Plant Biotechnology Journal (2014), pp. 1–14
doi: 10.1111/pbi.12218
System analysis of microRNAs in the development and aluminium stress responses of the maize root system Xiangpei Kong1, Maolin Zhang1, Xiangbo Xu2, Xiaoming Li3, Cuiling Li1 and Zhaojun Ding1,* 1
The Key Laboratory of Plant Cell Engineering and Germplasm Innovation, College of Life Sciences, Shandong University, Jinan, China
2
Maize Institute, Shandong Academy of Agricultural Sciences/National Maize Improvement Sub-Center/National Engineering Laboratory for Wheat and Maize, Jinan,
China 3
Shandong Provincial Key Laboratory of Agricultural Microbiology, College of Plant Protection, Shandong Agriculture University, Tai’an, China
Received 2 March 2014; revised 21 May 2014; accepted 23 May 2014. *Correspondence (Tel +86 531 8836 2351; email [email protected])
Keywords: Al stress, maize, miRNAs, root development.
Summary MicroRNAs (miRNAs) are a class of regulatory small RNAs (sRNAs) that down-regulate target genes through mRNA cleavage or translational inhibition. miRNA is known to play an important role in the root development and environmental responses in both the Arabidopsis and rice. However, little information is available to form a complete view of miRNAs in the development of the maize root system and Al stress responses in maize. Four sRNA libraries were generated and sequenced from the early developmental stage of primary roots (PRY), the later developmental stage of maize primary roots (PRO), seminal roots (SR) and crown roots (CR). Through integrative analysis, we identified 278 miRNAs (246 conserved and 32 novel ones) and found that the expression patterns of miRNAs differed dramatically in different maize roots. The potential targets of the identified conserved and novel miRNAs were also predicted. In addition, our data showed that CR is more resistant to Al stress compared with PR and SR, and the differentially expressed miRNAs are likely to play significant roles in different roots in response to environmental stress such as Al stress. Here, we demonstrate that the expression patterns of miRNAs are highly diversified in different maize roots. The differentially expressed miRNAs are correlated with both the development and environmental responses in the maize root. This study not only improves our knowledge about the roles of miRNAs in maize root development but also reveals the potential role of miRNAs in the environmental responses of different maize roots.
Introduction MicroRNAs (miRNAs) are a class of endogenous, 20–24 nucleotides, noncoding small RNAs (sRNAs), which negatively regulate target gene expression at the post-transcriptional levels through mRNA cleavage or translation inhibition (Bartel, 2004; Carrington and Ambros, 2003; He and Hannon, 2004; Li et al., 2013b). miRNAs are encoded by MIRNA genes that are transcribed into the primary miRNAs (pri-miRNAs) by RNA polymerase II, which, in turn, are cleaved twice by the RNase III enzyme Dicer-like 1 (DCL1) to yield mature miRNAs (Bartel, 2004; Jones-Rhoades et al., 2006; Mallory and Bouche, 2008). Accumulating evidence indicates that miRNAs play crucial regulatory roles not only in response to a broad variety of abiotic (Covarrubias and Reyes, 2010; Zhang et al., 2009a; Zhao et al., 2007; Zhou et al., 2013b) and biotic stresses (Eckardt, 2012; Ruiz-Ferrer and Voinnet, 2009) but also in response to nutrient stresses such as nitrate (Chiou, 2007; Vidal et al., 2010; Zhao et al., 2011), phosphate (Kuo and Chiou, 2011) and sulphate (Fujii et al., 2005; Liang et al., 2010) in plants. Over the past few years, knowledge about the roles of miRNAs in regulating root growth and development in response to developmental signals or different environmental cues has increased dramatically (Bian
et al., 2013; Khan et al., 2011; Kidner and Martienssen, 2005; Marin et al., 2010; Rubio-Somoza and Weigel, 2011). Recent studies showed that miR160-ARF10/ARF16/ARF17 module plays important roles in root cap formation, lateral root development and primary root growth in both Arabidopsis and rice (Gutierrez et al., 2009; Mallory et al., 2005; Meng et al., 2009; Wang et al., 2005). Interestingly, miR167–ARF6/ARF8 and miR160–ARF17 have opposite roles in the adventitious roots development through their control of the auxin homoeostatic enzyme GH3.6 in Arabidopsis (Gutierrez et al., 2009; Tian et al., 2004). In contrast to A. thaliana, miR167 positively regulate adventitious root development in rice (Meng et al., 2009). The miR164–NAC1 module affects lateral root initiation (Guo et al., 2005; Li et al., 2012; Meng et al., 2009), whereas miR390–TAS3–ARFs together with miR167–ARF8 and miR393–AFB3 are involved in the emergence and elongation of the lateral root (Marin et al., 2010; Meng et al., 2009; Vidal et al., 2010; Williams et al., 2005; Yang et al., 2006). Together, all these investigations demonstrate that certain miRNA plays crucial regulatory roles in root development. Maize (Zea mays) is one of the oldest and most important worldwide crops. It provides one of the most important human food supply and animal feed, and it is also crucial raw material
Please cite this article as: Kong, X., Zhang, M., Xu, X., Li, X., Li, C. and Ding, Z. (2014) System analysis of microRNAs in the development and aluminium stress responses of the maize root system. Plant Biotechnol. J., doi: 10.1111/pbi.12218
ª 2014 Society for Experimental Biology, Association of Applied Biologists and John Wiley & Sons Ltd
1
