Formation of Protein Disulfide Bonds Catalyzed by OsPDIL1;1 is Mediated by MicroRNA5144-3p in Rice

Abstract

Correct folding of proteins in the endoplasmic reticulum is important for their stability and function under stress. The protein disulfide isomerase (PDI) OsPDIL1;1 is a key protein-folding catalyst in rice (Oryza sativa L.). Here, microRNA5144 (osa-miR5144-3p) is reported to mediate the formation of protein disulfide bonds via targeting OsPDIL1;1 mRNA in rice seeds and seedlings during development and under conditions of abiotic stress, respectively. Expression analysis of transgenic rice and identification of cleavage sites showed that OsPDIL1;1 mRNA is a target of osa-miR5144-3p. Expression of osa-miR5144-3p and OsPDIL1;1 was shown to be inversely regulated in developing organs and under abiotic stress. The down-regulation of osa-miR5144-3p or overexpression of OsPDIL1;1 in transgenic rice showed increased total protein–disulfide bond content, compared with the wild type. This indicates that protein–disulfide bond formation is enhanced by down-regulation of osa-miR5144-3p or overexpression of OsPDIL1;1. These transgenic rice plants also displayed strong resistance to salinity and mercury stress, in comparison with the wild type. In contrast, the transgenic rice plants overexpressing osa-miR5144-3p or down-regulating OsPDIL1;1 had a lower protein–disulfide bond content; they were susceptible to abiotic stress and produced abnormal grains with small and loosely packed starch granules. These results indicate that protein–disulfide bond formation catalyzed by OsPDIL1;1 is modulated by osa-miR5144-3p in rice during development and is involved in resistance to abiotic stress.

Introduction

Disulfide bonds are required for the stability and function of a large number of proteins (Rietsch and Beckwith 1998). Although disulfides may be formed in multiple ways (Bulleid and Ellgaard 2011), their correct formation during folding of nascent peptides to yield native proteins in eukaryotic cells, namely oxidative protein folding, takes place mainly in the endoplasmic reticulum (ER) (Bulleid and Ellgaard 2011), where disulfides are formed in a process catalyzed by members of the protein disulfide isomerase (PDI) family (Liu and Howell 2010). During folding, cysteine residues that are in close proximity can form disulfides, even if they are not linked in the final native structure. Such non-native disulfides are prevalent in misfolded proteins, but can also be intermediates in normal folding. For native disulfides to form, non-native disulfides must be broken in a reaction that is also catalyzed by members of the PDI family. Hence, the PDI family plays a crucial role in both the formation and reduction of disulfides for correct folding of proteins entering the ER (Bulleid and Ellgaard 2011). However, under stressful conditions, the protein folding machinery in the ER may reach a limit as the demands for protein folding exceed the capacity of the system. Under these conditions, misfolded or unfolded proteins accumulate in the ER, triggering an unfolded protein response (Liu and Howell 2010). When protein misfolding occurs, unfolded proteins accumulate and aggregate in the ER, triggering a signal that selectively activates gene transcription. The activated genes can increase the folding or degradation of unfolded proteins by inducing the expression of genes related to those functions, such as ER-localized PDIs, in an attempt to maintain homeostasis of the ER (Kaufman 1999, Urade 2007, Liu and Howell 2010).

Typical PDI is the most prominent member of a family of closely related proteins (PDI-like) characterized by one, two or three thioredoxin-like active domains. The importance of PDI in health and disease has been examined in the fields of lipid homeostasis, hemostasis, infectious disease, cancer, neurodegeneration and infertility (Benham 2012). The plant PDI family also plays important roles in plant development and in plant responses to various types of stress (Lu and Christopher 2008b). Loss-of-function barley HvPDIL5-1 alleles at the recessive RESISTANCE TO YELLOW MOSAIC DISEASE 11 (rym11) resistance locus confer broad-spectrum resistance to multiple strains of Bymoviruses (Barley yellow mosaic virus and Barley mild mosaic virus), which seriously threaten winter barley production in Europe and East Asia. Therefore, HvPDIL5-1 alleles play a central role in resilient virus resistance in barley (Yang et al. 2014). PpPDI1 from Phytophthora parasitica is essential for inducing cell death in Nicotiana benthamiana leaves and it might be a virulent factor of P. parasitica contributing to plant infection (Meng et al. 2015). The chloroplast-localized AtPDI6 plays a role in the chloroplast. AtPDI6-knockdown plants displayed higher resistance to photoinhibition than the wild-type plants when exposed to a 10-fold increase in light intensity (Wittenberg et al. 2014). Plant PDIs have also been shown to be involved in the folding and deposition of seed storage proteins (d’Aloisio et al. 2010).

The rice genome contains at least 19 PDI-like sequences (OsPDILs). However, only eight OsPDILs were predicted to have two redox-active cysteine pairs: OsPDIL1;1–1;4, OsPDIL2;1–2;3 and OsPDIL5;1. OsPDIL5;1 is short, with only 147 amino acid residues; therefore, the rice genome was believed to encode seven PDI proteins (Houston et al. 2005, Onda and Kobori 2014). Three rice PDI family members (OsPDIL1;1, OsPDIL1;4 and OsPDIL2;3) have been studied. OsPDIL1;1 showed the highest catalytic activity for oxidative RNase refolding, followed by OsPDIL1;4 and OsPDIL2;3, but only OsPDIL1;1 showed obvious catalytic activity for reduction of the insulin–disulfide bond (Onda and Kobori 2014). Localization analysis showed that OsPDIL1;1 was uniformly dispersed in the luminal space of the ER, but OsPDIL2;3 was strongly detected in rings within the ER (Onda et al. 2011). Both OsPDIL1;1 and OsPDIL2;3 play a role in seed development (Han et al. 2012, Onda et al. 2011). The ospdil1;1 mutant produces small grains with floury endosperms caused by the loose packing of starch granules; therefore, OsPDIL1;1 is considered to play an important role in starch synthesis (Han et al. 2012). Knockout of OsPDIL1;1 also caused formation of aggregates containing proglutelins through non-native intermolecular disulfide bonds (Kim et al. 2012). However, only the OsPDIL2;3 knockdown mutant exhibits aberrant accumulation of prolamins in the type I protein body (PB-I); therefore, OsPDIL1;1 and OsPDIL2;3 are considered not to be functionally redundant in sulfhydryl oxidations of structurally diverse storage proteins, and to play distinct roles in protein body development (Onda et al. 2011).

