UBC Faculty Research and Publications

Complex phylogenetic distribution of a non-canonical genetic code in green algae Cocquyt, Ellen; Gile, Gillian H; Leliaert, Frederik; Verbruggen, Heroen; Keeling, Patrick J; De Clerck, Olivier Oct 26, 2010

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

Item Metadata


52383-12862_2010_Article_1542.pdf [ 1.35MB ]
JSON: 52383-1.0223453.json
JSON-LD: 52383-1.0223453-ld.json
RDF/XML (Pretty): 52383-1.0223453-rdf.xml
RDF/JSON: 52383-1.0223453-rdf.json
Turtle: 52383-1.0223453-turtle.txt
N-Triples: 52383-1.0223453-rdf-ntriples.txt
Original Record: 52383-1.0223453-source.json
Full Text

Full Text

RESEARCH ARTICLE Open AccessComplex phylogenetic distribution of a non-canonical genetic code in green algaeEllen Cocquyt1*, Gillian H Gile2, Frederik Leliaert1, Heroen Verbruggen1, Patrick J Keeling2, Olivier De Clerck1AbstractBackground: A non-canonical nuclear genetic code, in which TAG and TAA have been reassigned from stopcodons to glutamine, has evolved independently in several eukaryotic lineages, including the ulvophycean greenalgal orders Dasycladales and Cladophorales. To study the phylogenetic distribution of the standard and non-canonical genetic codes, we generated sequence data of a representative set of ulvophycean green algae andused a robust green algal phylogeny to evaluate different evolutionary scenarios that may account for the origin ofthe non-canonical code.Results: This study demonstrates that the Dasycladales and Cladophorales share this alternative genetic code withthe related order Trentepohliales and the genus Blastophysa, but not with the Bryopsidales, which is sister to theDasycladales. This complex phylogenetic distribution whereby all but one representative of a single natural lineagepossesses an identical deviant genetic code is unique.Conclusions: We compare different evolutionary scenarios for the complex phylogenetic distribution of this non-canonical genetic code. A single transition to the non-canonical code followed by a reversal to the canonical codein the Bryopsidales is highly improbable due to the profound genetic changes that coincide with codonreassignment. Multiple independent gains of the non-canonical code, as hypothesized for ciliates, are also unlikelybecause the same deviant code has evolved in all lineages. Instead we favor a stepwise acquisition model,congruent with the ambiguous intermediate model, whereby the non-canonical code observed in these greenalgal orders has a single origin. We suggest that the final steps from an ambiguous intermediate situation to anon-canonical code have been completed in the Trentepohliales, Dasycladales, Cladophorales and Blastophysa butnot in the Bryopsidales. We hypothesize that in the latter lineage an initial stage characterized by translationalambiguity was not followed by final reassignment of both stop codons to glutamine. Instead the standard codewas retained by the disappearance of the ambiguously decoding tRNAs from the genome. We correlate theemergence of a non-canonical genetic code in the Ulvophyceae to their multinucleate nature.BackgroundThe genetic code, which translates nucleotide triplets intoamino acids, is universal in nearly all genetic systems,including viruses, bacteria, archaebacteria, eukaryoticnuclei and organelles [1,2]. However, a small number ofeubacterial, eukaryotic nuclear, plastid and mitochondrialgenomes have evolved slight variations of the standard orcanonical genetic code [3,4]. Most codon reassignmentshave been traced to changes in tRNAs, either by singlenucleotide substitution, base modification, or RNA editing.Two main models have been proposed to explain the evo-lutionary changes in the genetic code [reviewed in [4-7]].The “codon capture” model [8] proposes that bias in GCcontent can eliminate certain codons from the entire gen-ome, after which they can reappear through randomgenetic drift, and become reassigned (“captured”) by muta-tions of noncognate tRNAs. This mechanism is essentiallyneutral, that is, codon reassignment is accomplishedwithout the generation of aberrant and non-functionalproteins. The “ambiguous intermediate” model is a non-neutral model, which posits that codon reassignmentoccurs through an intermediate stage where a particularcodon is ambiguously decoded by both the cognate tRNAand a tRNA that is mutated at locations other than at the* Correspondence: EllenE.Cocquyt@UGent.be1Phycology Research Group and Center for Molecular Phylogenetics andEvolution, Ghent University, Krijgslaan 281 S8, 9000 Ghent, BelgiumFull list of author information is available at the end of the articleCocquyt et al. BMC Evolutionary Biology 2010, 10:327http://www.biomedcentral.com/1471-2148/10/327© 2010 Cocquyt et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative CommonsAttribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction inany medium, provided the original work is properly cited.anticodon, after which the mutant tRNA may take overdecoding of the ambiguous codon if it is adaptive [9]. Thetwo models are not mutually exclusive and codon reas-signments may be driven by a combination of differentmechanisms [10,11]. For example, ambiguous intermediatestages may be preceded by strong GC bias [12].Similar evolutionary mechanisms of gradual codonreassignment have been suggested to apply to reassign-ment of stop codons to sense