Springer-VerlagTokyohttp://www.springer.de102650918-94401618-0860Journal of Plant ResearchJ Plant ResLife Sciences22510.1007/s10265-005-0225-3
Phylogenetic relationships in Peniocereus (Cactaceae) inferred from plastid DNA sequence data
ORIGINAL ARTICLE
Received: March 22, 2005 / Accepted: July 4, 2005 / Published online: September 6, 2005The Botanical Society of Japan and Springer-Verlag2005
Salvador Arias • Teresa Terrazas • Hilda J. Arreola-Nava •
Monserrat Vázquez-Sánchez • Kenneth M. Cameron
S. Arias (*)Jardín Botánico, Instituto de Biología, Universidad Nacional Autónoma de México, 70-614 México, DF, 04510, MexicoTel. +52-55-56228988; Fax +52-55-56229046e-mail: [email protected]
T. Terrazas · H.J. Arreola-Nava · M. Vázquez-SánchezPrograma de Botánica, Colegio de Postgraduados, Montecillo, Estado de México, 56230, Mexico
K.M. CameronThe Lewis B & Dorothy Cullman Program for Molecular Systematics Studies, The New York Botanical Garden, Bronx, NY, 10458-5125, USA
Abstract The phylogenetic relationships of Peniocereus(Cactaceae) species were studied using parsimony analysesof DNA sequence data. The plastid rpl16 and trnL-F regionswere sequenced for 98 taxa including 17 species of Penio-cereus, representatives from all genera of tribe Pachycer-eeae, four genera of tribe Hylocereeae, as well as from threeadditional outgroup genera of tribes Calymmantheae,Notocacteae, and Trichocereeae. Phylogenetic analyses sup-port neither the monophyly of Peniocereus as currently cir-cumscribed, nor the monophyly of tribe Pachycereeae sincespecies of Peniocereus subgenus Pseudoacanthocereus areembedded within tribe Hylocereeae. Furthermore, theseresults show that the eight species of Peniocereus subgenusPeniocereus (Peniocereus sensu stricto) form a well-supported clade within subtribe Pachycereinae; P. serpenti-nus is also a member of this subtribe, but is sister toBergerocactus. Moreover, Nyctocereus should be resur-rected as a monotypic genus. Species of Peniocereus subge-nus Pseudoacanthocereus are positioned among species ofAcanthocereus within tribe Hylocereeae, indicating thatthey may be better classified within that genus. A numberof morphological and anatomical characters, especiallyrelated to the presence or absence of dimorphic branches,are discussed to support these relationships.
Peniocereus (A. Berger) Britton and Rose is a relativelysmall genus of Cactaceae consisting of approximately 18species in the subfamily Cactoideae. Representing >80% ofall cacti, with 1,500–1,800 species (Barthlott and Hunt1993), Cactoideae is the most diverse and species rich of thefour cactus subfamilies. Species of Peniocereus are generallycharacterized as semierect, low shrubs growing prostrate orclimbing over other plants. Their thin stems are either cylin-drical or possess a few ribs, their trunks are short but woody,and their tuberous roots are usually described as turnip-likeor dahlia-like. Peniocereus species are found from thesouthwestern United States and northern Mexico toHonduras and Nicaragua, but most of them are Mexican indistribution. With one notable exception in northeasternMexico, they principally inhabit arid regions and tropicaldeciduous or subdeciduous forests of the Pacific slope(Bravo-Hollis 1978; Anderson 2001).
Peniocereus was first described by Berger (1905) as amonotypic subgenus of Cereus Mill., represented by Cereusgreggii Engelm. as its type species. Britton and Rose (1909)later elevated the taxon to a genus status, as Peniocereusgreggii. Since then, several new species have been assignedto the genus, either as newly discovered species weredescribed or as transfers from other genera. Sánchez-Mejorada (1974a, b) presented a detailed taxonomic historyof the genus and, at the same time, constructed an infrage-neric classification for it. This system was based primarilyon characters related to the presence or absence ofdimorphic branches, and suggested that Peniocereus couldbe divided into two subgenera. One of these subgenera,Peniocereus subgenus Peniocereus, contained species withsimple ribbed stems, and with adult branches similar inmorphology to juvenile ones. The second subgenus, Penio-cereus subgenus Pseudoacanthocereus, contained thosespecies developing branches with several ribs when young,but either angular with few ribs or cylindrical withoutconspicuous ribs at all when adults. Complementaryinformation based on flower form (salverform or tubular-
Used Mac Distiller 5.0.x Job Options
This report was created automatically with help of the Adobe Acrobat Distiller addition "Distiller Secrets v1.0.5" from IMPRESSED GmbH. You can download this startup file for Distiller versions 4.0.5 and 5.0.x for free from http://www.impressed.de. GENERAL ---------------------------------------- File Options: Compatibility: PDF 1.2 Optimize For Fast Web View: Yes Embed Thumbnails: Yes Auto-Rotate Pages: No Distill From Page: 1 Distill To Page: All Pages Binding: Left Resolution: [ 2400 2400 ] dpi Paper Size: [ 594.994 784.916 ] Point COMPRESSION ---------------------------------------- Color Images: Downsampling: Yes Downsample Type: Bicubic Downsampling Downsample Resolution: 150 dpi Downsampling For Images Above: 225 dpi Compression: Yes Automatic Selection of Compression Type: Yes JPEG Quality: Medium Bits Per Pixel: As Original Bit Grayscale Images: Downsampling: Yes Downsample Type: Bicubic Downsampling Downsample Resolution: 150 dpi Downsampling For Images Above: 225 dpi Compression: Yes Automatic Selection of Compression Type: Yes JPEG Quality: Medium Bits Per Pixel: As Original Bit Monochrome Images: Downsampling: Yes Downsample Type: Bicubic Downsampling Downsample Resolution: 600 dpi Downsampling For Images Above: 900 dpi Compression: Yes Compression Type: CCITT CCITT