Preliminary study on molecular phylogeny of Saccoglossa and a compilation of their food organisms

2004·12.6 MB·English

by  Katharina Händeler

Bonnerzoologische Beiträge Band 55 (2006) Heft 3/4 Seiten 231-254 Bonn, November2007
Preliminary study on molecular phylogeny of Sacoglossa
and a compilation of their food organisms*
Katharina Händeler'' & Heike Wägele"-)
''Institut für Evolutionsbiologie und Ökologie, Rheinische Friedrich-Wilhelnis-Universität, Bonn, Germany
-'Zoologisches Forschimgsmtiscum Alexander Koenig, Bonn, Germany
*Paperpresented tothe 2nd Intemational Workshopon Opisthobranchia, ZFMK, Bonn, Germany, September20th to 22nd, 2006
Abstract. The tlrst molecularanalysis ofthe Sacoglossabased on the 16Sr DNAgene (partial), with 39 species and 59
specimensanalysed, ispresented.Asaturationofsubstitution isobserved, aswellasconflictinthedataconcerningcer-
tain taxa. Therefore the phylogenetic relationship presented here has to be considered as preliminary. Many results are
congruent with the analysis ofJensen (1996) based on morphological characters. Plakobranchacea are monophyletic.
Elysiellapusilla is nested within the genusElysia andtherefore synonymy ofElysiella with Elysia confirmed. Thepo-
sitionofPlakobranchiisremains unclear,butacloseraffinitytoElysia than to ThuridiHa seemsmost likely. Majordif-
ferences lie in theparaphyly ofthe Oxynoacea,but only few membersofthis taxon havebeen included. Acompilation
ofavailable data on food algae is presented and the data are discussed under the light ofthe phylogenetic relationship
presentedhere.
Keywords. Opisthobranchia, molecularsystematics, SplitsTree, Chlorophyta, evolution, 16S rDNA.
INTRODUCTION
1.
Sacoglossaisasmall groupofOpisthobranchiawithabout Sacoglossa is feeding nearly exclusively on siphonalean
250to 300describedspecies(Jensen 1997a).Animalsare orsiphonocladalean Chlorophyta. They penetrate the cell
mm
small and species largerthan 30 are rare. The mono- wall with theiruniseriateradula andsuckoutthecell sap.
phyly ofthis group is well supported (Jensen 1996) al- Noteworthy is the possibility ofseveral sacoglossans to
though their phylogenetic position relating to other incorporate the chloroplasts from their food intracellular
groups ofOpisthobranchia is unresolved (Dayrat et al. in their digestive gland and partially perform photosyn-
2001; Dayrat&Tillier2002; Grande etal. 2004; Von- thesis. This was first recorded by Kawaguti & Yamasu
nemann et al. 2005; Wägele & Klussmann-Kolb (1965) and most recently analysed by Evertsen et al.
2005). (2007) (for a review see Rumpho et al. 2000). For sever-
al sacoglossans reception ofpolysaccharides and amino
ThefirstandonlyphylogenyofSacoglossabasedon mor- acides from these chloroplasts is documented (Trench
phological datawaspresentedbyJensen (1996)applying 1973; Williams & Walker 1999 afterGreene & Musca-
Hennigian principles. Her analysis was perfomied on tine 1972). In addition, many members ofthe Sacoglos-
genus level, exceptforthegenusElysia with fourinclud- saareableto incoiporatesecondarymetabolitesfromtheir
ed species. One ofher results was a differentiation into foodandusetheseasachemicaldefense(Cimino& Ghis-
two major clades, the shelled Oxynoacea and the shell- ELiN 1998; CiMiNO et al. 1999; Gavagnin & Fontana
less Plakobranchacea. Thelattercomprisethemonophyle- 2000; Marín & Ros 2004; Cimino & Gavagnin 2006).
tic Plakobranchoidca and the monophyletic Limapon-
tioidea. Cylinclrobulla appeared as the most basal taxon Jensen (1997a: 313) was the first who discussed a possi-
within her analysis. Gosliner (1995) performed a mor- ble con'elation between the evolution ofmoiphological
phologically based analysis concentrating on the genus charactersofthe Sacoglossaandthepossibleevolution of
Thuridilla. An analysis of the family Plakobranchidae the food plants. She argued that more investigations are
based on a genetic marker (COI) was published lately needed: "Rigorous application ofcladistic methodology
(Bass & Karl 2006). No further phylogenetic analyses in coevolution analysis requires fully resolved phyloge-
have been performed up to now. nies ofthe hosts" as well as the 'associates"." hitention
232 Katharina Händelhr & Heike Wägele: Phyiogeny ofSacoglossa
ofthis studywasto fill thisgapby analysingthephyioge- 60''C followed by cooling down to 4°C. Reaction pro-
ny of Sacoglossa with molecular data and by including ductswerepurifiedbyadding40)il watercontainingDex-
many more species. New available data on food organ- tranblau, 5|il natriumacetate (3M, pH4.6) and I25|ilpure
isms in Sacoglossa are presented, which might help to ethanol, centrifuging for 15min, taking supernatant,
elaborate co-evolutionary processes. adding 500)il fresh 70% ethanol, centrifuging for 5min,
taking supernatant, andwere dried at 37°C for30-40min.
