Arctoa (2002) 11: 31-43 PARAPHYLY OF BRYOPHYTES AND CLOSE RELATIONSHIP OF HORNWORTS AND VASCULAR PLANTS INFERRED FROM ANALYSIS OF CHLOROPLAST rDNA ITS (cpITS) SEQUENCES ÏÀÐÀÔÈËÅÒÈ×ÍÎÑÒÜ ÌÎÕÎÎÁÐÀÇÍÛÕ È ÁËÈÇÊÎÅ ÐÎÄÑÒÂÎ ÀÍÒÎÖÅÐÎÒÎÂÛÕ È ÑÎÑÓÄÈÑÒÛÕ ÐÀÑÒÅÍÈÉ: ÄÀÍÍÛÅ ÀÍÀËÈÇÀ ÏÎÑËÅÄÎÂÀÒÅËÜÍÎÑÒÅÉ ÕËÎÐÎÏËÀÑÒÍÛÕ ÑÏÅÉÑÅÐΠðÄÍÊ T.KH.SAMIGULLIN1, S.P.YACENTYUK1, G.V.DEGTYARYEVA2, K.M.VALIEHO- ROMAN1, V.K.BOBROVA1, I.CAPESIUS3, W.F.MARTIN4, A.V.TROITSKY1, V.R.FILIN2, A.S.ANTONOV1 Ò.Õ.ÑÀÌÈÃÓËËÈÍ1, Ñ.Ï.ßÖÅÍÒÞÊ1, Ã.Â.ÄÅÃÒßÐÅÂÀ2, Ê.Ì.ÂÀËÜÅÕÎ- ÐÎÌÀÍ1, Â.Ê.ÁÎÁÐÎÂÀ1, È.ÊÀÏÅÑÈÓÑ3, Â. Ô. ÌÀÐÒÈÍ4, À.Â.ÒÐÎÈÖÊÈÉ1, Â.Ð.ÔÈËÈÍ2, À.Ñ.ÀÍÒÎÍÎÂ1 Abstract Phylogenetic analysis of nucleotide sequences of chloroplast rDNA internal transcribed spacer (cpITS) regions (cpITS2-4) of 14 species of liverworts, 4 species of hornworts, 20 species of mosses, 7 species of lycopods and 2 species of algae was carried out. Phylogenetic trees constructed by maximum parsimony, maximum likelihood and neighbour-joining methods indicated that bryophytes are not monophyletic. The cpITS data suggest that the hepatic lineage branches most deeply in the land plant topology and that mosses are monophyletic, forming the sister group of lycopods and hornworts. Within the mosses, a conspicuous deletion with distinct phylogenetic distribution was observed in the cpITS3 region. This deletion is absent in other land plants, including the enigmatic genus Takakia, and marks Takakia as an evolutionary lineage distinct from mosses. Ðåçþìå Ïðîâåäåí ôèëîãåíåòè÷åñêèé àíàëèç íóêëåîòèäíûõ ïîñëåäîâàòåëüíîñòåé ñïåéñåðîâ õëîðîïëàñòíîé ðÄÍÊ (cpITS2-4) ó 14 âèäîâ ïå÷åíî÷íèêîâ, 4 âèäîâ àíòîöåðîòîâûõ, 20 âèäîâ ìõîâ, 7 âèäîâ ïëàóíîâèäíûõ è 2 âèäîâ õàðîâûõ âîäîðîñëåé. Ñîãëàñíî òîïîëîãèè ôèëîãåíåòè÷åñêèõ äåðåâüåâ, ðåêîíñòðóèðîâàííûõ ìåòîäàìè ìàêñèìàëüíîãî ïðàâäî- ïîäîáèÿ, áëèæàéøåãî ñâÿçûâàíèÿ è ìàêñèìàëüíîé ýêîíîìèè, ìîõîîáðàçíûå íå ìîíîôèëåòè÷íû. Íà êëàäîãðàììàõ ñðåäè íàçåìíûõ ðàñòåíèé ïåðâûìè îòäåëÿþòñÿ âåòâè, âåäóùèå ê ïå÷åíî÷íèêàì. Ìõè ïðåäñòàþò êàê ìîíîôèëåòè÷åñêàÿ ãðóïïà, ñåñòðèíñêàÿ ê êëàäå, îáúåäèíÿþùåé ïëàóíîâèäíûå è àíòîöåðîòîâûå.  íóêëåîòèäíîé ïîñëåäî- âàòåëüíîñòè cpITS3 ó ìõîâ îáíàðóæåíà ïðîòÿæåííàÿ äåëåöèÿ, îòñóòñòâóþùàÿ ó îñòàëüíûõ íàçåìíûõ ðàñòåíèé, âêëþ÷àÿ Takakia. INTRODUCTION branch distinct from vascular plants, but beyond Until recently, bryophytes (plants in which that there was little consensus concerning precise the gametophytic stage dominates during their evolutionary affinities. In various schemes, life cycle) most often were treated as the phylum bryophytes were viewed as sharing either a Bryophyta and subdivided into three classes, common or a different algal ancestor as vascular Anthocerotae, Hepaticae and Musci. This phylum plants, they were viewed either as descendants was generally considered as an evolutionary or as ancestors of vascular plants. Furthermore, 1 – A.N.Belozersky' Research Institute of physico-chemical biology, Moscow State University, Moscow 119992 Russia – ÍÈÈ Ôèçèêî-õèìè÷åñêîé áèîëîãèè èì. À.Í. Áåëîçåðñêîãî, ÌÃÓ, Ìîñêâà 119992 [fax: 7-(095)-939-31-81; e- mail:[email protected]] 2 – Biological Faculty of Moscow State University, Moscow 119992 Russia – Áèîëîãè÷åñêèé ôàêóëüòåò ÌÃÓ, Ìîñêâà 119992 3 – Botanisches Institut, Neuerheimer Feld 360, Heidelberg 69120,Germany 4 – Institute of Botany III, University of Duesseldorf, Universitaetsstr. 