Past and present geographic distributionsof plant taxa have been widely studied and have provided numerous insightsregarding evolutionary history. In concert, phylogenetic studies examining uniqueand disjunct distribution patterns have further clarified relationships withingroups of plants (Drew and Sytsma, 2012). East Asian/North Americandisjunctions are present in approximately 65 plant genera (e.g.
, Deng et al.,2015; Tiffney and Manchester, 2001; Wen, 1999). Current evidence suggests thatthese disjunctions have occurred since the mid-Miocene, approximately 12million years ago (Tiffney and Manchester, 2001). This distribution pattern representsan ongoing opportunity to examine divergence and dispersal patterns of varioustaxa. Robust phylogenetic analyses, as opposed to classical morphologicalstudies, are needed determine the degree of genetic differentiation that hasoccurred since a given dispersal event. The subtribe Nepetinae (Lamiaceae; mintfamily) is an ideal model to address questions related to East Asian/NorthAmerican disjunctions (Harley et al., 2004; Vogelmann, 1984).
The subtribe Nepetinae (sensu Drew &Sytsma, 2012) contains 13 genera that are almost exclusively found in Eurasiaand the Mediterranean region (Zielinska and Matkowski, 2014). Within thesubfamily, three genera, Agastache, Dracocephalum, and Meehania, sharea disjunct Eurasian/North American distribution. Two of the Nepetinae generafound within North America, Dracocephalumand Meehania, are represented by asingle species in North America, with the remaining species occurring inEurasia. Agastache is the only memberof the subtribe with a primarily New World distribution, with only one speciesfound in eastern Asia (Fuentes-Granados et al.
, 1998; Lint and Epling, 1945).Hence, Agastache provides an ideal study group to examine relationships and dispersalevents because of its unique distribution (within Nepetinae) and gene regions canbe easily amplified (B. Drew, personnel observation). Agastache is comprised of 22 species of herbaceousperennials (Zielinska and Matkowski, 2014) and divided into two morphologicallydistinct sections, Brittonastrum and Agastache (Sanders, 1987; Lint and Epling,1945). These sections are primarily defined by staminal orientation andgeographical distribution (Sanders, 1987; Lint and Epling, 1945). The 14 species of section Brittonastrumare native to the southwestern United States and central and western Mexico.Section Agastache consists of 8 species, which are mostly native to Eastern, North-Central,and Western, North America.
The exception is A. rugosa, which is native to eastern Asia (Fuentes-Granados et al,1998; Lint and Epling, 1945).In this study, evolutionary processes andrelationships within the subtribe Nepetinae were examined by creating aphylogenetic tree using the nuclear ribosomal internal and external transcribedspacer regions (ITS and ETS) and two pentatricopeptide repeat (PPR) gene regions. This studyrepresents the most comprehensive phylogenetic study of Nepetinae, both interms of sequence data and taxonomic coverage, to date, and will provideinsights into relationships within the subtribe and distribution patterns of taxawithin. Specifically, this study (1) creates a more detailed phylogeny of thesubtribe using sequences from herbarium and field collected specimens, (2) determinesthe time of dispersal to North America for the three genera with EastAsian/North American disjunctions, (3) proposes a method of dispersion (e.g.,Bering strait or similar landbridge) for Agastache, and (4) determinesif the dispersal events of Dracocephalum, Meehania, and Agastachefrom Eurasia to North America have the same pattern and timing.
Materials and Methods Sampling and Outgroupssection Atotal of 120 accessions, recovered from herbaria and field collections, wereincluded in this study. Of these accessions, all were sampled from the tribeMentheae, and 116 were sampled from within subtribe Nepetinae. All 13 genera inNepetinae were sampled: Agastache,Cedronella, Dracocephalum, Drepanocaryum, Glechoma, Hymenocrater, Hyssopus,Lallemantia, Lophanthus, Marmoritis, Meehania, Nepeta, Schizonepeta (Drewand Sytsma 2012).
When possible, multiple representations of a genus were used.Effort was made to collect and includemultiple accessions from genera that are large or have broad geographicdistributions (e.g., Dracocephalum,Nepeta).
The outgroup used was composed of Lycopus uniflorus, Lycopus maackianus, Horminum pyrenaicum, and Prunella vulgaris (Drew & Sytsma2012). DNA extraction,amplification, and sequencing DNAwas extracted from herbarium specimens and collected silica-dried plantmaterial that was collected using the DNeasy Plant Mini Kit (Qiagen, Valencia,California, USA) according to the manufacturer’s specifications. Onemodification to the protocol was made and involved heating the extracts at 65?Cfor 30 minutes (instead of 10) to break down secondary compounds that couldinterfere with subsequent PCR. Cycle sequencing reactions were performedfollowing the procedure of Drew and Sytsma (2012). The data were analyzed usingPE-Biosystems of Sequencing Analysis at University of Arizona.
Two pentatricopeptiderepeat (PPR) gene regions, AT1G09680 and AT3G09060 were amplified using theprimers as described in Yuan et al. (2010; AT1G09680) and Drew and Sytsma(2013; AT3G09060). Internal transcribed spacers (ITS) were amplified using theprimers Leu1 (Baldwin, 1992) and ITS4 (White et al., 1990). For amplificationof herbarium specimens, internal primers ITS2 and ITS3 (White et al.
, 1990)were used. External transcribed spacers (ETS) were amplified and sequenced asdescribed in Drew and Sytsma (2012). Sequence analyses, Phylogeneticanalysis, and divergence time estimation Allsequences were edited in Sequencher 5.
4.6 (Gene Codes, Ann Arbor, Michigan,USA) prior to being manually aligned using Mesquite 3.3 (Maddison and Maddison,2016). Gaps were treated as missing data and indels were not coded.Phylogenetic analyses for both data sets were performed using two approaches -Bayesian Inference (BI) and Maximum likelihood (ML). BI utilizing Markov ChainMonte Carlo (MCMC) was conducted using MrBayes v.
3.2.6 (Ronquist andHuelsenbeck, 2003), with defaults used for the related parameters.
Allcharacters were unordered and equally weighted, with gaps being treated asmissing data. Garli was used for the ML analyses of the data set. Divergencetimes were estimated using BEAST v1.8.3 (Drummond et al., 2012), whichestimates phylogenies and divergence times simultaneously.
Calibration points Calibrationpoints were based off of the 95% CI of both the nrDNA analyses from Drew andSytsma (2012). We constrained the crown of subtribe Nepetinae with an age rangeof 27.4-37.4 myr (Drew and Sytsma, 2012). Biogeography Ancestral area estimation was conducted usingStatistical Dispersal-Vicariance Analysis (S-DIVA).
S-DIVA compliments DIVA andstatistically evaluates the alternative ancestral ranges at each node in a treeaccounting for phylogenetic uncertainty and uncertainty in DIVA calculations(Yu et al., 2010). The nrDNABayesian Inference tree was used to reconstruct ancestral area for all membersof the subtribe.