Molecular Phylogeny and Biogeography of Ribes (Grossulariaceae), with an Emphasis on Gooseberries (subg. Grossularia)

Gooseberries are often distinguished from currants as a distinct genus (Grossularia) or subgenus (Ribes subg. Grossularia), but recent molecular phylogenetic analyses of chloroplast and nuclear data disagree as to the monophyly of this group. We report new sequence data from the 18–26S nuclear rDNA ITS and ETS regions and from the chloroplast psbAtrnH intergenic spacer that, in combination with previously reported data, suggest subg. Grossularia is monophyletic and nested within Ribes. Two main lineages are evident within subg. Grossularia, corresponding to the true gooseberries (subg. Grossularia sect. Grossularia) and a clade of glabrous-styled western North American gooseberries (subg. Grossularia sect. Robsonia, subg. Hesperia, Lobbia). Biogeographic analyses based on DIVA optimizations suggest a western North American origin for subg. Grossularia, with subsequent dispersal to east Asia giving rise to a well-supported clade of Asian gooseberry species in sect. Grossularia. This example contrasts with the well-documented pattern of dispersal from Asia to North America, and highlights the need to investigate additional groups distributed widely through the Northern Hemisphere. Ribes L. comprises approximately 150 species of shrubby plants with strikingly diverse oral and fruit features. The infrageneric classiŽcation for Ribes varies among the primary treatments (e.g., Janczewski 1907; Coville and Britton 1908; Berger 1924; Rehder 1940; Sinnott 1985), creating a difŽcult taxonomy (reviews in Spongberg 1972; Sinnott 1985). One of the more notable and consistent divisions within the genus is between the currants (subg. Ribes) and the gooseberries (subg. Grossularia). Currants are mostly spineless shrubs bearing multi-owered racemes with jointed pedicels, whereas gooseberries have nodal spines and sometimes bristly stems, bear few-owered racemes with non-jointed pedicels, and have highly reexed sepals (Senters and Soltis 2003). The differences between gooseberries and currants have led many researchers to recognize gooseberries at the subgeneric (subg. Grossularia) (Janczewski 1907; Sinnott 1985) or generic (Grossularia) (e.g., Coville and Britton 1908; Berger 1924) level. Previous phylogenetic analyses, however, are at odds regarding the status of the gooseberries as a natural group (Fig. 1). Sequence data from the nuclear 18S–26S rDNA internal transcribed spacer (ITS) region indicated monophyly of subg. Grossularia (Senters and Soltis 2003), while restriction site data from the 18S–26S nuclear rDNA region (Messinger et al. 1993) and from two chloroplast regions indicated polyor paraphyly (Fig. 1; Messinger et al. 1999). The chloroplast data strongly resolved the morphologically intermediate spiny currants (subg. Ribes sect. Grossularioides) as a paraphyletic grade subtending only the true gooseberries (subg. Grossularia sect. Grossularia), while a clade of glabrous-styled gooseberries (subg.Hesperia, Lobbia, and Grossularia sect. Robsonia) were weakly placed along a separate lineage. A second study using ITS sequence data also supported the monophyly of the true gooseberries (sect. Grossularia), but did not sample glabrous-styled gooseberries or spiny currants (Fenton et al. 2000). Given the long tradition of recognizing the gooseberries as a distinct genus (e.g., Coville and Britton 1908; Berger 1924) or subgenus (Janczewski 1907; Sinnott 1985), one goal of this study was to clarify the monophyly of subg. Grossularia. A Žrst step towards addressing the monophyly of the gooseberries (subg. Grossularia) is to combine the available chloroplast (Messinger et al. 1999) and ITS (Senters and Soltis 2003) datasets, and to obtain additional data to resolve apparent conicts