1, 40225 Duesseldorf, Germany 32 SAMIGULLIN T. KH. & AL. some authors defined bryophytes narrowly to Nishiyama, Kato, 1999; Renzaglia & al., 2000; include liverworts and mosses only, in which case Nickrent & al., 2000). The second hypothesis bryophytes and hornworts were considered as suggests that the basal land plant group includes independent branches stemming from a common liverworts, whereby the hornworts are the ancestor with vascular plants. Alternative views youngest representatives among the bryophytes hypothesized that two groups diverged early in and share a common ancestor with vascular land plant evolution: one including the hornworts plants (Lewis & al., 1997; Qiu & al., 1998). The and a separate lineage including mosses and view that hornworts are younger than liverworts vascular plants. As a further alternative, the and branched off form the main stem of land various groups of bryophytes were sometimes plant phylogeny before the divergence of mosses considered as independent phyla originating by and vascular plants (Mishler, Churchill, 1984; parallel evolution from various ancestors Kenrick & Crane, 1997a; Graham & al., 2000) possessing isomorphic change of generation (for does not find support from recent DNA sequence reviews of these various evolutionary scenarios, analyses. see: Meyer, 1958; Crandall-Stotler, 1984; Smith, Molecular methods and the analysis of 1986; Kenrick & Crane, 1997a). additional markers can help discriminate between Recent analysis of morphological and these competing hypotheses. To obtain additional molecular data sets suggest that bryophytes are information on phylogenetic relationships of the paraphyletic and placed at the base of land plant most ancient land plants, we obtained and clade (Kenrick & Crane, 1997a; Qiu, Palmer, 1999; extended the set of molecular data by analyzed Mishler, 2000). Currently, many authors favor nucleotide sequences from the chloroplast rDNA the notion that bryophytes share a common internal transcribed spacers (cpITS2, cpITS3 and ancestor with an ancestral charalean stem, where- cpITS4), as well as of the 4.5S and 5S rRNA by the relationships among bryophytes and genes located in the same operon. vascular plants still remain obscure. MATERIALS AND METHODS The analysis of morphological and bio- DNA was isolated from fresh (in some cases chemical characters alone holds little hope of dried) plant tissues. The list of analyzed species resolving bryophyte phylogeny. Paleobotanical and GeneBank accession numbers for the corres- data from the early phases of land plant ponding ITS sequences are given in Table 1. diversification are comparatively poor in DNA was isolated by CTAB method unambiguous characters and demonstrate rapid (Murray, Thompson, 1980). Amplification and morphological diversification of the land flora sequencing of the analyzed region was carried pioneers. As a consequence, various groups of out as described earlier (Goremykin & al., 1996; Devonian plants are difficult to assign Samigullin & al., 