such asArabidopsis thaliana,年代orghum bicolor,Zea maysandPopulus trichocarpa．High conservation of this gene family throughout land plant evolution suggests that it is important for plant growth and development. Currently we are beginning to understand the biological functions of severalAHLs. The evolutionary history of this gene family, however, has still barely been explored.

Members of the AHL proteins contain two conserved structural units, the AT-hook motif and thePlant andProkaryoteConserved (PPC) domain, the latter being also annotated as theDomain ofUnknownFunction #296(DUF296) [32]. Since the functions of this domain have been partially revealed [33], hereafter, we will refer it only as the PPC domain. The AT-hook motif enables binding to AT-rich DNA and has been identified in various gene families both in prokaryotes and eukaryotes, including theHighMobilityGroupA(HMGA) proteins in mammals [34]. The AT-hook motif uses a conserved palindromic core sequence, Arg-Gly-Arg, to bind to the minor groove of AT-rich B-form DNA. Upon binding with DNA, this core sequence adopts a concave conformation with close proximity to the backbone of the DNA, with both arginine side chains firmly inserting into the minor groove [35].

The second functional unit of the AHL proteins is the PPC domain, which is approximately 120 amino acids in length and exists as a single protein in Bacteria and Archaea [32]. Crystal structures of several bacterial and archaeal PPC proteins suggested that the prokaryotic PPC proteins form a trimer [36],[37]. In land plants, the PPC domain has been identified in AHL proteins where it is located at the carboxyl end relative to the AT-hook motif(s) [32]. The PPC domain is responsible for the nuclear localization of the AHL proteins as well as protein-protein interactions among AHL proteins and with other common interactors, such as transcription factors. It may suggest a role in regulating transcriptional activation by the AHL proteins in plants [33].

Members of theAHLfamily regulate diverse aspects of growth and development in plants. Most of the studies are from the analyses ofArabidopsis thaliana．年代everalAHLs are suggested to regulate the homeostasis of phytohormones, especially gibberellins [38], jasmonic acid [39] and cytokinins [40]. Two members of theArabidopsis thaliana AHLgene family,年代UPPRESSOR OF PHYTOCHROME B-4 #3(年代OB3/AHL29) andESCAROLA(ESC/AHL27), repress hypocotyl elongation for seedlings grown in the light [41]. As adults, theAtAHLover-expression plants develop enlarged organs, such as expanded leaves, flowers and fruits as well as delayed flowering and senescence [41]. Similar functions have also been proposed forAtAHL22, andHERCULES(HRC/AHL25) [42],[43].Arabidopsis thaliana ESC/AtAHL27andAHL20have also been implicated in the regulation of plant defense responses [44],[45].

In this study, we identified members of theAHLgene family in the completely sequenced genomes of 19 land plant species, ranging from the mossPhyscomitrella patensand the lycophyte年代elaginellato a variety of monocot and dicot species in the Phytozome database [46]. A closer look at their protein sequences revealed that these land plant AHL proteins can be divided into three types (Type-I, -II and -III) based on a combination of the number and composition of its two structural units, the AT-hook motif(s) and the PPC domain. The Type-IAHLs form one clade; while the Type-II and -IIIAHL在一起形成一个单独的分支。系统发育安娜lysis of theAHLgenes in basal plants suggests that such divergence between the two clades dated between the appearance of chlorophytes and mosses. In this study, we have further identified that theAHLgene family in land plants evolved into 9 phylogenetic sub-families. Finally, we have proposed an evolutionary scenario for theAHLgene family in land plants.

Early divergence in the land-plant AHL protein family

Members of theAHLgene family contain two functional units, the AT-hook motif and the PPC domain [32]. In order to identify theAHLgenes in land plant species, we performed searches against the Phytozome database using theAHLnucleotide and amino acid sequences fromArabidopsis thaliana[46]. We further added the retrieved results as additional queries to perform further searches to identifyAHLgenes from the genomes of 19 plant species (Figure1a, Additional files1,2and3)．

Initial phylogenetic analysis of the retrieved AHL proteins in this study suggested that all of the land-plant AHL proteins evolved into two major clades (Figure1b). This distinct division into two monophyletic clades could also be observed in phylogenetic analysis when using just theAHLgenes fromArabidopsis thaliana[32],[33],[38],[41] andOryza sativa[47]. Analysis of all theAHLgenes identified in this study in the moss and lycophytes reveals a similar distribution into these two clades. This further suggests that the division between these two branches dated before the divergence of mosses from the rest of the land plants.

