Integrated agronomic and chloroplast genomic analysis reveals forage potential and evolutionary insights in Pennisetum centrasiaticum

Abstract

The Yarlung Zangbo River Basin in China has a perennial grass species, Pennisetum centrasiaticum, which shows potential for improving feed quality. Researchers assembled two distinct collections, designated P. centrasiaticum—DB and P. centrasiaticum—NB, to investigate the growth characteristics and genetic variations within this species. Their objective was to comprehend the variance in phenotypic features and to sequence whole chloroplast genomes, which would illuminate evolutionary relationships and possibilities for selective breeding. The scientists assessed both agronomic and nutritional aspects of the grass and used sophisticated high-throughput methods to sequence the whole chloroplast genomes. P. centrasiaticum—DB demonstrated beneficial traits, including increased height (71.58 cm), a greater number of internodes (4.17), elongated ears (9.33 cm), an elevated ear weight ratio (7.91%), and higher ether extract content (16.40%), alongside reduced crude fiber content (386.30 g/kg), indicating improved palatability. The chloroplast genomes of P. centrasiaticum—DB (138,291 bp) and P. centrasiaticum—NB (138,283 bp) exhibited standard four-part architectures, with minor variations in the LSC, SSC, IR sections, GC content, and SSR distribution. A phylogenetic study classified Pennisetum and Cenchrus into separate clades, confirming their distinction at the species level. The differences in agronomic data and chloroplast genome data further clarify the distinctions between these two types of P. centrasiaticum, providing theoretical support for their future utilization. The P. centrasiaticum—DB germplasm has significant potential for producing high-quality fodder, while the chloroplast genome data provide fresh insights into the evolutionary linkages and molecular evolution of Pennisetum species.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-33018-2.

Introduction

The genus Pennisetum is one of the important genera of the tribe Paniceae¹. There are 80–140 Pennisetum species from the tropics and subtropics of both hemispheres^2,3. Pennisetum species occupy a wide range of habitats: savannas, woodlands, and woody places in warmer regions of America, Africa, and Asia, and the genus is remarkably diverse in Africa^3,4. Many of the species in this genus are economically significant and are domesticated for different purposes, such as cereal production⁵, pasture⁵, soil stabilization⁶, and ornamental cultivation⁷. P. centrasiaticum belongs to the Gramineae, Pennisetum genus. It is a tetraploid plant with karyotype formula 2n= 4x = 36⁸. As a C₄ plant⁹, this grass is generally capable of higher rates of CO₂ uptake than C₃ species, especially at low CO₂ concentrations¹⁰. In the Qinghai-Tibet Plateau, concentrated precipitation during the growing season and sufficient light can improve the low-temperature tolerance of C₄ plants, which may benefit C₄ plants in high-altitude areas¹¹. Chloroplast (cp) is the unique photosynthetic organelle of plant cells with its own genome^12,13. For most plants, their cp genomes have a highly conserved organization¹⁴, possess a quadripartite structure of large-single-copy (LSC) region, small-single-copy (SSC) region, and two copies of inverted repeat (IR) regions that separate the LSC and SSC apart^15,16. So, the cp genomes were widely used for the phylogenetic analysis among the plant species^17–19. Limited genomic and phylogenetic research has been conducted on Pennisetum centrasiaticum, despite the species’ commercial and ecological importance¹⁹. Its chloroplast genome has only been partially studied, and little is known about the amount of genetic and phenotypic diversity within the species. Its taxonomic location and the success of breeding and conservation initiatives for germplasm depend on our ability to comprehend its cp genome features better^20,21. P. centrasiaticum is an excellent forage with good nutritional value²², widely distributed on roadside, deserted land, and forest margins²³, and is common on slopes and grazed by sheep and goats in Kashmir hills²⁴. Moreover, P. centrasiaticum is frost and drought-resistant²⁵, which was recommended for the reclamation of degraded areas in Poland²⁶. P. centrasiaticum is widely distributed in China’s Yarlung Zangbo River Basin^27,28. In the Flora of China, taxonomists described the leaves of P. centrasiaticum as "The P. centrasiaticum leaves are narrow, the length of leaves ranged from 10 to 25 cm, and the width of leaves ranged from 5 to 8 mm, both sides of the blade have no bristles"²⁹. Nonetheless, there has been no research on the heritability of morphological variation in this species, especially regarding characteristics like leaf bristles that might indicate ecological adaptation. Explaining phenotypic variations and evaluating germplasm resources for fodder enhancement and ecological restoration are both made more challenging by the lack of genetic data^20,30. In August 2020, an ecological investigation was conducted in the upper and middle reaches of the Yarlung Zangbo River of P. centrasiaticum distribution areas to identify grass species that have potential value for ecological restoration and grazing. A population of P. centrasiaticum with dense bristles, which we named P. centrasiaticum—DB, was found in Danang County, Shannan City, Tibet, China. The germplasm is different from P. centrasiaticum, with no bristles, which we named P. centrasiaticum—NB. Now that these different germplasms have been found, we need to compare their phenotypes and genomes to see if the morphological variance is due to evolutionary divergence or adaptation to the environment³¹. This study measured phenotypic traits of P. centrasiaticum—DB and P. centrasiaticum—NB. High-throughput sequencing technology was used to assemble the chloroplast genomes and analyze the structural characteristics of two P. centrasiaticum germplasms. A phylogenetic tree was constructed to provide insights into the evolution of P. centrasiaticum. By analyzing the ecological and forage values of P. centrasiaticum and its chloroplast genome, more information can be provided for subsequent research on the protection of plant diversity, and it can serve as guidance for selecting suitable parents for future hybridization. The current knowledge gap in chloroplast genomics is being filled by this integrated method, which also improves our understanding of phenotypic variety, evolutionary linkages, and possible uses of P. centrasiaticum in fodder production and ecological restoration.

