Ammopiptanthus S. H. Cheng, a genus of Fabaceae with a unique evergreen broad-leaved habit found in arid northwest China, has been found either to contain two species—A. mongolicus (Maxim. ex Kom.) S. H. Cheng and A. nanus (Popov) S. H. Cheng—or only one species—A. mongolicus. Thus, the taxonomic status of this genus, and the species within it, is unclear. In this paper, we clarify the taxonomic relationship in Ammopiptanthus by analyzing morphological evidence (leaf characters, plant height, and canopy coverage), karyotypes (new indices), and new molecular data (ITS 1–4 gDNA markers as well as trnH-psbA, trnL-trnF, and trnS-trnG cpDNA markers), together with datasets from earlier papers, which described many other traits of Ammopiptanthus species, including genetic diversity, geographic differentiation, spatial distribution patterns, and mating systems. Our findings support the conclusion that the genus Ammopiptanthus contains two independent species, and we offer a corresponding revision of the position of Ammopiptanthus in the Flora of China.
The genus Ammopiptanthus S. H. Cheng (Fabaceae) has been viewed as a Tertiary relict genus, with up to two dominant species, A. mongolicus (Maxim. ex Kom.) S. H. Cheng and A. nanus (Popov) S. H. Cheng (Zheng, 1959; Liu, 1995; Wu & Wu, 1996; Ge et al., 2005; Li & Yan, 2011). Both species of Ammopiptanthus are endangered, and their habitats are stony and/or sandy deserts where the annual precipitation ranges from 100 to 160 mm. Ammopiptanthus nanus is mainly distributed in Ulugqat County in western Xinjiang Province, China, and extends west to Kyrgyzstan, growing in a narrow altitudinal strip between 1800 and 2800 m. In contrast, A. mongolicus is mainly distributed in the Alxa Desert in Inner Mongolia, China, and extends north to southern Mongolia, where it is widely distributed (Shi et al., 2009; Li & Yan, 2011; Zhang et al., 2015). Both species are insect-pollinated and flower profusely from early April to late May. Inflorescences of A. mongolicus and A. nanus produce 12 to 16 and 10 to 14 flowers, respectively, and upon seed maturation, heavy seeds are dispersed by gravity within a short distance of the parent plant (Ge et al., 2005; Li et al., 2006; Chen et al., 2009a, 2009b). Many useful genes have been cloned from A. mongolicus and A. nanus for their idiomatic physiological characters, including genes for molybdenum cofactor sulfurase (AnMCSU) (Yu et al., 2015) and betaine aldehyde dehydrogenase (BADH) from A. nanus (Yu et al., 2014) and a new dehydration-responsive element-binding protein (DREB) transcription factor (AmDREB2.1) (Li et al., 2015), the cold-response gene AmDUF1517 (Gu & Cheng, 2014), AmDHN (Sun et al., 2013), and a cold-induced galactinol synthase (AmGS) (Song et al., 2013) from A. mongolicus. The transcriptome database of A. mongolicus and A. nanus has been analyzed under different conditions (Zhou et al., 2012; Liu et al., 2013; Wu et al., 2014; Gao et al., 2015). However, the use of Ammopiptanthus as a genetic resource is impaired by the lack of taxonomic clarification in this genus.
There are disputes concerning both how many species exist within the genus Ammopiptanthus as well as its taxonomic relationship with other genera in the Thermopsidae (Cui, 1998; Wang et al., 2006; Wei & Lock, 2010; Li & Yan, 2011; Zhang et al., 2015). The leaf margins, plant height, and fruit morphology are the main characters used to distinguish between A. mongolicus and A. nanus (Jiang, 1989; Li & Yan, 2011), with leaf morphology being the key character used by taxonomists (Jiang, 1989; Cui, 1998; Li & Yan, 2011). Because both species within Ammopiptanthus are diploid and have a high degree of general morphological similarity (Zheng, 1959; Pan & Huang, 1993; Ge et al., 2005), several studies have investigated whether there are, in fact, two distinct species in Ammopiptanthus (Jiang, 1989; Ge et al., 2005; Li et al., 2006; Li & Yan, 2011). To date, intraspecific comparisons of morphological characters have focused on plant height, leaf size, leaf form (in particular the conspicuousness of lateral veins), leaflet number, and fruit morphology. We present these data in Table 1 and population data for the two species in Figure 1.