Group: 4 Anti-Alias To Gray: No Compress Text and Line Art: Yes FONTS ---------------------------------------- Embed All Fonts: Yes Subset Embedded Fonts: No When Embedding Fails: Warn and Continue Embedding: Always Embed: [ ] Never Embed: [ ] COLOR ---------------------------------------- Color Management Policies: Color Conversion Strategy: Convert All Colors to sRGB Intent: Default Working Spaces: Grayscale ICC Profile: RGB ICC Profile: sRGB IEC61966-2.1 CMYK ICC Profile: U.S. Web Coated (SWOP) v2 Device-Dependent Data: Preserve Overprint Settings: Yes Preserve Under Color Removal and Black Generation: Yes Transfer Functions: Apply Preserve Halftone Information: Yes ADVANCED ---------------------------------------- Options: Use Prologue.ps and Epilogue.ps: No Allow PostScript File To Override Job Options: Yes Preserve Level 2 copypage Semantics: Yes Save Portable Job Ticket Inside PDF File: No Illustrator Overprint Mode: Yes Convert Gradients To Smooth Shades: No ASCII Format: No Document Structuring Conventions (DSC): Process DSC Comments: No OTHERS ---------------------------------------- Distiller Core Version: 5000 Use ZIP Compression: Yes Deactivate Optimization: No Image Memory: 524288 Byte Anti-Alias Color Images: No Anti-Alias Grayscale Images: No Convert Images (< 257 Colors) To Indexed Color Space: Yes sRGB ICC Profile: sRGB IEC61966-2.1 END OF REPORT ---------------------------------------- IMPRESSED GmbH Bahrenfelder Chaussee 49 22761 Hamburg, Germany Tel. +49 40 897189-0 Fax +49 40 897189-71 Email: [email protected] Web: www.impressed.de
funnelform), areole density of the floral tube, and seeds(periclinal cells smooth or striated) was presented byBravo-Hollis (1978) in support of recognizing these twosubgenera as distinct.
More recently, there has been a debate in the literatureas to how many species are circumscribed by the genusPeniocereus. The exact phylogenetic position of the genuseither as a member of tribe Hylocereeae (Bravo-Hollis1978), Echinocereeae (Buxbaum 1975), or Pachycereeae(Anderson 2001) has also been debated. In fact, the mono-phyly of Peniocereus is also not certain since there are anumber of species that have either been included in it orreassigned to smaller segregated genera such as CullmanniaDistefano, Neoevansia W.T. Marshall, and Nyctocereus (A.Berger) Britton and Rose. For example, the species thatBuxbaum (1975) recognized as Peniocereus viperinus,because it has a floral tube, nectar chamber, and smooth andlarge seeds similar to other species Peniocereus, has had aconfused taxonomic history. It was originally classified as aspecies of Wilcoxia, but then elevated to monotypic genusstatus as Cullmannia viperina by Distefano (1956) becauseof its diurnal salverform flowers and smooth seeds. Like-wise, Peniocereus striatus and P. zopilotensis have been asource of taxonomic uncertainty. They were originally clas-sified as species of Neoevansia, together with a third species,N. lazaro-cardenasii, by Marshall (1941) on the basis ofhaving slender branches (less than 10 mm wide), and verylow and flattened ribs (Sánchez-Mejorada 1973). Buxbaum(1975) rejected this treatment and transferred both N. stri-ata and N. zopilotensis into Peniocereus because of theirextended perianths, floral tube form, and seed size. In a finalexample, Nyctocereus was erected as a genus by Britton andRose (1909) for a group of species with large funnelformflowers and elongated filaments. According to Bravo-Hollis(1978), Nyctocereus comprises four well-documented spe-cies. In the past decade, however, one of these species (N.chontalensis) has been transferred to Selenicereus, and theremaining three (N. serpentinus, N. hirschtianus, and N.oaxacensis) have been transferred into Peniocereus, by anad hoc working party of the International Organization forSucculent Plant Study (IOS, Hunt and Taylor 1991).Although the monophyly of Peniocereus has not beenbrought into question formally, Wallace’s (2002) cladogramsbased on rbcL and rpl16 nucleotide sequence data showedthat part of the Peniocereus/Nyctocereus lineage was alliedto subtribe Pachycereinae (tribe Pachycereeae), whereasanother part was associated with tribe Hylocereeae.
In terms of the position of Peniocereus within subfamilyCactoideae, Buxbaum (1958) suggested that the genus is abasal lineage of tribe Hylocereeae, most closely related toNyctocereus and Brachycereus on account of their large,white, nocturnal, flowers. Barthlott and Hunt (1993), on theother hand, considered Peniocereus to be a member of tribeEchinocereeae, but presented no evidence for this opinion,whereas Anderson (2001) proposed the inclusion of Penio-cereus into tribe Pachycereeae. This latter placement wasrejected by Terrazas and Arias (2002), who analyzed a setof anatomical and morphological characters for members ofCactoideae, and found no structural evidence to support arelationship of Peniocereus to Pachycereeae.
In order to assess the monophyly of Peniocereus and itsrelationships to tribes Pachycereeae and Hylocereeae, weconducted large-scale molecular phylogenetic analyses ofCactoideae that included 21 samples of Peniocereusrepresenting species from both subgenera recognized bySánchez-Mejorada (1974a, b). These analyses also includedsamples of those genera historically associated with Penio-cereus as discussed above (i.e., Cullmannia, Neoevansia, andNyctocereus), along with taxa identified by previous phylo-genetic analyses to be part of this large clade of cacti.Sequence data from the chloroplast rpl16 intron, trn intronand the trnL-F intergenic spacer were collected for all ofthese taxa since these gene regions have proven informativein reconstructing phylogenetic relationships among Cacta-ceae in other studies (Arias et al. 2003).
