UBC Faculty Research and Publications

Ancient origin of the biosynthesis of lignin precursors Labeeuw, Leen; Martone, Patrick T; Boucher, Yan; Case, Rebecca J May 21, 2015

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-13062_2015_Article_52.pdf [ 2.53MB ]
JSON: 52383-1.0223559.json
JSON-LD: 52383-1.0223559-ld.json
RDF/XML (Pretty): 52383-1.0223559-rdf.xml
RDF/JSON: 52383-1.0223559-rdf.json
Turtle: 52383-1.0223559-turtle.txt
N-Triples: 52383-1.0223559-rdf-ntriples.txt
Original Record: 52383-1.0223559-source.json
Full Text

Full Text

RESEARCHsbctunbthwrigidity and strength to stems [4,5], as demonstrated by propanoid pathway, which can generate precursors toLabeeuw et al. Biology Direct  (2015) 10:23 DOI 10.1186/s13062-015-0052-y(p-coumaryl alcohol), through a series of reduction2E9, CanadaFull list of author information is available at the end of the articlelignin-deficient mutants with low structural support[6,7]. The hydrophobic nature of lignin makes it imper-meable to water and is thus crucial for a vascular system,enabling transport of water throughout the plant [5]. Inaddition, lignin has been suggested to protect plantsa diverse group of compounds including flavonoids,coumarins, quinones, and monolignols (Figure 1). Thestarting point of this pathway is the amino acid phenyl-alanine, which is deaminated to form cinnamic acid,followed by p-coumaric acid and p-coumaroyl-CoA[11]. The lignin specific pathway then uses p-coumaroyl-CoA to produce the simplest monolignol, H monolignol* Correspondence: rcase@ualberta.ca1Department of Biological Sciences, University of Alberta, Edmonton, AB, T6Gary cell wall and provides structural support, givingphylogenetic analysis of these genes to decipher the evolution and origin(s) of lignin biosynthesis.Results: Enzymes involved in making p-coumaryl alcohol, the simplest lignin monomer, are found in a variety ofphotosynthetic eukaryotes, including diatoms, dinoflagellates, haptophytes, cryptophytes as well as green and redalgae. Phylogenetic analysis of these enzymes suggests that they are ancient and spread to some secondarilyphotosynthetic lineages when they acquired red and/or green algal endosymbionts. In some cases, one or more ofthese enzymes was likely acquired through lateral gene transfer (LGT) from bacteria.Conclusions: Genes associated with p-coumaryl alcohol biosynthesis are likely to have evolved long before thetransition of photosynthetic eukaryotes to land. The original function of this lignin precursor is therefore unlikely tohave been related to water transport. We suggest that it participates in the biological defense of some unicellularand multicellular algae.Reviewers: This article was reviewed by Mark Ragan, Uri Gophna, Philippe Deschamps.Keywords: Lignin, Monolignol, Algae, Evolution, Haptophyte, Chlorophyte, Rhodophyte, Diatom, Cryptophyte,DinoflagellateBackgroundLignin is a complex and highly recalcitrant form of car-bon often thought to be one of the key innovations ofland plants, allowing the movement of plants fromaquatic habitats to terrestrial ecosystems [1,2]. It is thesecond most abundant biopolymer on earth, after cellu-lose [3]. In land plants, lignin is deposited in the second-from biological attacks. Being recalcitrant to biologicaldegradation, increased lignification can offer protectionwhen a plant is damaged [8]. The precursors of lignin(p-coumaric acid, p-coumaroyl-CoA, p-coumaraldehyde) andits monomers (monolignols p-coumaryl alcohol, coniferylalcohol and sinapyl alcohol) have also been suggested tohave antimicrobial properties [9,10].Lignin biosynthesis starts with the general phenyl-Ancient origin of the bioprecursorsLeen Labeeuw1, Patrick T Martone2, Yan Boucher1 and ReAbstractBackground: Lignin plays an important role in plant struof the hallmarks of land plants. The recent discovery of ligon the evolution of its biosynthetic pathway, which coulddetermine the taxonomic distribution of the lignin biosynof algae and their closest non-photosynthetic relatives, as© 2015 Labeeuw et al.; licensee BioMed CentrCommons Attribution License (http://creativecreproduction in any medium, provided the orDedication waiver (http://creativecommons.orunless otherwise stated.Open Accessynthesis of ligninecca J Case1*ral support and water transport, and is considered onein or its precursors in various algae has raised questionse much more ancient than previously thought. Toesis genes, we screened all publicly available genomesell as representative land plants. We also performedal. This is an Open Access article distributed under the terms of the Creativeommons.org/licenses/by/4.0), which permits unrestricted use, distribution, andiginal work is properly credited. The Creative Commons Public Domaing/publicdomain/zero/1.0/) applies to the data made available in this article,Labeeuw et al. Biology Direct  (2015) 10:23 Page 2 of 21reactions that modify the side chain [7] (Figure 1). Produc-tion of the more structurally complex G (coniferyl alcohol)and S (sinapyl alcohol) monolignols requires additionalenzymes using intermediates of H monolignol synthesis assubstrates, which perform O-methylation and hydroxyl-ation at various sites on the phenolic ring [2]. After theirbiosynthesis inside the cell, the monolignols are trans-ported to the cell wall through an unknown mechanism,before undergoing cross-linking of the monomers to formthe lignin polymer. This process, called lignification, is notwell understood but involves simple radical coupling reac-tions between monolignols and a growing lignin polymerFigure 1 The lignin biosynthesis pathway. Coloured dots represent the preused for the enzyme are: phenylalanine ammonia-lyase (PAL); cinnamate 4-hyd(CCR); cinnamyl alcohol dehydrogenase (CAD); peroxidase (PER); and laccase (Lfunction or its substrate is unknown, or that the presence of a particular compdemonstrated.[12]. The simple coupling reactions are believed to be car-ried out by peroxidases and laccases, enzymes that are di-verse and abundant in eukaryotes [3,13].The lignin biosynthesis pathway has been extensivelystudied in vascular plants, and was thought to be a hall-mark of this group [14]. Gymnosperms (softwood) areusually primarily composed of G lignin [15], angiospermdicots (hardwood) are composed of predominantly Gand S lignins while monocots (such as grasses) usuallyhave more H lignin content [3,13]. Investigation of socalled “lower” (non-vascular) plants revealed the evolutionof lignin biosynthesis to be more ancient than the originsence of a given enzyme in a specific taxonomic group. The abbreviationsroxylase (C4H); 4-coumarate:CoA ligase (4CL); cinnamoyl-CoA reductaseAC). Question marks (?) indicate that the enzyme responsible for a specificound in a given taxonomic group is hypothesized but has not beenLabeeuw et al. Biology Direct  (2015) 10:23 Page 3 of 21of vascular plants. Selaginella moellendorfii (spikemoss),which is part of the most ancient lineage of vascularplants, the lycophytes, can synthesize all three types of lig-nin, including the supposedly-derived S lignin, which issynthesized using an enzyme that is absent from all otherland plants [16,17]. Therefore, S. moellendorfii likely de-veloped this ability independently of angiosperms, whichwere previously the only organisms believed capable ofsynthesizing S lignin. Even the moss Physcomitrellapatens, a representative of the most ancient group of landplants, the non-vascular bryophytes, encodes all ligninbiosynthesis enzymes necessary to synthesize H and G lig-nins in its genome [4]. Although true polymerized ligninhas yet to be found in this organism, it contains mono-lignols and lignin-like molecules, [4,18,19]. More surpris-ing still is that H, G and S lignins have recently beenidentified in the cell walls of the calcified red alga Calliar-thron cheilosporioides (Rhodophyta) [20]. Given that landplants are separated by a large evolutionary distance fromrhodophytes, convergent evolution of lignin biosyntheticgenes has been suggested as an explanation of how C.cheilosporioides acquired lignified cells [13]. There is someevidence that lignin might also be found in other algae be-sides rhodophytes, but it is still heavily debated. Homologsof some of the genes for lignin biosynthesis have alsobeen found in green algae and diatom species [19].Charophytes, a class of freshwater streptophyte algae, havebeen suggested to contain lignin-like compounds [21,22],as have some brown algae [23-25]. Later studies arguedagainst this, claiming that brown algae contain only phe-nol compounds and not specifically lignin [12,26].Recent research has found that the lignin precursor, p-coumaric acid, as well as key genes of the lignin biosyn-thesis pathway are present in the haptophyte Emilianiahuxleyi [27]. Another study also found coumaric acid aswell as flavonoid compounds derived from it, in a dia-tom (Phaedactylum tricornutum), a haptophyte (Dia-cronema lutheri) and several green algae (Tetraselmissuecica, Chlorella vulgaris and Haematococcus plu-viaris) [28]. It is surprising that haptophytes and dia-toms contain a lignin precursor, as well as the genespotentially encoding enzymes to convert it into mono-lignol. Hapthophytes and diatoms are evolutionarilyvery distant to plants, red and green algae and to eachother. The potential for these two phylogenetically dis-parate phytoplanktons to synthesize monolignols sug-gests that the pathway for lignin synthesis might haveevolved in ancient oceans and may vastly predate theorigin of land plants. To investigate this possibility, wescreened all available algal genomes for the presence ofthe lignin biosynthesis pathway, in addition to a selec-tion of representative species of land plants and othereukaryotes belonging to various supergroups (unikonts,archaeplastids, stramenopiles, alveolates, cryptophytes,haptophytes). Although known genes required for thesynthesis of G and S lignins were mostly absent outsideof land plants, homologs of those needed for makingthe monomer of the simplest form, p-coumaryl alcohol(H monolignol), were found in green algae, red algaeand glaucophytes, as well as multiple organisms in super-groups other than Archaeplastida. Phylogenetic analysis ofthe p-coumaryl alcohol biosynthesis genes suggests anearly origin for this metabolic process, followed by at leastthree examples of independently evolved enzymes allow-ing organisms to make more complex derivatives, G and Smonolignols, in red algae, club mosses and the ancestor ofvascular plants.Results and discussionThe lignin biosynthetic pathway has a conserved andtaxonomically widespread coreAn extensive screen for homologs of the known ligninbiosynthesis genes was performed across all domains oflife, with a specific focus on eukaryotes (Tables 1 and2). Previous research has focused on lignin biosynthesisin Arabidopsis [29-32] and other model land plants[2,19,33-35], so only representative species of this grouphave been included in our search. As expected, all the lig-nin biosynthesis genes were found in all land plants forwhich genomes are available, with the exception of feru-late 5-hydroxylase (F5H), which is absent from the onlybryophyte (moss) in our dataset, Physcomitrella patens[7]. Surprisingly, four gene families had a wide distributionacross the various eukaryotic