Codon Usage Bias in the Chloroplast Genome of the Traditional Chinese Medicinal Plant Corydalis yanhusuo

1 Introduction

Chloroplasts are specialized organelles in green plant cells that carry out photosynthesis. The chloroplast genome contains both protein-coding genes and non-coding regulatory regions, and it participates in essential biological processes—including light capture, electron transport, photophosphorylation, and carbon fixation—through the expression of numerous photosynthesis-related proteins and enzymes [1]. Because of features such as a highly conserved structure, maternal inheritance, and a generally low rate of recombination, chloroplast genomes are widely applied in research on plant phylogeny, molecular evolution, and phylogeography [2]. In recent years, chloroplast genomics has emerged as a major research area in plant biology, and genomes from diverse plant species have now been sequenced and characterized. These studies are of great importance for advancing our understanding of photosynthesis [3], genetic traits [4], species evolution [5], and environmental adaptation [6].

Codon usage bias is defined as the phenomenon in which organisms differ in the frequency with which they use synonymous codons that encode the same amino acid during protein translation. Studies have shown that the usage frequency of synonymous codons exhibits preferences at multiple levels, including interspecific, intraspecific, and intergenic levels [7]. It is generally recognized that differences in the cellular abundance of various tRNAs directly influence the frequency of synonymous codon usage; specifically, the greater the abundance of a tRNA, the more frequently its corresponding codon is used. To satisfy the need for rapid synthesis of large quantities of proteins, highly expressed genes in organisms preferentially employ codons that match high-abundance tRNAs [8]. Additionally, other factors such as mutational pressure, genomic GC content, and protein structure also influence codon usage bias [9]. Therefore, analyzing codon usage bias helps deepen our understanding of molecular evolution in different species and the mechanisms regulating gene expression.

Corydalis yanhusuo W. T. Wang is a perennial herbaceous plant belonging to the genus Corydalis in the family Papaveraceae. Its dried tubers are used as the traditional Chinese medicinal material “Yuanhu”, which has notable efficacy in promoting blood circulation, regulating qi, and relieving pain [10]. Corydalis is the largest genus in Papaveraceae, with approximately 288 species recorded in China; members of this genus are widely distributed across both northern and southern China, particularly in the southwest region [11]. At present, chloroplast genome sequences of dozens of Corydalis species—such as C. yanhusuo, C. turtschaninovii [12], C. temulifolia [13], C. edulis, C. shensiana [14], and C. fangshanensis [15]—have been published and used for phylogenetic and evolutionary analyses, greatly enriching the available genetic information on Corydalis species.

Current studies on C. yanhusuo mainly focus on processing methods [16], chemical constituents [17], pharmacology and clinical efficacy [18], whereas studies on the codon usage bias of its chloroplast genome remain relatively limited. In this study, the chloroplast genome of C. yanhusuo was obtained through resequencing, and CodonW 1.4.2 software, together with EMBOSS online tools, was used to analyze the codon base composition characteristics of protein-coding genes and to examine the factors influencing codon bias. An in-depth investigation of the chloroplast genome of C. yanhusuo will contribute to further research on its genetic evolution and chloroplast genetic engineering.

2 Methods

2.1 Plant Materials

Fresh tubers of C. yanhusuo were collected on May 10, 2022, from Lefeng Village, Shangyuanguan Town, Chenggu County, Hanzhong City, Shaanxi Province, China (latitude 33°5′15″ N, longitude 107°13′49″ E; altitude 481 m). Tubers with smooth surfaces and no mechanical damage were selected and planted in plastic flowerpots filled with nutrient substrate, and the surface was covered with an appropriate amount of substrate. All tubers were cultured at 20°C and a relative humidity of 50–70%, under a 12 h light (2000 lux)/12 h dark photoperiod, with timely watering. The plant material was identified as C. yanhusuo (Papaveraceae, Corydalis) by Associate Professor Xinjie Yang from the Herbarium of Shaanxi University of Chinese Medicine (Fig. 1). When the plants reached a height of 10–15 cm and had fully expanded leaves, fresh, healthy leaves were collected, rapidly frozen in liquid nitrogen, and sent to Sangon Biotech (Shanghai) Co., Ltd. for high-throughput sequencing of the chloroplast genome.

Figure 1: Morphological characteristics of Corydalis yanhusuo.