2 Xiangpei Kong et al. plant for energy production (Kong et al., 2011). The root system of maize displays a complex root structure comprising several root types, which include embryonic primary and seminal roots, and postembryonic shoot-borne and lateral roots (Hochholdinger et al., 2004; Orman-Ligeza et al., 2013). Embryonic primary roots can be observed only 2–3 days after germination. Seminal roots which emerge from the scutellar node become visible approximately 7 days after germination. Shoot-borne roots are initiated from underground and aboveground nodes of the stem and are called crown and brace roots, respectively. Crown roots start to emerge approximately 10 days postgermination, whereas brace roots become visible much later, approximately 6 weeks after germination. Lateral roots which initiate in the pericycle of all above major roots also belong to the postembryonic root system. The embryonic root system in maize plays important roles during the early stages of plant development, whereas the postembryonic shoot-borne root system makes up the major backbone of the maize root stock during the later developmental stage (Orman-Ligeza et al., 2013). However, how the different maize roots are developmental and spatially regulated is not well understood. Genetic studies have identified and characterized 9 specific root mutants in maize. Jenkins (1930) reported the first maize mutant rt1, which reduces shoot-borne root formation. Another maize shoot-borne root mutant rtcs lacks the initiation of the seminal and the shoot-borne roots (Taramino et al., 2007). More recently, Majer et al. (2012) reported that RTCS encodes a LATERAL ORGAN BOUNDARIES domain (LBD) protein and ZmARF34 binds the promoter of RTCS to activate downstream gene expression, which is involved in the initiation and maintenance of seminal and shoot-borne roots. The lateral rootless1 (lrt1) mutant affects the initiation of lateral roots (Hochholdinger and Feix, 1998), while the slr1 and slr2 mutants show reduced lateral root elongation only in the embryonic roots (Hochholdinger et al., 2001). Recently, Woll et al. (2005) characterized another lateral root mutant rum1, which is deficient in the initiation of the seminal roots and the lateral roots in the primary root. Map-based cloning revealed that RUM1 encodes a truncated ZmIAA10 sequence that lacks the auxin degron sequence. ZmIAA10 can interact with ZmARF25 and ZmARF34, which are required for the initiation of seminal and lateral root initiation in primary roots (Woll et al., 2005). Three mutants (rth1, rth2 and rth3) have been identified with defects in root hair elongation (Wen and Schnable 1994). Currently, 321 mature maize miRNAs have been registered on miRBase (release 20) (Zhang et al., 2009b), and a number of miRNAs with specific function have been characterized in maize (Mica et al., 2006; Zhang et al., 2006b, 2009b). For example, zma-miR172 down-regulates an APETALA2-like gene, glossy15, to promote vegetative phase change in maize (Lauter et al., 2005). More recently, genomewide studies using high-throughput approaches have identified some new miRNAs and explored the roles of miRNAs in response to diverse environmental cues in maize, including short-term and long-term waterlogging stress (Liu et al., 2012b; Zhai et al., 2013), salt (Ding et al., 2009), low nitrate (Trevisan et al., 2012; Xu et al., 2011; Zhao et al., 2013) and low phosphorus (Pei et al., 2013; Zhang et al., 2012). Furthermore, the roles of miRNAs were also investigated in different tissues and organs, such as seeds (Ding et al., 2012; Kang et al., 2012; Li et al., 2013a; Wang et al., 2011), ears (Ding et al., 2014; Liu et al., 2014), endosperm (Gu et al., 2013), pollen (Li et al., 2013c) and brace root (Liu et al., 2012a). Although
miRNAs play crucial regulatory roles in root development, little information, to our knowledge, is available to form a complete view of miRNAs in the development of the maize root system. Furthermore, the mechanisms underlying the regulation of miRNAs in different maize root types remain elusive. And several molecular mechanisms have been recently reported in Arabidopsis about how primary and lateral roots have differential growth responses in response to environmental cues such as salt, gravity and plant nutrition availability (Duan et al., 2013; Rosquete et al., 2013; Tian et al., 2014; Vidal et al., 2013), which imply the differential regulation mechanisms between primary and lateral roots. These investigations also suggest the importance of the comparative investigations to our fully understand the molecular mechanism of root growth and development. Therefore, to better understand the development of the maize root system that has a typical fibrous root system, it is necessary for us to compare the roles of miRNAs in different maize root types. In this study, we generated and sequenced four maize small RNA libraries from the early developmental stage of primary roots (PRY), the later developmental stage of maize primary roots (PRO), the early developmental stage of seminal roots (SR), and the early developmental stage of crown roots (CR) using Solexa high-throughput sequencing technology and obtained a comprehensive view of the variation in miRNAs’ expression patterns in the four samples. Subsequently, we also investigated the correlation between the expression profiles of miRNAs and their targets. Finally, we detected the response of miRNAs to aluminium (Al) toxicity in these three root types (PR, SR and CR). Our studies will not only accelerate our understanding of the regulatory roles of miRNAs in maize root development but also provide deeper insight into the molecular mechanisms of different root types in response to environmental stimuli.
