Syst.Geogr. Pl. 71: 25-35(2001) "r

Molecularsystematics in thegenus (Dictyotales, Phaeophyta): a firstattempt based on restrictionpatterns of the InternalTranscribed Spacer 1 ofthe rDNA (ARDRA-ITS1)

OlivierDe Clerckaxc,Paul De Vosb,Monique Gillisb & EricCoppejansa a OnderzoeksgroepAlgologie, Laboratorium voorPlantkunde, Universiteit Gent, K.L.Ledeganckstraat 35,B - 9000Ghent, Belgium bLaboratorium voorMicrobiologie, Universiteit Gent, K.L.Ledeganckstraat 35,B - 9000Ghent, Belgium cCorresponding author

Abstract. - This studyrepresents a firstattempt to reconstructthe phylogenetic relationships in the genus Dictyotaby means of an RFLP analysis of theamplified Internal Transcribed Spacer 1 of the ribosomalregion (ARDRA-ITS1). To overcomeDNA extractionproblems, a newprotocol based on a combination of a CTAB methodand a Sephaglass Bandprep Kit is presented.Seven species, mostly includingmore than one specimen,were included.Padina boergeseniiwas used as an outgroup.The inferredphylogenetic picture was congruentwith the morphological-anatomicaldata. Specimens belonging to the same species invariablygroup togetherand differencesamong geographically isolated populations are reflectedin thetree. Parsimony-based bootstrap and jackknifesupports for basal nodes are generally very low and phylogeneticrelationships between species are poorly resolved.This poor resolutioncan probablybe attributedto theconsiderable variationin lengthof the ITS] regionamong thespecies. Thisvariation causes difficultiesin sequence alignmentsand calls into question the use of thismarker as a general tool in thegenus.

Key words: Dictyota,Phaeophyta, marine algae, molecularsystematics, ARDRA, ITS]

1 Introduction

Systematicstudies in the genus Dictyota (Dictyotales, Phaeophyta) have been seriously hampered by the lackof easily defined characters distinguishing species and by considerable intraspecific morphological variation.The lack of understanding ofthe variation of vegetative characters has resulted in developmen- taland other epigenetic morphological forms being described as differentspecies (Phillips 1992). Recent efforts,which culminated in a revisionof the members of the genus from the northern Atlantic Ocean (Hbrnig& al. 1992a,b),have lead to a betterunderstanding of this morphological variability. The Australianrepresentatives of the genus, characterised by a multilayeredmedulla (former genus Di- lophus),were studied by Phillips (1992) whoemphasised the use ofhitherto neglected characters such as thenumber of sporangial stalk cells and the size ofmature tetrasporangia inthe distinction of species. However,whereas these new morphological characters proved very useful to discriminatenotoriously difficultspecies [e.g. D. fastigiataSonder and D. gunniana(J.Agardh) Hirnig, Schnetter&

Subjectto copyright. All rightsreserved. O 2001National Botanic Garden of Belgium Permissionfor use mustalways be obtainedfrom the National Botanic Garden of Belgium. ISSN 1374-7886

This content downloaded on Thu, 10 Jan 2013 04:02:20 AM All use subject to JSTOR Terms and Conditions Syst.Geogr. Pl. 71 (2001)

Prud'hommevan Reine], for a majorityof tropical species such characters are non-discriminative (De Clerck& Coppejans1999; De Clerck1999). Mosttropical representatives ofDictyota have nearly identicalreproductive structures and thallus anatomy, and can generallybe distinguishedonly on the basis ofoverall habit and sizes of interdichotomies andcells. As DNA analysishas beenshown to be a veryeffective tool to investigaterelationships between organismsor between groups of organisms, phycologists have also begunto apply a varietyof genomic techniques.Hardly any results based on moleculardata exist for Dictyotales. Chloroplast DNA was partlycharacterised for D. dichotoma(Hudson) Lamouroux by Kuhsel & Kowallik(1985, 1987) anda partialsequence of the SSU rDNA (873 nucleotides,3'- end) of Taonia atomaria(Woodward) J.Agardhwas published by Tan & Druehl(1993). Lee & King(1996) triedto reconstruct the phylogeny of severalgenera of theDictyotales, but a criticalexamination of thedata showed that the reported sequences do notoriginate from . This studyinitiates the use of molecularmarkers for taxonomicpurposes in Dictyota.Comparison of rDNA characteristicsbetween different species, has shown to be a valuableapproach for phylogenetic inferences in thealgae (Zechman& al. 1990). However,the taxonomic resolution ofthe Small and Large Subunits is in brownalgae limited to the level oforders and families and to a lesserextent to genera(Saunders & Druehl1992, 1993; Tan & Druehl 1993;Saunders & Kraft1995; Rousseau & al. 1997).Characteristics ofthe Internal Transcribed Spacers (ITS 1 andITS2) offera much better taxonomic resolution at species and generic levels (Stache-Crain & al. 1997; Peters& al. 1997). Therefore,this study attempts to characterisethe ITS of Dictyotaand assessesits use in reconstructing phylogenetic relationships. It was ourinitial intention to compareboth theITS 1 andITS2, but amplification ofthe ITS2 regionproved to be problematicand no PCR products of sufficientquality could be obtainedwith DNA fromseveral specimens. Hence thisstudy only investigatesthe applicability of the ITS 1 regionto unravelthe phylogeny of Dictyota. Specifically, we performedan AmplifiedrDNA RestrictionAnalysis (ARDRA) (Heyndrickx& al. 1996) ofthis DNA region.In order to obtain the highest discrimination several restriction enzymes were used. The banding patternswere combined and the obtained composite pattern was analysedphylogenetically. It is beyond thescope ofthis introductory study to reconstructa complete phylogeny of the genus, because only a limitednumber of species was included.Nevertheless, preliminary hypotheses on interrelationshipsare discussed.