Each monophyletic clade defines one type of PPC domain in land plant AHL proteins

Examination of the PPC domains revealed that their protein sequences share unique characteristics within each of the two AHL phylogenetic clades (Figure1b, Additional file4)．The Clade-A AHL proteins share the same type of PPC domain (hereby named "Type-A PPC domain"). Clade-B AHL proteins share another type of PPC domain (hereby named "Type-B PPC domain").

In order to further examine the divergence between the PPC domains in AHL proteins, we performed a sequence logo analysis. The Type-A PPC Domain in Clade-A generally starts with Leu-Arg-Ser-His (Additional file4a); while the Type-B PPC domain in Clade-B generally starts with Phe-Thr-Pro-His (Additional file4b). Both types of PPC domains in AHL proteins are further followed by stretches of amino acid residues with moderate conservation. Examination of both types of PPC domains in the identified AHL proteins revealed that they contain a consensus conserved Gly-Arg-Phe-Glu-Ile-Leu motif (Additional file4a, b). It is also interesting to note that the coding sequences of this motif always exists at the immediate beginning of one exon region in the intron-containing Type-B PPC/DUC296 domains. The sequence upstream of the conserved six amino acids in Type-B PPC domains is generally Thr-Tyr-Glu, while it is generally Thr-Lys-His upstream of the six amino acids in Type-A PPC domains. The sequences downstream of the conserved six amino acids in both types of PPC domains are similar to each other.

Conserved functions of PPC domains in AHL proteins in land plants

In order to understand the biological functions of the PPC domains in the AHL proteins, we cloned two full-lengthAHLgenes from the bread wheatTriticum aestivumand one PPC domain from a soybeanGlycine max AHLgene (Gm06g01650.1) (Additional file5)．AlthoughGm06g01650.1is only a partial gene, it together with the cloned wheatAHLs and twoArabidopsis thaliana AHLs encode proteins that all contain a Type-I AT-hook motif and a Type-A PPC domain (Additional files5and6)．They share the same arrangement of secondary structural elements and tertiary structures with each other, as well as with their counterparts in prokaryotes and the moss,Physcomitrella patens(Figure2a and2b). A careful examination reveals that their PPC domains all exhibit aβ₁-α-β₃-β₇-β₄-β₅-β₆-β₂secondary structural arrangement, suggesting possible conserved biological functions of this domain among multiple species.

To test the hypothesis that the PPC domain may share conserved biological regulatory functions, we overexpressed this domain fromGm06g01650.1driven by the35Sconstitutive promoter in wild-typeArabidopsis thaliana．Multiple homozygous over-expression lines containing single-locus insertions exhibited longer hypocotyls in white light comparing with wild-type controls (Figure2c). This long-hypocotyl phenotype is similar to the one demonstrated by seedlings over-expressing the PPC domain fromArabidopsis thalianaAtAHL29/SOB3 [33], suggesting that shared conserved biological functions exist betweenGlycine maxandArabidopsis thaliana AHLs.

Arabidopsis thaliana AHLs have been suggested to suppress hypocotyl growth in the light [33],[41]. Therefore, the long-hypocotyl phenotype exhibited by over-expressing theGm06g01650.1PPC domain may be conferred through the disturbance of the growth suppression roles ofArabidopsis thaliana AHLgenes. To test this hypothesis, we examined if the PPC domain of Gm06g01650.1 can physically interact with theArabidopsis thalianaAHL proteins using a targetedlexA-based yeast two-hybrid assay (Figure2d,e). Using 1.25 mm 3-amino-1, 2, 4-triazol that prevented transcriptional auto-activation by SOB3/AtAHL29 in the bait protein, we demonstrated that SOB3/AtAHL29 fromArabidopsis thalianaand the PPC domain ofGlycine maxGm06g01650.1 can interact with each other (Figure2d,e).

Type-I and -II AT-hook motifs exist in AHL proteins

Two types of AT-hook motifs (Type-I and -II) are found in the AHL proteins (Figure3a,b; Additional file7) [33],[34]. Both types of AT-hook motifs in the AHL proteins share the same conserved Arg-Gly-Arg core and use this conserved palindromic core to bind the minor groove of AT-rich B-form DNA [35]. Clade-A AHLs contain only one copy of the Type-I AT-hook motifs; while, in Clade-B, some of the AHLs contain only one copy of the Type-II AT-hook motifs and the rest contain both types of AT-hook motifs.

