Comparative Chloroplast Genomics of Litsea Lam. (Lauraceae) and Its Phylogenetic Implications

: Litsea Lam. is an ecological and economic important genus of the “core Lauraceae” group in the Lauraceae. The few studies to date on the comparative chloroplast genomics and phylogenomics of Litsea have been conducted as part of other studies on the Lauraceae. Here, we sequenced the whole chloroplast genome sequence of Litsea auriculata, an endangered tree endemic to eastern China, and compared this with previously published chloroplast genome sequences of 11 other Litsea species. The chloroplast genomes of the 12 Litsea species ranged from 152,132 ( L. szemaois ) to 154,011 bp ( L. garrettii ) and exhibited a typical quadripartite structure with conserved genome arrangement and content, with length variations in the inverted repeat regions (IRs). No codon usage preferences were detected within the 30 codons used in the chloroplast genomes, indicating a conserved evolution model for the genus. Ten intergenic spacers ( psb E– pet L, trn H– psb A, pet A– psb J, ndh F– rpl 32, ycf 4– cem A, rpl 32– trn L, ndh G– ndh I, psb C– trn S, trn E– trn T, and psb M– trn D) and ﬁve protein coding genes ( ndh D, mat K, ccs A, ycf 1, and ndh F) were identiﬁed as divergence hotspot regions and DNA barcodes of Litsea species. In total, 876 chloroplast microsatellites were located within the 12 chloroplast genomes. Phylogenetic analyses conducted using the 51 additional complete chloroplast genomes of “core Lauraceae” species demonstrated that the 12 Litsea species grouped into four sub-clades within the Laurus - Neolitsea clade, and that Litsea is polyphyletic and closely related to the genera Lindera and Laurus . Our phylogeny strongly supported the monophyly of the following three clades ( Laurus–Neolitsea , Cinnamomum–Ocotea , and Machilus–Persea ) among the above investigated “core Lauraceae” species. Overall, our study highlighted the taxonomic utility of chloroplast genomes in Litsea , and the genetic markers identiﬁed here will facilitate future studies on the evolution, conservation, population genetics, and phylogeography of L. auriculata and other Litsea species.


Introduction
The Lauraceae or the Laurel family is a large monophyletic family in the order Laurales and encompasses approximately 2500 to 3000 species from around 50 genera [1,2]. Systematic classification schemes and the tribe-and genus-level phylogenetic relationships within this family have long been controversial [1,[3][4][5][6]. The focal genus of the present study, Litsea Lam., belongs to the "core Lauraceae" or "core Laureae" group in the sense of Chanderbali et al. [1] and Rohwer and Rudolph [7]. As currently circumscribed, Litsea contains around 408 species, with Litsea cubeba (Lour.) Pers. as its type species [8], Neolitsea. Due to the unsettled generic delimitation of Litsea and Lindera, they cautioned against assessing the delimitations of these genera from morphology. On the basis of the matK gene of chloroplast DNA and nuclear internal transcribed spacer (nrITS), Li et al. [3] analyzed the phylogenetic relationship of "core Laureae" group and found that Litsea, Actinodaphne, Neolitsea, and Lindera were polyphyletic but poorly bootstrap-supported. Fijridiyanto and Murakami [24] reconstructed the phylogenetic tree of Litsea and its related genera using the nuclear taxonomic marker rpb2, which supported the fact that Litsea was not monophyletic and closer to Lindera.
Over the last decade, the rapid advances of high-throughput sequencing or nextgeneration sequencing (NGS) technologies have yielded tremendous genome-scale data for angiosperm species and have improved the development of plastid phylogenomics and phylogenetic resolution of many land plants [25][26][27]. Sequencing complete chloroplast genome, for instance, has proven to be informative and effective in resolving complex phylogenetic relationships at a wide range of taxonomic levels [25][26][27]. Specific to the intrageneric phylogeny of the Lauraceae, these advances have revealed previously unknown phylogenetic relationships among genera. For example, Song et al. [6] employed the method of plastid phylogenomics to construct a comprehensive and robust phylogenetic tree of Lauraceae, finding that species of Litsea, Actinodaphne, Iteadaphne, Laurus, Lindera, Neolitsea, and Parasassafras grouped into one clade. Within the clade, four Litsea species (L. glutinosa, L. monopetala, L. magnoliifolia, and L. tsilingensis), Laurus nobilis, and four Lindera species (L. communis, L. glauca, L. megaphylla, and L. nacusua) were located in one sub-clade, while Lindera obtusiloba with other three Litsea species (L. cubeba, L. panamonja, and L. pierrei) formed another sub-clade. Tian et al. [28] investigated the phylogeny of the "core Lauraceae" group with their whole chloroplast genomes and found that three Litsea species (L. monopetala, L. glutinosa, and L. tsinlingensis), Laurus nobilis, and Lindera megaphylla formed one clade with high bootstrap. Using whole plastome genomes, Zhao et al. [29] also found strong support for a close relationship between the genera Litsea, Laurus, and Lindera, and suggested that the clade containing Litsea, Laurus, and Lindera was a sister to the "Cinnamomum-Ocotea clade" consisting of the genera Cinnamomum, Nectandra, and Sassafras. These studies have strongly indicated the polyphyletic nature of Litsea and their unresolved phylogenetic positions within the Lauraceae.
Litsea auriculata Chien et Cheng, the focal species of this study, is a rare deciduous tree and precious medicinal tree restricted and endemic to montane regions in the Zhejiang and Anhui provinces of eastern China [30,31]. Due to its narrow distribution and small number of wild populations, L. auriculata has been listed as "near threatened" by the IUCN and "Grade III Key Protected Wild Plant" by the Chinese Plant Red Book [32]. Thus far, research on L. auriculata has primarily focused on its seed germination, ecophysiology, and population genetics [30,[33][34][35], while studies on the conservation genetics and genomics of the species is lacking. Therefore, we sequenced and reported the complete chloroplast genome of Litsea auriculata for the first time and conducted comparative chloroplast genomics of the genus Litsea with published chloroplast genomes of 11 other Litsea species. Moreover, we developed potential genetic markers such as microsatellites and DNA barcodes for the genus Litsea. Furthermore, to shed light on the systematic placement of Litsea within the "core Lauraceae" or "core Laureae" group, we reconstructed maximum likelihood (ML) and Bayesian inference (BI) phylogenetic trees of these Litsea species and an additional 51 published plastomes of "core Lauraceae" species. The outcomes here will provide useful genomic resources for further studies on conservation and utilization of L. auriculata and other Litsea species and can lay a solid foundation for understanding the phylogeny of Litsea.

