﻿Lithocarpusdahuensis (Fagaceae), a new species from Fujian Province based on morphology and genomic data

﻿Abstract Lithocarpusdahuensis, a new Fagaceae species from Fujian Province, China, is described and illustrated. The new species is morphologically similar to L.konishii, but its oblanceolate leaf blade has more pairs of acute teeth on the margin, denser lateral veins, smaller cupules enclosing up to 1/4–1/3 of the nut, and its nut is only half as long as those of L.konishii. The plastome of L.dahuensis was 161,303 bp in length and displayed the typical quadripartite structure. Phylogenetic analyses distinguished L.dahuensis from L.konishii with strong support based on whole plastome and nrITS, respectively.


Introduction
The subfamily Quercoideae, of the family Fagaceae, consists of seven genera, including Castanea, Castanopsis, Chrysolepis, Lithocarpus, Notholithocarpus, Quercus, and Trigonobalanus, containing 1,135 species (The plant list 2022). Lithocarpus consists of 341 species, making it the second largest genus after Quercus. These species are widely distributed in tropical and sub-tropical broad-leaved evergreen forests throughout East and Southeast Asia, extending to New Guinea (Cannon 2001). The center of diversity is in East to Southeast Asia, with 123 species in China, mainly distributed in Guangdong, Guangxi, and Yunnan (Huang et al. 1999), 58 species in Thailand (Strijk et al. 2014), and 121 species in Vietnam (Ngoc et al. 2022).
Typically, Lithocarpus has spirally arranged leaves, which are glabrous, coriaceous, oblong-elliptical to oblong in shape, with the entire margin or with teeth along the margin. Its flowers are white to pale yellow. The male flowers are either solitary or in clusters of three or more, with campanulate or cup-shaped perianths, usually 6-lobed, partially united, and 12 stamens. The female flowers are usually solitary or in clusters of two to five, but only one or two of them are well developed, and they have perianths similar to male flowers, but smaller, and with 12 staminodes (Strijk et al. 2014). The sessile cupules are cup-shaped to discoid, with triangular to rhomboid bracts arranged in a diamond pattern on the cupule surface, enclosing the nuts completely or partially. The nuts are oblate to depressed with a concave or convex scar. Sometimes, the nut scar is concave at the margin but conspicuously convex at the center (Huang et al. 1999).
This paper describes a new species of Lithocarpus that was discovered during a field survey in a landscape forest behind the Xuefeng Village in Dahu Town, Minhou County, Fuzhou City in May 2017. It grows in well-preserved native broad-leaved evergreen forests in a valley and has leaves that resemble those of Quercus engleriana Seem, which were easily overlooked but were of constant concern. At the end of May 2018, the plants began to bloom and develop erect male inflorescences. By the end of September, fruit-bearing specimens were collected and measured. The newly found species is similar to L. konishii from Taiwan but differs in the leaf, cupule, and nut characters, as noted below. Considering the morphological differences, molecular data, and geographical isolation, here we describe it as a new species, Lithocarpus dahuensis.

Morphological description
The morphological description of the new species was based on the study of specimens collected in 2019 from various locations. A Stereoscopic Zoom Microscope (Carl Zeiss, Axio zoom. v.16, Germany), equipped with an attached digital camera (Axiocam), and a Digital caliper were used to record the sizes of the morphological characters. Field observations provided habitats and phenology for the new species.

DNA extraction and sequencing
In this study, total DNA was extracted from fresh leaves of the new species using a DNeasy Plant Mini Kit (Qiagen, Valencia, CA, USA). Purified total DNA of the new species was fragmented, and genome skimming was performed using next-generation sequencing technologies on the Illumina Novaseq 6000 platform. The sequencing was conducted by Berry Genomics Co. Ltd. (Beijing, China) using 150 bp paired-end reads with a 480 bp insert size, resulting in 11.58 GB of reads.

Genome assembly, annotation and analysis
The phylogenetic position of the new species was determined through the analysis of nrITS and whole plastome sequences. The nrITS (ITS1-5.8S-ITS2) was assembled using GetOrganelle v1.7.5, with -R of 7 and k-merset of "35, 85, 115". The emb-plant_nr library was selected as the reference genome database, then annotated and visualized using Geneious v2021.2.2.
The paired-end reads were filtered and assembled into a complete plastome using a GetOrganelle v1.7.5.0 (Jin et al. 2020) with appropriate parameters, including a Kmerset of "21,45,65,85,105", and a word size of 0.6. Following previous studies, our workflow includes five key steps as well: 1. Mapping reads to seed and assembling seedmapped reads for parameter estimation; 2. Recruiting more target-associated reads through extending iterations; 3. Conducting de novo assembly; 4. Roughly filtering fortarget-like contigs; 5. Identifying target contigs and exporting all configurations (Bankevich et al. 2012;Camacho et al. 2009;Jin et al. 2020;Langmead and Salzberg 2012). Graphs of the final assembly were visualized by Bandage (Wick et al. 2015) to assess their completeness. Gene annotation was performed using CPGAVAS2 (Shi et al. 2019) and PGA (Qu et al. 2019). The different annotations of protein coding sequences were confirmed using BLASTx. The tRNAs were checked with tRNAscan-SE v2.0.3. Final chloroplast genome map was created using OGDRAW.

