New insights into the taxonomy of tribe Euclidieae (Brassicaceae), evidence from nrITS sequence data

Abstract As currently delimitated, the species-rich mustard tribe Euclidieae DC. (Brassicaceae) comprises 28 genera and 152 species distributed primarily in Asia. To date, no tribe-wide comprehensive phylogenetic analysis has been conducted. In this study, sequence data from the nuclear ribosomal internal transcribed spacer (nrITS) region of 82 species in all 28 genera of Euclidieae were used to test its monophyly and infer inter- and intra-generic relationships within. Phylogenetic analyses revealed that Rhammatophyllum and Sisymbriopsis are embedded within Solms-laubachia s.l., and Solms-laubachia lanuginosa (Eurycarpus lanuginosus) fell outside the tribe. Therefore, Solms-laubachia s.l. as currently recognized is not monophyletic and its generic delimitation needed further study. Besides, our results suggest that the genera Lepidostemon, Neotorularia, and Tetracme are polyphyletic.


New insights into the taxonomy of tribe Euclidieae (Brassicaceae), evidence from nrITS sequence data Introduction
The Brassicaceae (Cruciferae) includes 52 tribes, 341 genera, and 3,997 species (Bras-siBase: https://brassibase.cos.uni-heidelberg.de/, accessed 5 February 2018, Koch et al. 2012, Kiefer et al. 2014) distributed worldwide, primarily in the temperate regions (Al-Shehbaz 1984). The family is academically and economically important (Franzke et al. 2011, Al-Shehbaz 2012, Huang et al. 2016, as it contains the well-known model plant species Arabidopsis thaliana (L.) Heynh. and many crops (e.g., cabbage, cauliflower, turnip, rape, canola, radish, wasabi) and ornamentals (e.g., species of Lobularia Desv., Iberis L., Hesperis L., Matthiola W. T. Aiton). Although the family is easily recognized morphologically, it is often difficult to assign an individual plant to a given genus, and there are tremendous controversies on generic delimitations and tribal assignments , Al-Shehbaz 2012. The total number of tribes and genera they include varied among different systems. For example, Schulz (1936) and Janchen (1942) recognized 19 and 15 tribes, respectively. By contrast, others (e.g., Zhou et al. 2001, Appel and adopted an alphabetical arrangement of genera. The use of molecular sequences to infer phylogenetic relationships during the past two decades have greatly improved our understanding of the evolution within the Brassicaceae. That led Al-Shehbaz et al. (2006) to propose the first phylogenetic tribal classification system based on prior molecular studies, especially the chloroplast ndhF sequences , and had since been expanded to include 52 tribes (Al-Shehbaz 2012, German and Friesen 2014, Chen et al. 2016. Generic boundaries had also been redefined, and studies on Solms-laubachia s.l. (Yue et al. 2006(Yue et al. , 2008, Eutrema R. Br. (Al-Shehbaz and Warwick 2005), Microthlaspi F. K. Mey. (Ali et al. 2016), to name a few, demonstrate that trend.
Because taxa sampling varied in previous studies, the interrelationships among genera of Euclidieae varied a great deal. In order to gain knowledge of phylogenetic relationship of the tribe, we conducted the first comprehensive study that included representatives of all genera.

Plant materials and molecular data
This study comprised 33 genera and 88 species, including 28 genera and 82 species of Euclidieae. Forty-nine ITS sequences of 37 species were newly generated here, and all others were downloaded from GenBank (Table 2 and Appendix 1). Six species of Lineage III (sensu Beilstein et al. 2006), namely Sterigmostemum sulphureum (Banks & Sol.) Bornm. and S. billardieri (DC.) D. A. German (Anchonieae), Bunias erucago