evident in these previous studies. Additional data may result in a convergence onto similar topologies, indicating that the apparent conict was weak or due to insufŽcient data, or alternatively, may reinforce the apparent conict between the chloroplast and nuclear genomes, indicating possible hybridization and introgression events in the history of the group (Rieseberg andWendel 1993; Avise 1994). Additional datasets reported here include sequence data from the nuclear encoded 18S–26S rDNA external transcribed spacer region (ETS) and from the chloroplast encoded psbA-trnH intergenic spacer region. Previous studies reported levels of variation in the ETS similar to or greater than those typical of the ITS regions (Baldwin and Markos 1998; Bena et al. 1998; Linder et al. 2000). Studies have also reported sufŽcient variation within the psbA-trnH intergenic spacer region for phylogenetic reconstruction within genera (e.g., Sang et al. 1997; Mast and Givnish 2002; Mort et al. 2002). A second goal of this study was to consider Ribes 78 [Volume 29 SYSTEMATIC BOTANY FIG. 1. A simpliŽed representation of prior phylogenetic hypotheses for Ribes. Chloroplast restriction site data from Messinger et. al.(1999) suggest that gooseberries (subg. Grossularia) are non-monophyletic. ITS data from Senters and Soltis (2003) support the monophyly of gooseberries. in the context of Northern Hemisphere biogeography. The Northern Hemisphere problem has recently attracted attention from a phylogenetic perspective, by both zoologists (e.g., Sanmart‡́n et al. 2001) and botanists (e.g., Manos and Donoghue 2001). It has become clear that oristic similarities among the major areas of endemism, such as eastern Asia and eastern North America (see Boufford and Spongberg 1983; Wen 1999), were established in several ways (e.g., movement through Beringia or the North Atlantic) and at different times in different groups (e.g., Tiffney 1985a; Manchester 1999; Wen 1999; Xiang et al. 1998, 2000; Donoghue et al. 2001; Fritsch et al. 2001; Manos and Stanford 2001; Tiffney and Manchester 2001; Xiang and Soltis 2001). Several recent studies have highlighted plant groups that appear to have diversiŽed initially in Asia and subsequently moved to North America via the Bering Land Bridge, apparently at several different times (e.g., Donoghue et al. 2001; Xiang and Soltis 2001). This iterative trans-Beringian movement resembles the pattern described for mammals earlier in the Tertiary (Beard 1998), and generally suggests that Asia has been a primary source area for Northern Hemisphere diversity. However, this may reect the sample of taxa that has been examined to date, and other patterns, including movement fromNorth America to Asia, have also been described (see examples in Sanmart‡́n et al. 2001). It is noteworthy that Northern Hemisphere plant groups that are especially diverse today in western North America have seldom been the focus of phylogenetic biogeographic analyses. Ribes is very broadly distributed around the Northern Hemisphere, and extends south in the mountains of South America, but is especially diverse in western North America, both in terms of the number of species and the representation of major subclades. Likewise, subg. Grossularia, which is the focus of our analysis, occurs around the Northern Hemisphere, but is most diverse in western North America. SpeciŽcally, numerous glabrous-styled gooseberries (subg. Hesperia, Lobbia, and sect. Robsonia) are distributed in western North America, while the true gooseberries (sect.Grossularia) are found throughout North America and Asia (including Taiwan and Japan), with one species in Europe. Better knowledge of phylogenetic relationships in Ribes, and especially inGrossularia, wouldmake it possible to assess geographic patterns of diversiŽcation and directions of movement