Materials and methods

Plant materials

P. centrasiaticum was first identified and described by Tzvelev, Nikolai Nikolaievich, in 1968. The P. centrasiaticum—DB and P. centrasiaticum—NB were identified and collected from two adjacent counties (Danang: 29°19′22″W, 91°21′18″E, 3581 m; Sangri: 29°16′57″W, 91°53′45″E, 3535 m) by the author, Litian Zhang, in the Yarlung Zangbo River Basin in Tibet, China, in 2020. The voucher specimens of P. centrasiaticum—DB (Zhang 20,200,833) and P. centrasiaticum—NB (Zhang 20,200,834) were kept in the Herbarium of Shigatse Baiyacheng Agricultural Products Processing Co., Ltd. Seeds of the two germplasms were collected, and a germination experiment was conducted in March 2024. The meadow managers at the Shigatse Forestry and Grassland Bureau permitted this field study. The work was performed with the approval of, and in cooperation with, the Shigatse Grassland Workstation and local herders. We confirm that the field studies did not involve endangered or protected species, and the permission number is SFGB20200521. Morphological traits derived from the Flora of China were used to confirm the species’ identification. To guarantee accuracy in germplasm identification, field samples typically consisted of 20–30 individuals per community.

Phenotypic data scoring

A stereomicroscope took photos of 2 germplasms in the seedling stage (ANDONSTAR: ADSM302) in 2024. In each sample site, three quadrats of 5 m × 5 m were set up in P. centrasiaticum communities, and plant materials of P. centrasiaticum in two sample sites were used to determine nutrient content. Crude protein (CP), ether extract (EE), crude fibre (CF), and crude ash (Ash) were determined according to Feed analysis and quality test technology³². Two representative P. centrasiaticum individual tussocks in each quadrat (6 tussocks in each sample site) were selected for phenotypic evaluation. The measured traits include plant height (PH), stem diameter (SD), number of above-ground internodes (NAGI), average internodes length (AIL), the ratio of the last internode (RLI), ear length (EL), ear width (EW), heading rate (HR), a leaf aspect ratio of reproductive stems (LARRS), a leaf aspect ratio of vegetative stems (LARVS). The plant materials were separated into different parts, stored in paper bags, heated at 105 °C for 30 min to kill the living cells, and then dried at 65 °C until a constant mass was achieved. Then, dry weight (DW) and ear weight ratio (EWR) were measured. In our previous research, we found that there are only very minor differences in soil nutrient and moisture levels between the two sampling sites, so there is no need to analyze them in this paper³³. Excel 2016 was used for data entry, descriptive statistics, ANOVA, and t-tests were calculated using SPSS version 20.0. At the flowering stage, we measured all phenotypic traits to make sure they were consistent between individuals. For all attributes, a total of 12 plants per germplasm were measured, with six plants collected from each sampling site. To compare germplasms, one-way ANOVA was employed, with significance being determined at p < 0.05. All measurements were taken in accordance with established protocols for evaluating forage, guaranteeing consistency. The abbreviations and descriptions of all traits are shown below (Table 1).

Table 1.

Description and abbreviation of factors scored in different sample sites.

