In Southern South America, Evidence from AFLP and ITS Sequence Data Author(S): Alexandra M

Home , Helwingia chinensis, Ilex amelanchier, Ilex anomala, Ilex collina, Ilex cornuta, Ilex glabra

Molecular Analyses of the Genus Ilex (Aquifoliaceae) in Southern South America, Evidence from AFLP and ITS Sequence Data Author(s): Alexandra M. Gottlieb, Gustavo C. Giberti and Lidia Poggio Source: American Journal of Botany, Vol. 92, No. 2 (Feb., 2005), pp. 352-369 Published by: Botanical Society of America, Inc. Stable URL: https://www.jstor.org/stable/4123880 Accessed: 18-12-2018 17:13 UTC

REFERENCES Linked references are available on JSTOR for this article: https://www.jstor.org/stable/4123880?seq=1&cid=pdf-reference#references_tab_contents You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at https://about.jstor.org/terms

Botanical Society of America, Inc. is collaborating with JSTOR to digitize, preserve and extend access to American Journal of Botany

This content downloaded from 157.92.6.34 on Tue, 18 Dec 2018 17:13:49 UTC All use subject to https://about.jstor.org/terms American Journal of Botany 92(2): 352-369. 2005.

MOLECULAR ANALYSES OF THE GENUS ILEX (AQUIFOLIACEAE) IN SOUTHERN SOUTH AMERICA, EVIDENCE FROM AFLP AND ITS SEQUENCE DATA'

ALEXANDRA M. GOTTLIEB,1,3 GUSTAVO C. GIBERTI,2 AND LIDIA POGGIO'

'Universidad de Buenos Aires, Facultad de Ciencias Exactas y Naturales, Ciudad Universitaria, Pabell6n II, 4to piso, Departamento de Ecologifa, Gen6tica y Evoluci6n, Lab. 62, C1428EHA, Buenos Aires, Argentina; and 2Instituto de Quifmica y Metabolismo del Ffirmaco (IQUIMEFA), Junin 956, 1113-Buenos Aires, Argentina

In order to clarify the relationships among southern South American (sSA) representatives of the genus Ilex, an amplified fragment length polymorphism (AFLP) analysis was accomplished. In addition, the phylogenetic relationships of the species were studied using ribosomal internal transcribed spacer (ITS) sequence data alone and in combination with AFLP data, taking into account the possible existence of paralogous sequences and the influence of alignment parameters. To explore stability of phylogenetic hypotheses, a sensitivity analysis was performed using 15 indel-substitution models. Within each species assayed, the AFLPs allowed the recognition of several diagnostic bands. Furthermore, the AFLP analysis revealed that individuals belonging to the same morpho-species formed coherent clades. In addition, some cases of geographical association were noted. Studies on ITS sequences revealed divergence between data obtained herein and sequence data downloaded from GenBank. The sensitivity analyses yielded different interspecific hypotheses of relationships. Notwithstanding, analyses of the ITS data alone and in combination with AFLPs, rendered clades stable to variation in the analytical parameters. Topologies obtained for the AFLPs, the ITS data alone and the combined analyses, demonstrated the existence of a group formed by I. argentina, I. brasiliensis, I. brevicuspis, L integerrima, and L theezans, and that L dumosa and I. paraguariensis were distantly related to the former. Incongruence with traditional taxonomical treatments was found.

Key words: AFLP; direct optimization; Ilex; ITS; sensitivity analysis.

The plant genus Ilex L., which belongs to the holly family, Reissek, L microdonta Reissek, L paraguariensis A. St. Hil., Aquifoliaceae Bartl., comprises deciduous and evergreen I. pseudobuxus Reissek, I. taubertiana Loes., and I theezans bushes or trees with economic importance as crops and C. or- Martius ex Reissek. Of those mentioned, I. paraguariensis namentals. These plants are functionally dioecious; rudimen- is the most relevant from an economic perspective. In Argen- tary pistils in staminate flowers and sterile stamens in pistillate tina, southern Brazil, Paraguay, and Uruguay, the aerial parts flowers are borne on different plants. A remarkable morpho- of this plant are used to prepare a very popular beverage called logical diversity has been described in the genus, including mate, which is highly appreciated for its flavor and stimulating trees to creeping prostrate plants or epiphytic shrubs. It properties has due to its caffeine and theobromine contents (Filip more than 400 species, dispersed mainly in America and Eur-et al., 2001). asia, but also in Oceania and Africa (Giberti, 1994b). About a hundred years after Linnaeus (1753) published the South America is considered one of the main areas of di- diagnosis for the genus Ilex, Gray (1856) attempted to estab- versification of Ilex, together with East Asia (Loesener, lish 1942; an infrageneric classification. In 1861, Reissek published Cuenoud et al., 2000). In southern South America (sSA) the greatestthe species inventory for the time, followed by Hooker species are mostly found in areas with tropical or subtropical (1862) who reported the occurrence of about 145 species climates, mainly in northeastern Argentina, southeastern worldwide. Brazil Loesener (1891, 1901, 1908, 1942) published a and eastern Paraguay. Twelve species have been described series for of seminal studies recording almost 280 species of Ilex, sSA (Loesener, 1901; Lillo, 1911; Edwin and Reitz, 1967; dispersed Gi- globally, and recognizing several subgeneric ranks berti, 1998), namely, Ilex affinis Gardner, I. argentina althoughLillo, L under an unconventional hierarchical arrangement. brasiliensis (Sprengel) Loes., I. brevicuspis Reissek, SinceI. cha- then, several authors have recognized some inaccuracies maedryfolia Reissek, L dumosa Reissek, L integerrima in(Vell.) Loesener's system and proposed emendations in the systematics of the genus (Rehder, 1927; Hu, 1949, 1950; Krtissmann, 1962; Baas, 1975; Lobreau-Callen, 1975; Martin, 1977). Spe- ' Manuscript received 11 December 2003; revision accepted 17 September cies delimitation in Ilex, based on overall morphological sim- 2004. ilarity, is complicated due to the various concepts used by The authors are indebted to L. D. Belingheri, S. D. Prat Kricun, and R. M. Mayol from EE. INTA Cerro Azul (Misiones, Argentina) who kindly different provided authors to define the taxa and to the lack of complete specimens from living collections; to N. Jouv6 and P. Rubio of University information, of particularly for evergreen species (Galle, 1997). Alcald (Madrid, Spain) who supported the sequencing, to M. E. Rol6n In from order to investigate the relationships among the species the "Establecimiento Las Marias" (Corrientes, Argentina) who facilitatedof Ilex found in sSA, we implemented the amplified fragment commercial lines of "mate," to S. A. Catalano and V. A. Confalonieri for length polymorphism (AFLP) technique (Vos et al., 1995). In their invaluable suggestions. We also want to thank three anonymous review- addition, we studied the phylogenetic relationships of the spe- ers whose suggestions helped to improve this manuscript. This work was supported from grants of the "Agencia Nacional de Investigaciones Cientifi- cies using ribosomal internal transcribed spacer (ITS) se- cas y Tecnicas" (Argentina) (PICT 01-4443) and from CONICET (PIP 0559/ quence data alone and in combination with AFLP data, taking 98). into account the possible existence of paralogous sequences SAuthor for correspondence. E-mail: [email protected]. and the influence of alignment parameters.

352

This content downloaded from 157.92.6.34 on Tue, 18 Dec 2018 17:13:49 UTC All use subject to https://about.jstor.org/terms February 2005] GOTTLIEB ET AL.-MOLECULAR ANALYSES OF SOUTH AMERICAN ILEX 353

Beside the general use of AFLP markers in the estimation ments. Because parameter choice is arbitrary but unavoidable of genetic diversity and differentiation among individuals, when doing algorithmic DNA sequence comparisons, its ex- populations, and plant species, they have also been ploration success- becomes essential (Giribet and Wheeler, 1999; Gi- fully used in phylogenetic analysis between closely ribet related and Ribera, 2000). Herein, we used ITS sequences to species (Cervera et al., 1998; Kardolus et al., 1998; study Mace the et phylogenetic relationships of sSA taxa and to explore al., 1999; Negi et al., 2000; Ren and Timko, 2000; the Koopman influence of applying different insertion/deletion (indel or et al., 2001). Because of the rapidity with which reliable gap) and high substitution models (sensitivity analysis sensu resolution markers can be generated, the AFLP technique Wheeler, is 1995) on phylogenetic inference. Phylogenetic anal- considered a powerful tool for molecular studies (Mueller yses were and accomplished through direct optimization of DNA Wolfenbarger, 1999). In addition, the high ratio of sequencespolymor- (Wheeler, 1996). The sensitivity analysis is a way phism generated per PCR experiment (the multiplex to explore ratio) the data and to discern between stable relationships (Powell et al., 1996) and the random distribution of most (those supported throughout the parameters range) and unsta- AFLP bands across the genome (Zhu et al., 1998) are consid- ble relationships (those appearing only under particular con- ered key features of this technique. Several drawbacks have ditions), allowing the formulation of more robust hypotheses been detected for AFLPs (Robinson and Harris, 1999), which (Giribet and Ribera, 2000) than in customary phylogenetic introduce bias into the estimates, whether analyzed by phe- analyses. Direct optimization is a method that does not dis- netics or cladistics. connect the alignment step from the tree-building step and that A fundamental requirement in phylogenetic studies based generalizes phylogenetic character analysis to include indel on nucleic acid or protein sequences is that the genes com- events (Giribet, 2001, 2003). By doing this, indels appear as pared are orthologous (originating from organismal cladogen- transformations (not as states) linking ancestral and descendent esis) (Alvarez and Wendel, 2003). Plant rDNA consists of nucleotide sequences (Giribet and Wheeler, 1999). thousands of copies or paralogues (derived from gene duplication events) spanning several loci, which evolve through MATERIALS AND METHODS gene conversion, unequal crossing over, and perhaps repeat amplification (Baldwin et al., 1995). ITS sequences have prov- Specimens studied-The list of sSA flex taxa employed, together with their en useful in species-level phylogenetic studies of a wide range accession numbers, geographical origin and GenBank accession numbers of of taxa (Alice and Campbell, 1999; Downie et al., 2000; Lia ITS sequences obtained in this study are presented in Appendix 1 (see Sup- et al., 2001, among others). Within-individual rDNA poly- plemental Data accompanying the online version of this article). Plant material was principally obtained from the Estaci6n Experimental INTA Cerro Azul morphisms may occur when concerted evolution is not fast (EEINTACA) Germplasm Bank (Misiones, Argentina). Leaf samples were enough to homogenize repeats in the face of high rates of also collected from natural populations from Northern Argentina; voucher mutation and/or recent interspecific hybridization (Campbell specimens of these samples were deposited at the BACP Herbarium (Centro et al., 1997). Intraspecific variation concerning ITS sequences de Estudios Farmacol6gicos y Botfinicos, CEFyBO, Buenos Aires, Argentina). has been detected in several plant groups (Baldwin et al., 1995 Furthermore, leaf samples of commercial lines of mate were obtained from and references therein; Buckler and Holtsford, 1996a, b; Buck- the Establecimiento Las Marias (Corrientes, Argentina). Young leaf samples ler et al., 1997; Campbell et al., 1997; Hartmann et al., 2001; were collected from healthy plants (without evidence of fungal or insect in- Mayol and Rosell6, 2001). Active ITS regions have functional fection) and were silica-gel preserved. Additional 42 ITS sequences were constraints that can help discriminate between functional anddownloaded from GenBank (http://www.ncbi.nlm.nih.gov); for their geo- nonfunctional paralogous (i.e., pseudogene) sequences (Buck- graphical origin and accession numbers see Appendix 2 (Data Supplement ler et al., 1997). For instance, nonfunctional ITS paralogues accompanying the online version of this article). Helwingia japonica (C. P. could accumulate random substitutions, which will destabilize Thunberg ex A. Murray) Dietrich and Helwingia chinensis Batalin were cho- hairpins and reduce secondary structure stability. When non- sen as the outgroup. According to several molecular studies (Cu6noud et al., functional paralogous ITS sequences are unknowingly includ- 2000; Powell et al., 2000; Savolainen et al., 2000; Albach et al., 2001; Bremer ed in the phylogenetic analyses (to the exclusion of appropriate et al., 2002), Helwingia and Ilex are closely related members of Aquifoliales. orthologous comparisons), the resulting gene tree may con- Loeseners' (1891, 1901, 1908, 1942) original dispositions of subgeneric ranks found organismal divergence events with a tracking of the his- were used in this work (Fig. 19). The species flex microdonta, I. pseudobuxus, tory of gene duplication (Alvarez and Wendel, 2003). Erro- and I. taubertiana, were only included in the sequence analyses because accessions of these taxa were unavailable in sufficient number for AFLP anal- neous assessment of orthology and paralogy will lead to phylogenetic incongruence. Hence, discrimination of inactive par- ysis. Only two sSA taxa remain to be studied, namely flex affinis and I. chamaedryfolia, because these have not been found in the field since the alogues and pseudogenes becomes crucial. Another critical step in molecular phylogenetic studies is 1980s (G. Giberti, Instituto de Quifmica y Metabolismo del Firmaco, IQUI- MEFA, personal communication). Attempts to obtain good quality DNA from the alignment of sequence data, in which primary homology herbarium/air-dried specimens were unsuccessful. statements are being established. However, it is typically performed without considering the effects of analytical parame- DNA extraction-Leaves were ground to a fine powder in liquid nitrogen ters, such as indel and base transformation values, on the phy- and then placed in a microtube. The DNeasy Plant kit (QIAGEN Inc., Valen- logenetic inference. In the case of multiple sequence align- cia, California, USA) was used for DNA extraction following the manufac- ments, different parameters may yield alternative alignments, turer's instructions. DNA was kept at 40C until use and then stored at-200C. and these primary homology hypotheses may support distinct topologies (Doyle and Davies, 1998). Indeed, Morrison and AFLP technique--The AFLP methodology was carried out using the AFLP Ellis (1997) showed that different multiple sequence alignment Analysis System I Starter Primer kit (Gibco BRL, Life Technologies, Carls- methods were responsible for the variation encountered among bad, California, USA) as described in the instruction manual. Eight selective resultant phylogenetic trees topologies, and this variation wasprimers were combined as follows: E + AGG, M + CTT; E + ACT, M + more important than the one found when different methods CAC; of E + AGC, M + CAT; and E + ACC, M + CAG. Other primers phylogenetic reconstruction were applied to the same align- combinations assayed were discarded because they generated profiles in which

