The discontinuous geographic distribution pattern of plants in the north temperate zone has been a focus of biogeographic research, especially concerning the mechanisms behind the formation of such a pattern and the spatial and temporal evolution of this intermittent distribution pattern. Hypotheses of boreotropical origin, land bridge migration, and out-of-Tibet have been proposed to explain the formation of the discontinuous distribution pattern. The distribution of Lonicera shows a typical Europe-Asia-North America discontinuous distribution, which makes for a good case study to investigate the above three hypotheses. In this study, we inferred the phylogeny based on plastid genomes and a nuclear data set with broad taxon sampling, covering 83 species representing two subgenera and four sections. Both nuclear and plastid phylogenetic analyses found section Isika polyphyletic, while sections Nintooa, Isoxylosteum, and Coelxylosteum were monophyletic in subgenus Chamaecerasus. Based on the nuclear and chloroplast phylogeny, we suggest transferring L. maximowiczii and L. tangutica into section Nintooa. Reconstruction of ancestral areas suggests that Lonicera originated in the Qinghai-Tibetan Plateau (QTP) and/or Asia, and subsequently dispersed to other regions. The aridification of the Asian interior may have facilitated the rapid radiation of Lonicera in the region. At the same time, the uplifts of the Tibetan Plateau appear to have triggered the spread and recent rapid diversification of the genus on the QTP and adjacent areas. Overall, our results deepen the understanding of the evolutionary diversification history of Lonicera.

We collected leaf material from 54 samples of Lonicera, including two subgenera (Lonicera and Chamaecerasus), and four sections (section Isika, section Nintooa, section Isoxylosteum, and section Coeloxylosteum). We performed genome skimming for the 54 species of Lonicera. Newly collected tissue of Lonicera was approved by Hainan University (Hainan, China) and met local policy requirements. We downloaded plastome sequences for 29 species of subgenera Chamaecerasus and Lonicera, and ten additional species as outgroups from GenBank as detailed in Table 1. Thus, a total of 93 plastomes spanning subgenus Chamaecerasus and subgenus Lonicera were used, including representative samples for four sections of Lonicera.

We extracted total genomic DNA from leaves using a modified cetyltrimethylammonium bromide (CTAB) method (Doyle and Doyle, 1987). We assessed the quality and quantity of DNA using an ultramicro spectrophotometer (ultramicro nucleic acid analyzer) and 1% agarose gel electrophoresis. We quantified each DNA sample with an Agilent 2100 BioAnalyzer (Davis California USA). DNA samples with a total content of at least ≥0.8 ug were selected for library construction and whole genome sequencing. Paired-end sequencing libraries with insert sizes of 300–500 bp were constructed and sequenced at the Beijing Genome Institute (BGI Shenzhen China) using the BGISEQ-500 platform. Raw reads were filtered and trimmed using SOAPfilter v2.2 using the following standard parameters: (1) read pairs were removed when the unknown base content in a single sequencing read exceeded 10% of the length of the read and when the number of low quality (≤ 10) bases exceeded 40% of the length; (2) filtered readings generated by PCR replication.

We used GetOrganelle v2022.1.1 using default settings with the embplant_pt database and embplant_nr database (see online manual at https://github.com/ Kinggerm/GetOrganelle) for the assembly of the whole chloroplast genome and the assembly of the rDNA cistron (including ITS and ETS), respectively. Assemblies were visualized using Bandage v0.8.1 (Wick et al., 2015) for elongated assembly maps (FASTG) and selected target assembly maps (GFA) to determine if loops were formed and if the final assembly was complete and circular. The obtained plastome sequences in fasta format were then initially annotated with Geneious Prime v2021.2.2 (Kearse et al., 2012) to detect possible inversions/rearrangements with species within the same section used as reference. The start codon, stop codon and intron/exon boundaries were manually edited. Unipro UGENE v1.32 was used to determine the IR boundaries. All newly annotated plastome sequences were submitted to GenBank (see Table 1 for accession numbers). We summarized the length, GC content, and number of genes in the entire plastome as well as for its different regions (e.g., large/small single-copy regions, and inverted repeat regions; Table 2, 3).

Notepad++, MAFFT