Abbreviated name	Full name	Description
P. centrasiaticum—DB	P. centrasiaticum with dense bristles	Germplasm of P. centrasiaticum with dense bristles
P. centrasiaticum—NB	P. centrasiaticum with no bristle	Germplasm of P. centrasiaticum with no bristles
P. centrasiaticum—XZ	P. centrasiaticum in Xizang	Germplasm of P. centrasiaticum collected in Shigatse, Tibet, China
P. centrasiaticum—NX	P. centrasiaticum in Ningxia	Germplasm of P. centrasiaticum collected in Ningxia Province of China
PH	Plant height (cm)	The height of P. centrasiaticum
SD	Stem diameter (cm)	The diameter of the main stem
NAGI	Number of above-ground internodes	The above-ground internodes of the main stem
AIL	Average internode length (cm)	The average internode length of the main stem above ground
EL	Ear length (cm)	The length of the ear
EW	Ear width (cm)	The width of the ear
HR	Heading rate (%)	The heading stems number to all stems number
LARRS	Leaf aspect ratio of reproductive stems	The ratio of leaf length to leaf width in reproductive stems
LARVS	Leaf aspect ratio of vegetative stems	The ratio of leaf length to leaf width in vegetative stems
EWR	Ear weight ratio (%)	The weight of the ear to the total dry weight of P. centrasiaticum
SLR	Stem-to-leaf ratio	The weight of the stem to the weight of the leaf
FDR	Fresh-to-dry ratio	The fresh weight to dry weight of P. centrasiaticum
CP	Crude protein (%)	Crude protein content of P. centrasiaticum
EE	Ether extract (g/kg)	Ether extract content of P. centrasiaticum
CF	Crude Fiber (g/kg)	Crude Fiber content of P. centrasiaticum
Ash	Crude Ash (%)	Crude Ash content of P. centrasiaticum

DNA sequencing and assembly

Genomic DNA was extracted from desiccated leaves of two P. centrasiaticum germplasms with a TGuide Plant Genomic DNA Extraction Kit. The integrity and purity of DNA were evaluated before the establishment of the library. Sequencing libraries were constructed, and high-throughput sequencing was conducted by Genesky Biotechnologies Inc., Shanghai, China, utilizing the Illumina HiSeq platform (Illumina, San Diego, CA, USA) under project number 20B1104C. Fastp (0.19.5) and R (3.6.1) were used to evaluate the raw data quality, and raw reads were assembled into contigs using metaSPAdes³⁴. Then, the chloroplast contigs were identified using blast + (2.9.0), referencing the Setaria italica cp genome sequence (GenBank accession number: MK348604). The representative chloroplast sequence contigs were ordered and oriented according to the reference chloroplast genome. A NanoDrop 2000 spectrophotometer and a Qubit 3.0 fluorometer were used to measure the concentration and purity of the DNA. The library was built using paired-end reads of 150 bp and has an insert size of about 350 bp. Only readings with a quality score of 30 or above were kept for assembly after running metaSPAdes (v3.13.0) with the default parameters. Both germplasms were sequenced to a depth of more than 100 × across the chloroplast genome, guaranteeing accurate assembly.

Annotation and analysis of the cpDNA sequences

The plastome protein-coding genes, rRNAs, and tRNAs were annotated using CPGAVAS2, and OGdraw (version 1.1.1) was employed to create the chloroplast genome map. The complete cp genome of P. centrasiaticum germplasms with their gene annotation was deposited in NCBI GenBank with accession numbers OL672847.1 (P. centrasiaticum—DB) and OL672848.1 (P. centrasiaticum—NB), respectively. The chloroplast genomes of P. centrasiaticum—XZ (GenBank: MW421597)¹⁹, P. centrasiaticum—NX (GenBank: MW442087), Cenchrus ciliaris (GenBank: MH286942), Cenchrus purpureus (GenBank: MF594682), Cenchrus americanus (GenBank: NC 024,171), and Oryza sativa (GenBank: KM103369) were downloaded from NCBI to enrich our phylogenetic data and outgroup data, respectively. The online mVISTA program (http://genome.lbl.gov/vista/ mvista/submit.shtml) was applied to compare the complete cp genome of P. centrasiaticum to the other obtained cp genomes with the shuffle-LAGAN mode^35,36, and the genome of P. centrasiaticum—NB was used as the reference genome. The whole genome sequence alignment was performed using MAFFT v7.017³⁷ and manually adjusted in MEGA-X³⁸. Then, the whole genome sequence was used to construct the Neighbor-Joining (NJ) phylogenetic tree by MEGA-X with a bootstrap set to 1000. All chloroplast genome alignments were performed using MAFFT with the default settings. The Shuffle-LAGAN alignment approach was employed with default parameters for the mVISTA comparison. Due to the minimal sequence divergence in the chloroplast genomes and NJ’s suitability for low-variation datasets, NJ was chosen. In order to determine how robust clades were, bootstrap analysis with 1000 replicates was carried out. Clade separation was determined utilizing the whole chloroplast genome alignment and the resultant NJ phylogenetic tree. A group was identified as a separate clade only when it had a bootstrap support value of 70% or above. mVISTA similarity patterns were employed to confirm greater sequence similarity within clades compared to those between clades. Oryza sativa functioned as the outgroup to ascertain the root position and substantiate the divergence of the principal lineages. To ensure structural completeness before phylogenetic analysis, the chloroplast genome’s GC content and gene composition were also summarized.

