A mutation in POLR2A gene associated with body size traits in Dezhou donkeys revealed with GWAS

Abstract

The Dezhou donkey is a famous local donkey breed in China. The aim of the present study was to identify the genes associated with the body size traits of the Dezhou donkey and facilitate the breeding activities of the donkeys. A total of 349 donkeys from 2 generations (113 individuals in F0 and 236 in F1) were analyzed with restriction-site-associated DNA sequencing. A genome-wide association study revealed that the region between 13.7 and 15.6 Mb of chromosome 13 is significantly associated with body sizes. Candidate genes related to body size development, including POLR2A, CHRNB1, FGF11, and ZBTB4, were identified. The results of GO and KEGG analysis indicated that the genes involved in many GO terms were related to metabolic processes and developmental processes. Additionally, a T>C mutation (Chr13:14312485) was found at intron 10 of the POLR2A gene. The association analysis showed significant differences among genotypes for the size traits. The body size of the individuals with the TT genotype was significantly higher than that with the CC genotype. The results showed that the polymorphism of POLR2A has the potential to be used as a marker in the breeding programs of the Dezhou donkeys.

Loci significantly associated with donkey body size traits were identified with genomic data in the study, which can serve as a marker to improve donkey production performance.

Introduction

The Dezhou Donkey is an important indigenous breed in Shandong Province in the east of China, which was named for the prefecture of the city where the Dezhou Donkeys are mainly distributed (Seyiti and Kelimu, 2021). The Dezhou Donkey has a large stature and was regarded as having great potential for breeding into a superior breed for the production of meat and milk (Lai et al., 2021; Wang et al., 2022b). Body sizes are closely related to important economic traits, so it is crucial to study the genetic mechanisms associated with body size and the outcome will facilitate the breeding activities of the Dezhou Donkey (Metzger et al., 2013; Chen et al., 2020).

The genes regulating body sizes have been extensively studied in domestic animals. Special haplotypes of IGF1 gene were identified to be associated with body size in dogs (Sutter et al., 2007). Further studies showed that 46% to 52.5% of the variation of body size in dogs could be explained by 7 markers of the 6 genes of GHR, HMGA2, SMAD2, STC2, IGF1, and IGF1R (Rimbault et al., 2013). A genome-wide association study (GWAS) with the microarray genotyping data of 220 sheep revealed that 46 SNPs across 14 chromosomes were significantly associated with body size in sheep. These SNPs were associated with genes such as NCAPG, LCORL, and HMGA2 (Posbergh and Huson, 2021). A meta-analysis of the cattle genome showed that variations in 163 regions accounted for 13.8% of the phenotypic variation of body size (Bouwman et al., 2018). Selection signaling analysis based on whole genome resequencing data in pigs revealed 3 strong QTL-containing selection regions localizing to the NR6A1, PLAG1, and LCORL genes associated with increased body length and vertebrae number (Rubin et al., 2012). Similarly, there have been increasing studies related to the body size of horses (Asadollahpour Nanaei et al., 2020). A GWAS indicated that the genes NCAPG/LCORL, HMGA2, and ZFAT explained 83% of the body size variation in horses (Makvandi-Nejad et al., 2012). Another study found that an SNP within the enhancer of the TBX3 gene was a causative mutation of the variation in body size among Chinese domestic horses (Liu et al., 2022a).

A number of candidate genes for body size in donkeys were also identified in recent years and some of them overlapped with those from other livestock. Selection signal analysis of whole genome resequencing data from 12 donkey breeds revealed the NCAPG/LCORL is closely related to withers height traits (Shen et al., 2021). The LCORL/NCAPG, FAM184B, TBX3, and IHH genes were also identified as the main candidate genes for donkey body sizes in a study with selective signaling analysis of genomic data (Liu et al., 2022b). An A>G mutation was identified in the second intron of the TBX3 gene, of which the AA and GG genotypes differed significantly in body size traits in donkeys (Wang et al., 2021). Moreover, several other genes were also reported to be associated with donkey body sizes, such as NR6A1, CDKL5, and DCAF16 (Fang et al., 2019; Zhou et al., 2020). However, the genetic mechanism of the donkey body size traits, which is regulated by a network of genes, is far from being fully explained, and further in-depth studies are still necessary to be conducted to clarify the underlying mechanism of the important traits.