Materials and methods
Taxon sampling
A total of 17 Peniocereus species from the 18 recognized byAnderson (2001), plus one undescribed, and one cultivatedspecies (P. marnierianus Backeb.) were sampled (Table 1).In addition, representative species from every genus of tribePachycereeae and several from tribe Hylocereeae (sensuAnderson 2001) were also sequenced. The outgroup taxafor this phylogenetic study included three genera from thetribes Calymmantheae (i.e., Calymmanthium substerile),Notocacteae (i.e., Eulychnia castanea), and Trichocereeae(i.e., Harrisia earlei). These three taxa have been shown torepresent clades distantly related to Pachycereeae andHylocereeae in more broadly sampled molecular phyloge-
Table 1. Classification of Peniocereus according to four authors
Taxa Britton and Rose (1920) Bravo-Hollis (1978) Gibson and Nobel (1986) Anderson (2001)
Cullmannia – Transferred to Wilcoxia Transferred to Peniocereus Transferred to PeniocereusNeoevansia – 2 spp. Transferred to Peniocereus Transferred to PeniocereusNyctocereus 5 spp. 6 spp. ca. 6 spp. Transferred to Peniocereus
and SelenicereusWilcoxia 4 spp. 6 spp. Transferred to Echinocereus Transferred to Echinocereus
netic studies of the Cactaceae (Nyffeler 2002; Wallace2002).
Most of the tissue from the sampled taxa was collectedfrom the epidermis of fresh stems in native populations, andthen immediately dried in silica gel. Sequences from 43members of the tribe Pachycereeae used in this study werepreviously published (Arias et al. 2003; Arreola-Nava et al.unpublished), and thus downloaded from GenBank. A fewsamples were obtained from cultivated material in the livingcollections of the Desert Botanical Garden (DES) andHuntington Botanical Garden (HBG), as indicated inTable 2, which includes complete voucher information andGenBank accession numbers.
DNA extraction, amplification and sequencing
Total DNA was extracted using the Dneasy kit (QIAGEN,Valencia, CA, USA) according to the manufacturer’sinstructions. Aliquots were then stored at −20°C. Targetregions were amplified in 25 µl volumes using standardpolymerase chain reaction (PCR) protocols that includedthe addition of 10X buffer, dNTPs, BSA, and betaine. Weachieved the highest quality amplifications of trnL-F andrpl16 using the following thermal cycling profile for30 cycles: 95°C for 50 s, 60°C for 50 s, 72°C for 90 s. ThetrnL-F region, including trnL(UAA) intron, trnL exon 2 andintergenic spacer between trnL and trnF, was amplifiedusing primers c and f as designed by Taberlet et al. (1991).These same primers, along with primers d and e (Taberletet al. 1991), were used for cycle sequencing. The rpl16region, including a part of rpl16 exon 2, rpl16 intron, rpl16exon 1, and the intergenic spacer between rpl16-rps3, wasamplified as a unit using primers rpl16-1216F and rps3-42R(or sometimes rpl16-18R; see Asmussen 1999). These sameprimers along with rpl16-584F and rpl16-957F were used forcycle sequencing. In all cases, the resulting PCR productswere purified using QIAquick spin columns (Qiagen,Valencia, CA, USA) according to the manufacturer’s pro-tocols. Cycle sequencing reactions (32 cycles: 96°C for 10 s,50°C for 5 s, 60°C for 3 min) were performed using theBigDye Terminator Mix. These reactions resulted incomplete forward and reverse strands of the target regionsfor nearly all sequences. Centri-Sep Sephadex columns(Princeton Separations, Adelphia, NJ, USA) were usedaccording to the manufacturer’s instructions to removeexcess dye terminators and primer from the cycle sequenc-ing products. The products were subsequently dehydratedin a vacuum centrifuge, resuspended in a mixture of forma-mide and loading dye, and loaded onto a 5% denaturingpolyacrylamide gel. Samples were run for 9 h on an AppliedBiosystems ABI 377XL automated DNA sequencer, andthe resulting chromatograms were edited using Sequencher4.1 (GeneCode).
Phylogenetic analysis
All sequences were initially aligned with CLUSTAL X(Thompson et al. 1997) using the default conditions for gap
opening penalty (value of 10) and gap extension penalty(value of 0.05). The aligned sequences were then importedinto MacClade 4.0 (Maddison and Maddison 2000) forminor manual adjustment. The resulting alignments wereanalyzed using the parsimony criterion in PAUP* 4.0b10(Swofford 2002), with unordered and unweighted characterstates. Gaps were treated in two different ways: (1) codedas missing data, and (2) added to the aligned sequences asindel-based characters. GapCoder (Young and Healy 2003)was used to automate this procedure, following the simplegap coding method of Simmons and Ochoterena (2000).Parsimony analyses of the separate and combined trnL-Fand rpl16 data involved heuristic search strategies of 1,000random addition sequence replicates using equal weightsand tree bisection-reconnection (TBR) branch swapping,but keeping only ten trees per replicate in order to discoverpossible islands of maximum parsimony. Resulting treesfrom this first phase of analysis were then used as startingtrees in a second phase of analysis in which all equallyparsimonious trees were collected (MULTREES in effect).Support values for the relationships discovered in eachanalysis were calculated by performing both bootstrap (bts)and jackknife (jck) analyses. Ten thousand heuristic searchreplicates were executed using the TBR branch-swappingalgorithm. To assess the level of congruence between thedata sets, we employed the incongruence length difference(ILD) test (Farris et al. 1995), implemented in PAUP* asthe partition homogeneity test. One hundred heuristicsearch replicates were performed with all characters equallyweighted and TBR branch swapping.
