Identification of stably expressed reference genes for expression studies in Arabidopsis thaliana using mass spectrometry-based label-free quantification

Abstract

Arabidopsis thaliana has been used regularly as a model plant in gene expression studies on transcriptional reprogramming upon pathogen infection, such as that by Pseudomonas syringae pv. tomato DC3000 (Pst DC3000), or when subjected to stress hormone treatments including jasmonic acid (JA), salicylic acid (SA), and abscisic acid (ABA). Reverse transcription-quantitative polymerase chain reaction (RT-qPCR) has been extensively employed to quantitate these gene expression changes. However, the accuracy of the quantitation is largely dependent on the stability of the expressions of reference genes used for normalization. Recently, RNA sequencing (RNA-seq) has been widely used to mine stably expressed genes for use as references in RT-qPCR. However, the amplification step in RNA-seq creates an intrinsic bias against those genes with relatively low expression levels, and therefore does not provide an accurate quantification of all expressed genes. In this study, we employed mass spectrometry-based label-free quantification (LFQ) in proteomic analyses to identify those proteins with abundances unaffected by Pst DC3000 infection. We verified, using RT-qPCR, that the levels of their corresponding mRNAs were also unaffected by Pst DC3000 infection. Compared to commonly used reference genes for expression studies in A. thaliana upon Pst DC3000 infection, the candidate reference genes reported in this study generally have a higher expression stability. In addition, using RT-qPCR, we verified that the mRNAs of the candidate reference genes were stably expressed upon stress hormone treatments including JA, SA, and ABA. Results indicated that the candidate genes identified here had stable expressions upon these stresses and are suitable to be used as reference genes for RT-qPCR. Among the 18 candidate reference genes reported in this study, many of them had greater expression stability than the commonly used reference genes, such as ACT7, in previous studies. Here, besides proposing more appropriate reference genes for Arabidopsis expression studies, we also demonstrated the capacity of mass spectrometry-based LFQ to quantify protein abundance and the possibility to extend protein expression studies to the transcript level.

Keywords: RT-qPCR; abscisic acid; expression study; jasmonic acid; label-free quantification; pathogen infection; reference gene; salicylic acid.

Figure 1

Label-free quantification (LFQ) intensities of the candidate reference genes having stable abundances between 0 and 3 days post-inoculation (dpi) with Pseudomonas syringae pv. tomato DC3000 (Pst DC3000) in 5-week-old A. thaliana ecotype Col-0. The protein abundance levels of these candidate reference genes, calculated from LFQ using Proteome Discoverer v2.4 (Thermo Fisher Scientific, Waltham, MA, United States), were not statistically different between 0 and 3 dpi (p > 0.05). Three individual plants were pooled as one biological replicate, each with two technical repeats of LFQ analysis. Error bar represents the standard deviation of a total of six technical repeats based on three biological replicates.

Figure 2

Ranges of expression levels of the candidate reference genes and reference genes commonly used in previous studies. (A) A boxplot showing the range of expression values [log₂(C_t value)] of the candidate reference genes and the reference genes commonly used in previous studies obtained by RT-qPCR when subjected to Pst DC3000 at 0 dpi and 3 dpi. (B–D) Boxplots showing the range of expression values [log₂(C_t value)] of the candidate reference genes when subjected to jasmonic acid (JA) (B), salicylic acid (SA) (C), and abscisic acid (ABA) (D) treatments and the corresponding mock control treatments. The solid line inside each box represents the median expression value and the lower and upper edges of the boxes denote the 25^th and 75^th percentiles, while the whiskers represent the maximum and the minimum values. Each dot represents the expression value calculated from each technical replicate of RT-qPCR. Three technical repeats were performed for each biological replicate, with two biological replicates in total.

Figure 3

geNorm analysis of the average stability values of the candidate reference genes predicted in this study and reference genes commonly used in previous studies upon Pst DC3000 infection. Average expression stability values (M) of the remaining control genes during the step exclusion of the least stable control among the reference gene candidate upon Pst DC3000 infection using geNorm. The higher the average M, the lower rank of the expression stability of the reference gene candidate.

Figure 4

Pairwise variation (V) of the candidate reference genes predicted in this study and reference genes commonly used in previous studies calculated by geNorm. The pairwise variation representing (V_n/V_{n + 1}). The threshold value of V for accessing the optimal number of the reference genes for RT-qPCR normalization is 0.15. n represents the number of reference gene required for expression normalization in RT-qPCR.

Figure 5

Comprehensive gene stability ranking of the predicted reference gene candidates and the commonly used reference gene for Pst DC3000 inoculation. The stability ranking was generated using the RefFinder. A lower stability ranking refers to a higher stability of the reference gene. Gene names in black: candidate reference genes reported in this study; gene names in red: reference genes commonly used in previous studies. ^#TUB2_1 and TUB2_2 refer to two different primer pairs for TUB2 (Wu et al., 2019; Gao et al., 2022). ^{^}UBC21_1 and UBC21_2 refer to two different primer pairs for UBC21 (Cuéllar Pérez et al., 2014; Gao et al., 2022)^{. *}ACT1_1 and ACT1_2 refer to two different primer pairs for ACT1 (Kim and Hwang, 2014; Zhang et al., 2015).

Figure 6

geNorm analysis of the average expression stability values of the candidate reference genes under Pst DC3000, JA, SA, and ABA treatments. Average expression stability values (M) of the remaining control genes were calculated by a stepwise exclusion of the least stable control gene among the reference gene candidates under each treatment using geNorm. The higher the M value, the lower is the ranking of the reference gene candidate in terms of expression stability.

Figure 7

Pairwise variations (V) of the candidate reference genes. The pairwise variations (V) of the candidate reference genes calculated using geNorm, and V = V_n/V_{n + 1}, where n is the number of reference genes required for expression normalization in RT-qPCR. The threshold value of V for achieving the optimal number of reference genes for RT-qPCR normalization is 0.15. All the candidate genes had V values well below the threshold value under all the treatments, indicating a minimal number of reference genes from among these candidates is required for normalization in RT-qPCR.