The phylogenetic utility of the nuclear protein-coding gene EF-1a for resolving recent divergences

12 Pages·2010·0.98 MB·English

2010.TheJournalofArachnology38:9–20
The phylogenetic utility of the nuclear protein-coding gene EF-1a for resolving recent divergences in
Opiliones, emphasizing intron evolution
Marshal Hedin, Shahan Derkarabetian, Maureen McCormack and Casey Richart: Department of Biology, San Diego
State University, San Diego, California 92182-4614, USA. E-mail: [email protected]
Jeffrey W. Shultz: Department of Entomology, University of Maryland, College Park, Maryland 20742-4454, USA
Abstract. Our focus was to design harvestmen-specific PCR primers to target both introns and exons of the nuclear
protein-codinggeneElongationFactor-1alpha(EF-1a).Wetestedthisprimersetontengenerarepresentingallprimary
lineagesofOpiliones,withsetsofclosephylogeneticrelatives(i.e.,setsofseveralcongeners)includedtospecificallyassess
utility at shallow phylogenetic levels. Our research also included the collection of parallel mitochondrial protein-coding
DNAsequencedatasetsforthecongenericsetstocomparerelativeratesofevolutionandgenetreecongruenceforEF-1a
versusmitochondrialdata.Theharvestmenprimersresultedinsuccessfulamplificationfornineoftentestedgenera.Exon
sequences for these nine genera appear orthologous to previously-reported EF-1a Opiliones sequences, which were
generatedusingRT-PCRmethods.Newly-generatedexonsequencesareinterruptedbythreeseparatespliceosomalintrons;
two introns are restricted to one or two genera, but a third intron is conserved in position across all surveyed genera.
PhylogeneticanalysesofEF-1anucleotidedataforcongenericsetsresultingenetreesthataregenerallycongruentwith
mitochondrialgenetrees,withEF-1aphylogeneticsignalcomingfrombothintronandexonsites,andresolvingapparently
recentdivergences(e.g.,asrecentasonemillionyearsago).Overall,thecombinationofgeneorthology,conservedintron
position, and gene tree congruence at shallow levels suggest that this gene region will prove generally useful for both
phylogeographicandspecies-levelphylogeneticanalysesinOpiliones,complementingalready-documentedutilityathigher
taxonomiclevels.
Keywords: Molecularsystematics,phylogeography,genetree,orthology,elongationfactor-1alpha
Both phylogenetics and taxonomy obviously have been three novel nuclear genes (H3, H4, U2 snRNA) to address
impactedbyadvancesinmolecularbiology,andmostmodern relationships in sandokanid laniatoreans. Although the focus
systematic analyses published today include some molecular was at slightly higher taxonomic levels, several congeneric
component.InthearachnidorderOpiliones,severalmolecular comparisonsweremade.Again,despitetheclearutilityinthe
systematicstudieshavebeenpublishedthatconsiderrelatively multigenicperspectiveusedinthisstudy,thesegenefragments
ancient phylogenetic divergences (e.g., Giribet et al. 1999, individually are small (330, 160, 130 basepairs (bp), respec-
2002; Shultz & Regier 2001). Fewer studies have considered tively), and our examination of these data suggests minimal
molecular phylogenetic divergence within a species, or variation among congeneric species.
between closely related species, and those published have Here we assess the phylogenetic utility of the nuclear
relied mostly on mitochondrial DNA sequence data. While protein-coding gene Elongation Factor-1 alpha (EF-1a) for
informative, a mitochondrial-only perspective has limitations resolving relatively shallow opilion phylogenetic divergences.
(Ballard & Whitlock 2004; Rubinoff & Holland 2005), and is In metazoans, EF-1a is a core gene involved in protein
quickly becoming obsolete in a molecular systematics world synthesis – this fact facilitates use in molecular phylogenetics
dominated by multigenic datasets. (e.g., the gene is expected to be found in all taxa, have
Ahandfulofnucleargeneshavebeenusedtoaddressrecent conserved function, etc). This gene is commonly used in
divergences within Opiliones. Those utilized offer a mix of hexapodsystematics(summarizedinCaterinoetal.2000),and
pros and cons. Thomas & Hedin (2008) used ribosomal ITS has also seen limited use in arachnids (Shultz & Regier 2001,
sequences at the intraspecific level in Fumontana deprehendor 2009; Hedin & Maddison 2001). Given our interest in
Shear (Laniatores), and although well-behaved in this case, resolving recent divergences, we expect most phylogenetic
this gene region has been shown to be problematic in other information to come from either silent substitutions at third
arthropodtaxabecauseofalackofconcertedevolution(e.g., codon exon positions (e.g., Cho et al. 1995; Reed & Sperling
Harris & Crandall 2000; Bower et al. 2009). Ribosomal 28S 1999), or from introns (e.g., Danforth et al. 1999; Hedin &
sequences have been used in various harvestmen taxa, but in Maddison 2001). Here we present the results of a commonly
our experience these data show limited variability at the employed strategy: to develop PCR primers that reside in
shallowest levels (e.g., see Hedin & Thomas 2010). Shultz & relatively conserved exons, but span variable introns that are
Regier (2009) used reverse transcriptase polymerase chain conservedinposition(seePalumbi1996).Wehavediscovered
reaction(RT-PCR)methodstogenerateexonnucleotidedata several EF-1a introns in Opiliones, some of which are
for the nuclear protein-coding genes EF-1a and POL II, and universally conserved in position, but variable at shallow
applied these data to species-level relationships in the genus phylogenetic levels. The primers developed permit consistent
Caddo. While these data are informative at shallow levels, and robust amplification of a fast-evolving gene region using
gatheringthedatarequirestheextracostandexpertiseneeded standard PCR methods, allowing potentially rapid collection
to conduct RT-PCR. Finally, Sharma & Giribet (2009) used of population-level samples (e.g., many individuals from
9
10 THEJOURNALOFARACHNOLOGY
Table 1.—PCR primers. Primers markedwithanasterisk weredesignedafter sequence collection using moregeneral primers. Therelative
position(POS)ofprimers(59end)isshownbasedonalignmentwithexonsequencesofShultz&Regier(2001).Ambiguitycodesarestandard.
