Introduction

Chinese yam (Dioscorea opposita Thunb.) is a monocot belonging to the Dioscoreaceae family1. It has been an underutilized crop that is being neglected globally2. The yam tubers are highly nutritious, containing secondary metabolites that offer health benefits such as regulating blood sugar levels and controlling cholesterol, fat absorption and hypertension3,4,5. Rich in resistant starch, polysaccharides, and steroidal saponins, Chinese yam has become a popular vegetable in Southeast Asia, contributing to an improved and diversified the local diet6. In China, yam has been used as both food and traditional medicine for thousands of years, underscoring its economic value and potential for the development of health-related products. In recent years, the cultivation area of Chinese yam has expanded significantly, with the planting area of Huaishan yam in China exceeding 200,000 hectares annually. The planting area of Huaishan yam in China has reached more than 200,000 hm2. However, challenges such as soil microbial imbalance, soil nutrient depletion, high soil temperatures, and the accumulation of pathogenic bacteria hinder yam cultivation, becoming major factors restricting its production and industrialization. With the expansion of northern yam species to the West, the planting area of these varieties has increased rapidly in regions such as Shaanxi, Gansu and Ningxia, Xinjiang and western Inner Mongolia7. The western region of Inner Mongolia offers rich land resources, sandy soil, abundant sunlight, and significant temperature differences between day and night, making it an ideal ___location for yam cultivation. However, factors such as the short frost-free period, drought, and saline-alkali soil conditions negatively impact yam production in this area. As a result, breeding and screening yam varieties suited for these challenging conditions have become urgent priorities. Additionally, studying the genetic resistance of yam to abiotic stresses and analyzing their gene expression are essential steps in developing varieties suitable for local production.

With the advancement of molecular biology, the research on plant functional gene expression has become a prominent focus.. Real-time quantitative PCR (RT-qPCR) is a highly effective method for quantitatively analyzing target gene expression. It has emerged as a reliable technique for detecting and measuring gene expression levels in molecular biology research. While genome and transcriptome sequencing can also detect gene expression, RT-qPCR is widely favored for its high sensitivity, flexibility, and efficiency8,9,10.

The quantitative process is influenced by factors such as the quality of RNA and cDNA, primer specificity, and PCR amplification efficiency11. Consequently, the quantitative analysis of target genes requires stable internal reference genes for normalization12,13,14. Internal reference genes are expected to exhibit stable expression across various conditions. However, their expression levels can vary significantly depending on environmental factors, plant tissues, and physiological states15,16. Therefore, selecting stable reference genes specific to each species and treatment method is essential to improve the accuracy of gene expression calibration and standardization17. Genes that maintain basic cellular functions and encode essential proteins are frequently used as reference genes, as their products are necessary for cellular activity and are typically expressed under a wide range of conditions. These genes include the 18S ribosomal RNA (18S) gene18, ubiquitin (UBQ)19, actin(ACT)20, and glycero-aldehyde-3-phosphate dehydrogenase (GAPDH) gene21. GADPH is not an optimal reference gene for ABA treatment in Isodon rubescens. However, it is suitable for use in the roots of Avena sativa L. under drought stress and the leaves under salt stress22,23. For Iris lactea var. Chinensis, the most suitable internal reference genes under cadmium stress are EIF-5A and UBC. Previous studies have identified TIP41, CYP, PGK, GAP, and PP2Acs as the most suitable internal reference genes under abiotic stresses, such as salt, drought, and cold24. Among different varieties and tissues of Zanthoxylum bungeanum, UBQ and UBA expression are the most stable during different development stages of fruit25. Screening for stable reference genes using q-PCR analysis has been conducted in many plants, including sweet potato26, Lilium henryi Baker27, Broussonetia papyrifera28, and Allium sativum L. (garlic) 29. The evaluation of reference genes is based on Ct values from qRT-PCR, with widely used software including using the delta-Ct method30, geNorm31, NormFinder32, BestKeeper33, and RefFinder34. Despite the maturity of reference gene screening methods, there are limited studies on reference genes specific to yam and its tissues under various treatments. This lack of research significantly restricts the effective application of reference genes in yam studies.