RNA polymerase II subunit A (POLR2A) is a vital gene regulating the development of mammals. The gene encodes the largest catalytic subunit in the RNA polymerase II complex, which is stably expressed in the testis, skin, and bone marrow (Chai et al., 2019). A review of medical research has revealed that the POLR2A gene is highly expressed in gastric cancer tissues and can promote the migration and proliferation of cancer cells (Jiang et al., 2021). Mutation in the POLR2A gene is associated with neurodevelopmental syndrome (Giacomini et al., 2022). POLR2A, as a housekeeping gene, is differentially expressed in tissues such as skeletal muscle in older and younger populations, which could help in the treatment of skeletal muscular dystrophy (Touchberry et al., 2006). Our study reveals a major and previously undescribed role of POLR2A associated with donkey body sizes with GWAS. It has the potential to be a valuable marker for the breeding of Dezhou Donkey.

In this study, GWAS was applied to identify the candidate genes for the body size traits of Dezhou donkeys based on restriction-site-associated DNA sequencing (RAD-seq) data of a large population of donkeys. The outcome of the study will facilitate the conservation and selection of Dezhou donkeys.

Materials and Methods

The samples were obtained following the principles approved by the Animal Care and Use Committee of China Agricultural University (permit number: XK257).

Sample collection and genomic DNA extraction

The study samples were obtained from the Shandong Dong’e Black Donkey Breeding Center. A total of 349 Dezhou donkeys were phenotyped for withers height, body length, chest girth, and cannon circumference (Supplementary Table S1). These donkeys included 113 and 236 donkeys from the F0 and F1 generations, respectively. The F0 generation consisted of 100 females and 13 males, while the F1 generation could be categorized into 13 families, including 119 males and 117 females. Genomic DNA were extracted with the Tissue Genomic DNA Extraction Kit (Tiangen Biotech Co., Ltd, Beijing) using skin tissues collected by a minimally invasive skin sampling method. The concentration and purity of DNA were quantified using NanoDrop 2000 spectrophotometer and stored at −20 °C.

RAD-seq library preparation, sequencing, and data quality control

RAD-seq was performed by Beijing Allwegene Technologies Co., Ltd using the Illumina Hiseq 2500 sequencing platform with paired-end 150 sequencing. The sequencing output of raw data were 641 Gb, while the clean data, following quality control filtering and quality control, totaled 515 Gb, with Phred values of Q20 ≥ 90% and Q30 ≥ 85%. The RAD-seq data of 349 Dezhou donkeys has been deposited in GenBank under BioProject accession PRJNA1125120. Quality control was conducted using Trimmomatic-0.39 software, a Java-based program, to filter low-quality reads. The filtered, high-quality sequencing data were compared to the reference genome (ASM303372v1, GCA_003033725.1) using BWA software. The results were formatted, sorted, de-duplicated, and indexed by SAMTOOLS. Population SNP detection and extraction were performed using GATK software. Mutations with low-quality values were removed using vcftools, and SNPs with more than 20% missing data (––max-missing 0.8) or MAF lower than 5% (––maf 0.05) were discarded.

SNPs were filtered using PLINK 1.9 (Purcell et al., 2007), and the following SNPs were discarded: (1) SNPs were missing more than 10% of their genotype data; (2) SNPs with a minor allele frequency <1%; (3) Samples with SNP call rates lower than 80%; and (4) SNPs with Hardy-Weinberg equilibrium P-value < 1E−6. The GWAS was performed with the remaining 433,767 SNPs and 294 individuals after quality control (Supplementary Table S2).

GWAS for body-size traits

The GWAS was conducted with GEMMA software based on univariate and multivariate mixed linear models, respectively (Zhou and Stephens, 2012). The mixed linear model included fixed effects such as population genetic structure, individuals’ age, and sex, as well as random effects like individual genetic relationship. The univariate linear mixed model for each SNP is as follows:

$${\displaystyle \begin{array}{l}{\displaystyle y=X\alpha +Z\beta +W\mathrm{\mu }+e}\end{array}}$$

Where y is the observations of body size traits, X is the matrix of fixed effects, α is the estimated parameter for the fixed effects, Z is the indication matrix of SNP, β is the marker effect of the SNP, W is the indicator matrix for the random effects, µ is the predicted random individual, and e is the random variance, which follows (0, δe²).

The multivariate-based GWAS analysis is also based on the statistical approach of a mixed linear model, which is modeled as follows:

$${\displaystyle \begin{array}{l}{\displaystyle Y=WA+x{\beta }^T+U+E}\end{array}}$$

Y is the n × d phenotype matrix composed of phenotypic data, n is the number of samples, d is the number of phenotypes, W is an n × c covariate matrix (fixed effects), A is a c × d coefficients matrix corresponding to the covariate matrix, x is an n-vector of genotypes, β is the effect sizes of the SNPs for the d traits, U is the random effects of n × d matrix and E is an n × d matrix of errors.