This content downloaded from 157.92.6.34 on Tue, 18 Dec 2018 17:13:49 UTC All use subject to https://about.jstor.org/terms 354 AMERICAN JOURNAL OF BOTANY [Vol. 92 the amplification products were sequence too dense replicates. to allow Characters reliable scoring, were equally or alter- weighted and unordered, and all natively, generated too few amplification other settings products. were defaults. Selective The amplification reliability of the clades was evaluated by products were mixed with an equalbootstrap volume analysis of dye (Felsenstein, reagent (98% 1985) [v/v] in form- PAUP* with 1000 replications and amide, 10 mM EDTA, 0.025% all[w/v] other bromophenol settings as described blue and earlier. 0.025% Trees [w/v] were inspected with the program xylene cyanol). Five microlitres TreeViewwere separated (Page, by1998). electrophoresis The unrooted in resulting a Model strict consensus tree was root- S2 apparatus (Gibco BRL Sequencing ed using System, I. paraguariensis Life Technologies) to be consistent through with ITS trees. 6% (w/v) polyacrylamide gels containing 5 M urea, in 1 X TBE buffer (89 mM Tris, 89 mM boric acid, 2ITS mM sequence EDTA, studies-WepH 8). A 30-330 characterized bp AFLP andDNA identified ITS paralogues in Ladder (Gibco BRL, Life Technologies) sSA Ilex species size marker by examining was included their twice G + Cor content, secondary structure three times in each electrophoresis stabilities, run. Thus,and nucleotide the size of substitution AFLP bands patterns scored and using the tree-based ap- ranged from 50 to 330 bp. Gels proach were stainedadvocated with by silver Bailey nitrate et al. (Bassam(2003). Toet discriminate the presence of al., 1991). presumed active sequences from paralogues and putative pseudogenes, we first followed the procedure proposed by Mayol and Rosell6 (2001). There- ITS amplification-Amplification fore, of pairwise ITS 1, 5.8S nucleotide gene, and divergences ITS 2 was werecar- calculated between the 21 ITS ried out using primers ILEXFP sequences (5'-AACAAGGTTTCCGTAGGTGA-3', obtained herein and the 10 sequences retrieved from GenBank that Powell et al., 2000) and ITS-4 (Whitebelong etto al.,the 1990).same sSAAmplifications Ilex taxa. The were values per- were calculated using the Ki- formed in 50 [pL using 0.225 mMmura of two-parameter each dNTP (Promega model (Kimura,Corp., Madison, 1980) with DNAdist 3.5c program im- Wisconsin, USA), 20-25 pmoles plemented of each primer, in BioEdit 5 .L (Hall, of Thermophilic 1999). Additionally, DNA ITS sequences were screened Polymerase 10 X Buffer (Promega using Corp.), BioEdit 1.5 (Hall, mM 1999)MgC12, for 1.25 length units variation, of Taq G + C content, and the pres- DNA polymerase (Promega Corp.), ence of I fxLa structural of genomic motif DNA in ITStemplate, 1 (i.e., GGCRY(4and to 7 n)GYGYCAAGGAA; sterile double-distilled water. Conditions Liu and Shardl, for PCR 1994). amplification This conserved were those motif of is often associated with part of Vilgalys and Hester (1990). Amplifications the hairpin andwere loop carried structures out in thatan Eppendorf may serve as a critical signal in the en- Mastercycler (Perkin-Elmer Corp., zymatic Foster processing City, California, of the ribosomal USA). Control RNA (Liu and Shardl, 1994). Further- samples without DNA template more, were optimalincluded and in eachsuboptimal PCR run. folding PCR structures,products and associated free energy were checked by electrophoresis values of 1-[LL(AG) insamples Kcal/mol, in 1% of agarose ITS 1 andgels ITS in 12 Xwere recorded. Fold predictions TAE buffer (0.04 M Tris, 0.114% were glacial made atacetic the acid,Quickfold 1 mM web EDTA, server pH (Zuker,8.0). 2003) of the MFOLD pro- Quantification of PCR products gram, was versionestimated 3.1. by In comparison all cases, foldings with known were conducted at 370C by use of a amounts of a DNA molecular-size search marker within (Lambda 5% of theEcoRI/HindIII, thermodynamic Promega optimality set. Because each se- Corp.) included in duplicate in quenceall gels. could Gels adoptwere morestained than with one ethidium folding structure,bro- maximum and minimum mide for 15-20 min and photographed AG values underwere recorded UV light. for In each some sequence, cases, best together with the number of all amplification results were achieved possible by(optimal adding and 6% suboptimal bovine serum foldings) albumin secondary structures. However, no (BSA, Promega Corp.) to the PCR attempts reaction were mix. made PCR to products study the were RNA purified structures per se. These screenings using a QIAGEN Gel Extraction were kit extended (QIAGEN to the Inc.). remaining Both strands representatives of the of Ilex and outgroup. complete ITS region (ITS 1-5.8S-ITS 2) were sequenced in Perkin Elmer/ Applied Biosystems automatic Phylogeneticsequencing units and (ABIsensitivity Prism 310analyses-Parsimony or 377) at was chosen as the op- the Molecular Biology Facility timalityof the Universidad criterion, asde implementedAlcaldi (Madrid, in theSpain). program for Direct Optimization Electropherograms were proofread POY, with version the software3.0.11 (Gladstein BioEdit Sequence and Wheeler, Align- 1997; Wheeler and Gladstein, ment Editor (Hall, 1999). Boundaries 2000; Deof Laetthe codingand Wheeler, and spacer 2003). regions To save were computation time, as suggested determined by comparison with in published the POY sequencesuser's manual of Ilex. and In ingeneral, Giribet when (2001), the sequence data set was two or more orthologous ITS partitioned. sequences were The availablefirst partition per sSA comprises taxon (see 257 to 265 bp, and the second Appendix 1; Supplemental Data partition accompanying has ca. the 340 online bp. Approximately version of this 45 ar- positions of the 5.8S gene, rep- ticle), these were used to produce resenting consensus missing sequences data inin severalBioEdit sequences (Hall, 1999) downloaded from the GenBank, to speed-up computational times were of excludedphylogenetic from analyses. the analyses. A parameter space of two analytical variables was examined, i.e., gap cost and transversions/transitions (Tv/Ts) ra- Data analyses-AFLP-Air-dried tio gels cost. were Precisely, scanned we and used banding gap costs patterns = 1, 2, 4, 8 and 10, and Tv/Ts costs were registered with the Gel-Pro = 1 (equal4 Analyzer weights), Program 2 (transversions (Media Cybernetics, receive twice the weight than transi- Maryland, USA). Each AFLP band, tions) regardless and 4. All of combinations its relative intensity,of transformation was costs used were in accor- considered as a dominant allele dance at a uniquewith the locus. triangle In some inequality cases, fragments (Wheeler, 1993, 1995). Each particular were scored as missing data because tree charactersearch was states performed could not as be follows: interpreted 50 independent heuristic searches unambiguously. The data were starting extracted from as athe table best as of either five independentpresent (1) orWagner trees were conducted via absent (0). Monomorphic bands subtree-pruning (bands present inand all regrafting individuals (SPR) of a followedspecies) by TBR, retaining two trees and diagnostic bands (monomorphic per step, bands as exclusivelyspecified in present the following in one species) commands: "-norandomizeoutgroup were discriminated within each -buildsperreplicate species and across 5the -replicates entire data 50 set. -maxtrees Box 2 -slop 2 -checkslop 5". plots of mean values; SE and SD Commandswere calculated "-slop" and plottedand "-checkslop" with the program were employed to reduce error de- STATISTICA '99 edition (Kernel rived release from 5.5 the A, heuristicStatSoft, operations1999). For these(Wheeler, 2002). Step matrices (i.e., calculations, accessions with missing transformation data in one ormatrices) more primer were specifiedcombinations in POY using the command "-mo- were excluded. Pairwise genetic lecularmatrix" distance (D) with between an argument all individuals for the given was step matrix. Strict consensus estimated according to the complementary trees were calculated value of Neiusing and the Li's command: (1979) sim- "-poystrictconsensustreefile". ilarity coefficient, implemented An inimplied PAUP* alignment(Swofford, was1998): generated D = 1 - (2n,/1 in each case by the command: [n; + ni]), where n,i is the number "-impliedalignment". of shared fragments Approximate between two Bremer individ- support values (Bremer, 1988) uals i and j, and n, and n, are thewere total calculated number with of fragments a heuristic in approximation individual i procedure implemented in and j, respectively. From these POYvalues, (command: mean intra- "-bremer"). and interspecific Finally, distance strict consensus trees and single trees values were obtained. Furthermore, were usedthe AFLPto obtain data 50%set for majority 120 accessions rule consensus tree. Implied alignments belonging to eight taxa was analyzed and trees using were maximum deposited parsimony at TreeBase in PAUP*. (S996). We specifically used nodal Initial trees in the heuristic stabilitysearch by as tree-bisection-reconnection defined by Giribet (2003), i.e., (TBR) the degree to which nodes are branch-swapping algorithm were affected built by stepwise variation with in the500 analytical random additionconditions. Robustness was used here-

This content downloaded from 157.92.6.34 on Tue, 18 Dec 2018 17:13:49 UTC All use subject to https://about.jstor.org/terms February 2005] GOTTLIEB ET AL.-MOLECULAR ANALYSES OF SOUTH AMERICAN ILEX 355

TABLE 1. Number and relative abundance of AFLP monomorphic (m) TABLE 2. Number of AFLP monomorphic (m) and diagnostic (d) and diagnostic (d) bands obtained per primer combination for se- bands recognized per Ilex species. lected Ilex taxa.

d/m X Total m/total d/m X Species m d 100 (%) Primer combination bands m x 100 (%) d 100 (%) Ilex argentina 58 19 33 E + AGG/M + CTT 224 28 12.5 12 43 L brasiliensis 60 13 22 E + ACT/M + CCA 210 57 27 30 53 I. brevicuspis 40 19 47.5 E + AGC/M + CAT 193 42 22 21 50 L dumosa 13 8 61.5 E + ACC/M + CAG 209 58 28 31 53 L integerrima 74 13 17.5 Total 836 185 22 94 51 L paraguariensis 38 17 44.7 L theezans 69 5 7

in as a measure of stability RESULTS of nodes to fluctuation in parameters values across an analytical landscape (Ogden and Whiting, 2003). Two sets of analyses AFLP were analysis-The performed. four primer combinations First, generated we undertook a phylogenetic analysis of all ITS sequences, 836 scoreable bands, of whichincluding 22% were monomorphic those (Table tentatively determined as paralogues and pseudogenes. 1). All 120 individuals To explore tested shared a single the constant influence band. of alignment parameters on the relative placement Primer combinations of E + ACT/Mthese + CCA and sequences,E + ACC/M five different combinations of the two variables were + CAG yieldedemployed. a nearly identical number In ofparticular, monomorphic these combinations were: 111 (i.e., gap cost = 1, andTv diagnostic = bands.1 Amongand the sevenTs species 1), included 241, in 421, 811 and 1021. Then, we performed a broader this study,sensitivity 94 diagnostic bands were distinguished, analysis which ac- using only presumably active ITS sequences under 15 countedanalytical for more than 50% of conditionsthe monomorphic fragments (namely, 111, 121, 141, 211, 221, 241, 411, 421, 441, recorded.811, The proportion821, of monomorphic841, and1011, diagnostic 1021 and 1041). Each of the 15 tree searches was performed bands detected varied amongas theindicated species (Table 2). Ilex inte-earlier. gerrima had the highest number of monomorphic bands, Combined analysis-The whereas the highest combined number of diagnostic bands analysis was obtained involved AFLP data and ITS sequences. The sensitivity for I. argentina and I.analysis brevicuspis. In the box plotwas in Fig. 1,performed in POY as indicated previously, with the exception I. argentina gave thethat highest mean a numberthird of AFLP filebands consisting of AFLP data per species was included in all calculations. (mean = 175, SD = 29.9), whereas The the lowest new number dataof matrix for species was extracted from the original AFLP bands was matrix found in I. brevicuspis for (mean individuals= 124, SD = 6.1). and consisted of eight taxa and 578 characters. A band Intra- and was interspecific considered distance values derived from AFLPpresent in a taxon (1) if it was recorded in more than data80% are presented of in Tablethe 3. Excluding individuals, I. dumosa, for which whereas it was considered absent in a taxon (0) if it was the highest recorded intraspecific distance invalue wasless obtained (0.1223,than 20% of the individuals. Otherwise, bands were considered SD = 0.0408), as remaining polymorphic taxa had similar values. When in-(0,1). The weights of AFLP changes were set so that nucleotide terspecific mean distances characters were compared, I. integerrima were and never upweighted relative to AFLP data. The computer I. programtheezans appeared to be the leastPOY distant pairallows of taxa (0.1014, nonsequence character changes to be simultaneously optimized Table 3). with molecular data on multiple trees.

220

180. 0 CO 2 00 ......

F-- o

120 ...... Mean+SD Mean-SD m Mean+SE 100 Mean-SE I. argentina I. brevicuspis I integerrima I theezans I. brasiliensis I. dumosa I. paraguariensis O Mean

Ilex species

Fig. 1. Comparison of mean number of AFLP bands obtained for selected Ilex species.

This content downloaded from 157.92.6.34 on Tue, 18 Dec 2018 17:13:49 UTC All use subject to https://about.jstor.org/terms 356 AMERICAN JOURNAL OF BOTANY [Vol. 92

TABLE 3. Intra- (diagonal line in bold) and interspecific mean distance values (Nei and Li, 1979) for sSA Ilex taxa based on AFLP data.