The fluorescently labeled products were size sorted and
MATERIALS AND METHODS
2. visualized using an ABI PrismTM 377 DNA Sequencer
by Applied Biosystems. The sequences were extracted
2.1. Biological material with Sequencing Analysis 3.0. The sequences ofthe re-
verse primerwere complemented with Factura 2.0.1 and
58 specimens including 38 species ofsacoglossans were sequences ofboth strands were compared and synchro-
usedforthephylogenetic reconstruction. Locality, dateof nised with Sequence Navigator.
collection and accession numbers are indicated in Table
1. According to the results ofJensen (1996) and unpub- The gene fragments were aligned by applying Muscle3.7
lished data of the senior author, Cylincirohiilla beaiiii (Edgar2004a; Edgar2004b)andtheresultingalignment
(Cylindrobullidae, Opisthobranchia) was chosen as out- wasverifiedby eye. Lengthofthealignmentwas478base
group species. Cylindrobulla and Sacoglossa share pairs.
synapomoi-phies(presenceofashell, adductormuscle,and
a lamellate gill (Jensen 1996)), but also lack synapomor-
phiesofthe Sacoglossa(e.g. ascus),and showplesiomor- 2.3 Phylogentic reconstruction
phic characters (e.g. triangular shaped teeth, large shell).
A Bayesian analysis was performed with MrBayes 3.1.2
(Huelsenbeck & RoNQuiST 2001; Huelsenbeck &
2.2. DNA extraction, amplification and sequencing BoLLBACK 2001). Maximum Parsimony analysis.Neigh-
bor-joining analysis and Maximum likelihood analysis
DNAwasextracted from alcohol-preserved specimensby have been perfonned with PAUP* 4.0b10. Modeltest 3.7
means ofNucleoSpin"-' Tissue-Kit by Machery-Nagel or (Posada & Buckley2004) selected asbest-fit model the
Dneasy-' Blood andTissue Kitby Quiagen,guidedbythe GTR+G+I-model (AIC). Therefore, neighbour-joining
enclosed protocol. Asegment ofthe 16Sr DNAgene was analysis was conducted on a GTR-model and Maximum
amplified with primer pairs, LR-J-12887 5'-GGA GCT Likelihood analysis and Bayesian analysis on a
CCG GTTTGAACTCAGATC-3' and LR-N-13398 5'- GTR+G+I-model. Saturation ofsubstitutions was tested
CGG CCG CCT GTT TAT CAAAAA CAT-3\ Amplifi- against patristic distances as well asagainstdistancesob-
cation reactions (50|.il) consisted of39.15^1 ddHiO, 5|.il tained with theGTR-model.Aposteriori statistical analy-
lOX PCRBufferwithoutMgCU (Fermentas),4nrMgCl2 sis could not be perfonned due to long calculation time
(25mM), 0.15pl Taq-Polymerase (5U/|al). 0.4|ul ofeach ofbootstrapping methods. For visualizing conflict in the
Primer (lOprnol/^l) and 0.5^1 DNA. The PCR was car- data set, which seemedobvious afterthe analyses, Splits-
ried out in the Gene Amp PCR System 9600 by Perkin Tree4(HusON& Bryant2006)wasapplied.ANeighbour-
Elmer'-' under following conditions: 95°C for 240s, fol- Net analysis was perfonned applying the K2P-model.
lowedby9Touch-down-cylcesof45sat94°C,45sat 56(-
1)°C, 90sat 72°C, followedby25 amplification-cyclesof
RESULTS
45s at 94°C, 45s at 48°C, 90s at 72°C and a final exten- 3.
sion at 72°C for lOmin followedbycoolingdown to4°C.
DNA
Amplicons were purified by means ofNucleoSpin"*' Ex- Saturation in substitutions ofthe partial 16Sr ob-
trat II by Machery-Nagel, guided by the enclosed proto- tainedby usingtheGTRmodel isshown in Figure 1. The
col.Themassoftheampliconswasestimatedbycompar- graph approachesaplateauindicatingasaturation ofsub-
ingethidiumbromide stainingintensity of5|.il ofeachpu- stitutions. When splittingthedatasetintosubgroups(fam-
rified reaction. ily orgenera), saturationwas lessevidentformostgroups,
but not e.g. the genus Elysia or TIniridilla. Substimtion
Cycle sequencing reactions (Cycle Sequencing Kit rateswerealso lesssaturated, when usingp-distances (not
BigDye'- Terminator vl.l by Applied Biosystems) were shown here).
cairiedoutintheGeneAmp PCR System 9600by Perkin
Elmer® under following conditions: 96°C for 120s fol-
lowed by 15 cycles of 10s at 96°C, 5s at 50°C, 150s at
—
Bonnerzoologische Beiträge 55 (2006) 233
3.1. Phylogenetic reconstruction Oxynoacea arc basal and paraphyletic. The families Juli-
idae and Oxynoidae again are not monophyletic. Similar
A maximum parsimony analysis produced 81 shortest as in the MP analysis, Tamanovalva is the sistertaxon to
topologies with a length of 1401 steps. The 50%-Major- thePlakobranchacea.ThepositionoíOxynocviridis is un-
ity-Ruletree is shown in Figure2. Within themonophyle- clear.
tic Plakobranchacea, the family Polybranchiidae and the
polybranchiid genus Cyerce are also monophyletic. The Two identical trees were found in the Maximum-likeli-
monophyletic genus Ercolauia appears as the sistertaxon hood analysis; the resulting tree is shown in Figure 3.
