The 10 Arabidopsis CBL and 26 CIPK genes were downloaded from TAIR (http://www.arabidopsis.org) as queries to search against turnip genomes (http://www.bioinformatics.nl/brassica/index.html?data=bras_tp%2Fdata&loc=A03%3A22821424..22821483&tracks=DNA%2CTranscripts%2CGenes&highlight=). The domains and functional sites in each protein were examined with ps_scan.pl. All CBL protein sequences containing the EF-hand calcium-binding domains (PS50222), as well as all CIPK protein sequences with protein kinase domains (PS50011) and the NAF/FISL motif (PS50816) were extracted as candidates. All candidates were used to search against the GenBank nonredundant protein database. Homology analysis between turnip and Arabidopsis was conducted using DNAMAN software. Physicochemical parameters including the MW, theoretical PI, grand average of hydropathicity, and the number of amino acids were calculated using the ProtParam tool of ExPaSy (http://web.expasy.org/protparam/; Gasteiger et al., 2005). Putative EF-hand was predicated using a simple modular architecture research tool (http://smart.embl-heidelberg.de/). Myristoylaton and palmitoylation motifs were predicted using PlantsP (http://mendel.imp.ac.at/myristate/SUPLpredictor.htm) and CSS-Palm 3.0 software (Ren et al., 2008), respectively. The diagram of the intron/exon structures of BrrCBL and BrrCIPK was analyzed using the online Gene Structure Display Server (http://gsds.cbi.pku.edu.cn/; Hu et al., 2015). Subsequently, the MEME program was used to search for conserved motifs in the turnip BrrCBL and BrrCIPK protein sequences (Bailey et al., 2006).

The BrrCBL and BrrCIPK protein sequences were aligned using the MAFFT version 7 program, and phylogenetic trees were constructed using the MEGA 5.0 software with the neighbor-joining method and the 1,000 bootstrap test replicates (Tamura et al., 2011).

To map the locations of BrrCBL and BrrCIPK genes in turnip, the chromosomal distribution of turnip genomic sequences were generated by circus software (http://www.circos.ca/software/download/circos/). Then, two types of gene duplications were recognized, namely, segment duplication and tandem duplication (Flagel and Wendel, 2009). The synonymous (Ks) and nonsynonymous (Ka) substitution rates were estimated by the codeml program of PAML4 (Yang, 2007). The divergence time (T) of BrrCBL and BrrCIPK gene pairs was calculated as T = Ks/2λ (divergence rates λ = 1.5 × 10⁻⁸ for Arabidopsis; Cao et al., 2011).

The coding region (CDS) of BrrCBL genes was cloned and fused to a binary vector pRI101-AN DNA, with a green fluorescent protein (GFP) and driven by the Cauliflower mosaic virus (CaMV) 35S promoter, forming a 35S:BrrCBL-GFP construct. After confirmation by sequencing, these constructs were separately transferred into Agrobacterium tumefaciens EHA105 by electroporation for infiltration into leaves of Nicotiana benthamiana (Wydro et al., 2006). For the control, we tested the subcellular localization of GFP alone in the leaf cells of N. benthamiana. Freshly made transformed Agrobacteria cell cultures were resuspended in media containing 10 mM MES–KOH (pH 5.6), 10 mM MgCl₂·6H₂O, and 100 μM/L acetosyringone (AS, Sigma), adjusted to an OD₆₀₀ of 0.6–0.8. Next, 3 mL of culture was obtained by sterile single-use syringes to inject into the abaxial air space of N. benthamiana leaves. The leaf section near the injection site was squashed 3 days after infiltration. Fluorescence images were detected using a laser-scanning confocal microscope (Olympus Optical Co. Ltd., Tokyo, Japan). GFP was excited with a 488 nm laser line, and emissions were captured using a 505–530 band-pass filter.

The estimated expression profiles expressed in fragments per kilobase of exon model per million mapped reads values for each BrrCBLs and BrrCIPKs from different developmental stages were obtained from RNA-Seq data in the reanalyzed turnip (Li et al., 2015). Samples were obtained in the early stage before cortex splitting (ES1 and ES2), cortex-splitting stage (CSS1 and CSS2), and secondary root-thickening stage (RTS1 and RTS2) in turnip. These stages were described in detail in a previous study (Li et al., 2015). The normalization and hierarchical clustering analysis of gene expression patterns were performed on the basis of Pearson coefficients with average linkage using the Genesis software (version 1.7.1; Sturn et al., 2002).