2 Materialand methods

2.1 DNA extraction Algalsamples were desiccated in silica gel (Chase & Hills 1991).Voucher specimens were processed as herbariumspecimens and depositedin GENT (table 1). Totalcellular DNA wasextracted from 20-50 mg dry material. Samples were ground in liquid nitrogen in 800 pl of2% CTAB extractionbuffer (100 mMTris-HC1, pH 8.0, 1.4 mMNaC1, 20 mMEDTA, pH 8.0,2% w/vCTAB, 2 pC mercapto-ethanol).After 2 hoursof incubationat roomtemperature, two choloroform-isoamylalcohol (24:1) extractionswere performed.The DNA wasthen bound to Sephaglassbeads (20 pl) withthe addition of 1 mlof Gel Solubilizer(6N Nal). Following centrifugation(1 min, 15,000 g) thepellet containing the glass beads and the DNA wasthen washed in 160pI ofwash buffer, after whichI mlof 50% EtOH wasadded. The samples were placed overnight at4'C, centrifuged(1 min, 15.000 g) andair dried. DNA was elutedfrom the glass beads by addition of 100pl of0.1 X TE or H20 at55?C

2.2 PolymeraseChain Reaction PCR wasperformed in50 pl reactionvolumes, containing 5 pl oftemplate DNA, 200 pM dNTP's,0.5 pM ofeach primer, 1.5 mM MgClzstandardEurotaq reaction buffer and 1.25 units of Taq DNA polymerase(Eurogentec, Belgium). The PCR ina Perkin-Elmer thermocycler(Gene Amp PCR System 9601) proceeded as follows: an initial denaturation stepat 94?C for 4 minfollowed by 33 cyclesof I minat 94?C, 2 minat 55?C and2 minat 72?C. Amplificationproducts were checked for correct molecular weight and yieldon EtBr-stained1% agarosegels. Several primer combinations were tested, of which the TW5-5.8S IR couplewas themost successful.Reaction productswere also obtainedwith TW7-LSU 15R, ITS 1-LSU115R, ITS 1-JO6and TW7-JO6 (fig. l, table 2). TW5,TW7, ITS 1 andITS4 areuniversal primers initially designed by White & al. (1990). JO6was designed by Van Oppen & al. (1995). 5.8S 1R, 5.8S1F and LSU15R are Phaeophytaspecific primers designed by Milller& al. (1998) and Peters& Burkhardt (1998).

26

This content downloaded on Thu, 10 Jan 2013 04:02:20 AM All use subject to JSTOR Terms and Conditions De Clerck,De Vos, Gillis& Coppejans,Molecular systematics ofDictyota (Phaeophyta)

Table 1. Collectionsites and referencenumbers of the samples ODC andHEC referto theherbarium collections (GENT) ofthe first and last author respectively. Species Locality Herbarium number D. bartayresianaLamouroux Tanzania,Dar es Salaam,Mbudya Island ODC 466 Tanzania,Pemba, Tondooni, Verani ODC 490 Tanzania,Zanzibar, Zanzibar, Chwaka ODC 710 Tanzania,Zanzibar, Pongwe ODC 715 Papua NewGuinea, Motupore Island ODC 272 D. ceylanicaKOtzing Tanzania,Dar es Salaam,Pongojo Island ODC 688 D. cervicornisKOtzing Tanzania,Zanzibar, Chwaka ODC 707 Tanzania,Zanzibar, Pongwe ODC 713 D. crispataLamouroux Tanzania,Zanzibar, Ras Fumba ODC 332 D. dichotoma(Hudson) Lamouroux France,Bretagne ODC 176 The Netherlands,Zeeland ODC 114 SouthAfrica, Cape Town HEC 10912 SouthAfrica, Cape Town ODC 822 D. humifusaH6rnig, Schnetter & Coppejans Tanzania,Zanzibar, Matemwe ODC 738 D. pulchellaH6rnig & Schnetter Panama,San Bias FieldStation ODC 594 Padinaboergesenii Allender & Kraft Tanzania,Pemba, Tondooni, Verani ODC 499

ITS4 5.8SIR JO06LSUIll5R

ssu ITS1 5.8S ITS2 LSU

TW5B TW7B ITS1 5.8S1F

Figure1. Positionof the primers in the nuclear encoded rDNA used for amplification.