Plant Sampling and DNA Extraction of L. auriculata
We sampled fresh young leaves of L. auriculata from a cultivated tree at the Shanghai Chenshan Botanical Garden, Shanghai, China (31 • 4.2609 N, 121 • 10.9040 E), and lodged a voucher specimen (accession number: TMMJZ20200904) at the herbarium of Nanjing University. Total genomic DNA of L. auriculata was extracted from approximately 5 mg of the silica-dried leaf tissue using a modified CTAB method [36]. Subsequently, the integrity of DNA was evaluated by agarose gel electrophoresis and validated using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). NanoDrop LITE spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, USA) was employed to measure the concentration of DNA.

Illumina
Paired-End Sequencing, De Novo Assembly, and Annotation of the Chloroplast Genome of L. auriculata The high-quality genomic DNA of L. auriculata was used to construct Illumina pairedend (2 × 150 bp) library and implement the low-coverage shotgun sequencing in a lane of HiSeq Xten platform (Illumina, San Diego, CA, USA) at The Beijing Genomics Institute (BGI, Shenzhen, China). Approximately 6 Gb of raw data were sequenced, and the clean data were obtained employing NGS QC Tool Kit by removing adapter sequences and low-quality reads with a Q-value ≤ 20. We de novo assembled the complete chloroplast genome of L. auriculata via the GetOrganelle pipeline [37] using the clean data. The chloroplast genome was then automatically annotated using Plastid Genome Annotator [38], with manual adjusting and confirmation of the annotated protein coding genes (CDS) in Geneious Prime (http://www.geneious.com/ (accessed on 10 May 2021), Biomatters Ltd., Auckland, New Zealand). The annotated tRNA genes were further verified using the tRNAscan-SE [39] with default parameters. Afterwards, the resulting annotated cp genome of L. auriculata was submitted to The National Center for Biotechnology Information (NCBI; https:// www.ncbi.nlm.nih.gov/ (accessed on 11 May 2021)). The circular cp genome physical map of L. auriculata was drawn employing the Chloroplot (https://irscope.shinyapps.io/ chloroplot/ (accessed on 5 May 2021)) [40], with subsequent manual editing.