Seeds of turnip were grown in soil pots and 1/4 Hoagland's nutrient solution (pH 5.5) under controlled conditions (28°C day/25°C night cycle, relative humidity of 75–80%, 200 mmol photons m⁻² s⁻¹ light intensity). After 3 weeks of germination, drought stress was executed by withholding water for 7 days. For MeJA, high salinity, and cold stress, seeding was exposed to MeJA (1 mM), NaCl (100 mM), and cold temperature (4°C). The roots and leaves of seedlings were harvested after treatment for 3 h, with 0 h as the control. The samples, including the root, stem, leaf, and flower, were harvested at the flowering stage. All samples were immediately frozen in liquid nitrogen and then stored at −80°C for RNA extraction.

Total RNA samples were isolated using the Eastep® Super Total RNA Extraction Kit (Promega, Madison, WI, USA). RNA was quantified by NanoDrop1000 (NanoDrop Technologies, Inc.) with integrity checked on 0.8% agarose gel. Approximately 5 μg RNA was reverse transcribed using the GoScript Reverse Transcription System (Promega, Madison, WI) to generate cDNA. qRT-PCR was conducted in triplicate with different cDNAs synthesized from three biological replicates of different tissues and treatments using FastStart Universal SYBR Green Master (Rox, Roche, Indianapolis, IN, USA) and a 7,500 Sequence Detection System (Applied Biosystems, USA). The reaction parameters for thermal cycling were as follows: 95°C for 10 min, followed by 40 cycles of 94°C for 5 s, and 60°C for 15 s. B. rapa tubulin beta-2 chain-like (LOC103873913) was amplified as an internal control. The primers used for qRT-PCR are listed in Supplementary Table 1. The relative gene expression levels were obtained by dividing the extrapolated transcript levels of the target genes by the levels of the internal control from the same sample. The results were obtained from the comparison of the treatment with the control using independent-samples T-test. Statistical analysis was performed using the software IBM SPSS Statistics 20.0.

Y2H assays were conducted using the MatchMaker Y2H system (http://www.clontech.com/). The CDS regions of BrrCBL and BrrCIPK genes were first subcloned into pGADT7 and pGBKT7 vectors, respectively. The primers used for vector construction are listed in Supplementary Table 1. Then, the plasmids of BrrCBL and BrrCIPK were sequentially transformed into yeast strain AH109 through the lithium acetate method following the protocol described in the Yeast Protocols Handbook (Clontech). The transformed yeast cells were grown on medium as follows: (1) lacking leucine and tryptophan SD−Trp−Leu to serve as a positive control for transformation and loading; (2) lacking leucine, tryptophan, and histidine SD−Trp−Leu−His to test for PPIs under low stringency, and (3) lacking leucine, tryptophan, histidine, and adenine SD−Trp−Leu−His−Ade to test for interactions under stringent conditions. Cell growth was recorded at 48 h intervals over the course of 6 days. The method of assessment of the interaction strength between sets of proteins was as follows (Kleist et al., 2014). Red boxes indicate vigorous growth on -LTHA plates. Orange boxes indicate weak growth on -LTHA plates. Yellow boxes indicate robust growth on -LTH plates but no growth on -LTHA plates. Pale yellow boxes indicate weak growth on -LTH plates but with each CBL–CIPK interaction conferring superior growth to that in the empty vector control. Gray boxes indicate only growth on -LT plates.

For BiFC assays, the CDS regions of selected BrrCBL and BrrCIPK genes were sub-cloned into 35s-SPYCE and 35s-SPYNE vectors, respectively. The primers used for vector construction are listed in Supplementary Table 1. The fusion construct was transferred into A. tumefaciens strain EHA105, then into tobacco leaves by Agrobacterium infiltration (Wydro et al., 2006). Infiltrated leaves were observed with a laser scanning confocal microscope (Olympus Optical Co. Ltd., Tokyo, Japan).