tothePolybranchiidae.The familiesPlakobranchidaeand Plakobranchacea is monophyletic, as well as the families
Limapontiidaearenotmonophyletic,thegeneraElysiaand Plakobranchidae and Polybranchiidae. Boselliidae is sis-
Thuridillaarealsoparaphyletic. Plakohrauchiisgroupsas tertaxon to the Plakobranchidae. The genus Thuridilla is
sistertaxon of several Elysia species. The shelled monophyletic. ElysiellapusillawithElysiabenettaeassis-
Oxynoaceaarebasal. Theyareparaphyletic as well as the tertaxon groups in the genus Elysia. rendering this genus
family Juliidae and Oxynoidae. The position ofOxyuoe paraphyletic. In this analysis Plakobranchus ocellatus is
viridis is unresolved. the sistertaxon to all £'/v¿'/(7-species (including Elysiella)
andmonophyletic ThuridillatothisPlakolvanchus/Elvsia
In the Neighbor-joining analysis (not presented here), clade. Limapontiidaeagain isparaphyletic. Contrai7 tothe
Plakobranchacea are monophyletic, as well as the family other analyses, Ercolauia is not monophyletic. The two
Polybranchiidaeandthegenus Thuridilla. Bosellia mimet- species represent the most basal Plakobranchacea. The
ica (Boselliidae) groups in the family Plakobranchidae, Oxynoacea and Oxynoidae are paraphyletic again, but
rendering this family paraphyletic. Limapontiidae is pa- contrary to all other analyses, the .luliidae with the gen-
raphyleticagain, sinceErcolauiaandStiligerdonotgroup era Julia and Taiuauovalva represent a monophyletic
together. The genus Elysia again comprisesElysiella and group. The position ofOxynoe viridis is unresolved.
Plakobranchus and is polyphyletic in this analysis.
Plakobranchiis is sistertaxontothesame£"/v.s/i/-complex The cladogram based on the analysis with MrBayes is
as in the maximum parsiomony analysis. Similar to the shown in Figure4. Plakobranchaceaismonophyletic, but
maximum parsimony analysis, Elysia henettae is the sis- resolution within that taxon is low. Elysia (without E.
tertaxontoElysiellapiisilla. MonophyleticErcolauia ap- benettae),Elysiella/Elysia bennettaecomplex andPlako-
pears as the most basal genus of the Plakobranchacea. branchusformamonophyletic clade,buttheirpositionis
p-distances
0.9
-|
0,8 -
0,7 -
0.6 -
0.5 -
0.4 -
0.3 -
0.2 - ^^'^''l-rirtÖl^^^^^^
—
0 , 1 1 1 I 1 1 1 1
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0.8 0.9
d-Distances GTR
Fig. 1. Saturationofsubstitutionofthe 16SrDNAgene (partial) used in thisanalysis. The straight line isthebisector. Distances
are calculated as patristic distances (y-axis) against d-distances calculated by applying the GTR model (x-axis).
Katharina Hände-ler & Heike Wägele: Phylogeny ofSacogiossa
Cyercespec. 2
Cyercespec. 1
Cyercenigricans
Cyerce antillensis Polybranchiidae
Cyerce edmundsi
Poiybranchiacf. orientalis
Mourgonaosumi
Ercolaniaspec. 1
Limapontiidae
Ercolaniaspec. 2
Boselliamimetica 1
Boselliamimetica 2 Boselliidae
Boselliamimetica 3
Placida dendrítica 1
Placida dandrñica 2 Limapontiidae
Placida dendrítica 3
Thurídilla carísoni 1
Thuridilla carísoni2
Thurídillakathae
Plakobranchidae
Thurídilla vatae
Thurídillaalbopustulosa
Thurídillahoffae
Thuridillagracilis
Thurídillaratna 1
Thurídillaratna 2
Thurídillabayerí Plakobranchidae
Thurídillalineolata
Thurídillahopei1
Thurídillahopei2
Elysia tímida 1
Elysia tímida2
Elysia tímida3
Elysiaobtusa
Elysia tomentosa 1
Elysia tomentosa2
Elysiamacnaei
Elysiaspec.5 Plakobranchidae
Elysiaspec.1
Elysiaamata 1
Elysiaomata2
Elysiaomata3
Elysiaamakusana
Plakobranchusocellatus1
Plakobranchusocellatus2
Elysiellapusilla 1
Elysiellapusilla2
Elysiellapusilla3 Plakobranchidae
Elysiellapusilla4
Elysiabenettae
Elysia virídis 1
Elysia virídis2 Plakobranchidae
Elysia virídis3
Tamanovalvalimax ]Juliidae
Julia exquisita ]Jiiliidae
Volvatella virídis ]Volvatellidae
Lobigervirídis 1
Lobigervirídis2 Oxynoidae
Lobigervirídis3
Cylindrobullabeauii
Oxynoe virídis ] Oxynoidae
2. Sacogiossa; 16Sr DNAgene: Maximum Parsimony analysis, 50% Majority Rule consensus tree.
Bonnerzoologische Beiträge 55 (2006) 235
unresolved. This complex and two Tliiiriclilla complexes Derbesia and Cladophora. The only two references for
(representing all Plakobranchidae) and Bosellia form the Plakobranchusocellatusarebased on observations in the
next higher unresolved monophyletic clade. laboratory (Jensiín 1980aafterSwitzcr-Dunlap 1975; Mi-
rose 2005 after Adachi 1991). This species feeds on
Polybranchiidaeandwithinthatfamilythegenus Cyerce, Chlorodesmis and Udotea. Interesting is the food organ-
is monophyletic. The family Limapontiidae and the ism ofPattyclaya, which, according to Jensen (1992), is
limapontiid genus Ercolauia are paraphyletic. The Caulerpa.