Results Deep sequencing and data analysis To identify the role of miRNAs in the regulation of maize root development and architecture, we constructed four maize small RNA libraries from PRY, PRO, SR, and CR. For all four small RNA libraries, over ten million reads were generated through sequencing. Solexa raw data were available at Gene Expression Omnibus [GEO: GSE53865]. After removing sequences of low quality, adaptor contamination and RNAs smaller than 18 nucleotides, 10721082 highquality clean reads, 18–30 nucleotides in length, representing 3073394 unique sRNAs was obtained from PRY, 11057522 quality reads representing 3407520 unique sRNAs was obtained from PRO, 11572637 quality reads representing 3494355 unique sRNAs was obtained from SR, and 10920069 quality reads representing 3714028 unique sRNAs was obtained from CR (Table S1). These high-quality sRNAs were used for further analyses. It was also shown that 8765985 (81.76%), 8259650 (74.7%), 8915412 (77.04%) and 8384414 (76.78%) of the clean reads in PRY, PRO, SR and CR, respectively, were perfectly mapped to the maize genome (B73 RefGen_v2, release 5b.60) using SOAP. In addition, sequences corresponding to rRNAs, tRNAs, snoRNAs and snRNAs were identified. We also identified small RNAs positioned at repeat loci using Tag2Repeat software and annotated them as repeat-associated sRNAs. Known miRNAs accounted for 1165785 (10.87%) sequences representing 588 (0.02%) unique sRNAs, 1053854 (9.53%) representing 620 (0.02%) unique sRNAs, 856506 (7.4%) representing 587
ª 2014 Society for Experimental Biology, Association of Applied Biologists and John Wiley & Sons Ltd, Plant Biotechnology Journal, 1–14
miRNAs in maize root development 3 (0.02%) unique sRNAs and 767414 (7.03%) representing 590 (0.02%) unique sRNAs in PRY, PRO, SR and CR, respectively (Table S1). The well-represented unannotated sRNAs were further used to predict novel miRNAs. In maize, the size of the sRNAs was not evenly distributed. As shown in Figure 1, 24-nucleotide small interfering RNAs (siRNAs) were the predominant size class in maize roots, followed by 21and 22-nucleotide sRNAs. This observation indicated that siRNAs, which are mostly 24 nucleotides long, are relatively more prevalent in roots, which is consistent with previous reports on maize and other plants such as Arabidopsis, peanut, cotton, rice and Medicago truncatula (Liu et al., 2012a,b; Wei et al., 2010; Xue et al., 2013).
(a)
Identification of conserved miRNAs in maize Previous studies indicate that several miRNA families are evolutionarily conserved across many plant species and play important roles in plant development. To identify conserved miRNAs in our dataset, all sRNA sequences were Blastn searched against the known maize mature miRNAs and their precursors in the miRNA database miRBase. One hundred and thirty of 172 known maize miRNA precursors were detected in PRY sample, 141 in PRO sample, 134 in the SR sample and 138 in the CR sample. In total, we identified 246 mature miRNAs, belonging to 28 miRNA families in four libraries (Table S2). One hundred and ninety-four of these known mature miRNAs were expressed in the PRY sample, 217 in the PRO sample, 202 in the SR sample and 213 in the CR sample (Figure 2a). One known miRNA family, miR482, was not successfully detected in our data sets. Maize miRNA families displayed significantly varied expression levels. zmamiR156, which consists of 12 members, was expressed with the highest abundance, followed by zma-miR166 and zma-miR168, which possess 14 and 2 members, respectively.
(b)
(c)
Prediction of novel maize miRNAs One of the most important features of high-throughput sequencing is that it helps us to explore novel miRNAs in the sRNA transcriptome. We used Mireap software to predict novel miRNAs by exploring the secondary structure, the Dicer cleavage site and the minimum free energy of the unannotated sRNAs that could be mapped to the maize genome and could be subjected to a rigorous secondary structural analysis of their precursors. In total,
(d)
Figure 1 Length distribution of small RNAs identified from PRY, PRO, SR and CR libraries.
Figure 2 Venn diagram of differentially expressed conserved miRNAs (a, c) and novel miRNAs (b, d) in different root types.
ª 2014 Society for Experimental Biology, Association of Applied Biologists and John Wiley & Sons Ltd, Plant Biotechnology Journal, 1–14
4 Xiangpei Kong et al. 32 novel miRNAs were identified based on the rules (Tables S3 and S4). Compared with novel maize miRNAs identified in the two recent sequencing in maize root (Liu et al., 2012a; Zhai et al., 2013), 26 of 32 potentially novel miRNAs identified in the present study were first reported in maize root development (Table S5). The length of the newly identified miRNAs ranged from 21 bp to 23 bp, and the negative folding free energies of these pre-miRNA hairpin structures ranged from 165.1 to 21.9 (kcal/mol) with an average of 70.13 kcal/mol, which is similar to the results from recent studies in maize and other plants. In addition, our newly identified plant pre-miRNAs had a higher MFEI (minimal folding free energy (MFE) index) (0.60–1.71) (Table S3), with an average of 1.02, than that for tRNAs, rRNAs and mRNAs (Zhang et al., 2006a). Of the predicted novel miRNAs, 20 were expressed in PRY, 18 in PRO, 26 in SR and 19 in CR (Figure 2b).