The microRNA osa-miR5144-3p was isolated from rice developing grains; it has been identified in leaves, roots and in vitro cultured rice embryogenic calli (Chen et al. 2011,, Campo et al. 2013, Cheah et al. 2015). No homolog of osa-miR5144 has been identified from other species, and little is known about its functions at present. Human miR-148a inhibited the proliferation and promoted the paclitaxel-induced apoptosis of ovarian cancer cells, an effect which may be partly attributed to direct targeting of PDIA3 (Zhao et al. 2015). In vivo experiments with mice and worms showed that induction of ER stress correlated with decreased miR-322 abundance and increased PDIA6 mRNA abundance (Groenendyk et al. 2014). However, little is known about how plant PDIs are regulated, except that the expression of four AtPDI genes was reduced in the atbzip60 mutant (Lu and Christopher 2008a).

Here, we report that osa-miR5144-3p regulates OsPDIL1;1 expression and plays a role in protein folding during rice development and in rice resistance to abiotic stress.

Results

osa-miR5144-3p targets OsPDIL1;1 mRNA

Until now, osa-miR5144-3p has only been isolated from rice (Zhu et al. 2008). In order to identify the targets of osa-miR5144-3p, transgenic rice overexpression of mature osa-miR5144-3p (osa-miR5144-3p-ox) was generated in the wild-type japonica cultivar Zhonghua 11 (ZH11) background, using an artificial microRNA technique (http://wmd3.weigelworld.org/cgi-bin/webapp.cgi). Quantitative real-time PCR (qRT-PCR) analysis showed that >10 independently transformed osa-miR5144-3p-ox lines displayed at least a 2-fold up-regulation of osa-miR5144-3p, compared with the expression level in ZH11 (Fig. 1A). OsPDIL1;1 (LOC_Os11g09280) was predicted as one of the 10 putative target genes of osa-miR5144-3p by psRNATarget (http://plantgrn.noble.org/psRNATarget/) (Supplementary Table S1). qRT-PCR analysis also showed that OsPDIL1;1 was down-regulated among the 10 predicted genes in leaf blades of osa-miR5144-3p-ox lines (Fig. 1B), while the other nine genes did not display down-regulation in osa-miR5144-ox lines (Supplementary Fig. S1A). In addition, the expression of the other three OsPDIL genes (OsPDIL1;2, OsPDIL1;3 and OsPDIL1;4) was not affected in osa-miR5144-3p-ox lines (Supplementary Fig. S1B). This indicates that overexpression of osa-miR5144-3p reduces the expression of OsPDIL1;1 specifically, and suggests that OsPDIL1;1 mRNA is targeted by osa-miR5144-3p. To confirm further that OsPDIL1;1 is a target gene of osa-miR5144-3p, transgenic rice down-regulation of osa-miR5144-3p (STTM5144-3p) was generated by the short tandem target mimic (STTM) strategy (Yan et al. 2012). qRT-PCR analysis showed that the abundance of osa-miR5144-3p was effectively reduced in STTM5144-3p lines (Fig. 1C). This consequently led to the enhanced expression of OsPDIL1;1 (Fig. 1D).

Fig. 1

MicroRNA osa-miR5144-3p targets OsPDIL1;1 mRNA in rice. (A–D) Relative expression level of osa-miR5144-3p and OsPDIL1;1 measured by qRT-PCR in leaves of transgenic rice. The wild-type rice Zhonghua 11 (ZH11) and transgenic rice plants osa-miR5144-3p-ox (overexpressing osa-miR5144-3p) and STTM5144-3p (down-regulating osa-miR5144-3p) were grown in Hoagland’s solution for microRNA and RNA extraction. U6 and e-EF-1a were used as microRNA and mRNA reference genes, respectively. Means ± SD (n =15–20) represent three biological repeats. (E) Confirmation of osa-miR5144-3p-directed cleavage of OsPDIL1;1 mRNA. The arrows on miRNA:mRNA alignments indicate the cleavage sites and clone numbers of 10 sequencing clones identified in ZH11 by 5'-RLM-RACE. The asterisks below the alignment are cleavage sites detected by the degradome sequencing analysis by StarScan (Liu et al. 2015), and the detected cleavage frequencies (target RPM, target per million transcripts) are shown beside.

To identify the cleavage site of osa-miR5144-3p on OsPDIL1;1 mRNA, a 5'-RLM-RACE (5'-RNA ligase-mediated rapid amplification of cDNA ends) assay was performed using total RNA isolated from wild-type ZH11 grown under normal conditions. Sequencing results of the 5'-RLM-RACE clones showed that five of the 10 sequenced clones were cleaved in the complementary region of the osa-miR5144-3p and OsPDIL1;1 mRNA (Fig. 1E), while four clones showed consistency of the cleavage sites identified by degradome sequence analysis (Fig. 1E). The degradome search results showed that the detected cleavage frequencies in the 9th and the 11th on the osa-miR5144-3p:OsPDIL1;1 mRNA alignment site were 1.2 and 1.2 RPM (reads per million) (Fig. 1E), respectively, in wild-type rice seedlings grown under normal conditions (Li et al. 2010). These data indicate that osa-miR5144-3p can direct the cleavage of OsPDIL1;1 mRNA to regulate its abundance. All the above assays demonstrated that OsPDIL1;1 is a specific target of osa-miR5144-3p in rice.