in the group. MATERIALS AND METHODS Samples. New data reported here include 12 ITS sequences, 73 ETS sequences (from 57 species), and 53 psbA-trnH sequences. Previous studies provided an additional 67 ITS sequences (Senters and Soltis 2003) and restriction site data from two ampliŽed chloroplast regions (rbcL to accD and rpoC1 to rpoC2) for 32 species (Messinger et. al. 1999). Our sampling was weighted towards subg. Grossularia, the focus of this study. We included all of the true gooseberry species (sect. Grossularia) recognized by Sinnott (1985), all of the Asian true gooseberry species recognized by Berger (1924), representatives from each of the three glabrous-styled gooseberry taxa (sect. Robsonia, subg. Hesperia, Lobbia), and both spiny currant species (sect. Grossularioides). In total, 82 of the approximately 150 species of Ribeswere included in this study (Table 1), representing all of the infrageneric taxa recognized by Berger (1924), but with nomenclature following Sinnott (1985). Itea, represented by I. virginica and I. ilicifolia , was chosen as the outgroup based on previously published studies (Soltis et al. 1990, 1993; Morgan and Soltis 1993; Soltis and Soltis 1997), in which Ribes, Itea, and Saxifragaceae s.s. appeared to be closely relatedmembers of Saxifragales. 2004] 79 SCHULTHEIS AND DONOGHUE: PHYLOGENY AND BIOGEOGRAPHY OF RIBES DNA Isolation. Total DNA was isolated from fresh, silica-desiccated, and herbarium material. All extractions were performed using QiaGen DNeasy Plant Mini kits, following manufacturer’s instructions, except that plant material was ground in warmed AP1 buffer (provided in QiaGen kit) rather than liquid nitrogen. Some DNA aliquots were provided by Senters and Soltis (University of Florida), as indicated in Table 1. AmpliŽcation and Sequencing. PCR and sequencing reactions were performed using a Perkin-Elmer CorporationGeneAmp9600 or MJ Research Inc. DNA Enginey Thermal Cycler. PCR products were cleaned using QiaGen QiaQuick PCR PuriŽcation kits. Sequencing reactions were performed with ABI Prism BigDye Terminator Cycle Sequencing Ready Reaction Kits, using half or full reactions, and were cleaned using either EtOH/NaOAc precipitation, following recommendations in BigDye kits, or Edge Biosystems Performa DTRy Gel Filtration systems. In some cases, sequences were improved with the addition of DMSO to sequencing reactions (approximately 6% Žnal volume). Sequencing reactions were resolved on 5% polyacrylamide gels with an Automated Biosystems (ABI) 377 sequencer, loaded with standard or Rapid Load membrane combs (The Gel Company). Sequencher version 3.1.1 was used to examine and edit sequence chromatograms. ITS ampliŽcation products for the 12 new sequences reported in this studywere generatedusing primers ITS4 (White et al. 1990) and ITS-I (59-GTCCACTGAACCTTATCATTTAG-39; designedbyL. E. Urbatsch, Louisiana State University) with the followingparameters: initial denaturation (978C, 1 min), followed by 35 cycles of denaturation (978C, 10 sec), annealing (488C, 30 sec), extension (728C, 20 sec increasing 4 sec with each cycle), and concluding with a Žnal extension (728C, 7 min). PCR reactions contained 2.5mL 10X AmpliTaq buffer (Perkin-Elmer), 2.5mL 10mM dNTPs, 2.5mL 25mM MgCl2, 1.25mL BSA, 1.25mL 10mM ITS4 primer, 1.25mL 10mM ITS-I primer, 0.1mL AmpliTaq DNA polymerase (Perkin-Elmer), and 1–10ng DNA, to a total volume of 25mL. AmpliŽcation products were sequenced with ITS4, ITS2 (White et al. 1990), ITS-I and ITS3B (the reverse complement of ITS3; White et al. 1990). Most the length of ITS 2 was sequenced in both directions. Sequencing primers for ITS 1 often produced short sequences with minimal overlap, such that sequences were largely based on a single strand. A portion of the 5.8S region could not be recovered fromR. cynosbati (41 bp) or R. echinellum (77 bp), and was thus coded as N’s in the matrix. Of the 67 ITS sequences obtained from Senters and Soltis (2003), Žve were missing a large portion of ITS 1 (approx. 