Comparative Chloroplast Genome Analysis of Litsea
Chloroplast genome comparison among the 12 Litsea species was carried out under the Shuffle-LAGAN mode via mVISTA program (http://genome.lbl.gov/vista/mvista/ (accessed on 8 May 2021)) [41] to elucidate the level of sequence divergence, using the annotation of L. acutivena chloroplast genome as a reference.
Additionally, codon usage together with relative synonymous codon usage (RSCU) [44] value was estimated for all protein-coding genes (the genes with sequence length less than 300 bp and repeated genes were eliminated) of 12 whole chloroplast genomes of Litsea via CodonW v1.4.2 (http://codonw.sourceforge.net/ (accessed on 8 May 2021)) [45]. The two unique codons (AUG and UGG) and the three stop codons (TAA, TAG, and TGA) have no degeneracy and were deleted from the data before analysis.

Mining of cp Microsatellite Markers and Hypervariable Regions of Litsea
We used the MIcroSAtellite (MISA) perl script [46] to exploit simple sequence repeats (SSRs) within the chloroplast genomes of the studied 12 Litsea species, setting the parameters with thresholds of 10 repeat units for mononucleotide SSRs; 6 repeat units for dinucleotide SSRs; and 5 repeat units for tri-, tetra-, penta-, and hexa-nucleotide SSRs. Genus-targeted polymorphic SSRs among these 12 Litsea species were selected under the following three criteria: (1) SSRs located in the homologous regions, (2) SSRs that possessed the same repeat units, and (3) the number of repeat units was variable.

Phylogenetic Analysis
To ascertain the phylogenetic position of Litsea within the "core Lauraceae" or "core Laureae" group, we employed 12 Litsea together with other 50 "core Lauraceae" species representing the Laurus-Neolitsea, Cinnamomum-Ocotea, and Machilus-Persea clades as ingroups and Caryodaphnopsis tonkinensis as an outgroup according to the newly updated classification of Lauraceae by Song et al. [6]. Phylogenetic analyses were conducted on the whole chloroplast genome sequences of the above 63 species. The nucleotide sequences were aligned using the default parameters in the MAFFT v7 (https://mafft. cbrc.jp/alignment/software/ (accessed on 10 May 2021)) [48] and followed by some manual adjustments.
We inferred the phylogenetic relationships among the above-studied species using the maximum likelihood (ML) method as implemented in RAxML-HPC v8.2.8 [49] on the CIPRES Science Gateway website (https://www.phylo.org/ (accessed on 11 May 2021)) and the Bayesian inference (BI) method as implemented in MrBayes v3.1.2 [50] under the unpartitioned strategy. A corrected Akaike information criterion (AIC) value was used to determine the best-fitting model of nucleotide substitution and sequence evolution via jModelTest v2.1.10 [51,52], resulting the optimal model of GTR + I + G for both subsequent ML and BI phylogenetic analysis.
Bayesian analyses were conducted using two separated runs of the Markov chain Monte Carlo (MCMC) algorithm for 1 million generations and tree sampling every 1000 generations. The first 25% of sampled trees were discarded as burn-in, and the 25% best-scoring trees were used to construct the consensus tree and to estimate the posterior probabilities (PPs). Convergence was determined by estimating the average standard deviation of the split frequencies (<0.01). To construct the ML tree, we also ran two searches to ensure identical topologies and ML nodal support was calculated with 1000 bootstrap (BS) replicates for each run. Topologies of the above phylogenetic trees were visualized using the Interactive Tree of Life (iTOL) v4 (https://itol.embl.de (accessed on 11 May 2021)) [53].