Table 2. Primersused for amplifications Primer Direction Length(bp) Position Sequence(5'-3') TW5B forward 22 1541 AACTTAAAGAAATTGACGGAAG TW7B forward 23 1829 GGGGCAATAACAGGTCTGTGATG ITS1 forward 19 2217 TCCGTAGGTGAACCTGCGG 5.8S1F forward 20 2784 TGATGATTCACTGGATTCTG 5.8S1R reverse 20 2804 CAGAATCCAGTGAATCATCA J06 reverse 21 3134 ATATGCTTAAGTTCAGCGGGT ITS4 reverse 20 3148 TCCTCCGCTTATTGATATGC LSU115R reverse 20 3245 CTCTCCAGACTACAATTCGG

2.3 AmplifiedrDNA Restriction Analysis (ARDRA) A totalof 5 restrictionenzymes was selected: DpnlI, HpaII, MroI, RsaI andTsp509I. As therestriction site of MroIl (a hexa-cutter), islocated near the end of the SSU rDNA,it enabled us to verify the PCR productfor correct identity. Indeed the smaller fragment of thisdigest is 490-492 bp long and represents the conservative 3'-end of the SSU. Thelonger fragment is variable in size and represents theITS 1 region.Digests were completed according to themanufacturers' instructions. Reaction volumes were either 20 or 30 pl, dependingonthe concentration ofthe PCR product.A 20 pl reactionvolume consists of 10 pllPCR product,7.5 pl H20,2 pllreaction bufferand 5 Unitsenzyme. Digested DNA wassize fractionated byhorizontal gel electrophoresis at90 V (3 V/cm)for 130 min in 1 X TBE-buffer(Heyndrickx & al. 1996).Samples were loaded on an agarosegel (15 X 10 cm,20 wells)of 2% (w/v)Metaphor agarose(FMC Bioproducts,Rockland). The outer two wells were left empty and of the 18 remainingwells, 15 wereused for the rDNA digestsand 3 (one in themiddle and one on eachborder of thegel) forthe molecular weight marker. The markerwas a combinationof twoseparate ladders each with a differentrange: the Biorad 100 bp molecular(100 to 1000bp) andBiorad 500 bp

27

This content downloaded on Thu, 10 Jan 2013 04:02:20 AM All use subject to JSTOR Terms and Conditions Syst.Geogr. P1. 71 (2001)

molecularladders (500 to 3000 bp). Banding patterns were analysed with GelCompar 4.0 (AppliedMaths, Kortrijk, Belgium) in three steps:conversion, normalisation and numerical analysis. Normalisation ofthe patterns is ofprimary importance to achievereliable databasesby which an objectivecomparison of the banding patterns can be performed.All ARDRApatterns were normalised using thecombination ofthe two ladders. After normalisation the patterns obtained with each of the five restriction enzymes were assembled. A similaritymatrix between each pair of combined patterns was calculated with a bandposition tolerance of 1.0%using the Dice similaritycoefficient (SD) accordingto the formula 2nAB/(nA+nB) with nAB, the number of bands common for pattern A andB, nA thetotal number of bands in pattern A, nB thetotal number of bands in pattern B. Dendrogramswere constructed from the similarity matrices,using the UPGMA (Unweighted Pair Group Method using arithmetic Averages) clustering algorithm. In parallelwith the distancemethod the dataset was also analysed using parsimony analysis. Phylogeneticrelationships were inferred using PAUP* 4.0b2 (Swofford 1999). was usedas an outgroup. Most-parsimonioustrees were constructed using the heuristic search option, random sequence addition (500 replicates)and TBR branchswapping. Support for individual internal branches was determined by a bootstrap(Felsenstein 1985) and parsimony jack- knifinganalysis (Farris & al. 1996)as implementedinPAUP (100 replicates).

Table 3. Sizes offragments in thedigests obtained with all enzymesused

Species & Restrictionenzyme Herb. nr. Tsp5091 Mrol Dpnll Hpall Rsal Total

D. bartayresiana 1544/257/134 1461/490 649/427/232/145 1008/308/273/174 528/224/200/189 1951 ODC 466

D. bartayresiana 1488/260/137 1461/490 637/427/232/145 1003/308/273/174 523/221/202/190 1951 ODC 490

D. bartayresiana 1488/260/137 1426/492 637/431/237/150 998/305/268/169 523/224/205/189 1918 ODC 710

D. bartayresiana 1506/260/136 1470/490 637/432/237 988/304/269/174 523/222/203/188 1960 ODC 715

D. bartayresiana 919/461/262/134 1384/490 654/421/232/142 507/220/199/183 523/225/212/198 1874 ODC 275

D. ceylanica 1291/262/235/136 1515/492 798/437/336/314/240/175 526/343/288/143 1525/519 2007 ODC 688

D. cervicornis 449/258/131/118 776/492 433/275 /236/181/143 397/309/222/170/109 /85 525/342/289/141 1268 ODC 707

D. cervicornis 446/265/134/123 754/492 418/263/235/183/145 382/307/222/171/108/86 692/549/297/221 1246 ODC 713

D. crispata 433/255/130/115 758/494 430/293/240/159/132 433/269/167/142 519/318/287/160 1252 ODC 332

D. dichotoma 972/478 1376/492 1247/430/238 512/327/284/142 1360/528 1868 Bretagne

D. dichotoma 973/474/266 1376 /490 1276/431/236 310/281/173 1368/518 1866 Zeeland

D. dichotoma 943/467/264 1376/492 637/278/326/192 449/259/132/121 1321/526 1868 var. intricata Cape Town D. dichotoma 943/474/266 1376/492 637/278/326/192 449/261/132/121 1328/520 1868 var. intricata Cape Town D. humifusa 1798 1810/492 892/433/390/237 391/275 1908/519 2302 ODC 738