OxynoaceaandOxynoidaeareparaphyletic. TheJuliidae
arenotresolvedandthepositionofOxynoeviridisremains Polybranchiidae feedon speciesoftheorders Biyopsidales
again unclear. and Caulerpales (Bryopsidophyceae) except for Mour-
goiui Cymopolia (Jensen 98 Clark 994)andAceta-
( 1 1; 1
The analysis ofthe dataset with SplitsTree visualises the bularia (Hamatani 1994), Dasycladales, Dasyclado-
incongmencies in the dataset and the resulting incongru- phyceae). Hennaeidae also exhibitabroaderspectrum of
encies in the different cladograms. Figure 5 shows a net- food organisms: The genus Hei-maea is characterized by
workconnectingall terminal taxa. Paralleledges ofsame feeding on Rhodophyta (Graham 1955 after Pelseneer
length within the network visualise a split between two 1935; Taylor 1968; Taylor 1971; Kremer & Schmitz
groups. Membersofeachgroupsharesamecharacters(as- 1976; Kremer & Schmitz 1976 after Comet & Marche-
sumed apomoiphic nucleotides for that group). Fit value Marchad 1951; Jensen& Clark 1983; Jensen 1993a, b;
ofthe analysis is 95.06, indicating that actually all con- Williams&Walker 1999afterJensen 1983)exceptHer-
flictsarevisualisedinthefigure. No longbranchtaxacan maea vancouverensis consuming Istlunia nervosa, a di-
be identified. Very well supported is the split between atom (Williams& Gosliner 1973).Aplysiopsis feedson
Oxynoacea + Cylindrobulla and Plakobranchacea, indi- algae belonging to the orders Cladophorales (Jensen
catedbythelongedges.Alsowell supportedaresplitsbe- 1980aafterConor 1961;Jensen 1980aafterGreene 1970;
tween species (e.g. Placida dendritica. Elysia tiinida or Jensen 1993a; Trowbridge 1993;Jensen 1995; Williams
Elysia viridis) and the rest ofthe sequences. No split be- & Walker 1999 after Jensen 1983) or Ulvales (Jensen
tween outgroup taxon Cylindrobulla and Sacoglossa ex- 1980aafterConor 1961; Jensen 1980aafterGreene 1970;
ists(Fig. 5 andFig. 6,the lattershowingthe analysiswith Trowbridge 1993), Costasiella feedson algaebelonging
areduceddata set forbettervisualisation). But a split be- to the orders Bryopsidales (Clark & Busacca 1978;
tween Cylindrobulla + Tamanovalva Umax + Oxynoe Jensen 1980a, 1981; Clark et al. 1981; Clark 1984;
viridis against all other sequences can be obsei"ved (Fig. Jensen 1993a; Clark 1994; Williams & Walker 1999
5). Phylogenetic signal (shown as splits) is already low after Jensen 1980, 1981, 1983) and Caulerpales (Jensen
on genus level (e.g. Ercolania), or not existing (e.g. Cy- 1981; Williams&Walker 1999afterJensen 1980, 1981,
erce, Elysia or Thuridilla). A separate analysis of all 1983) (Bryopsidophyceae), Cladophorales (Williams &
Thuridilla species (Fig. 7) showsthe close affinities ofT. Walker 1999 after Jensen 1980, 1981, 1983) and
ratna and T. gracilis, whereas T. bayeriappears separate. Vaucheria (Hetcrokontophyta) (Jensen 1990a, b).
Limapontiidae show the most extraordinary food range.
Species ofthe genus Limapontia feed on Cladophorales
3.3. Food spectrum in Sacoglossa (Graham 1955 afterGascoigne 1952;Jensen 1975, 1980a
after Gascoigne 1956), but also Bryopsidales (Jensen
A compilation of available data on food organisms of 1975, 1980a after Gascoigne 1956), Ulvales (Graham
sacoglossans is presented in Table 2. Food organisms of 1955 after Gascoigne 1952) and the heterokontophyte
Cylindrobullaarenotknown.All membersofOxynoacae Vaucheria (Graham 1955 after Gascoigne 1954; Gas-
feed on species of Caulerpa. In contrast, species ofthe coigne & Sartory 1974; Jensen 1980a after Gascoigne
Plakobranchacea have a much wider food spectrum: 1956; Hartog 1959). Alderia also feeds on Vaucheria
(Evans 1953; Graham 1955 afterGascoigne 1954; Clark
Members ofthe family Boselliidae seem to be restricted 1975; Jensen 1980a after Hartog 1959; Krug & Manzi
totheorderCaulerpales(Bryopsidophyceae),mainlyfeed- 1999), but there is one reference for the cladophoralean
ing on Halimeda. Within Plakobranchidae, Elysia shows Rhizoclonium (Graham 1955 afterGascoigne 1954). The
the widest range on food organisms, containing Het- genus Placida mainly feeds on species ofBiyopsidales
erokontophyta (Vcnicheria. Padirni. Biddulphia) and an- (Brüel 1904; Graham 1955 after Gascoigne 1954;
giospermes(seagrasses).Elysia timida andElysiafilicau- Bleakney 1989 after Thompson 1976 and after Millen
da feed on Acetahularia (Dasycladales). Kreimer and 1980; Jensen 1980a; Jensen 1980a after Monselise &
Janke(1988) claimedthatchloroplastsofthealgalgenus Mienis 1977 and after Schmekel 1982, Jensen 1981,
Codium have been found in Elysia timida as well. Only 1990b; Marzo, di et al. 1993; Trowbridge 2004 after