100bp for R. giraldii and R. montigenum; approx. 240 for R. missouriense; approx. 130 for R. watsonianum; approx. 230 for R. rubrum). Sequences from the external transcribed spacer of 18–26S nuclear rDNA (ETS) were obtained following the strategy of Baldwin and Markos (1998). Ribes odoratum (subg. Ribes sect. Symphocalyx) and R. hirtellum (subg. Grossularia sect. Grossularia) were chosen for initial long-distance PCR of the inter-genic spacer region (IGS) in order to design an internal primer conserved across Ribes. Long-distance PCR of the IGS was conducted with the following parameters: initial denaturation (948C, 1 min), followed by 35 cycles of denaturation (948C, 30 sec), and combined annealing and extension (728C, 6 min). PCR reactions contained the following components: 2.5mL 10X KlenTaq LA Polymerase Mix (Clontech Laboratories, Inc.), 2.5mL 10mM dNTPs, 2.5mL 10mM 18S-IGS primer (Baldwin and Markos 1998), 2.5m L10mM 26S-IGS primer (Baldwin and Markos 1998), 0.5mL KlenTaq LA DNApolymerase, and 1–10 ng DNA, to a total volume of 25 mL. AmpliŽcationproducts were sequenced in one direction using primer 18S-E (Baldwin and Markos 1998), yielding approximately 480bp of readable sequence. Primer ETS-Rib1 (59GAACTGTTGTCGCGTGCGTCGT39) was designed 59 of the 18S-E priming site, froma conserved region within ETS. Attempts were made to sequence further into the ETS region using primer ETS-Rib2 (59 ACGACGCACGCGACAACAGTTC 39), the reverse complement of ETS-Rib1, but readable sequences were only obtained from R. hirtellum, hindering efforts to identify a conserved site further 59 of ETS-Rib1. Short-distance PCR of the 39 portion of the ETS region was performed using the ETS-Rib1 primer in conjunction with the 18SETS primer (Baldwin and Markos 1998) under the following parameters: initial denaturation (948C, 2 min), followed by 35 cycles of denaturation (948C, 30 sec), annealing (658C, 1 min), extension (728C, 1.5 min), and concluding with a Žnal extension (728C, 7 min). PCR reactions contained 2.5mL10XAmpliTaq buffer (PerkinElmer), 2.5mL 10mM dNTPs, 2.5mL 25mM MgCl2, 0.5mL 10mM ETS-Rib1 primer, 0.5mL 10mM 18S-ETS primer, 0.1mL AmpliTaq DNA polymerase (Perkin-Elmer), and –10ng DNA, to a total volume of 25mL. Cleaned ampliŽcation products were sequenced in both directions with primers ETS-Rib1 and 18S-ETS. Itea ilicifolia is a partial sequence, missing the Žrst 107 bp. Primers psbAF and trnHR (Sang et al. 1997) were used for both ampliŽcation and sequencing of the psbA-trnH intergenic spacer region. PCR reaction parameters were: initial denaturation (948C, 2 min), followed by 35 cycles of denaturation (948C, 30 sec), annealing (618C, 30 sec), extension (728C, 1 min), and concluding with a Žnal extension (728C, 7 min). PCR reactions contained 2.5mL 10X AmpliTaq buffer (Perkin-Elmer), 2.5mL 10mM dNTPs, 2.5mL 25mM MgCl2, 1.25mL 10mM psbAF primer, 1.25mL 10mM trnHR primer, 0.1mL AmpliTaq DNA polymerase (Perkin-Elmer), and1–10ng DNA, to a total volume of 25mL. For some samples, ampliŽcation was aided by the addition of 2.5mL BSA to the PCR reaction. Ribes viburnifolium was not successfully sequenced, possibly due to homopolymer strings close to both the 59 and 39 primer sites. A central portion of approximately 159 bp could not be recovered from Itea virginica and was coded as N’s in the matrix. Sequence identities were veriŽed by performingBLASTsearches (National Center for Biotechnology Information, National Institutes of Health) using sequences from Ribes burejense. Alignment. Sequences were initially aligned in Clustal X Multiple Sequence Alignment Program version 1.81 under the default settings, then adjusted by eye using