Expression profiles of miRNAs in different maize root types To elucidate the potential regulatory roles of miRNAs in maize root development, we analysed their differential expression profiles. As shown in Figure 2a,b, 171 known and 14 novel miRNAs were expressed in all four libraries, whereas six known and two novel miRNAs were specifically expressed in PRY, 10 known and three novel miRNAs in PRO, one known and six novel miRNAs in SR and 15 known and one novel miRNAs in CR. Next, we analysed the differential expression profiles of miRNAs in four libraries. As shown in Figures 2c,d and 3, 75 conserved and eight novel miRNAs exhibited differential expression between PRY and PRO. A total of 51 conserved and two novel miRNAs, including zmamiR159c/d, zma-miR167, zma-miR172, zma-miR395, zmamiR528, zma-miRN111 and zma-miRN69, were up-regulated in PRO compared with PRY, whereas 24 conserved and six novel miRNAs, including zma-miR156, zma-miR529, zma-miRN117, zma-miRN24 and zma-miRN110, were reduced in PRO. Next, we compared the normalized expression levels of miRNAs from PRY and SR; 68 conserved and nine novel miRNAs were differentially expressed in PRY and SR, of which 22 conserved and five novel miRNAs, including zma-miR172, zma-miR528, zma-miR167a/b/c/ d, zma-miRN111 and zma-miRN173, showed higher expression levels in SR than in PRY (Figures 2c,d and 3). To identify the roles of miRNAs and their involvement in embryonic primary roots and postembryonic crown roots, we compared miRNA expression profiles between PRY and CR. A total of 74 conserved and 8 novel miRNAs showed a differential expression pattern. Among them, 42 conserved and 3 novel miRNAs, which includes zma-miR167a/b/c/d, zma-miR164a/b/c/d/g, zma-miR395, zmamiRN111, zma-miRN148 and zma-miRN235, were highly abundant in CR. The other 32 conserved and 5 novel miRNAs, including zma-miR156, zma-miR528, zma-miR169o, zmamiR160, zma-miRN110, zma-miRN117 and zma-miRN24, were highly expressed in PRY. In conclusion, zma-miR167, zmamiR172 and zma-miRN111 were more abundant in PRO, SR and CR than in PRY, suggesting that these miRNAs might mainly function in the later root developmental stage in maize. In contrast, the levels of zma-miR156, zma-miRN43, zma-miRN110 and zma-miRN24 were reduced in PRO, SR and CR, implying that these miRNAs might mainly function in the early root developmental stage in maize. Furthermore, the expression of zma-miR164, zma-miR395 and zma-miRN235 was up-regulated in CR but not in SR, suggesting that these miRNAs may play
crucial roles in crown root development. zma-miR528 was found to be up-regulated in SR but down-regulated in CR, indicating that zma-miR528 may be a key factor in the development of embryonic and postembryonic roots. To confirm the expression patterns of the differentially expressed miRNAs using Solexa deep sequencing, we randomly selected 11 miRNAs and performed stem-loop quantitative realtime PCR (qRT-PCR) (Figure 4). Most of our qRT-PCR results were consistent with the sequencing data. For example, the expression of zma-miR395 was up-regulated in PRO and CR as compared with PRY.
Target prediction for maize miRNAs and function analysis Previous studies have indicated that miRNAs inhibit gene expression by binding to protein-coding mRNAs. To characterize the functions of the differentially expressed miRNAs in maize roots, we first predicted their targets (Tables S6 and S7). For miRNAs differentially expressed between PRO and PRY, 335 target genes predicted for 78 miRNAs (70 conserved and eight novel miRNAs) and 289 target genes predicted for 72 miRNAs (63 conserved and nine novel miRNAs) were differentially expressed in SR compared with PRY, whereas 245 target genes were predicted among 77 miRNAs (69 conserved and 8 novel miRNAs) in CR compared with PRY. To gain a better understanding of the roles of these miRNA targets in maize roots, a gene ontology (GO) analysis was performed using the GO database, which organizes information based on cellular component, molecular function and biological process categories (Figure S1). In the case of the target genes of miRNAs differentially expressed between PRO, SR, CR and PRY, there was an obvious enrichment of C-5 methylation of cytosine (GRMZM2G092497, one target of zma-miR156b-3p) in the biological process category, while ATPase activity was enriched in the molecular function category for the target genes of conserved miRNAs. Finally, nuclear heterochromatin, SWI/SNF complex and CCAAT-binding factor were clearly overrepresented in the cellular component category for the target genes of conserved miRNAs differentially expressed between PRO, SR, CR and PRY, respectively. Taken together, these results indicate that these miRNAs targets may play important roles in maize root development at the transcriptional level. We next performed KEGG pathway analyses on the miRNA targets that might be involved in maize root development. Our analyses revealed prominent involvement of the RNA transport and mRNA surveillance pathway, followed by the metabolic pathways and plant hormone signal transduction pathways (Tables S8). Previous studies have shown that plant hormones, such as auxin, brassinosteroid (BR) and jasmonic acid (JA), play an important role in root development (Petricka et al., 2012). Thus, our data further support the notion that nutrient metabolism and plant hormones play essential roles in maize root development. We also analysed the expression of 22 predicted target genes of 8 selected miRNAs using qRT-PCR. As shown in Figure 5, the expression of most of the miRNAs, such as zma-miR167 and zmamiRN235, was inversely correlated with the expression of their targets. However, some target gene expression was positively related to their corresponding miRNAs. It is possible that these genes were regulated by their corresponding miRNAs through translation inhibition. Furthermore, the expression of most novel miRNAs was negatively associated with their targets.
ª 2014 Society for Experimental Biology, Association of Applied Biologists and John Wiley & Sons Ltd, Plant Biotechnology Journal, 1–14
miRNAs in maize root development 5
Figure 3 Clustering of miRNAs differentially expressed conserved miRNAs in different root types. Red indicates that the miRNA has a higher expression level in the treatment samples; green indicates that the miRNA has a lower expression in the treatment samples; and grey indicates that the miRNA has no expression in at least one sample.
miRNAs are differentially expressed in maize roots in response to Al stress Our deep sequencing data showed that miRNAs play essential roles in maize root development and architecture, and the expression profiles of miRNAs showed divergent dynamics in different maize root types. These results prompted us to test whether the environmental response were different among these three root types (PR, SR and CR).