D. pulchella 407/373/360/297/262/1685/492 745/438/261/240 433/269/168/143/123 1628/509 2177 ODC 594 137

P. boergesenii 798/259/133 866/492 428/189 519/220/200/186 519/406/232/114 1358 ODC 499

28

This content downloaded on Thu, 10 Jan 2013 04:02:20 AM All use subject to JSTOR Terms and Conditions De Clerck,De Vos, Gillis& Coppejans,Molecular systematics ofDictyota (Phaeophyta)

I I I I I I 50 60 70 80 90 100 Dictyotabartayresiana ODC 710 Dictyotabartayresiana ODC 715 Dictyotabartayresiana ODC 466 Dictyotabartayresiana ODC 490 Dictyotabartayresiana ODC 272 Dictyotadichotoma Bretagne

Dictyotadichotoma Zeeland Dictyotadich. var. intricata Cape Town Dictyotadich. var. intricata Cape Town Dictyotaceylanica ODC 688

Dictyotapulchella ODC 594 Dictyotahumifusa ODC 738 Dictyotacervicornis ODC 707 Dictyotacervicornis ODC 713 Dictyotacrispata ODC 332 Padinaboergesenii ODC 499

Figure 2. UPGMA treeof seven species (16 isolates) using the Dice similaritycoefficient; tolerance was set at 1.0%.

3 Results

3.1 Bandingpatterns Theamplified ITS 1 regionshowed much variation in length between the different species. The shortest one (D. cervicornis,ODC 713, 1246 bp) being1056 bp shorterthan the longest one (D. humifusa, ODC 738, 2302 bp). Also betweenisolates of a singlespecies some variation, up to 86 bp, was observed(e.g. D. bartayresiana).All specimenswere digested with the same set of five enzymes which resultedin 11-21bands per specimen (table 3). Digestionwith MroI always resulted in twobands of whichthe lowermolecular weight band (490-492 bp) occurredin all specimenswhile the higher molecularweight band varied from 754 bp inD. cervicornisto 1685 bp inD. pulchella.Digestion with Tsp5091resulted in 3-4 bandsin mostspecimens except for D. humifusain whichonly a singlelarge bandwas found(1798 bp). The DpnII andHpaII enzymerevealed 4 to6 bandsper digest; RsaI 2 to4 bandsper digest.

3.2 Pheneticanalysis The distancetree was constructedbased on combinedpatterns of the individual digests (fig. 2). Below a cut-offvalue of 50% similaritythe first lineage contains D. cervicornisand D. crispataas wellas the outgroup,P. boergesenii.The secondlineage groups the remaining Dictyota specimens. At about 58% similaritythe latter lineage can be subdividedin a groupwhich unites D. bartayresianaand D. dicho- toma on the one handand D. ceylanica,D. humifusaand D. pulchellaon theother hand. All D. bartayresianaisolates are grouped with a minimumsimilarity of90% betweenthe specimen from Papua

29

This content downloaded on Thu, 10 Jan 2013 04:02:20 AM All use subject to JSTOR Terms and Conditions Syst.Geogr. Pl. 71 (2001)

New Guinea(ODC 272) and theTanzanian specimens (ODC 466, 490, 710, 1713). The banding patternsof the Tanzanian D. bartayresianaspecimens were very similar, 97-100%. The D. dichotoma group,can be subdividedin a SouthAfrican and a northeastAtlantic lineage which show relatively low similarities(78%). Betweenthe northeast Atlantic isolates a similarityof 90% was observed,while specimensfrom Cape Town,South Africa were identical in bandingpattern. D. ceylanica,D. humifusa andD. pulchellagroup together at about63% similarityand D. ceylanicaappears to be a sistertaxon of D. pulchella at 70% similarity.

Dictyotabartayresiana ODC 710

89/-85 Dictyotabartayresiana ODC 715 Dictyotabartayresiana ODC 466 68/76 Dictyotabartayresiana ODC 490 60/58 Dictyotabartayresiana ODC 272 Dictyotacervicornis ODC 707 98/96 59/57 Dictyotacervicornis ODC 713 Dictyotacrispata ODC 332 81/75 Dictyotaceylanica ODC 688

Dictyotapulchella ODC 594 96 /84 Dictyotadichotoma Zeeland Dictyotadichotoma Bretagne 83/68 100/100 Dictyotadich. var. intricata Cape Town Dictyotadich. var. intricata Cape Town

Dictyotahumifusa ODC 738 Padinaboergesenii ODC 499

Figure 3. Strictconsensus tree resulting from parsimony. Bootstrap and jackknifevalues (100 replicates), indicatedabove internodalbranches, are shownfor brancheswith support 2 50%.

3.3 Parsimonyanalysis Maximumparsimony produced 4 mostparsimonious trees of 116 steps.These differedonly in the relativeplacement of the respective specimens of D. bartayresiana.The overalltopology of the strict consensustree (fig. 3) is similarto the distance tree. There is perfectagreement with the distance tree on theposition of the terminal nodes. The five specimens of D. bartayresianaform a singleclade, with the specimenfrom Papua New Guinea(ODC 272) beinga sisterto the East African specimens (ODC 466, ODC 490, ODC 710, ODC 715). D. cervicornisand D. crispata grouptogether with the D. bartayre- sianaclade, but bootstrap (60) andjacknife (58) valuesare relatively low. D. ceylanicaand D. pulchella forma well-supportedclade butthe position of thisgroup remains unresolved. The D. dichotoma complexis well-supportedas a monophyleticgroup. The position of D. humnifusa,which groups as sister tothe D. dichotomaclade, however, is notsupported above 50%.