OsPDIL1;1 is involved in protein–disulfide bond formation in vivo (Onda and Kobori 2014). The ER is a protein-folding compartment, especially for protein–disulfide formation (Kaufman 1999, Bulleid and Ellgaard 2011). Therefore, if OsPDIL1;1 is involved in protein–disulfide formation in vivo, it would be expected to be localized to the ER. Transient expression of OsPDIL1;1::EGFP (enhanced green fluorescent protein) fusion protein in rice protoplasts showed that OsPDIL1;1 was co-localized with the ER marker ER-rk (Nelson et al. 2007) (Fig. 2A–D), whereas free EGFP was not matched with the ER-rk signal (Fig. 2E–H). This indicates that OsPDIL1;1 is indeed targeted to the ER, which confirmed the subcellular location of OsPDIL1;1 proposed by Onda et al. (2011).

Fig. 2

OsPDIL1;1 is targeted to the ER. (A–D) OsPDIL1;1::EGFP (A) co-localized with the ER marker ER-rk::mCherry (B) in transiently transformed rice protoplasts; (D) merged image of (A) and (B). (E–H) Free EGFP (E) was not co-localized with the ER marker ER-rk::mCherry (F) in rice protoplasts as a control; (H) merged image of (E) and (F).

Abundance of OsPDIL1;1 transcript is inversely related to the expression level of osa-miR5144-3p in rice

The promoter of the OsaMIR5144 gene was used to drive the expression of β-glucuronidase (GUS) to investigate the OsaMIR5144 gene expression pattern (Fig. 3A). The GUS staining analysis showed that the OsaMIR5144 gene was strongly expressed in leaf blades and leaf sheaths, but weakly in young roots and spikelet of panicles. To investigate how OsPDIL1;1 expression is regulated by osa-miR5144-3p during rice development, the correlation between the levels of expression in each case was assayed using parallel analysis by qRT-PCR in rice organs. The results showed that the expression level of mature osa-miR5144-3p in leaf blades, leaf sheaths and culms was higher than that in roots and mature panicles (Fig. 3B). This result indicated that the mature osa-miR5144-3p expression pattern in developing rice is consistent with the expression pattern of the OsaMIR5144 gene obtained by GUS staining (Fig. 3A). In contrast, the expression level of OsPDIL1;1 in leaf blades and leaf sheaths was lower than that in roots, culms and panicles (Fig. 3C). These results suggest that the abundance of the OsPDIL1;1 transcript is inversely related to the expression of osa-miR5144-3p in leaves, panicles and roots, but not in culms. The leaf diurnal expression analysis showed that osa-miR5144-3p and OsPDIL1;1 displayed opposite diurnal expression patterns (Fig. 3D, E). OsPDIL1;1 is strongly expressed during daytime, while osa-miR5144-3p is highly expressed at night. All these data on expression analysis demonstrate that the expression level of OsPDIL1;1 is regulated by osa-miR5144-3p expression during rice development.

Fig. 3

Correlation between levels of expression of osa-miR5144-3p and OsPDIL1;1. (A) GUS staining in different organs of P_Osa-MIR5144:GUS flowering plants. (B–E) qRT-PCR analysis of osa-miR5144-3p and OsPDIL1;1 expression in wild-type ZH11 rice. Plants were grown under natural conditions for RNA and microRNA extraction. U6 and e-EF-1a were used as microRNA and mRNA reference genes, respectively. Means ± SD (n =15–20) represent three biological repeats. (B and C) osa-miR5144-3p (B) and OsPDIL1;1 (C) expression in different organs of ZH11. Pa, panicle. (D and E) Dinural expression of OsPDIL1;1 (D) and osa-miR5144-3p (E) in leaves of ZH11.

Expression of osa-miR5144-3p and OsPDIL1;1 is inversely regulated under abiotic stress

We explored whether the abundance of OsPDIL1;1 mRNA is also regulated by osa-miR5144-3p under abiotic stress (Fig. 4). Parallel expression analyses of precursor and mature forms of osa-miR5144-3p and OsPDIL1;1 in ZH11 were performed after one of various abiotic stress treatments using qRT-PCR. The expression level of the precursor (Fig. 4A) and mature (Fig. 4B) forms of osa-miR5144-3p decreased after three abiotic stress treatments, i.e. 150 mM NaCl, 37.5 μM HgCl₂ and 42°C; in contrast, the abundance of OsPDIL1;1 mRNA increased after the three treatments (Fig. 4C). Moreover, contrary to the trend described under the three stress treatments, dark treatment caused up-regulation of osa-miR5144-3p and down-regulation of OsPDIL1;1. Low temperature (4°C) treatment did not alter the expression levels of mature forms of osa-miR5144-3p and OsPDIL1;1, but the expression level of the precursor form of osa-miR5144-3p was down-regulated by the 4°C treatment. In addition, other putative target genes of osa-miR5144-3p did not display up-regulated expression under these stress conditions; four of the nine putative target genes showed down-regulation under some abiotic stresses (Supplementary Fig. S1D). These data suggest that osa-miR5144-3p responds to salt, mercury, dark and high temperature stresses to modulate the abundance of OsPDIL1;1 mRNA.