The full-length coding sequences of BrrCBL9.1 and BrrCBL9.2 were each inserted into a binary vector pRI101-AN DNA fused with a GFP and driven by Cauliflower mosaic virus (CaMV) 35S promoter. Consequently, a 35S: BrrCBL-GFP construct was obtained. Arabidopsis plants were transformed using the floral-dip method with the A. tumefaciens strain EHA105 (Clough and Bent, 1998). The assayed plants (overexpression, cbl9 mutant, complemented lines, and WT) were grown on 1/2 solid Murashige and Skoog (MS) medium containing NaCl (100 mM) for stress analysis. After 8 days, the seedling root lengths were measured. For expression analysis of corresponding BrrCIPK genes specially interacted with BrrCBL9.1 and BrrCBL9.2 in overexpression, cbl9 mutant, and complemented lines, the assayed plants were grown on MS medium for about 10 days. Then they were transferred to soil in a growth chamber (photoperiod of 16-h light/8 h darkness) at 21°C. After 2 weeks of germination, seedings were exposed to NaCl (100 mM) and were harvested after treatment for 3 h, with 0 h as the control. All samples were immediately frozen in liquid nitrogen and then stored at −80°C for RNA extraction.

We conducted a genome-wide identification of CBLs and CIPKs in turnip. Accordingly, we used Arabidopsis 10 CBL and 26 CIPK CDS sequences (Zhang et al., 2014) as queries to search against the published genome of turnip. After the search, we identified 19 CBLs and 51 CIPKs, the names assigned to these genes, and their phylogenies and sequence similarities to corresponding individual AtCBLs and AtCIPKs (Tables 1, 2). The full-length turnip CBL proteins were 194 (BrrCBL5) to 249 (BrrCBL10.1) amino acids long, except BrrCBL3.3 with 449 amino acids (Table 1). All the BrrCBLs contained four EF-hand motifs, which provide the structural basis for calcium binding. BrrCBL1.1, -1.2, -4.1, -4.2, -4.3, -5, -9.1, and -9.2 proteins each harbored a conserved N-myristoylation and palmitoylation motif (MGCXXS/T) that may be responsible for membrane association (Batistic et al., 2008). The BrrCIPKs identified in our study ranged in isoelectric points (PIs) from 5.49 (BrrCIPK16) to 9.28 (BrrCIPK23.1), with coding sequences of 352–524 amino acids (Table 2). These ranges suggest the diversity of the biochemical properties of BrrCIPKs. Among the identified BrrCBL and BrrCIPK genes, the CDS of 15 BrrCBLs and 47 BrrCIPKs were successfully amplified by PCR from cDNA prepared from turnip as the template using available primers (Supplementary Table 1). We further predicted the localization of 15 BrrCBL proteins by confocal laser microscopy (Supplementary Table 1). The BrrCBL1.1, -1.2, -2.2, -3.2, -4.3, -5, -8, -9.1, and -9.2–GFR fusion proteins each emitted a green fluorescent signal in the plasma membranes and nuclei of the epidermal cells of N. benthamiana leaves. By contrast, BrrCBL2.1, -3.1, -4.2, -6, -10.2, and -10.4–GFR were detected in plasma membranes. Although, lacking myristoylation and palmitoylation motifs, BrrCBL3.2 and BrrCBL8 localized in the membrane. The dynamic localization of CIPKs was determined by their specific CBL partners (Batistic et al., 2010; Batistic and Kudla, 2012). We did not analyze the subcellular localization of the CIPK protein.

To gain insights into the phylogenetic analysis of both CBL and CIPK families, we generated two phylogenetic trees for each family. The BrrCBL proteins were divided into three groups, namely, groups I, II, and III (Figure 1). We also compared the evolutionary relationship between turnip and Arabidopsis in terms of the CBLs (Supplementary Figure 2A). All BrrCBLs were clustered closely with AtCBL orthologs. However, CBL7 was present in Arabidopsis but absent in turnip. To analyze the structural characteristics of the BrrCBLs, we mapped gene structures, including exons and introns, on the basis of the turnip genome sequence. The results showed that all the BrrCBLs were intron rich, with six to eight introns (except BrrCBL3.3 with the 15 introns). Hence, the exon/intron numbers of the CBL gene potentially vary. Moreover, we analyzed the EF-hand motifs for the BrrCBL proteins (Supplementary Figure 3). BrrCBL1, -3, -4, -5, and -8 possessed no canonical EF hand, whereas BrrCBL6, -7, and -10 possessed one and BrrCBL1 and -9 harbored two canonical EF hands. The space of each EF-hand motif was absolutely conserved in the BrrCBL proteins. Up to 23 amino acids were observed to lie between EF1 and EF2, 25 amino acids between EF2 and EF3, and 32 amino acids between EF3 and EF4.