forone species ofThuridilla, T. hopei, dataareavailable: Willan & Morton 1984, after Bum 1989 and afterTrow-
5 2
236 Katharina Hándeler & Heike Wäüele: Pliylogeny ofSacoglossa
bridge 1998b)and Codiwn (Brühl 1904afterHecht 1895; Ercolania prefers Cladophorales (Clark 1975; Jensen
Graham 1955 alter Gascoigne 1954; Clark 1975; 1980a; Jensen 1980aafterTrinchese 1872, afterRao 1937,
Trowbridge2004afterWiilan & Morton 1984,afterBum afterSchmekel 1968, afterRasmussen 1973 andafterUsu-
1989 and afterTrowbridge 1998b; Behrens 2004) (B17- ki 1977; Jensen 1981; Marín & Ros 1992 afterMarin&
opsidophyceae). Brüll (1904) referenced to Trínchese Ros 1988; Jensen 1993a; Marzo, di et al. 1993; Clark
1876 thatPlacida dendritica feeds on Ulva doubtingthis 1994; Jensen 1999; Grzymbowski et al. 2007), one ref-
fact at the same time. The observation that Placida den- erence for Caulerpa (Jensen 1980aafterEdmunds 1963)
driticaateanothersacoglossa(Marcus du Bois-Reymond is not confirmed again. Stiligerfuscovittatus feeds onred
1972)seemstobe aparticularcase. Placidakingstoniex- algae (Jensen 1980a afterLance 1962). Unusual feeding
tended its diet to Cladopliora (JENSEN 1981 ). The genus strategies in the family show CaIHopea oophaga. Olea
hansineeusisandStiligervesiciilosiis, theyconsumeeggs
of other Opisthobranchia (Jensen 1986; Lemche 1974;
Crane 1971; Jensen 1999 after Haefelfmger 1962).
éElysialimida 1
Elysiatímida 2 In Figure 8, the available data of higher food taxa are
—EEllyyssiisatoíbmitduasa3 mapped onto the cladogram ofthe Maximum Likelihood
V1— Elysiamacnaei analysis. Two references are not included: Kremer and
Elysiaspec.5
'— Elysiasped Janke (1988)detennined chloroplasts in Elysia tiniidaas
Elysiaornata 1 chloroplastsofCodiuni neitherdescribingtheoriginofthe
Elysiaornata 2
Elysiaornata 3 included REM-slide nor giving a reference. Elysia tími-
Elysiaamakusana da was never described feeding on Codiuni, a misdeter-
4 'rEEllyyssiiaattoommeennttoossaa 21 mination of the chloroplasts is therefore likely. Brüel
Elysiaviridis 3 (1904)alreadydoubtedTrinchese'sinformationthatPlaci-
EEllyyssiiaavviirriiddiiss 21 dadendritica feedson Ulva. Trinchese'sobservationwas
Elysietiapusilla 1 never confmned over the following years.
r|I Elysiellapusilla 2 Plakobranchldae
r-y Elysiellapusilla 3
j'—' E7l7y9siEelllyasipausbielnlaett4ae DISCUSSION
Plakobranchusocellatus 1 4.
Plakobranchusocellatus 2
Triundillagracilis
Thuridillaratna 1 4.1. Phylogeny ofSacoglossa
Thuridillaratna 2
Thuridillabayen
^ Thuridillalineolata TheOxynoacea. comprisingthe shelledsacoglossans, are
I Thuridillahopei 1 paraphyletic in all phylogenetic analyses. Applying
L—' TThhuurriiddiillllaahvoapteaie2 SplitsTrce, there exists no infomiation for a split Cylin-
—
I Thuridillaalbopustulosa drobulla against Sacoglossa. But according to the veiy
J I' TThhuurniddiillllaaccaarrllssoonnii 21 long edges, there is much information for a split Cylin-
L Thundillakathae drobulla + Oxynoacea against Plakobranchacea. Jensen
Thuridillahoffae
Boseltiamimetica 1 (1996) mentioned an eversible oral tube, a long, rod-
Boselliamimetica 2 shaped preradular tooth and a large female genital papil-
Boselliamimetica 3
Cyercespec. la as synapomorphies for the Oxynoacea. These charac-
1[Jj'———CCIy'yeCeryrcecorecanentiisglphlececna.snis1s Polybranchilclae taesrsouatrgernooutpparnedseinntcilnudCiynlginCdyrloibnudlrloab.iUdlsainagsapacretpohfaltahespiind-
'-I ' —CyeProcleyberdamnucnhdisaiof.orientalis gi'oupisnotapplicableforthisgene, sinceitshowsalready
- Mourgonaosumi a high substitution rate for the present data set. But the
Placidadendritica 1 positionofCylindrobidlaasmemberoftheOxynoaceahas
Placidadendritica 2 Limapontildae
Placidadendritica 3 to be reconsidered by usinga more conservative gene, as
-ErEcroctoalnainaiassppeecd. 13LLiimmaappoonnttiillddaaee 16S has proven to be. COI is also not apt for this inves-
Volvatellavindis 3Volvatellidae tigation, since it shows high substitution rates on nu-
- JuliaexquisTiatmaanovalvaUmax Juliidae cleotide level and is not informative on amino acid level
Lobigerviridis 1 (unpublished data ofthe authors). The unresolved posi-
-LLoobbiiggeerrvviirniddiiss 23 Oxynoidae tion oíOxynoeviridisatthebaseofthecladogramsispe-
Cylindrobullabeauii culiar. The genus nevergroups with Lobiger, both usual-
- Oxynoeviridis 3Oxynoidae lyunitedundertheOxynoidae.Acontaminationisunlike-
Fig. 3. Sacoglossa; 16Sr DNA gene: Maximum Likelihood ly due to high affinity with opisthobranch sequences.