Conservation of Litsea Chloroplast Genomes
The Illumina HiSeq Xten platform produced 40,992,320 clean paired-end reads for the chloroplast genome de novo assembly of L. auriculata, and the mean sequencing coverage of the chloroplast genome estimated by the GetOrganelle pipeline was 159×. The complete chloroplast genome of L. auriculata was 152,377 bp in length and displayed a quadripartite structure consisting of a pair of inverted repeat regions (IR with 20,015 bp) divided by two single-copy regions (LSC, 93,533 bp, and SSC, 18,814 bp; Figure 1, Table 1). The overall GC content of the chloroplast genome was 39.2%, with the corresponding values of 37.9%, 33.9%, and 44.4% for the LSC, SSC, and IR regions, respectively. Moreover, there were a total of 113 unique genes, including 79 protein-coding genes (CDS), 30 transfer RNA (tRNA) genes, and 4 ribosomal RNA (rRNA) genes. Among these genes, 10 protein-coding genes and six tRNA genes contained a single intron, while three protein-coding genes possessed two introns. The gene rps12 was trans-spliced: the exon at the 5 end was located in the LSC region, whereas the 3 exon and intron were located in the IR regions. Moreover, the ψ ycf 1 and ψ ycf 2 were identified as pseudogenes because of the partial duplication (Table 1). The chloroplast genome of L. auriculata was deposited in GenBank (NCBI) with the accession number MW355498. Comparative chloroplast genomics can provide insights into the mechanism of chloroplast evolution of plant species, including the structural rearrangements and gene features [54][55][56][57]. In our results, all the 12 Litsea chloroplast genomes exhibited the typical and canonical quadripartite structure akin to the majority of angiosperms and other taxa in the Lauraceae [5,6,58]. Variation in chloroplast genome size was very small among 12 Litsea species, ranging from 152,132 (L. szemaois) to 154,011 bp (L. garrettii). The length of their LSC region varied from 93,119 (L. szemaois) to 93,827 bp (L. elongata), their SSC region from 18,799 (L. monopetala) to 18,936 bp (L. mollis), and their IR region from 20,015 (L. auriculata) to 20,744 bp (L. garrettii). The overall GC content was 39.2% across all species except for L. japonica and L. elongata (39.1%) ( Table 1). Gene contents of all 12 chloroplast genomes of Litsea species were also highly conserved, and they all encoded an identical set of 113 genes with 30 tRNA genes. Specifically, 15 genes possessed one intron (atpF, ndhA, ndhB, petB, petD, rpl2, rpl16, rpoC1, rps16, trnA-UGC, trnG-UCC, trnI-GAU, trnK-UUU, trnL-UAA, and trnV-UAC), and three genes had two introns (clpP, rps12, and ycf 3) ( Table 2).
Additionally, the LSC/IRb (JLB), IRb/SSC (JSB), SSC/IRa (JSA), and IRa/LSC (JLA) boundaries/junctions of the 12 Litsea chloroplast genomes (Figure 3) across the 12 Litsea species were highly conserved with minor variations. Specifically, the genes ycf 2, ndhF, ycf 1, trnH-GUG, and psbA were distributed at the boundaries of LSC/IR and SSC/IR in all 12 Litsea species. The entire ycf 2 crossed the LSC/IRb boundary corresponding to the pseudogene fragment ψ ycf 2 with 3051-3186 bp truncated at the IRa/LSC border, and the gene ycf 1 was located in the junctions of SSC/IRa among all 12 Litsea species, meaning their IRa boundaries all extended into ycf 1 gene with the extension of the pseudogene fragment ψ ycf 1 ranging from 1371 (L. mollis) to 1397 bp (L. monopetala) in the IRb/SSC (JSB) boundaries. In addition, the distance between IRb and gene ndhF ranged from 16 (L. garrettii) to 57 bp (L. japonica), and the length between IRa and gene trnH-GUG varied from 3 (L. mollis) to 23 bp (L. cubeba). Correspondingly, the expansion and contraction of the IR regions further explained the length variations in the chloroplast genomes of 12 Litsea species, and this phenomenon was quite common in most previously studied angiosperms and the other Lauraceae species [60][61][62]. The comparison results of codon preference indexed by the value of the relative synonymous codon usage (RSCU) showed that 12 Litsea species had a common preference for the usage of 30 codons within the 61 shared codons used in the chloroplast genomes (RSCU > 1), and the majority of them were characterized with adenine-thymine ending (Table S1), which was in line with the findings in other Lauraceae plants such as Cinnamomum bodinieri Levl. [63], indicating the relative conservation of the cp genes in the genus Litsea. However, the preference of the usage of three codons (AUA, GCC, and GGU) varied among the different species of this genus: Litsea auriculata, L. garrettii, L. glutinosa, L. monopetala, L. pungens, and L. szemaois seldom selected AUA when encoding isoleucine; L. cubeba, L. garrettii, L. glutinosa, and L. pungens preferred using GCC than other Litsea species; and L. elongata, L. mollis, and L. monopetala had bias on GGU (Table S1). This phenomenon might be explained by the biological characteristics of species or genes formed in the process of long-term evolution and adaptation to the environment and the result of the combined effects of selection, mutation, and drift. Other influencing factors may also include genome size, number of introns, tRNA abundance, and gene expression levels [64,65].