Results
trnL-F
Sequence length for the trnL intron plus trnL-F spacerranges from 767 bp in Peniocereus cuixmalensis to 1,151 bpin Polaskia chichipe. Aligned sequence length for the trnL-F data set is 1,315 bp, but a total of 1,201 characters wereused for the analysis after trimming the ends of the data setto avoid problems with considerable missing data. Figure 1shows the strict consensus of 885 equally parsimonious treesrecovered in the analysis that included 61 gaps coded as 0or 1. A summary of the matrix and tree statistics for thisgene region is presented in Table 3.
Five major clades are evident, and two of these containPeniocereus species. Peniocereus subgenus Pseudoacantho-cereus (sensu Sánchez-Mejorada 1974b) plus Acanthocereusand other Hylocereeae samples are supported as a strongclade (94% bts/89% jck) herein referred to as clade I.Within this clade Hylocereus, Weberocereus, and Selen-icereus, including S. chontalensis, represent a strongly sup-ported subclade (97% bts/92% jck). One species ofPeniocereus, P. hirschtianus, and three species of Acantho-cereus form a second monophyletic subclade (62% bts/56%jck), and they are sister to a third well-supported subclade(84% bts/74% jck) containing ten species of Peniocereus
Table 2. List of taxa investigated in this study, with voucher specimens (CHAPA Herbario Colegio de Postgraduados, DES Desert BotanicalGarden, HBG Huntington Botanical Garden, IBUG Herbario, Instituto de Botánica, Universidad de Guadalajara, MEXU Herbario Nacionalde México, NY New York Botanical Garden, OAX Herbario, CIIDIR Oaxaca, Instituto Politécnico Nacional), and collection locality (CU Cuba,GT Guatemala, MX Mexico, PR Puerto Rico, NI Nicaragua, US United States)
Tribe PachycereeaeAcanthocereus chiapensis Bravo Guzmán 949 (MEXU), Chiapas, MX DQ099916 DQ099985A. griseus Backeb. Arias 1183 (MEXU), Chiapas, MX DQ099917 DQ099986A. horridus Britton and Rose Arias 1177 (MEXU), GT DQ099918 DQ099987A. occidentalis-SIN Britton and Rose Arreola 1393 (IBUG), Sinaloa, MX DQ099919 DQ099988A. occidentalis-CHI Britton and Rose Guzmán 1002 (MEXU), Chiapas, MX DQ099920 DQ099989A. tetragonus-NL (L.) Hummelinck Arias 1450 (MEXU), Nuevo León, MX DQ099921 DQ099990A. tetragonus-VER (L.) Hummelinck Hernández sn. (IBUG), Veracruz, MX, DQ099922 DQ099991Backebergia militaris (Audot) Sánchez-Mej. Arias 1339 (CHAPA), Michoacán, MX AY181628 AY181609Bergerocactus emoryi (Engelm.) Britton and Rose Arias 1307 (MEXU), Baja Cal., MX DQ099925 DQ099994Carnegiea gigantea (Engelm.) Britton and Rose NYBG-sn., cult., Arizona, US AY181619 AY181591Cephalocereus apicicephalium Dawson Terrazas 632 (CHAPA), Oaxaca, MX DQ099927 DQ099996C. columna-trajani (Karw. ex Pfeiff.) Schum. Arias 1377 (CHAPA), Puebla, MX AY181648 AY181599C. nizandensis (Bravo and MacDoug.) Buxb. Terrazas 633 (CHAPA), Oaxaca, MX DQ099928 DQ099997C. senilis (Haw.) Pfeiff. Terrazas 529 (CHAPA), Hidalgo, MX AY181638 AY181616Dendrocereus nudiflorus (Engelm. ex Sauvalle)
Britton sunamb RoseAreces 57 (NY), Guantanamo, CU DQ099929 DQ099998
Echinocereus enneacanthus Engelm. subsp.brevispinus (W.O. Moore) N.P. Taylor
Arias 1453 (MEXU), Nuevo León, MX DQ099931 DQ100000
E. leucanthus N.P. Taylor Terrazas 410 (MEXU), Sonora, MX DQ099932 DQ100001E. parkeri N.P. Taylor subsp. parkeri Arias 1437 (MEXU), Nuevo León, MX DQ099933 DQ100002E. pectinatus (Scheidw.) Engelm. Guzmán 1234 (MEXU), San Luis Potosí, MX DQ099934 DQ100003E. poselgeri Lem. Arias 1452 (MEXU), Nuevo León, MX DQ099935 DQ100004E. schmollii (Weingart) N.P. Taylor Arias 91 (MEXU), Querétaro, MX DQ099936 DQ100005E. triglochidiathus Engelm. subsp. triglochidiathus Earle sn. (DES), New Mexico, US DQ099937 DQ100006Escontria chiotilla (Weber ex Schum.) Rose Terrazas 370 (CHAPA), Puebla, MX AY181622 AY181608Leptocereus quadricostatus (Bello) Britton and Rose Arias 1464 (MEXU), Cabo Rojo, PR DQ099942 DQ100011Lophocereus gatesii M.E. Jones Hamann s.n., cult., Baja Cal. Sur, MX AY181637 AY181601L. schottii (Engelm.) Britton and Rose Terrazas 474 (CHAPA), Baja Cal. Sur, MX AY181620 AY181613Myrtillocactus geometrizans (Martius) Console Terrazas 557 (CHAPA) Querétaro DQ099943 DQ100012M. schenckii (J. Purpus) Britton and Rose Terrazas 500 (CHAPA), Puebla, MX AY181633 AY181607Neobuxbaumia euphorbioides (Haw.) Buxb. ex Bravo Hamann s.n., cult., Tamaulipas, MX AY181635 AY181595N. macrocephala (Weber ex Schum.) Dawson Arias 1048 (MEXU), Puebla, MX DQ099944 DQ100013N. mezcalaensis (Bravo) Backeb. Terrazas 533 (CHAPA), Guerrero, MX AY181645 AY181600N. multiareolata (Dawson) Bravo et al. Terrazas 531 (CHAPA), Guerrero, MX AY181644 AY181594N. polylopha (DC.) Backeb. Terrazas 530 (CHAPA), Hidalgo, MX AY181644 AY191597N. scoparia (Poselger) Backeb. Hamann s.n., cult., Oaxaca, MX AY181625 AY181596Pachycereus grandis Rose Terrazas 534 (CHAPA), Puebla, MX AY181646 AY181605Pachycereus hollianus (Weber) Buxb. Arias 1373 (CHAPA), Puebla, MX AY181623 AY181603Pachycereus lepidanthus (Eichlam) Britton and Rose Cseh s.n., cult., Guatemala, GT AY181627 AY181618Pachycereus marginatus (DC.) Britton and Rose Arias 1372 (CHAPA), Puebla, MX AY181627 AY181618P. pecten-aboriginum (Engelm.) Britton and Rose Terrazas 535 (CHAPA), Guerrero, MX AY181624 AY181515P. pringlei (Watson) Britton and Rose Arias 1348 (CHAPA), Baja Cal. Sur, MX AY181642 AY181589P. tepamo S. Gama and S. Arias Arias 1150 (MEXU), Michoacán, MX AY181647 AY181593P. weberi (Coulter) Backeb. Terrazas 532 (CHAPA), Guerrero, MX AY181631 AY181614Peniocereus castellae Sánchez-Mej. Arreola 1462 (IBUG), Jalisco, MX DQ099945 DQ100014