Taxa I. argentina L brasiliensis I. brevicuspis I. dumosa L integerrima paraguariensis I theezans

L argentina 0.0738 I. brasiliensis 0.1582 0.0718 I. brevicuspis 0.2061 0.1996 0.0633 I. dumosa 0.241 0.2657 0.2570 0.1223 L integerrima 0.1801 0.1267 0.2079 0.2577 0.0739 I. paraguariensis 0.2211 0.2394 0.2449 0.2303 0.2271 0.0692 L theezans 0.1616 0.1096 0.1869 0.2491 0.1014 0.2367 0.0624

Of the 836 binary characters, zans, 749no distinct were differences parsimony were observed between infor- our se- mative. The heuristic search resulted quences and those in from 148 GenBank. most Instead, parsimonious major differences trees (MPT) with a length of were 5260found between (CI our = nucleotide 0.1561, data and RI sequences = 0.66). I bras- The strict consensus tree derived from AFLP data showed that iliensis AJ492661, I. brevicuspis AJ492663, I. dumosa all individuals of the same taxon grouped accordingly (Fig. 2). AJ492656, L integerrima AJ492664, and I. pseudobuxus Accessions of L theezans yielded a coherent group (BV, boot- AJ492660. These downloaded data showed a markedly low G strap value; BV = 93%), closely related to the clade L inte- + C content and a considerable reduction in stability of the gerrima (BV = 95%). Within L brasiliensis (BV = 66%), two most probable secondary structure, both for ITS 1 and ITS 2. subclades were distinguished, one comprised accessions from Ilex brevicuspis AJ492662 and I. dumosa AJ492657 were not Argentina and Paraguay (BV = 83%), and the other subclade completely comparable to other co-specific data due to a dif- involved individuals from Brazil (BV = 82%). In contrast, ference in length. Unfortunately, these sequences were incom- specimens of L argentina formed a highly supported clade plete lacking at least, 60 bp from the 5' end of ITS 1. Not- (BV - 97%). The clade L brevicuspis (BV = 100%) showed withstanding, their ITS 2 had comparable values to our se- several internal nodes unresolved and poorly supported. With- quences in all factors considered. Excluding these incomplete in the I. dumosa clade (BV = 99%), some geographical struc- sequences, no significant differences in length were encoun- turing was detected. The first subclade (BV < 50%) involved tered. Among North American and East Asian sequences, individuals from Argentina and Paraguay representing both va- some variations in the screened parameters were found (Table rieties of I. dumosa (I. dumosa var. dumosa and L dumosa var. 5). In particular, the values obtained for ITS 2 of L kiusiana guaranina, the latter indicated with an arrow in Fig. 2), while make us question its functionality. The lack of additional se- the other subclade comprised solely Brazilian accessions (BV quences impeded us from making a final decision. 68%). Regarding L paraguariensis (BV = 100%), the ma- The universal ITS 1 core motif of Liu and Schardl (1994) jority of Argentine, Brazilian, and Paraguayan accessions of was screened in all 55 sequences, and some variants were de- mate appeared intermingled (BV = 96%); still, most internal tected. Particularly in Ilex, 13 variants of this sequence were relationships were poorly supported. Commercial lines of mate recognized and arbitrarily named (Table 6). The types B and (SI-16, SI-19, SI-167 and G-18) formed a clade with a spec- D were the most frequent ones (21/53 and 11/53, respectively). imen from Paraguay (BV = 59%). Another subclade involved This scrutiny revealed that all six I. dumosa sequences lines SI-49, Y-383, and two accessions from Paraguay (BV (AY183479, AY183480, AY183481, AY183488, AY183489, 66%). The relationship between L brasiliensis, I. integerrima, and AJ492657) had a different motif that lacks the final ade- and L theezans was highly supported by the data (BV - nine. As to the GenBank data (L brasiliensis AJ492661, L 100%). Likewise, the relationship of L argentina and I. brev- brevicuspis AJ492663, L dumosa AJ492656, I. integerrima icuspis to the aforementioned taxa was also well-supported by AJ492664, and I. pseudobuxus AJ492660), these exhibited a the data (BV = 91%). motif that substantially departs from the core sequence, con- tributing to a less stable secondary structure (Tables 5 and 6). ITS sequence studies-Nucleotide divergence calculations Two of these (L dumosa AJ492656 and L brevicuspis showed that in several cases sSA ITS sequences retrieved from AJ492663) were previously reported as paralogous (Manen et GenBank were highly divergent from those obtained in this al., 2002). report (Table 4). For instance, similar divergence values (ca. 14-15%) were found for ITS sequences of L brasiliensis, I. Phylogenetic and sensitivity analyses-The input data set integerrima, and I. pseudobuxus. While I. brevicuspis for the phylogenetic analysis comprised 53 ingroup sequences AJ492663 diverged with our ITS data in ca. 11%, accession and two outgroup, also including putative divergent paralogues AJ492656 of I. dumosa, was ca. 21% divergent from L du- and pseudogenes. Analytical conditions 241, 421, and 1021 mosa AJ492657 and with the five sequences of L dumosa ob- yielded a single MPT of length 2437, 2796, and 4707, re- tained herein. On the other hand, for L theezans, L microdonta, spectively (not shown), whereas conditions 811 and 111 ren- I. brevicuspis AJ492662, and I. dumosa AJ492657, the range dered two and three MPTs of length 3617 and 1346, respec- of nucleotide variation detected was approximately under 5%. tively (not shown). The 50% majority rule consensus tree of In all cases, divergence values calculated among co-specific Fig. 3, condenses the information display in all MPT encoun- sequences obtained in this study were lower than 2%. The tered. Regarding sSA taxa, sequences of L argentina, I. bras- length, G + C content, free energy values of the RNA tran- iliensis, L brevicuspis, I. integerrima, I. microdonta, I. pseu- scripts, and the presence of a conserved structural motif were dobuxus, L taubertiana, L theezans, I. microdonta AJ492665, also evaluated in ITS 1 and ITS 2 of sSA and worldwide and L theezans AJ492666 appeared related in all the condi- distributed Ilex taxa (Table 5). For I. microdonta and I. thee- tions assayed. However, under conditions 111 and 241, se-

This content downloaded from 157.92.6.34 on Tue, 18 Dec 2018 17:13:49 UTC All use subject to https://about.jstor.org/terms February 2005] GOTTLIEB ET AL.-MOLECULAR ANALYSES OF SOUTH AMERICAN ILEX 357

86 Sta. Catarina[16F] Sta. Catarina [22 M 95 Rio GrandedoSul [167r Sta. Catarin [192 M]

52 Sta. Sta. Catarina Catarina[16] [19* i. theezans Sta. Catarina [22] Msiones [46] 93 69 Sta. Catarina (14] Sta. Catarin [169] Parana [65] 96Parana [73r 76 Parana [69]

95 Parna[6] 87 Parana[56] . integerrima Parana [62] Parana [72r 57 Msiones [221 (2)1 92 isiones [221 (3)1] Msiones [221 (1)]' 97 AltoPaana[226] sione[SN9fr] I. brasiliensis 66 _ Alto Parana [241] "83 Caaguazu Canindeyu [250] [230] Parana [59 F (2)] 82 Parara [59M] Parana [59 F (1)] Tucuman [109 M] Tauunan [110F]* Tucuman [110 MI Tucunan [111 F] Tucuman [112 M

SaltaSalta [207 (1)1 [207 (2)] argentina Tucurnan [111] 97 Tucuman [111 M Salta [540] Salta [541] 91 69 Msines [604 M (2)] Misiones[31] Misiones [3 F] Msiones [3 M 9 Misiones [6 F] Misiones [6 M 100 isiones[4M]

87 ii"nas[654M M nes604M [1[b (1)] Ibrevicuspis Misiones [604 F]

Msiones [2 M 58siones[5 F]* 99 siones[5 M (1)]

100 9 isiones [5 M MPsiones (2)] [595 F] Msiones [2 F]

64 Misiones [222(3)] Msiones[222 (4)] Msiones [222 (2)] Misiones [SN 6fr] Alto Parana [243 64 Afto Paana [227 F] 59 AltoParana[235] Alto Parana [235 F] 100 Misiones [8F] Misiones [8 M] 80 Misiones [610 F]

72 Msiones [44] 60M siones[611 Misiones [606 M]F] 61 siones[7 M (1)]. dumosa 89 Misiones [7 M (2)] Misiones[605 F] Msiones [607 M4 100 r 6siones[49 F] isiones [49 MK 9 Sta. Catarina [13] 99 91 Rio Grande do Sul [166] 5 Para na[55]

95 100 Parana[86] 59 arana[66]

Parana [90] 100 Parana [9] 76 Para na[76] 84 Parana [93] Rio Grande do Sul [168]

Sta.Misiones Catarina [50] [30] Parana [217] MAsiones[35] 79 MUsiones [220] Msiones [51] Alto Parana [224] 99 [1-74progeny] 51 [474 ] 59 [1-74F] Sta. Cataina [27 F] Rio Grande do Sul [138 F] 62 Sta.Catarina[28M] 65 Sta. Catarina[28 F] Parana [100 F]

Msionessiones [47] , I. paraguariensis Msiones [37 M* Parma [92]* 88 [SI-16 M] 96 [SI-19 F] [G-18 F] 59 [SI-67 F] Caaguazu [254] Misiones [SN8r] 100 sMones [34F] M9siones [34 M6 Misiones [164 Isiones [1 F] Canindeyu [228] 81 [SI-49 ] 66 Caaguazu [252] [Y-383 M6

Fig. 2. Maximum parsimony strict consensus tree of selected Ilex species derived from AFLP data. Numbers above branches indicate bootstrap values (1000 replicates). Accessions, as presented in Appendix 1 (see Supplemental Data accompanying the online version of this article), are indicated between brackets for clearness. F, stands for female; M, stands for male. The arrow indicates terminals belonging to I. dumosa var. guaranina. *, indicates specimens used for sequencing.

This content downloaded from 157.92.6.34 on Tue, 18 Dec 2018 17:13:49 UTC All use subject to https://about.jstor.org/terms 358 AMERICAN JOURNAL OF BOTANY [Vol. 92

TABLE 4. Intraspecific nucleotide divergence between Ilex species derived from Kimura two-parameter model (Kimura, 1980). Each individual sequences is identified by its corresponding GenBank accession number. Bolded sequences were obtained in this study.

Ilex argentina AY183482 AJ492655 L brasiliensis AY183490 AY183483 0.0015 0.0946 AJ492661 0.1426 AY183482 0.0927

L integerrima AY183477 AJ492664 L microdonta AY183484 AJ492665 AY183478 0.0078 0.1409 AY183485 0.0015 0.0271 AY183477 0.154 AY183484 0.0286

L paraguariensis AY183491 L pseudobuxus AY183486 AY183492 0.0092 AJ49260 0.1494

L brevicuspis AY183494 AY183496 AJ492662 AJ492663 AY183495 0.0185 0.0184 0.0262 0.1249 AY183494 0.0031 0.0068 0.1146 AY183496 0.005 0.109 AJ492662 0.1072

I. theezans AY183497 AY183493 AJ492666 AY183498 0.0127 0.0338 0.0515 AY183497 0.0159 0.0222 AY183493 0.0197

L dumosa AY183488 AY183481 AY183480 AY183479 AJ492657 AJ492656 AY183489 0.0046 0.0082 0.0113 0.0123 0.0117 0.215 AY183488 0.0049 0.0081 0.0092 0.0067 0.2114 AY183481 0.0086 0.0098 0.0108 0.2276 AY183480 0.008 0.0088 0.2059 AY183479 0.005 0.2143 AJ492657 0.2073

quence I. The phylogeneticbrevicuspis analysis restricted to active ITS sequenc- AJ492662 appeared to be related to our allele with es, comprised 46 ingroup sequencesa and two outgroup.low Strict Bremer support value (2 in condition 111 and 4 in consensus condition trees and single MPTs obtained under each com- 241, not shown). Neither the two sequences of L theezans bination of gap cost and Tv/Ts ratio cost are shown in norFigs. both sequences of I. microdonta formed monophyletic 4-18. The 50% majority rule consensus tree (Fig. groups.19, on the However, both L theezans appeared as sister in left), condenses40% the information display in ofall MPT encoun- the parameters sets, while both L microdonta appeared tered. The overall as topology of the treessister was not altered by the in 100% of the conditions. ITS sequences of L brasiliensis exclusion of the pseudogenes, although its resolution in- AJ492661, I. brevicuspis AJ492663, I. integerrima creased. RegardingAJ492664, sSA taxa, the clade composed by I. argen- and L pseudobuxus AJ492660 formed a clade under tina, I. brasiliensis, L brevicuspis, theL integerrima, L microdon- five analytical conditions explored. This clade received ta, L pseudobuxus,Bremer L taubertiana, and L theezans, including support values that ranged from 36 to 81, depending sequences I. microdonta AJ492665on and L theezans AJ492666the analytical conditions. These values represented the (hereafter, sSA clade) appearedsecond in 100% of the conditions as- to third highest values compared to all values obtained sayed. This clade received Bremer supportwith values that ranged each condition. Hence, these sequences never formed from 16 to 58, depending monophyletic on the analytical condition. These groups with their corresponding coun- terparts. values represent Although the third to fifth highest values compared to the placement of L dumosa AJ492656 varied with all values theobtained in each condition. Asparameters, before, the pairs (L it never appeared monophyletic with the remaining brasiliensis + I. theezans) and (I. integerrima + I. theezans L dumosa. On the other hand, I. dumosa AJ492657 appeared as sister to both L dumosa obtained here- AJ492666) were found throughout the explored weights, al- in, in 40% of the conditions, whereas it was placed as sister though both subclades received Bremer supports from 7 to 24, to L brevicuspis AJ492662 in 60% of the conditions. In these which represent the seventh to less than tenth highest values. latter cases, the Bremer support values were extremely high, Similarly, the subclade (L microdonta + I. pseudobuxus + I. ranging from 94 to 389. Ilex argentina AJ492655 never ap- microdonta AJ492665) appeared in 100% of the analyses, but peared related to its counterpart, while the placement of L its Bremer support was either the tenth highest value or lower. paraguariensis was highly dependent on parameter sets. As a result, The formation and placement of subcladeseveral (I. brevicuspis + I. features suggest that accessions L brasiliensis AJ492661, I. brevicuspis AJ492663, I. dumosa taubertiana) was dependent on analytical conditions; it was AJ492656, I. integerrima AJ492664, and I. pseudobuxus found in 80% of the conditions, and it formed a monophyletic AJ492660, could be considered as pseudogenes: the high nu- group with (I. microdonta + I. pseudobuxus + L microdonta cleotide divergence with co-specific data, the low G + C con- AJ492665) in only 53% of the conditions. The sister sequence tent, the departure from the conserved motif in ITS 1, the less to the sSA clade, in 67% of the parameters sets assayed, was thermodynamic stability in the RNA structure, and their basal represented solely by Asian L pedunculosa, but with low Bre- position on phylogenetic trees with respect to sSA counter- mer support values. Solely under analytical condition 811 (Fig. parts. Therefore, only L theezans AJ492666, I. microdonta 13), I. paraguariensis, both varieties of I. dumosa, and East AJ492665, and L argentina AJ492655, retrieved from Gen- Asian I. crenata and L mutchagara, appeared as sister group Bank, were further used as alternative functional ITS types. to the sSA clade, with a Bremer support of 29 (seventh highest

This content downloaded from 157.92.6.34 on Tue, 18 Dec 2018 17:13:49 UTC All use subject to https://about.jstor.org/terms February 2005] GOTTLIEB ET AL.-MOLECULAR ANALYSES OF SOUTH AMERICAN ILEX 359

TABLE 5. Results from ITS 1 and ITS 2 sequence comparisons.