Recently, however, Wei and Lock (2010) in Flora of China merged Ammopiptanthus mongolicus and A. nanus into one species, A. mongolicus, with a morphological description that just combined the two species' characters, which causes a taxonomic problem for the genus. A new taxon A. kamelinii Lazkov has been added to the genus from Kyrgyzstan (Lazkov, 2006) and remains a taxonomic puzzle; the key taxonomic features are shown in Table 1.
Many traits have been studied to examine the degree of interspecific difference, if any, between Ammopiptanthus mongolicus and A. nanus, including endogenous hormone regulation (Zheng et al., 2010), genetic diversity and geographic differentiation patterns (Ge et al., 2005), karyotypic differences (Feng & Song, 1987; Pan & Huang, 1993; Liu et al., 1996; Song et al., 2003), mating system differences (Li et al., 2006; Li & Tan, 2007; Chen et al., 2009a, 2009b; Li & Yan, 2011), community ecology (Y. Zhang et al., 2006; J. Zhang et al., 2010), and chemical element and organic molecule content (Yu et al., 2010; Zheng et al., 2010; Feng et al., 2011).
In this paper, we clarify this taxonomic problem using three key pieces of evidence: (1) the key morphological characters (leaf length, width, form, and size, as well as plant height and canopy shape); (2) new karyotypic analyses including interchromosomal indices; and (3) a molecular phylogeny based on nuclear internal transcribed spacer (ITS) data, as well as chloroplast intergenic spacer (trnH-psbA, trnL-trnF, and trnS-trnG) data. Taken together, this evidence will clarify the evolutionary relationship of Ammopiptanthus mongolicus and A. nanus and will yield a highly specific morphological description of these taxa.
Materials and Methods
A total of 22 populations of the two species were sampled, covering almost the entire geographical range of its distribution in China. Seven populations were sampled in western Xinjiang Province (Ammopiptanthus nanus, 35 individuals), and 15 populations were sampled in the Alxa Desert, including Inner Mongolia, Ningxia, and Gansu Provinces (A. mongolicus, 75 individuals) (Table 2). Five individuals were collected per population. Fresh leaves and fruit were gathered from each individual and were dried in silica gel shortly after collection. Longitude, latitude, and altitude were recorded for each sample. Soil samples from three locations (from two comers and the middle along the diagonal of the quadrat) were collected using 10 × 10 m quadrats (Song, 2001). Soil samples were taken from depths of 0–100 cm in dune areas, in order to compare and analyze soil component characteristics at various depths. At each location, the number of each species was recorded and, within each plot, every specimen for a focal species was collected and identified.
phenotyping
Summary information for the 22 populations is shown in Table 2. Five individual plants in each population and 20 leaves from each plant were selected for further analyses. Calipers were used to measure leaf length (LL) and width (WL), but leaf form (FL = LL/WL) and size (SL = LL × WL) were also calculated. Total plant height (H) and canopy coverage area (CA) were recorded for the aboveground portion of each individual. See the morphological characters and data listed in Table 3. Statistical analysis was performed using Excel 2007 and SPSS v.15.0 (SPSS Inc., Chicago, Illinois, U.S.A.). Quartile plots were drawn using Origin 8.5 (Origin Lab, Northampton, Massachusetts, U.S.A.).
Table 1.
Comparison of morphological characters from Flora Xinjiangensis (Li & Yan, 2011) and Lazkov (2006) of Ammopiptanthus mongolicus (Maxim. ex Kom.) S. H. Cheng, A. nanus (Popov) S. H. Cheng, and A. kamelinii Lazkov.
karyotyping
We obtained photos of chromosomes from Pan and Huang (1993) and Liu et al. (1996) and analyzed these with new indices, including six individuals from four species: one individual from each of Ammopiptanthus mongolicus and A. nanus published by Liu et al. (1996), as well as individual data from Pan and Huang (1993). To these, karyotype data for five individuals of A. mongolicus and A. nanus were added, which were from the unpublished data obtained from Pan Borong (Fig. 2). Karyotype data were analyzed using the Leica (Wetzlar, Germany) CW4000 Karyo software, and one individual of Piptanthus nepalensis (Hook.) Sweet, one individual of Thermopsis barbata Benth. (Liu et al., 1996), and two individuals of Salweenia wardii Baker f. (Yue et al., 2011) were used for confamilial outgroup comparisons (Liu et al., 1996). Karyotype formula was determined by chromosome morphology based on centromere position according to Levan classification (Levan et al., 1964). The following parameters were estimated to characterize the karyotypes numerically: long arm (LA), short arm (SA), total length (TL) [LA + SA], arm ratio (AR) [LA/SA], and centromeric index (CI) [SA/ (LA + SA) ×100]. For each taxon, karyograms were drawn based on length of chromosome size (arranged large to small). Karyotype asymmetry was estimated by several methods, including the karyotype asymmetry index (as K%) (Arano, 1963), total form percent (TF%) (Huziwara, 1962), Rec and Syi indices (Venora et al., 2002), interchromosomal asymmetry index (A2) (Zarco, 1986), dispersion index (DI) (Levan et al., 1964), degree of asymmetry of karyotype (A index) (Watanabe et al., 1999), asymmetry index (AI) (Paszko, 2006), coefficient of variation of chromosome length (CVCL = A2 × 100), and the mean centromeric asymmetry (MCA = A × 100) (Peruzzi et al., 2009). Asymmetry indices were compared with Pearson correlation (MCA as the x-axis and CVCL as the y-axis). Statistical analysis was performed using Excel 2007 and SPSS v.15.0. Scatterplots were produced by Origin 8.5.
