A RING membrane-anchor E3 ubiquitin ligase gene is co-expressed with steroidal glycoalkaloid biosynthesis genes in tomato

Abstract

RING membrane-anchor (RMA) E3 ubiquitin ligases are involved in endoplasmic reticulum (ER)-associated protein degradation, which mediates the regulated destruction of ER-resident enzymes in various organisms. We determined that the transcription factor JASMONATE-RESPONSIVE ETHYLENE RESPONSE FACTOR 4 (JRE4) co-regulates the expression of the RMA-type ligase gene SlRMA1, but not its homolog SlRMA2, with steroidal glycoalkaloid biosynthesis genes in tomato, perhaps to prevent the overaccumulation of these metabolites.

Plants produce a diverse array of specialized metabolites that provide them with adaptive advantages (Weng et al. 2021). Since bioactive metabolites can become toxic to plants at high levels, the carefully controlled production of these compounds is required. In the legume plant barrel clover (Medicago truncatula), the production of triterpenoid saponins is negatively controlled by the regulated destruction of the rate-limiting enzyme 3-hydroxyl-3-methylglutaryl CoA reductase (HMGR) in the mevalonate pathway. This degradation is mediated by the RING membrane-anchor (RMA) E3 ubiquitin ligase MAKIBISHI 1 (MKB1) (Japanese word for caltrop, a spiked weapon), which is involved in endoplasmic reticulum (ER)-associated degradation (ERAD) (Pollier et al. 2013). ERAD is an evolutionarily conserved mechanism for protein quality control that includes protein ubiquitination in the ER (carried out by the ligase complex) and subsequent proteasomal degradation (Christianson and Carvalho 2022). As an example of negative feedback regulation, when saponin biosynthesis is elicited by the defense-related plant hormone jasmonate (JA), M. truncatula MKB1 (MtMKB1) expression is also induced by JA, suppressing HMGR activity to avoid the runaway accumulation of membrane-disrupting saponins (Pollier et al. 2013).

Steroidal glycoalkaloids (SGAs) are cholesterol-derived specialized metabolites containing steroidal aglycones attached to hydrophilic sugar moieties. SGAs function as defense compounds in Solanum species such as tomato (S. lycopersicum) and potato (S. tuberosum) (Cardenas et al. 2015). The SGA and triterpenoid saponin biosynthesis pathways share an early branch that includes the mevalonate pathway, which produces an isoprenoid unit that is then converted to 2,3-oxidosqualene. In tomato and potato, the transcription factor JA-RESPONSIVE ETHYLENE RESPONSE FACTOR 4 (JRE4) coordinately upregulates numerous SGA biosynthesis genes, enabling the JA-elicited accumulation of defense compounds (Cardenas et al. 2016; Nakayasu et al. 2018; Thagun et al. 2016). The tomato RMA E3 ubiquitin ligase gene SlRMA1 (Solyc10g008410), also called MKB1 (SlMKB1), is regulated by JRE4 (Nakayasu et al. 2018; Thagun et al. 2016). This finding suggests that homologous RMA-type E3 ligases are involved in post-translational regulation of the related pathways in distinct lineages of the Solanaceae and legume families. Here we characterize the structural and expressional properties in various regulatory contexts of SlRMA1 and its homologs to get insights into the function of RMA E3 ubiquitin ligases in relation to SGA biosynthesis in tomato.

SlRMA1 and related proteins contain a single HC-type RING finger domain and a C-terminal membrane anchoring domain required for ER retention (Supplementary Figure S1) (Kosarev et al. 2002). The less-conserved regions outside of these two domains are shorter in SlRMA1 than in SlRMA2 or MtMKB1 (Supplementary Figure S1). We constructed a phylogenetic tree of SlRMA1 and related RMA proteins, finding that three other tomato proteins, Solyc07g063200, Solyc10g006960, and Solyc12g013800, are more closely related to MtMKB1 than to SlRMA1 (Figure 1A). Based on this result, we renamed Solyc10g008410 (previously named SlMKB1; Nakayasu et al. 2018) as SlRMA1 and Solyc10g008400 as SlRMA2. SlRMA1 and its closest homolog SlRMA2 are adjacent in the tomato genome (subfamily Solanoideae) (Figure 1B). The genes orthologous to SlRMA1 and SlRMA2 are also located next to each other in the genomes of potato (subfamily Solanoideae) and pepper (Capsicum annuum, subfamily Solanoideae) at syntenic positions. Whereas two RMAs are present in these Solanum species, a cluster of three RMA genes (including two SlRMA2-like genes) is present in pepper (Figure 1). Notably, orthologs of SlRMA2 but not SlRMA1 are present in the genomes of tobacco (Nicotiana tabacum, subfamily Nicotianoideae) and petunia (Petunia axillaris, subfamily Petunioideae) (Figure 1A). No orthologs of both SlRMA1 and SlRMA2 were found in Arabidopsis thaliana and M. truncatula. One possible scenario is that the ancestral copy of SlRMA1 arose through the duplication of SlRMA2, which occurred in ancestors common to species of the subfamily Solanoideae after their separation from the other subfamilies. It is intriguing to address when and how functional differentiation among the duplicates and possible recruitment of RMA1 to regulation of Solanum-specific SGA biosynthesis happened after the gene duplication.