Results

Surface characteristics of two P. centrasiaticum germplasms under stereomicroscope

When the two germplasms had grown to the four-leaf stage, the seedlings were removed from the pots, and their leaf sheaths and leaves were observed under a stereomicroscope. White bristles were observed on the surface of P. centrasiaticum DB leaf sheaths, whereas the leaf sheaths of P. centrasiaticum NB were found to be smooth and without bristles (Fig. 1A). The differences between the two germplasms were further reflected in the leaves. On the surface of P. centrasiaticum—DB, high-density bristles were observed on both sides of the leaves (Fig. 1B). The nature of these bristles, when compared to trichomes seen on Poaceae leaves, led researchers to conclude that they are epidermal outgrowths that are neither glandular nor glandular-like.

Fig. 1.

Leaf-sheaths (A) and leaves (B) of two P. centrasiaticum germplasms under a stereomicroscope at the seedling stage. The upper part of the figure is P. centrasiaticum—DB, and the lower part is P. centrasiaticum – NB. Note: Scale bar = 1 mm.

Analysis of phenotypic traits of P. centrasiaticum germplasms

Table 2 displayed the average values of phenotypic characteristics for the two P. centrasiaticum germplasms. P. centrasiaticum—DB was found to exhibit higher values in PH (71.58 cm), SD (0.16 cm), NAGI (4.17), EL (9.33 cm), HR (42.99 percent), LARRS (36.43), LARVS (36.87), EWR (7.91 percent), SLR (1.45), and EE (16.40 percent). Significant differences were observed between the two germplasms in PH, NAGI, EL, EWR, and EE, with P. centrasiaticum—NB being outperformed in growth and nutrient parameters. In contrast, P. centrasiaticum—DB was recorded to have lower AIL (15.18 cm), EW (0.67 cm), FDR (2.22), CP (7.60 percent), CF (386.30 g/kg), and Ash (8.06). These results indicated that P. centrasiaticum—DB had better palatability for cattle than P. centrasiaticum—NB, due to its significantly lower CF. These findings indicated that the two germplasms varied significantly in nutritional and agronomic characteristics, with lower CF values in P. centrasiaticum—DB, suggesting that it may have better palatability.

Table 2.

Mean value of phenotypic traits of two P. centrasiaticum germplasms.

Phenotypic traits	P. centrasiaticum—DB	P. centrasiaticum—NB
PH (cm)	71.58 ± 5.31a	55.77 ± 3.27b
SD (cm)	0.16 ± 0.01a	0.15 ± 0.01a
NAGI	4.17 ± 0.31a	3.17 ± 0.17b
AIL (cm)	15.18 ± 0.90a	15.78 ± 0.70a
EL (cm)	9.33 ± 0.68a	5.82 ± 0.47b
EW (cm)	0.67 ± 0.03a	0.68 ± 0.04a
HR (%)	42.99 ± 8.39a	28.98 ± 3.25a
LARRS	36.43 ± 3.45a	30.51 ± 3.73a
LARVS	36.87 ± 6.94a	23.13 ± 2.45a
EWR (%)	7.91 ± 1.31a	4.30 ± 0.67b
SLR	1.45 ± 0.17a	1.39 ± 0.23a
FDR	2.22 ± 0.11a	2.32 ± 0.18a
CP (%)	7.60 ± 0.71a	7.78 ± 0.04a
EE (%)	16.40 ± 0.69a	12.20 ± 0.52b
CF (g/kg)	386.30 ± 6.99a	396.00 ± 2.60a
Ash (%)	8.06 ± 0.14a	8.22 ± 0.07a

Note: Trait marked with "a" after values measured in different plants indicated no significant difference. Traits marked with "a" after the value measured in one plant and "b" after the value measured in another plant indicated significant differences (P < 0.05).

Genome organization and features

The complete chloroplast genome of P. centrasiaticum germplasms displayed a circular DNA with a typical quadripartite structure (Fig. 2), with 138291 bp in P. centrasiaticum—DB, larger than 138283 bp in P. centrasiaticum—NB.

Fig. 2.