Figure 1.
Variation of key taxonomic characters, including leaf traits (leaf length, width, form, and size), as well as plant height and canopy area in different populations of Ammopiptanthus S. H. Cheng (A. mongolicus (Maxim. ex Kom.) S. H. Cheng: 15 populations; A. nanus (Popov) S. H. Cheng: seven populations).
molecular phylogeny
Total genomic DNA was extracted from dried leaf tissue using a modified 29 CTAB method (Rogers & Bendich, 1985; Doyle & Doyle, 1990). The ITS region (ITS-1, 5.8S rDNA, ITS-2) was amplified and sequenced using primers ITS-1 and ITS-4 according to White et al. (1990), and the spacers trnH-psbA, trnS-trnG, and trnL-trnF were amplified and sequenced using the primers and protocols of Sang et al. (1997), Shaw et al. (2005), and Taberlet et al. (1991), respectively. Amplification products were purified using PCR Product Purification Kits (Shanghai SBS, Biotech Ltd., China). Sequencing reactions were conducted with the forward and reverse primers of the amplification reactions, using the DYEnamic ET Terminator Kit (Amersham Biosciences, Little Chalfont, Buckinghamshire, U.K.), with an ABIPRISM 3730 automatic DNA sequencer (Shanghai Sangon Biological Engineering Technology & Services Co., Ltd., Shanghai, China). Electropherograms were edited and assembled using SEQUENCHER 4.8 (Gene Codes, Ann Arbor, Michigan, U.S.A.).
Following the phylogeny analysis of Ammopiptanthus by Wang et al. (2006) and Zhang et al. (2015), for our phylogenetic analysis we selected species from the genus Thermopsis R. Br., including T. licentiana E. Peter, T. alpina (Pall.) Ledeb., T. smithiana E. Peter, T. mongolica Czefr., T. turkestanica Gand., T. chinensis Benth. ex S. Moore, T. rhombifolia (Nutt. ex Pursh) Richardson, T. villosa (Walter) Fernald & B. G. Schub., and T. montana Nutt. ex Torr. & A. Gray, as well as other confamilial species including Piptanthus concolor Harrow ex Craib, P. leiocarpus Stapf, P. laburnifolius (D. Don) Stapf, P. nepalensis, Vuralia turcica Uysal & Ertuğrul, Baptisia tinctoria (L.) R. Br., B. australis (L.) R. Br., Sophora flavescens Aiton, S. tomentosa L., and S. davidii (Franch.) Skeels. Spartium junceum L. was used as the outgroups. All DNA data for these species were sourced from GenBank, and the accession numbers for these samples are shown in Table 2.
Sequences were initially aligned using ClustalX (Thompson et al., 1994) followed by manual adjustments using SeAl ver. 2.0 (Rambaut, 2002). Phylogenetic analyses were conducted on each DNA region as well as on the combined nuclear and plastid datasets. The jModelTest 2.1.4 (Posada, 2008; Darriba et al., 2012) was run for each dataset to determine the best fitting model of sequence evolution. We used Akaike information criterion (AIC) values to rank model fit for Bayesian analyses. Phylogenetic relationships were inferred using Bayesian inference (BI) as implemented in MrBayes 3.1.2 (Ronquist & Huelsenbeck, 2003). Maximum parsimony (MP) relationships were determined using PAUP version 4.0b10 (Swofford, 2003).