Aluminium (Al) toxicity severely limits crop production in the acidic soil (pH
doi: 10.1111/pbi.12218
System analysis of microRNAs in the development and aluminium stress responses of the maize root system Xiangpei Kong1, Maolin Zhang1, Xiangbo Xu2, Xiaoming Li3, Cuiling Li1 and Zhaojun Ding1,* 1
The Key Laboratory of Plant Cell Engineering and Germplasm Innovation, College of Life Sciences, Shandong University, Jinan, China
2
Maize Institute, Shandong Academy of Agricultural Sciences/National Maize Improvement Sub-Center/National Engineering Laboratory for Wheat and Maize, Jinan,
China 3
Shandong Provincial Key Laboratory of Agricultural Microbiology, College of Plant Protection, Shandong Agriculture University, Tai’an, China
Received 2 March 2014; revised 21 May 2014; accepted 23 May 2014. *Correspondence (Tel +86 531 8836 2351; email [email protected])
Keywords: Al stress, maize, miRNAs, root development.
Summary MicroRNAs (miRNAs) are a class of regulatory small RNAs (sRNAs) that down-regulate target genes through mRNA cleavage or translational inhibition. miRNA is known to play an important role in the root development and environmental responses in both the Arabidopsis and rice. However, little information is available to form a complete view of miRNAs in the development of the maize root system and Al stress responses in maize. Four sRNA libraries were generated and sequenced from the early developmental stage of primary roots (PRY), the later developmental stage of maize primary roots (PRO), seminal roots (SR) and crown roots (CR). Through integrative analysis, we identified 278 miRNAs (246 conserved and 32 novel ones) and found that the expression patterns of miRNAs differed dramatically in different maize roots. The potential targets of the identified conserved and novel miRNAs were also predicted. In addition, our data showed that CR is more resistant to Al stress compared with PR and SR, and the differentially expressed miRNAs are likely to play significant roles in different roots in response to environmental stress such as Al stress. Here, we demonstrate that the expression patterns of miRNAs are highly diversified in different maize roots. The differentially expressed miRNAs are correlated with both the development and environmental responses in the maize root. This study not only improves our knowledge about the roles of miRNAs in maize root development but also reveals the potential role of miRNAs in the environmental responses of different maize roots.
Introduction MicroRNAs (miRNAs) are a class of endogenous, 20–24 nucleotides, noncoding small RNAs (sRNAs), which negatively regulate target gene expression at the post-transcriptional levels through mRNA cleavage or translation inhibition (Bartel, 2004; Carrington and Ambros, 2003; He and Hannon, 2004; Li et al., 2013b). miRNAs are encoded by MIRNA genes that are transcribed into the primary miRNAs (pri-miRNAs) by RNA polymerase II, which, in turn, are cleaved twice by the RNase III enzyme Dicer-like 1 (DCL1) to yield mature miRNAs (Bartel, 2004; Jones-Rhoades et al., 2006; Mallory and Bouche, 2008). Accumulating evidence indicates that miRNAs play crucial regulatory roles not only in response to a broad variety of abiotic (Covarrubias and Reyes, 2010; Zhang et al., 2009a; Zhao et al., 2007; Zhou et al., 2013b) and biotic stresses (Eckardt, 2012; Ruiz-Ferrer and Voinnet, 2009) but also in response to nutrient stresses such as nitrate (Chiou, 2007; Vidal et al., 2010; Zhao et al., 2011), phosphate (Kuo and Chiou, 2011) and sulphate (Fujii et al., 2005; Liang et al., 2010) in plants. Over the past few years, knowledge about the roles of miRNAs in regulating root growth and development in response to developmental signals or different environmental cues has increased dramatically (Bian
et al., 2013; Khan et al., 2011; Kidner and Martienssen, 2005; Marin et al., 2010; Rubio-Somoza and Weigel, 2011). Recent studies showed that miR160-ARF10/ARF16/ARF17 module plays important roles in root cap formation, lateral root development and primary root growth in both Arabidopsis and rice (Gutierrez et al., 2009; Mallory et al., 2005; Meng et al., 2009; Wang et al., 2005). Interestingly, miR167–ARF6/ARF8 and miR160–ARF17 have opposite roles in the adventitious roots development through their control of the auxin homoeostatic enzyme GH3.6 in Arabidopsis (Gutierrez et al., 2009; Tian et al., 2004). In contrast to A. thaliana, miR167 positively regulate adventitious root development in rice (Meng et al., 2009). The miR164–NAC1 module affects lateral root initiation (Guo et al., 2005; Li et al., 2012; Meng et al., 2009), whereas miR390–TAS3–ARFs together with miR167–ARF8 and miR393–AFB3 are involved in the emergence and elongation of the lateral root (Marin et al., 2010; Meng et al., 2009; Vidal et al., 2010; Williams et al., 2005; Yang et al., 2006). Together, all these investigations demonstrate that certain miRNA plays crucial regulatory roles in root development. Maize (Zea mays) is one of the oldest and most important worldwide crops. It provides one of the most important human food supply and animal feed, and it is also crucial raw material
Please cite this article as: Kong, X., Zhang, M., Xu, X., Li, X., Li, C. and Ding, Z. (2014) System analysis of microRNAs in the development and aluminium stress responses of the maize root system. Plant Biotechnol. J., doi: 10.1111/pbi.12218
ª 2014 Society for Experimental Biology, Association of Applied Biologists and John Wiley & Sons Ltd
1