Figure 1. Phylogenetic relationships and gene organizations of RING membrane-anchor (RMA) E3 ubiquitin ligases. (A) Phylogenetic tree of RMA proteins. Close homologs of SlRMA1 (Solyc10g008410) and SlRMA2 (Solyc10g008400) were retrieved from the genomes of potato (St; Solanum tuberosum, v6.1), pepper (Ca; Capsicum annuum, cv. CM334 release 1.55), tobacco (Nt; Nicotiana tabacum, Nitab v4.5 Edwards 2017), and Petunia axillaris (Pa, v1.6.2). HsRMA1 (NM_006913.4) from human (Hs; Homo sapiens) was included as the outgroup. Tomato proteins (ITAG release 4.0) are shown in red. Branch points that encompass proteins closely related to SlRMA1 or SlRMA2 are marked by circles. The amino acid sequences of full-length proteins were aligned using MUSCLE (Edgar 2004). An unrooted phylogenetic tree was constructed using the neighbor-joining algorithm with MEGAX (Kumar et al. 2018) in a default setting. The percentage support from 1,050 bootstraps is indicated at the branch nodes. The scale bar indicates the rate of amino acid substitutions per site. (B) Gene organization of SlRMA1, SlRMA2, and their orthologs from potato and pepper. The orders and orientations of the genes are depicted schematically. Genes homologous to each other are connected by broken lines. Non-RMA genes that are highly conserved among species are indicated by black arrows. The numbers of unrelated genes present between the indicated genes are shown in brackets.

The loss-of-function tomato mutant jre4-1 shows a drastic decrease in SGA accumulation and the expression of SGA biosynthesis genes (Nakayasu et al. 2018). In parallel with SGA metabolism genes, SlRMA1 is downregulated in the leaves and roots of this mutant (Nakayasu et al. 2018). To address whether homologs of SlRMA1 are regulated by JRE4, we examined the transcript levels of SlRMA2 and other genes by reverse-transcription quantitative PCR (RT-qPCR) using the leaves of wild-type (cv. Micro-Tom) and jre4-1 plants. Following total RNA isolation and cDNA synthesis, qPCR was carried out with the primers shown in Supplementary Table S1 as described (Shoji et al. 2022). Tomato EF1α was used as a reference gene. Although they were included in the phylogenetic tree (Figure 1A), the tomato genes Solyc02g078320, Solyc10g006960, and Solyc12g010670 were not analyzed since their expression levels are very low based on reported data (Tomato Genome Consortium 2012). While SlRMA1 was clearly downregulated in jre4-1, as reported previously (Nakayasu et al. 2018), we determined that all of the examined homologs, including SlRMA2, Solyc07g054080, Solyc07g063200, and Solyc12g013800, are expressed at similar levels in the wild type and jre4-1 (Figure 2A), indicating that JRE4 regulates SlRMA1 but not its homologs. In line with the differential regulation of SlRMA1 and SlRMA2 by JRE4, when proximal promoter regions (−1,000 to −1; counted from the first ATG) were computationally searched as described (Thagun et al. 2016), a putative JRE4-binding P box element (5′-TAGCACACACCA-3′ at −744 to −735) was found in SlRMA1, while SlRMA2 has no such binding sites.

Figure 2. Expression patterns of SlRMA1 and SlRMA2 in tomato. Relative transcript levels were measured by RT-qPCR. The primers used are listed in Supplementary Table S1. Average values and standard deviations of three biological replicates are shown. (A) Expression in the wild type and jre4-1. The leaves from 5-week-old plants were used for analysis. Expression levels in wild-type tissue were set to 1. Significant differences relative to the controls were assessed by Student’s t-test. * p<0.05; ** p<0.01. (B) Gene expression in leaves treated with methyl jasmonate (MeJA), 1-aminocyclopropane-1-carboxylic acid (ACC), or both. Detached leaves from 5-week-old tomato plants were submerged for 24 h in solutions containing either or both chemicals, provided at 100 µM. Expression levels in the mock-treated controls were set to 1. Different lowercase letters indicate significant differences between values at p<0.05, as determined by one-way analysis of variance (ANOVA) followed by a Tukey–Kramer test. (C) Expression in various tissues of wild-type tomato. Open flowers, leaves, roots, and fruits at different ripening stages were examined. While values of flowers, leaves, and roots are shown in white bars, black and light grey bars represent data of green fruits and fruits after breaker stage, respectively. Expression levels in tissues with the highest expression were set to 1. Different lowercase letters indicate significant differences between values at p<0.05, as determined by one-way ANOVA followed by a Tukey–Kramer test.