1998). Nucleotide sequences unequivocally to either the bryophytes or the were aligned manually by using Sed program vascular plants on the basis of their preserved of the Vostorg package (Zharkikh & al., 1990). morphological traits (Edwards & al. 1995; Sites with dubious positional homology were Edwards & al. 1998; Wellman & al., 1998; excluded from phylogenetic analysis. The Edwards & Axe, 2000). resulting data matrix was analyzed by programs Although DNA-systematics (genosystematics) Tree-puzzle (Strimmer, von Haeseler, 1996) and has been applied to study the problem of early PAUP* version 4.0b8 (Swofford, 2000). evolution of land plants in the end of the past Maximum parsimony (MP) analysis century, the current data have yet to produce a involved a heuristic search conducted with unanimously supported view of early plant PAUP* using TBR (tree-bisection-recon- diversification. Molecular data are consistent with nection) branch swapping with character states at least two competing hypothesis. Under the specified as unordered and equally weighted. first view, hornworts would be the most ancient 100 random addition replicates were performed. land plants and mosses (or mosses together with In the parsimony analyses all gaps were treated liverworts) form a clade sister to vascular plants as missing data. (Malek & al., 1996; Beckert & al., 1999; Garbary, Bootstrap analyses (Felsenstein, 1985) were Renzaglia, 1998; Hedderson & al., 1998; performed to assess the degree of support for Paraphyly of bryophytes inferred from cpITS 33 particular branches on the trees. Bootstrap nucleotide composition of cpITS2-4 of the values were calculated from 100 replicate analyzed plants appears to be rather uniform analyses with TBR branch swapping and 5 and thus suitable for phylogenetic analysis. replicates of random addition sequence of taxa. Tamura-Nei distances between all the pairs of 5000 most parsimonious trees from each replicate sequences were lower than 0.5, and transition/ were saved. transversion ratio was 2.9. Maximum likelihood (ML) phylogeny Phylogenetic trees constructed by ML, NJ estimation was explored using the Tree-puzzle and MP methods are presented in Figs. 2-5. These program. The Tamura-Nei (1993) model of topologies are not completely identical, but the substitutions with eight Gamma rate categories relative position of the main bryophyte groups and approximate parameter estimation was used. is constant across trees, irrespective of the Distance trees were calculated using the reconstruction algorithm and the method of neighbor-joining (NJ) method (Saitou, Nei, tree optimization used. 1987) as it implemented in PAUP*. The In all trees, mosses, hornworts and lycopods Tamura-Nei (1993) model of substitutions was form a clade with high bootstrap support (94%, used. Rates across sites were assumed to follow 82%, 75% for ML, NJ and MP trees, respectively). gamma distribution. 1000 bootstrap resamplings Among mosses the basal position is occupied were performed. Insertions and deletions were by Sphagnum and Andreaea. Buxbaumia, not taken into account. Polytrichales (Atrichum, Pogonatum, Poly- RESULTS trichum) and Tetraphis branch off subsequently.