30

This content downloaded on Thu, 10 Jan 2013 04:02:20 AM All use subject to JSTOR Terms and Conditions De Clerck,De Vos, Gillis& Coppejans,Molecular systematics of Dictyota(Phaeophyta)

4 Discussion

DNA extractioninbrown algae in general and in the Dictyotales in particular poses important difficulties as mostmethods that are successfulin othergroups of organisms cannot be appliedwith success. The methodapplied in thisstudy, a combinationof a CTAB method(Doyle & Doyle 1987; Rogers& al. 1989) anda SephaglasBandprep Kit (Pharmacia), was successfulfor a numberof taxa, but yielded no amplifiableDNA forothers. In generalthe method was unsuccessfulwhen applied to thick,coarse, parenchymatousthalli (e.g. Spatoglossumasperum J.Agardh,Stoechospermum polypodioides (Lamou- roux)J.Agardh and zonale (Lamouroux) Papenfuss). No DNA couldbe extractedfrom a numberof medium sized Dictyota species either (e.g. D. ciliolataSonder ex Kiitzing,and D. friabilis Setchell).Perhaps these failures can be relatedto the amount of contaminating polyphenols, which show considerablevariation within and amongspecies (Targett & Arnold1998). To improvethe DNA extractionmethod for applicability to a widervariety of Dictyotalean genera, either the present method shouldbe modifiedor moretime-consuming methods should be used.The lattercan takeup to several daysand require special equipment (e.g. CsCl densitygradient ultracentrifugation). Quality of the PCR productsis oftenrelated to the purity of the DNA samples.If the template DNA is ofhigh quality, PCR willyield a pureproduct of sufficient quantity that can be reproduciblydigested with restriction enzymes. However,amplification with an impureDNA templatewill rarely result in goodPCR products.The generatedbanding patterns are then frequently difficult to interpretas the bands are weak and because of thepresence of indiscrete bands which may form a smearover the gel. Even in DNA ofsufficient quality the yieldof thePCR forthe ITS 1 regionwas alwaysmuch higher than for the ITS2 region,despite several attemptsto adjustthe PCR conditions.No immediateexplanation could be found.Perhaps secondarystructures of theInternal Transcribed Spacer hinder the amplification of thelocus. The existenceof secondarystructures in thisnon-coding part of the genome has beendemonstrated for a numberof algal groups(Coleman & Mai 1997; Mai & Coleman1997; Peters& al. 1997),but their functionstill remains unclear. Secondary structures, such as stem-loops,are knownto renderPCR amplificationand sequencing more difficult (Saunders & Druehl1993).

The ARDRA-ITS1 techniqueproved successful in thegenus Dictyota. Restriction patterns are highly reproducibleand specimens that belong to a singlespecies based on morphologicalobservations group togetherin thetrees obtained. Additionally, geographical differences are reflectedin therestriction patterns.A singlespecimen of D. bartayresianafrom Papua New Guineashowed a somewhatlower similaritytothe other 4 specimensof the species from Tanzania, but still grouped in the same lineage. In a similarway the D. dichotomavar. intricataspecimens from South Africa were distinct from the Europeanrepresentatives ofthe same species. The tendency to reveal geographic variation in restriction patternscould indicate that Dictyotales are a ratherold lineagewithin the Phaeophyta. Moreover, the observationthat sequence data of the ITS 1 regionof D. cervicornis(De Clerck,unpublished) could not be alignedto similar sequences of other brown algae might be indicativeof an earlydifferentiation ofthe order.ITS regionsof other brown algae are generally remarkably conserved. Saunders & Druehl(1993) noticeda 90% similarityinthe ITS 1 regionbetween two kelp species from two morphologically diverse families,Alaria marginataPostels & Ruprechtand Postelsia palmaeformisRuprecht. Similarly, ITS 1 and ITS2 sequencesof several species of the Desmarestiaceae only diverged 0.2-14.6% (Peters & al. 1997). It is mostlikely that the divergence patterns observed from restriction patterns in thegenus Dictyotawill be mirroredby a muchhigher divergence in sequencedata of the same molecule. This is also reflectedin thelength differences of thePCR productsamong the different isolates. Neither Saunders& Druehl(1993), Stache-Crain& al. (1997) norPeters & al. (1997) reportedlarge inserts or deletions.Length differences as observedin thegenus Dictyota have been commonlyobserved in Chlorophyta(Bakker & al. 1995), buthave neverbeen reported for any group of Phaeophyta.The observedhigh divergences of the ITS regionin Dictyota affect its use inreconstructing phylogenies. On