As expected, the phylogenetic tree divided the BrrCIPKs into two groups, namely, group A (intron-rich) (18 BrrCIPKs) and group B (intron-less) (33 BrrCIPKs); this finding was consistent with previous evolutionary analyses (Kolukisaoglu et al., 2004) (Figure 2). The structural differences in the BrrCIPKs may allow BrrCIPK genes to function differently because the functional structural gene domains determine the gene's function (Albrecht et al., 2001). All of the 51 BrrCIPKs were clustered closely with AtCIPK orthologs (Supplementary Figure 2B). In the Arabidopsis CIPKs (Batistic and Kudla, 2004), all BrrCIPKs contained domain structures similar to that of AtCIPK24, including an N-terminal protein kinase catalytic (PKC) domain and a C-terminal regulatory domain (NAF/FISL motif) was identified (Supplementary Figure 4A). Previous studies showed that the PKC domain in the N-terminal and a short motif called NAF/FISL motif located in the C-terminal regulatory domain were necessary and sufficient for mediating interactions with CBLs (Albrecht et al., 2001; Guo et al., 2001; Hashimoto et al., 2012). Thus, amino-acid sequences were produced to gain insight into the BrrCIPK functional domains within each residue position.

We further searched for the conserved motifs in both BrrCBL and BrrCIPK proteins using the MEME program to analyze the diversity of motif compositions. A total of 15 conserved motifs designated as motif 1 to motif 15, respectively, were identified within the genes (Supplementary Figures 5, 6). These motifs may also help predict the genes' functions. Our analysis revealed that motif 10 contained the N-myristoylation motif of 12 BrrCBL proteins (Supplementary Figure 5). For the BrrCIPKs, motifs 1 and 2 were found within the PKC domain. Motif 8 in the protein–protein interaction (PPI) domain was shared by all BrrCIPKs. Motifs 7 and 11, also shared by all the BrrCIPK members, were found within the region of the NAF/FISL domain and showed relatively conserved amino acids (N, A, F, I, S, and L; Supplementary Figure 6). The amino acid residues at the 2nd, 3rd, 4th, 7th, and 10th sites of the NAF/FISL motif were rather conserved (Supplementary Figure 4B). The phylogenetic analysis together with the structure analysis presented herein may facilitate the functional annotation and study of CBLs and CIPKs in turnip.

To determine the relationship between gene duplication and genetic divergence in the CBL or CIPK gene family in turnip, we investigated the chromosomal location of BrrCBL and BrrCIPK (Figure 3). All the BrrCBLs and BrrCIPKs were located at the top sections of chromosomes. Among the 10 chromosomes, chromosome 1 contained the maximum number of seven BrrCBL genes. Chromosomes 4, 5, and 7 carried no BrrCBL. Chromosomes 6, 8, and 10 each contained only one BrrCBL gene, and all other chromosomes contained 2–4 BrrCBLs (Figure 3A). Given the phylogenetic analyses and an entire genome analysis of gene duplications, we observed that six pairs of BrrCBL were putative paralogs, including five segmental duplications (BrrCBL1.1/1.2, BrrCBL2.1/2.2, BrrCBL3.1/3.3, BrrCBL4.1/4.3, and BrrCBL9.1/9.2) and one tandem duplication (BrrCBL10.2/10.3) (Table 3). Furthermore, we found that some paralog genes were located on different chromosomes or at different locations on the same chromosome (Figure 3B). Two pairs of BrrCBL3.1/3.3 and BrrCBL10.2/10.3 were located on the same chromosomes at different locations. Meanwhile, other paralogs were located on different chromosomes. Among all the genes studied, the BrrCIPK genes were found variably distributed on the turnip chromosomes. Chromosomes 5 and 9 contained the maximum and the minimum numbers, respectively, of BrrCIPK genes. A total of 17 pairs of putative BrrCIPK paralog genes were produced by segmental duplication, including 14 duplication events between chromosomes (BrrCIPK1.1/1.2, BrrCIPK2.2/2.3, BrrCIPKV6.1/6.2, BrrCIPK7.1/7.2, BrrCIPK9.1/9.2, BrrCIPK10.1/10.2, BrrCIPK12.1/12.2, BrrCIPK13.2/13.3, BrrCIPK20.1/20.2, BrrCIPK21.1/21.2, BrrCIPK22.1/22.2, BrrCIPK23.1/23.2, BrrCIPK5/25, and BrrCIPK26.1/26.2), as well as three duplication events within the same chromosome (BrrCIPK4.1/4.2, BrrCIPK11.1/11.4, and BrrCIPK17.1/17.4; Figure 3B, Table 3). These observations suggest the large-scale segmental duplication events involved in the expansion of BrrCBL and BrrCIPK gene families in turnip.