supergroups and were notrestricted to land plants. These include: 4-coumarate:CoAligase (4CL); cinnamoyl-CoA reductase (CCR); cinnamylalcohol dehydrogenase (CAD); and caffeoyl-CoA O-meth-yltransferase (CCoAMT) (Tables 1 and 2).The first three of these enzymes (4CL, CCR andCAD) catalyze consecutive steps in lignin biosynthesisand are sufficient to produce p-coumaryl alcohol fromcoumaric acid (Figure 1). Homologs of all three enzymeshave similar taxonomic distributions, being found mostlyin marine photosynthetic algae in addition to land plants.Representatives of green algae, red algae, glaucophytes, di-atoms, dinoflagellates, haptophytes and cryptophytes, aswell as the non-photosynthetic oomycetes, harbor homo-logs of these three enzymes. Although oomycetes are non-photosynthetic, they are believed to share a photosyntheticancestor with other stramenopiles such as brown algaeand diatoms [36]. If the 4CL, CCR and CAD homologs inthese diverse eukaryotes indeed catalyze the same bio-chemical reactions as their plant homologs, the capacityto make at least the precursor of the simplest form of lig-nin (H lignin) would be much more widespread than cur-rently thought (land plants and red algae).We have performed functional prediction analysis onall homologs of 4CL, CCR and CAD homologs usinglaDLabeeuw et al. Biology Direct  (2015) 10:23 Page 4 of 21Table 1 Distribution of lignin biosynthesis genes in archaepOrganism Phylum PAL C4H 4CL CCR CAARCHAEPLASTIDSLand plantsthe Argot2 [37] and ESG [38] packages, which are amongthe best performing functional annotation programs avail-able [39] (Figure 2). Homologs of plant CCR and CADwere predicted to have a conserved function with moder-ate to high confidence in at least one representative of thegreen algae, red algae, diatoms, dinoflagellates, hapto-phytes and cryptophytes. For CAD, alcohol dehydrogenaseArabidopsis thaliana Streptophyta + + + + +Oryza sativa Streptophyta + + + + +Physcomitrella patens Streptophyta + + + + +Picea abies Streptophyta + + + + +Selaginella moellendorffii Streptophyta + + + + +Green algaeBathycoccus prasinos Chlorophyta +GBotryococcus braunii Chlorophyta +EC. reinhardtii Chlorophyta + +Chlorella sp. NC64A Chlorophyta +G +GChlorella variablilis Chlorophyta +G +G +GCoccomyxa subellipsoidea Chlorophyta +G + +Dunaliella salina Chlorophyta +EHaematococcus pluvialis Chlorophyta +EM. pusilla CCMP1545 Chlorophyta +GMicromonas sp. RCC299 Chlorophyta +GO. lucimarinus Chlorophyta +Ostreococcus sp.RCC809 Chlorophyta +GOstreococcus tauri Chlorophyta +GPolytomella sp. Chlorophyta +GPolytomella parva Chlorophyta +EPrototheca wickerhamii Chlorophyta +EVolvox carteri Chlorophyta +G +Red algaeC. tuberculosum Rhodophyta +G +G +GChondrus crispus Rhodophyta +G +G +GCyanidioschyzon merolae Rhodophyta +G +GGaldiera sulphararia Rhodophyta +G +GPorphyridium cruentum Rhodophyta +E +E +EGlaucophytesCyanophora paradoxa Glaucophyta +G +G +GFootnote = + Present in both genome sequence and EST library, +E Present in ESTdetermined by a reciprocal BLASTP hit with an E-value < 1x10-30 or less using the charNCBI, JGI, and Congenie databases. The abbreviations used for the enzyme can b4-hydroxylase (C4H); 4-coumarate:CoA ligase (4CL); cinnamoyl-CoA reductase (CCp-coumarate 3-hydroxylase (C3H); caffeic acid O-methyltransferase (COMT); caffeoshikimate esterase (CSE) and Selaginella moelledorfii F5H (smF5H); peroxidase (PERstid genomesHCT C3H COMT CCoAMT F5H CSE smF5H PER LACfunction was often predicted along with cinnamyl alcoholdehydrogenase function with comparable confidence. Theprediction for the specific function of 4-coumarate:CoAligase was difficult for 4CL homologs, being weak even forenzymes that have been biochemically characterised ashaving that function (such as several of Arabidopsis thali-ana paralogous 4CL enzymes) [7]. However, ligase+ + + + + + + ++ + + + + + + ++ + + + + + ++ + + + + + + ++ + + + + + + + ++G+E+G +G+G +G+G +G+ +G +G+G+G+G+G+G +G+G+G+Elibrary only, +G Present in genome sequence only. Presence (+) isacterized land plant A. thaliana or S. moellendorffii gene as a query, searching thee described as follows: phenylalanine ammonia-lyase (PAL); cinnamateR); cinnamyl alcohol dehydrogenase (CAD); p-hydroxycinnamoyl-CoA (HCT);yl-CoA O-methyltransferase (CCoAMT); ferulate 5-hydroxylase (F5H); caffeoyl); and laccase (LAC).haLabeeuw et al. Biology Direct  (2015) 10:23 Page 5 of 21Table 2 Distribution of lignin biosynthesis genes in non-arcOrganism Phylum PAL C4H 4CL CCRSTRAMENOPILESfunction was predicted with high or very high confidencefor all organisms with predicted CCR and CAD functions.Although homologs of all three enzymes are also found inglaucophytes and oomycetes, functional predictions wereweak or absent for one or more of these enzymes in thesetwo taxonomic groups.DiatomsFragilariopsis cylindrus Bacillariophyta + +EPhaeodactylum tricornutum Bacillariophyta + +T. pseudonana Bacillariophyta +GPelagophyteA. anophagefferens Heterokontophyta +GBrown algaeEctocarpus siliculosus Phaeophyceae +GEustigmatophytesN. gaditana Eustigmatophyceae +G +GChrysophytesOchromonas danica ChrysoophyceaeOomycetesAlbugo laibachii Heterokontophyta +G +GPhytophthora infestans Heterokontophyta + +Phytophthora sojae Heterokontophyta +ALVEOLATESApicomplexaCryptosporidium muris Apicomplexa +GCryptosporidium parvum Apicomplexa +GTheileria parva ApicomplexaToxoplasma gondii ApicomplexaDinoflagellatesSymbiodinium minutum Dinoflagellata +G +GCiliatesParamecium tetraurelia Ciliophora +GT. thermophila Ciliophora +G +GHAPTOPHYTESE. huxleyi CCMP1516 Haptophyta + +CRYPTOPHYTESGuillardia theta Cryptophyta +G +GHemiselmis andersenii CryptophytaFootnote = + Present in both genome sequence and EST library, +E Present in EST librareciprocal BLASTP hit with an E-value < 1x10-30 or less using the characterized land plaMarine Genomics Unit (http://marinegenomics.oist.jp/genomes/gallery/) and Parameciuenzyme can be described as follows: phenylalanine ammonia-lyase (PAL); cinnamate 4cinnamyl alcohol dehydrogenase (CAD); p-hydroxycinnamoyl-CoA (HCT); p-coumarateO-methyltransferase (CCoAMT); ferulate 5-hydroxylase (F5H); caffeoyl shikimate esteraslaccase (LAC).eplastid genomesCAD HCT C3H COMT CCoAMT F5H CSE smF5H PER LACMost enzymes involved in lignin biosynthesis are multi-functional or have multiple, slightly divergent, paralogouscopies with different functions, either with the direction ofthe reaction catalyzed reversed or a change in substrate af-finity. This is especially true for CAD, which catalyzes thelast step in monolignol biosynthesis. SAD (sinapyl alcohol+ ++ +E+G+E+G ++ ++ ++++G++ + +G+G +Gry only, +G Present in genome sequence only. Presence (+) is determined by ant A. thaliana or S. moellendorffii gene as a query, searching the NCBI, JGI, OISTm (http://paramecium.cgm.cnrs-gif.fr/) databases. The abbreviations used for the-hydroxylase (C4H); 4-coumarate:CoA ligase (4CL); cinnamoyl-CoA reductase (CCR);3-hydroxylase (C3H); caffeic acid O-methyltransferase (COMT); caffeoyl-CoAe (CSE) and Selaginella moelledorfii F5H (smF5H); peroxidase (PER); andDLabeeuw et al. Biology Direct  (2015) 10:23 Page 6 of 214CL* CCR CAARCHAEPLASTIDSLand plantsArabidopsis thaliana F F FOryza sativadehydrogenases) cannot be differentiated from CADphylogenetically and both enzymes display some of theother’s specific activity [40,41]. Also, there are many en-zymes with CAD activity (oxidation of an aldehyde to analcohol) as well as alcohol dehydrogenase activity (reduc-tion of an alcohol to an aldehyde), such as yeast ADH6and ADH7 enzymes [42]. Given the functional flexibilityin these protein families, functional switches are likely toPhyscomitrella patensPicea abiesSelaginella moellendorffiiGreen algaeBathycoccus prasinosChlamydomonas reinhardtii Chlorella sp. NC64AChlorella variablilisCoccomyxa subellipsoideaHaematococcus pluvialisM. pusilla CCMP1545 Micromonas sp. RCC299O. lucimarinus Ostreococcus sp.RCC809 Ostreococcus tauriPolytomella sp.Polytomella parvaPrototheca wickerhamiiVolvox carteri Red algaeCalliarthron tuberculosumChondrus crispusCyanidioschyzon merolaeGaldiera sulpharariaPorphyridium cruentumGlaucophytesCyanophora paradoxaVery high coHigh confideModerate coBelow modeHomolog preFFigure 2 Functional prediction of the p-coumaryl alcohol biosynthesis pand ESG) need to predict the correct function for it to be annotated as such.biochemically characterized. No single fungi or bacteria harbour homologs forepresentatives of these taxonomic groups. An empty space indicates that nofrom the ESG software. * for 4CL, as even enzymes with biochemically demonby the softwares used, the figure indicates prediction of ligase activity.4CL* CCR CADSTRAMENOPILESDiatomsFragilariopsis cylindrusP.  tricornutum have occurred frequently throughout eukaryotic evolution,making exact functional predictions difficult without bio-chemical data.p-coumaryl alcohol synthesis is likely common inphotosynthetic eukaryotesHomologs for all genes of the 4CL-CCR-CAD pathway,responsible for the synthesis of p-coumaryl alcohol fromT. pseudonanaBrown algaeEctocarpus siliculosusEustigmatophytesN. gaditanaOomycetesAlbugo laibachiiPhytophthora infestansPhytophthora sojaeALVEOLATESApicomplexaCryptosporidium murisCryptosporidium parvum CiliatesParamecium tetraureliaT. thermophilaDinoflagellatesSymbiodinium minutumHAPTOPHYTESE. huxleyi CCMP1516 CRYPTOPHYTESGuillardia thetaFUNGI FBACTERIA F Fnfidence: >70%nce: >50%nfidence: >= 30%rate confidence: <30%sent, no functional predictionathway genes. Both programs used for functional prediction (Argot 2An "F" indicates that an enzyme from that taxonomic group has beenr all three enzymes, but all the enzymes are found individually in somehomolog is found in a particular species. Confidence values are derivedstrated function are not annotated as such with significant confidenceLabeeuw et al. Biology Direct  (2015) 10:23 Page 7 of 21p-coumaric acid, occur in several widely divergenteukaryotic taxonomic groups (stramenopiles, haptophytes,cryptophytes, dinoflagellates and archaeplastids). This doesnot agree with a previously proposed origin in early landplants [3]. Several alternative hypotheses could explain thisdistribution. The genes of this pathway (which will be calledthe p-coumaryl alcohol biosynthesis pathway henceforth)could have been present in eukaryotic ancestors predatingthe origin of some or all of these groups, and lost in line-ages in which the pathway is absent. The superfamilies towhich 4CL (adenylate-forming enzymes), CCR (adenylatereductase) and CAD (medium-chain dehydrogenase/reduc-tases) belong are most certainly ancestral to eukaryotes, be-ing widespread in all eukaryotic supergroups [35,43-47].However, close homologs to characterized plant 4CL, CCRand CAD enzymes with a correctly predicted function arenot as ubiquitous. The complete p-coumaryl alcohol bio-synthesis pathway is present almost solely in photosyntheticeukaryotes, making a scenario in which they evolve in anancient eukaryotic ancestor and are subsequently selectivelylost in non-photosynthetic taxonomic groups unlikely. Amore parsimonious explanation could be that this pathwaywould have originated in an ancestor of green and red algaeand was subsequently transferred to other taxonomicgroups in the same way they acquired the