31

This content downloaded on Thu, 10 Jan 2013 04:02:20 AM All use subject to JSTOR Terms and Conditions Syst.Geogr. Pl. 71 (2001) species andpopulation level the ITS regionmay provide a goodresolution by means of the ARDRA technique,but more distant relationships (e.g. betweengenera) could be affectedby convergences, parallelismand non-homologyofbands with a less reliabletree as thedirect result. Towards future sequencework the large differences inlength could pose seriousproblems as theindividual sequences maybe extremelydifficult to align. Thedistance and parsimony analyses result in trees with similar topologies. Especially towards the terminalnodes, both analyses are highlycongruent. Internal nodes show a numberof incongruences. Fromthe bootstrap and jackknife supports it becomes clear, however, that most of the internal nodes are onlyweakly supported, limiting the value of the ARDRA-ITS 1 techniquefor uncovering intraspecific relationshipsin Dictyota. D. cervicornisand D. crispatashare a numberof morphological characters not encountered in the otherrepresentatives ofthe genus included in this analysis such as combinationof surface proliferations and involucratesporangia. Detailed study of the paraphyses which surround the antheridial sori of D. crispatahave shown to be ofa differentnature than those observed in all otherDictyota species, being multicellularand pigmented(De Clerck1999; De Clerck& Coppejans,unpublished results). These morphologicalobservations are reflectedin the distance tree, where both species form a distinctlineage togetherwith P. boergesenii,which was selectedas an outgroup.In theparsimony analysis, however, D. cervicornisand D. crispatagroup together with D. bartayresiana,although the bootstrap (60) and jackknifing(58) supportsare low. At present it, therefore, remains uncertain whether the genus Dictyota as currentlydefined is polyphyleticand a newgenus should be describedto accommodate D. cervicornis andD. crispata.The bandingpatterns of P. boergeseniiand D. cervicornis-crispatamight be superfi- ciallysimilar but in fact not homologous. TheDictyota dichotoma complex has been studied in detailby Hornig & Schnetter(1988), resulting in thedescription of D. pulchellaHirnig & Schnetter.The latterspecies was erectedto accommodate thetropical Caribbean representatives ofthe D. dichotomacomplex after hybridisation experiments showedthem to be reproductivelyisolated from European warm-temperate D. dichotoma specimens. On morphologicalgrounds, however, D. pulchellais particularlyhard to differentiatefrom genuine D. dichotomadue to substantial amounts of morphological variation in bothspecies (Hornig & al. 1992b). D. ceylanicaKiitzing, originally described from Sri Lanka and widely represented inthe Indian Ocean, is morphologicallyindistinguishable from the Caribbean D. pulchella(De Clerck1999). Ouranalysis shows bothspecies to be sistertaxa, which could then have diverged from a Tethyanancestor. It is surprisingthat both taxa are only distantly related to the warm temperate D. dichotoma,to which they aremorphologically very similar. The position of D. humifusaremains unresolved. In thedistance tree D. humifusagroups together with D. ceylanicaand D. pulchella;the parsimony analysis places D. humifusain a basalpolytomy.

5 Conclusion

The use of restrictiondata to inferphylogenies has been subjectto considerabledebate. The main criticisminvolves homology of bandsand treatmentof insertionsand deletions(for more complete discussions,see Templeton1983; DeBry & Slade 1985;Albert & al. 1992). Generallygene sequences providea betterresolution and theresulting phylogeny are less subjectto criticism.Despite these criticisms,trees based on restrictiondata are in generalcongruent to trees based on genesequences in bacteria(Heyndrickx & al. 1996).The phylogeny presented in this study should, however, be interpreted withcare. The distance and parsimony analysis are largely congruent at theterminal nodes, but differ in theplacement of the D. cervicornis-crispataclade and in the position of D. humifusa.Terminal nodes seemto be wellsupported as indicatedby bootstrap and jackknife values, but the reliability of the trees decreasestowards the base. Otherfactors, such as thelimited number of taxa included, may also affect

32

This content downloaded on Thu, 10 Jan 2013 04:02:20 AM All use subject to JSTOR Terms and Conditions De Clerck,De Vos, Gillis& Coppejans,Molecular systematics of Dictyota(Phaeophyta) the value ofthe present phylogenetic picture. Most of the species used can be regardedas common, tropicalrepresentatives of thegenus that are differentiatedmorphologically only by sizes of inter- dichotomiesand corticaland medullarycells (exceptfor D. cervicornisand D. crispata).Warm temperatespecies, (not included in theanalysis) are generallybetter characterised by well-defined differencesin theirreproductive morphology (arrangement ofsporangia in sori,number of sporangial stalkcells, involucra) or cortical and medullary structure (De Clerck& Coppejans1999). Inclusionof thesespecies could affect the topology of the tree considerably. Therefore, the position of D. cervicornis and D. crispata,which render the genus Dictyota paraphyletic in theUPGMA analysis,should be consideredpreliminary. Although, morphological investigations have revealedthat the presence of multicellular,pigmented paraphyses is uniquewithin the genus (De Clerck& Coppejans,unpublished results),a better hypothesis for the true relationships ofD. cervicornisand D. crispatato the rest of the genuscan onlybe obtainedby inclusion of a seriesof related genera. Only then will it become clear whetherthe genus is trulypara- or polyphyletic and the description of new genera is justified.Inclusion of species of relatedgenera could also bringnew insights in therelationship between Dictyota and Pachydictyon.The lattergenus was createdto accommodatespecies characterised by a multilayered cortex.Future research should also examinethe relationships between Dictyota species with a multilay- eredmedulla (former genus Dilophus) and those with a unilayeredmedulla in order to provide supple- mentaryevidence either for or against the merging of Dilophus with Dictyota by Hoirnig & al. (1992a) or inopposition to thisidea. Theobserved wide divergences ofthe ITS regionof Dictyota species suggests that a moreconserved geneticlocus might be moreappropriate for unravelling the phylogenetic relationships of thisgroup of organisms.The lengthvariations and insertions and deletions in particularrender the ITS 1 regionless suitablefor phylogenetic research at lowertaxonomic ranks. In Rhodophytathe large subunit of the RuBisCoenzyme (rbcL) proved very appropriate for inferring phylogenies at theordinal and family level (Freshwater& al. 1995; Fredericq& al. 1996, Hommersand& al. 1999). In Phaeophytathe only publisheddata come from a studyof Ectocarpus and Kuckuckia (Stache-Crain & al. 1997),where the RuBisCo spacerwas usedto comparesequence divergence of geographically isolated populations. In thisrespect it is difficulttopredict the suitability of therbcL region in the genus Dictyota.