One cold and three incrementally heated Markov chain Monte Carlo (MCMC) chains were run for 2,000,000 generations. Trees were sampled every 100 generations. A partitioned Bayesian analysis of both the nuclear and the plastid datasets was also implemented by applying the previously determined models to each data partition (Brown & Lemmon, 2007). For each dataset, MCMC runs were repeated twice to avoid spurious results. All Bayesian analyses produced split frequencies of less than 0.01, showing convergence between the paired runs. The first 2000 to 5000 trees before stationarity were discarded as burn-in, and the remaining trees were used to construct majority-rule consensus trees. Furthermore, a neighbor-net analysis was carried out considering the uncorrected p-distance between individuals and the same outgroup species from the Bayesian analyses, using SplitsTree 4.13.1 (Huson & Bryant, 2006). Branch support was tested using bootstrapping (Felsenstein, 1985) with 1000 replicates.
Results
phenotype
The results listed in Table 3 and Figure 1 show that the minimum canopy area (CA) of our sample of Ammopiptanthus mongolicus individuals is 0.0009 m2 and the maximum is 18.68 m2, while the minimum height is 0.09 m and the maximum is 2.40 m. For A. nanus, the minimum CA is 0.0020 m2 and the maximum is 6.43 m2, while the minimum height is 0.05 m and the maximum is 1.32 m. The statistical results show that while there are significant differences in CA between Ammopiptanthus species (P = 0.001) the variation in CA between A. mongolicus and A. nanus is a taxonomically reliable trait that can be used to distinguish between these two species.
Table 3.
The key taxonomic and phenotypic characters in Ammopiptanthus mongolicus (Maxim. ex Kom.) S. H. Cheng and A. nanus (Popov) S. H. Cheng.
The result in Table 3 and Figure 1 shows that variation in four leaf characters significantly differs between Ammopiptanthus mongolicus and A. nanus. The LL of A. mongolicus (14.98–40.01 mm; 24.30 ± 0.22 mm) is somewhat similar to that of A. nanus (14.41–41.57 mm; 25.05 ± 0.17 mm), but these values significantly differ (P = 0.001). However, our statistical models show highly significant differences between all other leaf characters (WL, FL, and SL) (P = 0.001). The results show the WL of A. mongolicus (5.03–19.14 mm; 9.86 ± 0.14 mm) is significantly narrower than that of A. nanus (7.23–24.01 mm; 15.68 ± 0.11 mm), and also that the FL of A. nanus (0.96–3.72; 1.62 ± 0.01) is more suborbicular than that of A. mongolicus (1.45–5.38; 2.57 ± 0.04). Moreover, the SL of A. mongolicus (98.99–586.55 mm2; 243.62 ± 4.97 mm2) is significantly smaller than that of A. nanus (144.57–903.64 cm2; 400.80 ± 4.85 cm2).
karyotype analyses
According to Liu et al. (1996), Ammopiptanthus mongolicus is distributed in Qinggeletu, Inner Mongolia. Individuals are diploid, and the karyotype formula was 2n = 2x = 18 = 6m (2SAT) + 10sm (2SAT) + 2st. The karyotype was 3A. In contrast, A. nanus in Liu et al. (1996) is distributed in Ulugqat, Xinjiang; individuals are diploid, and the karyotype formula was 2n = 2x = 18 = 8m (4SAT) + 8sm + 2st. The karyotype was 2A.
Our karyotypic data for Ammopiptanthus species are sourced from two studies. First, we include data from Pan and Huang (1993), in which A. mongolicus was collected from Turpan Eremophytes Botanical Garden, Chinese Academy of Sciences (TURP) and A. nanus was collected at field sites in Ulugqat, Xinjiang. For A. mongolicus and A. nanus, respectively, the karyotype formulae were 2n = 2x = 18 = 10m + 4sm (SAT) + 2sm + 2st and 2n = 2x = 18 = 8m + 2m (SAT) + 2sm + 2sm (SAT) + 4sm (SC), and both of the karyotypes were 2A. We obtained karyotypic data for 10 additional individuals by analyzing chromosome photographs of Ammopiptanthus species provided by Pan Borong (unpublished data, Fig. 2). In these photographs, all individuals were found to be diploid, and all karyotypes were 2A. Karyotype formulae are shown in Table 4.
We also included data from Piptanthus nepalensis and Thermopsis barbata for comparison (Liu et al., 1996). Piptanthus nepalensis is distributed in Mangkang, Tibet, and was found to be diploid, with a karyotype formula of 2n = 2x = 18 = 6m (2SAT) + 10sm + 2st and the karyotype 2A. Thermopsis barbata was harvested near Lhasa, Tibet, and was found to be diploid, with a karyotype formula of 2n = 2x = 18 = 14m (2SAT) + 4sm and the karyotype 1A.