capacity tophotosynthesize: through taking up a red or green alga in asecondary endosymbiotic event (endosymbiotic gene trans-fer or EGT) [36]. Another possibility is that several genes inthis pathway were acquired independently by lateral genetransfer (LGT) from bacteria or heterotrophic eukaryotes invarious photosynthetic lineages. Phylogenetic analysis ofthe 4CL, CCR and CAD enzymes were performed to differ-entiate between these hypotheses on the origin(s) of the p-coumaryl alcohol biosynthesis pathway.p-coumaryl alcohol biosynthesis could have originated inan ancient archaeplastidThe p-coumaryl alcohol biosynthesis pathway seems tobe ancestral to green algae, but with frequent loss of thismetabolic function throughout this taxonomic group(Figure 2). All green algal species are found within a sin-gle well-supported (97%) clade in the 4CL tree (mixedwith red algae and secondarily photosynthetic organ-isms) (Figure 3) and in two clades in the CCR tree andthree clades in the CAD tree (Figure 4 and 5). The mul-tiple green algal clades in the CAD tree likely representconservation of a different paralog in two major greenalgal lineages; prasinophytes and the core chlorophytes[48]. The core chlorophytes are monophyletic and groupwith land plants, while the prasinophytes are divided intwo clades, the result of multiple divergent paralogs be-ing present in Bathycoccus prasinos and Ostreococcuslucimarinus. In the CCR tree, the two green algal cladesare both composed of core chlorophytes, as well asoverlapping in their species content and are proximal toeach other. They likely represent paralogy originatingafter the divergence of core chlorophytes from ancestralgreen algae. Despite paralogy being observed in greenalgae for all three genes, the presence of p-coumaryl al-cohol biosynthesis gene homologs in various speciesfrom two major green algal groups suggests the presenceof this pathway is ancestral. Green algae are never partof clades containing organisms from other taxonomicgroups besides secondarily photosynthetic species andred algae, suggesting the absence of LGT for p-coumarylalcohol biosynthesis gene homologs in this lineage. Thefact that only two of fifteen green algal genomes or ESTlibraries screened contain homologs to all three p-cou-maryl alcohol biosynthesis enzymes suggests this func-tion has been frequently lost and is likely non-essential,as compared to its crucial role in most land plants.The story is different for red algae. In this group, only4CL seems to be ancestral, being found in a single clade,which also contains green algae and secondarily photo-synthetic organisms (Figure 3). The only exceptions tothis are two of the five 4CL paralogs in Calliarthrontuberculosum clustering in bacterial clades, one of themvery strongly with a marine α-proteobacteria groupknown as the roseobacter clade, indicating a likely LGTfrom bacteria. For both CCR and CAD, red algae arepolyphyletic, with some paralogs weakly clustering withhomologs from bacteria or heterotrophic eukaryotes. Itis therefore difficult to say whether the pathway was an-cestrally found in red algae or some of its enzymes wereacquired by LGT. Like in green algae, only a small pro-portion of red algal genomes investigated have homologsfor all three enzymes (one in five). Although the fact thatthe red alga Calliarthron cheilosporioides, a close relativeto C. tuberculosum, can synthesize p-coumaryl alcohol iswell established [20], this function does not seem to bean essential feature of red algae, making the loss of itlikely. This is perhaps visible in the specialized role oflignin in the uncommon lignified joints, genicula, of C.cheilosporioides [20].These results suggest that all three genes of the p-coumaryl alcohol biosynthesis pathway are likely tohave been present in at least the shared ancestor ofplants and green algae, but possibly earlier, beforethe speciation of the red algal ancestor (Figure 6). Afew additional features of the phylogenies support anearlier origin than the ancestor of land plants. TheCAD homologs found in core chlorophytes clusterwith land plants (albeit with modest 74% bootstrapsupport). Also, although both the CCR and CADphylogenies likely contain hidden ancestral paralogyand differential loss, this is not the case for the 4CLtree. Paralogy is evident only for red and green algaeand confined to a single clade (with the exception ofLabeeuw et al. Biology Direct  (2015) 10:23 Page 8 of 21copies likely acquired from bacteria by C. tuberculo-sum). As red and green algal species are polyphyleticinside this clade, duplication(s) of the 4CL homologlikely occurred in their ancestor and therefore wouldpredate these lineages. It is difficult to say if thepathway could have been present in the archaeplastidFigure 3 Maximum likelihood phylogeny of the 4-coumarate:CoA ligaseAmino acid sequences were aligned with MUSCLE and the tree compiled usinLuciferase, a 4CL homolog [79], was used as the outgroup. Gene names are incenzyme has been biochemically characterized in this organism [29,72,73].ancestor itself, before divergence of the glaucophytes.Glaucophyte is the earliest branching lineage in archae-plastids [49] and could potentially be quite informative onthe origin of the p-coumaryl alcohol biosynthesis pathwayin this group. Unfortunately, although it harbours homo-logs for all three enzymes of the pathway, none of(4CL) enzyme from the p-coumaryl alcohol biosynthesis pathway.g RaxML. Numbers above branches refer to bootstrap values above 50%.luded next to taxa when function could be predicted. * indicates that theFigure 4 (See legend on next page.)Labeeuw et al. Biology Direct  (2015) 10:23 Page 9 of 21the diatom Phaeodactylum tricornutum strongly groupsAlthough ancient paralogy coupled with differential lossR)d us thn aa wLabeeuw et al. Biology Direct  (2015) 10:23 Page 10 of 21with bacteria (91%) and was likely acquired from amember of this domain by LGT. The cryptophyte singleCCR homolog position in the tree is unresolved. TheCAD homologs of secondarily photosynthetic organismsthem has strong functional prediction and their positionin phylogenies is unresolved. It is therefore not possible todetermine if the pathway is ancestral to archaeplastidswith information currently available.The p-coumaryl alcohol biosynthesis pathway likelyspread through endosymbiotic gene transferFour very disparate groups of secondarily photosyn-thetic organisms have at least one representative withall three enzymes of the p-coumaryl alcohol biosyn-thesis pathway with the correct functional predictions: di-noflagellates, diatoms, hapthophytes and cryptophytes.The large majority of 4CL, CCR and CAD homologsfound in these organisms cluster with each other (despitebeing from widely divergent taxonomic groups) or withred or green algae.There is no observed paralogy of the 4CL gene in sec-ondarily photosynthetic organisms. Diatoms and thehaptophyte weakly cluster together (51%), the crypto-phyte clusters strongly (100%) with another secondarilyphotosynthetic organism (Ectocarpus silicosis, a brownalgae) and with green algae, while the dinoflagellate isfound inside the same well-supported (97%) mixed greenand red algal clade. For CCR, the haptophyte has twoparalogs, one clustering strongly with the dinoflagellate(98%) and the other with both the dinoflagellate and thediatom Fragilariopsis cylindrus (100% support). Otherdiatom CCR paralogs weakly cluster with green algae(50% support), while the single CCR homolog found in(See figure on previous page.)Figure 4 Maximum likelihood tree of cinnamoyl-CoA reductase (CCAmino acid sequences were aligned with MUSCLE and the tree compile50%. 3-hydroxysteroid dehydrogenase, a CCR homolog [63], was used aresearch including bona fide CCR compared to CCR or CCR-like genes. I(DFR) genes were included [19,45]. Gene names are included next to taxbeen biochemically characterized in this organism [30,69,70].show a similar pattern to 4CL and CCR. A dinoflagellateand a hapthophyte paralog strongly cluster togetherwithin a green algal clade that also includes the crypto-phyte (100% support). The second dinoflagellate CADparalog weakly clusters with the diatom F. cylindrus sin-gle CAD homolog (59% support) and the second hapto-phyte paralog groups strongly with green algae (100%).The underlying pattern is clear: a recurring clustering ofsecondarily photosynthetic organisms from disparatetaxonomic groups with each other or green or red algae.Although the exact pattern of species clustering variescan often make phylogenies misleading, it is very unlikelythat it would result in similar patterns of taxonomicallyunrelated secondarily photosynthetic organisms clusteringwith each other or with green/red algae for three differentgenes. The recurrent co-clustering of dinoflagellates, hap-thophytes and diatoms in 4CL, CCR and CAD phyloge-nies is more parsimoniously explained by a commonorigin. The cryptophyte, Gulliardia theta, on the otherhand, does not cluster directly with these other secondar-ily photosynthetic organisms in any phylogeny, suggestingan independent origin. More evidence is needed to con-firm the exact origin of these genes and the number ofevents in which they might have been acquired by second-arily photosynthetic organisms.LGT could have impacted the evolution of ligninprecursors biosynthesisPrevious studies have suggested that at least one gene inthe lignin biosynthesis pathway, phenylalanine ammonialyase (PAL) (Figure 1), was likely acquired through LGTfrom soil bacteria to an ancestor of land plants [50].LGT is likely to also have influenced the evolution of p-coumaryl alcohol biosynthesis. Some of the 4CL homo-logs present in the red alga C. tuberculosum might havebetween phylogenies (a likely result of ancient paralogyand differential loss), it suggests that most 4CL, CCR andCAD homologs present in secondarily photosynthetic spe-cies have been acquired by EGT from a red or green algaeor their ancestors. This also implies that the p-coumarylalcohol biosynthesis pathway, or at the very least its com-ponent genes, are ancient, predating the diversification ofvarious major eukaryotic taxonomic groups such as thedinoflagellates, hapthophytes, cryptophytes and diatoms.enzyme from the p-coumaryl alcohol biosynthesis pathway.sing RaxML. Numbers above branches refer to bootstrap values abovee outgroup. The various classes of CCR are shown based on previousddition, genes showing high similarity to dihydroflavonol reductasehen function could be predicted. * indicates that the enzyme hasbeen affected by this phenomenon. Indeed, the 4CLphylogeny contains a strongly supported clade composedof the red alga C. tuberculosum grouping with bacteria,including many sequences derived from roseobacters(Figure 3). This suggests a horizontal 4CL gene transferfrom a roseobacter to a red alga (Figure 3), giving C.tuberculosum extra copies of 4CL in addition to those itlikely inherited from archaeplastid ancestors. Close phys-ical associations have been shown to promote LGT, andtwo roseobacter species are known to live intracellularlyand intercellularly within red macroalgae [51,52]. SomeFigure 5 (See legend on next page.)Labeeuw et al. Biology Direct  (2015) 10:23 Page 11 of 21red algal species even depend on bacteria for growth ormorphogenesis, highlighting the intimacy of this rela-tionship [53,54]. This bacteria-red algae clade in the 4CLtree also includes a functionally characterized gene fromthe bacterium Streptomyces coelicolor, which has beenshown to have 4CL activity [55]. Argot2 and ESGalso predict all roseobacter and C. tuberculosum 4CLhomologs in this clade to have 4CL function, aprediction only made for these bacteria, red alga andsome land plants enzymes in our datasets. This makes itlikely that these laterally transferred homologs have true4-coumarate:CoA ligase function. For a LGT event froma bacterium to a eukaryote to be confirmed, bacterialgenes need to be found inserted next to genuineeukaryotic genes. Unfortunately, the C. tuberculosum4CL homologs are found on very small contigs in theGymnospermsLycophytesAngiospermsBryophytesChlorophytes Rhodophytes  GlaucophytesDiatoms  PelagophytesPhaeophytesEustigmatophytesChrysophytesOomycetes  ApicomplexaCCHFAArchaeplastidsStramenopilesAlveolateH+G+SH+G+SH?H+G+SH?