A specific consensus sequence, Gly-Ser-Lys-Asn-Lys, was observed at the carboxyl end of the Arg-Gly-Arg core sequence in the Type-I AT-hook motifs (Figure3a, Additional file7a,b). The conservation of these downstream sequences is more significant in the AHLs that only contain this type of AT-hook motif. However, these sequences are more variable in other AHLs that also possess a Type-II AT-hook motif (Additional file7b)。只有短共识氨基酸延伸,参数-Lys-Tyr, could be observed downstream of the conserved Arg-Gly-Arg core sequences of the Type-II AT-hook motifs in clades of both AHLs (Figure3b, Additional file7c,d). The conservation of these downstream sequences is similar among the AHLs in either clade (Additional file7c,d).

Three types of AHL proteins in land plants

Based on a combination of type and number of the AT-hook motif(s) and the PPC domain, all the AHL proteins identified in this study can be further classified into three types (Type-I, -II and -III AHLs) (Figure3c). The Type-I AHL proteins contain one Type-I AT-hook motif and one Type-A PPC domain. The Type-II AHL proteins contain two AT-hook motifs (one additional Type-II AT-hook motif at the N-terminus of the Type-I AT-hook motif) and one Type-B PPC domain. Finally, the Type-III AHL proteins contain one Type-II AT-hook motif and one Type-B PPC domain. Clade-A is comprised of the Type-IAHLgenes, while Clade-B is comprised of the Type-II and -IIIAHLgenes. Both clades haveAHLgenes fromPhyscomitrella patens(moss) forming a sister clade to the rest of the members of the clade, indicating an early divergence between the Type-IAHLs and the other two types ofAHLgenes.

Type-I and -II AHLs found in flowering plants were present in early-diverged land plants

In order to understand the evolutionary origin of theAHLgenes, we also performed searches forAHLgenes in chlorophytes. Neither anyAHLgenes nor genes encoding the PPC domain could be identified in the current release of theChlamydomonas reinhardtiiandVolvox carterigenomes (Figure1a) [46],[48],[49]. Surprisingly, we were able to identify only onePPC基因编码的PPC领域没有n associated AT-hook motif(s) inMicromonas pusilla CCMP1545[50] andOstreococcus lucimarinus[51] (Additional file8)．To further examine the presence of thePPCgene in picoeukaryotic species, we further examined the genome of an additional picoeukaryotic strainOstreococcus tauri[52]. Similarly, only a single copy of thePPCgene could be identified (Additional file8)．This is similar to the case observed in bacterial and archaeal genomes, where each species contains only onePPCgene which encodes a single protein (Additional file8) [32].

We further examined the genomic sequences of theAHLgenes and found that the Type-II and -IIIAHLgenes generally contain introns, while the Type-IAHLgenes lack introns in their genomic sequences. This suggests that it is likely that the intron-less Type-IAHLgenes in land plants is the ancestral form from which the two intron-containing types are derived. In each species, there are generally more Type-IAHLgenes in number than either of the other two types (Figure1a). Compared to other families, the Poaceae species have a lower percentage of Type-IIIAHLgenes, includingZea mays[53],Oryza sativa[54],[55] andBrachypodium distachyon[56]. Notably, in年代orghum bicolor[57] we could not detect any Type-IIIAHLs (Figure1a). It is likely that the Type-IIIAHLs arose latest since the mossPhyscomitrella patensand lycophyte年代elaginella moellendorffiicontain only Type-I and -IIAHLs (Figure1a).

Plant introns have been suggested to play important roles in regulating the expression of their associated genes through alternative splicing [58]-[60], nonsense-mediated mRNA decay [61], or intron-mediated transcriptional enhancement [62]. In order to understand the biological functions of the introns in Type-II and -IIIAHLs, we extracted the intron sequences fromArabidopsis thaliana AHLs and examined their capabilities to enhance the transcription of their associated genes using the IMEter 2.0 server [63]. The first introns of severalAtAHLs demonstrated at least a moderate ability to enhance the transcription of their genes (Additional file9a-c). Particularly, the first introns inAtAHL4,6and14are predicted to strongly enhance their transcription.

Monophyletic Clade-A contains type-I AHLs

The early divergence between and significant divergence within the twoAHLclades made analyzing them separately necessary to obtain reliable amino acid alignments. We first performed Bayesian inference analysis on the retrieved Clade-AAHLs. The Clade-AAHLs in land plants is comprised of Type-IAHLs that we have organized for discussion convenience into five subfamilies (Subfamilies A1, A2, A3, A4 and A5) (Figures4and5)．

In order to better understand the evolutionary events which occurred among these five subfamilies, we reconciled the obtained Bayesian tree with the land-plant species tree and inferred whether the internal nodes within the Clade-A Bayesian tree were associated with gene duplication, gene loss, or lineage divergence events. Since their emergence in land plants, theAHLs within this clade have undergone multiple gene duplication events in the early plant lineages. The Subfamily A1-A5AHLs emerged from lineage divergence events after the divergence of lycophyteAHLs and from the rest of vascular plants and further expanded via a series of gene-duplication/divergence events in angiosperms. The emergence of Subfamily A1, A3 and A5AHLs started via gene-duplication events; while, Subfamily A2 and A4AHLs emerged via speciation events.