MacClade 4 sequencing editor features (Maddison and Maddison 2000). AllRibes sequences were readily alignable by eye, as were Itea sequences, but alignment between Ribes and Itea was often difŽcult. Ambiguous regions were aligned so that the number of informative sites was minimized. Base pairs that could not be assigned with conŽdence due to weak or noisy signal were coded with ‘‘N’’ or with IUPAC-IUB ambiguity symbols. Gaps inserted for alignment in regions of inferred indels were coded as ’’–’’ and treated as missingdata. Seven large indels coded as present (1) or absent (0) were added to the psbA-trnH matrix. The seven coded indels included three insertions in Itea (positions 79–88, 199–207, 418–434), two deletions in Itea (positions 110–114, 359–372), an insertion consisting primarily of TA repeats in members of sect. Grossularia (positions 279–295), and a deletion in R. himalense, R. manshuricum, and R. rubrum (positions 111–116). Datasets are available at TreeBASE (study accession number s1001). Dataset Combinability. Overlap among the four datasets (ITS, ETS, psbA-trnH, restriction sites) was not complete. Of the 84 species included in this study, 50 were represented in at least three of the four datasets (Table 1). A series of partition homogeneity tests (Farris et al. 1995) were performed in PAUP* ver. 4.0 (Swofford 2001) to assess dataset combinability. Each test consisted of 100 replicates employing heuristic searches with simple taxon addition, TBR branch swapping, maxtrees set to 1000, and invariant characters excluded. Initial partition homogeneity tests indicated that the psbA-trnH and restriction site datasets were combinable (p50.66), while all other dataset combinations were incongruent (p# 0.01). To explore potential sources of incongruence, we conducted additional homogeneity tests with individual taxa or groups of taxa excluded based on differential and well-supported (bootstrap.70%) placement in trees, or on differential placement regardless of support levels (deQueiroz et al. 1995). For the ETS and ITS comparison, taxa were excluded one at a time and in groups if they represented composites of multiple accessions (Table 1) and exhibited differential placements in ETS versus ITS trees. Samples of R. oxyacanthoides subsp. oxyacanthoides were also excluded in ETS versus ITS comparisons because there was divergence between the two included ETS sequences (sequenced for this study), thus it was not clear which should be combined with the ITS sequence (obtained 80 [Volume 29 SYSTEMATIC BOTANY TABLE 1. Taxa included in this study, with GenBank accession numbers. Following Messinger et al. (1999), taxonomy corresponds to Sinnott (1985) with Grossularia subg. Hesperia and subg. Lobbia treated under Ribes. ITS sequences noted in boldface are from Senters and Soltis (2003). Senters and Soltis provided the DNA for italicized sequences. Restriction site data are from Messinger et al. (1999). Voucher information is cited as per Senters and Soltis (2003) or Messinger et al. (1999) when the data or DNA came from those sources. Herbarium codes follow the Index Herbariorum, Eighth Edition. NPGR refers to the USDA-ARSNational Plant GermplasmRepository. UCBG refers to the University of California Botanical Garden. RBGE refers to the Royal Botanic Garden, Edinburgh. Subg. Ribes L. Sect. Berisia Spach (Alpine Currants) R. acuminatum (Hook. F.&Thomson) Jancz. Chase 3585 (K): ITS AF426376. R. alpinum L. Arnold Arboretum 678–80E: ETS AY138021; psbA-trnH AY138094. Chase 3587 (K): ITS AF426378. NPGR 6640: restriction sites. R. diacanthum Pallas Arnold Arboretum 1852–81B: ETS AY138022; psbA-trnH AY138095; ITS AY138047. NPGR 34 (Messinger 315; OSC): restriction sites. R. giraldii Janczewski Arnold Arboretum 609–74B: ETS AY138023; psbA-trnH AY138096. M. Mort s.n. (WS): ITS AF426381. R. glaciale Wall. UCBG 91.0285: ETS AY138024; psbA-trnH AY138097; ITS AY138048. R. komarovii Pojark. UCBG 91.078: ETS (1): AY138025. Arnold Arboretum 1094–82A: ETS (2): AY138026; psbA-trnHAY138098. R. maximowiczii Batalin T. S. Elias 10940 (RSA): ITS AF426380. NPGR 267: restriction sites. R. orientale Desf. Boufford et al. 28317 (A): ETS AY138027; psbA-trnH AY138099. R. tenue Janczewski Boufford et al. 27604 (A): ETS AY138028; psbA-trnH AY138100. Chase 3610 (K): ITS AF426377. R. vilmorini Janczewski Chase 3612 (K): ITS AF426379. Sect. Calobotrya (Spach) Jancz. (Ornamental Currants) R. afŽne H.B.K. M. Medina 2517 (NY): ITS AF426326. R. brandegeei Eastwood s.c. 350568 (WS): ITS AF426331. R. cereum Douglas UCBG 93.1213: ETS AY138013; psbA-trnH AY138087. NPGR 237.001: restriction sites. Ross 3425 (NY): ITS AF426328. R. ceriferum Coville and Rose s.c. 05017194 (MO): ITS AF426333. R. ciliatum H. and B. Diggs 2625 (NY): ITS AF426327. NPGR 670.001 (Messinger 311; OSC): restriction sites. R. dugesii Greenman Siplivinsky 3939 (WS): ITS AF426329. R. glutinosum Benth. Chase 3594 (K): ITS AF426340. R. indecorum Eastwood UCBG 86.0903: ETS AY138014; psbA-trnH AY138088. Mort 1370 (WS): ITS AF426336. R. malvaceum Smith UCBG 91.1481: ETS AY138015; psbA-trnH AY138089. Johnson s.n. (WS): ITS AF426338. R. mogollonicum Greene NPGR 294.001: ITS AF426332; restriction sites. R. neglectum Rose Villarrea 4940 (NY): ITS AF426330. R. nevadense Kellogg UCBG 89.1635: ETS AY138016; psbA-trnH AY138090. Mort 1373 (WS): ITS AF426339. R. sanguineum Pursh UCBG 90.0193: ETS AY138017; psbA-trnHAY138091. Mort 1372 (WS): ITS AF426335. NPGR 46: restriction sites. R. tortuosum Benth Breedlove 62230 (MO): ITS AF426325. R. viscosissimum Pursh Grimes 1878 (NY): ITS AF426334. NPGR 281.001 (Fredricks 394; OSC): restriction sites. R. wolŽi Rothrock Siplivinsky 4587 (NY): ITS AF426341. Sect. Coreosma (Spach) Jancz. (Black Currants) R. americanum Mill. NPGR 93: restriction sites. Nee 24196 (NY): ITS AF426375. R. bracteosum Douglas UCBG 89.1645: ETS AY138033; psbA-trnH AY138103; ITS AY138049. R. fragrans Pallas s.c. 4378976 (MO): ITS AF426373. R. hudsonianum Richardson B. Ertter 3807 (NY): ITS AF426372. var. petiolare (Douglas) Jancz. NPGR 278 (Fredricks 390; OSC): restriction sites. R. janczewskii Pojark. Chase 3597 (K): ITS AF426370. R. nigrum L. NPGR 215.001 (OSC): ITS AF426374. R. viburnifolium A. Gray UCBG 65.1431: ETS AY138034. NPGR 762.001: ITS AF426371; restriction sites. Sect. Grossularioides (Jancz.) Rehd. (Spiny, or Gooseberry-stemmed Currants) R. lacustre (Pers.) Poiret Lesica 4710 (NY): ETS (1): AY138018; ITS AF426366. Arnold Arboretum 777–93A: ETS (2): AY138019; psbA-trnH AY138092. NPGR 45: restriction sites. R. montigenum McClatchie Bugham & Miller s.n. (WS): ETS AY138020; psbA-trnH AY138093; ITS AF426367. NPGR 864.001 (Messinger 254; OSC): restriction sites. Sect. Heritiera Jancz. (Dwarf Currants) R. erythrocarpum Coville and Leiberg NPGR 860.001 (Messinger 249; OSC): ITS AF426342 ; restriction sites. R. howellii Greene NPGR 449.001 (Messinger 333; OSC): ITS AF426343 ; restriction sites. R. glandulosum Grauer UCBG 89.0750: ETS (1): AY138035. Schultheis 589–00 (YU): ETS (2): AY138036; psbA-trnH AY138105. NPGR 231: restriction sites. Chase 3605 (K) ( 5 R. prostratum L’Her.): ITS AF426345. R. laxiorum Pursh NPGR 439: restriction sites. Goodrich 19052 (WS): ITS AF426344. Sect. Parilla Jancz. (Andine Currants) R. andicola Jancz. Friere-Fierro 2577 (NY): ITS AF426368. R. fasciculatum Sieb. & Zucc. UCBG 88.0615: ETS (1): AY138043; psbA-trnH AY138104. Arnold Arboretum 1879: ETS (2): AY138044. Chase 3592 (K): ITS AF426346. R. ovalifolium Jancz. Gentry s.n. (MO): ITS AF426369. R. valdivianum Phil. Messinger 314 (OSC): restriction sites. Sect. Ribes L. (Red Currants) 2004] 81 SCHULTHEIS AND DONOGHUE: PHYLOGENY AND BIOGEOGRAPHY OF RIBES