codons, in which a mutatedtRNA may initially recognize and eventually capture astop codon from the cognate release factor [5]. Stopcodons may be particularly prone to reassignment eitherbecause they are less prevalent than sense codons (occur-ring only once per gene) and therefore cause minimaldetrimental effects if they are reassigned, or becausechanges to release factors can occur rapidly [4,13].Only five eukaryotic lineages are known to have evolvednon-canonical nuclear genetic codes, including ciliates,hexamitid diplomonads, fungi (in the genus Candida andmany ascomycetes), polymastigid oxymonads, and greenalgae (in the Dasycladales and Cladophorales). By far themost common variant is the reassignment of the stopcodons TAG and TAA (TAR) to glutamine, which hasoccurred independently in hexamitid diplomonads [14],several ciliates [15,16], polymastigid oxymonads [17,18]and dasycladalean and cladophoralean green algae [19-21].Interestingly, this TAR®Gln reassignment has never beenobserved in prokaryotes or organelles.The aim of the present study is to document the phylo-genetic distribution of the non-canonical genetic code ingreen algae, with emphasis on the ulvophycean relatives ofthe Dasycladales and Cladophorales. Our approach con-sists of screening green algal nuclear genes for the pre-sence of non-canonical glutamine codons and interpretingthe evolution of the genetic code and glutamine codonusage in a phylogenetic framework. The apparently com-plex phylogenetic distribution of a non-canonical code inclosely related ulvophycean lineages offers a unique oppor-tunity to study the mechanisms of the protein translationalmachinery that may have led to codon-reassignment.ResultsThe dasycladalean genera Acetabularia, Batophora andParvocaulis, and the cladophoralean genera Cladophoraand Chaetomorpha have been shown previously to usethe TAR®Gln genetic code [19-21]. We have character-ized one or more of 8 different genes from 21 taxarepresenting the breadth of ulvophycean diversity [thisstudy and [22], see also Additional files 1 and 2]. Fromthese data, we found that the non-canonical glutaminecodons appear in nuclear-encoded genes of additionalmembers of the both Dasycladales (genus Bornetella)and Cladophorales (genera Boergesenia, Boodlea, Clado-phora, Dictyosphaeria, Phyllodictyon, Siphonocladus,Valonia), as well as Trentepohliales (genus Trentepoh-lia) and the genus Blastophysa, which is currently notassigned to an order. We inferred that an organism usesa non-canonical code if TAR codons were found athighly conserved positions where land plants and othergreen algae encode glutamine. Only TGA is used asstop codon in EST sequences that we obtained for Cla-dophora (40 S ribosomal protein S9, GQ421515; OEE1,GQ421494). Moreover, other sequences deposited inGenBank from Cladophora (GapA, DQ270261; EF-1a,EF551321) and Acetabularia (PsbS, BK006014) termi-nate with TGA codons only. The presence of TGA asthe only stop codon for species having TAR at con-served glutamine positions further supports our infer-ence of a non-canonical genetic code.The presence of the standard code was determined forthe genus Ignatius and the order Bryopsidales (generaCaulerpa, Bryopsis) by the presence of only canonicalglutamine codons and the use of all three stop codons.The Ignatius actin gene has TAG as stop codon, whilethe b tubulin and HPS90 genes have TAA as stopcodons. For Caulerpa, both b tubulin and HPS90 geneshave TAA as stop codons, so we did not observe the useof TAG for stop. In publicly available sequences fromBryopsis all three stop codons are used: TAA in ribonu-clease Bm2 gene (AB164318), TAG in lectin precursorand oxygen evolving protein of photosystem II genes(EU410470 and AB293980), and TGA in the bryohealinprecursor gene (EU769118). Taken together, this showsthat the Bryopsidales use the standard genetic code.The occurrence of the standard and TAR®Gln code isplotted onto the reference phylogeny in Figure 1. TheStreptophyta, prasinophytes, Trebouxiophyceae andChlorophyceae possess the standard genetic code. Withinthe class Ulvophyceae, the standard code is found in theorders Ulvales, Ulotrichales, Bryopsidales and the genusIgnatius whereas the orders Dasycladales, Cladophorales,Trentepohliales and the genus Blastophysa have a non-canonical code. However, the taxa with the TAR®Glncode do not form a monophyletic group (Figure 1).If both the phylogeny and the distribution of the geneticcodes shown in Figure 1 are correct, then more than onegain and/or loss event of the non-canonical code must bepostulated. To examine the validity of the phylogeny, weperformed AU topology tests to evaluate whether the datarejected a topology in which all taxa with the non-canoni-cal code formed a monophyletic group. Specifically, wefound that the topology where Bryopsidales is sister to allUlvophyceae with the TAR®Gln code is rejected withhigh significance (ΔlnL = 27.9; p < 0.0001).The estimated evolution of glutamine codon usage fre-quencies is shown in Figure 2 (only the green algalclasses Chlorophyceae, Trebouxiophyceae and Ulvophy-ceae are shown). In general, the canonical codon CAGCocquyt et al. BMC Evolutionary Biology 2010, 10:327http://www.biomedcentral.com/1471-2148/10/327Page 2 of 9Figure 1 Phylogenetic distribution of the standard and non-canonical genetic codes in green algae. The occurrence of a