Within each subfamily of Clade A,AHLgenes from Euphorbiaceae, Salicaceae, Fabaceae, Rosaceae, Brassicaceae and Poaceae families could all be observed, suggesting they may have evolved from one subfamily-specific most common ancestral gene and later functional divergence occurred among these subfamilies. In the extant plant species, theAHLgenes have undergone extensive gene-duplication/loss events (Table1)．The gene duplication events in several extant plant species, such asGlycine max[64] andMalus domestica[65], are probably associated with their recent whole genome duplication events. On the contrary, in several other plant species includingRicinus communis,Carica papaya,Vitis viniferaand monocot species, theAHLgene phylogenies show drastic gene loss events.

Monophyletic Clade-B contains type-II and -III AHLs

Clade-B of theAHLgene family is comprised of Type-II and Type-IIIAHLs (Figures6and7)．The Type-IIAHLs from the early diverging mossPhyscomitrella patensand lycophyte年代elaginella moellendorffiiconstitute a clade at the base of the phylogenetic tree (Figure6)．The angiosperm portion of Clade-B can be divided into four subfamilies (Subfamilies B1, B2, B3 and B4).

In Subfamilies B1 and B4, members of the Type-IIIAHLs tend to group together and form Type-IIIAHLsub-clades (highlighted with gradient shaded box). Individual members of Type-IIAHLs can be observed within the Subfamily B4 Type-III AHL sub-clades. This indicates possible regaining of the Type-I AT-hook motif within this subfamily, suggesting that not all Type-I AT-hooks are homologous. Individual Type-IIIAHLs also exist within the Type-IIAHLsub-clades (such as Subfamilies B2, B3 and B4). This suggests an independent loss of the Type-I AT-hook motifs by AHL proteins within these subfamilies. Taken together, this indicates there are close evolutionary relationships between these two types ofAHLs with, apparently, multiple transitions from Type-II to Type-IIIAHLs, and from Type-III to Type-IIAHLs. The genomes of the mossPhyscomitrella patensand lycophyte年代elaginella moellendorffiido not contain Type-IIIAHLs, suggesting that the loss of the Type-I AT-hook motif in Clade-B occurred after lycophytes diverged from the rest of vascular plants (Figures1a and6)．

年代imilar to their counterparts in Clade A, the Clade BAHLs also experienced multiple gene duplication and loss events during angiosperm diversification (Figures6and7)．年代ubfamily B1-B4AHLs emerged from lineage divergence events and further expanded via multiple gene duplication/loss/divergence events (Table1)．In each extant plant species, Clade-BAHLs experienced similar numbers of gene duplication/loss events as their counterparts in Clade-A, suggesting shared evolutionary pressure between the two clades.

Members of each AHL monophyletic clade share similar expression patterns

To test the hypothesis that Clade-A and -BAHLs evolved independently, we examined the expression patterns of theAHLs inArabidopsis thalianausing Genevestigator V3 [66]. Based on their expression patterns across various tissues at different developmental stages, the 29Arabidopsis thaliana AHLs can be clearly distinguished into two groups (Additional file10)．仔细检查表明,ⅱ型and -IIIAtAHLs tend to share similar expression patterns. Type-II and -IIIAtAHLs, which constitute the Clade-BAHLs, are primarily expressed during seed and flower development. They are only moderately expressed in other tissues. On the other hand, Type-IAtAHLs, which constitute the Clade-AAHLs, are primarily expressed during vascular tissue and root development, which are distinctly different from the expression patterns observed for Type-II and -IIIAHLs. Such distinct expression patterns between the two clades ofAHLs can also be observed inZea mays(Additional file11)．

TheAHLgene family was first described about 10 years ago, as a group of plant-specific genes encoding proteins containing one or two copies of the AT-hook motif and a 120-amino-acid PPC domain [32]. In this study, AHL proteins have been identified in various plant species, including the early diverging mosses and lycophytes, as well as several angiosperm families [46]. We have further classified the AHL proteins into three types based on the number and composition of these two domains. Accordingly, both the AT-hook motifs and PPC domains of the AHL proteins can be classified into two types based on the phylogenetic analysis performed in this study.