Figure 2.
Scatterplot of MCA and CVCL for Ammopiptanthus mongolicus (Maxim. ex Kom.) S. H. Cheng, A. nanus (Popov) S. H. Cheng, Piptanthus nepalensis (Hook.) Sweet, Thermopsis barbata Benth., and Salweenia wardii Baker f.: S. wardii (a) and S. wardii (b). Am-1–5, Am (a), Am (b): A. mongolicus; An-1–5, An (a), An (b): A. nanus; see the other data source from Table 4.
Analysis of karyotype asymmetry indices showed values in Table 4 for Syi, Rec, A2, DI, A, AI, CVCL, and MCA, respectively. All Ammopiptanthus mongolicus and A. nanus individuals showed the same chromosome number (2n = 18), but each species had a different karyotype and formula (Pan & Huang, 1993; Liu et al., 1996). According to the data from Liu et al. (1996), karyotype indices of species from the genus Ammopiptanthus are intermediate between the patterns of Piptanthus nepalensis and Thermopsis barbata, and all karyotype indices show that A. mongolicus and A. nanus are more closely related to each other than to either P. nepalensis or T. barbata. According to the data by Pan and Huang (1993) and the photo data provided by Pan Borong (unpublished data, Fig. 2), the karyotype indices of A. mongolicus and A. nanus are different (Table 3). Scatterplots of MCA and CVCL (Fig. 2) values, as well as DI values (Table 4), designed to assess and visualize overall classification strength, show a bifurcation in the genus Ammopiptanthus that corresponds to the two putative species (Fig. 2).
molecular phylogeny
Phylogenetic analyses were conducted on each DNA region, as well as on the combined nuclear and plastid datasets (ILD test, P = 0.074). We used AIC values to rank model fit for Bayesian analyses. The 4-gene combined dataset included 37 samples and 3117 bps; MP and BI produced trees with the same topology, and so only the BI tree is presented (Fig. 3). In this tree, Ammopiptanthus is resolved as monophyletic, and the relationship is well supported (100% bootstrap support and 1.00 posterior probability). We identified two lineages within the genus Ammopiptanthus, a finding that was well supported by both the MP and BI analyses. These putative clades correspond to A. mongolicus (PP = 0.98, BI; PP = 99, MP) and A. nanus (PP = 0.99, BI; PP = 100, MP). Thermopsis did not form a monophyletic group, since T. licentiana, T. alpina, and T. smithiana were placed outside of the majority species (T. mongolica, T. turkestanica, etc.) and clustered instead with Piptanthus Sweet. The other species of Thermopsis, including T. chinensis, T. rhombifolia, T. villosa, and T. montana, were mixed with Vuralia turcica, Baptisia tinctoria, and B. australis. The results regarding “core Thermopsidae” members (i.e., Piptanthus, Anagyris L., Thermopsis, and Baptisia Vent.) were consistent with the findings of Crisp et al. (2000), Heenan et al. (2004), and Wang et al. (2006). The Sophora clade, which included three species (S. davidii, S. flavescens, and S. tomentosa), was found to be embedded within Thermopsidae and to be sister to Ammopiptanthus.
Table 4.
Comparison of the values obtained from the different karyological parameters used.
As shown in the network analysis in Figure 4, all individuals of Ammopiptanthus were clustered in the same group (bootstrap support = 94.8), but A. mongolicus (bootstrap support = 14.8–65.4) and A. nanus (bootstrap support = 19.6–68.9) were found to be systematically distinct.
Discussion
In the past two decades, Ammopiptanthus populations have been subject to rapid demographic decline, mainly due to increasing anthropogenic pressure within their natural range and low intrinsic growth rates (Zheng, 1959; Yang et al., 2004; Ge et al., 2005; Li et al., 2006; Xiao & Jia, 2007; X. L. Ma et al., 2011; Y. Y. Ma et al., 2011; Wang et al., 2011). Recent surveys have discovered that it is relatively difficult to find Ammopiptanthus seedlings in the wild and, therefore, these two species have been categorized as “endangered” and given protected status in China (Wu & Wu, 1996; Ge et al., 2005).