+G?+S?H+G+SH?EH+G+S lignin from monolignols  p-coumaric acid from phenylalanineRDinoflagellates H?p-coumaryol alcohol fromp-coumaric acid  Heterotrophic EukaryotesBacteria CADCCR 4CL Bacteria Bacteria CCR Loss Loss OR  OR  (See figure on previous page.)Figure 5 Maximum likelihood tree of cinnamyl alcohol dehydrogenase (CAD) enzyme from the p-coumaryl alcohol biosynthesis pathway.Amino acid sequences were aligned with MUSCLE and the tree compiled using RaxML. Numbers above branches refer to bootstrap values above 50%.Sorbitol dehydrogenase, a CAD homolog [80], was used as the outgroup. The various classes of CAD are shown based on previous research, includingthe sinapyl alchohol dehydrogenases (SAD), which share some specific activity with CAD [40,41]. Gene names are included next to taxa when functioncould be predicted. * indicates that the gene has been biochemically characterized in this organism [31,42,71,74].Labeeuw et al. Biology Direct  (2015) 10:23 Page 12 of 21ABFigure 6 Major evolutionary events hypothesized in the evolution of the lconsensus of current phylogenetic analyses of the eukarotic domain [1,36,49]. Maour genome survey and phylogenetic analyses of the putative p-coumaryl alcohocatalyzing consecutive steps in the lignin biosynthesis pathway were found are cmargin for each taxonomic group, with the type of lignin (either H, G or S) specienzymes were found and fuctionally predicted but that there is currently no biocdirection of putative EGT and LGT events (solid arrows are used for events that ailiatesrytpophytesaptophytesunginimalssUnikonH?H?xcavateshizariamoebozoaacteriatsignin biosynthetic pathway across the eukaryotic tree. The tree is ajor events indicated by labeled arrows on the tree are hypothesized froml biosynthesis enzymes. Taxonomic groups in which three or more enzymesolored. The chemical detection of polymerized lignin is indicated in thefied. A question mark (?) indicates that some putative lignin biosynthetichemical evidence of polymerized lignin. Arrows indicate the origin andre conclusive, dashed arrows when events are hypothesized).Labeeuw et al. Biology Direct  (2015) 10:23 Page 13 of 21alga’s genome and further upstream and downstream se-quence data would be needed to determine if they have abacterial or algal context. It is therefore not possible to ex-clude that the C. tuberculosum 4CL genes clustering withroseobacters represent bacterial contamination present inits genome sequence, despite systematic screening of se-quence data to remove it [56]. Regardless of whether LGTbetween roseobacters and C. tuberculosum has taken place,the presence of 4CL in these bacteria raises the possibilitythat they could provide intermediates for the production ofp-coumaryl alcohol to their algal host. If roseobacters canmake p-coumaric acid (none of the known genes for doingso have yet been found in this group), the presence of 4CLwould theoretically enable them to produce p-coumaroyl-CoA and potentially provide it to their algal host.Another likely case of LGT is the acquisition of a pu-tative CCR (with very high confidence functional predic-tion) by the diatom P. tricornutum from bacteria. Theformer is found nested inside a bacterial clade withstrong support (91%). As other diatoms’ putative CCRscluster with green algae, it is likely that a bacterial CCRdisplaced the homolog from algal origin previouslypresent in P. tricornutum. Although LGT did not bring anovel gene to either C. tuberculosum or P. tricornutum(both had an existing homolog prior to LGT which waseither displaced or complemented), it likely had an effecton their secondary metabolism.The phenylpropanoid pathway is unique to land plantsand fungiThe phenylpropanoid pathway enzymes responsible forthe production of p-coumaric acid from the amino acidphenylalanine, PAL and cinnamate 4-hydroxylase (C4H),were only found in land plants and fungi and are clearlymissing from all other eukaryotic genomes screened(Tables 1 and 2, Figure 1). How is it possible for organ-isms to synthesize monolignols without these enzymes?The red alga Calliarthron can produce all types of lig-nin [20] and we could not find these enzymes encodedin the C. tuberculosum genome sequence (Table 1). Theproduct made by PAL and C4H from phenylalanine, p-coumaric acid, has been found in an axenic culture ofthe haptophyte E. huxleyi [27], which also lacks PALand C4H (Table 2). Since both Calliarthron and E. hux-leyi have 4CL, CCR and CAD homologs, there must beother, yet to be described, enzyme(s) capable of synthe-sizing p-coumaric acid and provide it as a substrate tothe 4CL-CCR-CAD pathway to produce p-coumaryl al-cohol. Furthermore, we could not find PAL or C4H genesin the genomes of any green alga or diatom (Tables 1 and2), although p-coumaric acid has been found in speciesfrom both of these groups [28].PAL was likely acquired from bacteria by the ancestorof land plants or fungi and later horizontally transferredbetween these two groups [50]. C4H is also uniquelyfound in land plants and fungi, with distant bacterial ho-mologs (data not shown). The combination of these twogenes is therefore likely a late invention/acquisition ofland plants and fungi. Land plants added two genes (C4Hand PAL) to the 4CL, CCR and CAD already present intheir ancestor (Figure 6). Whether PAL and C4H dis-placed an ancestral enzyme(s) synthesizing p-coumaricacid or those enzyme(s) are still present in land plant ge-nomes is currently unknown. The only other enzymeknown to synthesize p-coumaric acid is tyrosine ammonialyase (TAL), which has only been found in a few bacteria[50]. It can by itself convert the amino acid tyrosine tocoumaric acid, suggesting that a single unknown enzymecould carry the same function in eukaryotes lacking PALand C4H but which have 4CL, CCR and CAD, such as thehapthophyte E. huxleyi and the red alga Calliarthron. It isalso possible that enzymes analogous to PAL and/or C4Hexist, as PAL activity has been found in the green algaChlorella pyrenoidosa [57], but we could not find PALhomologous to plant enzymes in any of the Chlorella ge-nomes screened (Table 1).Expansion of the lignin biosynthesis pathway has occurredmultiple times independently on land and in the seaScreening of eukaryotic genomes revealed that except for4CL, CCR and CAD, all other lignin biosynthesis genesfound in land plants are missing from Calliarthron(Table 1), despite the clear presence of all three lignintypes in this red alga [20]. Assuming the capacity toproduce 4CL’s substrate p-coumaric acid, the presence of4CL, CCR and CAD genes theoretically enables thesynthesis of p-coumaryl alcohol as well as its intermedi-ates, which can be used as substrates for the synthesis ofG and S lignins (Figure 1). This makes convergentevolution of the ability to synthesize G and S lignins inCalliarthron simple, as it would only require the additionof two more enzymes to this core pathway. For example,the lycophyte S. moellendorfii only needed to add caffeicacid O-methyltransferase (COMT) and ferulate 5-hydroxylase (smF5H) to the enzymes needed for p-cou-maryl alcohol synthesis to be able to make both G and Slignins [17]. Also, not all land plants can make S lignin,and the presence of both producers and non-producersin various groups of plants suggests that this ability hasbeen gained and lost multiple times in land plants. Forexample, most gymnosperms do not produce S lignin, butsome can, such as Ginkgo biloba (maidenhair tree) [15].The fact that modifications of lignin production can easilyevolve from a genetic background found in various photo-synthetic eukaryotic lineages lends insight into how thered alga Calliarthron likely evolved the ability to produceG and S lignins. Whether this type of convergent evolu-tion has also happened in other lineages is an openLabeeuw et al. Biology Direct  (2015) 10:23 Page 14 of 21question. Provisional biochemical evidence for the pres-ence of p-coumaryl alcohol in brown and green algae, butabsence of G and S variants [4] suggests that numerouseukaryotes could have the capacity to only synthesize p-coumaryl alcohol or H lignin.ConclusionsThe widespread distribution of coumaryol biosynthesisgene homologs across various eukaryotic supergroupssuggests an ancient origin for this pathway. Although wecannot dismiss the possibility of its presence in an ancienteukaryotic ancestor and subsequent loss in all lineages inwhich it is absent, an origin in archaeplastids is more par-simonious. The ancient pathway for p-coumaryl alcoholsynthesis should contain one or more gene(s) that precede4CL, CCR and CAD, as it requires a source of p-coumaricacid, but these have yet to be discovered. Since p-couma-ric acid is found in the haptophytes E. huxleyi and D.lutheri, the diatom P. tricornutum and the green alga C.vulgaris [27,28], and none of these lineages contains anyhomologs of PAL and C4H, there is little doubt in the ex-istence of enzyme(s) with an analogous function(s). It isnot implied that any organisms carrying this ancient path-way can necessarily polymerize p-coumaryl alcohol (Hmonolignol) to form H lignin or even synthesize themore complex G and S monolignols, but they are likelyable to synthesize at least p-coumaryl alcohol. As allthe secondary photosynthetic organisms investigatedas well as most green and red algae are marine organ-isms, it is intriguing to consider an authentic marinesource of monolignols or lignin. As these compoundsand their degradation products are used as biomarkersto calibrate for terrestrial carbon input into marinesystems [58,59], marine sources of monolignols or lig-nin therefore have the potential to redefine our under-standing of the marine carbon cycle.The function of p-coumaryl alcohol in unicellular mar-ine photosynthetic eukaryotes such as diatoms, dinofla-gellates, hapthophytes, cryptophytes and some green andred algae, is unclear. The two main roles of lignins de-rived from p-coumaryl alcohol and other monolignols inland plants are water transport and structural support.Water transport systems are absent in unicellular algae.If the p-coumaryl alcohol likely produced by unicellularand photosynthetic eukaryotes is polymerized as ligninor lignans, it could also contribute to their structuralstrength, although other compounds such as silica andcellulose are