2.2 Chloroplast Genome Resequencing and Gene Screening

The chloroplast genome of C. yanhusuo was resequenced using the Illumina Novaseq 6000 platform. Following sequence assembly, annotation, and correction, the raw data were deposited in the Genome Sequence Archive of the National Genomics Data Center, Beijing Institute of Genomics, Chinese Academy of Sciences [19] (https://ngdc.cncb.ac.cn/gsa; Accession No.: CRA014252). To ensure data reliability, the coding sequences (CDS) were screened according to the following criteria: repeated sequences and sequences shorter than 300 bp were removed to avoid sampling bias, sequence lengths were ensured to be multiples of three, and each CDS was verified to start with the codon ATG and end with one of the stop codons TAA, TGA, or TAG. Ultimately, 50 chloroplast CDS of C. yanhusuo were selected for subsequent codon usage bias analysis.

2.3 Codon Composition Analysis

Based on the screened chloroplast CDS of C. yanhusuo, the GC content at the first, second, and third codon positions (designated GC1, GC2, and GC3, respectively) as well as the overall GC content of codons (GCall) were analyzed using the EMBOSS online tool (https://www.bioinformatics.nl/emboss-explorer/) and CodonW 1.4.2 software (https://codonw.sourceforge.net). Additional parameters, including the effective number of codons (ENC), relative synonymous codon usage (RSCU), codon adaptation index (CAI), and the frequency of guanine or cytosine at the third codon position (GC3s), were also calculated [20,21]. Pearson correlation analysis of these codon-related parameters was performed using SPSS 22.0 to assess relationships within the chloroplast genome of C. yanhusuo.

2.4 Neutral Plotting

A neutral plot analysis was performed using Excel 2019, with GC3 plotted on the x-axis and GC12 (the average of GC1 and GC2) on the y-axis. The factors influencing codon usage bias were inferred by examining the relationship among the first, second, and third codon positions. A significant correlation between GC12 and GC3 with a regression coefficient close to 1 indicates that codon bias is largely driven by mutation; otherwise, codon bias is mainly influenced by natural selection [22].

2.5 ENC-Plot Mapping

A scatter plot was generated with GC3s on the x-axis and the observed ENC values on the y-axis, and a standard curve of expected ENC values was added. The expected ENC was calculated using the formula: ENC = 2 + x + 29/[x2 + (1 − x)2], where x = GC3s. The closer the scatter points are to the standard curve, the stronger the influence of mutation on codon usage bias; conversely, the farther the points are from the curve, the more codon usage bias is determined by natural selection [23].

2.6 PR2-Plot Mapping

Parity Rule 2 (PR2) analysis was conducted using Excel. A scatter plot was generated with G3/(G3 + C3) on the x-axis and A3/(A3 + T3) on the y-axis, and horizontal and vertical symmetry lines were added. The center of the plot represents the point where A = T and G = C. If codon usage bias is mainly driven by mutation, the scatter points are expected to be evenly distributed around the plot [24].

2.7 Identification of Optimal Codons

Codons with RSCU values greater than 1 were identified as high-frequency codons. The 50 CDS sequences were ranked by their ENC values, and five genes with relatively low ENC (high-expression genes) and five with relatively high ENC (low-expression genes) were selected to construct high- and low-expression libraries, respectively. The ΔRSCU value was then calculated (ΔRSCU = RSCU of high-expression genes − RSCU of low-expression genes), and codons with ΔRSCU ≥ 0.08 were considered advantageous codons for high-expression genes [25]. The intersection of these high-frequency codons and high-expression ones was defined as optimal codons of the C. yanhusuo chloroplast genome.

3 Results and Discussion

3.1 Analysis of Codon Usage in the Corydalis yanhusuo Chloroplast Genome

The basic codon composition of the C. yanhusuo chloroplast genome is summarized in Table 1. Among the 50 screened genes, the number of synonymous codons ranged from 94 (rpl23) to 2224 (ycf2), with an average of 385.36, while the number of encoded amino acids ranged from 100 (ndhE, rpl23, and rps14) to 2311 (ycf2), with an average of 402.04. The ycf2 gene is the largest plastid gene in angiosperms and, owing to its long sequence and low nucleotide substitution rate, is frequently used in studies of angiosperm phylogenetic relationships. It has been reported that the ycf2 gene alone can generate a generally well-supported phylogenetic tree [26]. The protein hydrophobicity ranged from −1.122 (rps18) to 1.016 (ndhG), with an average of −0.083, while protein aromaticity (Aromo) ranged from 0.027 (rps11) to 0.183 (ndhC). These results indicate that the ndhG and ndhC genes, which encode reduced nicotinamide adenine dinucleotide (NADH) dehydrogenase subunit proteins, exhibit the highest hydrophobicity and aromaticity indices, respectively. Previous studies have shown that hydrophobicity and aromaticity are major factors influencing amino acid usage variation in the chicken proteome [27].

