Introduction

Gene duplication and alternative splicing (AS) are two major evolutionary mechanisms that can increase the functional diversification of genes (Jin et al. 2008; Yuan et al. 2009). Gene duplications, often followed by some sort of divergence, contribute to the establishment of new gene functions. Gene duplications are thought to function in the diversification of gene families (Freeling 2009). AS is another mechanism that contributes to gene diversity by generating different transcripts from a single mRNA by differential inclusion or exclusion of regions of the precursor mRNA (Reddy 2007). Results of recent studies indicate that AS is prevalent in plants, impacting at least 30 % of all expressed protein-coding genes (Chen et al. 2007; Wang and Brendel 2006).

Mitogen-activated protein kinase (MAPK) cascades are universal signaling modules in eukaryotes, including yeasts, animals, and plants (Heinrich et al. 2012; Liu et al. 2011; MAPK Group 2002; Samajova et al. 2011). A typical MAPK cascade consists of three protein kinases, MAPK kinase kinase (MAPKKK), MAPK kinase (MAPKK), and MAPK. MAPKKK, MAPKK, and MAPK sequentially phosphorylate the corresponding downstream substrates (MAPK Group 2002). MAPK cascades have been involved in plant growth and development, biotic and abiotic stresses (Samajova et al. 2011). The versatile roles of MAPK cascades can be explained, in part, by the fact that there are multiple genes in each of the three tiers of kinases, e.g., in Arabidopsis, there are 20 MAPKs, 10 MAPKs, and about 60 MAPKKKs (MAPK Group 2002). Other plant species have similar numbers. Gene family analysis has revealed that segmental duplication is expected to contribute to the expansion of the MAPK-cascade gene family (Koo et al. 2007; Lin et al. 2010).

The generation of a high number of different isoforms by AS in MAPK-cascade genes also plays an important role in regulating the function of MAPK cascades. The AS of MAPK-cascade genes may influence kinase activity, protein stability, subcellular localization, and binding properties with substrates and effectors. AS of MAPK-cascade genes is a feature common to both mammals and plants. AS of human MAP kinases JNK1, JNK2, and JNK3 yields products containing different carboxy-termini, and thus proteins exhibiting different roles (Gupta et al. 1996). ERK1 and p38 also undergo AS to generate different isoforms (Enslen et al. 1998; Yung et al. 2001). In Arabidopsis, two isoforms of cANP1 have been identified. The longer splice variant (cANP1L) has an intron-like sequence for the kinase-unrelated region. Protein encoded by cANP1L is functional in the mating pheromone-responsive signal pathway of yeast. However, retention of the intron-like sequence in cANP1L impairs protein activity when compared with the shorter protein in yeast (Nishihama et al. 1997). Another Arabidopsis gene, AtMPK13, has also been found to generate at least three splice variants, one with complete splicing of five introns, one with the fourth intron retained, and one with the fifth intron retained. Intron retention of AtMPK13 generates proteins with no kinase activity (Lin et al. 2010). In rice, OsMAPK5 has been demonstrated to generate at least two differentially spliced transcripts. OsMAPK5b is smaller than OsMAPK5a as a result of a 104-amino-acid deletion, leading to OsMAPK5b with no kinase activity (Xiong and Yang 2003). OsBWMK1, another MAPK of rice, generates three splice variants with different expression and subcellular localization (Koo et al. 2007). The different expression of OsBWMK1 splice variants is due to the result of alternative promoter usage (Koo et al. 2009). MIK is a maize gene coding for a GCK-like MAP4K (Llompart et al. 2003). MIK generates at least four different mature mRNAs that accumulate with particular expression during maize development (Castells et al. 2006). Analyses of the maize genome have revealed that, although the average number of introns per gene is about the same in corn and Arabidopsis, maize genes have more alternatively spliced isoforms (Alexandrov et al. 2009). Here, we showed that a previously uncharacterized ZmMPK4 (ZmMPK3-1b in this research) is one such splicing variant. ZmMPK3-1 and ZmMPK3-2 are segmentally duplicated genes in the maize genome, and the expression profiles of ZmMPK3-2 and ZmMPK3-1 splice variants differ.