already known fulfil this function in suchorganisms [60]. More likely functions that can be ful-filled by p-coumaryl alcohol are UV protection and mi-crobial defence. Phenolic compounds such as p-coumaric acid and its derivatives exhibit high UV ab-sorptivity and could potentially protect an organismagainst the damaging effects of sunlight [61]. Thelignin biosynthetic pathway has also been implicated inthe defence system of plants, as individual enzymes(e.g., CAD, CCR and CCoAMT) have been shown todefend against microbial attacks [8,9,62,63]. Intermedi-ates of the lignin biosynthesis pathway have also beenshown to have antimicrobial properties [10,64,65].Such a role in host defense or UV protection may haveprovided selection for the early evolution of lignin bio-synthetic pathway in the ocean, which was then co-optedfor water transport and structural strength in landplants faced with new selective pressures of an air-landenvironment.MethodsDatabase searchSequences for the enzymes in the phenylpropanoid andlignin specific pathways of Arabidopsis thaliana werefound in the Kyoto Encyclopedia of Genes and Genomes(KEGG) database (http://www.genome.jp/kegg/), as wellas previous literature [19,45,66]. Additional enzymes fromthe lignin pathway found only in Selaginella moellendorffiiwere obtained from the KEGG database and previousliterature [16]. Whenever possible, proteins that havebeen biochemically characterized were used as queries.These protein sequences [NCBI: NP_179765, NCBI:NP_188576, NCBI: NP_173872, NCBI: NP_001077697,NCBI: NP_173047, NCBI: NP_181241, NCBI: NP_180607,NCBI: NP_199704, NCBI: NP_200227, NCBI: NP_850337,NCBI: NP_195345, NCBI: XP_002963471, NCBI: AAB09228, NCBI: XP_002992167] were used to query the Na-tional Centre for Biotechnology Information (NCBI) data-base (http://www.ncbi.nlm.nih.gov/) for homologs, usingBLASTp and tBLASTn searches of the protein, genomeand expressed sequence tags (EST) databases. An e-valueof 10-30 or less was used as a stringent cut-off for hom-ology. A reciprocal BLASTp search of the hits on Arabi-dopsis thaliana proteins was then performed to confirmorthology. Searches of the algal genomes were carried outusing the Department of Energy (DoE) Joint Genome Insti-tute (JGI) database (http://genome.jgi.doe.gov) using thelatest releases as of April 2014 of the dataset created with‘all models’ gene prediction algorithms. The algal genomesincluded were: Chamydomonas reinhardtii, Chlorella sp.NC64A, Micromonas sp. RCC299, Micromonas pusillaCCMP 1545, Ostreococcus lucimanarinus, Ostreococcus sp.RCC809, Volvox carteri f. nagariensis, Fragilariopsis cylin-drus, Thalassiosira pseudonana, Phaeodactylum tricornu-tum, Aureococcus anophagefferens, and Emiliania huxleyiCCMP 1516. Additional algal genomes were retrieved fromNCBI, including: Bathycoccus prasinos, Botryococcusbraunii, Chlorella variablilis, Coccomyxa subellipsoidea,Dunaliella salina, Haematococcus pluvialis, Ostreococcuslucimarinus, Ostreococcus tauri, Polytomella sp., Polyto-mella parva, Prototheca wickerhamii, Chondrus crispus,Labeeuw et al. Biology Direct  (2015) 10:23 Page 15 of 21Galdiera sulphararia, Ectocarpus siliculosus, Nannochlor-opsis gaditana, Ochromonas danica, Guillardia theta, andHemiselmis andersenii. The genome of the dinoflagel-late Symbiodinium minutum was searched usinghttp://marinegenomics.oist.jp. Additionally, the genomesof the following reference land plants were searched inNCBI: Oryza sativa, Physcomitrella patens, Selaginellamoellendorffii. The genomes from non-photosyntheticspecies closely related to chloroplast-bearing taxa werealso specifically searched: Albugo laibachii, Phytophthorainfestans, Phytophthora sojae, Cryptosporidium muris,Cryptosporidium parvum, Theileria parva, Toxoplasmagondii, and Tetrahymena thermophile. Paramecium tetra-urelia, using http://paramecium.cgm.cnrs-gif.fr/db/tooland Picea abies using http://congenie.org/blastsearch,were also searched. Local searches against red algal ge-nomes were carried out for Calliarthron tuberculosum,Cyanidioschyzon merolae, and Porphyridium cruentum[56]. C. tuberculosum is very closely related to the otherspecies of Calliarthron, C. cheilosporioides, in whichlignin was physically identified [67]. In addition, DrAdrian Reyes performed a local search on Cyanophoraparadoxa genome [68].Functional predictionFunctional prediction was performed on all homologs ofenzymes in the phenylpropanoid and lignin specific path-ways found in our public database searches. Two proteinfunction prediction packages were used, Argot2 [37] aswell as ESG [38]. These are two of the top programs forprotein function prediction according to the ongoing Crit-ical Assessment of protein Function Annotation (CAFA)study [39]. Argot2 performs BLAST and HMMer searchesof sequence databases and then annotates the results withGO (Gene Ontology) terms retrieved from the UniProtKB-GOA database and terms which are then weighted usingthe e-values from BLAST and HMMer. The weighted GOterms, which can also be provided directly, are processedaccording to both their semantic similarity relations de-scribed by the Gene Ontology and their associated score.ESG recursively performs PSI-BLAST searches from se-quence hits obtained in the initial search from the targetsequence, thereby performing multi-level exploration ofthe sequence similarity space around the target protein.Each sequence hit in a search is assigned a weight that iscomputed as the proportion of the log(E-value) of the se-quence relative to the sum of log(E-value) from all the se-quence hits considered in the search of the same level, andthis weight is assigned for GO terms annotating the se-quence hit. The weights for GO terms found in the secondlevel search are computed in the same fashion. Ultimately,the score for a GO term is computed as the total weightfrom the two levels of the searches. The score for each GOterm ranges from 0 to 1.0.Phylogenetic analysisAs only land plants contained most genes found in thephenylpropanoid and lignin biosynthesis pathways, onlythe three core enzymes of the p-coumaryl alcohol biosyn-thesis pathway were further analyzed: cinnamyl alcoholdehydrogenase (CAD), 4-coumarate:CoA ligase (4CL),and cinnamoyl-CoA reductase (CCR). In order to en-sure the sequence dataset was complete, additional se-quences of enzymes from reference plant genomescharacterized in previous literature were added to thedatasets [19,29-31,40-42,45,69-74]. Protein sequenceswere imported into Geneious Pro v5.5 (Biomatters,New Zealand) [75] and multiple alignments were con-structed using MUSCLE [76]. The alignments werethen edited in Geneious. Core functional domains andmotifs were determined using the NCBI conserved do-main search on the Arabidopsis thaliana proteins aswell as previous literature [40,77]. Proteins in whichthese core motifs were not conserved were eliminated.Poorly aligned regions were manually edited. The align-ments were then imported into Randomized AxeleratedMaximum Likelihood v.7.0.4 (RAxML) [78] (http://sco.h-its.org/exelixis/software.html) to create a maximum likeli-hood phylogenetic trees (WAG substitution model, 100bootstrap replicates, gamma distribution parameter es-timated). The tree was formatted for presentation inFigTree v.1.3 (http://tree.bio.ed.ac.uk/software/figtree/).Reviewers’ commentsReviewer’s report 1Mark raganThe authors present computationally based evidence thatthree genes (4CL, CCR and CAD) likely to encode thesuccessive enzymatic steps that reduce the side chain of p-coumaric acid to the corresponding alcohol are present indiverse (mostly photosynthetic) protists, and in red andgreen algae. On the other hand, genes encoding the bio-synthesis of p-coumarate from phenylalanine are knownonly from land plants and fungi. The authors reasonablysuggest a scenario in which capability for stepwise reduc-tion of p-coumarate was transferred by secondary endo-symbioses from an ancient red or green alga to otherlineages. Conservative settings are used for sequencematching, and in general the results are interpreted withadequate caution.Despite the title and use of prejudicial terminology(monolignol) throughout the article, no new evidence ispresented that lignin is actually produced by organismsother than green plants, green algae and the red alga Cal-liarthron. The seeming coding capacity to synthesize p-coumarate and (in some diatoms, oomycetes and Emilia-nia) laccase is certainly suggestive, and the authors pointout potential functions of monolignols other than in gen-eration of a cross-linked structural polymer. NonethelessLabeeuw et al. Biology Direct  (2015) 10:23 Page 16 of 21it remains easy for the reader to over-interpret the results.I urge the authors to be much more sparing in their usageof the term monolignol.AU: We agree with the reviewer that the term mono-lignol can be prejudicial. We have replaced it with p-cou-maryl alcohol wherever possible. p-coumaryl alcohol andH monolignol are essentially the same chemical com-pound, but the former name does not implies that it isused to make lignin. We have also changed the title tomake it less prejudicial, simply referring to synthesis oflignin precursors instead of lignin.Just as CAD not clustering directly with land plantsenzymes does not necessarily mean that homologs fromother organisms have a different function, the inferredhomology of CAD between green algae and plants doesnot require shared derived function.AU: We agree that the clustering of green algae CAD withland plants is only suggestive of shared function and does notdemonstrate it. We have changed the text to remove implica-tions of shared function based on phylogenetic clustering.Although the authors use a conservative BLAST thresh-old, require reciprocal best hits (RBHs) and where possiblequery with sequences from biochemically characterisedproteins, they do not mobilise further evidence (e.g. shareddomain structure) potentially supportive of shared function,and rightly call attention to paralogy and functional overlapamong some lignin-biosynthetic genes in green plants.RBH does not confirm orthology (Methods, paragraph 1);it would be better to refer to database match partners sim-ply as such, or if needs be as putative homologs.AU: We have changed our use of terminology to avoidimplying orthology when it has not been clearly demon-strated. We have also performed additional functionalanalysis on all proteins of interest using the programsArgot2 [37] as well as ESG [38], two of the top programsfor protein function prediction according to the CAFAstudy [39] and included the results in Figure 2. We havealso removed from the phylogenies all proteins withoutthe correct function predicted.Regarding the title: the authors do not build a case formultiple origins, nor that this origin (or origins) oc-curred in the ocean.AU: We have changed the title to remove reference tomultiple origins and the ocean.Reviewer’s report 2Uri gophnaThe authors present an interesting and comprehensiveanalysis of the evolution of lignin biosynthesis, and inferan ancient origin for it in the ancestor of archaeplastids.The topic is of high interest and the phylogenetic ana-lysis is robust. I have two comments:The authors elaborate on lignin biosynthesis in the redalga Calliarthron that has a 4 CL homolog