POS
ForwardPrimers
EF1-OP1 59-CGTGGTATYACCATYGATATCAC-39 28
EF1-OP2 59-GATTTCATCAARAACATGATYAC-39 112
EF1-OP2SCLER 59-GATTTCATCAAGAACATGATTAC-39 112
*EF1-OP2ASAB 59-GCTGTGCTTATTGTTGCTGCYGG-39 157
*EF1-OP2BSAB 59-GGTACTGGTGAGTTTGAAGCTGG-39 178
EF1-OP3 59-TTTGARGAAATCCARAARGAAGT-39 322
EF1-OP3PHAL 59-TTTGAAGAAATCCAAAAGGAAGT-39 322
EF1-OP3SCLER 59-TTTGAGGAAATCCAGAAGGAAGT-39 322
EF1-OP4 59-TACATYAAGAAGATTGGTTA-39 352
EF1-OP4SCLER 59-TACATCAAGAAGATCGGTTA-39 352
*EF1-OP5LEIO 59-GGAGATAACATGTTGGAACAAAG-39 415
EF1-OP5PHAL 59-AACATGTTGGAACAAAGTACCCA-39 421
EF1-OP5LAN 59-AACATGYTGGAAGCTTCCTCC-39 421
EF1-OP5ISCH 59-AACATGTTGGARGCCAGYGC-39 421
EF1-OP6PHAL 59-CATCACCACTGAAGTTAAATCTG-39 672
ReversePrimers
*EF1-OP5LEIORC 59-CTTTGTTCCAACATGTTATCTCC-39 437
EF1-OPRC1PHAL 59-CAGATTTAACTTCAGTGGTGATG-39 692
EF1-OPRC1SCLER 59-CGGACTTGACCTCAGTGGTGATG -39 694
EF1-OPRC2 59-GANACGTTCTTNACRTTGAA-39 767
*EF1-OPRC2LEIO 59-GAAACGTTCTTAACATTGAA-39 767
EF1-OPRC2PHAL 59-ACGGAAACGTTCTTAACGTTGAA-39 770
EF1-OPRC3PHAL 59-ATGACCTGGGCRGTGAATTCTTC-39 854
*EF1-OPRC3LEIO 59-ATAACCTGGGCAGTAAATTCTTC-39 854
EF1-OPRC3SCLER 59-ATGACCTGAGCCGTGAACTCTTC-39 854
EF1-OPRC4 59-GAACTTGCANGCAATGTGAGC-39 935
*EF1-OPRC4LEIO 59-GAACTTGCAAGCAATGTGAGC-39 935
EF1-OPRC5PHAL 59-GGTTGTCTTCCAATTTCTTGCC-39 992
multiple populations from many species). We present this freezer.AllspecimensarecurrentlyhousedintheArthropods
primersetasaresourcetobeexplored,modified,andutilized Genomics Collection at San Diego State University. We
by other opilion systematists. extracted genomic DNA from leg tissues using the Qiagen
DNeasy Kit; these extracts were either suspended in 200 mi-
METHODS
croliters (ml) of Qiagen AE buffer, or dried down using a
Primerdesign.—Shultz&Regier(2001)generatednearlyfull- Speedvac and resuspended in a smaller volume of AE buffer.
lengthEF-1aexonnucleotidesequencesusingRT-PCRmethods Testing primer utility.—For some samples (Siro, Caddo,
foracomprehensivesampleofOpiliones.Thesesequenceswere Ortholasma,DendrolasmaandBishopella),weassessedprimer
downloaded from GenBank, then compiled and translated in utility in a preliminary manner. Here we conducted PCR
MacClade 4.08 (Maddison & Maddison 2003). We designed experimentstotestwhetherourprimersresultedinamplicons
primers in conserved exon regions that might also amplify of expected size, and if so, we directly sequenced these
introns, using known intron positions from spiders and amplicons. These sequences inform us about primer success,
hexapods as a guide (Table 1, Figure 1). Within conserved intron position, and relative intron size, but do not address
regions,wemanually(i.e.,withoutsoftware)designed20–25bp variability of the nucleotide data within a taxon (see below).