ITS 1 ITS 2

AG AG AG G + C AG optimal suboptimal Consensus G + C optimal suboptimal Taxa* Length (%) (Kcal/mol) (Kcal/mol) Structures sequence** Length (%) (Kcal/mol) (Kcal/mol) Structures

SOUTH AMERICA Ilex argentina 260 59.23 -99.7 -94.8 8 B 240 56.67 -90.4 -87.6 4 L argentina AJ492655 262 59.54 -104.7 -100 9 D 248 59.68 -87.4 -83.3 4 I. brasiliensis 262 62.6 -103.8 -98.9 12 B 242 58.68 -104.3 1 I. brasiliensis AJ492661 259 50.97 -84.6 -80.7 6 J 245 53.06 -87.9 -83.6 10 I. brevicuspis 263 61.22 -104 -98.8 14 B 242 57.85 -95 -90.3 4 I. brevicuspis AJ492662 203 63.55 -72.9 - 1 B 248 58.06 -97.5 -93.7 4 I. brevicuspis AJ492663 259 49.03 -74.3 -70.6 18 L 242 54.55 -84.6 -80.7 10 I. dumosa var. dumosa 263 60.84 -112.1 -106.5 8 G 241 58.09 -94.1 -90.5 11 I. dumosa var. guaranina 263 61.6 -114.7 -109.6 7 G 241 58.51 -99.9 -95 6 L dumosa AJ492657 205 63.41 -76.6 -73.1 4 G 247 58.7 -102.1 -97.1 5 L dumosa AJ492656 251 45.42 -60.9 -58.2 7 K 238 52.52 -70.6 -67.4 11 I. integerrima 264 62.12 - 105.5 - 100.4 10 B 240 58.33 - 88.4 - 84.9 3 L integerrima AJ492664 259 51.74 -86.4 -82.4 13 J 244 52.46 -87.6 -83.5 11 I. microdonta 265 62.64 -106.2 -101 18 B 244 59.02 -100.7 -96.4 4 L microdonta AJ492665 265 61.89 -108.5 - 103.4 10 B 248 60.89 - 102.5 -97.7 6 I. paraguariensis 269 61.71 -114.9 -109.2 18 F 240 59.17 -101.1 -96.7 2 I. pseudobuxus 265 63.77 -110 -104.7 11 B 242 60.74 -96.5 -91.7 9 I. pseudobuxus AJ492660 258 48.45 -82.4 -78.4 9 M 245 51.84 -83.8 -80.2 9 I. taubertiana 263 58.56 -97 -92.6 11 A 244 57.38 -98 -95.6 3 I. theezans 263 60.08 -96.5 -92.5 3 B 242 57.85 -94.4 -90.5 4 L theezans AJ492666 264 62.88 - 110 -104.7 20 B 249 59.04 -107.9 -102.7 2

EAST ASIA I. mitis 260 61.54 - 106.8 - 102.1 7 B 244 56.56 -97.6 -97.4 2 I. beecheyii 262 59.92 -113.3 - 108.7 7 B 223 59.19 -89.3 -84.9 8 I. buergeri 261 58.62 -104.5 -99.5 7 C 238 59.24 -84.3 -80.1 16 I. cornuta 261 61.3 -115.7 -110 12 B 237 58.23 -78.7 -75.3 14 I. crenata 260 60.77 -103.6 -98.5 18 A 239 62.34 -96.7 -92.2 23 I. dimorphophylla 261 62.07 -118.8 -113.3 10 B 237 59.07 -87.9 -83.8 15 I. integra 263 56.27 -100.4 -95.4 12 H 224 58.04 -91 -86.9 12 I. kiusiana 262 59.16 -105.3 - 100.7 7 C 228 53.07 -70.2 -66.9 5 L latifolia 260 59.62 -104.5 -99.3 8 D 238 57.98 -87.9 -85.5 2 I. liukiuensis 256 59.38 -97.5 -92.8 16 E 237 57.38 -78.9 -75 13 I. macropoda 257 59.14 -101.2 -96.9 8 C 239 55.65 -82.2 -78.7 13 L maximowicziana 260 60.77 -104.2 -100.3 13 B 237 59.49 -88.1 -83.8 13 I. mertensii 263 60.08 -113.8 -108.4 9 A 223 59.19 -89.3 -84.9 8 L micrococca 259 59.07 -97.9 -93.7 14 C 243 59.67 - 100.8 -96 4 I. mutchagara 262 61.45 -107.6 -102.4 11 A 239 64.44 -101.1 -96.1 15 L. pedunculosa 263 60.46 -113.8 -108.4 8 B 242 57.44 -101.6 -97 4 I. percoriacea 263 60.08 -113.8 -108.4 9 A 223 59.19 -89.3 -84.9 8 I. rotunda 260 63.85 -110 -104.8 11 B 248 64.92 -113.6 -108.6 7 I. rugosa 256 57.81 -94.4 -90.4 15 B 238 55.04 -87.8 -83.9 7 I. serrata 258 59.69 -96.8 -92.3 14 D 240 55 -92.4 -88 10 I. warbugii 260 59.62 -106.1 -101.5 5 D 240 58.33 -83.6 -79.5 19 L yunannensis 258 59.69 -102.8 -97.7 14 D 241 57.68 -94.3 -90.8 4 I. anomala 258 54.26 -92.7 -88.3 9 I 231 55.41 -83.7 -81.9 5 L canariensis 261 65.13 -120.8 -115.2 11 B 247 63.16 -105.9 -101.6 10

NORTH AMERICA I. amelanchier 257 59.53 -100.3 -96.8 9 B 242 52.89 -83.9 -79.9 6 I. cassine 262 59.54 -104.7 -100 9 D 249 60.24 -91.7 -87.2 8 L collina 260 58.08 -107.6 -102.9 9 D 234 54.7 -89.1 -84.9 8 I. decidua 260 58.08 - 101.4 -97.3 10 D 234 56.84 -96.2 -92.1 5 I. glabra 261 58.62 -94.1 -89.5 17 D 243 58.02 -101.8 -96.9 14 L. mucronata 256 59.38 -102.6 -97.8 9 B 242 54.13 -87.6 -83.8 5 L opaca 259 58.69 -105.7 -100.5 14 D 243 59.26 -91.5 -88.1 3 I. vomitoria 260 58.46 -102.7 -98.6 9 D 234 56.84 -96.2 -92.1 5

Helwingia japonica 264 56.44 -93.8 -89.2 10 X 227 53.74 -86.3 -82.8 7 H. chinensis 265 56.23 -94.2 -89.6 15 X 227 53.3 -85.6 -82.1 6

* Sequences in bold were obtained in this study (see Materials and Methods, and Supplemental Data accompanying the online version of this article). Sequences identified with GenBank accession numbers are those of Manen et al. (2002). For remaining sequences, see Supplemental Data accompanying the online version of this article. ** , see Table 6.

This content downloaded from 157.92.6.34 on Tue, 18 Dec 2018 17:13:49 UTC All use subject to https://about.jstor.org/terms 360 AMERICAN JOURNAL OF BOTANY [Vol. 92

TABLE 6. ITS 1 motives of Liu and Schradl (1994) found in 55 ITS Ilex microdonta 1 Ilex sequences and outgroup sequences (for detailed distribution 416 6I - 6 Ilex pseudobuxus of each type, see Table 5). 10% I. microdonta AJ492665

Type 60/01 0R% 8 3I. Ilex theezans integerrima AJ492666 GGCGCTGTTTGCGCCAAGGAA A 60% Ilex taubertiana GGCGCTGTCTGCGCCAAGGAA B 6 Ilex brevicuspis GGCGCTATTTGCGCCAAGGAA C GGCGCTATCTGCGCCAAGGAA D 60% 100% fIlex-9 Ilex theezans brasiliensis GGCGCTATCTGCGGGAACCAT E GGCGCTGTCTGTGCCAAGGAA F Ilex argentina GGCGCTGTATGCGCCAAGGACCCAT G 60% 21 I. brasiliensis AJ492661 GGCGCTTTTTGCGCCAAGGAA H 8-24 ibOl I. integerrima AJ492664 GGCGCTATATGTGCCAAGGAA I 60% 36-81 1 L.. . Lpseudobuxus AJ492660 GGTGTTATCTGTGCCAAGGAA J 100% L brevicuspis AJ492663 AACACTATCTACACCAAGGAA K GGCGTTATCTGTACCAAGGAA L Ilex pedunculosa CGACGTTATCTGTGCAAAGAA M 4-16 Ilex macropoda GGCGCAAACAAGCGCCAAGGAA X 100o LIlex micrococca 60% 60% Ilex yunnanensis Note: Bold and italics = changed bases compared to consensus motif Ilex serrata of Liu and Schardl (1994) see Materials and Methods. o% . Ilex rotunda Ilex mitis 60% L Ilex canariensis value). In the analytical space explored, L dumosa varieties Ilex liukiuensis consistently yielded a clade with the second highest Bremer value (18-99). The varieties appeared more closely related to 8o00 Ilex warburgii East Asian I. crenata and L mutchagara (73% of the condi- 80% Ilex buergeri Ilex latifolia tions) than to other sSA taxa, excluding I. paraguariensis. The latter taxa appeared basal in relation to remaining ingroup taxa, 80% -612-1 Ilex cornuta in eight of the 15 parameters sets assayed, whereas under six 3-6 100oLf Ilex dimorphophylla analytical conditions, it was closely allied to L dumosa, I. 100% Ilex maximowicziana crenata and L mutchagara. 80% Ilex beecheyi 1026 Ilex percoriacea Concerning North American taxa, the ITS sequences of L collina, I. decidua, and L vomitoria were related across all the 0-50 1% Ilex mertensti 60% 100% Ilex integra analytical space, although with variable Bremer support (from Ilex kiusiana 3 to 54), i.e., values ranging from the fourth to less than 10th 118-366 Ilex rugosa highest. The pair of taxa I. cassine and I. opaca formed a clade with I. argentina AJ492655 in 100% of the conditions, with 10% o L dumosa AJ492657 Bremer support values within the third and seventh highest Lo I. brevicuspis AJ492662 60% Ilex dumosa dumosa (13-49). Ilex amelanchier and L mucronata appeared mono- so% Ilex dumosa guaranina phyletic in 80% of the conditions. The location of L glabra varied considerably under different analytical conditions. Like- f0Ilex decidua 4-46 60 Ilex vomitoria wise, the placement of Hawaiian L anomala relies on analyt- 100% _Ilex collina ical conditions, appearing basal to several East Asian taxa in 53% of the conditions. 9- . argentina AJ492655 14-32 100 . Ilex cassine The Asian taxa formed two stable groups in all of the anal- 100% Ilex opaca yses performed. One was composed of sequences of L beech- Ilex amelanchier eyii, L integra, L kiusiana, I. mertensii, and L percoriacea, 60o Ilex mucronata and had variable Bremer support values ranging from 5 to 87 Ilex crenata (second to ninth higher values). Instead, the other stable clade 60% L Ilex mutchagara formed by I. cornuta, I. dimorphophylla, and L maximowic- I. dumosa AJ492656 ziana had lower Bremer support values. Whereas, Asian taxa I. buergeri, I. latifolia, I. liukiuensis, and I. warburgii appeared Ilex glabra Ilex anomala monophyletic in 93% of the conditions. The position of the Ilex paraguariensis remaining Asian taxa (i.e., I. macropoda, I. micrococca, I. rotunda, L serrata, and I. yunnanensis) varied with alignment Helwingia chinensis parameters. ITS sequences of African L mitis and Macarone- H. japonica sian L canariensis formed a clade in 60% of the parameter Fig. 3. Fifty percent majority rule consensus tree derived from the sen- space. sitivity analysis of Ilex ITS sequences, including divergent copies. Numbers below branches indicate percentage of formation of the node. Numbers above Combined analysis-Additional analyses were carried out branches represent ranges of Bremer support values. Clades found throughout the parameter space explored are marked with thick lines. combining AFLP data for eight taxa and nucleotide sequences of active ITS alleles. A 50% majority rule consensus tree is shown as a summary of the sensitivity analysis (Fig. 19, on the right). The lengths of strict consensus trees and single MPT

This content downloaded from 157.92.6.34 on Tue, 18 Dec 2018 17:13:49 UTC All use subject to https://about.jstor.org/terms February 2005] GOTTLIEB ET AL.-MOLECULAR ANALYSES OF SOUTH AMERICAN ILEX 361