This study aims to identify reliable internal reference genes for analyzing target gene expression in yam. Nine candidate reference genes were selected based on yam transcriptome data for gene standardization. To assess the suitability of the selected candidate genes and ensure the accuracy of the results, we used both the most stable and the most unstable internal reference genes identified in this study. We then measured the relative expression levels of WRKY genes in yam across different developmental stages, tissues, and stress conditions using RT-qPCR. This research provides a theoretical foundation for the subsequent screening of resilient yam varieties and further gene function studies.

Materials and methods

Experimental material

Dioscorea opposita variety ‘Dahechangyu’ was used as the experimental material in this study. The yam plants were grown at College of Horticulture and Plant Protection, Inner Mongolia Agricultural University, Hohhot, China, was used as experimental material in the present study. The experimental fields were managed according to standard agricultural practices. Tubers were sampled at intervals of 90, 105, 120, 135, 150 and 165 days after planting. In October 2023, the roots, stems, and leaves of the yam were collected at the peak of tuber expansion. Yam tubers that had been released from dormancy and were of similar size were selected and planted into plastic pots (top diameter 10 cm, bottom diameter 8 cm), which were filled with a substrate composed of peat and perlite (1:1 ratio). The plants were cultured in a chamber at 25 °C with a 16/8 h day/night photoperiod and 65–75% relative humidity. When the yam seedlings reached approximately 15 cm in height(about one month after planting), the following treatments were applied: For high temperature stress, seedlings were grown in the chamber at 36 °C; for cold stress, seedlings were grown in the chamber at 4 °C; for salt stress, seedlings were grown in water containing 200 mmol/L NaCl; for drought stress, seedlings were grown in water containing 20% PEG. Yam leaves were collected at 0, 12, 24, and 48 h after each treatment. For abscisic acid stress (ABA) and methyl jasmonate treatment (MeJA) stress, seedlings were grown under normal growth conditions and sprayed with ABA/MeJA (100 mmol/L). The leaves of yam were collected at 0, 1, 6, 12, and 24 h after treatment. All samples were immediately frozen in liquid nitrogen and stored at − 80 °C. All treatments were conducted in triplicate.

Total RNA isolation and cDNASynthesis

Total RNA was extracted using an RNA extraction kit (9769; TaKaRa, Beijing). The quality of the RNA samples was assessed using a NanoDrop 2000c (Thermo Fisher Scientific, USA) and gel electrophoresis. First-strand cDNA synthesis was performed using cDNA Synthesis SuperMix (RR047A; TaKaRa, Beijing) and stored at − 20 °C.

Selection of yam sequences, primer design and PCR optimization

The candidate reference genes were selected based on our transcriptome data35 and previous reports of reference genes used in studies of various plant species36,37,38. Primers were designed with Primer Premier 5.0, with melting temperatures between 51–60 °C, sizes of 17–20 bp, and approximately 50% GC content (Table 1). Before qRT-PCR, the primer pairs were tested using a standard PCR reaction on a Mastercycler Gradient (Eppendorf) to determine the optimal conditions. Amplicons of the expected size were verified by 1% agarose gel electrophoresis and ethidium bromide staining. The amplification efficiency (E) of the candidate genes was calculated using a standard curve (a fivefold dilution series of cDNA was used as the template) by q-PCR. The formula for efficiency is: E (%) = (10−1/slope–1) × 100%39.

Table 1 Primer sequences and PCR amplification characteristics information on the candidate reference genes of yam.
Full size table

qPCR analysis of reference genes

q-PCR was performed using a CFX Connect™ real-time system (Bio-Rad, Singapore) followding the instructions provided by TaKaRa Biotech (9769; TaKaRa, Beijing, China) for the Super Real Fluorescence Quantitative Premix Reagent-Enhanced Kit. The reaction mixture cinsisted of 10 µL of 2 × SYBR Green RT-qPCR Master Mix, 8.2 µL of ddH2O, 1 µL of cDNA, and 0.4 µL each of forward and reverse amplification primers (10 µmol/L), for a final volume of 20 µL. The PCR conditions included denaturation at 95 °C for 30 s, followed by 40 cycles of 5 s at 95 °C and 30 s at 60 °C. The melting curve was set to the instrument’s default. All samples were tested with three technical replicates and three independent biological replicates.