DNA extraction, PCR amplification, and sequencing
Total genomic DNA was extracted from silica gel-dried leaf materials using the Plant Genomic DNA Kit (Tiangen Biotech, Beijing, China) following the manufacturer's protocol. The ITS region was amplified with the primers ITS-18F as modified by Mummenhoff et al. (1997) and ITS4 (White et al. 1990). All polymerase chain reactions (PCR) were performed in a 25 μL volume consisting of 1-2μL sample DNA (approx. 1-10 ng), 12.5μL Premix Taq TM (Takara Biomedical Technology, Beijing, China), 1μL of 10 μM stock of each primer, adjusted to 25 μL with ddH 2 O. The PCR protocol of the ITS region involved a hot start with 4 min at 94 °C, and 30-32 cycles of amplification (1 min denaturing at 94 °C, 45-60 s annealing at 52-53 °C, 60-80 s extension at 72 °C), and a final elongation step for 10 min at 72 °C. The sequencing primers are the same as amplified primers.

Phylogenetic analyses
Original chromatograms were evaluated with Sequencher 4.1.4 (Gene Codes Corporation 2002) for base confirmation and contiguous sequences editing, and sequences were aligned with MAFFT v7.311 (Katoh et al. 2002, Katoh andStandley 2013) and were manually adjusted with MEGA 7.0.14 (Kumar et al. 2016). The aligned sequences were analyzed with maximum parsimony (MP) and Bayesian inference (BI). Maximum parsimony analysis were performed by PAUP* 4.0a161 (Swofford 2018) with all characters unweighted. Heuristic parsimony searches were conducted with 100 replicates of random addition of sequences to search for multiple islands of most parsimonious trees (Maddison 1991). Bootstrap analyses (BS) (Felsenstein 1985) to assess the relative support for monophyletic groups were calculated from 1000 replicates using a heuristic search with ten random-addition subreplicates, TBR branch swapping and MULPARS in effect. For Bayesian inference analysis, jModeltest v2.1.7 (Darriba et al. 2012) was used to select the best-fitted model of nucleotide substitution based on the Akaike information criterion (AICc), and the SYM+I+G model was selected for the ITS dataset. Bayesian inference based on the Markov chain Monte Carlo methods (Yang and Rannala 1997) was performed using MrBayes v3.2.6 (Ronquist et al. 2012), four simultaneous Monte Carlo Markov chains (MCMCs) were run for five million generations, and one tree sampled every 1000 generations. The first 1250 trees (25% of total trees) were discarded as burn-in. The remaining trees were summarized in a 50% majority-rule consensus tree, and the posterior probabilities (PP) were calculated.
The resolution of MP analysis was relatively weaker than the outcome of BI analysis. Only the topologies of Bayesian phylogenetic analysis were shown (Figure 1). The result clearly showed that all 28 genera of Euclidieae formed a moderately to strongly supported clade ( PP / BS = 0.99 / 61; Figure 1