Fig. 4

qRT-PCR analyses of the correlation between expression of osa-miR5144-3p precursor (A) and mature (B) forms and OsPDIL1;1 (C) in ZH11 under abiotic stresses. ZH11 seedlings were grown in IRRI solution for 4 weeks, and then subjected to the following abiotic stress treatments: 4 °C, 150 mM NaCl, darkness, 37.5 μM HgCl₂ and 42 °C. Samples were harvested at different time after treatments for isolation of total RNA or small RNA. U6 and e-EF-1a were used as microRNA and mRNA reference genes, respectively. Means ± SD (n =15–20) represent three biological repeats. The insert in (C) is an enlargement of the dark treatment.

osa-miR5144-3p is involved in rice seed development

OsPDIL1;1-RNAi (down-regulation of OsPDIL1) and OsPDIL1;1-ox (overexpression of OsPDIL1;1) (Supplementary Fig. S2A, B), as well as osa-miR5144-3p-ox and STTM5144-3p (Fig. 1), were generated to investigate the role of osa-miR5144-3p in rice. Under normal growth conditions, osa-miR5144-3p-ox and OsPDIL1;1-RNAi produced grains with a floury, chalky endosperm (Fig. 5A), and low 1,000-seed weight (Fig. 5G). Scanning electron microscopy (SEM) analysis revealed that the starch granules of osa-miR5144-3p-ox and OsPDIL1;1-RNAi were small and loosely packed (Fig. 5C, D), compared with those of ZH11 (Fig. 5B); while seeds of STTM 5144-3p and OsPDIL1;1-ox were of normal appearance (Fig. 5 E, F) and had a higher 1,000-seed weight (Fig. 5G) than that of ZH11. This showed that up-regulation of osa-miR5144-3p and down-regulation of OsPDIL1;1 cause a similar phenotype of seeds.

Fig. 5

Altered expression of osa-miR5144-3p and OsPDIL1;1 affects the development of rice seeds. Rice plants were grown in the field under natural conditions. (A) Picture of dry mature seeds. (B–F) Scanning electron microscopy analysis of mature endosperms. Scale bar = 1 cm in (A) and 200 μm in (B–F). (G) The 1,000-seed weight. (H) Correlation between expression of osa-miR5144-3p and OsPDIL1;1 in developing panicles by qRT-PCR analysis. P1–P5 are panicles at stages of pollen mother cell (P1), pollen meiosis (P2), mononuclear pollen (P3), mature pollen (P4) and 7 d after pollen fertilization (P5). Means ± SD (n=15–20) represent three biological repeats. An asterisk represents a significant difference (P < 0.05) in comparison with ZH11 by Student’s t-test.

To investigate the correlation between expression of osa-miR5144-3p and OsPDIL1;1 during seed development, changes in the level of expression in each case were analyzed in developing panicles (Fig. 5H). osa-miR5144-3p expression decreased, while that of OsPDIL1;1 increased as panicle development proceeded (Fig. 5H). These results demonstrate that osa-miR5144-3p plays a role in seed development by modulating OsPDIL1;1 expression.

osa-miR5144-3p and OsPDIL1;1 mediate rice tolerance to abiotic stress

osa-miR5144-3p and OsPDIL1;1 showed inverse expression under several abiotic stress conditions (Fig. 4); therefore, we checked the effects of the changes in their expression on rice tolerance to these stress situations (Fig. 6). Under normal growing conditions, transgenic seedlings which showed altered expression of OsPDIL1;1 and osa-miR56144-3p did not show differences in growth (Fig. 6A, E, G). However, under conditions of salt, high temperature, low temperature or HgCl₂ stress, OsPDIL1;1-ox and STTM-5144-3p rice seedlings displayed a higher degree of tolerance, while osa-miR5144-3p-ox and OsPDIL1;1-RNAi rice seedlings displayed a lower level of tolerance (Fig. 6B, D, F, H, I), when compared with ZH11. All osa-miR5144-3p-ox and OsPDIL1;1-RNAi seedlings died after 7 d in the dark, 10 d at 42°C, 3 d at 4°C and 10 d at 150 mM NaCl, but higher survival rates were observed for OsPDIL1;1-ox and STTM-5144-3p under all stress treatments tested, except for dark treatment (Fig. 6B–D, I). After 2 weeks at 37.5 μM HgCl₂, growth and new lateral roots in osa-miR5144-3p-ox and OsPDIL1;1-RNAi seedlings were severely hindered, but OsPDIL1;1-ox and STTM-5144-3p seedlings grew better and developed newer lateral roots than ZH11 (Fig. 6H, J). Roots and shoots of osa-miR5144-3p-ox and OsPDIL1;1-RNAi seedlings accumulated more mercury than their OsPDIL1;1-ox and STTM-5144-3p counterparts (Fig. 7A, B). The HgCl₂ shoot/root rate in osa-miR5144-3p-ox and OsPDIL1;1-RNAi was higher than that in OsPDIL1;1-ox and STTM-5144-3p, after HgCl₂ treatment (Fig. 7C), which may suggest that more mercury was translocated from roots to shoots in osa-miR5144-3p-ox and OsPDIL1;1-RNAi than in OsPDIL1;1-ox or STTM-5144-3p seedlings. Three high temperature tolerance-related genes (Bip, Hsp901 and Hsp70) were checked after high temperature treatment (42°C) (Fig. 8). Although they were all up-regulated in all seedlings after exposure to high temperature, Hsp90 and Hsp70 were up-regulated to a greater extent in STTM5144-3p and OsPDIL1;1-ox than in osa-miR5144-3p-ox or OsPDIL1;1-RNAi seedlings. Expression of OsPDIL1;1 was induced in the wild type, STTM5144-3p and OsPDIL1;1-ox by high temperature treatment, but it was not induced in osa-miR5144-3p-ox or OsPDIL1;1-RNAi. These results suggest that osa-miR5144-3p is involved in mediating rice tolerance to HgCl₂, salt, and high and low temperature stress conditions.