Stability analysis and validation of selected rreference gGenes

Delta-Ct, BestKeeper, geNorm, NormFinder and RefFinder software programs were used to analyze the expression stability of the candidate genes of in yam across different developmental stages, different tissues and different abiotic stresses. The geNorm and NormFinder analyses employed the 2−∆CT method for calculations, while BestKeeper analyzed the Ct values. RefFinder was employed for a comprehensive evaluation, determining the most stable reference gene(s). The geometric mean of all rankings was used to calculate the overall expression stability ranking, aiding in the selection of the most suitable reference genes for yam. Based on the selected reference genes, the expression patterns of WRKY during yam development and under abiotic stress were analyzed to further validate stability of the reference genes. The samples and procedure are the same as qPCR analysis of reference genes, and relative expression level was calculated using the 2−∆∆CT method.

Results

Primer specificity and amplification efficiency of candidate reference genes

Agarose gel electrophoresis results showed that the fragments for the nine candidate reference genes (F-box, ACT, GAPDH, PP2A, 18S, GUSB, UBQ, CHCI and EIF) (Fig. 1A) and two validation genes (Fig. 1B) were between 100 and 200 bp, which was consistent with the expected fragment sizes. Each product was well amplified with no primer dimers, indicating that the selected reference gene primers were suitable for use in subsequent experiments. Additionally, qRT-PCR was performed using cDNA templates from different samples, and the melting curve was used for evaluation. The melting curves of the nine candidate reference genes all showed single peaks, indicating specific primers and reproducible amplification curves between samples (Fig. 2).

Fig. 1
figure 1

Specificity analysis of nine candidate reference gene primers of Yam. M: 2000 marker; (A) PCR of nine gene primers for internal reference gene screening; (B) Two genes for internal reference gene validation.

Full size image
Fig. 2
figure 2

Melting curves of qPCR products of nine candidate reference genes of yam.

Full size image

The efficiency of primer amplification (E) was calculated based on the slope (K) and correlation coefficient (R2) derived from the standard curve(Supplementary materials). The amplification of the candidate genes ranged from 94.98% to 113.79% with all R2 values exceeding 0.98 (Table 1). These results indicate strong primer amplification efficiency, meeting the stringent criteria for qRT-PCR.

Analysis of candidate gene expression abundance

To assess the expression consistency of the nine potential reference genes across various tissues and different treatment conditions, the transcriptional abundances were represented as Ct values. The expression levels of candidate reference genes were determined by measuring the cycle threshold (Ct) values using RT- qPCR. The Ct value is inversely proportional to gene expression abundance, meaning a lower Ct value indicates higher gene expression. The Ct values of the nine candidate reference genes ranged from 6.45 to 34.53 across different tuber expansion periods, tissues, and abiotic stress conditions. The smallest variation in gene expression was observed for CHCI (2.79) and UBQ (3.34), while EIF (11.41) was the gene with the most variable levels of expression. During different tuber expansion periods of yam, the Ct values of the candidate genes ranged from 6.45 to 32.38. In different tissues, 18S had the lowest mean Ct value, while EIF had the highest. For different abiotic stresses, the coefficient of variation for 18S was relatively low, whereas EIF exhibited the highest coefficient of variation (Table 2). These finding suggest that none of the selected genes exhibited a constant level of expression across the different D.opposita samples tested. The ranking of the expression levels was as follow: 18S > GAPDH > ACT > CHCI > UBQ > GUSB > PP2A > F-box > EIF (Fig. 3). Overall, the results indicated that the gene expression levels showed significant divergence across all samples.

Table 2 Internal reference gene Ct value.
Full size table
Fig. 3
figure 3

Ct value distribution of nine candidate reference genes in different developmental stages, different tissues and abiotic stress of yam. The boxes indicate the 25th and 75th percentiles. The line across the box and each box indicate the median and mean Ct values, respectively.