that clustersorthologs from bacteria of the roseobacter clade. Whileit is tempting to suggest LGT from these bacteria, thatare often algal symbionts, the authors acknowledge thatthese homologs can be also " represent bacterial contam-ination present in its genome sequence". Regardless ofwhether the gene has stably integrated into the algal gen-ome or its enzyme product is obtained from the symbiontsby the alga without gene transfer, association with thesebacteria could contribute to the lignin synthesis of theeukaryote over evolutionary timescales. Additionally, hasthe role of this pathway been investigated in the roseobac-ter clade, and has it been connected to pigment productionin these marine bacteria that are often host-associated?AU: Although there is no current evidence of bacterialsymbionts of algae providing them with metabolic precur-sors for synthesis of bioactives or structural components,we acknowledge the possibility of such a situation. Afterscreening all currently available roseobacter genomes forlignin biosynthesis genes, we have only found 4CL is somespecies, while CCR and CAD were absent. We cannot ruleout that a yet to be identified roseobacter symbiont of algaewould have these two genes, as putative homologs havebeen found in some bacteria. If roseobacters can make p-coumaric acid (none of the known genes for doing so havebeen found in this group), the presence of 4CL would theor-etically enable roseobacters to produce p-coumaroyl-CoAand potentially provide it to their algal host. We now men-tion this possibility in the text.The authors conclude that "Monolignols and their po-lymerized forms, such as lignin, are also likely to haveoriginally had a different function than structural sup-port or water transport". While there are many func-tions, such as antibacterial activity, for intermediates inthe pathway, polymerized forms such as lignin are veryhydrophobic and resistant. Thus, structural support isthe most obvious role, even from the perspective of asingle-celled eukaryote. The authors should provide evi-dence to support their conclusion, or modify it.AU: We now mention the possibility that lignin, ifpresent in unicellular organisms, could contribute to thecell wall structure. However, several other compoundshave already been shown to fulfill that role in unicellularorganisms (silica, cellulose, etc) and lignin does not seemnecessary for this purpose. Also, we have not demon-strated that the organisms studied can make lignin, Weare suggesting that our evidence supports the productionof the H lignin precursor, p-coumaryl alcohol.Reviewer’s report 3Philippe deschamps (nominated by purificacion Lopez-Garcia)Labeeuw et al. present a genomic survey of the genescoding the enzymes known to be involved in the biosyn-thesis of lignins, as well as a phylogenetic analysis ofthree enzymes catalyzing the conversion of p-coumaricLabeeuw et al. Biology Direct  (2015) 10:23 Page 17 of 21acid into H monolignol. H monolignol and itssubstituted forms (G and S) are substrates for the syn-thesis of lignin in Viridiplantae. The purpose of the art-icle is to determine if monolignols biosynthesis is aspecificity of land plants or if there are evidences thatthe pathway existed before their diversification. In thelatter case, did the corresponding pathway emerged orwas acquired in the common ancestor of Archaeplas-tida? An additional goal is to determine if monolignolbiosynthesis related genes exist in other eukaryotic line-ages, and to understand how they evolved.From their surveys found in Tables 1 and 2, it is clearthat the synthesis of complex lignin is restricted to Strep-tophyta and that only a portion of the lignin pathway isfound in Chlorophyta and Rhodophyta. From the resultspresented in the same tables, we can observe that thereseems to be a correlation between the actual or former ex-istence of a plastid in a lineage and the presence of the 3enzymes catalyzing the synthesis of H monolignol. Fromthese observations, the authors hypothesize that thesethree enzymes were inherited by or developed in thecommon ancestor of Archaeplastida and transmitted toother photosynthetic and related heterotrophic lineagesby secondary endosymbioses. To test this hypothesis,they produced a phylogenetic tree for each gene familycorresponding to the three enzymes involved in Hmonolignol synthesis. In their interpretation of these trees,the authors see a strong support for their hypothesis.If I agree that the presence/absence pattern suggestssuch a scenario, none of the phylogenetic tree providedshow a topology supporting it. First, the monophyly ofArchaeplastida is never recovered. I know that this is al-most always the case in single gene trees, so this isnonetheless quite possible that the corresponding en-zymes were acquired in their common ancestor. Yet,there is an additional problem: as mentioned by the au-thors, enzymes involved in lignin biosynthesis are oftenduplicated and sub-functionalized. With this in mind,let's focus on the topology inside the Archaeplastidaclade: red algae are not recovered as coherent clades.Moreover, red and green algae possess several copies ineach gene family but these copies are shuffled, meaningthat they don't follow the expected species phylogeny.The authors use these duplications as a positive argu-ment in favor of a common gene ancestry in Archaeplas-tida. This conclusion is only possible if paralogousclusters are well defined and observable, but this is notthe case. Actually the three phylogenetic trees probablysuffer from hidden paralogies, and thus are prone to in-terpretation errors.AU: We have now revised our hypothesis to be less spe-cific in terms of the origins of p-coumaryl alcohol synthe-sis in eukaryotes. We now propose that it originatedeither in the shared ancestor of green algae and landplants, or possibly earlier in the archaeplastid lineage.This is in acknowledgement that red algae are not recov-ered in coherent clades and could have acquired theCAD and CCR genes by LGT. We acknowledge that hid-den ancient paralogy is likely present in the phylogeny ofthe CCR and CAD gene. However, our positive argumentin favor of an origin for lignin biosynthesis early inarchaeplastid evolution is made mostly using the 4CLgene. This gene does not present rampant paralogy, butonly displays it in a single strongly supported clade whichis composed solely of red and green algae as well as stra-menopiles and a dinoflagellate, which have secondaryred plastids. We agree that this is only suggestive of anorigin predating the split between red and green algae forthe 4CL and not enough to conclude it. We have modi-fied the manuscript to suggest that the pathway couldhave been assembled by several EGT and LGT eventsand did not necessarily originate in the ancestor ofarchaeplastids.So, in a context where Archaeplastida are not recov-ered as monophyletic and display a problematic internaltopology, it is impossible to pretend resolving any puta-tive endosymbiotic gene transfer from red algae towarda secondary photosynthetic lineage. Moreover, on eachtree, red points are displayed on nodes supposed to sup-port a case of secondary EGT; none of them are compat-ible with a secondary EGT from a red alga. Based on theobserved topology, one could propose with the same (ifnot more) likelihood that secondary photosynthetic line-ages acquired H monolignol synthesis related genes viaindependent LGT.AU: We originally considered a red algal origin to bethe most likely explanation for the presence of three pu-tative enzymes of the lignin biosynthesis pathway almostonly in organisms known to have secondary red plastids.However, we understand that our evidence for that spe-cific scenario is limited. We have therefore changed ourhypothesis substantially to also include the possibilities ofboth green algal origin (or origin from other ancestralalgae) and LGT for all three genes. The red circles on thethree trees now simply indicate a putative origin throughEGT from either red or green algae or an LGT from a dif-ferent source.Finally, unexpected heterotrophic species are presentwithin Archaeplastida, like Opisthokonts, Amoebozoaand Fungi. This further decreases the ability to correctlyinterpret the trees.AU: We have now performed functional predictionanalysis on all 4CL, CCR and CAD homologs found. Allhomologs without a correct functional prediction wereremoved from the dataset for phylogenetic analyses. Thiseliminated a number of non-orthologous homologs fromheterotrophic species. There are now very few statisti-cally supported clades containing archaeplastids orLabeeuw et al. Biology Direct  (2015) 10:23 Page 18 of 21secondarily photosynthetic organisms that also containheterotrophs. In the 4CL tree, the roseobacter clade isthe only clade with heterotrophs and archaeplastids,and is a clear case of LGT. In the CCR tree, only twoclades contain photosynthetic eukaryotes and hetero-trophs. One is a bacterial clade that also includes thediatom P. tricornutum and is a clear case of LGT frombacteria to a eukaryote. The other is a mixed clade withmoderate (70%) support including a single red alga(Porphyridium cruentum) as well as bacteria and het-erotrophic eukaryotes. This clade is now discussed in thetext as illustrating that red algae present a polyphyleticpattern in the CCR and CAD trees and cannot be used tomake inferences on the evolution of monolignol biosyn-thesis. In the CAD tree, a single supported clade showsgrouping of a heterotroph (yeast) with a photosyntheticeukaryote (red alga). It is at the very base of the tree andgroups paralogs from these organisms that are alcohol de-hydrogenases, not cinnamyl alcohol dehydrogenases (veryhigh confidence in functional prediction analysis). Thesehave not been removed from the dataset despite their clearfunctional prediction as alcohol dehydrogenases, as theyboth have paralogs in the tree with cinnamyl alcohol de-hydrogenase (CAD) function. Besides these few exceptions,the vast majority of nodes grouping archaeaplastids to-gether or archaeplastids with secondarily photosyntheticspecies is well supported in these trees and contains no het-erotrophs. We appreciate the reviewer raising this issueand have now included more detailed explanations ofmixed heterotrophs/autotrophs clades, which are in somecases the result of LGT. We have also pinpointed theseevents in Figures 3, 4 and 5.Additionally, I think that Figure 6 is highly problem-atic. First, as I explained above, there is no support forthe evolutionary scenario that is depicted. Secondly,such a ?consensus of current phylogenetic analyses ofthe eukarotic domain? does not exist. The relative pos-ition of the SAR, Haptophytes and Cryptophytes is stillhighly debated and their global monophyly was rejectedin recent studies (for example Burki et al. 2012 inPRS.B). Moreover, if the common origin of their plastidsfrom a single red alga seems possible (see Petersen et al.2014 in GBE), the number of subsequent endosymbiosesleading to the diverse secondary red lineages is stillundetermined.AU: We recognize that the phylogeny of eukaryotic su-pergroups is controversial in the scientific community.We originally depicted one of the popular theories for thisphylogeny: that a single endosymbiotic event was at theorigin of the red plastid