Fig. 6

Altered expression of osa-miR5144-3p and OsPDIL1;1 affects abiotic stress tolerance of rice. (A, E and G) Pictures of ZH11 and the altered osa-miR5144-3p- and OsPDIL1;1-expressing rice which were grown in Hoagland’s solution (A) or medium (G) under normal conditions. (B–D, F and H) Pictures of ZH11 and the altered osa-miR5144-3p- and OsPDIL1;1-expressing rice which were grown in Hoagland’s solution with 150 mM NaCl for 10 d (B), in darkness for 7 d (C), at 42 °C for 10 d (D), at 4 °C for 3 d (F) or in Hoagland’s medium with 37.5 μM HgCl₂ for 7 d (H). (I) Survival rate after abiotic stress treatments. (J) Length of shoot and roots after 37.5 μM HgCl₂ treatment. Means ± SD (n =15–20) represent three biological repeats. Asterisks represent significant differences (P < 0.01) in comparison with ZH11 by Student’s t-test (I, J).

Fig. 7

osa-miR5144 affects mercury allocation in rice. Rice seedlings with altered expression of osa-miR5144-3p and OsPDIL1;1 were grown in Hoagland’s solution with 37.5μM HgCl₂ for 7 d and then harvested for Hg measurement. (A and B) Hg accumulation in roots (A) and shoots (B) after Hg treatment. (C) Shoot/root Hg content ratio after Hg treatment. Means ± SD (n =15–20) represent three biological repeats. An asterisk represents a significant difference (P < 0.01) in comparison with ZH11 by Student’s t-test.

Fig. 8

Expression of high temperature tolerance-related genes (OsPDIL1;1, OsBip, OsHsp90 and OsHsp70) in rice seedlings with altered osa-miR5144-3p and OsPDIL1;1 expression. Seedlings were kept at 42 °C for 10 d prior to RNA isolation.

Formation of disulfide bonds is affected by osa-miR5144-3p

OsPDIL1;1 can control the stability and activity of target proteins by regulating formation and isomerization of protein–disulfide bonds (Onda et al. 2011). Therefore, we examined free (Fig. 9A) and total (Fig. 9B) sulfhydryl contents in rice leaves after altering the expression of osa-miR5144-3p and OsPDIL1;1, then calculated their protein–disulfide bond content (Fig. 9C) as the difference between total sulfhydryl and free sulfhydryl. Under normal growing conditions, total sulfhydryl content was significantly reduced in both osa-miR5144-3p-ox and OsPDIL1;1-RNAi, but it did not change in either OsPDIL1;1-ox or STTM5144-3p, when compared with ZH11 (Fig. 9A); however, the free sulfhydryl content in both osa-miR5144-3p-ox and OsPDIL1;1-RNAi was reduced to a lesser extent than in either OsPDIL1;1-RNAi or STTM5144-3p (Fig. 9B). Therefore, protein–disulfide bond content was lower in both OsPDIL1;1-RNAi and osa-miR5144-3p-ox, and higher in both STTM5144-3p and OsPDIL1;1-ox (Fig. 9C), than in ZH11. These results imply that more protein–disulfide bonds are formed in both STTM5144-3p and OsPDIL1;1-ox than in either OsPDIL1;1-RNAi or osa-miR5144-3p-ox. This suggests that altered expression of osa-miR5144-3p and OsPDIL1;1 affects folding of protein–disulfide bonds in rice.

Fig. 9

Formation of disulfide bonds is affected by osa-miR5144-3p. (A–C) Content of total sulfhydryl (A), free sulfhydryl (B) and protein–disulfide bonds (C) in the altered osa-miR5144-3p- and OsPDIL1;1-expressing rice under normal conditions, at high temperature or under mercury stress. The abundance of protein–disulfide bonds was calculated from total sulfhydryl (A) minus free sulfhydryl (B). Means ± SD (n =15–20) represent three biological repeats. Different letters in the coloumn represent significant differences (P < 0.01) from each other by Student’s t-test.

The total sulfhydryl content was reduced to a larger extent than that of free sulfhydryl after a stress treatment at high temperature (42°C) or with HgCl₂, when compared with that in ZH11 under normal growing conditions (Fig. 9A). In contrast, free sulfhydryl content was only slightly reduced (Fig. 9B) in both OsPDIL1;1-RNAi and osa-miR5144-3p-ox, while total sulfhydryl content was increased (Fig. 9A). On the other hand, free sulfhydryl content was severely reduced (Fig. 9B) in both STTM5144-3p and OsPDIL1;1-ox. Therefore, after stress treatments, contents of protein–disulfide bonds decreased in both OsPDIL1;1-RNAi and osa-miR5144-3p-ox, and increased in both STTM5144-3p and OsPDIL1;1-ox (Fig. 9C). These results imply that after HgCl₂ or high temperature treatment, more free sulfhydryl in both STTM5144-3p and OsPDIL1;1-ox, and less free sulfhydryl in both OsPDIL1;1-RNAi and osa-miR5144-3p-ox was utilized to form protein–disulfide bonds, than in ZH11. These observations indicate that formation of protein disulfide bonds is regulated by osa-miR5144-3p through modulation of OsPDIL1;1 expression under stress conditions.