2 Xiangpei Kong et al. plant for energy production (Kong et al., 2011). The root system of maize displays a complex root structure comprising several root types, which include embryonic primary and seminal roots, and postembryonic shoot-borne and lateral roots (Hochholdinger et al., 2004; Orman-Ligeza et al., 2013). Embryonic primary roots can be observed only 2–3 days after germination. Seminal roots which emerge from the scutellar node become visible approximately 7 days after germination. Shoot-borne roots are initiated from underground and aboveground nodes of the stem and are called crown and brace roots, respectively. Crown roots start to emerge approximately 10 days postgermination, whereas brace roots become visible much later, approximately 6 weeks after germination. Lateral roots which initiate in the pericycle of all above major roots also belong to the postembryonic root system. The embryonic root system in maize plays important roles during the early stages of plant development, whereas the postembryonic shoot-borne root system makes up the major backbone of the maize root stock during the later developmental stage (Orman-Ligeza et al., 2013). However, how the different maize roots are developmental and spatially regulated is not well understood. Genetic studies have identified and characterized 9 specific root mutants in maize. Jenkins (1930) reported the first maize mutant rt1, which reduces shoot-borne root formation. Another maize shoot-borne root mutant rtcs lacks the initiation of the seminal and the shoot-borne roots (Taramino et al., 2007). More recently, Majer et al. (2012) reported that RTCS encodes a LATERAL ORGAN BOUNDARIES domain (LBD) protein and ZmARF34 binds the promoter of RTCS to activate downstream gene expression, which is involved in the initiation and maintenance of seminal and shoot-borne roots. The lateral rootless1 (lrt1) mutant affects the initiation of lateral roots (Hochholdinger and Feix, 1998), while the slr1 and slr2 mutants show reduced lateral root elongation only in the embryonic roots (Hochholdinger et al., 2001). Recently, Woll et al. (2005) characterized another lateral root mutant rum1, which is deficient in the initiation of the seminal roots and the lateral roots in the primary root. Map-based cloning revealed that RUM1 encodes a truncated ZmIAA10 sequence that lacks the auxin degron sequence. ZmIAA10 can interact with ZmARF25 and ZmARF34, which are required for the initiation of seminal and lateral root initiation in primary roots (Woll et al., 2005). Three mutants (rth1, rth2 and rth3) have been identified with defects in root hair elongation (Wen and Schnable 1994). Currently, 321 mature maize miRNAs have been registered on miRBase (release 20) (Zhang et al., 2009b), and a number of miRNAs with specific function have been characterized in maize (Mica et al., 2006; Zhang et al., 2006b, 2009b). For example, zma-miR172 down-regulates an APETALA2-like gene, glossy15, to promote vegetative phase change in maize (Lauter et al., 2005). More recently, genomewide studies using high-throughput approaches have identified some new miRNAs and explored the roles of miRNAs in response to diverse environmental cues in maize, including short-term and long-term waterlogging stress (Liu et al., 2012b; Zhai et al., 2013), salt (Ding et al., 2009), low nitrate (Trevisan et al., 2012; Xu et al., 2011; Zhao et al., 2013) and low phosphorus (Pei et al., 2013; Zhang et al., 2012). Furthermore, the roles of miRNAs were also investigated in different tissues and organs, such as seeds (Ding et al., 2012; Kang et al., 2012; Li et al., 2013a; Wang et al., 2011), ears (Ding et al., 2014; Liu et al., 2014), endosperm (Gu et al., 2013), pollen (Li et al., 2013c) and brace root (Liu et al., 2012a). Although
miRNAs play crucial regulatory roles in root development, little information, to our knowledge, is available to form a complete view of miRNAs in the development of the maize root system. Furthermore, the mechanisms underlying the regulation of miRNAs in different maize root types remain elusive. And several molecular mechanisms have been recently reported in Arabidopsis about how primary and lateral roots have differential growth responses in response to environmental cues such as salt, gravity and plant nutrition availability (Duan et al., 2013; Rosquete et al., 2013; Tian et al., 2014; Vidal et al., 2013), which imply the differential regulation mechanisms between primary and lateral roots. These investigations also suggest the importance of the comparative investigations to our fully understand the molecular mechanism of root growth and development. Therefore, to better understand the development of the maize root system that has a typical fibrous root system, it is necessary for us to compare the roles of miRNAs in different maize root types. In this study, we generated and sequenced four maize small RNA libraries from the early developmental stage of primary roots (PRY), the later developmental stage of maize primary roots (PRO), the early developmental stage of seminal roots (SR), and the early developmental stage of crown roots (CR) using Solexa high-throughput sequencing technology and obtained a comprehensive view of the variation in miRNAs’ expression patterns in the four samples. Subsequently, we also investigated the correlation between the expression profiles of miRNAs and their targets. Finally, we detected the response of miRNAs to aluminium (Al) toxicity in these three root types (PR, SR and CR). Our studies will not only accelerate our understanding of the regulatory roles of miRNAs in maize root development but also provide deeper insight into the molecular mechanisms of different root types in response to environmental stimuli.