found in secondarily photosyn-thetic eukaryotes, which was in agreement with our inter-pretation of the data. We understand the concerns of thereviewer and have now changed Figure 6 substantially,including uncertainties about the relationship betweenmajor eukaryotic groups which are still under discussionin the scientific community. Since we have now changedthe article to reflect uncertainty about the origin of puta-tive lignin biosynthesis genes we have identified, the evo-lutionary events depicted have also been modified, andboth LGT and EGT events are now included.Reviewers’ comments – second round of reviewReviewer’s report 1Mark raganThe authors have now addressed my concerns.Reviewer’s report 2Uri gophnaNo further comments.Reviewer’s report 3Philippe deschamps (nominated by purificacion Lopez-Garcia)The revised version of the manuscript presented byLabeeuw et al. takes into account many of the suggestionmade by the reviewers. This has clearly enhanced someaspects. For instance, the general comment was thatsome of the conclusions or “announcements” were hastyconsidering the data that were actually presented (forexample on the ability to synthesize genuine lignin pre-cursors or on evolutionary interpretations of the phylo-genetic trees). This specific issue has been corrected.On the other hand, some decisions made during therevision process, in order to comply with some of thereviewer's comments, have indirectly lowered the rigorof the analysis.One question raised was, considering the amount ofgene paralogies detected in the phylogenetic trees forthe four studied gene families, can we be sure that thesegenes are function homologues? This is an interestingquestion, especially when the goal is to try to indirectlydemonstrate the presence of an active specific pathwayin some species. To answer this, the authors have de-cided to use several tools that compute a prediction ofthe actual function of a protein based on its primary se-quences. The authors mention this in their text, so Iwon't elaborate on the various issues of these predictionmodels, and on the fact that they are still not reliableenough to replace proper functional analyses. However, Imust insist on a more critical issue. Following theirfunctional predictions, the authors decided to removefrom their phylogenetic analyses all proteins that werenot predicted as functional homologs of the genes ofinterest. I quote : “We have now performed functionalprediction analysis on all 4CL, CCR and CAD homologsfound. All homologs without a correct functional predic-tion were removed from the dataset for phylogeneticanalyses. This eliminated a number of non-orthologoushomologs from heterotrophic species. There are nowLabeeuw et al. Biology Direct  (2015) 10:23 Page 19 of 21very few statistically supported clades containing archae-plastids or secondarily photosynthetic organisms thatalso contain heterotrophs (…) ”. What has been done hereis absolutely unacceptable. A sequence homolog is not ne-cessarily a functional analog, but the opposite also applies:homologous proteins that differ in their function shouldnot be considered as non-orthologous: orthology is de-fined by the sequence, not by the function. What the au-thors did here was to remove these sequences form theiranalyses and claim that the results are now really more inaccordance with their interpretations (for instance: nomore heterotrophic species in photosynthetic clades,which I pointed out as a potential flaw in their phylogen-etic study). In short: the authors do not realize that thisoperation is comparable to a falsification.Anyway, the phylogenetic trees presented in this revisedmanuscript still don't really support the conclusions. Thesignal is blurry, sequences are shuffled, probably due to ahigh degree of hidden paralogy. The authors wrote: “boththe CCR and CAD phylogenies likely contain hidden an-cestral paralogy and differential loss”. This means thatthey are OK to explain a patchy distribution by differentialloss for some clades. But they also ignore it completely forsome other clades where heterothrophic species mix withsecondary photosynthetic ones. For these latter, the au-thors think that photosynthetic species got the genes byE/LGT, and that heterothrophic species are just there,maybe due to isolated LGT or maybe because they are not“real orthologs”. Sorry but, strictly speaking, these mixedclades can also be explained (with the same likelihood) byan ancestral wide distribution followed by differential loss.My opinion is unchanged: The presence absence dis-tribution suggests that the pathway for p-coumaryl alco-hol synthesis in eukaryotes could have emerged inArchaeplastida and have been transmitted by endosym-biotic transfers to other lineages. But, on a pure tech-nical point of view, the phylogenetic analyses cannot beobjectively used to support this hypothesis. I even con-sider, due to the kind of changes that were made in thesephylogenetic analyses, that this revised version is tech-nically less correct then the original one.AU: We understand the reviewer’s concern regarding theremoval of homologs lacking functional predictions fromthe phylogenies of 4CL, CCR and CAD homologs. We haveadded these back to the phylogenies and these sequencesdid not significantly change the clades on which we basedour conclusions, key clusters remaining well supported sta-tistically. We agree with the reviewer that the trees gener-ally have a weakly supported backbone and that the CADand CCR trees are likely affected by some degree of ances-tral paralogy. However, our conclusions are based on nodesthat are well supported statistically (which is clearly indi-cated in the figures). Also, the phylogenetic pattern inwhich secondarily photosynthetic organisms from widelydivergent taxonomic groups strongly cluster together is re-peated in the trees of all three genes. We consider it highlyunlikely for the observed pattern to be the result of ancientparalogy and differential losses happening in the samecomplex way for three separate genes. In our opinion, it ismuch more likely the result of a common origin, as sug-gested in our hypothesis of EGT between an ancestralarchaeplastid and the ancestors of secondarily photosyn-thetic organisms. The reviewer also mentions the alternatepossibility that p-coumaryl alcohol biosynthesis gene ho-mologs were present in an ancient eukaryotic ancestor andlost in numerous lineages. We have now added a few state-ments in the text presenting this alternate hypothesis.Abbreviations4CL: 4-coumarate:CoA ligase; C3H: p-coumarate 3-hydroxylase;EGT: Endosymbiotic gene transfer; C4H: Cinnamate 4-hydroxylase;CAD: Cinnamyl alcohol dehydrogenase; CCoAMT: Caffeoyl-CoA O-methyltransferase; CCR: Cinnamoyl-CoA reductase; COMT: Caffeic acid O-methyltransferase; CSE: Caffeoyl shikimate esterase; F5H: Ferulate 5-hydroxylase; HCT: Hydroxycinnamoyl-CoA; LAC: Laccase; LGT: Lateral genetransfer; PAL: Phenylalanine ammonia-lyase; PER: Peroxidase;smF5H: Selaginella moelledorfii F5H (smF5H).Competing interestsThe authors declare that they have no competing interests.Authors’ contributionsLL carried out the bioinformatics analyses as well as drafting the manuscript.PM contributed novel gene sequences and drafting of the manuscript. YBcontributed to the phylogenetic analysis as well participating in the designof the experiment and drafting of the manuscript. RC conceived of thestudy, and participated in the analysis and drafting of the manuscript. Allauthors have read and approved the manuscript.Authors’ informationLL is a PhD student studying the influence of bacterial bioactive moleculeson phytoplankton physiology. PM is an Associate Professor of botanystudying biomechanics in macroalgae. YB is an Assistant Professor ofmolecular evolution and phylogenetics studying microbial evolution. RJC isan Assistant Professor of microbial ecology studying the interactionsbetween microalgae and bacteria.AcknowledgmentsThis work was supported by Natural Sciences and Engineering ResearchCouncil of Canada discovery grants to RJC, PTM and YB.Author details1Department of Biological Sciences, University of Alberta, Edmonton, AB, T6G2E9, Canada. 2Department of Botany and Biodiversity Research Centre,University of British, Columbia, BC, V6T 1Z4, Canada.Received: 7 November 2014 Accepted: 31 March 2015References1. Kenrick P, Crane P. The origin and early evolution of plants on land. Nature.1997;389:33–9.2. Weng J-K, Chapple C. The origin and evolution of lignin biosynthesis. NewPhytol. 2010;187:273–85.3. Boerjan W, Ralph J, Baucher M. Lignin biosynthesis. Annu Rev Plant Biol.2003;54:519–46.4. Espiñeira JM, Novo Uzal E, Gómez Ros LV, Carrión JS, Merino F, Ros BarcelóA, et al. Distribution of lignin monomers and the evolution of lignificationamong lower plants. Plant Biol. 2011;13:59–68.5. Campbell MM, Sederoff RR. Variation in lignin content and composition.Plant Physiol. 1996;110:3–13.Labeeuw et al. Biology Direct  (2015) 10:23 Page 20 of 216. Jones L, Ennos AR, Turner SR. Cloning and characterization of irregularxylem4 (irx4): a severely lignin-deficient mutant of Arabidopsis. Plant J.2001;26:205–16.7. Bonawitz ND, Chapple C. The genetics of lignin biosynthesis: connectinggenotype to phenotype. Annu Rev Genet. 2010;44:337–63.8. Moura JCMS, Bonine CAV, de Oliveira Fernandes Viana J, Dornelas MC,Mazzafera P. Abiotic and biotic stresses and changes in the lignin contentand composition in plants. J Integr Plant Biol. 2010;52:360–76.9. Tronchet M, Balagué C, Kroj T, Jouanin L, Roby D. Cinnamyl alcoholdehydrogenases-C and D key enzymes in lignin biosynthesis, play an essentialrole in disease resistance in Arabidopsis. Mol Plant Pathol. 2010;11:83–92.10. Keen NT, Littlefield LJ. The possible association of phytoalexins withresistance gene expression in flax to Melampsora lini. Physiol Plant Pathol.1979;14:265–80.11. Zhong R, Ye Z-H. Transcriptional regulation of lignin biosynthesis. Plant SignalBehav. 2009;4:1028–34.12. Lewis NG, Yamamoto E. Lignin: occurrence, biogenesis and biodegradation.Annu Rev Plant Physiol Plant Mol Biol. 1990;41:455–96.13. Vanholme R, Demedts B, Morreel K, Ralph J, Boerjan W. Lignin biosynthesisand structure. Plant Physiol. 2010;153:895–905.14. Weng J-K, Banks JA, Chapple C. Parallels in lignin biosynthesis. CommunIntegr Biol. 2008;1:20–2.15. Uzal EN, Gómez Ros LV, Pomar F, Bernal MA, Paradela A, Albar JP, et al.The presence of sinapyl lignin in Ginkgo biloba cell cultures changes ourviews of the evolution of lignin biosynthesis. Physiol Plant.2009;135:196–213.16. Weng J-K, Li X, Stout J, Chapple C. Independent origins of syringyl lignin invascular plants. Proc Natl Acad Sci U S A. 2008;105:7887–92.17. Weng J-K, Akiyama T, Bonawitz ND, Li X, Ralph J, Chapple C. Convergentevolution of syringyl lignin biosynthesis via distinct pathways in the lyco-phyte Selaginella and flowering plants. Plant Cell. 2010;22:1033–45.18. Gómez Ros LV, Gabaldón C, Pomar F, Merino F, Pedreño MA, Barceló AR.Structural motifs of syringyl peroxidases predate not only the gymnosperm-angiosperm divergence but also the radiation of tracheophytes. New Phytol.2007;173:63–78.19. Xu Z, Zhang D, Hu J, Zhou X, Ye X, Reichel KL, et al. Comparative genomeanalysis of lignin biosynthesis gene families across the plant kingdom. BMCBioinformatics. 