non-canonicalgenetic code (TAR®Gln) is indicated with gray squares. The taxa with the non-canonical code do not form a monophyletic group Threealternative scenarios can explain this phylogenetic distribution: (1) A single origin of the non-canonical code along the branch leading to theorders Trentepohliales, Dasycladales, Bryopsidales, Cladophorales and the genus Blastophysa and a subsequent reversal to the universal code inthe Bryopsidales (indicated with black arrow and cross). (2) Three independent gains of the non-canonical code in the Trentepohliales, theDasycladales and the Cladophorales + Blastophysa (indicated with gray arrows). (3) A stepwise process of evolution of the non-canonical codewith a single initiation of the process along the branch leading to the orders Trentepohliales, Dasycladales, Bryopsidales, Cladophorales and thegenus Blastophysa, followed by a completion of the process in all lineages except the Bryopsidales (black arrow combined with gray arrows). Thereference phylogeny of the green plant lineage was obtained by maximum likelihood inference of a 25% site stripped dataset containing 7nuclear genes, SSU nrDNA and plastid genes rbcL and atpB [22]. Numbers at nodes indicate ML bootstrap values (top) and posterior probabilities(bottom); values below respectively 50 and 0.9 are not shown.Cocquyt et al. BMC Evolutionary Biology 2010, 10:327http://www.biomedcentral.com/1471-2148/10/327Page 3 of 9is most commonly used, followed by the canonicalcodon CAA. The dominance of the canonical glutaminecodons CAG over CAA is also reflected in the non-canonical glutamine codons, TAG being much morecommon than TAA. The bias towards CAG and TAG islikely a product of the overall bias of these species and/or genes for GC residues, leading to increased use of Gin the third codon position. Changes from canonical tonon-canonical glutamine codons require a single transi-tion (C ® T) at the first codon position.DiscussionOur results reveal a broad distribution of a non-canonicalnuclear genetic code in the Ulvophyceae, where gluta-mine is encoded by canonical CAG and CAA codons aswell as non-canonical TAG and TAA codons. Surpris-ingly, we find that taxa with this non-canonical code donot form a monophyletic group according to the see-mingly robust phylogeny of the organisms (Figure 1). Ifthe inferred phylogeny is indeed correct, three alternativescenarios can explain the distribution of the code on thattree: (1) a single origin of the non-canonical code alongthe branch leading to the orders Trentepohliales, Dasy-cladales, Bryopsidales, Cladophorales and the genusBlastophysa and a subsequent reversal to the standardcode in the Bryopsidales (Figure 1: indicated with blackarrow and cross), (2) three independent gains of the non-canonical code in the Trentepohliales, the Dasycladalesand the Cladophorales + Blastophysa (Figure 1: indicatedwith gray arrows), and (3) a stepwise process of evolutionFigure 2 Estimated ancestral frequencies of glutamine codon usage. Canonical codon CAG is most commonly used, followed by the othercanonical codon, CAA. Among the non-canonical codons, TAG is used more commonly then TAA. The estimate ancestral frequencies of non-canonical codon usage along the nodes of interest are indicated with arrows.Cocquyt et al. BMC Evolutionary Biology 2010, 10:327http://www.biomedcentral.com/1471-2148/10/327Page 4 of 9of the non-canonical code with a single initiation of theprocess along the branch leading to the orders Trente-pohliales, Dasycladales, Bryopsidales, Cladophorales andthe genus Blastophysa, followed by a completion of theprocess in all lineages except the Bryopsidales (Figure 1:black arrow combined with gray arrows). Alternatively,because changes in the genetic code are so rare, the pos-sibility that the reference phylogeny is wrong should notbe passed over too easily. More specifically, if oneassumes that the current position of the Bryopsidales iswrong and that in reality this group is sister to all thetaxa with a non-canonical code, only a single transitionto a non-canonical genetic code would have to beinvoked. In what follows, we will discuss each of thesepossibilities in more detail and report on some cytologi-cal correlates of the non-canonical genetic code.Phylogenetic uncertaintyOur phylogenetic tree is based on the most comprehen-sive dataset currently available for the Ulvophyceae. It isinferred from a concatenated dataset including sevennuclear genes, SSU nrDNA and two plastid genes usingmodel-based techniques with carefully chosen partition-ing strategies and models of sequence evolution andapplication of a site removal approach to optimize thesignal-to-noise ratio [22]. Our tree shows a sister rela-tionship between Dasycladales and Bryopsidales withmoderate to high statistical support (BV 81 and PP 1.00).This relationship is concordant with a recently published74-taxon phylogeny of the green lineage based on SSUnrDNA and two plastid genes [23]. Both phylogeniesshow major improvements in taxon and gene samplingwithin the Ulvophyceae compared to previously pub-lished phylogenies, which were either based on a singlemarker, did not include the Bryopsidales, or could notresolve the relationships among the Bryopsidales, Dasy-cladales and Cladophorales [24-26]. Based on the datasetused to infer our reference tree, the alternative topologyin which the Bryopsidales are