Results Deep sequencing and data analysis To identify the role of miRNAs in the regulation of maize root development and architecture, we constructed four maize small RNA libraries from PRY, PRO, SR, and CR. For all four small RNA libraries, over ten million reads were generated through sequencing. Solexa raw data were available at Gene Expression Omnibus [GEO: GSE53865]. After removing sequences of low quality, adaptor contamination and RNAs smaller than 18 nucleotides, 10721082 highquality clean reads, 18–30 nucleotides in length, representing 3073394 unique sRNAs was obtained from PRY, 11057522 quality reads representing 3407520 unique sRNAs was obtained from PRO, 11572637 quality reads representing 3494355 unique sRNAs was obtained from SR, and 10920069 quality reads representing 3714028 unique sRNAs was obtained from CR (Table S1). These high-quality sRNAs were used for further analyses. It was also shown that 8765985 (81.76%), 8259650 (74.7%), 8915412 (77.04%) and 8384414 (76.78%) of the clean reads in PRY, PRO, SR and CR, respectively, were perfectly mapped to the maize genome (B73 RefGen_v2, release 5b.60) using SOAP. In addition, sequences corresponding to rRNAs, tRNAs, snoRNAs and snRNAs were identified. We also identified small RNAs positioned at repeat loci using Tag2Repeat software and annotated them as repeat-associated sRNAs. Known miRNAs accounted for 1165785 (10.87%) sequences representing 588 (0.02%) unique sRNAs, 1053854 (9.53%) representing 620 (0.02%) unique sRNAs, 856506 (7.4%) representing 587
ª 2014 Society for Experimental Biology, Association of Applied Biologists and John Wiley & Sons Ltd, Plant Biotechnology Journal, 1–14
miRNAs in maize root development 3 (0.02%) unique sRNAs and 767414 (7.03%) representing 590 (0.02%) unique sRNAs in PRY, PRO, SR and CR, respectively (Table S1). The well-represented unannotated sRNAs were further used to predict novel miRNAs. In maize, the size of the sRNAs was not evenly distributed. As shown in Figure 1, 24-nucleotide small interfering RNAs (siRNAs) were the predominant size class in maize roots, followed by 21and 22-nucleotide sRNAs. This observation indicated that siRNAs, which are mostly 24 nucleotides long, are relatively more prevalent in roots, which is consistent with previous reports on maize and other plants such as Arabidopsis, peanut, cotton, rice and Medicago truncatula (Liu et al., 2012a,b; Wei et al., 2010; Xue et al., 2013).
(a)
Identification of conserved miRNAs in maize Previous studies indicate that several miRNA families are evolutionarily conserved across many plant species and play important roles in plant development. To identify conserved miRNAs in our dataset, all sRNA sequences were Blastn searched against the known maize mature miRNAs and their precursors in the miRNA database miRBase. One hundred and thirty of 172 known maize miRNA precursors were detected in PRY sample, 141 in PRO sample, 134 in the SR sample and 138 in the CR sample. In total, we identified 246 mature miRNAs, belonging to 28 miRNA families in four libraries (Table S2). One hundred and ninety-four of these known mature miRNAs were expressed in the PRY sample, 217 in the PRO sample, 202 in the SR sample and 213 in the CR sample (Figure 2a). One known miRNA family, miR482, was not successfully detected in our data sets. Maize miRNA families displayed significantly varied expression levels. zmamiR156, which consists of 12 members, was expressed with the highest abundance, followed by zma-miR166 and zma-miR168, which possess 14 and 2 members, respectively.
(b)
(c)
Prediction of novel maize miRNAs One of the most important features of high-throughput sequencing is that it helps us to explore novel miRNAs in the sRNA transcriptome. We used Mireap software to predict novel miRNAs by exploring the secondary structure, the Dicer cleavage site and the minimum free energy of the unannotated sRNAs that could be mapped to the maize genome and could be subjected to a rigorous secondary structural analysis of their precursors. In total,
(d)
Figure 1 Length distribution of small RNAs identified from PRY, PRO, SR and CR libraries.
Figure 2 Venn diagram of differentially expressed conserved miRNAs (a, c) and novel miRNAs (b, d) in different root types.
ª 2014 Society for Experimental Biology, Association of Applied Biologists and John Wiley & Sons Ltd, Plant Biotechnology Journal, 1–14
4 Xiangpei Kong et al. 32 novel miRNAs were identified based on the rules (Tables S3 and S4). Compared with novel maize miRNAs identified in the two recent sequencing in maize root (Liu et al., 2012a; Zhai et al., 2013), 26 of 32 potentially novel miRNAs identified in the present study were first reported in maize root development (Table S5). The length of the newly identified miRNAs ranged from 21 bp to 23 bp, and the negative folding free energies of these pre-miRNA hairpin structures ranged from 165.1 to 21.9 (kcal/mol) with an average of 70.13 kcal/mol, which is similar to the results from recent studies in maize and other plants. In addition, our newly identified plant pre-miRNAs had a higher MFEI (minimal folding free energy (MFE) index) (0.60–1.71) (Table S3), with an average of 1.02, than that for tRNAs, rRNAs and mRNAs (Zhang et al., 2006a). Of the predicted novel miRNAs, 20 were expressed in PRY, 18 in PRO, 26 in SR and 19 in CR (Figure 2b).