Discussion
Our results suggested that Solms-laubachia s.l. is not monophyletic, within which both Rhammatophyllum and Sisymbriopsis are embedded. Besides, S. lanuginosa fell outside of the clade. The closeness of Solms-laubachia, Rhammatophyllum, and Sisymbriopsis was revealed in previous studies (e.g., Belstein et al. 2006, Warwick et al. 2007, German et al. 2009). However, these studies only sampled one or two representative species of each genus and therefore did not reach a convincible conclusion on their generic status. By contrast, this study sampled 23 of 33 species of Solms-laubachia, five of seven of Rhammatophyllum, and three of four of Sisymbriopsis, representing thus far the most complete taxa sampling on these three genera.
Solms-laubachia had recently been subjected to a series of studies, including taxonomy (Lan andZhou 1981, Al-Shehbaz andYang 2000), cytology (Yue et al. 2003, molecular phylogeny (Yue et al. 2006(Yue et al. , 2008, and biogeography (Yue et al. 2009). As traditionally circumscribed, this genus contained nine to thirteen species distributed from Southwest China to East-Himalayan. However, molecular phylogenetic studies demonstrated that Desideria Pamp. and Phaeonychium O. E. Schulz should be included in it, and that led to greatly expanding of the morphological and geographic boundaries of Solms-laubachia. For example, previously delimited Solms-laubachia species have entire, pinnately veined leaves and latiseptate fruit, whereas the expanded Solms-laubachia also has palmately veined leaves, and terete fruit. The geographic distribution of Solms-laubachia s.l. is also expanded westward into Central Asia.
By contrast, Sisymbriopsis includes annual, biennial or perennial herbs primarily with stalked and 1-or 2-forked to dendritic trichomes, pinnately lobed to coarsely dentate or rarely subentire basal and cauline leaves, and linear, flattened and latiseptate fruit with torulose valves and complete septum. Its four species are distributed in Afghanistan, China (Qinghai, Xinjiang, and Xizang), Kyrgyzstan, and Tajikistan (Al-Shehbaz et al. 1999, Al-Shehbaz andGerman 2016).
Although our results suggest combining Solms-laubachia, Rhammatophyllum, and Sisymbriopsis into one monophyletic genus, merging these three genera into one will make it vastly heterogeneous morphologically (Table 3).The combined genus would be highly variable by encompassing nearly all habit types in the family, nearly all petals colors, and almost all inflorescence types, and would be almost impossible to delimit morphologically. Alternatively, one could keep both Rhammatophyllum and Sisymbriopsis as separate monophyletic genera (Figure 1), and split Solms-laubachia s.l. into several well-delimited smaller genera depending on how different the species cluster together. Because our phylogenetic analyses was based on single ITS sequence fragments, infra-generic relationships can be satisfactorily resolved only by further studies dealing with cpDNA and other single-copy nuclear markers.
As for the outlier Solms-laubachia lanuginosa, its three accessions formed a clade clustered with Bunias erucago (Buniadeae), Hesperis sibirica, and H. isatidea (Hesperideae). Because it fell out of Solms-laubachia and the remainder of Euclidieae, we suggest restoring its previous status in the genus Eurycarpus Botsch. The incongruence between taxonomic treatments based strictly on morphology call for the need to draw generic limits and relationships after conducting adequate molecular phylogenetic analyses. Identifying the tribal position of Eurycarpus is beyond the scope of this paper, and it will be conducted in the near future with nuclear and chloroplast sequences data.
The monospecific genus Metashangrilaia was established based on M. forrestii (W.W.Sm.) Al-Shehbaz & D.A.German, a species used to be put in Braya. Previous molecular analyses revealed that it had very distinct ITS sequences and formed a wellsupported clade sister to the rest of Braya (Warwick et al. 2004). Besides, it showed great morphological divergences from other Braya species German 2016). All these led Al-Shehbaz and German (2016) to separate it from Braya and accommodate it in the newly established Metashangrilaia. This study provides further evidence on a strong sister taxon relationship between Metashangrilaia and Shangrilaia (Figure 1), supporting the decision by Al-Shehbaz and German (2016).
Our results also suggest that Neotorularia, Tetracme, and Lepidostemon are not monophyletic. Of the four species sampled from Neotorularia, the generic type N. torulosa clustered with N. contortuplicata, and they were sister to Streptoloma desertorum Bunge, while N. tetracmoides and N. dentata each formed an independent clade (Figure 1). The three sampled Tetracme formed two separate clades, one of which was T. contorta and T. quadricornis, whereas the other was T. secunda and Octoceras lehmannianum (Figure 1). The non-monophyly of both genera is congruent with previous studies (Warwick et al. 2004(Warwick et al. , 2007. Finally, Lepidostemon used to be a monospecific genus, the type species is L. pedunculosus Hook. f. & Thomson. It was expanded by Al-Shehbaz (2000, to include six species endemic to the Mid-western Himalaya (Al-Shehbaz 2015). The ITS sequence of L. glaricola (H. Hara) Al-Shehbaz (Couvreur et al. 2010) did not fall with our newly sequenced L. rosularis (K. C. Kuan & Z. X. An) Al-Shehbaz in one clade. However, due to limited data and low solution of ITS sequences, further studies with extensive sampling and more molecular markers are needed to clarify the taxonomic circumscription of the non-monophyletic genera -Neotorularia, Tetracme, and Lepidostemon.