The GEMMA software employed the null hypothesis of β = 0 in conducting significance tests on each SNP, utilizing the Wald hypothesis test method. The P-value derived from the GWAS was subjected to correction for multiple tests using the Bonferroni method, with the objective of reducing the probability of false positives (Nicodemus et al., 2005). Manhattan plots and QQ plots were generated using the R package “qqman.” The thresholds for statistical significance were set at −log10 (5 × 10⁻⁸) and −log10 (1 × 10⁻⁵), representing genome-wide significance and suggestive significance, respectively (Turner, 2014; Schaid et al., 2018). The percentage of phenotypic variation explained by SNP was calculated using GCTA software and self-coded R language code.

Gene annotation and enrichment analysis

Gene annotation was conducted within a 50-kb region upstream and downstream of the physical location of the significantly associated SNP loci using BEDTools software (Quinlan and Hall, 2010). GO and KEGG pathway enrichment analyses were performed for significantly associated genes using KOBAS software (Xie et al., 2011). The P-value of the enrichment result were corrected for multiple tests using the Benjamini method.

Association analysis of gene polymorphisms with phenotypes

A total of 175 donkeys, aged 3 years or older, were selected from the 349 individuals sequenced for association analysis of genotype and phenotype data (Supplementary Table S3). Association analyses were performed using general linear modeling in SPSS software, in which genotype effect was a fixed factor and phenotype value was a dependent variable. The model of analysis is as follows:

$${\displaystyle \begin{array}{l}{\displaystyle y_{ijk}=\mu +G_i+{\alpha }_j+{\beta }_k+p_l+e_{ijkl}}\end{array}}$$

y _ijk is the measured value of the body size traits, μ is the population mean, G_i is the genotype effect, α_j is the sex effect, β_k is the age effect, p_l is the kinship effect, and e_ijk is the random residual.

Results

GWAS for the donkey body size traits

A GWAS was performed with 294 Dezhou donkeys for the body size traits using univariate and multivariate linear mixed models. The association analysis for the withers height trait revealed 6 potentially associated SNPs on chromosome 13, forming a single association signal interval ranging between 13.7 and 14.7 Mb on chromosome 13 (Figure 1a). Similarly, the analysis of the body length trait also showed an extremely significant association region on chromosome 13, specifically ranging from 14.3 to 14.7 Mb (Figure 1b). The trait for chest girth showed potential association with SNPs loci on chromosomes 3, 5, 13, 28, and 29 (Figure 1c). Nine identified SNPs loci were significantly associated with the cannon circumference trait at the genomic level. These loci were mainly located on chromosome 13 and formed a strong association signal interval ranging between 13.7 and 15.6 Mb (Figure 1d). The uni-trait and multivariate GWAS analyses showed similar results, both having highly significant association signal peaks on chromosome 13 (Figure 1e).

Figure 1.

Manhattan plots indicate the significant threshold and suggestive threshold for GWAS of different body size traits: (a) withers height, (b) body length trait, (c) chest girth, (d) cannon circumference, and (e) analysis on all of the body size traits.

Gene annotation and enrichment analysis of significant SNP

50-kb upstream and downstream of each significant SNP were genetically annotated based on the donkey reference genome. The genes annotated within the interval (Chr13: 13.7 to 14.7 Mb) significantly associated with the withers height trait included POLR2A, ZBTB4, CHRNB1, and FGF11 (Table 1). The analysis of the body length trait yielded similar results to the withers height trait, and the genes annotated by the significant SNPs were almost the same as those for the withers height trait. The significant association interval involving the cannon circumference trait (Chr13: 13.7 to 15.6 Mb) overlapped with the intervals associated with the withers height and body length traits (Table 1).

Table 1.