sister to all taxa with anon-canonical code is significantly less likely than theML tree as shown by AU tests. As a consequence, wemust conclude based on all available information that theulvophycean taxa with the non-canonical code form aparaphyletic group and one of the more complex evolu-tionary scenarios for the gain of the non-canonical codehas to be invoked.Gain-reversal hypothesisA reversal from the non-canonical to the standard geneticcode is unlikely for several reasons, indeed arguably moreunlikely than the original change to the non-canonicalcode. Following a transition to the non-canonical code,TAR glutamine codons would be present in many codingsequences. In order to revert to the canonical code, thesecodons would all have to revert to canonical CAR codonsor they would terminate translation, with obvious detri-mental effects. The improbability of this reversal isenforced by the increased frequency of the codons: stopcodons appear only once per gene, whereas glutamine ispresent many times per gene on average. Our ancestralstate estimates indicate a non-negligible usage frequencyof both non-canonical codons along the ancestral nodes ofinterest (Figure 2C, D, indicated with arrows). Theseresults must be considered with caution because of theintrinsic limitations of ancestral state estimation [27] andthe fact that the non-independence of the evolution of thegenetic code and glutamine codon usage cannot be cap-tured by the Brownian motion model. The possibility of areversal from a non-canonical to a canonical code cannotbe ruled out entirely. For example, this could have beenpossible through an ambiguous intermediate stage inwhich TAR codons are both recognized by normal gluta-mine tRNAs and release factors. It is worth mentioninghere that gain-reversal scenarios have been demonstratedin a number of cases. For example in arthropod lineagemitochondrial genomes, reassignment of AGG from Ser toLys has been shown to occur at the base of the arthropodsand has been reversed to the normal code several timesindependently. These reversals have been correlated withmutations in anti-codons of the tRNA-Lys/-Ser and withlow abundance of the AGG codon [28]. Nevertheless, themajor argument against a reversal from a stop to sensereassignment as observed in the Ulvophyceae is the sud-den appearance of many internal stop codons. In addition,it has been suggested that the non-canonical TAR®Glncode is more robust to error than the standard code,further reducing the likelihood of reversals [29]. Takentogether, the reversal of a non-canonical code to the stan-dard code appears highly unlikely.Multiple independent gainsSeveral independent acquisitions of non-canonical codeshave been reported for ciliates [4,30-32]. Stop codonreassignments are surprisingly frequent in this group oforganisms: the same non-canonical TAR®Gln code hasevolved several times, another non-canonical code inwhich TGA codes for tryptophan evolved twice, a non-canonical code that translates TGA to cysteine evolvedonce and a fourth non-canonical codes which translatesTAA into glutamic acid has been reported for three cili-ate species. In the present study, the distribution of thenon-canonical code in the phylogenetic tree wouldrequire three gains: in the Trentepohliales, the Dasycla-dales and the Cladophorales + Blastophysa. Contrary tothe situation in ciliates, however, Ulvophyceae onlyevolved a single type of non-canonical code and theydid so in closely related lineages, and even within theciliates it has been suggested that the reassignment ofCocquyt et al. BMC Evolutionary Biology 2010, 10:327http://www.biomedcentral.com/1471-2148/10/327Page 5 of 9TAR codons to glutamine in some oligohymenophoranlineages and glutamic acid in others may share a com-mon origin [32]. It seems unparsimonious to suggestthat such a rare occurrence as codon reassignmentcould take place three times independently in three clo-sely related lineages, but the possibility cannot beexcluded altogether.Stepwise acquisition modelSeveral studies have suggested that stop codon reassign-ment is a gradual process requiring changes to tRNA andeukaryotic release factor 1 (eRF1) genes [reviewed in[4,5,10]]. In several eukaryotes, the two glutamine tRNAsrecognizing CAG and CAA are also able to read TARcodons by a G·U wobble base pairing at the third antico-don position [33]. Under normal circumstances (standardcode), the eukaryotic release factor outcompetes the glu-tamine tRNAs when it comes to binding TAG and TAA.Mutations in glutamine tRNA copies that allow them tobind TAG and TAA may increase the capacity to trans-late TAG and TAA to glutamine. This leads to an inter-mediate step in which both eRF1 and the mutated tRNAscan easily bind to TAR glutamine codons. The ciliateTetrahymena thermophila has three major glutaminetRNAs, one that recognizes the canonical CAR codons,and two supplementary ones that recognize the non-canonical TAR codons. These two supplementary tRNAswere shown to have evolved from the normal glutaminetRNA [16]. A similar situation likely exists in diplomo-nads [14]. In order to alter the genetic code, mutationsare required not only in glutamine tRNAs but also ineRF1 so they no longer recognize TAG and TAA codons.In ciliates it has been shown that eRF1 sequences arehighly divergent in domain 1 between species with acanonical and non-canonical