Materials and Methods

Plant Materials, Growth Conditions, and Agrobacterium-Mediated Transient Gene Expression

Maize cultivar Zhengdan958 was used in this research. Maize seedlings (14-days-old) were cultivated and treated as described previously (Gu et al. 2010; Liu et al. 2012). Briefly, maize seeds were washed several times with tap water and soaked in distilled water for germination. Seedlings were grown in Hoagland’s solution (pH 6.0) under greenhouse conditions at 22/26 °C (night/day) and a photoperiod of 14/10 h (day/night) for 2 weeks.

Extraction of RNA and DNA

RNA and DNA was extracted as previously described (Liu et al. 2012). Total RNA was extracted using Trizol (Invitrogen, Carlsbad, CA) method. Genomic DNA was isolated by the CTAB method.

Northern and Southern Blot Analyses

Total RNA (30 μg) and genomic DNA (30 μg) were used for northern and Southern blots, respectively. Hybridization analyses were performed as described (Liu et al. 2012). Probes were labeled with [a-32P]dCTP according to the manufacturer’s instructions (Primer-a-Gene Labeling System, Promega, Madison, WI). The primers used in the hybridization analyses are listed in Supplementary Table S1.

Dot Blot Assay

To confirm the specificity of probes used for northern blot analysis, we performed dot blot assay as described by Park et al. (2010). Plasmid DNA containing cDNAs of either the ZmMPK3-1a, ZmMPK3-1b or ZmMPK3-2 genes was blotted onto nylon membranes (Amersham, Pharmacia, Little Chalfont, UK). The membranes were hybridized with [a-32P]dCTP-labeled probes. Washes and detection were performed as above for northern blots.

Semi-Quantitative RT-PCR Analysis

Total RNA was extracted using Trizol (Invitrogen) from samples collected at various intervals, incubated with RNAse free DNAse I (Invitrogen) overnight, and reverse transcribed employing SuperScript II (Invitrogen) according to the manufacturer’s protocols. Zmactin was used as control for equal loading. Primers are listed in Supplementary Table S1.

Bioinformatic Analysis

The genomic context, and chromosomal location of ZmMPK3-1 and ZmMPK3-2 were performed using sequences from the maize genome sequencing project (http://www.maizegenome.org). Intron and exon organization of ZmMPK3-1 and ZmMPK3-2 were determined on http://www.phytozome.net/ (Goodstein et al. 2012). Sequence alignments of introns were displayed as a sequence logo with WebLogo (http://weblogo.berkeley.edu/logo.cgi) (Crooks et al. 2004).

Results and Discussion

The Previously Identified ZmMPK4 is One of Two Spliced Transcripts of ZmMPK3-1

Maize ZmMPK4 was a previously isolated cDNA with no characterized function (Berberich et al. 1999). The cDNA of ZmMPK4 was 1,647 base pairs (bp) long and encoded a protein of 406 amino acids. Remarkably, this gene encoded a protein with an unexpected insertion of 30 amino acids between subdomains IX and X (Berberich et al. 1999). This insertion has not been found in other plant MAPKs identified so far (Supplementary Fig. S1). In our attempt to test the expression of ZmMPK4, we detected two DNA bands in leaves and the predominantly amplified DNA products were about 500 bp, although the primers were expected to generate DNA products of 590 bp. This result was confirmed by DNA marker and DNA products from ZmMEK1 (589 bp) (Fig. 1a).

Fig. 1a–e
figure 1

Genomic analysis of ZmMPK3-1 and ZmMPK3-2 genes. a Expression analysis of ZmMEK1 and ZmMPK3-1. ZmMEK1 and ZmMPK3-1 were amplified from root (R), stem (S), or leaf (L) by 38 cycles of PCR. ZmMEK1 was used to mark the position of the 589 bp band. Arrow indicates the splicing variant of ZmMPK3-1b. b Southern blot analysis of ZmMPK3-1 and ZmMPK3-2. Genomic DNA (30 μg) from leaves was digested with EcoRI or HindIII, separated by electrophoresis on an agarose gel, and blotted onto nylon membranes. The common region or divergent region of ZmMPK3-1 and ZmMPK3-2 was used as a probe. c Exon–intron organization and chromosome localization of ZmMPK3-1 and ZmMPK3-2. The size of genomic DNA was reduced by the ratio indicated. Genomic DNA is presented as lines (introns) and boxes (exons). Start (ATG) codon, stop (TAG) codon and numbers of base pairs are marked. The retention of the third intron in ZmMPK3-1b is outlined. d Sequence logo and comparative analysis of the exon–intron junction of ZmMPK3-1 and ZmMPK3-2. Sequence alignments of the exon–intron junction (ten nucleotides) were displayed as a sequence logo with WebLogo (http://weblogo.berkeley.edu/logo.cgi) (Crooks et al. 2004). The height of each nucleotide represents the relative frequency of the nucleotide at that position. Boxes represent the position of the exon sequence, and the line indicates the intron sequence. e Sequence alignment of the third intron of ZmMPK3-1 from maize cultivar B73, Zhengdan 958, and Merit