Phylogenetic analysis
The phylogenetic relationship was constructed using Maximum likelihood (ML) analyses with the combined nrITS sequence. In total, 92 samples of Lithocarpus, Morella, Corylus and Carpinus were included in our analysis (Suppl. material 1: table S1). Three species of Morella rubra, Corylus fargesii and Carpinus cordata were used as outgroups (Wu et al. 2022). Each individual locus was aligned using MAFFT 7.310 (Katoh and Standley 2013) with default settings. All missing data were treated as gaps. The best nucleotide substitution model according to the Bayesian Information Criterion (BIC) was TNe+R3, which was selected by ModelFinder (Kalyaanamoorthy et al. 2017) implemented in IQTREE v.1.6.8. Maximum likelihood phylogenies were inferred using IQ-TREE (Nguyen et al. 2015) under the model automatically selected by IQ-TREE ('Auto' option in IQ-TREE) for 2000 ultrafast (Minh et al. 2013) bootstraps.

Genomic comparison with related species
The online tool IRscope (Amiryousefi et al. 2018) was employed to draw the genetic architecture of the IR/SC junctions. Then the sequences of 12 Lithocarpus species were aligned using MAFFT7.310 (Katoh and Standley 2013), the nucleotide diversity (Pi value) of single copy genes and intergenic regions was estimated by DnaSP v.6 (Rozas et al. 2003).

Comparative analysis of the plastomes
The plastome of Lithocarpus dahuensis was compared to those of the other 11 Lithocarpus species. The plastome size of these species is very similar (Table 2), ranging from 161,020 bp for L. balansae to 161,974 bp for L. polystachyus. These genomes displayed a typical circular quadripartite structure consisting of a pair of IR regions (25,606 bp to 25,899 bp) separated by an LSC region (90,407 bp to 90,731 bp) and an SSC region (18,239 bp to 19,255 bp) ( Table 2). The overall GC content was identical (~36.7%) across all compared plastomes, and was clearly higher in the IR region (~42.7%) than in the other regions (LSC ~34.5% and SSC ~30.8%), possibly because of the high GC content of the rRNA that was located in the IR regions.
Taxonomic notes. The following morphological characteristics were used to classify the species, including the acute teethed leaf blade margins, a concave nut scar, and cupules that do not completely enclose the nut. There are four other plants share similar characteristics with L. dahuensis, including cupule encrustation (the cupule base is sessile, encasing the base of the nut or about half of it) and a fruit umbilicus (the surrounding margin of the fruit umbilicus is clearly concave), the differences between which are shown in the key.
Conservation status. During our fieldwork from 2017 to 2022, fruit-bearing large trees of Lithocarpus dahuensis were only found in the landscape forest of the Xuefeng village valley, Dahu town, Minhou County, Fujian Province, China. They were also found in the surrounding secondary coniferous and broad-leaved mixed forest, but these were mostly small trees that sprouted after the large trees were felled and did not bear fruit. As the location was discovered to be the only known position, we suggest its placement in the Data Deficient category of IUCN (2022).

Discussion
Phylogenetic analysis was completed on the whole chloroplast genomes, and nrITS sequences of the Fagaceae species. Based on the well-supported phylogenetic trees (Figs 2, 3), Lithocarpus dahuensis is a new species most closely related to L. konishii. It is worth noting that L. konishii is found mainly in the central and southern regions of Taiwan, the eastern regions of Hainan and Hong Kong, all of which are islands. Nevertheless, L. dahuensis was found to be endemic in the mountains of central Fujian, separated by a strait and over 385 km from L. konishii. Lithocarpus konishii is found at altitudes between 100-1150 m, usually 500-700 m (Huang et al. 1999), whereas L. dahuensis only occurs at altitudes above 1000 m.
The new species has an overall morphology similar to Lithocarpus konishii from Taiwan (Huang et al. 1999). However, there are some obvious differences, especially in the morphology of the leaf, cupules and nuts, such as the oblanceolate leaf blade of L. dahuensis has up to 15 pairs of dense lateral veins and 7-10 pairs of acute teeth on the leaf margin from the 2 nd -3 rd lateral veins above the leaf base, whereas L. konishii has 5-8 pairs of obtuse teeth from the 3 rd lateral vein above the leaf base. L. dahuensis has a smaller cupule that encloses up to 1/4-1/3 of the nut, and its nut is only half as long as those of L. konishii (1.4-1.8 vs. 1.8-2.4 cm).
As a result of its simple, stable genetic structure and ease of sequencing, the chloroplast genome has become increasing popular for species identification, phylogeny reconstruction, demographic history tracing and species divergence studies (Liu et al. 2020). However, genomic information on Lithocarpus, particularly the complete chloroplast genome in the NCBI database was very limited. In the current study, we sequenced and assembled the whole chloroplast genome of the new Lithocarpus species and found that the chloroplast genome of L. dahuensis was 161,303 bp in length, within the expected range (107-218 kb) of most angiosperm chloroplast genomes (Heslop-Harrison 2017), and it had the typical quadripartite structure (Jansen et al. 2005;Daniell et al. 2016). The GC content of L. dahuensis chloroplast genome was low (36.75%), which is similar to that reported from other Fagaceae genomes (Yang et al. 2017;Hinsinger and Strijk 2019;Pang et al. 2019).
The chloroplast genome of the new species was compared with the other members of Lithocarpus to understand its structural variations and rearrangements. The results showed that all 12 Lithocarpus plastomes were remarkably similar in terms of size, genes, and genome structures. Genomic comparison between the species revealed a relatively higher level of divergence in non-coding regions than in coding regions, similar to what has been reported for the genus Quercus and Carya from the family Fagaceae (Li et al. 2018;Shen et al. 2022). We also identified a considerable number of variations in the noncoding chloroplast genome sequences, such as rbcL-accD, rpl20-rps12, trnK-rps16, trnF-ndhJ and ccsA-ndhD. These noncoding sites may be useful in understanding the ecological significance of the species in terms of spatial distribution and adaptability besides the evolutionary relationship of the new species within Fagaceae.