Effective plant conservation should be based on a comprehensive understanding of a plant species' ecological and taxonomic context. As new methods have been developed, the taxonomy of this genus has provided a clearer picture; newer analyses may also provide sufficient evidence for taxonomic revision at the genus level (Hong, 2016). The new methods that have provided new data are many and include, first, phenotyping of Ammopiptanthus specimens according to taxonomic features. The results of this paper show that two of these features, leaf form and canopy area, effectively support the division of the genus into two species. Plant height was previously used to distinguish the two species. According to the description provided by Li and Yan (2011), the height of A. mongolicus ranges from 1.5 m to 2 m, while that of A. nanus ranges from 0.4 m to 0.7 m. However, our observations indicate that plant heights were similar (P = 0.258). Where A. nanus is in general much shorter than that of A. mongolicus, it was found to have a significantly smaller canopy area than A. mongolicus (P = 0.001) (Table 3, Fig. 1). The LL (leaf form) index analysis showed that leaves of A. nanus are suborbicular, while those of A. mongolicus are subulate. Both of these observations are consistent with the description provided by Li and Yan (2011). Other taxonomic features (Li & Yan, 2011; Table 1) were not analyzed in this paper but may be useful as descriptive morphological characters in the future. The related species A. kamelinii (Lazkov, 2006) remains a taxonomic puzzle. For this species, whose key morphological characters are shown in Table 1, most of its characters, including plant height and leaf characters, are consistent with A. nanus, but leaflet number is instead similar to A. mongolicus, and legume margin character is unknown. Resolving the taxonomic status of A. kamelinii will likely require collection of chromosomal or genetic data.
Figure 3.
The majority-rule consensus tree from Bayesian inference and RAxML of Ammopiptanthus S. H. Cheng and other related species in Thermopsidae based on ITS 1–4 and three cpDNA sequences (trnH-psbA, trnL-trnF, and trnS-trnG). The numbers above the branches are the posterior probabilities.
Figure 4.
Neighbor-net analyses based on uncorrected p-distances. Numbers indicate bootstrap values over 1000 replicates.
Secondly, chromosomal data provided by Liu et al. (1996) and Pan and Huang (1993), including supplementary photo data from Pan (unpublished data), were reanalyzed using new indexes in this paper. The earliest studies regarding chromosomal characters (e.g., chromosome number, karyotype, etc.) of Ammopiptanthus species suggested that the chromosome numbers of A. mongolicus are 2n = 16, x = 8 (Feng & Song, 1987). However, more recent work has consistently shown the chromosome number 2n = 18, x = 9 for A. mongolicus (Song et al., 2003; Wang et al., 2006; Zhang et al., 2015), the validity of which is demonstrated again here. The incorporated karyotypes and formulae from previous work (Pan & Huang, 1993; Liu et al., 1996) with new analyses of karyotype asymmetry were determined by scatter diagrams based on MCA and CVCL (Fig. 2), along with other chromosomal indices (Table 4); the evidence supports the conclusion that there are two species in Ammopiptanthus. The systematic contexts of the relationships of Ammopiptanthus species, especially the relationship between these species and confamilials such as Piptanthus nepalensis and Thermopsis barbata, were also shown with their specific correlations (Fig. 2, Table 4).
Finally, new genetic data, including ITS and three chloroplast intergenic spacer markers (trnH-psbA, trnL-trnF, and trnS-trnG), further support the conclusion that there are two species in the genus Ammopiptanthus (Figs. 3, 4). The RAxML and Bayes phylogram, as well as a network analysis, show that Ammopiptanthus contains two relatively well-supported and independent clades. These analyses divide and clarify the populations and haplotypes into two identifiable regions, which designate two Ammopiptanthus species (Figs. 3, 4). This result is consistent with phenotypic analysis and karyotypic data and is also consistent with other phylogenetic studies distinguishing two species within the genus Ammopiptanthus (Wang et al., 2006; Zhang et al., 2015). The phylogenetic relationship of clades of the Thermopsidae, including Ammopiptanthus, Sophora L., Thermopsis, and Piptanthus, among others, can be further analyzed using these DNA markers and karyotypes in future work.
We combined the datasets analyzed here with other data, including comparisons in community composition, anatomical observations of vegetative organs, reproductive traits, the genetic diversity as quantified by esterase isozymes, phylogeographic analyses, inorganic element content, endogenous hormone levels, alkaloids and flavonoid levels, combinative amino acids, and the thermal hysteresis activity (THA) temperature of antifreeze proteins (AFPs). Taken together, we conclude that there are many consistent differences between Ammopiptanthus mongolicus and A. nanus, and, therefore, that A. mongolicus and A. nanus should be considered two independent species within the genus Ammopiptanthus (Yun & Zhang, 2004; Ge et al., 2005; Li et al., 2006; Yu et al., 2007; Shi et al., 2009; Zheng et al., 2010; Feng et al., 2011).