Figure 4. 2 Ilexpseudobums Figure 5. Il4xpseudabis L microdonta Gbk 1 1 4 E . microdorta Gbk 13 lexnicrdo~n ta 121 4 Ilex micrdonta 3 Ilex brevicurpis Ilex bmvicurpis lex taubertana Ilex taubertiam 7 Ilex bmriliensir 7 lekx brasiliensis 16 1- flexle theezaw 16 1 flex theezane 7 Ile.x ingerrima Ilex inegerrima 1. theezam Gbk I. theezanr Gbk Ilex argentina Ilex argetina 2 Ilexpeduruloa Ilex macropods 4 lex macropoda 4 lex amicrcocca 4 Ilex micrmoccca 4 Ilex amitis 1 lex yurmnensir 4 lIlex kx yi m ienris yumiti rs 4 lex serrata 1flex csanarieuis Ilex caiaenlwir 1 Ikefroue 2 Ilexurotcnda flex glabra flex glabra Ilex serrnda lex pedwulnc ia 4 Ilex likiuensis 9 Ilex crnta 11 x mutcqgara 31 flex Ilex buewcarbuii rgr fIlex dwmsa dumosa Ilex lkiblia 4 23 lex dmosa gmroina 4 Ilex cornuta lex lpraguarensris 2 2 Ilex dimorphophyla 4 Ilex maxhmowitzana 15 Ilex casine 115 1 Ilexargettina opaoca GBK 10 lexlex pecoriaceabeecheyi 7 Ilex dcidua 8 Ilex mertemii 3 3 Ilex vaoitoria flex collina 2 4 l ex ingram -lex khusiana Ilex rugosa 9 fl Ilex ,amelaxchier muconata 8 r Ilex amelanchier 5 lex liukiernsis Ilex mucronata 4 Ilex warl gii 6 Ilex anomdna flex buergeri 5 Ilekx decidua 2 flex rai osa 5 Ilex vonitoria 56 lex cornua 5 ex colina 10 I. argnti GBK 2 Jkxll 2 cmaxumowicniam 3 Iex dorphophy a 13 Ilex carine 11 flex beecheni 135 lex oplca llex pacauadens Il4 e mx percaea 4 Ilex crenata 16 I llexmwrtemii lex inegrag 3 llex mtchagara 18 Ilex dwnosa dumosa Il roagea Ilex dmosaguarina lex anomda Helwirgia chinensis r Helwirgia chinensir H. jconica H.jrovnica

7 lex mriliensis 11 Ilex bmriiensis 1 Ilex threzar SIlex theezars

SIlex iegerrima 2 lex 11 alxitcrregeimayagwenina 8141 L theezanr Gbk Figure 7. 23 L Narganr Gbk Figure 6. 1 k menfi L. peudobuxur 1413 .upleudohbaum 5 5 I mic rdonta Gbk 211 I. microdonta 4 Jlex microdata Gbk Ilex rmicedonata flex brevicurprs 4 Ilex brevicurpis lex taubertiaa 6 Ilex taubertiar flex pedwcwosa 1 lex perduncuosa 6 2 yk c ancrnsirs r1 4 lex carlensis Ilex rglbma Ilex yuwarmnis Ilex serrata 1 lex icrmcocca Ilex uacropodva SIlex micxrcocca 1 ael mitir 3 E Ilx cawine 2 2 J xchari 2 lk opaca Ilex percoriea 1 Ilex mucronata 5 flex rtendii 2 lkx serrata 1 lex mucgara l l69 aflxmenlaca Ilex glbra r ileJx beecheyi 3 2 Ilex Sm Ilex ckirLnautc 11 Ilx perc oriea Ilx blerelri S24 Ilex mertenrii IJ% instgra 4 lex corndaa Ilex khisiana Ilkx dimorphophyUa 11ex maximowicziana 2 3 Ie 3 war/rgIlex fabuiensis lex angosa Ilx buergier Ilex crenata 11 Ilexamutchagara 4 H Il acomwrma lejx dmosa dumosa 2 ItCeximorphophylla 22 lex ~umosa guraina 5 L Ilek maximowicziamna 5 Ilkxparagumensu - lex rugosa 8 Ilk amnomda S1 1 L IgentiGbk 8 Ilekx vomitoria Ilekx opaca 4 Ilex deckida 137 11ex decidua Ilex collina

Ilex colli~na 16l 7 Ilex crenata 8 r Ilex amelanchier 9 r Ilex liex mutchagara dumosa dumosa Ilex mucwonata 29 Iles dumosaguararina Ilex anomala Helwingia chiensis Helwigia chinenris H. japonica H. jqponica

Figs. 4-7. Results from the sensitivity analyses of Ilex taxa via direct optimization of ITS sequences. 4. Strict consensus tree of four most parsimonious trees (MPT) of length 1013, under parameters 111. 5. Strict consensus tree of eight MPT of length 1174, under parameters 121. 6. Strict consensus tree of 11 MPT of length 1297, under parameters 141. 7. Strict consensus tree of three MPT of length 1301, under parameters 211. Figure abbreviations: The transformation matrix stands for the step matrix used in each search (i.e., 241: gap cost = 2, Tv cost = 4, Ts cost = 1). Gbk, indicates sSA sequences retrieved from GenBank and discussed in the text.

This content downloaded from 157.92.6.34 on Tue, 18 Dec 2018 17:13:49 UTC All use subject to https://about.jstor.org/terms 362 AMERICAN JOURNAL OF BOTANY [Vol. 92

4 lex pseudobwua Figure 8. 7 micdntaGbk / 1ex Ilex microdonta peudobuxur- Figure 9. 7 7 micdonta 7 lex microdonta Ghk 221 lex brevicurpis 241 3 lex brevicupis 4 Ilex tauberdama 23 Ilex taubertiaw 11 Ilex brasiensis 12 Ilex brailenris 20 2 Ilex theezas 24 2 Ilex theezams 9 Ilex ilmegerrima Ilex integerrima I. theezapr Gbk 11 L theezas Gbk Ilex argentina flex argentna 7 Ilex macrpoda 10 Ilex macmpda Ilex microcca 4 5 Ilex micrococca 3 Ilex milis f5lex cmids 3 Ilex yumnenr!is S lex yunanensis Ilex serrata '4 flex serrata 5 Ilex rotunda flex rotunda 5 rlex glabra 4 lex kglabmr I. flex canariens Ilex Ilex canariensis pedmculosa Ilex pedmculosa 4 Ilex liAkiuensis 9 Ilex crenata 6 Ilex warburgii 71 Iex mautch~ara Ilex buergeri 2 21 lex doa 2doa 3 2 Ilex l klia 45 Ilex cdmragurwarna Ilex cornuta 4 Ilex parcqgaiemis 4 Ilex dmorphophylla 18 I arentina Ghk 8 Ilex maximowiczia'm 30 Idex carsine Ilex beecheyi 33Ilex opwa 15 lex percoriacea 12 Ilex tcidma 3 3 3 Ilex mertensii 6 5 lex vomiloria flex integra Ilkx collina 2 Ilex kiusiana 14 r Ilex amelanchier lex nrugaoa Ilex xanomala 1t flex 6 Ilex muconata linukienis 11 r Ile ameanchier 7 Ilex warburgii L Ilex mucronatac Ilex buergeri Ilex tlaolia 3 6 Ilex ecidza f7 Iex vomitoria flex mcornuna Ilex collina 8 Iex dmorphophylla 18 L. argentia Gbk 263 Ilk maximowicziam 17 Ilex carsine 3 Ilex beecheyi Ilex opaua Ilex percoriacea 4 Ilex crenata 3 3 Ilext mertemi 16 Ilex iuriana Ilex dumosa dumosa 221l 10 K Ikx muchcgara 29 Ilex cdunosarag msra IlxIlex i~ntegra hngaoa liex papguariendss Ilex anomala Helwi4gia chinensi H.jcpnicaHelwingia chinesir H. jqpoica

1 Ilex micllota

Figure 10. 8 &k peudobuxu l microdonta Gbk I. microdonta Gbk 6 15 Ilex integerrima Figure 11. 3 146 Ilex kxWe"doo integerrima 3 6 Ilex &evicuspis 33 3 Ilex brevicupis 411 L. ,hd ,, Gbk Ilex takbertidna 421 L ,he, Gbk Ilex tauberotira 4 16 Ilex brailiins L Ile5 t eemran Ilx theezaons 9 Ilex argentina 17 J lex aogentina ,9 1 k Ile brawszilns Ilex pedmculosa Ilex pedmculosa 9 r IIex canarienrir 3 4 Ilex canariensis 5 Ilex mitis llx miti Ilex ylaunenris 11 Ilwex macrpoda Ilex micrococca SI Ilex micjrocca Ilex macropoda Ilex yunanensis 1lex beecheyi S6 Ilex serrata Ilex pepcoriacea Ilex rounda Ilex mertensii Ilex gabra 4 x integra 4 Iex lidiuensis Slex iuria 8 Ilex warburgii 2 6 Ilex liukiwnsis 4 Ilex buergeri 3 Ilex warburgi 6 Ilex la&folia Ilex cornuta 2 5 Ilex IlJxbuergeri latolia 3 4 lex dmorphophyla 2Ilex corntaa lkx maximowicziaa 6 Ilex dmorphohla lex rugosa 6 rIlex maximowicziana 6 17 11 Il hceyi Ilex rugara Ilex percoriacea Ilex anomala 43 llex n mertensii 3 Il_ x integra 19 34 Ilex . argentiucarsine Gbk I/ex opia Ilex anomala 9 r Ilex ameanchier 15 Ilex 6 vomitorial/ex &cidua Ilex mucronarta 5 Ilex collina 5 r Ilex serrata 8 4 Ilex 17 Ilexdcidja wvomitoria Ilex rotunda Ilex collina 7 lex amelanc hier 7 1 25 Ilex cassine 33 I. argetina Gbk l ex mucrnata Ilex opaca 38 Ilex dnardma donora 6r Ilex crnata 199 Ilex glaba 278 7 Ilex murhura Ilex crenata &Ilx mmosra tdonsa Ilex donoraguoasaina Ilex Almnarguarrn I/ex nmntchugara Ilex ~peagunsris Ilex parnuariensis Helwirgia chinenvis Helwirgia chineNis H. japonica H. jaonica Figs. 8-11. Results from the sensitivity analyses of Ilex taxa via direct optimization of ITS sequences. 8. Single MPT of length 1533, obtained under parameter 221. 9. Strict consensus tree of two MPT of length 1842, under parameters 241. 10. Strict consensus tree of four MPT of length 1797, under parameters 411. 11. Strict consensus tree of six MPT of length 2083, under parameters 421.

This content downloaded from 157.92.6.34 on Tue, 18 Dec 2018 17:13:49 UTC All use subject to https://about.jstor.org/terms February 2005] GOTTLIEB ET AL.-MOLECULAR ANALYSES OF SOUTH AMERICAN ILEX 363

2 Ilex pseudobuxus 7 Ilex microdonta Figure Figure 12. 12.9 5 Ilexmicdonta L. micrdonta Gbk I microdonta Gbk 4 lex brevYCuspis Figure 13. 12 lex pseudob-ur 19 Ilex inlgerrima I. theezanr Gbk 441 1 Ilex tauberdana 17 Ilk brarilienis 811 1 lex taubertiana 34 2 Ilex theezans 22 Ilex brvicurpir Ilex integerrima 32 115 Ilex braliensis 13 . theezanr Gbk Ilex theezany 1 llex argentina Ilex argentina 12 Ilex macmlpoda 29 28 le crenata 8 Ilex micocca Ilex mutchagara 7 lex mitis 29 flex pLeaguaciensis 1 4 Ilex ylunanenrii 21 41 Ilex dumosadumosa 4, Ilexk serrata Ilex dumrosaguarcaina 14 lex romiunda 62 . argentna Gbk 11 lex glabra 40 Ilex carsine Ilex canariens Ilex opaca 4 lex pedmculosa 11 Ilex amelanchier 44 Ilex dumosa duosn a 14 lex mucromnta 7 lex donoraguaranina 9 1 lex glabra Ilex crenata Ilex rotunda Ilex muatchgara 8 Ilex canarienis 7 Ilex parcpq nieais S14 Ilex mitis Ilex pedunculosa 7 31 Ilex cassine 28 Ilex serrata Ilekx opaca Ilex yunnamenris 5 Ifex racida 10 1 lex macropoda Ilex omitoria flex micmeocca lex collina flex rugosa 21 r Ilx amelanchier 5 flex liukuensis I7ex mucrnata Ilex warburgii Ilex liriuensis Ilex buegeri 10 Ilex warburgii 12 flex nlia Ilex buergeri 13 1 lex mcornata 4 lex k lafolia 4 Ilx dimorphophyUa 1 Ilex cornuta Ilex maximowicziana 9 lex dmorphophy 4 Ilex colina 392 Ilex maximowicziana 5 30 Ilex decidua Ilex beecheyi Ilex vomitoria 23 lex percoriacea 29 Ilex beecheyi 3 5 Ilex mertensii 227 18 Ilexpercoriacea 2lex integra 36 Ilex tertemii Ilex kiuriana Ilex inlgra Ilex rugara Ilex k~idiana flex wanomala /lex anomcda Helwingia chinenis Helwirgia chinenis H japonica - H japonica

Ilex liku~ nris 4 Ilex microdona 10 lkx pseudobmus Figure 14. k berr"i Figure 15. ImicrodontaGbk fe x w comrna Figure 15 1 elxintgerrima 821 lex yrphophy 84153 kbx tm S Ilex brevicuspisrtiw L theezans Gbk 22 lex brasilicsis 24 flex ruglsa /&x maxmnowicziana 14 Ilextheezans 19 I9 x beecheyi lkx argentina Ile8 peroriacea lkx peduculosa Ilex mertensii 2I 3 2 integra 6 1 lkex 84 lexmacropoda micrococca //ekx hiuskm Ilex anomala Ilex ywm nreis 1316 Ilex caiensis 6 33 lex d cim flex miis 2 l x serrta I/ex vomitoria I1 k rotunda 3 Ilex mucrnata 20 x gabra ltx amekmchier 4 jler microdonta

2 I microdonta Gbk 1 Ilex mwlbari 11 Ile brevicuspis 3Ilex aburgeri 21 7 Ilex taubertna 11 lex corniga 213 lekx brarifensis 4 Ilea dmorphophs a LI theezanr Gbk 8 Ilex argeatina 13 13Ilex beeahei i Ilex poriacea 2 220 flr Jexcaiwiens Iflexpedunculoa Ilex mitis I64871" L/x ire gra 9 lex Ilex macropolda infegerrima 49 1 Ilex dumosar annla 6 fIlex micrococca 3 Ilex corinaa flex yunanemis 25 Ilex tcihar 3 10 r Ilexserrata Ile- x rotunda Illx vomitoria 31 69_ I.argee iGbk 3 Ilex caosine flerx opaca L lex opoca 364 3 lex cna flex mutchagara Ilex dumosadumosa l89 lex dumosaguarangina flex paraguariensis Helwingiachineneis F" jqpon,= H jiaponica

Figs. 12-15. Results from the sensitivity analyses of Ilex taxa via direct optimization of ITS sequences. 12. Strict consensus tree of two MPT of length 2540, under parameters 441. 13. Single MPT of length 2655, obtained under parameter 811. 14. Single MPT of length 3047, obtained under parameter 821. 15. Strict consensus tree of three MPT of length 3629, under parameters 841.