tree (GTR+G+I-model,AIC). SplitsTree visualizes uniting positions with Cylindrobul-
25
Bonner zoologische Beiträge 55 (20U6) 237
la and Tamanovalva Umax, which might present homo- In all analyses, the PlakobranchaceasensuJensen (1996)
plastic characters. The Juliidae are monophyletic only in are monophyletic and this is visualized very well by the
the Maximum Likelihood analysis, but support is present longedges separatingthistaxon from Oxynoacca+ Cylin-
intheSplitsTreeanalysis.AccordingtoJensen (1996)the drobulla. But this is not the case for the higher ranking
bivalved shell and small, paired pharyngeal pouches are taxa Plakobranchoidea and Limapontioidea. Whereasthe
autapomorphiesofthis family. It is mostunlikely thatthe fomiertaxonwasrecognizedat least in themaximum like-
bivalved shell has evolved several times independently, lihoodand MrBayesanalyses,the Limapontioideawith the
as several phylogenetic analyses presented here would genera Placida and Ercolania (both members of the
suggest. Limapontiidae) never appeared monophyletic. This is of
special interesttaking intoconsideration theongoingdis-
cussiononthe validity ofPlacida, Ercolania and Siiligcr
asthreeseparategenera(e.g. Marcus du Bois-Reymond
1982; Jensen 1985). hi Jensen's (1996) analysis, the
Limapontiidaeandthegenus Costasiella (Costasiellidac)
o.isElysiatímida 3
100i Elysiatímida 2 formapartlyunresolvedgroup. Limapontiidae ischarac-
—EEtlyyssiiaaotíbmtiudsaa1 terized by two visceral ganglia instead ofthree or more
^M— Elysiamacnaei (Jensen 1996).
^
— Elysiaspec.
' Elysiasped
Elysiaornata 1 Contraiy to Jensen's analysis, the Polybranchiidae form
miElysiaornata 2
Elysiaornata 3 a monophyletic group with exactly the same branching
' Elysiaamakusana patternofthe six species (threegenera) in all analyses, al-
100'rEEllyyssiiaattoommeennttoossa21 Plakobranchidac though no support can be seen in the splits graph. Inter-
Í Elysiavindis 3 estingisthesistertaxarelationshipofthetwoAtlantic Cy-
Elysiavindis 1
Elysiaviridis 2 erce species, C. antillensis and C. edniiiiulsi. Marcus du
¡Elysiellapusilla 1 Bois-Reymond (1982) and Hamatani 1994) considered
Elysiellapusilla 2 (
- Elysiellapusilla 3 Cyerce edmundsi Thompson, 1977 as synonymous with
EEllyyssiieallbaenpuestitlalea 4 Mourgonagerniaineae Marcus & Marcus, 1970.Assign-
i.oa— Plakobranchusocellatus 1 ment of Cyerce edmundsi. as described by Thompson
Plakobranchusocellatus 2
O.izThundillagracilis (1977),tothegenus Cyercecontradictsdiagnostic features
'0Thuridillaratna 1 ofthisgenus: „Thecerataprovedtocontainabundantlob-
Thuridillaratna 2
Thuridillabayeri ules ofthe digestive gland" (Thompson 1977: 137). Cy-
npThuTrhiudrilildailhlaopleiineo1lata Plakobranchidae erce is seperated from the other Polybranchiidae by hav-
Thuridillahopei 2 ingnobranchesofthedigestivegland inthecerata(Mar-
^—9— Thuridillavatae cus DUBois-Reymond 1982 afterEliot 1910).According
' Thuridillaalbopustulosa
Thuridillacarlsoni 1 to the results ofthis study, Cyerce edmundsi is related to
ThurTihduinldliallkaatchaarelsoni 2 Plakobranchidae Cyerce and not Mourgona. Are-arrangement ofthe gen-
Thuridillahoffae eraand species is too early, more taxa ofLimapontioidea
BBoosseelllliiaammiimmeettiiccaa 21 withthe families Polybranchiidae, Hermaeidae,Costasiel-
Boselliamimetica 3 lidac, and Limapontiidae have to be added in future in-
1.00 Cyercespec.
I Cyercespec.1 vestigations to clarify validity ofthese taxa.