We estimated the divergence time (T) of six pairs of turnip putative BrrCBL paralog proteins by measuring the synonymous (Ks) and nonsynonymous (Ka) mutation rates (Table 3). In the process, we used a divergence rate of 1.5 × 10⁻⁸ mutations per synonymous site per year. The estimated divergence time (T) for the turnip CBL paralogs was approximately between 6.3633 and 17.3700 million years ago (MYA), with an average duplication time of ~12.4256 MYA. Ka/ks (ω) was calculated for each pair of CBL paralog genes. The ω-values for all of the putative CBL paralogs with mean value of 0.134 were less than one. Thus, the five pairs of turnip CBL proteins were potentially under strong purifying selection pressure. However, one pair of turnip CBL genes (BrrCBL10.2/BrrCBL10.3, ω = 0.5547) attained relatively large ω-values. This result implied that the corresponding genes may have evolved rapidly from the last common ancestor. We also estimated the divergence time of 17 pairs of turnip putative CIPK paralog genes. Moreover, the estimated T for BrrCIPK paralogs was between 1.4467 and 28.1533 MYA, with an average duplication time of approximately 12.4194 MYA. Interestingly, the divergence time of three BrrCIPK paralogs (BrrCIPK6.1/6.2, BrrCIPK11.1/11.4, and BrrCIPK5/25) were estimated to have occurred recently (only approximately one). Schranz and Mitchell–Olds (Schranz et al., 2006) estimated the time of very early radiation of Brassicaceae at 34 MYA. Other authors proposed that the conserved linkage arrangements between B. rapa and A. thaliana diverged from a common ancestor at ~14.5–20.4 MYA (Ferguson et al., 2003). However, the divergence time of three pairs of BrrCIPK paralogs (BrrCIPK10.1/10.2, BrrCIPK22.1/22.2, and BrrCIPK5/25; 20.9600–28.1533 MYA) occurred precedent to the period of the origin of the B. rapa. The ω in all the putative BrrCIPK paralogs with mean value of 0.167 were less than one. Hence, the 17 pairs of turnip BrrCIPK genes were possibly under strong purifying selection pressure. By contrast, three pairs of BrrCIPK genes, namely, BrrCIPK7.1/7.2 (ω = 0.3621), BrrCIPK17.1/17.2 (ω = 0.3784), and BrrCIPK26.1/26.2 (ω = 0.4386), achieved relatively large ω-values, which indicate that they may have evolved rapidly from those of the last common ancestor.

We determined the spatial and temporal expression profiles of 19 BrrCBL and 51 BrrCIPK genes in different tissues of turnip (root, stem, leaf, and flower) by quantitative real-time PCR (qRT-PCR; Figure 4A). A heat map was created through the hierarchical clustering of the gene expression profiles of the BrrCBL and BrrCIPK genes. These profiles can be classified into two clusters by expression pattern. Cluster I consisted of 16 BrrCBLs and 28 BrrCIPKs with different expression patterns in different tissues. Some of these genes were highly expressed in certain tissue tested. BrrCBL1.1, -1.2, -4.2, and -4.3, as well BrrCIPK6.3, -10.2, and -17.1, almost showed peak transcript levels in the root. BrrCIPK13.2 and 23.3 were highly transcribed in the stem, whereas BrrCIPK3, -9.1, -15, -17.2, and -26.1 were highly transcribed in the leaf. These transcriptional patterns indicate that these genes may be involved in organ development and growth. By contrast, some genes, especially BrrCBL10.2, showed no transcript levels in all tissues. Cluster II included three BrrCBL and 23 BrrCIPK genes with high transcript abundance in the flowers. Hence, the genes in Cluster II may play important roles in flower development. Additionally, some paralogs showed opposite expression patterns. Examples include 4 out of 6 BrrCBL paralogs (BrrCBL2.1/2.2, BrrCBL4.1/4.3, BrrCBL10.2/10.3, and BrrCBL12.1/12.2) and 4 out of 17 BrrCIPK paralogs (BrrCIPK1.1/1.2, BrrCIPK2.2/2.3, BrrCIPK17.1/17.4, and BrrCIPK26.1/26.2) in different tissues (Figure 4A).