oligonucleotides in regions with approximately equal base WeconductedPCRexperimentsusingvariousforward/reverse
composition, and GC-rich 39 nucleotides corresponding to primer combinations (Table 2). All PCR reactions included
strictly conserved second codon positions. Most primers were 0.8 ml of genomic DNA with 0.08 ml Ex Taq polymerase
designed to work for all Opiliones, but in some cases we also (Takara Bio Inc.), and 2.5 ml each of Takara dNTP mix,
designedclade-specificprimers(e.g.,EF1-OP3PHALisaprimer Takara buffer (with Mg2+) and primers (at 2–4 mM concen-
that best matches the phalangioid sequences from Shultz & trations).Weusedtissueculturewater(Sigma-AldrichCo.)to
Regier2001,etc.).Wedesignedadditionalinternal,clade-specific bring reaction volumes to 25 ml. Experiments were conducted
primersafterinitialsequencingwithourmoregeneralprimers. on two different PCR machines, including a MJ Research
Samples.—We tested primer utility on ten genera (Table 2) PTC-100 (95u C, 2 min; 353 (95u C, 30 s; 53u C, 45 s; 72u C,
thatrepresentallprimaryhigher-levelcladeswithinOpiliones 45 s);72uC,10 min)andanEppendorfMastercyclerPersonal
(Fig. 2). Specimens were either preserved directly in 100% (94u C, 3 min; 353 (94u C, 45 s; 45u C, 45 s; 72u C, 1 min
EtOHinthefield,keptcold,andlatertransferredtoa280uC 15 s); 72u C, 5 min). Amplification products were visualized
freezer (or 220uC freezer inthe case of Caddo), or preserved on agarose gels, purified via polyethylene glycol (PEG)
directly in 100% EtOH in the lab and stored in a 280u C precipitation, and sequenced directly at the SDSU Micro-
HEDINETAL.—EF-1aUTILITYINOPILIONES 11
Figure 1.—Upper: Relative primer positions (see also Table1). Spider intron positions from Hedin & Maddison (2001), hexapod intron
positionsfromBrady&Danforth(2004).Smallcirclesdesignate100-bpintervals.Lower:Ampliconsizesandrelativeintronpositionsforopilion
sample. Small hashes above amplicon boxes for Ortholasma and Sabacon indicate alternative 59 primers used for some templates (see also
Table2).
chemical Core Facility (http://www.sci.sdsu.edu/dnacore/ Leiobunum,Sabacon,Sclerobunus/Cyptobunus,andAcuclavella.
sdsu_dnacore.html)onanABI Prism3100capillarymachine. Forthese same taxon setswealso collected parallel mitochon-
Both strands were determined for most templates using drial sequence datasets, allowing us to gauge the relative
PCR primers in sequencing reactions. For some Leiobunum variability of nuclear versus mitochondrial gene data, and ask
templates,weusedadditionalinternalprimersforsequencing. if independent phylogenetic analyses resulted in generally
We assembled and edited sequence contigs using Sequencher congruent results. For Sabacon, Sclerobunus/Cyptobunus, and
version4.5.Forsometaxa(seeResults),AT-richregions(and Acuclavella, we collected partial cytochrome oxidase I (CO1)
possible allelic length heterozygosity) within introns caused mitochondrial sequences, following standard techniques as in
prematurestopsinsomesequencingreads;sequencesforthese Thomas&Hedin(2008).ForwardprimersincludedC1-N-1510
templates were thus determined using single reads from or C1-N-1718S, combined with C1-J-2568 or C1-J-2776S (see
opposite sides of the intron. Also, for Acuclavella we were Hedin & Thomas 2010). For Leiobunum, we collected nearly
able to generate high-quality 780-bp sequence reads for the full-length mitochondrial ND1 sequences using the custom
entire exon plus intron amplicon, and thus used only a single primers LR-N-12945LEI (59- TGACCTCGATGTTGAAT-
sequence read for several of these templates. Sites with two TAA -39) and CB-J-11638LEI (59 - CCTWATAAACTAAT-
peaksofequalintensityonchromatogramswereinterpretedas CATTTAGC-39).
representing nucleotide heterozygosity, and scored as such in We conducted phylogenetic analyses at two different
data matrices using standard ambiguity codes. hierarchical levels. First, we conducted an exon-only analysis
To assess phylogenetic utility at shallow levels more that included the Shultz & Regier (2001) exon data, plus
rigorously, we collected EF-1a sequences for sets of close newly-generated exon-only data (i.e., introns removed). This
phylogenetic relatives (both within and between species) for exon matrix was aligned manually, and subject to a heuristic
12 THEJOURNALOFARACHNOLOGY
Table 2.—Taxonsample,includingvoucherandcollectionlocationinformation,mitochondrialsample,andsuccessfulEF-1aprimers.