This content downloaded from 157.92.6.34 on Tue, 18 Dec 2018 17:13:49 UTC All use subject to https://about.jstor.org/terms 364 AMERICAN JOURNAL OF BOTANY [Vol. 92

Figure 16. 15 le xpxeubux, Figure 17. 1 e,,,,,~b., I microdonta Gbk 5 1 microdonta Gbk 20 flkx brasileknis

Ilkx taubertiana 1011 22 Jkx"it"egem 1021 23 fle twheems, 18 Ikx integerrima 19 Ilex bvicupis 51 L theezns Gbk lex bmvicuspis 38 16 lkex brasiliensis 12 Ilex tcubertidwa 16 ex theezans Ilex argentina Ilex argentina Ilex peiurulosa lex pedrnculosa 8 8r Ikx macropoda S16 Ilex macropoda 8 Ilex microcca 12 Ilex micrxocca 2 8 Ilex ywmm is Ilex yunna e nis 14r lex cmriensis 23 Ilex camcien*is 6 Ilex mitis LIex mitis Ilex serrata 9 Ilex serrna lex rotunda lex rotacda llex beecheyi 24 16 Ilex Ilex danosaguarminadmnosadumnosa 31 Ilex percoriace 8 16 lex panqguinries 22 Ilex mertensii 5 lex integra 21 Ilex carnata Slex kiana 7 Ilex mutchaoara 2 lex lilkideneis 33 Ilex caxsine. argentina Gbk lex warburgi 6 fIlex opaca llex buergeri Ilexk mnuronata 9 Ilex lkbfolia 11 lexcornuda lke awnekanchier 11 4 lex morphophylla 17 l/ex glaba 33 18 Ilx mmbln iclna l ex rugoxa 18 Ilex canmain SIlex buerieri 8 16 Ilex dckida 811Ilek coiaa 54 Ilex dcidua 14 ex prcoranuta 33 Ilex mitoria Ilex meresimowicia 68 fle mucronuaa Ilex iregra Slex glabra Ilex aelaunchier 1 4 I1k collina 14 82 1.agentinaGbk 16 I/kx Ilex icuavomitoria 28 Hl w I rex ca hine 11 flex op 79 aca klx dwumsaadumosa 36f18 flex lex dwnosaguaramina crncaa IlexHelwxabeechen)i aintega Jx mutchagara l23 lex kiumapoda jlexpcaagumensa HelwirLgia chinenmis Helwingia chinenris .jimvnica LIH japonica

2 3 lex bratilensii

Figure 18.0 Ilex micrcrca 1fle lhe aecsas 1041 Ikex ethcanchr9 2 IHx I. itegerrimatheezans Gbk 12 10 8 lex2'4124 ypseudmbus kle vicuspLa 10 Ilex 10 Ikx taberlana laubertiana f2lex prcarsiends Ilex argentina 12 4 Ilex pediuculsa 10 ( Ilex micrcocca

1Ilex llkxmitd omtilors a 58 r Ilekserr1ta 1 0323 Ilex Ilex nacoropdarotunda 101 4 Ilex 2 mucronraa Ilex aelachier lex ingira

15 10 Ilex abwgii ILEX buergeri 10 5 I4 leiukiensisIlex cornlia 4 v5 vox morphophyia Ilex mariowiczima lkex rugusa 15 1 1 I lex becheyi =Ile peloriacea S Ilex crtensii

Il iritegra 1 5. _ lex dnosaguarana Ilex coirina H23 - wir x homitoria jI/l apicida

r-H..iaponica

Figs. 16-18. Results from the sensitivity analyses of Ilex taxa via direct optimization of ITS sequences. 16. Single MPT of length 3023, obtained under parameter 1011. 17. Strict consensus tree of two MPT of length 3491, under parameters 1021. 18. Strict consensus tree of two MPT of length 4127, under parameters 1041.

This content downloaded from 157.92.6.34 on Tue, 18 Dec 2018 17:13:49 UTC All use subject to https://about.jstor.org/terms February 2005] GOTTLIEB ET AL.-MOLECULAR ANALYSES OF SOUTH AMERICAN ILEX 365

Subgenus- Senes Section -Subsecton flex microdlonta Euilex Aquifolimn Microdontae Ecrodontac e 4-20 530 Ile pseudbuEuilex Liopus4 EuilexLioprinuc Excelsae Laxae 100% .n microdonta Gbk 100% Euilex Aquifolium Microdontae Eumnicrodontae 53% Ilex brevicuspis - 3o Euilex Aquifolitun Microdontae Etuncrodontae 80% flex taubertiana 1 3 Euilex Lioprinus Excelsae Laxae 7-23 Ilex integertima - -Euilex Aquifolium Megalae Pedicellatae 16-58 15-60 7-24100 Ilex I. thleezans theezans 60o Euilex Gbk Aqulfolitun 9396 MegalaeEuilex Pedicellatae Aquifolium Megalae Pedicellatae 100% 100% 7-4 Ilex zabrasiliensis 73'o Euilex10000 Aquifolium Megalae Pedicellatae 67% flex argentina 6 7 Euilex Aquifolium Microdontae Repandae Ilex pedulnculaosa Euilex Li opinuus Excelsae mbellifomes flex canariellsis Euilex Lioprinus Cassinoides 53% 60% lex niti 00 Euilex qcifolium Lemurn ses 87% fIlex Ilex micrococcaO macropoda 8000 5P Broeuaipnus PrinoidesMicroccoca

53% fleX vnlnanleS flex is serrataEuilex Lioprinus Prinus Euprinus Cassinoides

67% Ilex glabraL flex Euilex Liopnnus Itunda Cassinoides 80 Etilex Liop.inus Excelsae Umbellifonn.es Ilex linkinensis Euilex Aquifclium Microdlontae Repandae

93% 93% Ilex belexfIlex buetgeriwarbugii 93%oEuiex i Euilex Aquifolium Aquifolium Microdontae Microd..ontae Repandae Repandae Ilex lackfolia Euilex Aquifoliurm Aqufolioides Insignes 73%4-14 fle cO7rta 73-01 Euilex Aquifulirn Aquifolioides CNvodontae 19 100- Ilex dimotphophlvlla -1oo% 1- Euilex Aquifbliun Aquifolioides COxyodontae 100% flex m raximnoicziana 100 Euilex Aquifoliurn Microdontae Repandae

60% Ilex beechevi8-31 Euilex8-32 Aquifolium Aquifolioides COxyodontae Ilex percoriacea Euilex Paltona Polyphyllae 100% 100% 60% 60- lex mertensii 5-87 uilex Aquifoliu m Microdontae Repandae

100%5-87 lex integra 100%Euilex Aquifoliurn Aquifolioides Oxyodontae Ilex ki-usiana Euilex Aquifoliumn Aqtuiflioides Cxyoduontae Ilex ntgosa Euilex Aquifoliun Rugosae Ilex anomnala Byronia EubTronia 9-82 I argentina Gbk 9-82 Euilex Aquifoliiun Microdontae Repandae 13-49 100% Ilex Cassine 100 13-49 Euilex Liopnnus Cassinoides

100% Ilex opaca 100% Ilex 5 Euilexdecidua Liopninus P inus P Cassinoides Moides 3-54 60% ex vomitoia 3- 47 Euilex Aquifolitun Microdontae Vomitoriae 100% Ilex collina 100%. Pnnus Pninoides

135- Ilex amelanchier Prinus Prinoides

100% m0% Ilex muctronata1 10094 o- Ilex crenata Euilex Paltoria Polyphyllae I67 Ilex mutchagara ' Euilex Paltoria Polyphyllae 73% 18-9 Ilex dumosa drosa -112 Euilex Paltoria Polyphyllae 100% Ilex dumosa guaranina 100% I Ilex paraguariensis Euilex Aqtifoliums Microdontae Repandae Helwingia chinensis

SH. aponica Fig. 19. Fifty percent majority rule consensus tree derived from the sensitivity analyses of ITS sequences alone (on the left) and from the combined analysis (on the right). Numbers below branches indicate percentage of formation of the node. Numbers above branches represent ranges of Bremer support values. Clades found throughout the parameter space explored are marked with thick lines. Gbk indicates sSA sequences downloaded from the GenBank and discussed in the text. Loesener's subgeneric disposition is presented. obtained, ranged from 1356 to 7479. When compared with the pology was congruent with topology of ITS sequence alone, ITS alone analysis, we observed that in 10/15 of the explored as were Bremer support values for stable nodes. Main differ- parameter space, the number of MPTs found remained un- ences were found within the sSA clade, though some Asian changed; in 4/10 situations, the number of MPT was lower and North American taxa showed alternative placement. With- than in the ITS analysis alone; and only in one case, was the in sSA, the relationship between I. brasiliensis, I. integerrima, number of MPT higher (condition 411 rendered 10 MPT when and I. theezans was resolved in the combined analysis as with combining and 4 in ITS alone). In general, the combined to- AFLP data alone. In contrast, the combination of data did not

This content downloaded from 157.92.6.34 on Tue, 18 Dec 2018 17:13:49 UTC All use subject to https://about.jstor.org/terms 366 AMERICAN JOURNAL OF BOTANY [Vol. 92 resolve the relationship was between among the lowest both values obtained. I. microdonta Similarly, Gauer and alleles (which appeared monophyletic Cavalli-Molina in (2000), 47% using of RAPD the markers, conditions) found low di- and I. pseudobuxus. The placement vergence among of naturalI. argentina populations of L paraguariensisappeared from here slightly less dependent on Brazil. parameters. It seems reasonable that Similarly, the arguments used the by those addition of AFLP data slightly stabilized authors (the occurrence the relationship of gene flow favored by of fruit-eating I. dumosa with I. crenata and I. mutchagara. birds or recent fragmentation Notwithstanding, caused by deforestation and/orthe data combination did not help reduction in stabilizing of natural forests) arethe responsible parameter-depen- for the results of dent position of I. paraguariensis. both studies. Neither the analysis of ITS data alone nor the combined analysis, yielded trees that ITS sequencecompletely studies-The nucleotide agreed divergence compar-with Loesener's taxonomic classification. isonAlthough revealed that in several the cases, trees divergence betweenderived se- from ITS analysis under parameter quence sets data obtained 221 in this(Fig. study and8), sequence 441 data down-(Fig. 12), and 821 (Fig. 14) had the highest loaded from number GenBank, far exceeded of the valuesclades reported in the concordance with current classification, literature, none usually belowof 5%the (Baldwin supraspecific et al., 1995). Results ranks appeared monophyletic. presented here, as a rule, do not support the view that the ITS sequences retrieved from GenBank (those of Manen et al., DISCUSSION 2002), and our ITS sequences reflect intraspecific polymorphism. In general, the sSA ITS sequences of Manen et al. AFLP analyses-The AFLP technique is highly (2002) efficient were highly divergent from those obtained herein, have for detecting polymorphisms of restriction fragments lower G by + C PCRcontent, and were less thermodynamically stable. amplification. The difficulty in identifying homologous Previously, mark- Manen et al. (2002) reported two of their South ers, the potential non-independence of bands and Americanthe dominant ITS sequences as paralogous (L dumosa AJ492656 nature of AFLP markers, are the main disadvantages and L ofbrevicuspis the AJ492663). In this study, we gathered solid technique (Mueller and Wolfenbarger, 1999). However, evidence AFLP indicating that those sequences might be treated as allows the simultaneous screening of many different pseudogene high-res- rDNA alleles, or as highly divergent paralogues olution markers randomly distributed throughout alleles the genome.(sensu Baldwin, 1995), or as class II paralogues (sensu The combinations of selective AFLP primers Buckleremployed et al., in 1997) or deep paralogues (sensu Bailey et al., this study allowed the recognition of several diagnostic 2003). Furthermore, bands we report three additional ITS sequences within each of the seven species assayed. These would from Manens'be use- original data set as pseudogene rDNA alleles, ful for molecular identification of the plants, particularly i.e., I. brasiliensis at AJ492661, L integerrima AJ492664, and L early developmental stages. A relationship between pseudobuxus the num- AJ492660. In contrast, L microdonta AJ492665 ber of AFLP bands scored for each species and its and ploidy L theezans level AJ492666 might represent alleles within dif- was detected, even though it was not statistically ferent tested active (Fig. ribosomal arrays. Our phylogenetic analysis in- 1). For instance, L argentina had the most complex cluding banding divergent ITS alleles, confirmed our results based on pattern, and it is the sole tetraploid species (2n =examining 80) detected factors correlated to functionality. Different labo- in the region (Barral et al., 1995). ratory methods were employed in Manen's work and ours, Our AFLP analysis revealed that individuals belongingconcerning mainly to the amplification primers (universal primer the same morpho-species formed coherent clades plus (with specific mod- primer used herein, and both universal primers erate to strong bootstrap support). Therefore, in seven Manen's) clades and the sequencing (direct sequencing of double were distinguished corresponding to each of the strandedspecies DNAsur- was employed herein vs. cloning by Manen et veyed. The species I. brasiliensis, I. integerrima, andal., 2002). I. thee- These might account for the differential recovery zans had a strongly supported relationship. The branchingof pseudogenes. order in which I. argentina and I. brevicuspis associate Our with results I. support the idea that phylogeneticists using nu- brasiliensis, I. integerrima, and I. theezans was not clear undoubt- ITS sequences should be aware of the potential risks of edly defined by the data; nevertheless, the monophyly downloading of the sequences from GenBank and using them for five species was very well-supported. Instead, the phylogenetic relationship analysis without prior corroboration of their or- to I. dumosa and I. paraguariensis of the aforementioned thologous spe-nature. As stated by Mayol and Rosell6 (2001), cies was not clearly established. without a detailed inspection of some basic features, errors in An important aspect of our analysis is the observation evolutionary that inferences could be easily overlooked. Incorpo- AFLP data do not confirm the traditional distinction, rating estab-ITS paralogues can distort the phylogenetic signal lished by Loesener (1901), between I. dumosa (Bucklervar. dumosa et al., 1997) yielding spurious clades. and I. dumosa var. guaranina, which was based on the inflorescence type. It seems possible that morphological Phylogenetic features inferences and sensitivity analyses--Direct used to distinguish and define the varieties (Loesener, optimization 1901; of DNA sequence data yields improved testing Giberti, 1994a) are not associated with the differences of phylogenetic ob- hypotheses because provisional homologies served herein. Instead, our results suggest that within are optimized I. du- for each cladogram individually, not a priori and mosa, Argentine and Paraguayan populations wereuniversally derived as with the static homologies of multiple align- from a common ancestral population a posteriori ofments the (Wheeler, diver- 2003). The sensitivity analysis as proposed gence from Brazilian population. A similar association by Wheeler was(1995), results in the knowledge of nodal stability, observed within I. brasiliensis. On the contrary, showingno geograph- how an alignment parameter affects the outcome tree ical pattern was noted within I. paraguariensis, for topology which (Giribet, in- 2003). Stability, logically different from ternal nodes appeared, in general, poorly supported support, and canre- only be defined by reference to some factors (Go- solved. For this species, calculated mean intraspecific loboff distanceet al., 2003). A group found in all the parameter space