From the prokaryotic PPC proteins to the AHL proteins in land plants

The PPC domain found in the AHL proteins exists by itself as a single protein in prokaryotes [32]. Individual strains of Bacteria and Archaea contain one gene encoding a PPC protein (Additional file8)．This observation suggests a role for the PPC domain in fundamental biological processes that has been conserved since prokaryotes throughout evolution. It is intriguing to note that even in the eukaryotic photosynthetic phytoplankton, such asMicromonas pusila[50] andOstreococcus lucimarinus[51), PPC的蛋白质still exists as a single gene. This observation indicates that the association with an AT-hook motif is not necessary for the functions of the PPC protein/domain in prokaryotes and early eukaryotes.

The appearance of the AHL proteins may have occurred between the emergence of the embryophytes and tracheophytes (pointed out by the red star in Figure1a). The primitive AHL proteins emerged when the AT-hook motif fused with the PPC protein between the divergence of picoeukaryotes and the mossPhyscomitrella patens．These primitive proteins later diversified and evolved into two monophyletic clades that comprise the three types of modern AHL proteins found in land plants. However, the evolutionary history of the expansion and later diversifications of theseAHLgenes are yet unexplored.

Ancient events on the AHL evolutionary timeline in land plants

In order to better understand the expansion of the land-plant-specificAHLgenes, we hypothesized the evolutionary events (duplications and deletions) that occurred at common ancestors across land plants (Figure8)．In the embryophytes and tracheophytes, there were few gene duplication/loss events occurring after the emergence ofAHLgenes in bothAHLclades. However, both Clade-A and -BAHLs later experienced rapid expansion in angiosperms, which may be responsible for their large numbers in extant angiosperm species. During the emergence of the grass lineage, Clade-AAHLs exhibited more gene duplications than those in Clade-B. However, during the emergence of eudicots, Clade-BAHLs duplicated more rapidly.AHLs in Clade-B expanded in eudicots mainly through numerous gene duplication events; while those in Clade-A were also coupled with a few gene loss events. With the emergence of rosids, Clade-AAHLs duplicated more than their counterparts in Clade-B. Both clades later experienced dramatic gene losses during the emergence of Malvidae (Eurosids II).

The most dramatic difference between Clade-A and -BAHL年代出现在Fabidae (Eurosid的出现s I). Clade-AAHLs showed rapid birth-and-death events; while the Clade-B copies experienced only gene loss events. This is in direct contrast to theAHLgenes in the emergence of nitrogen fixing species. Clade-AAHLs endured rapid gene losses; while Clade-B copies experienced birth-and-death events. In Malpighiales and Brassicaceae, both clades also emerged through gene birth-and-death events.

A model for the evolutionary history of the AHL gene family in land plants

Based on the results in this study, we propose a model to describe the evolutionary history of theAHLgene family in land plants (Figure9)．In this model, thePPCgene existed by itself and encoded a PPC protein in prokaryotes as well as in early Viridiplantae. Prior to the divergence of extant embryophytes, the PPC domain became associated with a Type-I AT-hook motif to form a primitive intron-lessAHLgene. Another Type-II AT-hook motif was further acquired by this type of primitiveAHLbefore the divergence of mosses from the rest of land plants to form a second type ofAHLgene. This new type ofAHLfurther acquired introns in their genomic sequences. The emergence of both types ofAHLs occurred somewhere between the divergence of picoeukaryotes and mosses. These two types of primitiveAHLs duplicated, differentiated and further developed independently into members of Type-I and -IIAHLs in early land plants, defining the two clades (Clade-A and -B). This model is supported by the observation that only these two types ofAHLs could be found in mosses and lycophytes (Figure1)．Members of the intron-containing Type-IIAHLs further diversified, some losing the Type-I AT-hook motif while retaining the type-II AT-hook motif, forming the intron-containing Type-IIIAHLs. While we have a general idea of when these events occurred, more detailed sampling among green algae, particularly the streptophyte algae, and more land plant lineages (liverworts, hornworts, ferns, gymnosperms, and monocots other than grasses) is needed to fully resolve the timing of gains and losses of the AT-hook motifs and duplication/deletion events.

AHL genes belong to the conserved "Gene Tool Kit" in plant evolution

年代ince they originated and diversified early in land plant evolution, theAHLgenes also belong to the conserved "gene tool kit" of ancient plants. Throughout the evolution of land plants, theAHLs co-evolved with other "tool kit" members to regulate essential biological processes. The proposed co-evolution is supported by the observed genetic interactions with other ancient plant gene families, such as the NAC transcription factors and the MADS-box genes [33],[67]. This hypothesis is also supported by the observed physical interactions of AHL proteins with histones (H2B, H3 and H4), TCPs (TCP4, TCP13, and TCP14), ATAF2 and DELLA proteins [33],[68].