321
subgenus Pseudoacanthocereus together with two species ofAcanthocereus. Clade II is represented by Armatocereuslaetus and A. matucanensis as strongly supported sisterspecies (100% bts/99% jck). Another pair of species,Dendrocereus nudiflorus and Leptocereus quadricostatus,are supported as sisters (99% bts/98% jck), and are referredto as clade III. Seven species of Peniocereus subgenusPeniocereus represent a weakly supported (68% bts/65%
jck) group in subtribe Pachycereinae which is embeddedwithin tribe Pachycereeae (corresponding to clade IV).Also in clade IV, there are two samples of P. serpentinus,which are weakly supported (53% bts/- jck) as sister toBergerocactus. Taxa representing subtribe Stenocereinae(80% bts/70% jck) are also embedded within clade IV, andthis group of 20 species is sister to the seven sampled speciesof Echinocereus.
P. cuixmalensis Sánchez-Mej. Arreola 903 (IBUG), Jalisco, MX DQ099946 DQ100015P. fosterianus Cutak Reyes 2907 (OAX), Oaxaca, MX DQ099947 DQ100016P. greggii (Engelm.) Britton and Rose subsp. greggii García 953 (MEXU), Coahuila, MX DQ099948 DQ100017P. greggii (Engelm.) Britton and Rose
subsp. transmontanus (Engelm.) Backeb.Quirk s.n. (DES-1960.6704), Arizona, US DQ099949 DQ100018
Tribe TrichocereeaeHarrisia earlei Britton and Rose Saint Croix Bot. Garden, cult. (Fleming 118) DQ099939 DQ100008
Taxon Voucher GenBank accession
trnL-F rpl16
Table 2. Continued
322
Fig. 1. Strict consensus tree of 885 most parsimonious trees based on plastid trnL intron and trnL-F spacer sequences. The phylogenetic analysis includes gaps as recoded characters. Bootstrap/jackknife percentages (>50%) are given below branches. The positions of Peniocereus species are highlighted
323
rpl16
Individual sequence lengths range from 927 bp in Eulychniacastanea to 1,457 bp in Escontria chiotilla. Total alignedsequence length analyzed for the rpl16 data set after trim-ming the ends of the matrix was 1,357 characters, plus 77gaps coded. A nearly identical tree topology with compara-ble levels of support was obtained after analysis of the rpl16data (tree not shown). The analysis resulted in 17 equallyparsimonious trees (Table 3). Five clades were recovered.Within clade I, species of Peniocereus subgenus Pseudoa-canthocereus are monophyletic and sister to Acanthocereus,although Peniocereus hirschtianus is positioned within theAcanthocereus subclade, and two species of Acanthocereusare positioned within the Peniocereus subclade (just as theywere for trnL-F). Tribe Pachycereeae (clade IV) is sup-ported with 72% bts/69% jck. Within this clade eight speciesof Peniocereus, as seen for the trnL-F data, are recoveredin two subclades among subtribe Pachycereinae. Seven ofthose species are supported as a monophyletic group (75%bts/70% jck), as are the two samples of Peniocereus serpenti-nus from Michoacan and Oaxaca, but these two groups arenot closely related to each other.
Combined analysis
The sequence data from trnL-F and rpl16 were combinedsince the ILD indicated that there was no significant incon-gruence (P = 0.05) between them, and because the individ-ual gene trees were structurally similar, with no evidence of“hard incongruence”. The analysis resulted in 471 equallyparsimonious trees (Table 3), and one of those, chosen atrandom, is shown in Fig. 2 so that relative branch lengthscan be observed. Arrows are positioned on this single treeto indicate nodes that collapse in the strict consensus. Ingeneral the topology is similar to the results of each indi-vidual plastid tree, although the overall topology is morefully resolved and support levels are generally higher.Peniocereus subgenus Pseudoacanthocereus (nine species)and two taxa of Acanthocereus, A. chiapensis and A. griseus,form a monophyletic group sister to the remaining speciesof Acanthocereus plus Peniocereus hirschtianus. Theremaining species of Peniocereus are supported as membersof two different subclades: Peniocereus serpentinus assister to Bergerocactus emoryi, whereas seven species of
Peniocereus subgenus Peniocereus represent a differentmonophyletic group in subtribe Pachycereinae.