Voucher
Taxon no. Collectionlocation(Nlat.,Wlong.) mtDNA EF-1aprimers
Sironidae
Sirocf.kamiakensis OP2350 Idaho:IdahoCo.,FS311,SRoute14(45.6853, no OP3,OPRC4LEI
2115.5427)
Caddoidea:Caddidae
Caddoagilis OP2565 NewHampshire:CheshireCo.,PisgahStatePark no failed
(42.862,272.428)
Phalangioidea:Sclerosomatidae
Leiobunumaldrichi OP829 Mississippi:TishomingoCo.,TishomingoStatePark yes OP2,OPRC4LEI
(34.6054,288.1927)
L.aldrichi OP1069 Missouri:CalhounCo.,Marshall(42.3012, yes OP2,OP5LEIORC,OPRC4LEI
284.9674)
L.vittatum OP835 Tennessee:CumberlandCo.,CumberlandMtnSP yes OP2,OP4,OP5LEIORC,
(35.9013,284.9958) OPRC4LEI
L.speciosum OP1405 Tennessee:DavidsonCo.,westNashville(36.1227, yes OP2,OP4,OP5LEIORC,
286.9061) OPRC4LEI
L.calcar OP1394 NorthCarolina:ClayCo.,BigTuniCreek(35.1025, yes OP2,OP4,OP5LEIORC,
283.7007) OPRC4LEI
L.calcar OP814 Tennessee:CockeCo.,roadtoCosbyCG(35.7633, yes OP2,OPRC4LEI
283.2115)
L.undescribedspecies OP1383 Virginia:GraysonCo.,GraysonHighlandsSP yes OP2,OPRC4LEI
(36.6247,281.5013)
L.serratipalpe OP1080 Maryland:MontgomeryCo.,Unity,TusculumFarm yes OP2,OPRC4LEI
(39.2427,277.0872)
Troguloidea:Nemastomatidae
Ortholasmarugosum OP807 California:SanBenitoCo.,NEofPinacatePeak,off no OP2BSAB,OP3,OPRC4LEI
Hwy101(36.8602,2121.6126)
Ortholasmarugosum OP808 California:SanMateoCo.,SanBrunoMtnSP no OP3,OPRC4LEI
(37.6943,2122.4529)
Dendrolasmamirabile OP1000 Oregon:CurryCo.,Hwy33,EofGoldBeach no OP3,OPRC4LEI
(42.5465,2124.1273)
Ischryopsalidoidea:Ceratolasmatidae
Acuclavellacf.quattuor OP2233 Idaho:IdahoCo.,EagleMtTrailhead,Ssideof yes OPRC4only
LochsaRiver(46.4292,2115.1335)
A.merickeli OP2251 Idaho:IdahoCo.,alongheadwatersofRedRiver, yes OPRC4only
RedRiverRd(45.7853,2115.2026)
A.quattuor OP2255 Idaho:IdahoCo.,TributaryofCrookedCreek,FS yes OPRC4only
Rd222(45.5791,2115.4431)
A.cosmetoides OP2278 Idaho:IdahoCo.,CanyonCreekTrailhead,US12 yes OPRC4only
(46.2101,2115.5442)
A.cf.merickeli OP2345 Washington:JeffersonCo.,CedarCreek,EofUS yes OP2BSAB,OPRC4
(10147.7105,2124.4095)
A.cf.merickeli OP2347 Washington:LewisCo.,tributaryofIronCreek,FS yes OP2BSAB,OPRC4
Rd25(46.4033,2121.9902)
Ischryopsalidoidea:Sabaconidae
Sabaconsimoni OP2522 Italy:Piemont,Prov.Cuneo,nearMonesidiTriora yes OP2BSAB,OPRC4
(44.07381,7.75028)
S.cavicolens OP721 Tennessee:AndersonCo.,nearNorrisDamCave yes OP2BSAB,OPRC4
(36.221,284.09)
S.cavicolens OP657 Virginia:LeeCo.,CaveSpringRec.Area(36.803, yes OP2BSAB,OPRC4
282.92)
S.cf.cavicolens OP1283 NorthCarolina:MaconCo.,BullpenBridge(35.015, yes OP2BSAB,OPRC4
83.126)
S.cf.cavicolens OP691 NorthCarolina,HaywoodCo.,belowHeboMtn. yes OP2BSAB,OPRC4
(35.686,282.906)
S.cf.cavicolens OP1290 NorthCarolina:TransylvaniaCo.,LookingGlass yes OP2BSAB,OPRC4
Creek(35.297,282.767)
S.cf.cavicolens OP660 Virginia:ScottCo.,Hwy23/58,nearWeberCity no OP2,OPRC4
(36.6341,282.5604)
HEDINETAL.—EF-1aUTILITYINOPILIONES 13
Table 2.—Continued.
Voucher
Taxon no. Collectionlocation(Nlat.,Wlong.) mtDNA EF-1aprimers
S.cf.cavicolens OP1535 NorthCarolina:AveryCo.,HensonCreek,Nof no OP2,OPRC4
Ingalls(36.0374,282.0420)
Travunioidea:Sclerobuninae
Sclerobunus OP1056 Oregon:ClatsopCounty,EcolaSP(45.9221, yes OP3,OP4SCLER
nondimorphicus 2123.9767)
Cyptobunuscavicolens OP2143 Montana:JeffersonCounty,LewisandClark yes OPRC3SCLER;OPRC4
Caverns(45.8386,2111.8668)
Sclerobunusrobustus OP885 NewMexico:SandovalCounty,JemezMtns yes OP4SCLER,OPRC3SCLER
robustus (35.8384,2106.4044)
Sclerobunusr.robustus OP1149 Colorado:DoloresCounty,alongDoloresRiver yes OP4SCLER,OPRC3SCLER
(37.7679,2107.9871)
Sclerobunusr.robustus OP1122 Colorado:GunnisonCounty,SofGothic(38.9397, yes OP3,OPRC4
2106.4072)
Sclerobunusr.robustus OP1164 Colorado:CusterCounty,WetMtns(38.1335, yes OP4SCLER,OPRC3SCLER
21105.1791)
Grassatores:Phalangodidae
Bishopellalaciniosa OP320 Tennessee:JohnsonCo.,BackboneRockRecArea,S no OP3,OPRC4LEI
Damascus(36.5948,281.8173)
Bishopellalaciniosa OP108 NorthCarolina:JacksonCo.,UpperFalls, no OP3,OPRC4LEI
WhitewaterRiver(35.0336,283.0141)
Figure 2.—Opilionesphylogenywithtaxonsample.GeneraltreestructurefollowsGiribet&Kury(2007,fig.3.4).Theuncertainphylogenetic
placementofDyspnoiisindicatedbyadashedline,representingasistercladerelationshiptoeitherEupnoi(followingShultz&Regier2001),or
toLaniatores(e.g.,Giribetetal.1999).