Full size image

Upon cloning and sequencing, the DNA products from the larger band were found to be 590 bp and matched most of the ZmMPK4 sequence (90 bp insertion included). The DNA products from the smaller band were 498 bp, containing two highly similar sequences (93.37 %) and matched the parts of ZmMPK4 sequence with no 90 bp insertion. From our sequencing results, we presumed that the smaller band (498 bp) contains DNA products from two highly identical genes or nearly identical paralogs (NIPs) (Emrich et al. 2007). Before the availability of the maize genomic sequence (Wilson 2008), we conducted a first round Southern blot to test the copy number of ZmMPK4. As shown in Fig. 1b, we detected two bands in EcoRI- or HindIII-digested genomic DNA, indicating that ZmMPK4 was a two-copy gene. If the insertion was excluded, ZmMPK4 is the homolog of Arabidopsis AtMPK3 (Supplementary Fig. S2). To avoid confusion, and according to Arabidopsis MAPK nomenclature, we renamed ZmMPK4 as ZmMPK3-1. Without accounting for the 90 bp insertion, we found that the DNA products from the larger band (590 bp) fully match one product of the smaller band (498 bp). Therefore, we concluded that ZmMPK3-1 has at least two AS variants. We named the splicing variant without the 90 bp insertion ZmMPK3-1a, and the splicing variant with the 90 bp insertion ZmMPK3-1b.

ZmMPK3-1 and ZmMPK3-2 are Segmentally Duplicated Genes in the Maize Genome

The availability of the genomic sequence of maize B73 provides a convenient approach to analyze the genomic information of ZmMPK3-1. In silico analysis showed that ZmMPK3-1 is located on the long arm of chromosome 9 (Fig. 1c). In searching for a highly identical gene or NIP of ZmMPK3-1, we found another gene located on the short arm of chromosome 1 with 92.48 % identity to ZmMPK3-1 (Supplementary Fig. S3). Both these genes were homologs of Arabidopsis AtMPK3 and encoded proteins sharing 90.1 % identity (Supplementary Fig. S2, S4). The gene located on chromosome 1 was named ZmMPK3-2 (Fig. 1c). One product of the smaller band (498 bp) corresponded to just part of ZmMPK3-2. Therefore, the DNA products from the smaller band (498 bp in Fig. 1a) were a mixture of ZmMPK3-1 and ZmMPK3-2.

The genomic organizations of ZmMPK3-1 and ZmMPK3-2 were highly similar to each other. Both ZmMPK3-1 and ZmMPK3-2 consisted of five exons. The exon organization (sequence and length) was highly conserved. Furthermore, intron length at comparable positions was also highly similar. In silico analyses indicated that ZmMPK3-1 and ZmMPK3-2 were extremely closely related to each other and that the two genes have likely evolved as a result of a recent segmental gene duplication event. We conducted a second round of Southern blot analysis using a specific probe generated from the 3′ portion of ZmMPK3-2. The result showed that the genomic segment containing the ZmMPK3-2 gene contributes to the previously identified two-copy gene (Fig. 1b). During our analysis, Wang et al. (2010) published the a paper on ZmMPK3 (ZmMPK3-2 in our research). However, the latter authors mistakenly located ZmMPK3 (ZmMPK3-2) on chromosome 2. Furthermore, the primers used by Wang et al. match both ZmMPK3-1 and ZmMPK3-2. Therefore, the expression results from the research of Wang et al. can be recognized as a mixed analysis of ZmMPK3-1 and ZmMPK3-2.