The observed physical interactions of the AHL proteins with other non-AHL transcription factors as well as with themselves led to a recently proposed "enhanceosome" molecular model [33]. In this model, it is proposed that the AHL proteins interact with each other to form homo-/hetero-trimer complexes via their PPC domains [33],[69]. A conserved 6-amino-acid motif in the PPC domain from each monomer AHL protein acts together with those from the other two monomers to compose a quaternary domain. This domain in turn may mediate physical interactions with other transcription factors. In this study, over-expression of oneGlycine maxAHL PPC domain recapitulated the long-hypocotyl phenotype reminiscent of over-expressing itsArabidopsis thalianacounterpart (Figure2)．This indicates that the AHL PPC domains may serve evolutionarily conserved roles in regulating biological processes in multiple plant species. The AHL proteins share similar secondary and tertiary structures (Figure2)．In particular, the 6-amino-acid motif, Gly-Arg-Phe-Glu-Ile-Leu, is highly conserved in the PPC domains of AHLs from all land plants (Additional file4)．Therefore, it is possible that the functional conservation of the AHL proteins is achieved through the preservation of interacting partners among different plant species. In this study, we showed thatArabidopsis thalianaSOB3 / AtAHL29可以与之交互Glycine maxGm06g01650.1 PPC domain (Figure2d,e). This observation supports the hypothesis that the AHL proteins from different species can interact with each other via their PPC domains. It would be interesting to test if the preservation of physical interactions between AHL proteins is also conserved among those from more distantly related plant species, such as between AHLs from the mossPhyscomitrella patensand angiosperms, or from monocot and dicot plants. In addition, we have predicted the orthologous and homologousAHLgenes in the examined plant species (Additional files12and13)．It would be intriguing to examine if the orthologous/homologous AHLs share similar interactions, genetic and/or physical, with other non-AHL orthologous/homologous partners.

Biological functions of the AHL proteins at AT-rich chromosomal DNA

Besides the potential for shared physical interacting partners, the AHL proteins in the land plant species examined in this study also contain either one copy of Type-I or -II AT-hook motifs or both types. These two types of AT-hook motifs have also been found in the mammalian HMGA proteins [34]. Mammalian HMGA1 binds to AT-rich DNA and serves as an architectural protein which alters the local chromatin state and modulates gene expression through both protein-protein and protein-DNA interactions [70]-[73]. The similar possession of AT-hook motifs by both AHLs and HMGAs suggest that they may share binding affinities for AT-rich DNA.

This association of the AT-hook motif with the PPC domain is likely to physically direct these plant PPC domains to AT-rich chromosomal regions. This notion is supported by the observation thatArabidopsis thalianaAHL1 binds to the AT-rich scaffold/matrix attachment regions (S/MARs) and its AT-hook motif is indispensable for AHL1's DNA binding capacity [32]. The S/MARs have been suggested to primarily localize near the transcription start sites [74],[75] or correlate with the origins of DNA replication [76]. SeveralArabidopsis thalianaAHL proteins bind to gene promoter regions and serve as transcriptional regulators [38],[77]. Therefore, it is likely that the potential targeting of the PPC protein to the S/MARs is correlated with functions in gene transcriptional regulation. It would be interesting to examine and compare the biological functions of both PPC proteins in Bacteria and Archaea with the AHL proteins in land plants in order to shed light on the potential evolutionarily conserved functions of this domain.

In this study, we proposed an evolutionary hypothesis for the diversification ofAHLgenes, from a prokaryotic single-copy gene encoding the PPC protein lacking an AT-hook motif, to three types of land plant AHL proteins incorporating two types of PPC domains and two types of AT-hook motifs. However, the biological functions of these three types of AHL proteins still need to be determined. Further experiments need to be performed to reveal their binding sites along plant chromosomes and the corresponding biological regulatory roles. It should be noted that the inferred evolutionary events in this study are based on the retrieved full-lengthAHLsequences available from current releases of completely sequenced plant genomes. Further analysis should incorporate sequences from additional plant species (particularly ferns and gymnosperms) to improve our understanding of the diversification and functional evolution of the three types of AHL proteins.

In this study, over 500 full-lengthAHLgenes have been identified from 19 fully sequenced plant genomes, ranging from the early divergingPhyscomitrella patensand年代elaginellato a variety of monocot and dicot flowering plants. Our analyses suggest that theAHLs can be classified into three types (Type-I/-II/-III) based on the number and composition of their functional domains, the AT-hook motif(s) and PPC domain. We further inferred their phylogenies in land plants via Bayesian inference analysis. TheAHLgenes emerged in embryophytes and have evolved into two distinct clades with Type-IAHLs diversifying in Clade-A and the other two types together diversifying into Clade-B. Our study indicates that Clade-A and -BAHLs diverged before the separation of mossPhyscomitrella patensfrom the vascular plant lineage. In angiosperms, Clade-AAHLs expanded into 5 subfamilies; while, the ones in Clade-B expanded into 4 subfamilies.