This content downloaded from 157.92.6.34 on Tue, 18 Dec 2018 17:13:49 UTC All use subject to https://about.jstor.org/terms February 2005] GOTTLIEB ET AL.-MOLECULAR ANALYSES OF SOUTH AMERICAN ILEX 367 may nonetheless be poorly supported, and at the same and I.time, paraguariensis a were distantly related to the former spe- group with high "support" may be unstable to changescies. These in groupings were found to agree with the phenolog- some parameters (Goloboff et al., 2003). A drawback ical of data sen- obtained by Coelho and Mariath (1996). Two main sitivity analysis, is that the number of parameters strategies to be ex- of leaf longevity were distinguished by these au- plored is unbounded, and thus new or more gap : change thors, althoughcost some taxa were not included: subdeciduous ratios could always be added (Giribet and Wheeler, 1999).(for I. brevicuspis,Try- I. integerrima, I. theezans, I. microdonta, ing to generalize an evolutionary model for all taxa, and all L loci, taubertiana) and perennifoliar (for I. dumosa and I. and all regions may be simplistic; however, evaluating paraguariensis). two general parameters (gap/change ratio and Tv/Ts) is a Apartstarting from performing separate analyses of ITS sequence point (Maxmen et al., 2003). data and AFLPs, we carried out a combined phylogenetic anal- In this group of plants, analytical conditions rendered ysis. We dif- found that robust clades obtained with ITS data alone ferent hypotheses of relationships between species, were indicating not affected by the addition of AFLP data. Only sSA that the alignment parameters did influence the phylogenetic clade exhibited any changes, such as a better resolution of the inference. Notwithstanding, for the ITS sequence data relationship alone between I. brasiliensis, I. integerrima, and I thee- and in combination with AFLP data, the sensitivity zans. analysis Loesener (1901, 1908, 1942) and Baas (1975), based on showed that some clades were stable to the variation in ana- shared morphological characters (for instance, inflorescences lytical parameters and that those clades mainly agreed with in fasciculate corymbs, flower and fruit of large sizes, coria- geographical origin of the taxa involved. For instance, ceous the limbs with entire or pauci-aserrate leaf margins), pro- clade composed solely by southern South American taxa posed(the a close relationship among these taxa. Other clades sSA clade) namely, L argentina, I. brasiliensis, L brevicuspis, showed in the combined analysis an increment in their per- L integerrima, L microdonta, L pseudobuxus, L taubertiana, centage of formation throughout the analytical space. How- and L theezans was stable to parameter fluctuations (i.e., ever, ro- no major differences in support values were observed bust) in both analyses of ITS data alone and combined between (Fig. sequence data alone and the combined analysis. 19). It is noteworthy that neither the sequences of L micro- In spite of their taxonomic classification, I. pseudobuxus and donta nor those of I. theezans appeared monophyletic, which I. taubertiana failed to form a clade, suggesting that morpho- could be caused by ancestral polymorphism, introgression, logical or features shared by these taxa (the occurrence of long sampling errors. Another stable clade consisted of East Asianand slender flower peduncles; cfr. Edwin and Reitz, 1967) taxa L beecheyi, L integra, L kiusiana, L mertensii, andmight L have been developed by convergence. Similarly, I. mi- percoriacea. In addition, both stable clades were highly sup- crodonta and I. brevicuspis did not group in accordance with ported by the data in all conditions. On the other hand, theirEast taxonomic disposition, and thus morphological characters Asian clade formed by L cornuta, L dimorphophylla, thatand distinguishL these taxa from the rest could have arisen by maximowicziana and the clade concerning North American convergence. taxa I. collina, L decidua, and I. vomitoria were robust to Furthermore, our analyses revealed that changes in align- fluctuations but showed variable support values. Many ex- ment parameters resulted in alternative taxa being placed as amples exist in the literature showing cases like the ones pre- sister to the sSA clade and that Asian taxa were involved in sented here, decoupling support from stability in both direc- the majority of analytical conditions. The placement of I. par- tions (Giribet, 2003). aguariensis was likewise highly sensitive to alignment param- Results presented here contradict those of Manen et al. eter variation, appearing either related to the I. dumosa-I. cren- (2002). A possible explanation for the incongruence of both ata-I. mutchagara group, or as a third independent sSA line- studies would be based on the use of several highly divergent age (Figs. 7, 8, 10, 11, 14, 15, 17, and 18). With regard to I ITS sequences by Manen's team and on the analytical meth- dumosa, our results conflict with the work of Manen et al. odology employed. In this last respect, in the work of Manen (2002) in which L dumosa appeared closely related to I. ar- et al. (2002), sequences were aligned manually, gaps were gentina, with both nuclear and plastid sequence data. We treated as missing data and indels were coded as present or found no evidence of such a relationship, neither with AFLP absent, but without an explicit criterion for doing so. Manual data nor with ITS data. Instead, our results indicate a close alignments have been criticized for the lack of objectivity, lack relationship of I. argentina specimens to the remaining taxa of repeatability, and lack of criterion for exclusion data (e.g., of the sSA clade. It is worth mentioning that Manen's se- Gatesy et al., 1993; Giribet and Wheeler, 1999). In our study, quence of I. argentina AJ492655 appeared herein, robustly the analysis proceeds directly from sequence data to phylo- related and highly supported by the data to North American I. genetic reconstruction treating gaps as relevant evolutionary cassine and I. opaca. The fact that both ITS sequences of I. transformations (Giribet and Wheeler, 1999). argentina failed to form a monophyletic clade could be ex- Herein, majority rule consensus trees were presented be- plained by the coexistence within this tetraploid species of cause the lack of an appropriate external criterion (i.e., char- many types of active ITS copies due to an incomplete ho- acter congruence, as suggested by Wheeler, 1995 and Giribet mogenization or to a recent formation of the polyploid. Alter- and Wheeler, 1999) prevents us from choosing among com- natively, the possibility of mislabeled samples cannot be dis- peting phylogenetic hypotheses. Taxonomic congruence is carded. It would be interesting to reanalyze all the data set probably not a good external criterion when groups are defined used by Manen et al. (2002), including plastid sequence data, arbitrarily (Giribet and Wheeler, 1999). through optimization and sensitivity analyses, in order to eval- The AFLP topology (Fig. 2) and the majority rule consensus uate the stability of the phylogenetic hypotheses proposed. cladograms of ITS data alone and combined data (Fig. 19) Present results support the idea that current classification of I. were remarkably congruent in that they demonstrated the ex- argentina, closely allied to I. paraguariensis (Baas, 1975), istence of a group formed by L argentina, I. brasiliensis, I. may not be justified and that a re-classification should be ac- brevicuspis, L integerrima, and L theezans and that I. dumosa complished.

This content downloaded from 157.92.6.34 on Tue, 18 Dec 2018 17:13:49 UTC All use subject to https://about.jstor.org/terms 368 AMERICAN JOURNAL OF BOTANY [Vol. 92

Our phylogenetic analysis CERVERA, M.provides T., J. A. CABEZAS, J. C. SANCHA,no Eexamples MARTINEZ DE TODA, AND of complete agreement between J. M.stable MARTINEZ-ZAPATER. clades 1998. Application and of AFLPs the to the charac-taxonomical terization of grapevine Vitis vinifera L., genetic resources. A case of treatment of Loesener (1891, 1901, 1908, 1942). Other studies, study with accessions from Rioja (Spain). Theoretical and Applied Ge- such as those of Cu6noud netics 97: 51-59. et al. (2000) and Manen et al. (2002), indicated an inconsistency COELHO, G. C., AND J. E. A.of MARIATH. Loesener's 1996. Inflorescence morphology infrageneric of system with sequence data. Ilex L.This (Aquifoliaceae) inconsistency species from Rio Grande do Sul, Brazil. was Feddes also noted with Galle's system (not shown), Repertorium 107: 19-30. which basically summarized previous taxonomic works, CUINOUD, but P., M. A. DELusing PERO MARTINEZ, contradictory P.-A. LOIZEAU, R. SPICHIGER, S. criteria. The phylogenetic study of ANDREWS, Setoguchi AND J.-E MANEN. 2000.and Molecular Watanabe phylogeny and bioge- (2000) that ography of the genus Ilex L. (Aquifoliaceae). Annals of Botany, London comprises solely Asian Ilex 85: 111-122.taxa followed the taxonomic treatment of Yamazaki (1989). DE LAET,In J.,our AND W. C.opinion, WHEELER. 2003. Poy versiona discrepancy 3.0.11 (Wheeler, Glad- exists between Yamazaki's treatment stein and DeLaet, and 6 May 2003).sequence Computer software data and documentation presented by those authors. Contradictions distributed bybetween the authors at website, morphological http://research.amnh.org/scicomp/ and molecular data are somehow projects/poy.php. expected because, as explained by Systma (1990), reliance on DOWNIE, overall S. R., D. S. KATZ-DOWNIE, morphological AND K. SPALIK. 2000. A phylogeny similarity of as an approximation of phylogenetic Apiaceae tribe Scandiceae: relationships evidence from molecular ribosomal rather DNA in- than on ternal transcribed spacer sequences. American Journal of Botany 87: 76- shared derived character states 95. can generate misleading estimates. DOYLE, J. J., AND J. I. DAVIES. 1998. Homology in molecular phylogenetics: Our results draw attention to the need of a thorough taxo- a parsimony perspective. In D. E. Soltis, P. S. Soltis, and J. J. Doyle nomic revision of Ilex that would be based on a simultaneous [eds.], Molecular systematics of plants II. DNA sequencing, 101-103. cladistic analysis of all morphological and molecular charac- Kluwer, Boston, Massachusetts, USA. ters available, without disregarding the influence of analytical EDWIN, G., AND R. REITZ. 1967. Aquifoliaceas. In R. Reitz [ed.], Flora il- parameters on phylogenetic reconstruction. ustrada catarinense. Tipografia e Livraria Blumenauense S. A. Itajaf, Bra- zil. FELSENSTEIN, J. 1985. Confidence limits on phylogenies: an approach using LITERATURE CITED bootstrap. Evolution 39: 783-791. FILIP, R., P. LOPEZ, G. C. GIBERTI, J. CoussIO, AND G. FERRARO. 2001. ALBACH, D. C., P. S. SOLTIS, D. E. SOLTIS, AND R. G. OLMSTEAD. 2001. Phenolic compounds in seven South American Ilex species. Fitoterapia Phylogenetic analysis of asterids based on sequences of four genes. An- 72: 774-778. nals of the Missouri Botanical Garden 88: 163-212. GALLE, E C. 1997. Hollies: the genus Ilex. Timber Press, Portland, Oregon, ALICE, L. A., AND C. S. CAMPBELL. 1999. Phylogeny of Rubus (Rosaceae) USA. based on nuclear ribosomal DNA internal transcribed spacer region se- GATESY, J., R. DESALLE, AND W. C. WHEELER. 1993. Alignment of ambig- quences. American Journal of Botany 86: 81-97. uous nucleotide sites and the exclusion of systematic data. Molecular ALVAREZ, I., AND J. E WENDEL. 2003. Ribosomal ITS sequences and plant Phylogenetics and Evolution 2: 152-157. phylogenetic inference. Molecular Phylogenetics and Evolution 29: 417- GAUER, L., AND S. CAVALLI-MOLINA. 2000. Genetic variation in natural pop- 434. ulations of mate (Ilex paraguariensis A. St. Hil., Aquifoliaceae) using BAAS, P. 1975. Vegetative anatomy and the affinities of Aquifoliaceae, Sphen- RAPD markers. Heredity 84: 647-654. ostemon, Phelline and Oncotheca. Blumea 22: 311-407. GIBERTI, G. C. 1994a. Aquifoliaceae. In R. Spichiger and L. Ramella (Dirs.), BAILEY, C. D., T. G. CARR, S. A. HARRIS, AND C. E. HUGHES. 2003. Char- Flora del Paraguay, vol. 24, Editions des Conservatoire et Jardin bota- acterization of angiosperm nrDNA polymorphism, paralogy and pseu- niques de la Ville de Geneve, Missouri Botanical Garden, H61iographia dogenes. Molecular Phylogenetics and Evolution 29: 435-455. S A, Geneva, Switzerland. BALDWIN, B. G., M. J. SANDERSON, J. M. PORTER, M. E WOJCIECHOWSKI, GIBERTI, G. C. 1994b. Aquifoliaceae. In A. T Hunziker and A. M. Anton C. S. CAMPBELL, AND M. J. DONOGHUE. 1995. The ITS region of nu- [eds.], Flora fanerogaimica de Argentina, vol. 1, 1-8, Programa PRO- clear ribosomal DNA: a valuable source of evidence on angiosperm phy- FLORA (CONICET), Museo Botinico, IMBIV, C6rdoba, Estudio Sigma logeny. Annals of the Missouri Botanical Garden 82: 247-277. S.R.L., Buenos Aires, Argentina. BARRAL, G., L. POGGIO, AND G. C. GIBERTI. 1995. Chromosome numbers GIBERTI, G. C. 1998. Hallazgo de Ilex brasiliensis (Aquifoliaceae) en la Ar- and DNA content from Ilex argentina (Aquifoliaceae). Boletin de la So- gentina. Boletin de la Sociedad Argentina de Botdnica 33: 137-140. ciedad Argentina de Botdnica 30: 243-248. GIRIBET, G. 2001. Exploring the behavior of POY, a program for direct op- BASSAM, B. J., G. CAETANO-ANOLLiS, AND P M. GRESHOFF. 1991. Fast and timization of molecular data. Cladistics 17: S60-S70. sensitive silver staining of DNA in polyacrylamide gels. Analytical Bio- GIRIBET, G. 2003. Stability in phylogenetic formulations, and its relationship chemistry 196: 81-84. to nodal support. Systematic Biology 52: 554-564. BREMER, K. 1988. The limits of amino acid sequence data in angiosperm GIRIBET, G., AND C. RIBERA. 2000. A review of arthropod phylogeny: new phylogenetic reconstruction. Evolution 42: 795-803. BREMER, B., K. BREMER, N. HEIDARI, P. ERIXON, R. G. OLMSTEAD, A. A. data based on ribosomal DNA sequences and direct character optimization. Cladistics 16: 204-231. ANDERSBERG, M. KALLERSJO, AND E. BARKHORDARIAN. 2002. Phylo- genetics of asterids based on 3 coding and 3 non-coding chloroplast DNA GIRIBET, G., AND W. C. WHEELER. 1999. On gaps. Molecular Phylogenetics and Evolution 13: 132-143. markers and the utility of non-coding DNA at higher taxonomic levels. Molecular Phylogenetics and Evolution 24: 274-301. GLADSTEIN, G., AND W. C. WHEELER. 1997. POY: the optimization of align- BUCKLER, E. S., AND T. P. HOLTSFORD. 1996a. Zea systematics: ribosomal ment characters. American Museum of Natural History, New York, New ITS evidence. Molecular Biology and Evolution 13: 612-622. York, USA. BUCKLER, E. S., AND T P. HOLTSFORD. 1996b. Zea ribosomal repeat evolu- GOLOBOFF, P. A., J. S. FARRIS, M. KALLERSJO, B. OXELMAN, M. J. RAMIREZ, tion and substitution patterns. Molecular Biology and Evolution 13: 623- AND C. A. SZUMIK. 2003. Improvements to resampling measures of 632. group support. Cladistics 19: 324-332. BUCKLER, E. S., A. IPPOLITO, AND T. P. HOLTSFORD. 1997. The evolution of GRAY, A. 1856. Manual of the botany of the northern United States, 2nd ed., ribosomal DNA: divergent paralogues and phylogenetic implications. including Virginia, Kentucky, and all East of the Mississippi. G. P. Put- Genetics 145: 821-832. nam and Co., New York, New York, USA. CAMPBELL, C. S., M. E WOJCIECHOWSKI, B. G. BALDWIN, L. A. ALICE, HARTMANN, AND S., J. D. NASON, AND D. BHATTACHARYA. 2001. Extensive ri- M. J. DONOGHUE. 1997. Persistent nuclear ribosomal DNA sequence bosomal DNA genic variation in the columnar cactus Lophocereus. Jour- polymorphism in the Amelanchier agamic complex (Rosaceae). Molec- nal of Molecular Evolution 53: 124-134. ular Biology and Evolution 14: 81-90. HALL, T. A. 1999. BioEdit: a user-friendly biological sequence alignment