Since ZmMPK3-1 can undergo AS, and since ZmMPK3-1 and ZmMPK3-2 are highly similar to each other, it is reasonable to ask whether ZmMPK3-2 can undergo AS. Analysis of the sequencing results revealed only the 590 bp product matching part of ZmMPK3-1 among the 100 clones sequenced. We further analyzed the introns of ZmMPK3-1 and ZmMPK3-2 and found that most of the introns (7 of 8) were the GU–AG type—the type found most typically in plants (Sheth et al. 2006). However, the third intron of ZmMPK3-1 was the GC–AG type (Fig. 1c, d), indicating that AS of ZmMPK3-1 may correlate with the type of third intron. We next examined the introns of ZmMPK3-1 and ZmMPK3-2 in three different inbred lines (B73, Zhengdan 958, and Merit). Most of the introns in the three inbred lines were conserved. Although all of the third introns of ZmMPK3-1 in the three inbred lines were of the GC–AG type, the sequence of the introns did not fully match with one another (Fig. 1e), indicating that further evolution had occurred in the third intron of ZmMPK3-1 after the divergence of the sweet corn (Merit) and the corn lines B73 and Zhengdan 958.

Our analyses support the conclusion that ZmMPK3-1, but not ZmMPK3-2, can retain the third intron for AS. Although there is no evidence, it is possible that AS of ZmMPK3-1 is correlated with the type of third intron. It seems that GC–AG type intron here could influence splice site recognition or other mechanisms, leading to the retention of the intron.

Organ-Specific Expression of ZmMPK3-1a, ZmMPK3-1b, and ZmMPK3-2

We next tested the expression profile of ZmMPK3-1a, ZmMPK3-1b, and ZmMPK3-2 in maize seedlings. Although the coding sequences of ZmMPK3-1 and ZmMPK3-2 shared high identity, the 3′-untranslated region (UTR) of the two genes shared only 46.3 % identity (Supplementary Fig. S5). Therefore, gene-specific probes were generated based on a portion of the 3′-UTR. A dot-blot assay was used to test the specificity of the probes and showed that the ZmMPK3-1 probe did not cross-hybridize with plasmid DNA containing cDNA of ZmMPK3-2. The ZmMPK3-2 probe did not cross-hybridize with plasmid DNA containing cDNA of ZmMPK3-1a or ZmMPK3-1b (Fig. 2a).

Fig. 2a–g
figure 2

Expression analysis of ZmMPK3-2, ZmMPK3-1a, and ZmMPK3-1b. a Dot blot assay of probes for ZmMPK3-1 and ZmMPK3-2. Specific probes were generated from the 3′ UTR portions of ZmMPK3-1 and ZmMPK3-2. The probe for ZmMPK3-1 recognized plasmid DNA of both ZmMPK3-1a and ZmMPK3-1b. b RNA gel blot analysis of ZmMPK3-1 and ZmMPK3-2 expression in maize root (R), stem (S), and leaf (L). c RT-PCR analysis of ZmMPK3-1a and ZmMPK3-1b expression in maize root (R), stem (S), and leaf (L). d Time course analysis of ZmMPK3-1 and ZmMPK3-2 expression in leaf under abscisic acid (ABA) (100 μM) treatment. e RT-PCR analysis of ZmMPK3-1a and ZmMPK3-1b expression in leaf under ABA (100 μM) treatment. f Time course analysis of ZmMPK3-1 and ZmMPK3-2 expression in leaf under NaCl (200 mM) treatment. g RT-PCR analysis of ZmMPK3-1a and ZmMPK3-1b expression in leaf under NaCl (200 mM) treatment. For RNA gel blot analysis, root (R), stem (S), and leaf (L, untreated or treated with ABA/NaCl) were collected from 14-day-old maize seedlings. Total RNA (30 μg) was loaded into each lane of a 1.5 % formaldehyde RNA gel and separated by electrophoresis. Ethidium bromide staining of gels confirmed equal loading of RNA samples. For RT-PCR analysis, total RNA was isolated from roots of 14-day-old maize seedlings. Zmactin was used as control for equal loading. The reactions were amplified for cycles as indicated. The experiments were repeated at least three times with similar results and representative results are shown