code, which might suggestthat eRF1 can no longer recognize TAG and TAAcodons in the species with a non-canonical code [31,34].An additional mechanism that constrains the evolvabilityof the genetic code and therefore represents yet anotherpotential step in the process of codon reassignment isnonsense-mediated mRNA decay (NMD) [35,36]. NMDreduces errors in gene expression by eliminating mRNAsthat encode for an incomplete polypeptide due to thepresence of stop codons. In the case of a TAR®Gln codethis NMD mechanism would have to be altered also inorder to prevent degradation of mRNAs containing TAGand TAA codons.A gradual, stepwise acquisition model of codon reas-signment can reconcile the opposed and problematichypotheses of multiple gains versus a single gain withsubsequent loss. For example, the ambiguous intermedi-ate model [5,9] would explain the distribution of thenon-canonical code in Ulvophyceae as follows: mutationsin the anticodons of canonical glutamine tRNAs occurredonce along the branch leading to the orders Trentepoh-liales, Dasycladales, Bryopsidales, Cladophorales and thegenus Blastophysa (Figure 1, black arrow). The presenceof these mutated tRNAs allowed TAG and TAA codonsto be translated to glutamine instead of terminatingtranslation. At this step, the mutated tRNAs competewith eRF1 for the TAA and TAG codons. To completethe transition to the non-canonical code, subsequentmutation of the release factors preventing terminationfor TAG and TAA codons is required. If one assumesthat this step occurred three times independently in theTrentepohliales, Dasycladales and Cladophorales + Blas-tophysa (Figure 1, gray arrows), whereas the mutatedtRNAs decreased in importance or went extinct throughselection or drift along the branch leading to the Bryopsi-dales, the distribution of the non-canonical code in theUlvophyceae would be explained. A detailed comparisonof eukaryotic release factors and glutamine tRNAs in therespective clades of the Ulvophyceae is needed to verifythis evolutionary scenario.Alternatively, the observed distribution could also beexplained under a stepwise version of the codon capturemodel [13]. We then assume that the TAR stop codonsdisappeared from a common ancestor (Figure 1: blackarrow) and were subsequently reassigned independentlyto glutamine codons in the Trentepohliales, Cladophor-ales + Blastophysa and Dasycladales (Figure 1: greyarrows) and reappeared with their old function in theBryopsidales. Although there is some evidence for differ-ent GC usage patterns in Ulvophyceae and earlier-branching Chlorophyta [37], we consider codon capturean unlikely candidate to explain TAR®Gln codon reas-signments because a pressure towards either AT or GCacross the entire genome cannot explain the extinctionof both TAA and TAG codons.The stepwise acquisition model with ambiguous inter-mediates is expected to reduce organismal fitness duringintermediate stages due to competition between releasefactors and glutamine tRNAs resulting in aberrant pep-tides. However, the fact that several eukaryotes havenatural nonsense suppressor tRNAs that can translatestop codons, though generally at low efficiency [33], andthat these have been maintained over evolutionary time,suggest that readthrough may not be a severe problemand could even increase fitness during periods of envir-onmental instability [5]. Considering that the Ulvophy-cean orders diversified during the environmentallyinstable Cryogenian [37-39] and that early-branchingUlvophyceae occur in a variety of habitats includingmarine, freshwater and terrestrial ecosystems, one couldspeculate that genetic code ambiguity may have beenadvantageous during their early evolution.Cocquyt et al. BMC Evolutionary Biology 2010, 10:327http://www.biomedcentral.com/1471-2148/10/327Page 6 of 9Cytological correlates of non-canonical codeIn the ciliates, the multiple appearances of alternativecodes have been attributed to their nuclear characteristics[40]. Ciliates are unicellular organisms with two nuclei: asmall, diploid micronucleus which is not expressed andrepresents the germ line for DNA exchanges during thesexual process, and a large, polyploid macronucleus, whichis actively transcribed and ensures vegetative cell growth,but is not passed on to progeny after sexual conjugationand is replaced by a newly formed macronucleus after anumber of rounds of mitotic division. There is therefore atime lag between the occurrence of mutations in themicronucleus and the expression of these mutations in themacronucleus, and this has been postulated to be a contri-buting factor to why ciliates have evolved alternativegenetic codes more frequently [40]. In this context it isworth mentioning that hexamitid diplomonads, for whicha single origin of a non-canonical code has been shown,have two semi-independent but similar nuclei per cell[17]. The same genetic code is also used by the uninucle-ate enteromonads, but enteromonads are known to haveevolved from within diplomonads [41], so the code origi-nated in a binucleate ancestor. Several ulvophycean groupsare also characterized by multinucleate cells, namely theDasycladales, Bryopsidales, Cladophorales and Blastophysa[22]. However, it should be mentioned that the Trente-pohliales, which also uses the non-canonical code, is char-acterized by uninucleate cells. The Cladophorales andBlastophysa are branched filamentous