JRE4 and the SGA biosynthesis genes downstream of JRE4 are induced by JA in tomato, which is attenuated by ethylene treatment (Nakayasu et al. 2018). We therefore investigated the responses of SlRMA1 and SlRMA2 to JA and ethylene treatment in tomato. Accordingly, we detached the youngest fully expanded leaves from 5-week-old wild-type tomato plants and submerged them in Gamborg B5 medium containing 100 µM methyl jasmonate (MeJA) and 100 µM 1-aminocyclopropane-1-carboxylic acid (ACC) (Nakayasu et al. 2018), as ACC is readily converted into ethylene in planta (Adams and Yang 1979). Following treatment for 24 h in the dark, we measured transcript levels by RT-qPCR. As reported (Chen et al. 2022), tomato ERF.F5 gene (Solyc10g009110) was induced by either MeJA and ACC (Supplementary Figure S2), supporting the effective treatments. SlRMA1 and the SGA biosynthesis gene SlHMGR1 (Solyc02g082260) were induced 4.1- and 5.5-fold, respectively, in response to MeJA compared to the mock-treated controls, whereas MeJA had no significant effect on SlRMA2 expression (Figure 2B). The addition of ACC almost completely suppressed the MeJA-dependent induction of SlRMA1 and SlHMGR1 (Figure 2B). By contrast, except for 1.9-fold increase of SlRMA2 expression, ACC alone had no significant effect on the expression of these genes (Figure 2B).

The expression of SGA biosynthesis genes and JRE4 drastically decreases in tomato fruits during ripening (Cardenas et al. 2016; Thagun et al. 2016). We therefore examined the expression of SlRMA1, SlRMA2, and the SGA biosynthesis genes SlHMGR1 and GLYCOALKALOID METABOLISM 1 (SlGAME1, Solyc07g043490) (Itkin et al. 2011) in open flowers, leaves, roots, and fruits at different stages of development by RT-qPCR. We observed no organ-specific expression patterns, although expression levels did vary among organs (Figure 2C). SlRMA1, SlHMGR1, and SlGAME1 expression decreased during fruit maturation, with SlRMA1 and SlGAME1 expression mainly declining during the green fruit stages and SlHMGR1 expression clearly declining at the breaker stage (Figure 2C). By contrast, SlRMA2 expression fluctuated during fruit maturation with a peak around the mature green fruit stage, but the differences in expression levels among the stages were relatively moderate (less than 4-fold) compared to the other genes (Figure 2C).

In summary, we studied the phylogenetic relationships (Figure 1A), gene organizations (Figure 1B), and expression patterns (Figure 2) of SlRMA1 and SlRMA2 in tomato. We determined that JRE4 co-regulates the expression of SlRMA1 (but not that of SlRMA2) and SGA biosynthesis genes. The differential expression of SlRMA1 and SlRMA2 in various contexts points to their functional differentiation. In a legume plant, silencing of MtMKB1 led to the overaccumulation of saponin, causing severe morphological defects in hairy roots (Pollier et al. 2013). It would be interesting to examine the effects of altered SlRMA1 and SlRMA2 function on SGA metabolism and development in tomato to clarify the role of RMA-mediated ERAD in SGA regulation. To maintain sterol homeostasis, ERAD-mediated proteolytic regulation of HMGR, squalene monooxygenase, and other ER-resident enzymes is carried out by non-RMA E3 ligases in yeast (Saccharomyces cerevisiae) and mammalian cells (Christianson and Carvalho 2022). Sterols and their intermediates are sensed and stimulate ERAD activity mainly by leveraging E3 ligase activity (Christianson and Carvalho 2022). It remains to be determined whether ER-resident enzymes other than HMGR are subjected to ERAD by RMA ligases and whether any metabolites are involved in the ERAD-mediated regulation of defense metabolism in plants.

Abbreviations

ACC

1-aminocyclopropane-1-carboxylic acid

ANOVA

one-way analysis of variance

ER

endoplasmic reticulum

ERAD

ER-associated degradation

GAME1

GLYCOALKALOID METABOLISM 1

HMGR

3-hydroxyl-3-methylglutaryl CoA reductase

JA

jasmonate

JRE4

JA-RESPONSIVE ETHYLENE RESPONSE FACTOR 4

MeJA

methyl JA

MKB1

MAKIBISHI 1

RMA

RING membrane-anchor

RT-qPCR

reverse-transcription quantitative PCR

SGA

steroidal glycoalkaloid

Conflict of interest

The authors have no conflicts of interest to declare.

Author contribution

T.S. conceived the research plans and performed all experiments and data analyses. K.S. supervised the study. T.S. prepared the draft manuscript. T.S. and K.S. reviewed the contents and approved the final version of manuscript.

Data availability

All data are available upon request.