2009;10 Suppl 1:S3.20. Martone PT, Estevez JM, Lu F, Ruel K, Denny MW, Somerville C, et al.Discovery of lignin in seaweed reveals convergent evolution of cell-wallarchitecture. Curr Biol. 2009;19:169–75.21. Delwiche CF, Graham LE, Thomson N. Lignin-like compounds andsporopollenin in Coleochaete, an algal model for land plant ancestry.Science (80- ). 1989;245:399–401.22. Kroken S, Graham L, Cook M. Occurrence and evolutionary significance ofresistant cell walls in charophytes and bryophytes. Am J Bot. 1996;83:1241–54.23. Dovgan I, Medvedeva E. Change in the structural elements of the lignin ofthe brown alga Cystoseira barbata at different ages. Chem Nat Compd.1983;19:85–8.24. Dovgan I, Medvedeva E, Yanishevskaya E. Cleavage of the lignins of the algaCystoseira barbata by thioacetic acid. Chem Nat Compd. 1983;19:88–91.25. Reznikov V, Mikhaseva M, Zil’bergleit M. The lignin of the alga Fucusvesiculosus. Chem Nat Compd. 1978;14:645–7.26. Ragan M. Fucus “lignin”: A reassessment. Phytochemistry. 1984;23:2029–32.27. Seyedsayamdost MR, Case RJ, Kolter R, Clardy J. The Jekyll-and-Hydechemistry of Phaeobacter gallaeciensis. Nat Chem. 2011;3:331–5.28. Goiris K, Muylaert K, Voorspoels S, Noten B, De Paepe D, Baart GJ E, et al.Detection of flavonoids in microalgae from different evolutionary lineages.J Phycol. 2014;50:483–92.29. Raes J, Rohde A, Christensen JH, Van de Peer Y, Boerjan W. Genome-widecharacterization of the lignification toolbox in Arabidopsis. Plant Physiol.2003;133(November):1051–71.30. Costa MA, Collins RE, Anterola AM, Cochrane FC, Davin LB, Lewis NG. An insilico assessment of gene function and organization of thephenylpropanoid pathway metabolic networks in Arabidopsis thaliana andlimitations thereof. Phytochemistry. 2003;64:1097–112.31. Fraser CM, Chapple C. The phenylpropanoid pathway in Arabidopsis.Arabidopsis Book. 2011;9:e0152.32. Goujon T, Sibout R, Eudes A, MacKay J, Jouanin L. Genes involved in thebiosynthesis of lignin precursors in Arabidopsis thaliana. Plant PhysiolBiochem. 2003;41:677–87.33. Hamberger B, Ellis M, Friedmann M, de Azevedo SC, Barbazuk B, Douglas CJ.Genome-wide analyses of phenylpropanoid-related genes in Populustrichocarpa, Arabidopsis thaliana, and Oryza sativa: the Populus lignin toolboxand conservation and diversification of angiosperm gene families. Can JBot. 2007;85:1182–201.34. Yokoyama R, Nishitani K. Genomic basis for cell-wall diversity in plants. Acomparative approach to gene families in rice and Arabidopsis. Plant cellPhysiol. 2004;45:1111–21.35. Carocha V, Soler M, Hefer C, Cassan-Wang H, Fevereiro P, Myburg AA, PaivaJAP and Grima-Pettenati J. Genome-wide analysis of the lignin toolbox ofEucalyptus grands. New Phytologist. 2015. doi: 10.1111/nph.13313.36. Keeling PJ. The number, speed, and impact of plastid endosymbioses ineukaryotic evolution. Annu Rev Plant Biol. 2013;64:583–607.37. Falda M, Toppo S, Pescarolo A, Lavezzo E, Di Camillo B, Facchinetti A, et al.Argot2: a large scale function prediction tool relying on semantic similarity ofweighted Gene Ontology terms. BMC Bioinformatics. 2012;13 Suppl 4:S14.38. Chitale M, Hawkins T, Park C, Kihara D. ESG: Extended similarity groupmethod for automated protein function prediction. Bioinformatics.2009;25:1739–45.39. Radivojac P, Clark WT, Oron TR, Schnoes AM, Sokolov A, Graim K, et al. Alarge-scale evaluation of computational protein function prediction. NatMethods. 2013;10:221–7.40. Guo D-M, Ran J-H, Wang X-Q. Evolution of the cinnamyl/sinapyl alcoholdehydrogenase (CAD/SAD) gene family: the emergence of real lignin isassociated with the origin of bona fide CAD. J Mol Evol. 2010;71:202–18.41. Barakat A, Bagniewska-zadworna A, Choi A, Plakkat U, DiLoreto DS, Yellanki P,et al. The cinnamyl alcohol dehydrogenase gene family in Populous:phylogeny, organization, and expression. BMC Plant Biol. 2009;9:26.42. Larroy C, Parés X, Biosca JA. Characterization of a Saccharomyces cerevasiaeNADP(H)-dependent alcohol dehydrogenase (ADHVII), a member of thecinnamyl alcohol dehydrogenase family. Eur J Biochem. 2002;269:5738–45.43. Persson B, Kallberg Y, Bray JE, Bruford E, Dellaporta SL, Favia AD, et al. TheSDR (short-chain dehydrogenase/reductase and related enzymes)nomenclature initiative. Chem Biol Interact. 2009;178:94–8.44. Fulda M, Heinz E, Wolter FP. The fadD gene of Escherichia coli K12 is locatedclose to rnd at 39.6 min of the chromosomal map and is a new member ofthe AMP-binding protein family. Mol Gen Genet. 1994;242:241–9.45. Barakat A, Yassin NBM, Park JS, Choi A, Herr J, Carlson JE. Comparative andphylogenomic analyses of cinnamoyl-CoA reductase and cinnamoyl-CoA-reductase-like gene family in land plants. Plant Sci. 2011;181:249–57.46. Cukovic D, Ehlting J, VanZiffle JA, Douglas CJ. Structure and evolution of 4-coumarate:coenzyme A ligase (4CL) gene families. Biol Chem. 2001;382:645–54.47. Nordling E, Jörnvall H, Persson B. Medium-chain dehydrogenases/reductases(MDR): Family characterizations including genome comparisons and activesite modelling. Eur J Biochem. 2002;269:4267–76.48. Leliaert F, Verbruggen H, Zechman FW. Into the deep: New discoveries atthe base of the green plant phylogeny. BioEssays. 2011;33:683–92.49. Archibald JM. The puzzle of plastid evolution. Curr Biol. 2009;19:R81–8.50. Emiliani G, Fondi M, Fani R, Gribaldo S. A horizontal gene transfer at theorigin of phenylpropanoid metabolism: a key adaptation of plants to land.Biol Direct. 2009;4:7.51. Case RJ, Longford SR, Campbell AH, Low A, Tujula N, Steinberg PD, et al.Temperature induced bacterial virulence and bleaching disease in achemically defended marine macroalga. Environ Microbiol. 2011;13:529–37.52. Ashen J, Goff L. Molecular identification of a bacterium associated with gallformation in the marine red alga Prionitis lanceolata. J Phycol. 1996;297:286–97.53. Hanzawa N, Chiba SN, Miyajima S, Yamazaki A, Saga N. Phylogeneticcharacterization of marine bacteria that induce morphogenesis in the redalga Porphyra yezoensis. Fish Genet Breed Sci. 2010;40:29–35.54. Fries L. The influence of microamounts of organic substances other thanvitamins on the growth of some red algae in axenic culture. Br Phycol J.1970;5(March 2015):39–46.55. Kaneko M, Ohnishi Y, Horinouchi S. Cinnamate : Coenzyme A ligase fromthe filamentous bacterium Streptomyces coelicolor A3(2). J Bacteriol.2003;185:20–7.56. Chan CX, Yang EC, Banerjee T, Yoon HS, Martone PT, Estevez JM, et al. Redand green algal monophyly and extensive gene sharing found in a richrepertoire of red algal genes. Curr Biol. 2011;21:328–33.57. Chen K, Feng H, Zhang M, Wang X. Nitric oxide alleviates oxidative damagein the green alga Chlorella pyrenoidosa caused by UV-B radiation. FoliaMicrobiol (Praha). 2003;48:389–93.58. Hedges JI, Keil RG, Benner R. What happens to terrestrial organic matter inthe ocean? Org Geochem. 1997;27:195–212.Submit your next manuscript to BioMed Centraland take full advantage of: • Convenient online submission• Thorough peer review• No space constraints or color figure charges• Immediate publication on acceptance• Inclusion in PubMed, CAS, Scopus and Google Scholar• Research which is freely available for redistributionLabeeuw et al. Biology Direct  (2015) 10:23 Page 21 of 2159. Opsahl S, Benner R. Distribution and cycling of terrigenous dissolvedorganic matter in the ocean. Nature. 1997;386:480–2.60. Popper ZA, Tuohy MG. Beyond the green: understanding the evolutionarypuzzle of plant and algal cell walls. Plant Physiol. 2010;153:373–83.61. Gitz DC, Liu-Gitz L, McClure JW, Huerta AJ. Effects of a PAL inhibitor onphenolic accumulation and UV-B tolerance in Spirodela intermedia (Koch.).J Exp Bot. 2004;55:919–27.62. Boudet A-M. Lignins and lignification: Selected issues. Plant PhysiolBiochem. 2000;38:81–96.63. Lacombe E, Hawkins S, Van Doorsselaere J, Piquemal J, Goffner D,Poeydomenge O, et al. Cinnamoyl CoA reductase, the first committedenzyme of the lignin branch biosynthetic pathway: cloning, expression andphylogenetic relationships. Plant J. 1997;11:429–41.64. Barber MS, McConnell VS, DeCaux BS. Antimicrobial intermediates of thegeneral phenylpropanoid and lignin specific pathways. Phytochemistry.2000;54:53–6.65. Gunnison D, Alexander M. Basis for the resistance of several algae tomicrobial decomposition. Appl Microbiol. 1975;29:729–38.66. Vanholme R, Cesarino I, Rataj K, Xiao Y, Sundin L, Goeminne G, et al.Caffeoyl shikimate esterase (CSE) is an enzyme in the lignin biosyntheticpathway in Arabidopsis. Science (80- ). 2013;341:1103–6.67. Gabrielson PW, Miller KA, Martone PT. Morphometric and molecularanalyses confirm two distinct species of Calliarthron (Corallinales,Rhodophyta), a genus endemic to the northeast Pacific. Phycologia.2011;50:298–316.68. Price DC, Chan CX, Yoon HS, Yang EC, Qiu H, Weber APM, et al. Cyanophoraparadoxa genome elucidates origin of photosynthesis in algae and plants.Science (80- ). 2012;335:843–7.69. Lauvergeat V, Lacomme C, Lacombe E, Lasserre E, Roby D, Grima-PettenatiJ. Two cinnamoyl-CoA reductase (CCR) genes from Arabidopsis thaliana aredifferentially expressed during development and in response to infectionwith pathogenic bacteria. Phytochemistry. 2001;57:1187–95.70. Baltas M, Lapeyre C, Bedos-Belval F, Maturano M, Saint-Aguet P, Roussel L,et al. Kinetic and inhibition studies of cinnamoyl-CoA reductase 1 fromArabidopsis thaliana. Plant Physiol Biochem. 2005;43:746–53.71. Sibout R, Eudes A, Mouille G, Pollet B, Lapierre C, Jouanin L, et al. CinnamylAlcohol Dehydrogenase-C and -D Are the Primary Genes Involved in LigninBiosynthesis in the Floral Stem of Arabidopsis. Plant Cell. 2005;17(July):2059–76.72. Silber MV, Meimberg H, Ebel J. Identification of a 4-coumarate:CoA ligasegene family in the moss, Physcomitrella patens. Phytochemistry.2008;69:2449–56.73. Ehlting J, Büttner D, Wang Q. Three 4-coumarate: coenzyme A ligases inArabidopsis thaliana represent two evolutionarily divergent classes inangiosperms. Plant J. 1999;19:9–20.74. Mee B, Kelleher D, Frias J, Malone R, Tipton KF, Henehan GTM, et al.Characterization of cinnamyl alcohol dehydrogenase of Helicobacter pylori:An aldehyde dismutating enzyme. FEBS J. 2005;272:1255–64.75. Drummond A, Ashton B, Buxton S, Cheung M, Cooper A, Duran C, Field M,Heled J, Kearse M, Markowitz S, Moir R, Stones-Havas S, Sturrock S, Thierer T,Wilson A, Geneious v5.5. 2011. Available at http://www.geneious.com76. Edgar RC. MUSCLE: multiple sequence alignment with high accuracy andhigh throughput. Nucleic Acids Res. 2004;32:1792–7.77. Ehlting J, Shin JJK, Douglas CJ. Identification of 4-coumarate : coenzyme Aligase (4CL) substrate recognition domains. Plant J. 2001;27:455–65.78. Stamatakis A. RAxML-VI-HPC: Maximum likelihood-based phylogeneticanalyses with thousands of taxa and mixed models. Bioinformatics.2006;22:2688–90.79. Conti E, Franks NP, Brick P. Crystal structure of firefly luciferase throws lighton a superfamily of adenylate-forming enzymes. Structure. 1996;4:287–98.80. McKie JH, Jaouhari R, Douglas KT, Goffner D, Feuillet C, Grima-Pettenati J,et al. A molecular model for cinnamyl alcohol dehydrogenase, a plantaromatic alcohol dehydrogenase involved in lignification. Biochim BiophysActa. 1993;1202:61–9.Submit your manuscript at www.biomedcentral.com/submit


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items