Full size image

Estimation of stability by delta‑Ct

Since the nine candidate reference genes exhibited wide variations in expression levels across the different samples, statistical methods were employed to rank the stability of gene expression and determine the number of reference genes required for accurate gene expression profiling under specific experimental conditions. The stability of each reference gene was evaluated by ΔCt, with the gene exhibiting the lowest mean standard deviation (mSD) considered the most stable. The results of the ΔCt analysis showed that UBQ was the most stable gene across different developmental stages of tuber expansion as well as among the various tissues. GUSB demonstrated the highest stability in High-temperature treated samples, while ACT was the most stable in Low-temperature treated samples. GAPDH proved to be the most stable in NaCl-treated samples, and 18S showed the highest stability in PEG-treated samples (Fig. 4). EIF was identified as the least stable reference gene across all samples, including different tissues, High-temperature treated, Low-temperature treated, and NaCl-treated conditions.

Fig. 4
figure 4

Stability evaluation of nine reference genes analyzed using ∆Ct. (A) All samples; (B) Different stages of tuber developmental; (C) Different tissues; (D) High temperature-treated; (E):Low temperature-treated; (F) NaCl-treated; (G) PEG-treated; (H): ABA-treated; (I) MeJA-treated.

Full size image

geNorm analysis

The geNorm analysis evaluates the expression stability of candidate genes by calculating the M value (Expression Stability) for nine candidate reference genes under various developmental stages of tuber growth, different tissues, and abiotic stress conditions. Developed by Vandesompele et al. in 2002, geNorm is widely used to assist in screening reference genes31. A smaller M value indicates greater stability in gene expression. Genes with an M value below 1.5 are considered suitable for use as a reference genes. Figure 5 shows the changes in M values under different conditions and tissues. The expression stability values of all reference samples were assessed, and all nine candidate reference genes had M values below 1.5. In the different developmental stages, the lowest M values for UBQ and ACT indicated the highest stability, while the highest M of PP2A indicated the lowest stability. In different tissues, GAPDH and F-box exhibited the most stable expression, while GUSB showed the greatest variability. For High-temperature treatment, GAPDH and 18S were the most stable reference genes. In Low-temperature treated samples, the combination of GAPDH and UBQ was the most suitable, while PP2A was the least stable. F-box was identified as the most stable endogenous reference gene in ABA treatment and tissues. On the other hand, EIF ranked the lowest in expression stability across high-temperature, NaCl, PEG, and ABA treatments, making it the most unstable reference gene. Thus, EIF is not suitable as an internal reference gene for gene expression analysis.

Fig. 5
figure 5

Mean expression stability values (M) of candidate reference genes. Note: The most stable genes are listed on the right and the most unstable genes are listed on the left.

Full size image

Additionally, the paired variant V value of the standardized factor can be calculated by the software after introducing a new reference gene. geNorm determines the optimal number of candidate reference genes by calculating the ratio Vn/n + 1 = 0.15 to achieve more accurate and reliable results. If Vn/Vn + 1 < 0.15, then the ideal number of reference genes is n. Conversely, if Vn/Vn + 1 > 0.15, then the number should be n + 1. As shown in Fig. 6, the Vn/Vn + 1 values for different treatments were all less than 0.15, indicating that two reference genes are sufficient for accurate normalization of target gene expression.

Fig. 6
figure 6

Pairwise variation (V) analysis of candidate reference genes.

Full size image

NormFinder analysis

NormFinder evaluates the stability of candidate genes by calculating an expression stability value (SV), where a lower SV indicates greater stability (Table 3). In the different developmental stages of the tuber, ACT and UBQ exhibited the most stable expression, while PP2A was the most unstable. CHCI showed the most stable expression in different tissues, while GUSB exhibited the greatest instability. Under High-temperature treatment, GAPDH and 18S were the most stable gene, while EIF was the most unstable. In Low temperature-treated samples, F-box was the most stable, whereas EIF and PP2A were the least stable. In NaCl-treated samples, GUSB and F-box were the most stable, while EIF was the least stable. In PEG-treated samples, UBQ was stable, while EIF again showed the least stability. The NormFinder analysis of all samples revealed that EIF was the least stable across conditions. These results indicated that NormFinder and geNorm produced almost identical stability rankings for the candidate genes, although some minor differences were noted.