14 THEJOURNALOFARACHNOLOGY
parsimony search in PAUP v4.06 (Swofford 2002), using vectorbase.org/, gene ISCW020299) indicates that intron I is
stepwisetaxonadditionfrom1000randomreplicateswithtree shared in position between Ixodes and Sabacon. Intron I is
bisection-reconnection (TBR) branch swapping. For the relatively small in Sabacon (less than 100 bp), and includes 59
congeneric sets, we conducted exhaustive parsimony searches (GT) and 39 (AG) splice site signal sequences (see Fig. 4)
on both nuclear and mitochondrial matrices, conservatively consistent withproposedmetazoanconsensussignal sequences
treating EF-1a intron gaps as missing. Branch support was (Senapathyetal1990;Mountetal.1992).
assessed using nonparametric bootstrap analyses (Felsenstein Intron II was found in all Leiobunum and one Ortholasma
1985), comprising 1000 pseudoreplicates of a branch-and- sample(OP807,amplifiedwithOP2BSAB;OrtholasmaOP808
bound parsimony search for each congeneric set matrix. was amplified using downstream primers). These taxa are not
phylogeneticrelatives,anddifferconsiderablyattheexonlevel
RESULTS
(see Fig. 3). We did not detect this intron in Acuclavella and
Data availability, primer utility.—All sequences have been some Sabacon, despite using primers that spanned this region
deposited to GenBank (accession numbers GQ870643– (Fig. 1).Thissomewhatrandomphylogeneticpatternofintron
GQ870668; GQ872152–GQ872185). Both intron and exon presence/absence is most likely the result of differential intron
alignmentsareavailableatwww.treebase.org(studyaccession loss (a similar argument applies for intron I), although we
number S2469). cannotruleoutindependentintrongains(e.g.,seeRoy&Penny
Using various primer combinations (Fig. 1, Table 2), we 2006). Intron II varies in size (, 100–200 bp), and to infer
were able to amplify and generate EF-1a sequences for all canonical 59 (GT) and 39 (AG) splice site signal sequences in
surveyed genera, except for the caddoid Caddo agilis. We Ortholasmarequiresatwoaminoaciddeletion(seeFig. 4).We
attempted eight different primer combinations for a single returntothisinferenceintheDiscussion.
extraction from Caddo, using primers that worked well for Finally,intronIIIisfoundinallopilionsamplesthatwehave
othertaxa.Wesuspectthatthefailureofthissamplereflectsa sequenced. This intron is shared in position with spiders
differenceinsamplepreservation,asnotedabove.Overall,the (Habronattus, see Hedin & Maddison 2001), but not with
most successful primers included the forward primers EF1- Ixodes.IntronIIIvariesconsiderablyinsizeamongharvestmen
OP2, EF1-OP2BSAB, and EF1-OP3SCLER, combined with taxa,rangingfrom60to. 500basepairsinlength;allinclude
the reverse primer EF1-OPRC4LEI. The most 59 (EF1-OP1) canonical 59 (GT) and 39 (AG) splice site signal sequences
and 39 (EF1-OPRC5PHAL) of our designedprimers failed in (Fig. 4).
all PCR experiments. We generated high-quality intron I sequences for the taxa
Features of the data.—Newly generated sequences were reportedhere,butinotherSabacontaxaalong(6-bp)poly-T
manually aligned withthe exon-only nucleotide data of Shultz region proved difficult to sequence through (data not
&Regier(2001)toidentifythereadingframe,introninsertions, reported). This provided our motivation to design Sabacon-
and intron/exon boundaries. As a preliminary assessment of specificprimers(OP2BSAB)downstreamofintronI.IntronII
sequenceorthology,wefirstremovedallintrons.Theremaining includes at least two long poly-T regions, but we had little
exon data translate toexpected amino acids,and witha single difficultly in sequencing through this intron for several
exception (see below) can be aligned without gaps. A strict LeiobunumandasingleOrtholasma.Oursamplesizeislargest
consensus tree resulting from parsimony analysis of the exon- for intron III, where there is considerable variation in intron
only data recovers expected phylogenetic placements for the size (see above), nucleotide composition, and sequencing
newly-generatedEF-1asequences(Fig. 3),suggestingorthology difficulty. The short (Fig. 4) Acuclavella introns were not
of the genecopies used in this study. Ours is the first study to difficult to sequence through, although these possess both
include the genus Acuclavella in a higher-level molecular poly-T and poly-A regions. Also, despite the long Siro intron
analysis; we recover Acuclavella as sister to Ceratolasma, III, and several simple-sequence regions, sequence reads were
consistentwiththehypothesisofShear(1986,fig.1). clean. Other taxa presented mixed success, even within
Patterns of exon variability across divergent opilion taxa congeneric sets, where some samples resulted in clean bi-
(e.g., among representatives of primary clades) have been directional reads, whereas other reactions resulted in clean
addressedpreviously(Shultz&Regier(2001).Patternsofexon reads that failed abruptly at poly-A/T regions. The sequence
evolutionwithincongenericsetsaresummarizedinTable 3.As contigs for these latter samples comprise single strand reads
expected, almost all exon variation occurs at third codon from opposite directions, with minimal terminal overlap.
