As pa holoparasitic weed, broomrape has seriously threatened the production of economically important crops, such as melons, watermelons, processed tomatoes and sunflower, in Xinjiang in recent years. However, the species and distribution of broomrapes in Xinjiang are unclear. In the present study, the hosts of broomrape were mainly tomato, sunflower, melon, melon seeds, fresh tomatoes, peppers, pumpkins, and cowpea in Xinjiang. All the samples were divided into two groups based on a phylogenetic tree constructed using ITS sequences and UPGMA analysis. The first group is morphologically branched, and most samples have high homology with O. cf. aegyptiaca, while the second group is morphologically unbranched, and most samples have high homology with O. cernua var. cumana, based on ITS sequence analysis. More importantly, all samples were divided into two major clusters and seven subclusters based on the UPGMA dendrogram, and the main genetic variation mainly existed among the populations.
Based on the phylogenetic tree constructed using ITS sequences, sample P76 had a closer relationship with O. purpurea var. bohemica than to O. cf. aegyptiaca, although the morphological characteristics of sample P76 were extremely similar to those of samples from the same field. However, O. purpurea has only been reported in a few areas of Central Europe [18]. In addition, the results of this study showed that the broomrape is not monophyletic based on the nuclear ITS region of Xinjiang Province. This is in accordance with the results of other studies, based on ITS sequences and plastid marker analysis [19]. There were some small clades in the first major clade, which might be related to the hosts and their living circumstances, and had a wider host range and more changeable living circumstances than the second major clade because the host genotype might influence the evolution of parasitic plants [20, 21, 22, 23].
Based on the phylogenetic tree constructed using rps2 sequences, all the samples generated one major clade. The 13 samples formed one clade with high homology to O. crenata, and the other samples formed another major clade with high homology to O. cernua. This might be related to crop rotation and the exchange of seeds in different areas, because the rps2 gene can be transferred from members of O. s. str. to Phelipanche spp. by horizontal gene transfer via attack on the same host [17]. Although there were obvious differences in the phylogenetic relationships based on the ITS and rps2 regions, this implies that a close relationship occurred between O. cf. aegyptiaca and O. cernua to some degree. Compared with the phylogenetic tree constructed with rps2 sequences, the morphology of the samples agreed with the phylogenetic relationship based on the ITS region. This indicates that the evolution speed of ITS was faster than that of rps2.
O. var. cumana showed strong host specificity, parasitizing only sunflower and processed tomato, while O. cf. aegyptiaca parasitized a variety of crops, including processed tomatoes, melons, and melon seed, in addition to being unable to parasitize sunflower in Xinjiang Province. Interestingly, the rps2 genes of the two types of O. spp. samples showed a certain degree of homology, such as samples P2, P16, and P11 of O. cernua var. cumana and samples P82, P9, and P52 of O cf. aegyptiaca parasitizing the same host of processing tomatoes, which may indicate that processing tomatoes plays a certain role in mediating gene exchange.
Broomrape had a strong ability to adapt to different climatic conditions in Xinjiang Province, while the changing climate and host diversity perhaps promoted the evolution of broomrape. The climate change stability of southern Xinjiang was stronger than that of the northern Xinjiang, which may have resulted in the genetic stability of southern samples being stronger than that of northern samples. The ITS and rps2 values of the samples from the same region with different hosts were different, such as the samples from the 62nd regimen, 163rd regimen and Wuyi farms. The ITS and rps2 genes of the samples from different regions with the same host had a close relationship, for example, samples P3, P25, P28, and P42, which may be related to the frequent dispatching of crop seeds carrying O. spp. seeds within different regions. This result indicated that rps2 of broomrape is more conserved than ITS.
Similar to the phylogenetic tree constructed based on ITS sequences, all samples were clearly divided into two groups by UPGMA analysis based on ISSR molecular markers, which were correlated with the morphological characteristics. The first cluster had a wide host, and not all samples were single, with stem branching. All samples in the second cluster were single, and their host ranges were narrow. The same and near geographic origins mostly gathered together, which is similar to the results of Stoyanov et al [24]. Interestingly, sample P22 came from the 163rd regimen, and the host was chili, which was shorter than other samples of the same origin in terms of morphological characteristics; however, it clustered together with samples of the same origin. The chili may not be suitable as the host of broomrape, but the pressure of survival and breeding could oblige it to change its host range.