Examination of their expression patterns suggests that theAHLs within each clade share similar patterns of expression with each other. While, theAHLs between the two clades exhibit distinct expression patterns from each other, suggesting potential conserved biological functions within each clade since their divergence along land plant evolution.

Manipulating theAHLgenes has been suggested to have tremendous effects to positively affect agriculture through increasing seedling establishment, plant biomass and improving plant immunity [33],[42],[78],[79]. Our analyses suggest that theAHLgenes from different land plant species may share conserved functions in regulating plant growth and development. Over-expression of aGlycine maxAHL PPC domain inArabidopsis thalianarecapitulates the phenotype observed when over-expressing itsArabidopsis thalianacounterpart. Our study further suggest that such functional conservation may be due to conserved physical interactions among the PPC domains of AHL proteins. In the end, our analyses reveal a possible evolutionary scenario for theAHLgene family in land plants, which will facilitate the design of new studies probing their biological functions and subsequently lead to improvements in crop biomass production.

Data retrieval

The amino acid sequences as well as coding sequences for the members of theArabidopsis thaliana AHLgene family were retrieved from the TAIR website [32],[41],[80] and were further used as queries for gene search using BLAST, TBLASTN, BLASTP and PSI_BLAST forAHLgenes in the Phytozome database [46] within the related plant species with a cut-off E value set at 1e^-2．The obtained results were further used as additional queries. Only intact gene sequences comprised of both AT-hook motif(s) and PPC domain were included and used as additional queries to perform in-depth gene searches in the Phytozome database. For further phylogenetic analysis, only protein sequences were used.

Cloning of AHL genes from Glycine max and Triticum aestivum

Genomic DNA as well as mRNA were prepared fromTriticum aestivumseedlings using DNeasy plant mini kit (Qiagen) and RNeasy plant mini kit (Qiagen). cDNA was further prepared using iScript Advanced cDNA Synthesis Kit for RT-PCR (Bio-RAD). Primer pairs (TaAHL1: 5'-ATG GGG AGC ATG GAC GGC CAC CC-3' and 5'-CTA GAA TGA CGT CGG CGG AGG CCG C-3'; TaAHL3: 5'-ATG GCC ACC GGC AGC AGC AAG TGG TG-3' and 5'-TCA GAT GCC GCC TCC CTG GTG GCC TC-3') were used to cloneTaAHL1andTaAHL3from both prepared genomic DNA and cDNA, correspondingly, and examined to be free-of-introns. Amino acid sequences of TaAHL1 and TaAHL3 proteins were predicted from their coding sequences and were used for further phylogenetic analysis. The nucleotide sequences ofTaAHL1andTaAHL3have been deposited into NCBI GenBank (Accession numbers:TaAHL1/Taq1, KJ461850;TaAHL3/Taq3, KJ461851). Genomic DNA ofGm06g01650.1was prepared from 6 day-old seedlings using ZR Plant DNA MiniPrep kit (Zymo Research). Coding sequence of its PPC domain was cloned using the primer pair (5'-TCC CCC CGG G A TGA AGC CAC CCG TCA TAG TCA CGC GCG AC-3' and 5'-AAC TGC AGT CAA TCA TCA TCA TGC TGA TTC AAG G-3'). The amplicon and binary vectorpCHF3were digested by XmaI with PstI and ligated together. The resulted plasmid was subsequently transformed into agrobacteriumGV3101and further transformed intoArabidopsis thaliana Col-0by the floral dipping method [81]. Surface-sterilized seeds were sown on 0.5× Linsmaier and Skoog modified basal medium (1.0% w/v phytogel and 1.5% w/v sucrose) and grown for 5 days at 25°C under 25 μmol∙s^-1-m^-2white light in a Percival E-30B growth chamber.

Yeast two-hybrid assay

AlexA-based Y2H system was used to test protein-protein interactions in yeast. The targeted yeast two-hybrid assay was performed as described in [33].

年代equence alignment and phylogenetic analyses

The amino acid sequences of the Type-I AHL proteins were aligned using MUSCLE [82],[83] and were further manually adjusted. Bayesian inference analysis was performed with the MrBayes 3.2.1 on XSEDE tool on CIPRES Science Gateway for 20 million generations with convergence at 0.022 [84]. The amino acid sequences of the Type-II and -III AHL proteins were aligned and manually adjusted. Bayesian inference analysis was performed with the MrBayes 3.2.1 for 10 million generations with convergence at 0.017. Generations were both sampled every 10,000 generations and the first 25% was set as burn-in.