Full size image

Northern blot analysis showed that transcripts of ZmMPK3-2 accumulated predominantly in stems and leaves, whereas transcripts of ZmMPK3-1 accumulated mainly in roots (Fig. 2b). Although the probe of ZmMPK3-1 could detect both ZmMPK3-1a and ZmMPK3-1b, we detected only one band in our Northern blot analysis. The RNA band represented a mixture of ZmMPK3-1a and ZmMPK3-1b (Fig. 2b). Therefore, we used RT-PCR to test the specific expression of ZmMPK3-1a and ZmMPK3-1b. As shown in Fig. 2c. ZmMPK3-1b was expressed in stem and leaf and was almost undetectable in root. ZmMPK3-1a, however, was expressed in root, stem, and leaf, with higher amounts detectable in root. These results showed that ZmMPK3-1a, ZmMPK3-1b, and ZmMPK3-2 were expressed differentially in root, stem, and leaf of maize seedlings. The expression of ZmMPK3-1 in root was due to the transcript of ZmMPK3-1a. AS of ZmMPK3-1 occurred mainly in stem and leaf. The organ-specific expression of ZmMPK3-1 and ZmMPK3-2 suggested that, although the two genes share high sequence identity, they may play different roles in different tissues.

Expression of ZmMPK3-1a, ZmMPK3-1b, and ZmMPK3-2 in Response to ABA and NaCl in Leaves

We next tested whether expression of ZmMPK3-1a, ZmMPK3-1b, and ZmMPK3-2 could be regulated by external stimuli. In our preliminary experiments, we found that expression of ZmMPK3-1 and ZmMPK3-2 was clearly induced by ABA (100 μM) (Liu 2012) or NaCl (200 mM), but was not affected significantly by salicylic acid (SA; 1 mM), H2O2 (10 mM), 30 % PEG6000 (w/v), or methyl jasmonate (MeJA; 100 μM) (data not shown). ZmMPK3-1 was regulated slightly by high temperature (40 °C) (data not shown). Figure 2d–g shows the effects of ABA (100 μM) or NaCl (200 mM) on expression of ZmMPK3-1a, ZmMPK3-1b, and ZmMPK3-2 in maize seedlings within 24 h. ABA (100 μM) could induce expression of both ZmMPK3-2 and ZmMPK3-1 (Fig. 2d). The induction of ZmMPK3-2 by ABA (100 μM) was more transient (within 1 h) than that of ZmMPK3-1. As shown in Fig. 2e, the increased expression of ZmMPK3-1 in response to ABA (100 μM) was contributed mainly by ZmMPK3-1a, as ABA (100 μM) did not significantly alter the expression of ZmMPK3-1b within 24 h. The increased expression of ZmMPK3-2 and ZmMPK3-1 in response to NaCl (200 mM) was likely complementary. NaCl (200 mM) induced strong expression of ZmMPK3-2 at 1–3 h and down-regulated ZmMPK3-2 at 6 h. The increased expression of ZmMPK3-1 was detectable from 4 h to 24 h (Fig. 2f). No clear alteration of ZmMPK3-1b was observed within 24 h, indicating that the significant increase of ZmMPK3-1 was contributed by ZmMPK3-1a but not ZmMPK3-1b (Fig. 2g).

Expression analysis showed that the segmentally duplicated ZmMPK3-1 and ZmMPK3-2 had a dissimilar expression pattern in root, stem, leaf, and in response to ABA (100 μM) or NaCl (200 mM), indicating that unique functional roles may have been distributed to the two genes after evolutionary duplication.

Conclusion

In this research, we reported that previously uncharacterized ZmMPK4 (ZmMPK3-1b in this research) was one of the two splicing variants of ZmMPK3-1. Generation of ZmMPK3-1b was due to retention of the third intron (90 bp) of ZmMPK3-1 (Fig. 1c). ZmMPK3-1b was expressed in the stem and leaf of maize seedlings (Fig. 2c). In leaf, we did not detect significant alteration in expression of ZmMPK3-1b in response to ABA (100 μM), NaCl (200 mM), SA (1 mM), H2O2 (10 mM), 30 % PEG6000 (w/v), or MeJA (100 μM) (Fig. 2d–g and data not shown). ZmMPK3-1 and ZmMPK3-2 were segmentally duplicated genes in the maize genome. ZmMPK3-1 and ZmMPK3-2 proteins shared 90.1 % identity (Supplementary Fig. S4). ZmMPK3-1 and ZmMPK3-2 had a dissimilar expression pattern in root, stem, leaf, and in reponse to ABA (100 μM) or NaCl (200 mM), indicating that unique functional roles may have been distributed to the two genes after evolutionary duplication (Fig. 2). Further study is needed to elucidate the mechanism of AS of ZmMPK3-1 and to explore the biological function of ZmMPK3-1a, ZmMPK3-1b, and ZmMPK3-2.