The GC content at the first, second, and third codon positions, the frequency of G or C at the third codon position (GC3s), the ENC, and the CAI for the CDS of the C. yanhusuo chloroplast genome are summarized in Table 2. The GCall content ranged from 34.35% to 48.67%, with an average of 40.06%; GC1 ranged from 37.15% to 57.98%, averaging 47.84%; GC2 ranged from 25.22% to 59.33%, with an average of 40.21%; and GC3 ranged from 24.45% to 39.53%, with an average of 32.12%. It has been reported that the plastome of the closely related species C. adunca exhibits the highest total GC content (41.03%), which is higher than that of the chloroplast genomes of C. saxicola, C. hsiaowutaishanensis, and C. davidii [28]. Kim [12] analyzed the chloroplast genomes of 36 Corydalis species and found that their total GC content ranged from 40.1% to 41.5%, higher than that observed in other genera of Papaveraceae, such as Eschscholzia californica, Chelidonium majus, and Lamprocapnos spectabilis, as well as in various green algae and land plants [29].

Codon usage bias in plant chloroplast genomes is influenced by numerous factors, among which gene mutation and natural selection are the most significant. GC content serves as an important indicator for assessing mutation tendencies [30]. In this study, the GC content at the first, second, and third codon positions followed the order GC3 < GC2 < GC1, indicating that GC was unevenly distributed across the three codon positions in the C. yanhusuo chloroplast genome. This pattern suggests that synonymous codons are relatively conserved, with a preference for ending in A or U. Similarly, Ren [31] reported that the third codon position in all coding genes of C. tomentella had a high AT content of 65.83%, and 65.91% in C. saxicola, reflecting the low GC3 content in the chloroplast genomes of these two Corydalis species. Furthermore, our results are consistent with findings in other medicinal plants, such as Mesona chinensis [32], Dryas octopetala var. asiatica [33], and Aroideae species [34]. Sablok [35] reported that the chloroplast genomes of Pooid grasses exhibit a narrow, unimodal GC3 distribution. The low GC3 values not only indicate evolutionary equilibrium in plastid genomes but also suggest a predominance of cytosine (C) deamination relative to other evolutionary forces.

The ENC values of the CDS in the C. yanhusuo chloroplast genome ranged from 42.87 (rps12) to 61.00 (rpl23), with an average of 50.54. Only five genes—rps12, psbA, petD, petB, and ndhC—had ENC values below 45 (Table 2), indicating that codon usage bias in the C. yanhusuo chloroplast genome is generally weak. The ENC is a comprehensive measure of synonymous codon bias, quantifying the extent to which a gene deviates from equal usage of synonymous codons. This index is independent of gene length or amino acid composition. For a single gene, ENC values theoretically range from 20 to 61, with values ≤ 35 indicating strong codon bias; higher ENC values correspond to weaker codon usage bias [36]. In this study, the average ENC value of the selected genes in the C. yanhusuo chloroplast genome was 50.54, higher than that of Dryas octopetala var. asiatica (average ENC 47.20) [33] and various Camellia species (average ENC 48.48–48.51) [24], indicating that overall codon usage bias in the C. yanhusuo chloroplast genome is weak. Among the genes, rps12 had the lowest ENC value, suggesting relatively strong codon bias. The rps12 gene encodes the S12 protein of the 30S small subunit of chloroplast ribosomes. Notably, rps12 has been identified as a highly variable site in the chloroplast genome of the medicinal plant Alpinia japonica [37] and exhibits high codon usage divergence in Malus chloroplasts [38].