Table 3 Stable values of nine candidate reference genes in different treatments and tissues by NormFinder.
Full size table

BestKeeper analysis

In the BestKeeper analysis, candidate reference genes were ranked based on the standard deviations (SD) of the original Ct values under various coefficient of variation (CV) settings. Lower SD and CV values indicate better stability of reference genes. As shown in Fig. 7, UBQ exhibited the highest stability across different developmental stages, tissues, and under High-temperature conditions. ACT showed the highest stability under cold stress, while 18S was relatively stable under PEG stress. Under all treatments, the coefficient of variation for EIF was greater than 1, indicating a large degree of variation. These results demonstrate that the stability of gene transcripts varies considerably under different experimental conditions, highlighting the need for careful selection of reference genes.

Fig. 7
figure 7

Stability analysis of nine reference genes based on BestKeeper.

Full size image

ReFinder analysis

To address discrepancies in the results from the four analytical methods, the RefFinder online tool was used for comprehensive ranking. RefFinder integrates the outputs of multiple algorithms to identify the optimal reference genes under different experimental conditions. As shown in Table 4, RefFinder identified UBQ as the most stable reference gene across all treatments. The stability ranking for the nine candidate reference genes was as follows: UBQ > CHCI > F-box > ACT > GAPDH > 18S > PP2A > GUSB > EIF. For different developmental stages and tissues, PP2A emerged as the optimal reference gene, with UBQ consistently ranking in the top three. Under various abiotic stress treatments, UBQ was the most stable gene. In contrast, EIF was consistently identified as the least stable gene under all conditions.

Table 4 Comprehensive stability analysis of reference genes based on RefFinder in different experiment conditions.
Full size table

Stability verification of candidate reference genes

To validate the reliability of the selected reference genes, the expression of WRKY was analyzed using the identified stable and unstable reference genes. WRKY is a key plant-specific transcription factor families that regulates various aspects of plant growth, development, physiological processes, and responses to abiotic stresses. The stable reference genes UBQ and PP2A, as well as the unstable reference gene EIF, were used to measure the expression levels of WRKY46a and WRKY51 in yam under different developmental stages, across various tissues, and during abiotic stresses (Fig. 8). Results showed consistent expression levels of WRKY when stable reference genes (UBQ and PP2A) were used. In contrast, the expression levels varied significantly when the unstable reference gene (EIF) was used, underscoring its unreliability. These findings confirm the stability and accuracy of the selected reference genes, validating the robustness of this study’s experimental results.

Fig. 8
figure 8

Stability verification of candidate reference genes with WRKYs.

Full size image

Discussion

Identifying stable reference genes across different stages of plant development, tissues, and abiotic stress conditions is a critical prerequisite for studying gene functions and regulatory mechanisms in plant growth and development. With advancements in transcriptome data analysis, it is now possible to screen candidate reference genes systematically40. q-PCR is regarded as an effective tool for molecular biology research41. It is instrumental in evaluating gene expression levels and elucidating plant responses to various stressors. Reference genes serve as the cornerstone for accurate normalization in qPCR, enabling researchers to understand biosynthetic pathways and regulatory mechanisms. Over the years, many suitable reference genes have been identified in both animal and plant species42. However, there are currently no documented reference genes for Dioscorea opposita (commonly known as Chinese yam). Among its varieties, Dahechangyu yam is a widely cultivated type in both northern and southern China. Identifying and validating stable reference genes for yam will provide a robust foundation for further studies on gene expression regulation in this species.

It is reported that the optimal reference genes can vary depending on the species and specific conditions43. Among the nine reference genes analyzed, EIF exhibited the highest mean Ct value, while 18S had the lowest. Since lower Ct values indicate higher gene expression levels, 18S was expressed at a significantly higher level compared to the other genes. The smallest variation in gene expression was observed for CHCI (3.10) and UBQ (3.34), whereas EIF (11.41) displayed the highest variability across all samples. In order to explore the changes in gene expression levels in yam, geNorm, NormFinder and BestKeeper were used to evaluate the stability of reference gene transcripts28. The results demonstrated that there were some differences in the stability of the reference genes across the different software. The reason for this may be due to the differences in algorithms between each software44 and a similar trend was seen in reports from other species, such as in Rubus45 and Gleditsia microphylla46. In this study, geNorm ranked UBQ as the most stable reference gene across different developmental stages, as well as under Low-temperature, salt, and MeJA treatments. Conversely, EIF consistently ranked as the least stable gene in seven of the eight treatments, making it the most unreliable choice as an internal reference gene for gene expression analysis.