Gene annotation results of significant SNPs from the GWAS

	Chromos-ome	Position	P-Value	Genes
Withers height	Chr13	14312485	2.31E-09	POLR2A, SLC35G3, ZBTB4, CHRNB1, FGF11
	Chr13	14775333	4.48E-08	BAP18, RNASEK, ALOX12
	Chr13	13708938	3.74E-06	CTC1, AURKB, BORCS6, TMEM107, VAMP2, PER1
	Chr27	6568285	4.61E-06	DEFA1
	Chr12	68025118	4.70E-06	GRHPR, TXNL4A, Pqlc1
	Chr13	14260180	6.10E-06	FXR2, SOX15, Mpdu1, CD68, EIF4A1, Senp3, TNFSF13, TNFSF12
Body length	Chr13	14312485	5.76E-10	POLR2A, SLC35G3, ZBTB4, CHRNB1, FGF11
	Chr13	14775333	8.58E-10	BAP18, RNASEK, ALOX12
	Chr13	2187909	4.83E-06	TRIP11
	Chr03	1.2E+08	1.25E-05	ZBTB49, NSG1, STX18, Nova1
Chest girth	Chr28	10783708	2.84E-06	MTHFD1L, PARD3
	Chr29	34831338	9.53E-06	PARD3
	Chr05	97345075	1.40E-05	CCDC80, SLC35A5
	Chr03	1.54E+08	1.43E-05	GC
	Chr13	14312512	2.15E-05	POLR2A, SLC35G3, ZBTB4, CHRNB1, FGF11
Cannon circumference	Chr13	14312485	8.89E-13	POLR2A, SLC35G3, ZBTB4, CHRNB1, FGF11
	Chr13	14775333	1.40E-12	BAP18, RNASEK, ALOX12
	Chr13	14260180	3.02E-09	FXR2, SOX15, Mpdu1, CD68, EIF4A1, Senp3, TNFSF13, TNFSF12
	Chr13	15641402	2.35E-08	C1qbp, RID1, Rpl36a, DERL2, MIS12
	Chr13	14115228	1.56E-07	DNAH2, Efnb3, WRAP53
	Chr13	13708938	2.45E-07	CTC1, AURKB, BORCS6, TMEM107, VAMP2, PER1
	Chr13	15282925	4.09E-07	CAMTA2, INCA1, KIF1C, ZFP3
	Chr15	2046302	8.55E-07	CSTL1, Cst11, Napb
	Chr13	15349999	8.60E-07	ZNF594

The annotation that was significantly associated with the multivariate GWAS still related to the genes POLR2A, SLC35G3, ZBTB4, CHRNB1, and FGF11. Enrichment analysis of the annotated genes revealed significant enrichment in GO entries, including terms related to skeletal phylogeny (GO:0001501), macromolecular complex binding (GO:0044877), protein binding (GO:0005515), and signaling (GO:0007165; Figure 2). The KEGG enrichment analysis showed significant enrichment in the phospholipase D signaling pathway (hsa04072) and the PI3K-Akt signaling pathway (hsa04151; Figure 2). The annotated genes were significantly enriched pathways involved in cell proliferation, growth, and skeletal system development.

Figure 2.

The GO terms and KEGG pathways of the genes identified from the mixed multitrait GWAS analysis of body size traits.

Association analysis of polymorphisms of POLR2A with body size traits

The GWAS analysis revealed that locus Chr13:14312485 in the POLR2A gene was associated with withers height, body length, and cannon circumference In univariate association analyses, a T>C mutation was found in the 10th intron of the POLR2A (Supplementary Figure S1). This locus could explain 7.9% of the variation in body height and 2% of the variation in cannon circumference, respectively. Additionally, Chr13:14312512 was associated with chest girth. Linkage disequilibrium analyses indicated that the 2 SNPs were linked in the same haplotype block (Figure 3). The association analysis of different genotypes of Chr13:14312485 with the 4 body size traits indicated that there were significant differences in the withers height among the groups with CC, CT, and TT genotypes at the locus (Table 2). Similarly, for body length and chest girth, significant differences were observed between TT and CC genotypes. Additionally, the cannon circumference of individuals with TT or CT genotypes showed a significant difference compared to those harboring CC genotypes.

Figure 3.

Linkage disequilibrium plot for significantly associated SNPs.

Table 2.

Association analysis between g.Chr13:14312485 T>C polymorphism and body size traits of Dezhou donkeys

Genotype	CC (n = 80)	CT (n = 70)	TT (n = 25)
Withers height	134.52 ± 0.69c	137.41 ± 0.57b	140.28 ± 1.39a
Body length	133.76 ± 0.80b	136.46 ± 0.96ab	139.60 ± 1.52a
Chest girth	148.38 ± 0.83b	150.19 ± 0.92ab	151.48 ± 1.84a
Cannon circumference	15.54 ± 0.16b	16.99 ± 0.15a	17.06 ± 0.31a

The difference between the values with different superscript letters in the same row reaches a significant level (p < 0.05).

Abbreviation: SE, standard error.