Furthermore, the GC3s values of genes in the C. yanhusuo chloroplast genome ranged from 0.217 to 0.371, with an average of 0.292. The CAI values ranged from 0.110 to 0.305, with an average of 0.170 (Table 2), indicating relatively low gene expression levels in the C. yanhusuo chloroplast genome. Robbins [39] reported that gene expression, gene location, and nucleotide composition together account for over 50% of the variation in molecular evolutionary rates of plastid-encoded genes, imposing significant constraints on chloroplast molecular adaptation. Additionally, plastid genes involved in energy production exhibit slower evolutionary rates and stronger constraints than those associated with information processing, as demonstrated in 773 angiosperm plastid genomes [39]. In this study, photosynthesis-related genes—including atp, ccsA, ndh, pet, psa, psb, and rbcL—exhibited a higher average CAI value (0.183) compared with genetic system-related genes, such as matK, rpl, rpo, and rps (average CAI 0.153), indicating relatively higher expression levels and stronger adaptation of photosynthesis-related genes. This finding is consistent with previous studies on Solanum species [40]. Moreover, the psbA gene has been shown to possess a significantly higher codon adaptation level than expected, suggesting that selective constraints have shaped its codon usage during angiosperm evolution [41].

Correlation analysis of base composition parameters is illustrated in Fig. 2. The results indicated that GCall was significantly positively correlated with both GC1 and GC2 (p < 0.05), and GC1 was significantly positively correlated with GC2 (p < 0.05). In contrast, GC3 showed no significant correlation with either GC1 or GC2 (p > 0.05), suggesting notable differences between the third codon position and the first and second positions in the codon composition of the C. yanhusuo chloroplast genome. Additionally, GC3s were significantly positively correlated with GC3 (p < 0.05). The ENC value showed no significant correlation with GC1, a significant negative correlation with GC2 (p < 0.05), and a significant positive correlation with GC3 (p < 0.05). The ENC value, GC3, and GC3s contents were all significantly positively correlated, indicating that codon usage bias in the C. yanhusuo chloroplast genome is mainly influenced by the nucleotide composition at the third codon position.

Figure 2: Correlation analysis of codon parameters in the Corydalis yanhusuo chloroplast genome. Note: * represents a significant correlation (p < 0.05).

The analysis of RSCU values in the C. yanhusuo chloroplast genome is presented in Table 3. Among the 61 codons (excluding the stop codons UAA, UAG, and UGA), approximately 30 codons had RSCU values greater than 1. Of these, 12 codons, including UUA, GUA, and UCA, ended with A; 16 codons, such as UUU, CUU, and AUU, ended with U; and two codons, UUG and UCC, ended with G and C, respectively. In a previous study [12], about 30 codons had RSCU > 1 across the chloroplast genomes of 48 Corydalis species, with the highest RSCU observed for AGA (1.69). Consistently, in C. yanhusuo, the AGA codon, which encodes arginine (Arg), exhibited the highest RSCU value (1.72) among all 61 codons, fully agreeing with the earlier findings. Previous works [42] have shown that the translation efficiencies of three arginine codons—AGA, CGU, and CGA—differ significantly, with nearly a 10-fold variation; AGA exhibits very high translation efficiency, whereas CGA is translated at an extremely low rate. Meanwhile, the two codons AUG and UGG, which encode methionine (Met) and tryptophan (Trp), respectively, have RSCU values of 1, indicating no codon usage bias. In contrast, 29 codons, including UUC, CUC, and CUA, show low occurrence frequency and weak usage bias (RSCU < 1). These results indicate that most high-frequency codons in the C. yanhusuo chloroplast genome end with A or U, accounting for 93.33% of such codons. Similarly, Nie [43] reported that codons ending in A and/or T exhibit higher RSCU values (>1) than their synonymous counterparts in the chloroplast genomes of five Asteraceae species.

3.2 Neutral-Plot Analysis

The results of the neutral plot analysis for the CDS codon bases of the C. yanhusuo chloroplast genome are illustrated in Fig. 3. The x-axis (GC3) ranged from 0.2445 to 0.3953, and the y-axis (GC12) ranged from 0.3326 to 0.5767. Most genes were distributed above the diagonal, while only one gene, cemA, which encodes an envelope membrane protein, was located below the diagonal. The correlation coefficient between GC3 and GC12 was −0.091 (p > 0.05), indicating no significant correlation. The slope of the regression line was −0.1317, suggesting that base mutation contributes approximately 13.17% to codon usage bias. These results indicate that codon usage bias in the C. yanhusuo chloroplast genome is mainly influenced by natural selection. In evolutionary biology, mutation pressure and natural selection are the main forces driving species or gene evolution, and a regression coefficient close to zero implies that natural selection plays the dominant role in shaping codon usage bias [44].

Figure 3: Neutrality plot analysis of the Corydalis yanhusuo chloroplast genome.