To enhance the detection accuracy of target genes and address the variability among different software tools, Vandesompele31 recommends using the V value (Pairwise variation value) generated from geNorm as the threshold. This approach minimizes errors by incorporating multiple reference genes. The geNorm program typically sets the threshold at Vn/n + 1 = 0.15. While 0.15 is considered an ideal value, it is influenced by factors such as the number of genes analyzed and the type of test samples. Consequently, significant differences in the accuracy of screening results have been observed47. According to both NormFinder and BestKeeper, EIF was identified as the most unstable reference genes, though the rankings for the most stable genes differed between the two programs.

To minimize errors caused by differences in software analysis and human selection, integrated analysis was used to evaluate the results from different software tools. This approach provided a means to better reflect the stability of each reference gene and reduce discrepancies in software evaluations48. UBQ, F-box, and ACT were most frequently ranked as the best genes for normalization by the four programs, with UBQ occurring 17 times, F-box 8 times, and ACT 7 times. In contrast, EIF was identified as the least stable, being ranked as the most unstable 22 times. Under most experimental conditions, UBQ displayed high expression stability, consistent with findings from an earlier study on the ephemeral plant Arabidopsis pumila49. The UBQ gene belongs to the ubiquitin gene family, which plays a role in several essential biological processes, including cell cycle regulation, growth and apoptosis, signal transduction, and immune responses. Due to its high sequence homology and conserved nature, UBQ has been used as a reference gene in rice50 and black fungus51. In contrast, EIF was the most unstable gene and unsuitable as an internal reference for gene expression determination, a finding consistent with similar studies in eggplant52. The stability evaluation in this study revealed that no single reference gene was suitable for all experimental conditions. The most stable reference genes varied across groups and even between organs within the same group. Therefore, we recommend selecting the most appropriate reference genes for specific conditions rather than relying on a universal reference gene for normalization, despite the additional effort and time required.

WRKY transcription factors (TFs) play a critical regulatory role in plant stress signaling by controlling the expression of various stress-related genes, thereby enhancing plant stress resistance53. At the same time, it also plays an important regulatory role in plant growth and development54. To verify the accuracy and reliability of the selected reference genes, a normalization procedure was performed on the relative expression levels of the target genes WRKY46a and WRKY51 under different experimental conditions. As shown in Fig. 8, the expression patterns of the target genes differed significantly depending on whether stable or unstable reference genes were used for normalization. When stable reference genes were used either individually or in combination there was minimal variation in expression levels. In contrast, the use of unstable reference genes resulted in pronounced changes in the expression patterns, highlighting the importance of selecting appropriate reference genes for accurate normalization. Meanwhile, the unstable internal reference genes changed significantly. Although the expression trends observed with individual stable reference genes were consistent with those derived from combinations of stable reference genes, slight differences in expression levels were still evident. In contrast, the use of inappropriate reference genes resulted in entirely inconsistent expression profiles compared to those obtained with stable reference genes, with significant changes observed. This highlights the substantial impact of using unstable reference genes on qRT-PCR results and underscores that the relative expression levels of target genes depend on both the selection and number of reference genes used. Similar findings have been reported in various species, including Codonopsis pilosula55, Scutellaria baicalensis56, and Schima superba57.

Conclusion

In this study, we systematically identified and screened ideal reference genes for qRT-PCR analysis of target gene expression in the Dioscorea genus under diverse experimental conditions. Our findings revealed that two consistently stable genes, UBQ and PP2A, are suitable for accurately normalizing target gene expression in yam across all tested conditions. EIF is unsuitable for accurately normalizing internal reference gene. Furthermore, these genes were successfully employed for gene expression analysis in the study of WRKY transcription factors. These results not only provide a solid theoretical basis for functional genomics research but also establish a foundation for exploring the molecular mechanisms underlying biological functions in the Dioscorea genus.