positions,withatotalof116,14,and7variablesitesatthird, Regions of terminal overlap that proved ambiguous to score
first, and second positions, respectively. Heterozygosity at wereexcludedfromthephylogeneticanalysesreportedbelow.
exon positions is minimal, with less than 10 total sites scored Comparative phylogenetic utility.—The protein-coding mi-
as heterozygous for all newly-generated sequences. tochondrialdata(CO1,ND1)translatetoaminoacidswithout
We discovered three different introns in the harvestmen stop codons, and can be aligned without gap insertions. We
sample,herenamedintronsI–III.Therelativepositionofthese comparedthesemitochondrialdatatointronplusexonEF-1a
introns,intron size,andintron/exon nucleotide boundaries are dataforthefoursetsofclosephylogeneticrelatives.IntronsII
summarizedinFigures 1 and4. IntronIappearedonlyintwo (Leiobunumonly)andIII(alltaxa)werealignedmanually,but
Sabaconsamples(thoseamplifiedwithEF1-OP2primers;most for the six-taxon Sclerobunus/Cyptobunus set, introns from
SabaconwereamplifiedusingdownstreamOP2BSABprimers). two specimens (OP1056, OP2143) were too divergent to align
We also used the EF1-OP2 for Leiobunum (Fig.1), but these reliably. Here we conducted phylogenetic analyses on exons-
sequences lack this intron. Examination of EF-1a genomic onlyforthesix-taxonset,andanalysesofintrons-onlyforfour
sequences for the tick Ixodes scapularis (http://iscapularis. of six taxa.
HEDINETAL.—EF-1aUTILITYINOPILIONES 15
Figure 3.—Exon-only majority-rule consensus parsimony tree (N 5 920, L 5 2760). Higher level clade names are indicated, following
taxonomy of Pinto-da-Rocha & Giribet (2007). Sequences generated in this study are shown with OP numbers; those from Shultz & Regier
(2001)areshownwithassociatedGenBanknumbers.TheSclerobunussequence(AF240858)generatedbyShultz&Regier(2001)wasoriginally
misidentifiedasS.robustus-thissequenceactuallycorrespondstoS.nondimorphicus.
Results of the nine separate exhaustive parsimony searches comparisons (see Table 3, Fig. 5). Finally, analyses of the
(three for Sclerobunus/Cyptobunus, two for all other congeneric independently-evolving mitochondrial versus EF-1a data result
sets) are summarized in Fig. 5. These comparative analyses in generally similar gene tree estimates. Visual inspection
reveal several patterns of interest. First, all trees are well- indicates an overall pattern of topological and branch length
resolved, with all but one of the searches resulting in a single similarity across data partitions. Moreover, we used the
most-parsimonious tree. Second, there is considerably more qualitative framework of Wiens (1998) to assess topological
divergence depth in mitochondrial trees as compared to EF-1a congruence. Among all trees, thirteen taxon bipartitions are
trees, as indicated by differences in total tree length and stronglysupported(bootstrapproportionvalues.70;seeHillis
individual branch length estimates. This reflects the well- &Bull1993)bythemitochondrialdata.Ofthese13bipartitions,
documented higher rates of molecular evolution in animal 10 are also strongly supported by the EF-1a data (Fig. 5).
mitochondrialgenomesascomparedto‘‘average’’nucleargenes Importantly, there is no evidence for ‘‘strong incongruence’’
(e.g., Moriyama & Powell 1997; Lin & Danforth 2004). Third, (followingWiens1998),whereconflictingtaxonbipartitionsare
phylogeneticsignalintheEF-1adataiscomingfrombothintron stronglysupportedbytheparalleldatasets.
andexonsites.Thisismost-clearlyillustratedintheSclerobunus/
DISCUSSION
Cyptobunustaxonsetwhereweconductedindividualanalyseson
exons versus introns, but is also seen when comparing intron A general lack of genomic resources for Arachnida has
versus exon substitutions for specific ‘‘shallow’’ pairwise hampered development of the type of multigenic molecular
16 THEJOURNALOFARACHNOLOGY
Table 3.—Patterns of EF-1a sequence evolution. Variable intron allow more accurate reconstructions of population and
sites were only counted at unambiguously aligned positions without species’ history, providing the necessary framework for
indels. The following samples were not included in intron variation addressing ‘‘shallow’’ systematic questions that abound in
counts:Sabacon(OP660,OP1535,OP2522),Sclerobunus/Cyptobunus this diverse group (e.g., fine-scale historical biogeography,
(OP1056, OP2143). Values for specific ‘‘shallow’’ pairwise compar-
cryptic speciation, rapid character evolution, etc.).