Discussion
The trnL-F, rpl16, and combined two-gene phylogenetictrees are incongruous with recent taxonomic treatments forPeniocereus (Barthlott and Hunt 1993; Anderson 2001),since the genus is grossly polyphyletic. Species of Peniocer-eus are positioned in four different clades, and these arediscussed below (Figs. 3, 4).
Peniocereus subgenus Pseudoacanthocereus
Buxbaum (1958) and Bravo-Hollis (1978) considered theentire genus Peniocereus to be a member of the tribe Hylo-cereeae. Wallace (2002) found an association between partof his “Peniocereus/Nyctocereus clade” and the Hylocer-eeae, based on limited molecular data. Our results confirmthis observation, but only for the nine species of Peniocereussubgenus Pseudoacanthocereus (sensu Sánchez-Mejorada1974b). Separate and combined gene analyses consistentlysupport the nine species of Peniocereus subgenus Pseu-doacanthocereus together with Acanthocereus chiapensisand A. griseus as a monophyletic clade (92% bts) of tribeHylocereeae. Incidentally, these latter two taxa have almostidentical gene sequences, which could supports the hypoth-esis that they represent a single species, A. chiapensis, asGómez-Hinostrosa (personal communication) suggestsbased on the presence of hypodermal crystals, and flowerand fruit size. Additional studies including a more variablegene need to be performed to support either the recognitionof two species or a single one. These species from Peniocer-eus subgenus Pseudoacanthocereus as well as A. chiapensisand A. griseus grow in deciduous tropical forests exclusivelyfrom western to southern Mexico.
Based on our findings, Sánchez-Mejorada (1974a, b) wascorrect to recognize a group of Peniocereus species, charac-terized by a unique combination of morphologic features.This group was recognized not as a separate genus, butrather as a subgenus (i.e., P. subgenus Pseudoacanthocereus,which should not be confused with the genus Pseudoacan-
Table 3. Summary of results for individual and combined data matrices
thocereus Ritter from northeastern South America). One ofthe most characteristic features of this group is the presenceof dimorphic stems. Juvenile plants appear quite differentfrom adults in that they first start out producing stems with3–8 ribs, and then progressively reduce the number of ribsuntil they eventually exhibit entirely cylindrical stems at
maturity. Dimorphism also occurs in A. chiapensis accordingto descriptions by Bravo-Hollis (1978), and also judgingfrom observations of specimens collected in central Chiapas.Although dimorphic branches might appear to serve as asynapomorphy for this clade, similar morphological changesactually take place within the development of many species
Fig. 2. One of the 471 equally parsimonious trees from the cladistic analysis of combined plastid trnL intron, trnL-F spacer, and rpl16 sequence data. The phylogenetic analysis included gaps as recoded characters. The arrows indicate branches that collapse in the strict consensus tree. Numbers above branches are branch lengths, and below branches are bootstrap/jackknife percentages (>50%). Tribal and subtribal names are indicated and the positions of Peniocereus species are highlighted
325
in tribe Hylocereeae. The morphological change betweenjuvenile and mature stems can be very pronounced in Epi-phyllum and Disocactus (excluding Aporocactus), for exam-ple, the first stems are narrow and cylindrical, whereassecondary and mature ones are completely flattened. A lesspronounced morphological change occurs in Acanthocer-eus, Hylocereus, Selenicereus and Weberocereus. Ribs arealways evident in these genera, but the first stems are thinwith flexible spines, whereas the mature stems are wider,usually with fewer ribs, and with rigid spines or spinesabsent altogether. In other words, dimorphic stems are a
common character in the tribe, and we propose that it maybe necessary to study the development of this transforma-tion in more detail to assess homology of the trait, and tooffer some insight regarding its possible evolutionary originand importance to the systematics of this tribe.
Peniocereus hirschtianus and Acanthocereus
The consensus trees from both plastid genes and the com-bined data show one species of Peniocereus, P. hirschtianus,
Fig. 3a–f. Representative Peniocereus species. a P. lazaro-cardenasii(Terrazas 461), b P. serpentinus (Guzman 129A), c P. macdougallii(Bravo s.n.), d P. johnstonii, plant with tuberous root (Arias 1296A),
e P. maculatus, plants with tuberous roots (Sanchez-Mejorada 3927),f P. zopilotensis roots (T. Terrazas 212)
326
being sister to Acanthocereus, with weak to moderate sup-port (71% bts in the two-gene tree). As for some species ofAcanthocereus (e.g., A. horridus, A. tetragonus), Peniocer-eus hirschtianus is a Central American rather than aMexican species, so its position in the phylogeny makessense from a geographic standpoint. On the other hand,P. hirschtianus is morphologically very different fromAcanthocereus. It is much more similar to some species ofPeniocereus subgenus Pseudoacanthocereus in terms ofgrowth form, size, spination, and flower form, but thesecharacters may simply be plesiomorphies shared by thehypothetical ancestor of the clade.
The phylogenetic position of Acanthocereus has beenuncertain for some time. An affinity with Dendrocereus hasbeen suggested (Buxbaum 1958; Gibson and Nobel 1986),and some authors have even gone a long way to proposethat both genera be treated as a single genus (Hunt andTaylor 1990; Barthlott and Hunt 1993; Taylor in Hunt 1997).Acanthocereus has also been suggested as the sister genusof Leptocereus (Wallace 2002). Our results support none ofthese hypotheses. In fact Dendrocereus and Leptocereusappear as sister taxa to each other, but in a different cladefar removed from Acanthocereus. Complementary system-atics studies on the genus Acanthocereus by Gomez-Hinostrosa (personal communication) and on Leptocereusby Areces-Mallea (personal communication) are underwayto provide more thorough revisions and robust phyloge-netic studies for these genera.