Enrichment of Chloroplast DNA Genetic Resources of Litsea
Microsatellites or SSRs are ubiquitous short tandem repeats often consisting of repetitive sequences/motifs of 1-6 bp in length and can be commonly found in the genomes of diverse organisms [66]. Owing to high level of polymorphisms, reproducibility, and abundance in plant genomes, they are widely used as one of the most important and valuable molecular markers for plant population genetic analysis, evolutionary studies, and molecular marker-assisted selection in plant breeding programs [58,67]. Here, we detected a total of 876 chloroplast SSRs within the chloroplast genomes of the 12 Litsea species by MISA, with the number of chloroplast SSRs ranging from 67 (L. mollis) to 80 (L. auriculata). The majority of these chloroplast SSRs (548 out of 876 or 62.56%) were mononucleotides composed mainly of short polyadenine (polyA) or polythymine (polyT) repeats and much less often contained guanine (G) or cytosine (C) tandem repeats. The remaining chloroplast SSRs were comprised of complex nucleotide repeats (13.81%), dinucleotides (11.99%), tetranucleotides (9.25%), trinucleotides (1.83%), pentanucleotides (0.34%), and hexanucleotides (0.11%) ( Table 3). Note: P1-P6 represent SSRs of the mononucleotide, dinucleotide, trinucleotide, tetranucleotide, pentanucleotide, and hexanucleotide types, respectively. Pc represents a complex nucleotide repeat. LSC and SSC denote large and small single copy regions, respectively. The last row represents the percentage for each type of SSR. "-" indicates no data.
Consistent with many previous reports [6,9], the chloroplast SSRs in Litsea were mainly located in the LSC regions (77.51%). A smaller percentage of the chloroplast SSRs occurred in the SSC (16.55%) and IR (5.94%) regions, respectively. These chloroplast SSRs were also richer in the non-coding regions (84.7%), such as intergenic spacers and introns, than coding regions (15.3%). The complete detail of repeat types and locations of cpSSRs in each Litsea species are listed in Table S2. Moreover, chloroplast SSRs exhibited high diversity and variations in copy numbers [56,68], and thus we identified 10 (6 mononucleotides, 3 trinucleotides, 1 tetranucleotides) polymorphic chloroplast SSRs within the 12 Litsea species (Table S2). These chloroplast SSR molecular markers developed in our study will be useful in population genetics and evolutionary studies of the genus Litsea as well as the molecular breeding and conservation of this and related genera.

Phylogenetic Analysis
The phylogenetic trees resulting from the maximum likelihood and Bayesian inference analyses yielded identical tree topologies and showed that the studied Litsea species can be divided into the following four sub-clades on the basis of the node supports ( Figure 5). Sub-clade I (BS = 100, PP = 1.00) contained two sibling Litsea species (L. mollis and L. cubeba) and Lindera obtusiloba. The sub-clade II, which contained Actinodaphne lancifolia (L. coreana), L. auriculata, L. szemaois, L. dilleniifolia, L. monopetala, L. garrettii, L. elongata, and L. japonica, was sister to sub-clade I. Within sub-clade II, the grouping of L. auriculata, as well as its sister taxon Actinodaphne lancifolia, was well supported with a of BS value 80 and PP of 1.00 ( Figure 5). Six Litsea species (L. szemaois, L. dilleniifolia, L. monopetala, L. garrettii, L. elongata, L. japonica) formed a cluster into within sub-clade II, which was sister to the L. auriculata-A. lancifolia pair. The other two Litsea species (L. glutinosa, L. acutivena), Lindera megaphylla, and Laurus nobilis formed sub-clade III with full support of BS 100 and PP 1.00, within which Litsea acutivena was sister to Lindera megaphylla ( Figure 5). The sub-clade IV was sister to sub-clade III and contained the highly supported sister taxa of Litsea pungens and Lindera floribunda (with a BS of 100 and PP of 1.00), which were sister taxa to five other Lindera taxa (L. rubronervia, L. praecox, L. neesiana, L. sericea, L. reflexa) ( Figure 5). The new chloroplast genome phylogeny of Litsea enables us to elucidate the phylogenetic relationships among Litsea and related genera. First and foremost, the phylogenetic relationships between Litsea, Lindera, and Laurus have long been complex and controversial [3,76]. All three genera share similar morphological traits such as having umbellate inflorescences subtended by large involucral bracts, introrse anthers, the presence of two stipitate glands at the third whorl, and oval or spherical fruits [3,76]. Previously, Lindera and Litsea were differentiated on the basis of the number of anther cells [21], but our tree topology and recent work by Tian et al. [28] showed that this character may not be as taxonomically useful as previously thought for differentiating these genera. It is likely that future work with more comprehensive sampling of Litsea and Lindera will reveal the extend of the polyphyly of both genera.
Our phylogenetic tree also revealed the surprising relationship between Litsea auriculata and Actinodaphne lancifolia. A deeper look at the nomenclature of Actinodaphne lancifolia revealed that Litsea coreana H. Lév. is a synonym of this species [77]. The sessile umbels, persistent bracts, and scattered leaves of A. lancifolia in conjunction with our molecular data supports the reinstatement of Litsea coreana Lévl. as the correct name for the taxon. Additionally, our phylogenetic trees demonstrated full support (BS = 100, PP = 1.00) that the genus Actinodaphne was most related to the genus Neolitsea, specifically that Actinodaphne trichocarpa was sister to Neolitsea sericea and Neolitsea pallens, and Actinodaphne obovata was closely related with them, which is consistent with the research of Song et al. [6].
To sum up, our phylogenetic study resolved here revealed that the Litsea group (among the 12 species investigated) containing in the Laurus-Neolitsea clade was polyphyletic and closely related to the genera Lindera and Laurus. Moreover, our ML and BI analyses with complete plastid genomes grouped all ingroup taxa into three clades (Laurus-Neolitsea clade, Cinnamomum-Ocotea clade, Machilus-Persea clade), and our phylogenetic topologies further supported the monophyly of the above three clades, which was also in accordance with the previously comprehensive phylogenetic tree constructed by Song et al. [6] on the basis of the cp genome data. Therefore, whole plastome sequencing has been proven to display higher resolution and reliability in resolving phylogenetic relationships at a wide range of taxonomic levels. On the other hand, to illuminate the more transparent systematic placements of Litsea and its related genera, we recommend more intensive and comprehensive taxa sampling and the combination of other sources of molecular data such as the transcriptomes and whole genomes [78,79]. Considering hybridization, gene introgression, polyploidization, and stochastic and systematic errors may also complicate the use of molecular data as the taxonomic proof of phylogeny [72][73][74][75], and thus we propose an integrated analysis of phenotypic traits such as inflorescence characters and molecular data for better resolving and clarifying the Lauraceae systematic research.