Chloroplast genome maps of P. centrasiaticum germplasms. Genes inside the circle were transcribed clockwise; genes outside were transcribed counterclockwise. In the inner circle, the AT content was represented by light gray, and the GC content was represented by dark gray. Different colors categorize genes belonging to various functional groups.

All P. centrasiaticum chloroplast genomes shared the common feature, contained a large single copy region (LSC) of 81,286 bp in P. centrasiaticum—DB and P. centrasiaticum—NB in 81,309 bp, a small single copy region (SSC) of 12,417 bp in P. centrasiaticum—DB and 12,352 bp in P. centrasiaticum—NB, a pair of inverted repeat regions (IRs) separated by the LSC and SSC of 44,588 bp in P. centrasiaticum – DB and 44,622 bp in P. centrasiaticum—NB, as well as coding sequence (CDS) of 57,318 bp in P. centrasiaticum—DB and 57,249 bp in P. centrasiaticum—NB. The GC content of the whole chloroplast genome in two P. centrasiaticum germplasms was 38.60%, and the GC content in two P. centrasiaticum germplasms was 36.50%, 44.10%, and 39.10% in LSC, IR, and CDS, respectively. While in the SSC region, the GC content of P. centrasiaticum was 33.10% in P. centrasiaticum—DB and 33.00% in P. centrasiaticum—NB. There were 74 CDS genes, 29 tRNA genes, and four rRNA genes, of which 10 have introns, five protein-coding genes (atpF, ndhA, ndhB, rpl2, rps12), and five tRNA genes (trnA-UGC, trnI-GAU, trnK-UUU, trnL-UAA, trnV-UAC) in the two germplasms (Table 3, Table 4, Table S1, Table S2).

Table 3.

Summary of chloroplast genomes sequenced in P. centrasiaticum germplasms.

Genome characteristics	P. centrasiaticum—DB	P. centrasiaticum—NB
Genome (bp)	138,291	138,283
LSC (bp)	81,286	81,309
SSC (bp)	12,417	12,352
IR (bp)	44,588	44,622
CDS (bp)	57,318	57,249
Overall GC content (%)	38.60	38.60
GC content in LSC (%)	36.50	36.50
GC content in SSC (%)	33.10	33.00
GC content in IR (%)	44.10	44.10
GC content in CDS (%)	39.10	39.10
CDS genes	74	74
tRNA genes	29	29
rRNA genes	4	4

Table 4.

Chloroplast genome gene content and functional classification in P. centrasiaticum germplasms.

Category	Group of genes	P. centrasiaticum—DB and P. centrasiaticum—NB
Self-replication-related genes	ribosomal RNAs	rrn16, rrn23, rrn4.5, rrn5
	transfer RNAs	trnA-UGCa, trnC-GCA, trnD-GUC, trnE-UUC, trnF-GAA, trnfM-CAU, trnG-GCC, trnH-GUG, trnI-CAU, trnI-GAUa, trnK-UUUa, trnL-CAA, trnL-UAAa, trnL-UAG, trnM-CAU, trnN-GUU, trnP-UGG, trnQ-UUG, trnR-ACG, trnR-UCU, trnS-GCU, trnS-GGA, trnS-UGA, trnT-GGU, trnT-UGU, trnV-GAC, trnV-UACa, trnW-CCA, trnY-GUA
	ribosomal proteins (SSU)	rps11, rps12a, rps14, rps16, rps18, rps19, rps2, rps3, rps4, rps7, rps8
	ribosomal proteins (LSU)	rpl14, rpl2a, rpl20, rpl22, rpl23, rpl32, rpl33, rpl36
	RNA polymerase	rpoA, rpoB, rpoC1, rpoC2
Photosynthesis-related genes	photosystem I	psaA, psaB, psaC, psaI, psaJ
	photosystem II	psbA, psbB, psbD, psbE, psbF, psbH, psbI, psbJ, psbK, psbL, psbM, psbN, psbT, psbZ
	cytochrome b/f complex	petA, petB, petD, petG, petL, petN
	ATP synthase	atpA, atpB, atpE, atpFa, atpH, atpI
	Protease	clpP
	RubisCO large subunit	rbcL
	NADH dehydrogenase	ndhAa, ndhBa, ndhC, ndhD, ndhE, ndhF, ndhG, ndhH, ndhI, ndhJ, ndhK
Other genes	Maturase	matK
	other genes	ccsA, cemA
	hypothetical chloroplast reading frames(ycf)	ycf1, ycf15, ycf4
	Translational initiation factor	infA
	ORFs

^aIntron-containing genes.