seaweeds consistingof multinucleate cells with a few to thousands of nucleiarranged in non-motile cytoplasmic domains (siphonocla-dous organization). Members of the Bryopsidales andDasycladales have a siphonous organization: they consistof a single, giant tubular cell with a single giant nucleus orwith numerous nuclei, and complex patterns of cytoplas-mic streaming. The presence of multiple nuclei per cellmight provide an opportunity to experiment with thegenetic code because the cell as an entity might still beable to function normally and each nucleus can potentiallybe passed to the next generation. Despite the fact thatsome eukaryotes with a non-canonical code do not featuremultinucleate cells and that there are plenty of examplesof groups with multinucleate cells that have not evolvedalternative codes, this observation suggests that a multi-nucleate cytology may facilitate codon reassignments.ConclusionsWe demonstrate that the Dasycladales and Cladophoralesshare the TAR to glutamine reassignment with the relatedorder Trentepohliales and the genus Blastophysa, but notwith the Bryopsidales, which is sister to the Dasycladales.We discuss several alternative scenarios for the origin ofthis complex distribution of the non-canonical code:phylogenetic uncertainty, gain-reversal hypothesis, multi-ple independent gains and stepwise acquisition model.Considering the robustness of our phylogeny, the pro-found genetic changes that coincide with codon reassign-ments and the scarcity of codon reassignments, weconclude that a stepwise acquisition model is the mostlikely hypothesis. The present study has provided a frame-work to better understand the evolution of the geneticcode. Further insights will be gained by sequencing andanalyzing release factors and glutamine tRNAs of taxausing non-canonical codes.MethodsRNA isolation, polymerase chain reaction and sequencingTotal RNA was extracted from 43 taxa representing themajor lineages of the Viridiplantae as described previously[22]. Portions of seven nuclear genes (actin, GPI, GapA,OEE1, 40 S ribosomal protein S9 and 60 S ribosomal pro-teins L3 and L17) were amplified, cloned when necessaryand sequenced as described in Cocquyt et al. [[22] andAdditional file 1]. A histone H3 gene was amplified usingthe same PCR conditions with an annealing temperatureof 55°C. The primers were based on a Cladophora coelo-thrix cDNA sequence aligned with GenBank sequencesfrom green algae and land plants (His-F: 5’-ATG GCICGT ACI AAG CAR AC-3’ and His-R: 5’-GGC ATGATG GTS ACS CGC TT-3’). In addition, total RNA wasextracted from Ignatius tetrasporus and the bryopsidaleanspecies Caulerpa cf. racemosa as described previously[21]. Portions of actin, b-tubulin, and HSP90 genes includ-ing the stop codon were amplified from these taxa by 3’RACE using the First Choice RLM-RACE kit (Ambion)using nested degenerate primers (actin-outer: 5’-TACGAA GGA TAC GCA CTN CCN-3’ C and actin-inner:5’- GAG ATC GTG CGN GAY ATH AAR GA-3’; b-tubu-lin-outer: 5’-GAT AAC GAG GCT CTN TAY GAY ATHTG-3’ and b-tubulin-inner: 5’-CCT TTC CGA CGG CTNCAY TTY TT-3’; HSP90-outer: 5’-ATG GTC GAT CCNATH GAY GAR TA-3’ and HSP90-inner: 5’-GCT AAGATG GAG MGN ATH ATG AA-3’). ). These sequenceswere deposited in Genbank under accession numbersHQ332547-HQ332551.Genetic codesThe presence of a non-canonical nuclear genetic code ingreen algal taxa was detected by screening alignments ofnuclear genes for supposed stop codons at positionscoding for glutamine in other green plant taxa and bythe presence of only TGA as a functional stop codon atthe predicted 3’ end of genes. The presence of the stan-dard code was inferred if only canonical glutaminecodons were present and all three stop codons occurredat the predicted end of genes.Cocquyt et al. BMC Evolutionary Biology 2010, 10:327http://www.biomedcentral.com/1471-2148/10/327Page 7 of 9Molecular phylogeneticsWe constructed a reference phylogeny of the Viridiplan-tae based on a 10-gene alignment to study the phyloge-netic distribution of the standard and non-canonicalgenetic code [see Additional file 2]. For a detailed treat-ment of the methods used for tree reconstruction werefer to Cocquyt et al. [22]; we will only provide a sum-mary here. The phylogenetic analysis was carried out onan alignment consisting of the seven nuclear genes men-tioned above, together with SSU nrDNA and the plastidgenes rbcL and atpB. Histone genes were excluded fromthe analysis because they are known to be duplicatedacross genomes [42,43]. Phylogenetic analyses were car-ried out with model-based techniques (maximum likeli-hood and Bayesian inference) after selection of a suitablepartitioning strategy and models of sequence evolution.The model selection procedure proposed 8 partitions:SSU nrDNA was partitioned into loops and stems (2 par-titions) and nuclear and plastid genes were partitionedinto codon positions (2 × 3 partitions). GTR+Γ8 was thepreferred model for all partitions. Noise-reduction tech-niques were applied to counteract the erosion of ancientphylogenetic signal caused by fast-evolving sites. Thephylogenetic tree presented here is based on the 75%slowest-evolving sites [22].Evolution of glutamine codon usageThe evolution of glutamine codon usage was estimatedusing ancestral state inference techniques. The frequencyof the two canonical and two non-canonical glutaminecodons was calculated