Cyercenigricans
Cyerceantillensis Polybranchiidae
Cyerceedmundsi Within the usually monophyletic Plakobranchoidea, two
MoPuolrygbornaancohsiuamcif.orientalis families have been considered, the monogeneric Boselli-
Oiflacidadendritica 1 idae and the species rich family Plakobranchidae. The
123 Placidadendritica 2 Limapontiidae
' Placidadendritica 3 genus Bosellia comprising only five species (Jensen
- Ercolaniaspec.5 ]Limaponliidae 1997a)wasalwaysataxon todebate. Pruvot-Fol(1954)
Ercolaniaspec.1 ]Ltmapontiidae
- Tamanovalvalimax ]Juliidae settledBoselliawithinthefamily Polybranchiidae. She in-
Volvatellaviridis ]Votvateliidae teipretedthedorsal apertures on the surface as areminis-
- Juliaexquisita 3Juliidae
JLobigerviridis 1 cence oflost cerata: „Des «orifices» semés sur la partie
LLoobbiiggeerrvviirriiddiiss 32 Oxynoidae dorsalefontpenserque I'animaldécritestincomplet, c est-
Oxynoeviridis 1Oxynoidae á-direqu'il aperdusespapilles."(Pruvot-Fol 1954: 180).
- Cylindrobullabeauii However, Pruvot-Fol has not examined a specimen of
Fig. 4. Sacoglossa; 16SrDNAgene: BayesianAnalysis,50%- BTorsíenlclhieasbeywrhietrtseenlfinb1u8t91f.olPloorwtedmatnhne (fi1r9s5t8d)esrcer-idpetsicornibeodf
Majority Rule consensus tree with posterior probabilities
the species and denied the existence ofpores on the dor-
(GTR+G+I-inodcl,AIC).
238 Katharina Händeler & Heike Wägele: Phyiogeny ofSacoglossa
Fig. 5. Sacoglossa; Neighbornetwork analysis applying SplitsTree4 to visualizeconflict in the whole sacoglossan dataset. Par-
allel edges with same length indicatea splitand visualizethedistancesbetween the twogroups involved. Note the longedgesbet-
ween (he Plakobrancliacca and the Oxynoacea + Cyliiulrohulla. Fit value: 95,85.
sal surface. Marcus du Bois-Rfymond (1980) and later Elysia, Elysiella, Plakobranclnis and Thwidilla. But the
Thompson and Jaklin (1988) positioned Boselliu in the otherincludedmemberofBosellia,B. marcusi, represents
family Plakobranchidae, but Marcus du Bois-Reymond the sistertaxon to Elysia timida. The authors considered
( 1982) transfened it in its own family Boselliidae based a wrong detemiination oftheir animal.
on the difference in number ofchromosomes in Bosellia
(n=7) and other sacoglossans (n=17), as well as the lack The family Plakobranchidae appears monophyletic only
ofparapodia. Jensen (1996) criticized that the numberof in the Maximum Likelihood analysis. This is in accor-
chromosomes is only known for the species Bosellia dance with Jensen's analysis (1996). She mentioned as
miiuetica and not for other species ofthis genus but de- mostimpoilantautapomorphiestheparapodiainwhichthe
cidedtoretain the family status until thephylogenctic po- digestive gland and the reproduction systemreach. In the
sition of the genus is solved. According to our results. SplitsTree analysis suppoil ofthe split Plakobranchidae
Bosellia seems to represent an own evolutionaiy line, in- versus rest of Sacoglossa is not found at all. Within the
dependent ofthe family Plakobranchidae (with the gen- Plakobranchidae, thegenus Thwidilla ismonophyletic in
era Elysia, Thwidilla and Plakobranclnis), but its posi- the NJ- and ML-analysis and branching pattern is identi-
tion varies within the Plakobranchacea. The result ofthe cal. The Bayesian analysis produced two ofthese mono-
Maximum Likelihood analysis places Bosellia as sister- phyletic Thwidilla complexesbut couldnotresolvetheir
taxon ofPlakobranchidae, similarto the results obtained positions in detail. Also in the Splits analysis, no split
by Jensen (1996). Bass and Karl (2006) used Bosellia ThwidillalvQsi ofSacoglossa is visible. Gosliner (1995:
miiuetica andBoselliainareiisiin theirphylogenctic analy- 45) mentioned followingapomoiphic features: „elongate
sis ofPlakobranchidae. Similar to our results, B. miiuet- rhinophores, single,thick,darklypigmentedampulla,non-
ica fonns the sister taxon to the further included genera muscular, unamied penis, bursa copulatrix with seperate
Bonnerzoologische Beiträge 55 (2006) 239
gonopore and orange or red eggs." A darkly pigmented traspecific variability ofthegene.Accordingto oursplits
ampulla could not be confirmed in the Mediterranean analysis, there are affiliations also between T. bayeriand
Thwidillahopei(ownunpublisheddata). It ispossible that T. lineolala (Fig. 7). Comparison of Gosliner's data
onlyIndopacifícspecieshaveapigmentedampulla, since (1995) with ourdata is hardly possible due to small over-
Goslinerdealtwith these species. Elysia viridis also lays lap ofincluded species. In agreement is the sistertaxa re-
„reddish yellow" eggs when feeding on Chaetoinorplia lationship of Thuridilla vatae and Thuridilla albopustu-
(Trowbridge & Todd 2001: 222). One particular clade losa.