Comparative analysis of sequence variation

The variation in the chloroplast genome sequence of two P. centrasiaticum germplasms was visible in Fig. 3. Compared with the chloroplast genome of Cenchrinae species, the chloroplast genomes between Pennisetum and Cenchrus showed remarkable changes in many regions. Some genes, such as matK, rps16, rpoC2, ccsA, and ndhI, showed sequence variation; the divergent regions among the compared genomes were the rps16-trnQ, trnfM-trnT, psaA-trnS, atpB-rbcL, ycf4-cemA, petA-psbJ, psbE-petL, trnI-trnL, ndhF-rpl32, trnL-trnI. Three intergenic spacer sequences showed variation among P. centrasiaticum germplasms: rps3-psaI, ycf1-ndhF, and ycf1. The three chloroplast genomes of P. centrasiaticum showed tiny sequence alterations, showing that both the plastome architecture and gene order were preserved. Noncoding intergenic spacers were the only portions that varied somewhat between germplasms, although coding parts were substantially identical. This pattern is consistent with the fact that many Poaceae species exhibit minimal intraspecific plastome variation.

Fig. 3.

Visualization of alignment between three Pennisetum germplasms and three Cenchrus species, with P. centrasiaticum—NB as the reference. The horizontal axis indicates the coordinates within the chloroplast genome. The vertical scale indicates the percentage of identity, ranging from 50 to 100%. Genome regions were colour-coded as exon, UTR, conserved noncoding sequences (CNS), and mRNA.

SSR analysis

MISA v2.1 software (https://webblast.ipk-gatersleben.de/misa/) was used to screen the chloroplast genome of P. centrasiaticum—DB and P. centrasiaticum—NB. And SSR analysis was performed using default parameters in MISA v2.1 with motif definitions: (unit—size, min—repeats) 1—10, 2—6, 3—5, 4—5, 5—5, 6 – 5 (1 is the length of the SSR repeat unit, and 10 is the number of repetitions, the rest may be deduced by analogy). The two P. centrasiaticum germplasms possessed the same SSR loci of 49 and the same SSR type (Table S3). There were four complex repeat types, 29 single-nucleotide repeat types (14 T-base repeats, 12 A-base repeats, 3 C-base repeats, and only one G-base repeat), four two-base repeat types, one three-base repeat type, nine four-base repeat types, and two five-base repeat types (Fig. 4A). P. centrasiaticum—DB in the LSC region possessed 38 SSRs, while P. centrasiaticum—NB possessed 39 SSRs. In the IRb region, P. centrasiaticum—DB possessed two SSRs, while P. centrasiaticum—NB possessed only one SSR. In the SSC region, P. centrasiaticum—DB and P. centrasiaticum—NB possessed the same SSRs of 7. In the IRa region, P. centrasiaticum—DB and P. centrasiaticum—NB possessed the same SSRs of 2 (Fig. 4B). This SSR distribution was typical of chloroplast genomes, with most copies clustered in the LSC region. P. centrasiaticum showed exemplary cpSSR structural conservation, as evidenced by a high proportion of A/T mononucleotide repeats and an identical total number of SSRs amongst germplasms. There were no germplasm-specific SSRs discovered, given the criteria specified.

Fig. 4.

The SSR analysis of the chloroplast genomes of P. centrasiaticum—DB and P. centrasiaticum—NB. (A) Number of SSRs in different types; (B) Number of SSRs in the LSC, SSC, and IR.

Codon usage analysis

The number of codons in the chloroplast genome is 20,259 in P. centrasiaticum—DB, and 20,236 in P. centrasiaticum—NB. The codons encoding leucine (Leu) are the most abundant in the chloroplast genome of P. centrasiaticum, 2232 in P. centrasiaticum—DB, and 11.02% of the total number of codons, 2227 in P. centrasiaticum—NB, and 11.01% of the total number of codons. The number of codons encoding cysteine (Cys) is the rarest, with 227 each. The most frequently used codon is ATT, encoding isoleucine (Ile), with 838 each. The least frequently used codon is TGC encoding cysteine, with 54 each. There are two codons with RSCU = 1, encoding the codon ATG for methionine (Met) and the codon TGG for tryptophan (Trp), while the remaining codons exhibit specific preferences. Among the codons with RSCU value > 1, except for the codon TCC encoding serine and the codon TTG encoding leucine, which ends with G/C, the rest end with A/T, indicating that codons in the chloroplast genome of P. centrasiaticum prefer to end with A/T (Table S4).