Discussion

China has been one of the major countries with a large donkey population. However, mechanization of agriculture led to a sharp drop in the donkey population size in China. The number of donkeys in China decreased from 9.23 million in 2000 to a little more than 2 million in 2020 (http://www.stats.gov.cn/tjsj/). The Dezhou donkey, a famous indigenous breed in China, has also experienced a notable population decline in recent years. In the past two decades, there has been an increasing demand for donkey meat, milk, and skin for the production of traditional Chinese medicine named Ejiao (Wang et al., 2022a). Consequently, breeding specialized donkey breeds with excellent performance to meet market demand becomes an urgent task, which is of great significance to improve the economic benefits of donkey breeding and promote the development of the donkey industry (He et al., 2019).

Body size traits serve as pivotal criteria for assessing the growth and production performance of livestock (Kader et al., 2016). Enhanced selection for body size traits in breeding efforts can also improve skin and meat production in donkeys. With the development of molecular genetics, the methods of animal breeding have also been updated from conventional selection to molecular marker-assisted selection (MAS) and genomic selection (GS; Shendure et al., 2008; Walsh, 2021), and mining functional genes and molecular markers for body size traits of donkeys will provide the foundational work for the MAS and GS in donkeys. Previous studies have revealed a number of candidate genes and mutations which are associated with donkey body size. The 5 SNPs in the NCAPG-DCAF16 region of the Dezhou population have significant effects on growth traits, especially the rs008 locus, which has a significant correlation with important growth traits such as body weight and body height (Hou Haobin, 2019). Six of the 11 selected genes identified in 6 donkey breeds were related to body size, namely TBX3, NCAPG, LOCR, BCOR, CDKL5, and ACSL4 (Zhou et al., 2020). An association study on the polymorphisms and expression of MSTN showed a correlation between the gene and traits of growth in Chinese donkeys (Liu et al., 2017). The mutation in Cytb was considered as a molecular marker for the selection of rump width and body height in donkeys (Chen et al., 2009). A study analyzing the effects of ACSL3 gene polymorphism on the growth traits of the donkey indicated that the Indel locus in the intron region of the ACSL3 gene is related to the growth traits of male and female Dezhou donkeys (Lai et al., 2020). Furthermore, a recent study highlighted the SPRY2, COL9A2, MIR30C, HEYL, BMP8B, LTBP1, FAM98A, and CRIM1 genes, which are involved in the development of structural tissues such as bone and cartilage and were associated with body size traits in horses (Bastos et al., 2023).

A GWAS was conducted in the present study on the 4 body size traits, including withers height, body length, chest girth, and cannon circumference. A region of significant association, Chr13:13.7 to 15.6 Mb, was confirmed in both single-trait analyses and mixed multitrait analyses. The results of GO analysis indicated that the genes within the strongly associated region involved in many GO terms related to metabolic processes and developmental processes. The KEGG analysis significantly enriched the genes within pathways associated with cell proliferation, actin cytoskeleton remodeling, cell growth, and glycogen metabolism. Several important genes were identified, which do not overlap with the candidate genes reported previously. In our study, a previously undescribed gene, POLR2A, was found to correlate with body size in donkeys.

A T>C mutation (Chr13:14312485) of POLR2A gene was significantly associated with body size traits (P-values < −log10 (5 × 10⁻⁸)). The donkeys with TT or CT genotype at the locus tend to have higher values of the traits than those carrying CC. However, further studies including a large sample size, still need to verify the results due to the relatively small number of donkeys involved in the present study. A previous study has indicated that POLR2A plays a pivotal role in the treatment of osteoporosis and is associated with osteoclast differentiation, which may serve as a potential biomarker and therapeutic target (Sobacchi et al., 2013). The POLR2A gene has the potential to prevent and treat menopausal osteoporosis by reducing bone resorption and increasing bone mass in mice (Liu et al., 2021). A study also identified POLR2A in an associated region on chromosome 11 associated with the body size of sheep (Posbergh and Huson, 2021). A similar study in sheep demonstrated that POLR2A is a functional gene associated with sheep’s body weight (Kominakis et al., 2017). In addition, POLR2A expression may affect the rate of protein deposition in fish, which correlates with the rate of skeletal muscle growth in mandarin fish (Yao et al., 2024). Based on the results from both present and previous studies, the identified SNP, Chr13:14312485, may have the potential to be used for the breeding activities of Dezhou donkeys.

Conclusion

An SNP (Chr13:14312485) within the 10th intron of the POLR2A, which was significantly associated with donkey body size traits, was identified in our study and may be used as a potential molecular marker for the breeding programs of donkeys.

Glossary

Abbreviations

GWAS

genome-wide association study

POLR2A

RNA polymerase II subunit A

RAD-seq

restriction-site-associated DNA sequencing

Conflict of interest statement

The authors declare no real or perceived conflicts of interest.