This content downloaded from 157.92.6.34 on Tue, 18 Dec 2018 17:13:49 UTC All use subject to https://about.jstor.org/terms February 2005] GOTTLIEB ET AL.-MOLECULAR ANALYSES OF SOUTH AMERICAN ILEX 369

editor and analysis program for Windows 95/98/NT Nucleic AcidsPAGE, Sym- R. D. M. 1998. TreeView 1.5.2. Computer software and documentation posium Series 41: 95-98. distributed by the author at website, http://taxonomy.zoology.gla.ac.uk/ HOOKER, J. D. 1862. Ilicineae. In G. Bentham and J. D. Hooker [eds.], rod/rod.html. Gen- era plantarum, vol. 1, 355-357. A Black, London, UK. POWELL, W., M. MORGANTE, C. ANDRE, M. HANGLEY, J. VOGEL, S. TINGEY, Hu, S.-Y. 1949. The genus Ilex in China. Journal of the Arnold Arboretum, AND A. RAFALSKI. 1996. The comparison of RFLP, RAPD, AFLP and Harvard University 30: 233-344, 348-387. SSR (microsatellites) markers for germplasm analysis. Molecular Breed- Hu, S.-Y. 1950. The genus Ilex in China. Journal of the Arnold Arboretum, ing 2: 225-238. Harvard University 31: 39-80, 214-240, 241-263. POWELL, M., V. SAVOLAINEN, P. CUEINOUD, J.-E MANEN, AND S. ANDREWS. KARDOLUS, J. P., H. J. VAN ECK, AND R. G. VAN DEN BERG. 1998. The 2000. The mountain holly (Nemopanthus mucronatus, Aquifoliaceae) potential of AFLPs in biosystematics: a first application in Solanum tax- revisited with molecular data. Kew Bulletin 55: 341-347. onomy (Solanaceae). Plant Systematics and Evolution 210: 87-103. REHDER, A. 1927. Manual of cultivated trees and shrubs hardy in North KIMURA, M. 1980. A simple method for estimating evolutionary rate of base America exclusive of the subtropical and warmer temperate regions. substitutions through comparative studies of nucleotide sequences. Jour- Macmillan, New York, New York, USA. nal of Molecular Evolution 16: 111-120. REISSEK, S. 1861. Celastraceae, Ilicineae et Rhamneae. In C. E P. von Martius KOOPMAN, W. J. M., M. J. ZEVENBERGER, AND R. G. VAN DEN BERG. 2001. [ed.], Flora brasiliensis 11, fasc. 28, Lipsiae ap. E Fleischer, Leipzig, Species relationships in Lactuca s. 1. (Lactuceae, Asteraceae) inferred Germany. from AFLP fingerprints. American Journal of Botany 88: 1881-1887. REN, N., AND M. P TIMKO. 2000. AFLP analysis of genetic polymorphism KROSSMANN, G. 1962. Handbuch der Laubgeholze in zwei Banden, band II. and evolutionary relationships among cultivated and wild Nicotiana spe- Paul Parey, Berlin, Germany. cies. Genome 44: 559-571. LIA, V. V., V. A. CONFALONIERI, C. I. COMAS, AND J. H. HUNZIKER. 2001. ROBINSON, J. P., AND S. A. HARRIS. 1999. Amplified fragment length poly- Molecular phylogeny of Larrea and its allies (Zygophyllaceae): reticulate morphism and microsatellites: a phylogenetics perspective. In E. M. Gil- evolution and the probable time of creosote bush arrival to North Amer- let [ed.], Which DNA marker for which purpose? Available online ica. Molecular Phylogenetics and Evolution 21: 309-320. http://webdoc.sub.gwdg.de/ebook/y/1999/whichmarker/index.htm. LILLO, M. 1911. Descripci6n de plantas nuevas pertenecientes a la flora Ar- SAVOLAINEN, V., M. W. CHASE, S. B. HOOT, C. M. MORTON, D. E. SOLTIS, gentina. Anales de la Sociedad Cientifica Argentina 72: 171-175. C. BAYER, M. E FAY, A. Y. DE BRUJIN, S. SULLIVAN, AND Y.-L. QIu. LINNAEUS, C. 1753. Species plantarum, vol. 1. Impensis Laurentii Salvii, 2000. Phylogenetics of flowering plants based on combined analysis of Stockholm, Sweden. plastid atpB and rbcL gene sequences. Systematic Biology 49: 306-362. LIU, J.-S., AND C. L. SCHARDL. 1994. A conserved sequence in internal SETOGUCHI, H., AND I. WATANABE. 2000. Intersectional gene flow between transcribed spacer 1 of plant nuclear rRNA genes. Plant Molecular Bi- insular endemics of Ilex (Aquifoliaceae) on the Bonin Islands and the ology 26: 775-778. Ryukyu Islands. American Journal of Botany 87: 793-810. LOBREAU-CALLEN, D. 1975. Les variations dimensionnelles du pollen du STATSOFT, INC. 1999. STATISTICA version 5.5 A for Windows [computer genre Ilex (Aquifoliaceae) et leur rapport avec le climat. Bulletin de la program manual]. StatSoft, Inc., Tulsa, Oklahoma, USA. Also available societd botanique de France. Collection Palynologie 122: 179-199. online http://www.statsoft.com/ LOESENER, T. 1891. Vorstudien zu einer Monographie der Aquifoliaceae. Ver- SWOFFORD, D. L. 1998. PAUP*: phylogenetic analysis using parsimony (*and handlungen den Botanischen Vereins fiir die Provinz Brandenburg und other methods), Version 4b10. Sinauer Associates, Sunderland, Massa- die angrezenden Liindern 33: 1-45. chusetts, USA. LOESENER, T. 1901. Monographia aquifoliacearum. I. Nova Acta Academiae SYSTMA, K. J. 1990. DNA and morphology: inference of plant phylogeny. Caesareae Leopoldina-Carolinae Germanicae Naturae Curiosorum 78. Trends in Ecology and Evolution 5: 104-110. LOESENER, T. 1908. Monographia aquifoliacearum. II. Nova Acta Academiae VILGALYS, R., AND M. HESTER. 1990. Rapid genetic identification and map- Caesareae Leopoldina-Carolinae Germanicae Naturae Curiosorum 89. ping of enzymatically amplified ribosomal DNA from several Crypto- LOESENER, T. 1942. Aquifoliaceae. In H. Harms and J. Mattfeld [eds.], Die coccus species. Journal of Bacteriology 172: 4238-4246. Natiirlichen Pflanzenfamilien 2nd ed., 36-86, Wilhelm Engelmann, Leip- Vos, P., R. HOGERS, M. BLEEKER, M. REIJANS, T. VAN DE LEE, M. HORNES, zig, Germany. A. FRITJERS, J. POT, J. PELEMAN, M. KUIPER, AND M. ZABEAU. 1995. MACE, E. S., C. G. GEBHARDT, AND R. N. LESTER. 1999. AFLP analysis of AFLP: a new technique for DNA fingerprinting. Nucleic Acids Research genetic relationship in the tribe Datureae (Solanaceae). Theoretical and 23: 4407-4414. Applied Genetics 99: 634-641. WHEELER, W. C. 1993. The triangle inequality and character analysis. Mo- MANEN, J.-E, M. C. BOULTER, AND Y. NACIRI-GRAVEN. 2002. The complex lecular Biology and Evolution 10: 707-712. history of the genus Ilex L. (Aquifoliaceae): evidence from the compar- WHEELER, W. C. 1995. Sequence alignment, parameter sensitivity and the ison of plastid and nuclear DNA sequences and from fossil data. Plant phylogenetic analysis of molecular data. Systematic Biology 44: 321-331. Systematics and Evolution 235: 79-98. WHEELER, W. C. 1996. Optimization alignment: the end of multiple sequence MARTIN, H. 1977. The history of Ilex (Aquifoliaceae) with special reference alignment in phylogenetics? Cladistics 12: 1-9. to Australia: evidence from pollen. Australian Journal of Botany 25: WHEELER, W. C. 2002. Optimization alignment: down, up, error and im- 655-673. provements. In R. DeSalle, G. Giribet, and W. C. Wheeler [eds.], Tech- MAYOL, M., AND J. A. ROSELLO. 2001. Why nuclear ribosomal DNA spacers niques in molecular systematics and evolution, 55-59, Birshdiuser, Basel, (ITS) tell different stories in Quercus. Molecular Phylogenetics and Evo- Switzerland. lution 19: 167-176. WHEELER, W. C. 2003. Implied alignment: a synapomorphy-based multiple- MAXMEN, A. B., E K. BURNETT, E. B. CUTLER, AND G. GIRIBET. 2003. sequence alignment method and its use in cladogram search. Cladistics Evolutionary relationships within the protostome phylum Sipuncula: a 19: 261-268. molecular analysis of ribosomal genes and histone H3 sequence data. WHEELER, W. C., AND D. GLADSTEIN. 2000. POY. Program and documen- Molecular Phylogenetics and Evolution 27: 489-503. tation, available at ftp://amnh.org/pub/molecular. MORRISON, D. A., AND J. T. ELLIS. 1997. Effects of nucleotide sequence WHITE, T J., T. BRUNS, S. LEE, AND J. TAYLOR. 1990. Amplification and alignment on phylogeny estimation: a case study of 18 S rDNA of Ap- direct sequencing of fungal ribosomal RNA genes for phylogenetics. In icomplexa. Molecular Biology and Evolution 14: 428-441. M. Innis, D. Gelfand, J. Sininsky, and T. White [eds.], PCR protocols. MUELLER, U. G., AND L. WOLFENBARGER. 1999. AFLP genotyping and fin- A guide to methods and applications, 315-322. Academic Press, San gerprinting. Trends in Ecology and Evolution 14: 389-394. Diego, California, USA. NEGI, M. S., A. SINGH, AND M. LAKSHMIKUMARAN. 2000. Genetic variation YAMAZAKI, T. 1989. Aquifoliaceae. In Y. Sotabe, H. Hara, S. Watari, and T. and relationship among and within Withania species as revealed by Tominari [eds.], Wild flowers of Japan, vol. 2, Woody plants, 26-32, AFLP markers. Genome 43: 975-980. Heibonska, Tobya (in Japanese). NEI, M., AND W.-H. LI. 1979. Mathematical model for studying genetic ZHU, var-J., M. D. GALE, S. QUARRIE, M. T. JACKSON, AND G. J. BYAN. 1998. iation in terms of restriction endonucleases. Proceedings of the National AFLP markers for the study of rice biodiversity. Theoretical and Applied Academy of Sciences, USA 76: S269-S273. Genetics 96: 602-611. OGDEN, H. T., AND M. E WHITING. 2003. The problem with "the Paleoptera ZUKER, M. 2003. Mfold web server for nucleic acid folding and hybridization Problem": sense and sensitivity. Cladistics 19: 432-442. prediction. Nucleic Acids Research 31: 3406-3414.

This content downloaded from 157.92.6.34 on Tue, 18 Dec 2018 17:13:49 UTC All use subject to https://about.jstor.org/terms