Discussion

Formation of disulfide bonds is important for proper folding of nascent polypeptides into functional proteins, as well as for plant development and resistance to stress. PDI is one of the folding catalysts of formation, isomerization and reduction/oxidation of disulfide bonds (Houston et al. 2005). PDIs are involved in several processes, such as the unfolded protein response in the ER and folding and secretion of storage proteins. The large diversity of PDIs in higher eukaryotes is considered to be essential to maintain specificity for a wide variety of potential substrate proteins. So far in plants, only a few predicted PDIs have been characterized in detail (Aller and Meyer 2013). OsPDIL1;1 is one of the 19 rice protein–disulfide isomerase-like proteins (Houston et al. 2005), and has been experimentally demonstrated to possess the highest catalytic activity for both disulfide bond formation and disulfide bond reduction among three rice PDILs (OsPDIL1;1, OsPDIL1;4 and OsPDIL2;3) (Onda et al. 2011). OsPDIL1;1 plays an important role in formation of storage protein–disulfide bonds in rice endosperm (Onda et al. 2011, Han et al. 2012). However, little is known about the regulation of PDI genes. So far, we only know that AtbZIP60 is partially required for induction of AtPDI expression upon ER stress in Arabidopsis (Lu and Christopher 2008a). Here, we have demonstrated that OsPDIL1;1 is regulated by the microRNA osa-miR5144-3p during rice seed development and under abiotic stress to mediate formation of protein disulfide bonds.

osa-miR5144 is a rice-specific microRNA from the OsMIR5144 gene, with no homologs found in other species thus far. The mature form of osa-miR5144-3p has been isolated from developing grains, leaves, roots, stems and in vitro cultured embryogenic calli by microRNA sequencing (Zhu et al. 2008, Chen et al. 2011, Campo et al. 2013, Cheah et al. 2015). The mature form of osa-miR5144-3p was also detected by our qRT-PCR (Fig. 4B). We were also able to detect its precursor form using the PCR primers located on the stem–loop of the predicted osa-miR5144-3p precursor (Fig. 4A). These results showed that osa-miR5144-3p is actually a transcribed microRNA from the rice OsaMIR5144 gene. Many non-conserved miRNAs are thought to be non-functional because no predicted targets have been validated (Li et al. 2010). Here, osa-miR5144-3p was demonstrated to play various roles in rice by targeting OsPDIL1;1 mRNA. Overexpression of osa-miR5144-3p caused reduced expression of OsPDIL1;1 in the transgenic rice osa-miR5144-3p-ox (Fig. 1A, B). In contrast, down-regulation of osa-miR5144-3p caused up-regulation of OsPDIL1;1 in STTMR5144-3p transgenic rice (Fig. 1C, D). Under normal growing conditions, osa-miR5144-3p could direct OsPDIL1;1 mRNA cleavage at the osa-miR5144-3p/OsPDIL1;1 mRNA complementary site, which was confirmed by 5'-RLM-RACE and degradome sequencing data analysis (Fig. 1E). The 5'-RLM-RACE results showed that under normal growing conditions, only five of 10 clones were cleaved at the complementary region of osa-miR5144-3p and OsPDIL1;1 mRNA in leaves of ZH11; degradome sequencing analysis in rice seedlings showed that cleavage frequencies in the 9th and the 11th on osa-miR5144-3p:OsPDIL1;1 mRNA alignment sites were 1.2 and 1.2 RPM (reads per million), respectively (Fig. 1E). This may suggest that the cleavage frequency of OsPDIL1;1 mRNA may be relatively low under normal growing conditions, leading to a sustained basic expression of OsPDIL1;1 (Fig. 1B). In addition, enhanced expression of osa-miR5144-3p and reduced expression of OsPDIL1;1 in transgenic rice produced a similar phenotype among developing seeds and a common response to abiotic stress (Fig. 5, 6). These results suggest that the abundance of OsPDIL1;1 transcripts is regulated by osa-miR5144-3p.

osa-miR5144-3p is involved in rice seed endosperm development. Previously reported (Onda et al. 2011) and here confirmed ER-localized OsPDIL1;1 (Fig. 2) has been found to facilitate the oxidative folding of vacuole-targeted storage proteins in rice seeds, such as proglutelins and α-globulin, and to refold reduced, denatured RNase and α-globulin with a higher efficiency than OsPDIL2;3 in a GSH/GSSG ratio-dependent manner (Onda et al. 2011). AtPDIL1;3 is involved in the synthesis of transitory starch in the leaves of Arabidopsis (Lu and Christopher 2006). OsPDIL1;1 has also been found to play an important role in starch synthesis. Its absence is associated with ER stress in the endosperm, which is likely to underlie the formation of the floury endosperm in the T3612 mutant (an ospdil1;1 mutant) (Han et al. 2012). SEM analysis revealed that the starch granules in the endosperm of osa-miR5144-3p-ox and OsPDIL1;1-RNAi were round and loosely packed (Fig. 5B), similar to starch granules in the ospdil1;1 mutant (Han et al. 2012). The 1,000-seed weight of both osa-miR5144-3p-ox and OsPDIL1;1-RNAi was lower than in ZH11; in contrast, the 1,000-seed weight in both STTM 5144-3p and OsPDIL1;1-ox was higher than in ZH11 (Fig. 5C). Expression of osa-miR5144-3p and OsPDIL1;1 in developing panicles was inversely correlated (Fig. 5D). These results suggest that osa-miR5144-3p is involved in the formation of endosperm starch granules, and that it may also mediate folding of storage proteins in rice seeds. osa-miR5144-3p finally affects seed development by mediating OsPDIL1;1 expression. This suggests that normal expression of osa-miR5144-3p is also necessary for normal rice grain development.