Acknowledgements.- OlivierDe Clerckand Paul De Vos are indebtedto theFund for Scientific Research- Flanders(Belgium) (FWO) forpostdoctoral research and research director grants, respec- tively.Monique Gillis and Eric Coppejans are indebted to theFund for Scientific Research - Flanders (Belgium)(FWO) forresearch and personnel grants in projects G.0024.96 and G.0072.96, respectively, andto theUniversiteit Gent for financial and personnel grants in projectBOF 01106498.The authors alsowish to express their gratitude towards J.L. Olsen, W.T. Stam,and A.F. Petersfor providing help at severalstages of this research. The authorsare indebted to thereviewers, Victor Albert and Bruno de Reviers,for their constructive criticism.

References

Albert V.A., Mishler B.D. & Chase M.W. (1992) Character-stateweighting for restriction site data in phylogeneticrecon- struction,with an examplefrom chloroplast DNA. In SoltisP.S., SoltisD.E. & DoyleJ.J. (eds.) MolecularSystematics of Plants: 369-403.London, Chapman and Hall. BakkerF.T, OlsenJ.L. & Stam W.T. (1995) Evolutionof nuclear rDNA ITS sequencesin the Cladophora albida/sericea clade (Chlorophyta).J. Mol. Evol.40: 640-651. ChaseM.W. & HillsH.G. (1991) Silica gel:an idealmaterial for field preservation ofleaf samples for DNA studies.Taxon 40: 215-220.

33

This content downloaded on Thu, 10 Jan 2013 04:02:20 AM All use subject to JSTOR Terms and Conditions Syst.Geogr. Pl. 71 (2001)