for each species in the phyloge-netic tree. Codon frequencies were mapped along thereference tree using the ace function of the APE package[44]. This function estimates ancestral character states bymaximum likelihood optimization [45]. The branchlengths were based on ML estimates because we considerthem to be a more relevant approximation of the amountof codon usage evolution that can be expected to takeplace than absolute time [cf. [23]]. The output from APEwas mapped onto the reference tree with TreeGradientsv1.03 [46] to plot ancestral states as colors along a colorgradient.Topological hypothesis testingApproximately unbiased tests [AU test, [47]] were usedto test an alternative relationship between ulvophyceanorders as suggested by the distribution of the canonicalgenetic code (see results). Site-specific likelihoods werecalculated by maximum likelihood optimization in Tree-finder using the same model specifications as for phylo-genetic inference [22]. AU tests were performed withCONSEL V0.1i [48].Additional materialAdditional file 1: GenBank accession numbers. Table with theGenbank accession numbers.Additional file 2: Alignment. nexus file of the data matrix containing 7nuclear genes, SSU nrDNA and plastid genes rbcL and atpB.AbbreviationseRF1: eukaryotic release factor 1; BV: boostrap value; CAR: CAG or CAA;GapA: glyceraldehyde-3-phosphate dehydrogenase; GPI: glucose-6-phosphate isomerase; Gln: glutamine; HSP: heat shock protein; Lys: lysine;ML: maximum likelihood; OEE1: oxygen-evolving enhancer protein; PP:posterior probability; Ser: serine, TAR: TAG or TAA.AcknowledgementsWe thank Caroline Vlaeminck for assisting with the molecular work and WimGillis for IT support. Funding was provided by the Special Research Fund(Ghent University, DOZA-01107605), a grant from the Natural Sciences andEngineering Research Council of Canada (227301), an NSERC postgraduatedoctoral scholarship to GHG and Research Foundation Flanders postdoctoralfellowships to HV, FL. PJK is a Fellow of the Canadian Institute for AdvancedResearch and a Senior Investigator of the Michael Smith Foundation forHealth Research. Phylogenetic analyses were largely carried out on theKERMIT computing cluster (Ghent University).Author details1Phycology Research Group and Center for Molecular Phylogenetics andEvolution, Ghent University, Krijgslaan 281 S8, 9000 Ghent, Belgium.2Canadian Institute for Advanced Research, Department of Botany, Universityof British Columbia, Vancouver, V6 T 1Z4 Canada.Authors’ contributionsEC, ODC and FL designed the study. EC and GHG carried out lab work. ECand FL maintained algal cultures. EC and HV analyzed data. EC and ODCdrafted the manuscript and FL, HV, PJK and GHG contributed to theinterpretation of the data. All authors revised and approved the finalmanuscript.Received: 30 June 2010 Accepted: 26 October 2010Published: 26 October 2010References1. Woese CR: Order in the genetic code. Proc Natl Acad Sci USA 1965,54(1):71-75.2. Crick FHC: The origin of the genetic code. J Mol Biol 1968, 38(3):367-379.3. Jukes TH, Osawa S: Evolutionary changes in the genetic code. CompBiochem Phys B 1993, 106(3):489-494.4. Knight RD, Freeland SJ, Landweber LF: Rewiring the keyboard: Evolvabilityof the genetic code. Nat Rev Genet 2001, 2(1):49-58.5. Santos MAS, Moura G, Massey SE, Tuite MF: Driving change: the evolutionof alternative genetic codes. Trends Genet 2004, 20(2):95-102.6. Osawa S: Evolution of the Genetic Code. Oxford University Press, Oxford;1995.7. Sengupta S, Yang XG, Higgs PG: The mechanisms of codonreassignments in mitochondrial genetic codes. J Mol Evol 2007,64(6):662-688.8. Osawa S, Jukes TH: Codon reassignment (codon capture) in evolution.J Mol Evol 1989, 28:271-278.9. Schultz DW, Yarus M: Transfer RNA mutation and the malleability of thegenetic code. J Mol Biol 1994, 235(5):1377-1380.10. Sengupta S, Higgs PG: A unified model of codon reassignment inalternative genetic codes. Genetics 2005, 170(2):831-840.11. Koonin EV, Novozhilov AS: Origin and Evolution of the Genetic Code: TheUniversal Enigma. IUBMB Life 2009, 61(2):99-111.Cocquyt et al. BMC Evolutionary Biology 2010, 10:327http://www.biomedcentral.com/1471-2148/10/327Page 8 of 912. Massey SE, Moura G, Beltrao P, Almeida R, Garey JR, Tuite MF, Santos MAS:Comparative evolutionary genomics unveils the molecular mechanismof reassignment of the CTG codon in Candida spp. Genome Res 2003,13:544-557.13. Osawa S, Jukes TH, Watanabe K, Muto A: Recent evidence for evolution ofthe genetic code. Microbiological Reviews 1992, 56(1):229-264.14. Keeling PJ, Doolittle WF: A non-canonical genetic code in an earlydiverging eukaryotic lineage. EMBO Journal 1996, 15(9):2285-2290.15. Horowitz S, Gorovsky MA: An unusual genetic code in nuclear genes ofTetrahymena. Proc Natl Acad Sci USA 1985, 82(8):2452-2455.16. Hanyu N, Kuchino Y, Nishimura S, Beier H: Dramatic events in ciliateevolution - alteration of UAA and UAG termination codons to glutaminecodons due to anticodon mutations in two Tetrahymena tRNAs-Gln.EMBO Journal 1986, 5(6):1307-1311.17. Keeling PJ, Leander BS: Characterisation of a non-canonical genetic codein the oxymonad Streblomastix strix. J Mol Biol 2003, 326(5):1337-1349.18. de Koning AP, Noble GP, Heiss AA, Wong J, Keeling PJ: Environmental PCRsurvey to determine the distribution of a non-canonical genetic code inuncultivable oxymonads. Environ Microbiol 2008, 10(1):65-74.19. Schneider SU, de Groot EJ: Sequences of two rbcS cDNA clones ofBatophora oerstedii: structural and evolutionary considerations. CurrGenet 1991, 20(1-2):173-175.20. Schneider SU, Leible MB, Yang XP: Strong homology between the smallsubunit of ribulose-1,5-bisphosphate carboxylase/oxygenase of twospecies of Acetabularia and the occurrence of unusual codon usage. MolGen Genet 1989, 218(3):445-452.21. Gile GH, Novis P, Cragg D, Zuccarello GC, Keeling PJ: The distribution ofelongation factor-1alpha (EF-1α), elongation factor-like (EFL), and a non-canonical genetic code in the Ulvophyceae: discrete genetic characterssupport a consistent phylogenetic framework. J Eukaryot Microbiol 2009,56(4):367-372.22. Cocquyt E, Verbruggen H, Leliaert F, De Clerck O: Evolution andcytological diversification of the green seaweeds (Ulvophyceae).Molecular Biology and Evolution. Mol Biol Evol 2010, 9:2052-2061.23. Cocquyt E, Verbruggen H, Leliaert F, Zechman FW, Sabbe K, De Clerck O:Gain and loss of elongation factor genes in green algae. BMC Evol Biol2009, 9:39.24. Zechman FW, Theriot EC, Zimmer EA, Chapman RL: Phylogeny of theUlvophyceae (Chlorophyta): cladistic analysis of nuclear-encoded rRNAsequence data. J Phycol 1990, 26(4):700-710.25. López-Bautista JM, Chapman RL: Phylogenetic affinities of theTrentepohliales inferred from small-subunit rDNA. Int J Syst Evol Microbiol2003, 53:2099-2106.26. Watanabe S, Nakayama T: Ultrastructure and phylogenetic relationshipsof the unicellular green algae Ignatius tetrasporus and Pseudocharaciumamericanum (Chlorophyta). Phycol Res 2007, 55(1):1-16.27. Martins EP: Estimation of ancestral states of continuous characters: Acomputer simulation study. Syst Biol 1999, 48(3):642-650.28. Abascal F, Posada D, Knight RD, Zardoya R: Parallel evolution of thegenetic code in arthropod mitochondrial genomes. Plos Biol 2006, 4(5):e127.29. Kurnaz ML, Bilgin T, Kurnaz IA: Certain non-standard coding tables appearto be more robusr to error than the standard genetic code. J Mol Evol2010, 70:13-38.30. Baroin Tourancheau A, Tsao N, Klobutcher LA, Pearlman RE, Adoutte A:Genetic code deviations in the ciliates: evidence for multiple andindependent events. EMBO Journal 1995, 14(13):3262-3267.31. Lozupone CA, Knight RD, Landweber LF: The molecular basis of nucleargenetic code change in ciliates. Curr Biol 2001, 11(2):65-74.32. Sánchez-Silva R, Villalobo E, Morin L, Torres A: A new noncanonical nucleargenetic code: Translation of UAA into glutamate. Curr Biol 2003,13(5):442-447.33. Beier H, Grimm M: Misreading of termination codons in eukaryotes bynatural nonsense suppressor tRNAs. Nucleic Acids Res 2001,29(23):4767-4782.34. Inagaki Y, Blouin C, Doolittle WF, Roger AJ: Convergence and constraint ineukaryotic release factor 1 (eRF1) domain 1: the evolution of stop codonspecificity. Nucleic Acids Res 2002, 30(2):532-544.35. Baker KE, Parker R: Nonsense-mediated mRNA decay: terminatingerroneous gene expression. Curr Opin Cell Biol 2004, 16(3):293-299.36. Maquat LE: Nonsense-mediated mRNA decay: Splicing, translation andmRNP dynamics. Nat Rev Mol Cell Bio 2004, 5(2):89-99.37. Cocquyt E: Phylogeny and molecular evolution of green algae (Doctoralthesis). Ghent University; 2009.38. Butterfield NJ, Knoll AH, Swett K: Paleobiology of the NeoproterozoicSvanbergfjellet formation, Spitsbergen. Fossils Strata 1994, 34:1-84.39. Verbruggen H, Ashworth M, LoDuca ST, Vlaeminck C, Cocquyt E, Sauvage T,Zechman FW, Littler DS, Littler MM, Leliaert F, De Clerck O: A multi-locustime-calibrated phylogeny of the siphonous green algae. Mol PhylogenetEvol 2009, 50(3):642-653.40. Cohen J, Adoutte A: Why does the genetic code deviate so easily inciliates? Biol Cell 1995, 85(2-3):105-108.41. Kolisko M, Cepicka I, Hampl V, Leigh J, Roger AJ, Kulda J, Simpson AGB,Flegr J: Molecular phylogeny of diplomonads and enteromonads basedon SSU rRNA, alpha-tubulin and HSP90 genes: Implications for theevolutionary history of the double karyomastigont of diplomonads. BMCEvol Biol 2008, 8:205.42. Nei M, Rooney AP: Concerted and birth-and-death evolution ofmultigene families. Annu Rev Genet 2005, 39:121-152.43. Wahlberg N, Wheat CW: Genomic outposts serve the phylogenomicpioneers: designing novel nuclear markers for genomic DNA extractionsof Lepidoptera. Syst Biol 2008, 57(2):231-242.44. Paradis E, Claude J, Strimmer K: APE: Analyses of Phylogenetics andEvolution in R language. Bioinformatics 2004, 20(2):289-290.45. Schluter D, Price T, Mooers AO, Ludwig D: Likelihood of ancestor states inadaptive radiation. Evolution 1997, 51(6):1699-1711.46. Verbruggen H: TreeGradients v1.02. 2009 [http://www.phycoweb.net].47. Shimodaira H: An approximately unbiased test of phylogenetic treeselection. Syst Biol 2002, 51(3):492-508.48. Shimodaira H, Hasegawa M: CONSEL: for assessing the confidence ofphylogenetic tree selection. Bioinformatics 2001, 17(12):1246-1247.doi:10.1186/1471-2148-10-327Cite this article as: Cocquyt et al.: Complex phylogenetic distribution ofa non-canonical genetic code in green algae. BMC Evolutionary Biology2010 10:327.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 redistributionSubmit your manuscript at www.biomedcentral.com/submitCocquyt et al. BMC Evolutionary Biology 2010, 10:327http://www.biomedcentral.com/1471-2148/10/327Page 9 of 9


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