Expression profiles of miRNAs in different maize root types To elucidate the potential regulatory roles of miRNAs in maize root development, we analysed their differential expression profiles. As shown in Figure 2a,b, 171 known and 14 novel miRNAs were expressed in all four libraries, whereas six known and two novel miRNAs were specifically expressed in PRY, 10 known and three novel miRNAs in PRO, one known and six novel miRNAs in SR and 15 known and one novel miRNAs in CR. Next, we analysed the differential expression profiles of miRNAs in four libraries. As shown in Figures 2c,d and 3, 75 conserved and eight novel miRNAs exhibited differential expression between PRY and PRO. A total of 51 conserved and two novel miRNAs, including zmamiR159c/d, zma-miR167, zma-miR172, zma-miR395, zmamiR528, zma-miRN111 and zma-miRN69, were up-regulated in PRO compared with PRY, whereas 24 conserved and six novel miRNAs, including zma-miR156, zma-miR529, zma-miRN117, zma-miRN24 and zma-miRN110, were reduced in PRO. Next, we compared the normalized expression levels of miRNAs from PRY and SR; 68 conserved and nine novel miRNAs were differentially expressed in PRY and SR, of which 22 conserved and five novel miRNAs, including zma-miR172, zma-miR528, zma-miR167a/b/c/ d, zma-miRN111 and zma-miRN173, showed higher expression levels in SR than in PRY (Figures 2c,d and 3). To identify the roles of miRNAs and their involvement in embryonic primary roots and postembryonic crown roots, we compared miRNA expression profiles between PRY and CR. A total of 74 conserved and 8 novel miRNAs showed a differential expression pattern. Among them, 42 conserved and 3 novel miRNAs, which includes zma-miR167a/b/c/d, zma-miR164a/b/c/d/g, zma-miR395, zmamiRN111, zma-miRN148 and zma-miRN235, were highly abundant in CR. The other 32 conserved and 5 novel miRNAs, including zma-miR156, zma-miR528, zma-miR169o, zmamiR160, zma-miRN110, zma-miRN117 and zma-miRN24, were highly expressed in PRY. In conclusion, zma-miR167, zmamiR172 and zma-miRN111 were more abundant in PRO, SR and CR than in PRY, suggesting that these miRNAs might mainly function in the later root developmental stage in maize. In contrast, the levels of zma-miR156, zma-miRN43, zma-miRN110 and zma-miRN24 were reduced in PRO, SR and CR, implying that these miRNAs might mainly function in the early root developmental stage in maize. Furthermore, the expression of zma-miR164, zma-miR395 and zma-miRN235 was up-regulated in CR but not in SR, suggesting that these miRNAs may play
crucial roles in crown root development. zma-miR528 was found to be up-regulated in SR but down-regulated in CR, indicating that zma-miR528 may be a key factor in the development of embryonic and postembryonic roots. To confirm the expression patterns of the differentially expressed miRNAs using Solexa deep sequencing, we randomly selected 11 miRNAs and performed stem-loop quantitative realtime PCR (qRT-PCR) (Figure 4). Most of our qRT-PCR results were consistent with the sequencing data. For example, the expression of zma-miR395 was up-regulated in PRO and CR as compared with PRY.
Target prediction for maize miRNAs and function analysis Previous studies have indicated that miRNAs inhibit gene expression by binding to protein-coding mRNAs. To characterize the functions of the differentially expressed miRNAs in maize roots, we first predicted their targets (Tables S6 and S7). For miRNAs differentially expressed between PRO and PRY, 335 target genes predicted for 78 miRNAs (70 conserved and eight novel miRNAs) and 289 target genes predicted for 72 miRNAs (63 conserved and nine novel miRNAs) were differentially expressed in SR compared with PRY, whereas 245 target genes were predicted among 77 miRNAs (69 conserved and 8 novel miRNAs) in CR compared with PRY. To gain a better understanding of the roles of these miRNA targets in maize roots, a gene ontology (GO) analysis was performed using the GO database, which organizes information based on cellular component, molecular function and biological process categories (Figure S1). In the case of the target genes of miRNAs differentially expressed between PRO, SR, CR and PRY, there was an obvious enrichment of C-5 methylation of cytosine (GRMZM2G092497, one target of zma-miR156b-3p) in the biological process category, while ATPase activity was enriched in the molecular function category for the target genes of conserved miRNAs. Finally, nuclear heterochromatin, SWI/SNF complex and CCAAT-binding factor were clearly overrepresented in the cellular component category for the target genes of conserved miRNAs differentially expressed between PRO, SR, CR and PRY, respectively. Taken together, these results indicate that these miRNAs targets may play important roles in maize root development at the transcriptional level. We next performed KEGG pathway analyses on the miRNA targets that might be involved in maize root development. Our analyses revealed prominent involvement of the RNA transport and mRNA surveillance pathway, followed by the metabolic pathways and plant hormone signal transduction pathways (Tables S8). Previous studies have shown that plant hormones, such as auxin, brassinosteroid (BR) and jasmonic acid (JA), play an important role in root development (Petricka et al., 2012). Thus, our data further support the notion that nutrient metabolism and plant hormones play essential roles in maize root development. We also analysed the expression of 22 predicted target genes of 8 selected miRNAs using qRT-PCR. As shown in Figure 5, the expression of most of the miRNAs, such as zma-miR167 and zmamiRN235, was inversely correlated with the expression of their targets. However, some target gene expression was positively related to their corresponding miRNAs. It is possible that these genes were regulated by their corresponding miRNAs through translation inhibition. Furthermore, the expression of most novel miRNAs was negatively associated with their targets.
ª 2014 Society for Experimental Biology, Association of Applied Biologists and John Wiley & Sons Ltd, Plant Biotechnology Journal, 1–14
miRNAs in maize root development 5
Figure 3 Clustering of miRNAs differentially expressed conserved miRNAs in different root types. Red indicates that the miRNA has a higher expression level in the treatment samples; green indicates that the miRNA has a lower expression in the treatment samples; and grey indicates that the miRNA has no expression in at least one sample.
miRNAs are differentially expressed in maize roots in response to Al stress Our deep sequencing data showed that miRNAs play essential roles in maize root development and architecture, and the expression profiles of miRNAs showed divergent dynamics in different maize root types. These results prompted us to test whether the environmental response were different among these three root types (PR, SR and CR).
Aluminium (Al) toxicity severely limits crop production in the acidic soil (pH