Peniocereus subgenus Peniocereus and allies
This clade of eight taxa is moderately supported as mono-phyletic by the individual plastid DNA data sets, and the
combined analysis provides strong support (96% bts) for itsmonophyly and resolution into three lineages (Fig. 2). Twosubspecies of P. greggii, from the Chihuahuan and SonoranDeserts, respectively, are sister to each other; this pair issister to a subclade containing P. lazaro-cardenasii, P. zopi-lotensis, and P. viperinus. These species are distributed inthe Tehuacan Valley and Balsas Basin of Mexico. Theremaining three species, P. striatus, P. johnstonii, and P.marianus, are endemic to the Sonoran Desert, and form adistinct subclade of their own. These eight taxa have beenthe source of considerable taxonomic debate. Sánchez-Mejorada (1974a) considered Peniocereus subgenus Penio-cereus to consist of only P. johnstonii, P. marianus, and P.greggii (the type species of the genus) based on monomor-phic juvenile and adult stems, salverform flowers, areoles onthe floral tube near each other and displaying trichomes,and seeds with striate micro-relief. The same characters,however, are shared with Neoevansia, which was erected asa genus to accommodate P. lazaro-cardenasii, P. zopiloten-sis, and P. striatus based on their thinner stems, low ribs, andclustered tuberous roots (Marshall 1941; Sánchez-Mejorada1973). Our trees provide no support for the recognition ofNeoevansia as a genus. Likewise, we show with high levelsof support that P. viperinus is a member of Peniocereussubgenus Peniocereus, and is firmly embedded in this clade.As mentioned earlier, this species was at one time classifiedin Wilcoxia by Britton and Rose (1909) based on its tuber-ous roots, very slender stem, numerous low ribs, and innertepals being as long as the floral tube. Distefano (1956)removed the species from Wilcoxia and erected the genusCullmannia for it, focusing on its pubescent stems, red sal-verform flowers, red pulp, and slightly striate seeds.Although the red-pinkish diurnal flowers of P. viperinus areexceptional in Peniocereus, this species otherwise shares
Fig. 4a–c. Representative Peniocereus species. a P. viperinus, in flower (Arias 975), b P. oaxacensis, with dimorphic stems and flower (Guzmán958), c P. cuixmalensis, in fruit (Arias 823)
most of the features described for Peniocereus subgenusPeniocereus sensu Sánchez-Mejorada (1974a) and Neoevan-sia sensu Marshall (1941). Before ending the discussion ofthis clade, it is also worth noting that a relationship betweenP. viperinus and any Wilcoxia species is rejected by ourresults. In fact, the concept of Wilcoxia as a genus is stronglyrejected because all of its remaining species [W. albiflora(= Echinocereus leucanthus), W. poselgeri, and W. schmollii]are well supported as being embedded within Echinocereus,as Taylor (1985) suggested they would be.
Peniocereus serpentinus
Our results show that Peniocereus serpentinus, sampledfrom one population in southern and one in western Mex-ico, is a member of the subtribe Pachycereinae, as suggestedby Wallace (2002) in his overview of columnar cacti, but itcannot be classified within Peniocereus. Instead, the molec-ular cladogram supports the reinstatement of Nyctocereusas a monotypic genus to accommodate this species. Brittonand Rose originally described Nyctocereus as a genus (des-ignating Cereus serpentinus as the type species) in 1909.Three more species were eventually included in it (Bravo-Hollis 1978), but two of these were later transferred toPeniocereus (P. hirschtianus, P. oaxacensis) by Hunt andTaylor (1991). Interestingly, those two species are farremoved from P. serpentinus in our gene trees, being locatedin the Hylocereeae as discussed above. Kimnach (in Huntand Taylor 1991) transferred the third species of Nyctocer-eus, N. chontalensis, to Selenicereus. This move is supportedby our phylogenetic tree, in which Selenicereus chontalensisis sister to S. grandiflorus in the Hylocereeae. As for theexact placement of Peniocereus (= Nyctocereus) serpentinus,the strict consensus tree indicates only weak support for asister relationship to Bergerocactus. This monotypic genusfrom northwestern Mexico and the extreme southwesternUnited States is structurally quite different from P. serpenti-nus in terms of flower form, size, spination, anthesis, fruitsize, fruit color, and pulp consistency. These differences,however, may reflect another example of divergent pollina-tion and/or fruit dispersal syndromes, and we suggest thatcloser examination of vegetative morphology/anatomy beundertaken to more fully appreciate the relationship ofthese two unusual cacti.
The molecular study presented here includes a broadenough sampling of Cactaceae, including nearly all Penio-cereus species and taxa previously associated with Peniocer-eus, to show that a number of nomenclatural changes willbe required in order for the taxonomy of these plants toreflect their phylogeny better. There are at least two alter-native ways in which to treat the species of Peniocereussubgenus Pseudoacanthocereus and Acanthocereus as mem-bers of tribe Hylocereeae. A new genus could be erected toaccommodate them plus A. chiapensis. At the same time, P.hirschtianus would need to be transferred into Acanthocer-eus. Alternatively, all of these taxa could be treated withinan expanded concept of Acanthocereus.
Within subtribe Pachycereinae of tribe Pachycereeae,Nyctocereus should be reinstated as a monotypic genus toaccommodate Peniocereus serpentinus, and the genusPeniocereus should be restricted to those seven speciespreviously classified as Peniocereus subgenus Peniocereus,Neoevansia, and Cullmannia (or Wilcoxia viperina). Fur-thermore, the present molecular analysis provides moredetailed phylogenetic hypotheses for the relationshipswithin subtribe Pachycereinae than any previous study todate (e.g., Cornejo and Simpson 1997; Wallace 2002;Terrazas and Loza-Cornejo 2002; Arias et al. 2003).