Coleman A.W. & Mai J.C. (1997) RibosomalDNA ITS-I and ITS-2 sequencecomparisons as a tool forpredicting genetic relatedness.J. Mol. Evol.45: 168-177. DeBry R.W. & Slade N.A. (1985) Cladisticanalysis of restriction endonuclease cleavage maps within a maximum-likelihood framework.Syst. Zool. 34: 21-34. De Clerck0. (1999) A revisionof the genus Dictyota Lamouroux (Phaeophyta) in the Indian Ocean. Ph.D. thesis, Univeristeit Gent,unpublished. De Clerck O. & Coppejans E. (1999) Two newspecies of Dictyota (Dictyotales, Phaeophyta) from the Indo-Malayan region. Phycologia38: 184-194. DoyleJJ. & DoyleJ.L. (1987) A rapidDNA isolationprocedure for small quantities of fresh leaf tissue. Phyto. Bull. 19: 11-15. Farris J.S., AlbertV.A., Kallersjo M., LipscombD. & Kluge A.G. (1996) Parsimonyjackknifing outperforms neighbor- joining.Cladistics 12: 99-124. FelsensteinJ. (1985) Confidenceintervals on phylogenies:an approachusing the bootstrap. Evolution 39: 783-791. Fredericq S., HommersandM.H. & FreshwaterD.W. (1996) The molecularsystematics of someagar- and carrageenan- containingmarine red algae based on rbcLsequence analysis. Hydrobiologia 326/327: 125-135. FreshwaterD.W., FredericqS. & HommersandM.H. (1995) A molecularphylogeny of the Gelidiales (Rhodophyta) based on analysisof plastid rbcL nucleotide sequences. J. Phycol. 31: 616-632. HeyndrickxM., VauterinL., VandammeP., KerstersK. & De Vos P. (1996) Applicabilityofcombined amplified ribosomal DNA restrictionanalysis (ARDRA) patternsin bacterial phylogeny and . J. Microbiol. Meth. 26: 247-259. HommersandM.H., FredericqS., FreshwaterD.W. & HugheyJ. (1999) Recentdevelopments in thesystematics of the Gigartinaceae(Gigartinales, Rhodophyta) based on rbcLsequence analysis and morphological evidence. Phycol. Res. 47: 139- 151. Hirnig,I. & SchnetterR. (1988). Noteson Dictyotadichotoma, Dictyota menstrualis, Dictyota indica and Dictyota pulchella spec.nova (Phaeophyta). Phyton-Ann. Bot. Austria 28: 277-291 Hirnig I., SchnetterR., Prud'hommevan Reine W.F.,Coppejans E., Aschenbach-WegeK. & Over J.M. (1992a) The genusDictyota (Phaeophyceae) in theNorth Atlantic. I. A newgeneric concept and new species. Nov. Hedw. 54: 45-62. HiornigI., SchnetterR. & Prud'hommevan Reine W.F. (1992b) The genusDictyota (Phaeophyceae) in the North Atlantic. II. Keyto the species. Nov. Hedw. 54: 397-402. KuhselM. & KowallikK. (1985) The plastomeof a brownalga, . I. Physicalproperties and the Bam HI/Sal I/BglII cleavagemap. Pl. Mol. Biol. 4: 365-376. KuhselM. & KowallikK. (1987) The plastomeof a brownalga, Dictyota dichotoma. II. Locationof structural genes coding for ribosomalRNA's, the large subunit of ribulose- 1,5-biphosphate carboxylase/oxygenase andfor polypeptides ofphotosystems I and II. Mol. Gen. Genet.207: 361-368. Lee W.J.& King R.J. (1996) The molecularcharacteristics offive genera of (Phaeophyta) from Australia: based on DNA sequencesof Nuclear rDNA Internal Transcribed Spacer (ITS) andfrom 5.8S. Algae 11: 381-388. Mai J.C. & ColemanA.W. (1997) The internaltranscribed spacer 2 exhibitsa commonsecondary structure in greenalgae and floweringplants. J. Mol. Evol.44: 258-271. MiillerD.G., Parodi E.R, PetersA.F. (1998). Asterocladonlobatum gen. et sp.nov., a newbrown alga withstellate chloroplast arrangement,and its systematic position judged from nuclear rDNA sequences. Phycologia 37: 425-432. PetersA.F., Burkhardt E. (1998).Systematic position of the kelp endophyte Laminarionema elsebetiae (Ectocarpales sensu lato, Phaeophyceae)inferred from nuclear ribosomal DNA sequences.Phycologia 37: 114-120. Peters A.F., Van Oppen M.J.H., WienckeC., Stam W.T. & Olsen J.L. (1997) Phylogenyand historicalecology of the Desmarestiales(Phaeophyceae) support a southernhemisphere origin. J. Phycol. 33: 249-309. PhillipsJ.A. (1992) Taxonomyand reproduction in Australian species of Dilophus (Dictyotales, Phaeophyta). Aust. Syst. Bot. 5: 657-694. Rogers S.O., RehnerS., Bledsoe C., Mueller G.J. & AmmiratiJ.F. (1989) Extractionof DNA frombasidiomycetes for ribosomalDNA hybridisation.Can. J.Bot. 67: 1235-1243. RousseauF., Leclerc M.C., De ReviersB. (1997) Molecularphylogeny of European Fucales (Phaeophyceae) based on partial large-subunitrDNA sequence comparisons. Phycologia 36: 438-446. SaundersG.W. & DruehlL. (1992)Nucleotide sequences of the small-subunit ribosomal DNA genesfrom selected Laminariales (Phaeophyta):implications for kelp evolution. J. Phycol. 28: 544-549. SaundersG.W. & DruehlL. (1993) Revisionof the kelp family Alariaceae and the taxonomic affinities of LessoniopsisReinke (Laminariales,Phaeophyta). Hydrobiologia 206/261: 689-697. Saunders G.W. & KraftG.T. (1995) The phylogeneticaffinities of Notheiaanomala (Fucales, Phaeophyceae) as determined frompartial small-subunit rRNA gene sequences. Phycologia 34: 383-389. Stache-Crain B., Miiller D.G. & GoffL.J. (1997) Molecularsystematics of Ectocarpusand Kuckuckia(Ectocarpales, Phaeophyceae)inferred from phylogenetic analysis of nuclear- and plastid-encoded DNA sequences.J. Phvcol. 33: 152-168.

34

This content downloaded on Thu, 10 Jan 2013 04:02:20 AM All use subject to JSTOR Terms and Conditions De Clerck,De Vos, Gillis& Coppejans,Molecular systematics of Dictyota(Phaeophyta)

SwoffordD.L. (1999) PhylogeneticAnalysis Using Parsimony (*and OtherMethods). Version 4. Massachusetts,Sinauer Associates. Tan I.H. & Druehl L.D. (1993) Phylogenyof the northeast Pacific brown algal (Phaeophycean)orders as inferredfrom 18S rRNAgene sequences. Hydrobiologia 260/261: 699-704. TargettN.M. & ArnoldT.M. (1998) Predictingthe effects of brown algal phlorotannins on marineherbivores intropical and temperatureoceans. J. Phycol. 34: 195-205. Templeton A.R. (1983) Phylogeneticinference from restriction endonuclease cleavage site maps with particular reference to humansand apes. Evolution 37: 221-244. Van OppenMJ.H., Draisma S.G.A, Olsen J.L. & Stam W.T. (1995) Multipletrans-Arctic passages in the red alga Phycodrys rubens:evidence from nuclear rDNA ITS sequences.Marine Biology 115: 179-188. White T.J., Burns T., Lee S. & TaylorJ. (1990) Amplificationand directsequencing of fungalribosomal RNA genesfor phylogenetics.In Innis M.A., Gelfand D.H., SninskyJ.J. & WhiteT.J. (eds.) PCR protocols:315-322. New York. ZechmannF.W., Theriot E.C., ZimmerE.A. & ChapmanR.L. (1990) Phylogenyof the Ulvophyceae (Chlorophyta): cladistic analysisof nuclear-encoded rRNA sequence data. J. Phycol. 26: 700-710.

Manuscriptreceived October 2000; acceptedin revised version January 2001.

35

This content downloaded on Thu, 10 Jan 2013 04:02:20 AM All use subject to JSTOR Terms and Conditions