年代econdary and tertiary structure prediction

The amino acid sequences of the PPC domains were retrieved from the coding sequences ofGm06g01650.1, TaAHL1(Taq1) andTaAHL3(Taq3)．The secondary and tertiary structures were predicted using the RaptorX Structure Prediction Server [85],[86]. The tertiary structure figures were prepared using Pymol version 1.3 (The PyMOL Molecular Graphics System, Schrodinger, LLC).

Inference of gene duplication and loss event

The plant species tree was adapted from the one in the Phytozome database [46]. The gene trees obtained from Bayesian inference analysis for each of the two AHL clades were reconciled with the plant species tree individually by Notung 2.6 [87] with default parameters. The orthologous and paralogous genes were further inferred by the Notung 2.6 program [87].

Availability of supporting data

The wheatAHLgenes,TaAHL1andTaAHL3were deposited into NCBI GenBank (Accession numbers:TaAHL1/Taq1, KJ461850;TaAHL3/Taq3, KJ461851). All supporting data are included as additional files and have been uploaded to LabArchives, LLC. DOI: 10.6070/H4PC30B2.

JZ conceived of the study, participated in its design and coordination, collected the sequences, performed bioinformatics analysis, clonedGm06g01650.1-PPC序列,进行了相关的转基因研究nd wrote the manuscript. DSF performed the yeast two-hybrid analysis and participated in writing the manuscript. JQ clonedTaAHL1/Taq1andTaAHL3/Taq3from wheat. EHR participated in the design of the study, participated in performing the bioinformatics analysis and writing the manuscript. MMN participated in the design and coordination of the study and writing the manuscript. All authors read and approved the final manuscript.

AHL:

AT-hook motif nuclear localized

PPC/DUF296:

Plant and prokaryote conserved/domain of unknown function #296

年代OB3:

年代uppressor of phytochrome B-4 #3

ESC:

ESCAROLA

HRC:

HERCULES

Al:

Arabidopsis lyrata

At:

Arabidopsis thaliana

Bd:

Brachypodium distachyon

Bra:

Brassica rapa

Cpa:

Carica papaya

Cs:

Cucumis sativus

Gm:

Glycine max

Mdp:

Malus domestica

Mes:

Manihot esculenta

Mt:

Medicago truncatula

Os:

Oryza sativa

Pp:

Physcomitrella patens

Ppa:

Prunus persica

Pt:

Populus trichocarpa

Rc:

Ricinus communis

年代b:

年代orghum bicolor

年代m:

年代elaginella moellendorffii

Vv:

Vitis vinifera

Zm:

Zea mays

A:

Angiosperms

B:

Brassicaceae

Em:

Embryophyta

Eu:

Eudicots

F:

Fabidae (Eurosids I)

G:

Grasses

Mp:

Malpighiales

Mv:

Malvidae (Eurosids II)

NF:

Nitrogen fixing

T:

导管植物(维管植物)

这个项目是由农业和支持Food Research Initiative competitive grant # 2013-67013-21666 of the USDA National Institute of Food and Agriculture (to M. M. N.), the O.A. Vogel Wheat Research Fund (to M.M.N.) and the Washington Grain Commission (to M. M. N.). This project was also supported by Global Plant Sciences Initiative Research Fellowship (Washington State University, to J. Z.), Pacific Seed Association Fellowship (to J. Z.), Maguire International Seed Technology Fellowship (to J. Z.), Lindahl Memorial Scholarship (to J. Z.) and Roscoe & Francis Cox Scholarship (to J. Z.). We are also grateful for support from the Brubbaken and Reinbold Monocot Breeding Fund (to M. M. N.).

Affiliations

1. Molecular Plant Sciences Graduate Program, Washington State University, Pullman, WA, 99164, USA

   Jianfei Zhao, David S Favero, Eric H Roalson & Michael M Neff

2. Department of Crop and Soil Sciences, Washington State University, Pullman, WA, 99164, USA

   Jianfei Zhao, David S Favero, Jiwen Qiu & Michael M Neff

3. 年代chool of Biological Sciences, Washington State University, Pullman, WA, 99164, USA

   Eric H Roalson

4. Present Address: Department of Biology, University of Pennsylvania, Philadelphia, PA, 19104, USA

   Jianfei Zhao

Corresponding author

Correspondence toJianfei Zhao．

Competing interests

The authors declare that they have no competing interests.

这篇文章是发表在生物医学的许可证Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Reprints and Permissions