was found in all analysis: (7^ bayeri (T. ratna 2 (7^ ratna
1 IT.gracilis))). The speciesareonlydistinguishedbydif- Jensen (1997a) estimated the numberoíPlakobranchus-
ferent colouration (Gosliner 1995). Jensen (1992: 278) speciesfrom 1 to 14. Inthisstudytwo specimensofPlako-
found „slight differences in the size of the pharyngeal branchusocellatiiswere used.Although all analysespub-
pouch." Brodie and Brodie (1990) considered T. bayeri lished yet (Gosliner 1995; Jensen 1996; Bass & Karl
as synonymwith T. ratna. Gosliner(1995: 9) mentioned 2006)assign Plakobrauchustoothermembersofthefam-
that"Internally, there is littledifferencebetween Tlniridil- ily Plakobranchidae, its position varies considerably. A
la bayeri [and] T/nn-idilla ratna", "... these data strongly unique character ofthe genus in relation to other mem-
suggest that the described differences between T. bayeri bers ofPlakobranchidae is thedorsal position ofthe anus
and T. ratna are simply variations within a single species (Jensen 1992).
that is extremelyvariable in its colouration." Concerning
Thuridillagracilis. GosLiNER(1995)complainedaboutthe In all analyses on Plakobranchidae (Jensen 1996; Bass
insuftlcient description of Risbec in 1928, hampering a & Karl 2006; this study), Elysiellapusillu is always lo-
correctassignment. Rudman (2000) statedthat"Contrary cated amongElysia species, E. benettae usually fonning
to Gosliner (1995), I consider T.hayeri to be a synonym itssistertaxon. JensenandWells(1990: 324)emphasized
ofThuridillagracilis."Ourstudy suppoilsthesynonymi- the validityofanowngenusElysiella: "thisspecies shows
sation ofT. ratna with T. gracilis Risbec, 1928. But syn- so many differences from other species ofElysia that it
onymisation ofThuridilla bayeriwith Thuridillagracilis definitely belongs to a different genus." But all phyloge-
seems preliminary, since we have no information on in- netic analyses indicate the invalidity ofthis genus, there-
0.01
Volvatella viridis
Julia exquisita
Lobiger viridis 3
^
Lobiger viridis 2 Tamanovalva max
Ii
Lobiger viridis 1
Cylindrobulla beauii
Oxynoe viridis
Fig. 6. Neighbor network analysis applying SplitsTree4 to visualize conflict in the reduced data set, comprising only Oxynoa-
ceaand the outgroup taxon Cylindrobulla beauii. Note thatthere is no split separating Cylindrobulla beauii fromthe Oxynoacea.
Similarlongedgesconnecting Volvatella,.Juliaand Tamanovalva indicatethelackofsignalofthetaxonJuliidae. Fitvalue: 98.53.
240 Katharina Hándellr & Heike Wacíple: Phylogeny ofSaeogiossa
foreElysiella¡ntsUla isnamed from liereon asElysiapiisil- Comparing our results with the only available phyloge-
la (Bergh, 1872). Several species, described under the netic hypothesis on Sacoglossa elaborated by Jensen
genus name Elysia (Elysia dubia Eliot, 1904, Elysia luil- (1996), the results ofthe Maximum Likelihood analysis
imedae Macnae, 1954, Elysia latipes Marcus & Marcus, come closestto herresults. Both analyses supportmono-
1960 and Elysia maciiaei Marcus, 1982) are mentioned phyly ofthe Plakobranchacea, Plakobranchoidea, Plako-
to be synonymous to Elysiapiisilla (seeJensen & Wells branchidae as well as the generaElysia (withElysiella as
1990). Marcus du Bois-Reymond (1980) distinguished a synonym) and Thuridilla. Plakobranchus and Bösellia
material identified as Elysia halimedae by Baba (1957) form distinct evolutionary lines. Juliidae are also mono-
and Burn ( 1972) fromthe original E. halimedae Macnae, phyletic. The two hypotheses contradict in the paraphyly
1954 from South Africa by the presence of papillate (this study) versus monophyly (Jensen 1996) of the
rhinophores in the latter. She re-named the material Oxynoacea, Oxynoidae and Limapontiidae. But it has to
niisidcntified as E. halimedae and established the name be emphasizedthatthesetaxaareunderrepresented inour
E. maciiaeiMarcus, 1980. Ourstudy includesa specimen studyandftitureanalyseshavetobeperformedwithmore
fromtheZoologische Staatssammlung München, collect- representativesofthesegroups. Furthennore, othergenes
ed in Sulawesi and identified by courtesy by Michael have to be used for analysing deepernodes ofSacoglos-
DNA
Schrödi asE. macnaei. This specimendoesnotgroupwith sa, since the 16Sr gene (this study), as well as the
Elysia pusilla indicating that E. macnaei is not synony- COI gene (unpublishedresults) showsaturation onhigh-
mous with E. pusilla. er level.
Thuridilla hoffae
0.07
Thuridilla carlsoni 1
Thuridilla carlsoni 2
Thuridilla albopustulosa
Thuridilla hopei 1
Thuridilla hopei 2
Thuridilla vatae
Thuridilla lineolata
Thuridilla ratna 2
Thuhdilla gracilis + Thuridilla ratna 1 Thuridilla bayeri
Fig. 7. Neighbornetwork analysis applying SplitsTiee4 to visualize similarity ofspecies in a reduced data set, comprisingonly
members ofthe genus Thuridilla. Note the one edge leading to the 2 specimens of Thuridilla ratna and the one specimen ofT.
gracilis. No conflictoccurs, indicatingthe synonymyofthese 2 species. Thuridilla hayeriis separate, neverthelessthelongedges
characterizing the split Thuridillagracilis/T. ralua /T. hayeri indicate its closer affmity to T. gracilis, than the shorteredge cha-
racterizing the split T. hayeri/T. lineolata. Fit value: 98,12.