isons (involving samples with mitochondrial divergence values
The primer set and data reported here provide a starting
between2–3.2%)areshowninparentheses(seealsoFig.5).
pointforotherresearcherstoexplore,modify,andutilize.The
VariableIntronsites VariableExonsites exonmatrixnowavailable(www.treebase.org,studyaccession
number S2469) includes many conserved regions for primer
Taxon IntronII IntronIII Pos1 Pos2 Pos3
design, and because the matrix includes taxonomic coverage
Sclerobunus/Cyptobunus – 50(10) 4 2 44(1) for all higher-level harvestmen groups, it is easy to modify
Acuclavella – 16(1) 2 2 16(0) primers to match a group of interest. Furthermore, we have
Leiobunum 19(1) 11(3) 3 2 20(1) discovered an intron (intron III) that is most-parsimoniously
Sabacon – 9(1) 5 1 36(4)
reconstructedasbeingubiquitousinOpiliones,withobviously
conserved exon regions flanking this conserved position
intron. With minor experimentation, it should be easy for
systematics perspective that is now seen in many other researchers with access to standard PCR tools to generate
organismal groups (e.g., see Brito & Edwards 2008). In both intron and exon EF-1a data for Opiliones.
Opiliones, a core set of genes have been used at higher Themostconspicuousproblemthatwehaveidentifiedisthe
taxonomic levels (e.g., various mtDNA genes, 18S, 28S, presence of short simple sequence regions within introns,
histone 3, EF-1a), but nuclear gene choices at lower levels whichbecauseofapparentlengthheterozygosity,sporadically
are extremely limited. We needto develop additional rapidly- interrupt direct sequence reads. Here we would first recom-
evolving,informativenucleargenesforharvestmen.Thesewill menddesigningcustomprimersthatspanonlyasingleintron
Figure 4.—Intron/Exon boundaries. Sequences generated in this study are shown with OP numbers, all those without are from Shultz &
Regier(2001).Exonnucleotidesareshadedingrey.BasepositionsareshownrelativetoalignmentwithexonsequencesofShultz&Regier(2001).
HEDINETAL.—EF-1aUTILITYINOPILIONES 17
Figure 5.—Comparativemitochondrial(left)andEF-1a(right)parsimonytrees.Numberofparsimonyinferredchangesshownadjacentto
branches. TL 5 length of most-parsimonious tree. Bootstrap proportion values above 90 are indicated by black thick branches; proportion
valuesbetween70–90indicatedbygreythickbranches.Allsearchesresultedinasinglemost-parsimonioustree,exceptfortheLeiobunumEF-1a
matrix, which resulted in 3 most-parsimonious trees (we illustrate only one of these). For mitochondrial trees, pairwise comparisons with
divergencevaluesbetween2–3.2%(Kimura2-parameterdistances;Kimura1980)arehighlightedwithbold,italicizedtext.
(afterpreliminaryexplorationofintronstructureforthetaxon The data reported here reveal general congruence between
of interest). Depending upon the severity of length heterozy- nuclear versus mitochondrial gene trees (Fig. 5). Gene tree
gosity, it may be necessary to clone PCR products, which is congruence, or lack thereof, is a reflection of several
obviously more expensive and labor-intensive, but is still a interacting variables, including sampling density, the depth
widely used lab method. Other alternatives include allele- anddivergencehistoryofthegroupofinterest,andpatternsof
specificPCR,orbioinformaticsapproaches,suchastheuseof geneevolution(seeMaddison1997).Forexample,older,well-
software programs that allow phase determination of length spaced speciation events are expected to result in more gene
variableallelesresultingfromdirectsequencing(seeFlotetal tree congruence than more-recent, ‘‘compressed-in-time’’
2006; Flot 2007). radiations. Gene flow across species boundaries in recently
18 THEJOURNALOFARACHNOLOGY
separatedspeciescancauseincongruenceamonggenetrees,as align intron II splice site signal sequences (see Fig. 4). Such a
can comparisons of paralogous gene copies in multi-gene gap is never present in other exon sequences, even from close
families. phylogenetic relatives (e.g., Dendrolasma). This sequence may
Given our sample of taxa and exemplars within these taxa, represent aparalog, but the exon appears functional (no stop
we observed minimal incongruence, but again, this is an codons),andisplacedasexpectedphylogenetically.Aperhaps
empirical issue that is expected to vary from group to group. more plausible explanation is that we have misaligned the
In this context, one relevant question is how old are the intron boundary, but this requires non-canonical splice sites
harvestmen taxa within congeneric sets? We consider two for this intron. More data is needed to distinguish between
indirect lines of evidence for estimating the ages of these these alternative hypotheses. In general, there is little to no
groups. First, we have highlighted on Fig. 5 pairwise sample evidence for multiple copies of EF-1a in harvestmen.
comparisons that differ by 2–3 percent for mitochondrial Researchers in prior studies (Shultz & Regier 2001, 2009)
protein-coding genes. Applying a ‘‘standard’’ arthropod have not suggested this problem, relationships inferred from
mitochondrialmolecularclock(e.g.,2.3%pairwisepermillion exonsprovideexpectedphylogeneticresults,andgenetreesfor
years; Brower 1994; Pons et al. 2006) across trees for these congeneric sets are largely congruent with independent data.