To be more precise, these other datasets showed that: (1) Community composition differs in diversity between Ammopiptanthus mongolicus and A. nanus communities. The community of A. nanus has fewer flora components, simpler structure, fewer community types, and lower coverage and abundance. However, community evenness of A. nanus is higher than that of A. mongolicus. Relative to A. mongolicus, A. nanus has weaker environmental adaptability and a narrower distribution than A. mongolicus (Wang Jiancheng, unpublished data; Y. Zhang et al., 2006; Q. Zhang et al., 2007). (2) The anatomical observation of roots, stems, and leaves of A. mongolicus and A. nanus shows the proportion of cortical tissues in roots of A. nanus was larger than that in A. mongolicus, but that the opposite trend is true of fiber tissues. Furthermore, the proportion and composition of epidermal cells in stems differs between the two species; the leptocentric vascular bundle in A. nanus is more prominent than that in A. mongolicus (Yang & Wang, 1991). (3) Comparing A. mongolicus with A. nanus, the mean number of flowers in A. mongolicus is higher (F = 17.51, P < 0.01) and the duration of flowering is longer (F = 14.08, P < 0.01). However, A. nanus had a relatively higher flowering amplitude, and there were obvious differences in the frequency distributions of duration of flowering between the two species (Li et al., 2006; Li & Tan, 2007). (4) Spectrum analysis of esterase isozymes of A. mongolicus and A. nanus show that the similarity index of esterase zymogram between the two species is 0.67, which reflects a close relationship but not that these species are identical, and hence supports the conclusion that they should be considered to be independent taxa (Wei & Shi, 1995); inter-simple sequence repeat (ISSR) markers of the two species are genetically distinct from each other, as is indicated by the fact that analyses show 63% species-specific bands (Ge et al., 2005). As well, 89.81% of variance in Ammopiptanthus occurred between species, and correlation between genetic distance and geographical distance was significant (r = 0.757, P = 0.0001) (Su et al., 2016). (5) Interspecific comparisons of three endogenous plant hormones, gibberellic acid (GA3), indole-3-acid (IAA), and abscisic acid (ABA), reveal differences in hormone expression rates during seed germination, and the hormone levels of A. mongolicus were higher than those of A. nanus in response to increasing Pb2+ concentration (Zheng et al., 2010). The mean content of various metallic elements (B, Fe, Co, Ni, and Ti) showed significant differences between six populations of A. nanus and 22 populations of A. mongolicus (Shi et al., 2009). Moreover, the composition of alkaloids and flavonoids in leaves of A. mongolicus and A. nanus were found to be different (Feng et al., 2011), as were the percentage and contents of nine combinative amino acids in the leaves and seeds of A. mongolicus and A. nanus (Yun & Zhang, 2004). (6) Interspecific differences in hysteresis activity of AFPs have also been identified; at a protein concentration of 20 mg × ml-1, THA of AFPs in A. nanus is 0.46°C, while such activity in A. mongolicus is 0.9°C (Yu et al., 2007).
Below we formally treat Ammopiptanthus mongolicus and A. nanus as two independent species in Ammopiptanthus and make some minor revision to the Flora of China treatments (Wei & Lock, 2010).
Taxonomic Treatment
Ammopiptanthus S. H. Cheng, Bot. Zhurn. (Moscow & Leningrad) 44: 1381. 1959. TYPE: Piptanthus mongolicus Maxim, ex Kom. [= Ammopiptanthus mongolicus (Maxim, ex Kom.) S. H. Cheng].
Evergreen shrubs. Leaves digitately 1- or 3-foliolate; stipules subulate, mostly adnate to petiole; leaflets entire, silvery tomentose. Flowers in short racemes terminating branchlets; bracts small, deciduous. Calyx campanulate, 5-toothed, sub glabrous. Corolla yellow; petals subequal, clawed. Stamens 10; filaments free; anthers uniform. Ovary stipitate, with several ovules; style glabrous. Legume linear to oblong, flat, dehiscent. Seeds reniform, compressed, strophiolate.
Vernacular name. sha dong qing shu.
Distribution and habitat. Ammopiptanthus includes two species, distributed in the arid area in the northwest of China, Kazakhstan, Kyrgyzstan, and south of Mongolia.