It is remarkable that Bergerocactus, Nyctocereus, andPeniocereus show an affinity to the Pachycereus grouprecognized by Arias et al. (2003) since Nyctocereus andPeniocereus are characterized by stems of small diameter,and tend to be shrubby, prostrate or climbing plants withenlarged roots. This is in sharp contrast to the massive tree-like habit and slender roots exhibited by cacti of the Pachyc-ereus group (i.e., Carnegiea, Lophocereus, and Pterocereus).The results of these gene trees revealed that an extremedivergence in the contrasting morphology of these plantshas obscured their natural relationships, and that our clas-sification systems have focused too much on plesiomorphiccharacters, and characters prone to convergent and parallelevolution (often driven by the same habitat pressures orselection by animal pollinators). On the other hand, wehave failed to recognize adequately the systematic value ofgeographic distribution, stem development, and vegetativeanatomy among these cacti.
Acknowledgements This research was completed in part with grantfunding from UNAM (PASPA 154/2003) that allowed S. Arias to serveas a Visiting Scientist at The New York Botanical Garden. Furtherfinancial support was generously made possible by Consejo Nacionalde Ciencia y Tecnología (33064-V) and the Lewis B. and DorothyCullman Foundation. We thank the Desert Botanical Garden,Huntington Botanical Garden, New York Botanical Garden, and U.Guzmán and A. Areces-Mallea who provided samples from their livingcollections for this work, as well as Ivonne Sanchez del Pino andDamon Little for technical support.
References
Anderson EF (2001) The cactus family. Timber Press, PortlandArias S, Terrazas T, Cameron K (2003) Phylogenetic analysis of Pachyc-
ereus (Cactaceae, Pachycereeae) based on chloroplast and nuclearDNA sequences. Syst Bot 28:547–557
Asmussen CB (1999) Toward a chloroplast DNA phylogeny of theGeonomeae (Palmae). In: Henderson A, Borchsenis F (eds) Evolu-tion, variation, and classification of palms. Mem NY Bot Gard83:121–129
Barthlott W, Hunt DR (1993) Cactaceae. In: Kubitzki K (ed) Thefamilies and genera of vascular plants. Springer, Berlin HeidelbergNew York, pp 161–197
Berger A (1905) A systematic revision of the genus Cereus. Annu RepMO Bot Gard 16:57–86
Bravo-Hollis H (1978) Las cactáceas de México, vol. I. UniversidadNacional Autónoma de México, México D.F
Britton NL, Rose JN (1909) The genus Cereus and its allies in NorthAmerica. Contr US Nat Herb 12:413–437
Britton NL, Rose JN (1920) The Cactaceae, vol 2. Carnegie Institution,Washington
Buxbaum F (1958) The phylogenetic division of the subfamily Cere-oideae, Cactaceae. Madroño 14:177–206
Buxbaum F (1975) Gattung Peniocereus. In: Krainz H (ed) DieKakteen. Franckh’sche Verlagshandlung, Stuttgart,
Cornejo DO, Simpson BB (1997) Analysis of form and function inNorth America columnar cacti (tribe Pachycereeae). Am J Bot84:1482–1501
Distefano C (1956) Su un nuovo genere di Cactaceae: Cullmannia.Nuovo Giorn Bot Ital 63:158–161
Farris JS, Källersjö M, Kluge AG, Bult C (1995) Testing significance ofincongruence. Cladistics 10:315–319
Gibson A, Nobel PS (1986) The cactus primer. Harvard UniversityPress, Cambridge
Hunt H (1997) Further amendments to the CITES Cactaceae checklist.Cact Cons Init 3:12–15
Hunt D, Taylor N (1990) The genera of Cactaceae: progress towardsconsensus. Bradleya 8:85–107
Hunt D, Taylor N (1991) Notes on miscellaneus genera of Cactaceae.Bradleya 9:81–92
Maddison WP, Maddison DR (2000) McClade vers. 4.0. Sinauer,Sunderland
Marshall WT (1941) Cactaceae. Abbey Garden, PasadenaNyffeler R (2002) Phylogenetic relationship in the cactus family
(Cactaceae) based on evidence from trnK/matK and trnL- trnFsequences. Am J Bot 89:312–326
Sánchez-Mejorada H (1973) El género Neoevansia Marshall historia yrevisión. Cact Suc Mex 18:13–26
Sánchez-Mejorada H (1974a) El género Peniocereus, enmienda ydivisión. Cact Suc Mex 19:36–41
Sánchez-Mejorada H (1974b) Revisión del género Peniocereus (Lascactáceas). Dirección de Agricultura y Ganadería, Toluca
Simmons MP, Ochoterena H (2000) Gaps as characters in sequence-based phylogenetic analysis. Syst Biol 49:369–381
Swofford DL (2002) PAUP*. Phylogenetic analysis using parsimony(*and other methods), version 4.0b10. Sinauer Associates,Sunderland
Taberlet P, Gielly L, Pautou G, Bouvet J (1991) Universal primers foramplification of three non-coding regions of chloroplast DNA. PlantMol Biol 17:1105–1109
Taylor NP (1985) The genus Echninocereus. Kew Magazine Mono-graph, Kew
Terrazas T, Arias S (2002) Comparative stem anatomy in the subfamilyCactoideae. Bot Rev 68:444–473
Terrazas T, Loza-Cornejo (2002) Phylogenetic relationships of Pachyc-ereeae: a cladistic analysis based on anatomical-morphological data.In: Fleming TH, Valiente-Banuet A (eds) Columnar cacti and theirmutualists, evolution, ecology, and conservation. Arizona UniversityPress, Tucson, pp 66–86
Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG(1997) The ClustalX windows interface: flexible strategies for multi-ple sequence alignment aided by quality analysis tools. Nucleic AcidsRes 24:4876–4882
Wallace R (2002) The phylogeny and systematics of columnar cacti: anoverview. In: Fleming TH, Valiente-Banuet A (eds) Columnar cactiand their mutualists, evolution, ecology, and conservation. ArizonaUniversity Press, Tucson, pp 42–65
Young ND, Healy J (2003) GapCoder automates the use of indelcharacters in phylogenetic analysis. BMC Bioinformatics 4:6