3.3 ENC-Plot Analysis

As illustrated in Fig. 4, a small number of genes in the C. yanhusuo chloroplast genome were distributed close to the expected ENC curve, while most genes deviated from the theoretical values and were scattered on both sides of the curve. Notably, more than half of the genes were located below the standard curve, suggesting that their codon usage bias is likely more strongly influenced by natural selection. Overall, the codon usage bias in the C. yanhusuo chloroplast genome appears to be shaped by the combined effects of mutation pressure and natural selection.

Figure 4: ENC-Plot analysis of the Corydalis yanhusuo chloroplast genome.

3.4 PR2-Plot Analysis

The results of the Parity Rule 2 (PR2) analysis are illustrated in Fig. 5. The 50 gene loci were unevenly distributed across the four quadrants: 26 genes were located in the region where G3/(G3 + C3) > 0.5 (right side of the plot), and 32 genes had A3/(A3 + T3) values below 0.5 (lower portion of the plot). These patterns indicate that in the CDS codons of the C. yanhusuo chloroplast genome, guanine (G) is used more frequently than cytosine (C) at the third codon position, and thymine (T) is used more frequently than adenine (A). Moreover, codon usage bias appears to be influenced by both mutation and natural selection, with natural selection playing the dominant role.

Figure 5: PR2-Plot analysis of the chloroplast genome of Corydalis yanhusuo.

3.5 Analysis of Optimal Codons

Identifying optimal codons in the chloroplast genome can enhance translation efficiency and gene expression accuracy. In this study, the 50 CDS were ranked in ascending order of their ENC values. The genes rps12, psbA, petD, petB, and ndhC, with the lowest ENC values, were classified as the high-expression library, while rpoA, ndhE, cemA, ndhJ, and rpl23, with the highest ENC values, were assigned to the low-expression group (Table 2). The RSCU values of both groups and the corresponding ΔRSCU value were calculated, and optimal codons were identified by selecting those with ΔRSCU > 0.08 and RSCU > 1 [33]. As shown in Table 3, a total of 30 high-frequency codons were identified based on RSCU values greater than 1; of these, 28 codons ended with U or A, while only two codons, UUG and UCC, encoding leucine and serine, respectively, ended with G and C. This result is consistent with findings from the Caragana chloroplast genome [45]. In addition, 23 codons with ΔRSCU ≥ 0.08 were identified as high-expression codons (Table 4). By selecting codons that were both high-expression and high-frequency, 16 optimal codons were ultimately determined: UUA, AUU, GUU, GUA, UCU, AGU, CCU, ACU, GCU, CAA, AAA, GAU, UGU, CGU, CGA, and GGU. All of these optimal codons end with A or U, confirming that codons in the C. yanhusuo chloroplast genome preferentially terminate with A or U. This finding is consistent with the RSCU results described above and with observations in Codonopsis chloroplast genomes [46].

Previous studies have proposed that codon usage bias is a key factor for enhancing recombinant protein expression in heterologous expression systems. Natural selection shapes codon usage by favoring translation-optimal codons, thereby improving the translational efficiency of specific genes and creating fitness differences among synonymous codons [47]. Given that each plant cell contains numerous copies of the chloroplast genome, transgenes inserted into chloroplasts can be expressed at extremely high levels, reaching up to 70% of total leaf protein. Generally, stronger codon usage bias correlates with higher gene expression levels. By selecting optimal codons for constructing gene expression vectors, the efficiency of gene expression can be significantly enhanced [48]. Therefore, identifying optimal codons in the C. yanhusuo chloroplast genome may provide a foundation for optimizing gene expression and facilitating molecular breeding.

4 Conclusion

In this study, the chloroplast genome of the medicinal plant C. yanhusuo was resequenced, assembled, and analyzed, and codon usage bias was comprehensively investigated based on 50 protein-coding sequences. Analysis of GCall and GC3 contents, together with the observed pattern GC3 < GC2 < GC1, indicated that codons in C. yanhusuo chloroplast genome preferentially end with A or U. Further analyses, including calculations of ENC and RSCU values, as well as neutral-plot and ENC-plot assessments, confirmed that codon usage bias in this plant is relatively weak, with natural selection exerting a greater influence on codon usage patterns than mutation. In total, 30 high-frequency codons, 23 high-expression codons, and 16 optimal codons were identified in the chloroplast genome, with the third base of the optimal codons showing a strong preference for A or U. Overall, these results provide insight into the genomic composition and codon usage patterns of the C. yanhusuo chloroplast genome. This study establishes a foundational dataset that can support future research on gene expression, chloroplast genetic engineering, and the phylogenetic evolution of C. yanhusuo.