Key to the Species of Ammopiptanthus
la. Leaves 1-foliolate, occasionally 3-foliolate, rhombicelliptic, apex obtuse or acute; legumes with straight margin l.A. mongolicus (Maxim, ex Kom.) S. H. Cheng
lb. Leaves 3-foliolate, occasionally 1-foliolate, broadly elliptic to broadly ovate, apex obtuse, often mucronate; legumes with constricted and uneven margin 2. A. nanus (Popov) S. H. Cheng
1. Ammopiptanthus mongolicus (Maxim, ex Kom.) S. H. Cheng, Bot. Zhurn. (Moscow & Leningrad) 44: 1382. 1959. Basionym: Piptanthus mongolicus Maxim, ex Kom., Trudy Glavn. Bot. Sada, n.s. 34: 33. 1920. TYPE: China. Gobi Alaschanica: delivibus aridis lapidosis mountains et arenis firmis, Aug. 1884 (fr.), N. Przewalskii 3 (holotype, LE!).
Evergreen shrubs, 0.3-0.9(-2+) m tall; bark yellowish brown. Stems terete, weakly ridged, gray-puberulent at first, glabrescent. Leaves 1-foliolate, occasionally 3-foliolate; stipules small, triangular, adnate to petiole, silvery tomentose; petiole 4–15 mm; leaflets rhombicelliptic, 1.5-4 × 0.5-2 cm, densely silvery tomentose on both surfaces, lateral veins inconspicuous, base broadly cuneate to rounded, apex obtuse or acute. Flowers 4 to 15, in short, dense terminal racemes; bracts ovate, 5–6 mm, deciduous; pedicels ca. 1 cm, subglabrous, with 2 bracteoles at midpoint. Calyx 5–7 mm. Corolla yellow, ca. 2 cm, petals long-clawed. Ovary stipitate, glabrous. Legume linear-oblong, 3–8 × 1–2 cm, flat, apex acute to obtuse, margin straight; stipe 8–10 mm. Seeds 2 to 5, orbicular-reniform, ca. 6 mm diam.
Vernacular name, sha dong qing, meng gu sha dong qing.
Phenology. Ammopiptanthus mongolicus flowers April through June and fruits May through August.
Distribution and habitat. Ammopiptanthus mongolicus is found in sand dunes, gravel slopes, and terraces beside ravines in Gansu, Inter Mongolia, Ningxia, and Xinjiang Provinces in China; it also occurs in Kazakhstan, Kyrgyzstan, and the south of Mongolia.
2. Ammopiptanthus nanus (Popov) S. H. Cheng, Bot. Zhum. (Moscow & Leningrad) 44: 1384. 1959. Basionym: Piptanthus nanus Popov, Trudy Prikl. Bot. 26: 1. 1931. Podalyria nana (Popov) Popov, Consp. Veget. Fl. Karpat, 130. 1949. TYPE: China. Xinjiang: Kashi, Kisyl-Ssu supra urbem et Tjian-Schan Centrali: rupestribus et delivibus lapidosis, Apr. 1929 (fr.), M. Popov 28 (holotype, LE!).
Evergreen shrubs, 0.5-0.8(-1 +) m tall; bark yellowish brown. Stems terete, weakly ridged, gray-puberulent at first, glabrescent. Leaves 3-foliolate, occasionally 1-foliolate; stipules small, triangular, adnate to petiole, silvery tomentose; petiole 4–15 mm; leaflets broadly elliptic to broadly ovate, 1.5-4 × 0.6-2.4 cm, densely silvery tomentose on both surfaces, lateral veins conspicuous, base broadly cuneate to rounded, apex obtuse, often mucronate. Flowers 4 to 15, in short, dense terminal racemes; bracts ovate, 5–6 mm, deciduous; pedicels ca. 1 cm, subglabrous, with 2 bracteoles at midpoint. Calyx 5–7 mm. Corolla yellow, ca. 2 cm, petals long-clawed. Ovary stipitate, glabrous. Legume linear-oblong, 3-8 × 1-2 cm, flat, apex acute to obtuse, margin constricted and uneven; stipe 8–10 mm. Seeds 2 to 5, orbicular-reniform, ca. 6 mm diam.
Vernacular name, xiao sha dong qing, ai sha dong qing.
Phenology. Ammopiptanthus nanus flowers from April through June and fruits from May through August.
Distribution and habitat. Ammopiptanthus nanus is only found in gravel hillsides in Kashi of Xinjiang Province in China, and extends west to Kyrgyzstan.