individualgroupswouldindicatedivergencesrangingfrom1– However, the possibility of such paralogy should not be
8 million years ago (MYA). These are, of course, very rough dismissed or overlooked in future studies.
estimates,withmanyassumptionswhichmayormaynothold Varioustechnologiesarenowavailableforthegenerationof
(astandardclockappliesto Opiliones,noamong-lineagerate comparative nuclear sequence data. These include expressed
variation, etc). Second, there is strong evidence for ancient sequence tag (EST) libraries, next-generation sequencing
vicariance within Acuclavella, leading to the primary phylo- (Mardis 2008), development of anonymous nuclear loci (e.g.,
genetic separation of Olympic and Cascade mountain Noonan & Yoder 2009), etc. Ultimately, these technologies
populations (OP2345, OP2347) from all other populations in maysupercedestudiesthatfocusonasingle‘‘candidategene,’’
the IdahoRockies.In manyothertaxa fromthisregion (e.g., forexample,ifwecanultimatelygeneratewholegenomicdata
Carstens et al. 2004; Steele et al. 2005), this west/east forallsamplesofinterest.However,thesetechnologiesarestill
vicariance results from the Cascadian orogeny at approxi- too expensive for the average systematist, particularly in
mately 5 MYA, which suggests a divergence time window geographic areas where resources for systematics research are
generallyconsistentwiththeabovemolecularclockestimates. limited. Until then, we argue that studies of the phylogenetic
The above divergence time estimates suggest that EF-1a behavior of ‘‘candidate genes,’’ such as the research reported
datamightbeusedtoresolvedivergencesthatarequiterecent here, still have considerable utility in the arachnological and
in absolute time. Although not extensive, all pairwise sample broader systematics communities.
comparisons at the ‘‘one million year horizon’’ also show at
leastsomedivergenceinEF-1a(seeTable 3).Furthermore,we ACKNOWLEDGMENTS
notethatourphylogenetictreatmentofinsertionanddeletion
WethankJoeDeas,DamianElias,DaltonHedin,LarsHedin,
(indel) variation in introns was very conservative, as several
RobinKeith,JimStarrett,StevenThomasandJeffUnderwood
indel sites appeared phylogenetically informative. More
forassistanceincollectingspecimens.AxelScho¨nhoferprovided
sophisticated treatment of this class of variation is expected
important European Sabacon specimens. Personnel at Lewis &
to further increase the phylogenetic informativeness of this
Clark Caverns State Park were helpful in our collections of
gene regionfor recent divergences (e.g., Simmons & Ochoter-
Cyptobunuscavicolens.SpecimenscollectedintheGreatSmoky
ena 2000; Kawakita et al. 2003; Benavides et al. 2007).
MountainsNational Parkwere collectedundera permitissued
In terms of gene tree congruence, one variable that seems
to Steven Thomas. Fred Coyle, Steve Perlaky, and the Abbott
nottobeinfluencingourdataisthepresenceofmultiplegene
family provided excellent accommodations while collecting in
copies, and potential mixing of paralogous copies. The
southern Appalachia. J.W.S. was supported by the Maryland
paralogy/orthology issue is one of the most important
Agricultural Experiment Station. Funding was provided by a
problems negatively impacting the phylogenetic utility of
collaborative NSF grant awarded to J. Shultz (Award
nuclear protein-coding genes (reviewed in Sanderson &
No. 0640179)andM.Hedin(AwardNo. 0640173).
Shaffer 2002). Multiple gene copies are in fact known for
EF-1a in other arthropods (see Danforth et al. 1999),
including spiders (Hedin & Maddison 2001). ‘‘Deep paral- LITERATURECITED
ogy,’’ or the inadvertent comparison of divergent gene copies Ballard,J.W.&M.C.Whitlock.2004.Theincompletenaturalhistory
(e.g., gene copies resulting from gene duplication before the ofmitochondria.MolecularEcology13:729–744.
divergence of Opiliones), is apparently not an issue in our Benavides, E., R. Baum, D. McClellan & J.W. Sites, Jr. 2007.
data,asmixingsuchparalogswouldnotresultintheexpected Molecular phylogenetics of the lizard genus Microlophus (Squa-
systematic relationships that we see in the exon-only analysis mata: Tropiduridae): aligning and retrieving indel signal from
nuclearintrons.SystematicBiology5:776–797.
(Fig. 3).Analternativeisthatharvestmenspeciesarecarrying
Bower, J.E., R.D. Cooper & N.W. Beebe. 2009. Internal repetition
multiple copies of the EF-1a gene that are kept similar by
andintra-individualvariationintherDNAITS1oftheAnopheles
concerted evolution (as hypothesized for salticid spiders,
punctulatusgroup(Diptera:Culicidae):multipleunitsandratesof
Hedin & Maddison 2001); distinguishing such closely-related
turnover.JournalofMolecularEvolution68:66–79.
paralogs would be more difficult in a phylogenetic analysis.
Brady,S.G.&B.N.Danforth.2004.RecentintrongaininElongation
Theonly hintof such‘‘cryptic’’ paralogyistheOrtholasma factor-1a of colletid bees (Hymenoptera: Colletidae). Molecular
OP807 sequence, which requires a 2 amino-acid exon gap to BiologyandEvolution21:691–696.