Phylogenetic analysis between P. centrasiaticum and Cenchrus species

Phylogenetic analysis was conducted using complete cp genome sequences of four P. centrasiaticum germplasms and three Cenchrus species, with Oryza sativa used as the outgroup (Fig. 5). Pennisetum and Cenchrus species were clustered into two clades. Clade I was composed only of P. centrasiaticum (support rate 100 percent), and Clade II included the three Cenchrus species (support rate 99 percent). The species P. centrasiaticum has outstanding plastome conservation, with all accessions clustering into a single, monophyletic group. According to chloroplast genome data, Pennisetum and Cenchrus are obviously distinct genera, and significant bootstrap values support the classification of the two genera as different clades.

Fig. 5.

The phylogenetic tree was constructed based on the chloroplast genomes of four P. centrasiaticum and three Cenchrus accessions by the Neighbor-Joining (NJ) method, with Oryza sativa as the outgroup. Numbers above the branches represent the posterior probabilities.

Discussion

Survival strategies of P. centrasiaticum with bristles on the surface

Over the past decade, numerous studies have demonstrated that trichomes play a significant role in the interaction between plants and their environment. Among them, the density and length of the trichome are essential influencing factors³⁹. In winter, trichomes can absorb up to 80% of the radiation, causing the temperature of the leaves to be slightly higher than that of the surrounding environment. However, in rainy summer conditions, the trichome’s absorption of incident radiation decreases to around 30% to 40%, causing the temperature of the leaves to be lower than the environment⁴⁰. In addition, trichomes can cover some stomata, thereby reducing water loss, which plays a significant role in water retention and helps alleviate leaf temperature⁴¹. Moreover, trichome improves plant disease resistance⁴² and insect resistance⁴³. In Tibet, high altitude, low temperature, and intense radiation make the living environment for plants very harsh. With bristles on the surface, P. centrasiaticum—DB may possess more advantages in such harsh environmental conditions than P. centrasiaticum—NB.

Comparison of two P. centrasiaticum germplasms

P. centrasiaticum is widely distributed in arid and semiarid areas of Tibet, with spreading rhizomes, exhibits diverse drought resistance²⁵, and is valuable in mountain areas due to high biomass production and high grazing tolerance²⁶. In our study, the PH and NAGI of P. centrasiaticum—DB were significantly higher than those in P. centrasiaticum—NB. This grass possessed a significant morphological variation in the Yarlung Zangbo River Basin²⁰. Utilized as forage, P. centrasiaticum—DB possessed higher yield potential. The study of wild P. centrasiaticum in the Shannan area of Tibet⁴⁴, Gansu Province⁴⁵, and Inner Mongolia⁴⁶ of China showed that the plant possessed high protein and low fiber content, which could be considered a good pasture. Forages form a major part of dairy cows’ diets in most farming systems, sometimes containing significant amounts of fat and polyunsaturated fatty acids⁴⁷. Crude fat shows a positive correlation with the solid nonfat (SNF) and total solids (TS) in yak milk, which influences the quality of yak milk⁴⁸. In our study, the fat content of two P. centrasiaticum was higher than 10%, which provided more energy to sheep and yak in Tibet. Moreover, the fat content of P. centrasiaticum—DB was significantly higher than in P. centrasiaticum—NB, which provided a low-cost approach in comparison with diet supplementation strategies, such as oils and starch, which are designed to improve milk fatty acid profiles relative to their impacts on the health of humans that consume them⁴⁹. The HR and EWR were higher in P. centrasiaticum—DB than in P. centrasiaticum—NB, revealing that more P. centrasiaticum—DB seeds could be collected for further utilization.

Comparison of two P. centrasiaticum chloroplast genomes with other reported chloroplast genomes of Pennisetum

This study employed second-generation high-throughput sequencing technology, specifically the Illumina HiSeq 2500, to assemble and annotate the chloroplast genomes of two P. centrasiaticum germplasms. The chloroplast genomes of the two species were similar in structure to those of most angiosperms. Both have a typical quadripartite structure^17,50; their lengths were 138,291 and 138,283 bp, respectively. Both chloroplast genomes encode 107 genes, including 74 protein-coding sequences, 29 tRNAs, and four rRNA genes. In addition, the chloroplast genome structure, the number and type of genes, and GC content of two P. centrasiaticum samples were similar to those of the published chloroplast genome of P. centrasiaticum in Shigatse¹⁹, which revealed that the chloroplast genome structure of this species is relatively conserved. The SSRs of the chloroplast genome were efficient markers used in the study of population genetic structure and pedigree geography^51,52. In this study, 49 SSR loci were obtained in the chloroplast genomes of P. centrasiaticum. These SSR loci can provide candidate markers for genetic diversity and conservation genetics. The chloroplast genome SSR sites of the two species in this paper were mainly distributed in the LSC region and were primarily composed of A/T bases, similar to the distribution of SSR sites in the chloroplast genome of most angiosperms^17,53.