Conclusions
The chloroplast genome of Litsea auriculata was determined and charactered for the first time in this study, and we conducted the comparative plastid genomics and phylogenomics with that of 11 other Litsea species and 51 "core Lauraceae" species. Our results showed the high conservativeness of the plastid genomes of these 12 Litsea species regarding the canonical angiosperm quadripartite structure with no structural arrangement or gene inversion and the unbiased codon usage preferences. In the meantime, abundant genetic resources including 10 intergenic spacers (psbE-petL, trnH-psbA, petA-psbJ, ndhF-rpl32, ycf 4-cemA, rpl32-trnL, ndhG-ndhI, psbC-trnS, trnE-trnT, and psbM-trnD), 5 protein coding genes (ndhD, matK, ccsA, ycf 1, and ndhF), and 876 cpSSRs were developed here as potential DNA barcodes and variable molecular markers for further research of delimitation, phylogenetics, population genetics, and evolution of the genus Litsea, as well as the molecular breeding and conservation of this and the related genus. Our chloroplast genome phylogeny revealed that the 12 Litsea species investigated were polyphyletic, closely related to the genera Lindera and Laurus, and nested within the Laurus-Neolitsea clade. In addition, our analyses with complete chloroplast genomes grouped all ingroup taxa into three clades (Laurus-Neolitsea clade, Cinnamomum-Ocotea clade, Machilus-Persea clade), and our phylogenetic topologies further supported the monophyly of the above three clades previously delimited by Song et al. [6]. We conclude that whole chloroplast genome sequencing enables the development of useful molecular markers and allows for higher resolution and reliability in resolving phylogenetic relationships of the Lauraceae at a wide range of taxonomic levels.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/f12060744/s1, Figure S1: Comparison of the cp genomes among the 12 Litsea species via mVISTA using annotation of L. acutivena as a reference. Table S1: RSCU analysis of protein-coding regions in 12 Litsea species. Table S2: The identified cpSSR loci and genus-targeted polymorphic cpSSRs in 12 Litsea species. Table S3: The nucleotide variability value (π) of both coding regions and non-coding regions (including intergenic spacers and introns) in 12 Litsea species.

Data Availability Statement:
The data presented in this study are available in the article and Supplementary Materials. The whole chloroplast genomes data of this study are openly available in GenBank of NCBI at https://www.ncbi.nlm.nih.gov; accession number: MW355498 (accessed on 11 May 2021).