Knockdown of the chloroplast AtPDI6 conferred resistance to photoinhibition under high light intensity in Arabidopsis, which is probably due to AtPDI6 acting by attenuating, repressing or activating synthesis of the PSII core protein D1 in response to light and redox signals (Wittenberg et al. 2014). However, down-regulation of rice OsPDIL1;1 rendered the seeds less tolerant to abiotic stress (Fig. 6). This difference may be due to the location of OsPDIL1;1 in the ER and suggests that OsPDIL1;1 may be involved in protein folding in the ER. The inverse relationship between expression of osa-miR5144-3p and OsPDIL1;1 occurs in leaves and roots, during leaf diurnal change in developing panicles and under some conditions of abiotic stress (Figs. 3, 4, 5D). This may imply that expression of OsPDIL1;1 is regulated not only in seeds but also in other organs. Under normal growing conditions, vegetative growth of all transgenic plants with altered osa-miR5144-3p and OsPDIL1;1 expression did not show an obvious phenotype (Fig. 6A) that may have resulted from down-regulation of OsPDIL1;1; however, grain development of these transgenic plants was different (Fig. 5), maybe because of a relatively high expression of OsPDIL1;1 in the ovary, the embryo and the early endosperm of rice (Supplementary Fig. S1C). Under abiotic stress conditions, down-regulation of osa-miR5144 and up-regulation of OsPDIL1;1 made rice seedlings more tolerant to mercury, salt, and high and low temperature stress (Fig. 6). In contrast, up-regulation of osa-miR5144-3p and down-regulation of OsPDIL1;1 caused reduced tolerance of rice seedlings to these stress conditions (Fig. 6) when compared with ZH11. Overexpression of OsPDIL1;1 and down-regulation of osa-miR5144-3p reduced the accumulation of mercury in roots and shoots. This inverse expression of OsPDIL1;1 and osa-miR5144-3p in particular reduced translocation of mercury from roots to shoots (Fig. 7). These results showed that high expression of OsPDIL1;1 could prevent mercury uptake and translocation from the root to the shoot, which implies that correct folding of proteins can protect rice from accumulating mercury to a significant extent. Three high temperature tolerance-related genes (Bip, Hsp90 and Hsp70) were significantly up-regulated more in STTM5144-3p and OsPDIL1;1-ox than in osa-miR5144-3p-ox or OsPDIL1;1-RNAi after high temperature treatment (Fig. 8), which may lead to higher resistance to high temperature stress (Fig. 5D).

However, the inverse expression change of osa-miR5144-3p and OsPDIL1;1 under dark treatment was different from that observed under the other stress conditions tested (Fig. 4). Unlike the higher survival rate of STTM5144-3p and OsPDIL1;1-ox under the three stress conditions, dark treatment actually caused the death of STTM5144-3p and OsPDIL1;1-ox seedlings (Fig. 6). In addition, expression of the precursor form of osa-miR5144 was down-regulated by low temperature (Fig. 4B, C), while expression of the mature osa-miR5144-3p and OsPDIL1;1 was not affected by low temperature. This means that the regulation of the expression of osa-miR5144-3p observed as an effect of low temperature treatment is more complex (Fig. 4B, C). The altered expression of osa-miR5144-3p and OsPDIL1;1 in rice under low temperature showed a similar phenotype to that under HgCl₂, NaCl and high temperature (Fig. 6F). These data may imply that other unknown mechanisms are acting in these transgenic plants. In summary, at least under stresses of Hg, NaCl and high temperature, osa-miR5144-3p is involved in tolerance of rice to these abiotic stress conditions by mediating the expression of OsPDIL1;1. Resistance to abiotic stress and less accumulation of mercury in OsPDIL1;1-overexpressing rice and osa-miR5144-3p down-regulating rice, suggest that they are potentially useful traits for rice molecular breeding.

Materials and Methods

Vector construction and rice transformation

Mature 21 bp osa-miR5144-3p sequence (MI0018056) was downloaded from miRBase (http://www.mirbase.org). The vector construct with artificial osa-miR5144-3p was inserted downstream from the Ubi-1 promoter in pXU1301 (Xia et al. 2015a). The STTM method was used to reduce expression of osa-miR5144-3p, (Yan et al. 2012). The fragment GAAGCTTTCTTCTTGTGCTGTAGCTGAGGAGAGTTGTTGTTGTTATGGTCTAGTTGTTGTTGTTATGGTCTAATTTAAATATGGTCTAAAGAAGAAGAATATGGTCTAAAGAAGAAGAATCTTCTTGTGCTGCTACTGAGGAGAGGATCCA was synthesized by Invitrogen™, and inserted downstream from the Ubi-1 promoter in pXU1301 (Xia et al. 2015a). Therefore, two target mimic sequences (underlined) were in the fragment. For generating the OsPDIL1;1-RNAi rice plants, two fragments of the OsPDIL1;1 cDNA (284 bp) were amplified by PCR and transferred downstream from the Ubi-1 promoter in the rice RNA interfence (RNAi) vector pTCK303 (Wang et al. 2004). Full-length cDNA of OsPDIL1;1 was inserted downstream from the Ubi-1 promoter in pXU1301 to overexpress OsPDIL1;1. A 2,001 bp promoter fragment of the OsaMIR5144 gene was amplified by PCR for P_OsaMIR5144:GUS construction; this fragment was then inserted into pCAMBIA1301 to replace the 35S promoter via KpnI and NcoI.

All these constructs were introduced into Agrobacterium tumefaciens strain EHA105 and the japonica rice variety ZH11 was transformed by Agrobacterium-mediated transformation (Hiei et al. 1997). All primers and genes used in this study are listed in Supplementary Tables S1 and S2.

Subcellular localization of OsPDIL1;1

To study the subcellular localization of OsPDIL1;1 in rice cells, OsPDIL1;1::GFP and ER-rk::mCherry (Nelson et al. 2007) were transiently co-expressed in rice protoplasts prepared and transformed by the method described previously (Zhang et al. 2011). Fluorescent proteins were visualized using laser confocal microscopy (ZEISS-510; Zeiss).

Prediction and identification of osa-miR5144-3p mRNA targets

Putative targets of osa-miR5144-3p were searched using a web-based prediction program (miRU; http://plantgrn.noble.org/psRNATarget/).