Relationship between P. centrasiaticum and Cenchrus species based on the complete chloroplast genome

Cenchrus and Pennisetum were two closely related genera^3,4, characterized by a highly modified inflorescence³. The two genera differ in the morphology of bristles; bristles in Cenchrus always showed some degree of fusion at the base, while in Pennisetum they were usually free⁵⁴. Nevertheless, Pennisetum and Cenchrus were not clearly defined due to variability in the basal fusion of the bristles. So, Correll and Johnston⁵⁵ considered the genus Pennisetum as a part of Cenchrus. Sequences of rbcL, ycf1, ndhF⁵⁶, and intergenic spacers of trnG-trnR, trnL-trnF, psbK-psbI^57,58 in the plant chloroplast genome were often used in phylogenetic studies of species. Studies using chloroplast markers (ndhF, trnL-trnF)^59,60 found Cenchrus to be monophyletic and embedded within Pennisetum, trnL-trnF and rpl16 markers in the chloroplast⁶¹ were also used in the molecular phylogeny of Pennisetum and Cenchrus. These authors suggested that additional molecular markers were needed to confirm their findings. Based on the complete chloroplast genome, P. centrasiaticum and Cenchru species were divided into two clades in this study. Therefore, the genera Cenchrus and Pennisetum should be kept separate rather than merged into a single genus; however, the chloroplast genomes of other species of the Subtrib. Cenchrinae has not been sequenced and published, and its overall phylogenetic relationship has not yet been reconstructed. The characteristics of the chloroplast genome of P. centrasiaticum germplasms reported in this study provide essential genetic information for studying phylogeny and evolution.

Phenotypic variation despite highly conserved chloroplast genomes

Although there were very few structural or sequence variations in the chloroplast genomes of P. centrasiaticum—DB and P. centrasiaticum—NB, there was a noticeable difference in plant height, nutritional attributes, and seed characteristics between the two germplasms. Different types of grasses have this pattern in detail. Although the upland and lowland ecotypes of Panicum virgatum exhibit significant morphological and physiological differences, there is slight intraspecific variation in their chloroplast genomes⁶². Moreover, Crassulaceae, like other angiosperm families, have sluggish chloroplast genome evolution and usually keep their gene content and structure consistent across populations or ecotypes⁶³. Another study has shown that cpDNA does not often have enough variety to account for phenotypic diversity within species^62,64. Biomass, nutrient content, and reproductive capability are among the many agronomic parameters regulated by nuclear genes. Chloroplast development, photosynthesis, stress tolerance, and growth responses are all controlled by the nuclear genome, which encodes about 95% of chloroplast proteins⁶⁵. A further important factor in the expression of traits is the interaction between the nuclear and plastid genomes; in cases when the two genomes are genetically distinct, nuclear genome divergence can lead to noticeable phenotypic variations⁶⁶. Different morphological and nutritional traits are present in P. centrasiaticum—DB and P. centrasiaticum—NB germplasms due to the influence of environmental gradients in Tibet, including heat, radiation, and moisture. Accordingly, the modest evolutionary rate and functional stability of cpDNA are mirrored in P. centrasiaticum’s preserved chloroplast genome. The striking variations in fat content, seed recovery rate, and forage production, on the other hand, are most likely due to physiological flexibility and nuclear genetic divergence. Ecotypes of fodder grasses exhibit substantial dietary diversity without observable plastome variations, a finding that is consistent with previous reports⁶². The findings show that cpDNA is a reliable tool for studying evolutionary relationships and tree of life patterns, but it cannot account for the differences in crop yields observed in the two germplasms.

Author contributions

Conceptualization, L.T.Z., Y.S.M., and A.H.E.; writing—original draft preparation, L.T.Z., Y.S.M., A.H.E., L. Y. H., L. T., Y. L. L., S. X., and M. D. D.; writing—review and editing, L.T.Z., S.A.S., D.M.L., S.H.Y., H.Y.W., Y.S.M., X.L.W., and A.H.E.; Funding, L.T.Z. and Y.S.M.; project administration, L.T.Z. and Y.S.M.; All authors contributed to the article and approved the submitted version.

Funding

The Ph.D supported this research. Fund Project of Yibin University (2023QH08), Solid-state Fermentation Resource Utilization Key Laboratory of Sichuan Province (2022GTYY11), and The Chief Scientist Program of Qinghai Province 2024-SF-101.

Data availability

The complete cp genome of *P. centrasiaticum* germplasms with their gene annotation presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found at NCBI GenBank with accession numbers OL672847.1 ( *P. centrasiaticum*—DB) and OL672848.1 ( *P. centrasiaticum*—NB), respectively.

Declarations

Competing interests

The authors declare no competing interests.