RNA extraction and reverse transcription–PCR (RT–PCR)

The extraction of small RNA and total RNA from rice, reverse transcription and qRT-PCR amplification were performed as previously described (Xia et al. 2012). To differentiate the expression patterns of the precursor and mature form of osa-miR5244-3p, different primer pairs and a reverse transcription strategy were utilized. PCR primers located at the stem–loop of osa-miR5144 were used to detect the expression level of the precursor form. The mature form was reverse transcribed with an anchor primer (GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACCTTCTTGT), and subsequently detected by primer pairs (TCTCCTCAGCAGCACAAGAAG, GTGCAGGGTCCGAGGTATTC) specific to the transcribed cDNA copy. Data of OsPDIL1;1 expression under the entire life cycle were verified with the data from RiceXpro (http://ricexpro.dna.affrc.go.jp/). All qRT-PCR data were presented from three or four biological repeat experiments.

5'-RLM-RACE assay

Total RNA from leaf sheaths of the wild-type ZH11 at the tillering stage was directly ligated to a synthesized RNA adaptor (GCUGAUGGCGAUGAAUGAACACUGCGUUUGCUGGCUUUGAUGAAA). Other processing for 5'-RLM-RACE was as previously described (Xia et al. 2015b).

Scanning electron microscopy

SEM was used for observation of starch granules in seeds. The methods were as previously described (Liu et al. 2006). Mature seeds were cut using a sharp blade and sputter-coated with gold–palladium in six 30 s bursts (JEE-420, Hitachi) in preparation for SEM (JSM-6360LV, Hitachi).

Mercury treatment

In order to test the mercury poisoning effect on seed germination, seeds sterilized with 10% NaOCl were planted in 1/2 Murashige and Skoog (MS) culture medium with 37.5 μM HgCl₂ and germinated for 7 d. In order to test the mercury poisoning effect on seedling growth, rice seedlings were grown in pots for 2 weeks at 28°C, kept on a 16/8 h light/dark cycle and watered with Hoagland’s solution (Yoshida et al. 1976). They were then transferred to Hoagland’s medium with 37.5 μM HgCl₂ added for 1 week.

High temperature treatment

Four-week-old rice seedlings grown in Hoagland’s solution (Yoshida et al. 1976) were transferred for 10 d to a chamber at 42°C and kept there on a 16/8 h light/dark cycle.

Salt treatment

To examine the effect of salt on seedling growth, 4-week-old seedlings were transferred into Hoagland’s solution with 150 mM NaCl for 10 d.

Low temperature treatment

Four-week-old rice seedlings grown in Hoagland’s solution (Yoshida et al. 1976) were transferred to a chamber for 3 d at 4°C and kept on a 16/8 h light/dark cycle; at the end of the 3 d period, they were grown at room temperature on the same light/dark regime for one more week.

Mercury content assay by inductively coupled plasma-mass spectrometry (ICP-MS)

Rice seedlings were grown for 4 weeks in pots containing 16 liters of Hoagland’s solution (Yoshida et al. 1976) at 28°C and kept on a 16/8 h light/dark cycle. Then, uniform seedlings were transferred to Hoagland’s solution containing 37.5 μM HgCl₂ for 7 d. After this treatment, all roots were dipped in 20 mM Na₂-EDTA for 20 min, washed thoroughly with Milli-Q water, dried at 105°C for 20 min and finally left at 70°C overnight until constant weight. Dried plants were ground to powder in an acid-cleaned agate mortar of MM400 Retsch (Germany) and passed through a 1 mm sieve. A 200 mg aliquot of ground fresh plant samples was weighed into digestion vials, mixed with 5 ml of HNO₃ and 0.5 ml HCl, then digested in Anton Paar Multiwave Pro and finally diluted up to 50 ml. Next the solution was diluted to 1/1,000 and measured by Agilent Technologies 7700x Series ICP-MS. Internal standards were added to ensure accuracy and precision. Standard solutions at 1 μg l⁻¹ Hg were measured every 20 samples to monitor the stability of ICP–MS.

Degradome sequence analysis

StarScan (Liu et al. 2015) was used to find sliced miRNA targets from degradome sequence data. The mature osa-miR5144-3p sequence was used as input to search the Japonica rice seedling degradome data set (GSE17398) (Li et al. 2010).

Data analysis

All results are presented as the mean ± SD of three replicates. Statistical analysis was performed using PASW statistics 18.0, including one-way analysis of variance (ANOVA) and the least significant difference (LSD) test to separate means. A difference was considered significant at P < 0.05 and highly significant at P < 0.01.

Supplementary Data

Supplementary data are available at JXB online.

Funding

This work was supported by the National Natural Science Foundation of China [grant No. 31671659/31701403/31772384]; Guangdong Agriculture Department of China [grant No. Yuenongji 201742; the National Key Research and Development Program of China [grant No. 2017YFD0100100]; and Guangdong ‘Pearl River Talents Plan’-Postdoctoral Project.

Acknowledgments

We thank X.X. Zhang for help with growing rice.

Disclosures

The authors have no conflicts of interest to declare.

Abbreviations

Abbreviations
 
• EGFP

  enhanced green fluorescent protein

 
• ER

  endoplasmic reticulum

 
• GUS

  β-glucosidase

 
• ICP-MS

  inductively coupled plasma-mass spectrometry

 
• miRNA

  microRNA

 
• osa-miR5144-3p

  Oryza sativa miR5144-3p

 
• OsPDI, Oryza sativa

  protein disulfide isomerase

 
• qRT-PCR

  quantitative real-time PCR

 
• 5'-RLM-RACE

  5'-RNA ligase-mediated rapid amplification of cDNA ends

 
• RNAi

  RNA interference

 
• SEM

  scanning electron microscopy

 
• STTM

  short tandem target mimic method