The expression levels of BrrCBL and BrrCIPK genes were reanalyzed using publicly available RNA-sequence data from three different developmental stages (Li et al., 2015). These stages include the average of the early stage before cortex splitting (ES1 and ES2), the cortex-splitting stage (CSS1 and CSS2), and the secondary root-thickening stage (RTS1 and RTS2) in turnip corresponding to ES, CSS, and RTS, respectively (Figure 4B). Likewise, the transcripts of the BrrCBL and BrrCIPK genes can be detected in different developmental stages and can be divided into three clusters by expression pattern. The eight BrrCBL and 18 BrrCIPK genes were assigned to Cluster I, except BrrCIPK13.1, -19, and -20.1, which were relative highly expressed in the secondary RTS. Five BrrCBL and 10 BrrCIPK genes were found in Cluster II, and they were almost exclusively expressed in the early stage before cortex splitting (ES), except for BrrCIPK2.2 and BrrCIPK13.3. Among the Cluster III genes, 6 BrrCBL and 23 BrrCIPK genes were highly expressed in the early stage before cortex splitting (ES) and CSS. By contrast, the expression of BrrCBL3.2 was not detected and BrrCBL4.3 was expressed at a very low level. These results show that some of the genes were modestly expressed in developmental stages. Meanwhile, some genes were specifically and abundantly expressed in certain developmental stages. Therefore, such genes may play important role during developmental stages. We also found that a subset of paralogous genes were differentially expressed in the three developmental stages (Figure 4B). BrrCBL9.1 was expressed abundantly in the CSS and RTS stages, whereas BrrCBL9.2 was specifically expressed in the ES. The expression patterns of two BrrCBL paralogous pairs (BrrCBL1.1/1.2 and BrrCBL4.1/4.3) and six BrrCIPK paralogous pairs (BrrCIPK1.1/1.2, BrrCIPK2.2/2.3, BrrCIPK7.1/7.2, BrrCIPK13.2/13.3, BrrCIPK7.1/7.2, and BrrCIPK23.1/23.2) also diverged dramatically in developmental stages. Overall, the expression patterns of BrrCBL and BrrCIPK genes at different tissues and different developmental stages may indicate the functional divergence between paralogs.

We then further investigated the expression profiles of BrrCBL and BrrCIPK genes under stress conditions. We also evaluated the possible functional divergence of paralog genes. To achieve these goals, we conducted qRT-PCR analysis. The transcript level of the BrrCBL1.1 gene showed the most significant increases among those of other genes after various stress treatments (Figure 5). By contrast, BrrCBL2.1, -4.2, and -4.3 expression was remarkably downregulated in four stress treatments. Under NaCl treatment, the BrrCBL1.1, -3.2, -4.1, -5, -6, and -10.2 transcript levels significantly increased. Meanwhile, BrrCBL1.2, -3.3, -9.2, and -10.4 expression was still slightly upregulated after NaCl treatment. Under the drought treatment, BrrCBL1.1, -1.2, -2.2, -4.1, -6, -9.1, and -10.3 expression significantly induced. In addition, cold stress significantly upregulated the expression levels of BrrCBL1.1, -2.2, -8, -9.2, -10.1, and -10.4. In response to MeJA treatment, BrrCBL1.1, -3.2, -5, -6, -8, -9.1, -9.2, and -10.2 were highly upregulated. Additionally, we investigated the expression profiles of BrrCBL paralog gene pairs and found that they exhibited different expression patterns under different stress conditions. BrrCBL9.1 expression was downregulated, whereas BrrCBL9.2 expression was upregulated by NaCl and cold treatments. The results suggest the functional divergence of the BrrCBL paralog pairs.

For the 51 BrrCIPK genes assayed under four stress treatments, the most striking induction was observed for 34 BrrCIPK genes after MeJA treatment, followed by 30 BrrCIPK genes after NaCl treatment. Among these genes, BrrCIPK2.2, -3.2, -5, -8, -9.1, -10.2, -18, -19, and -20.1 exhibited remarkable MeJA-induced upregulation. BrrCIPK10.2, -20.2, -23.2, and -26.1 also exhibited remarkable NaCl-induced upregulation. Drought treatment induced the expression of 24 BrrCIPKs as that under cold treatment. Strikingly, BrrCIPK17.4 was markedly upregulated in response to drought. The transcript levels of BrrCIPK9.1 and BrrCIPK21.1 showed remarkable increases after cold treatment. Conversely, the transcript levels of BrrCIPK17.1 gene were significantly decreased under the four treatments (Figure 6). Several paralog gene pairs (such as BrrCIPK1.1/1.2, BrrCIPK4.1/4.2, and BrrCIPK10.1/10.2) demonstrated divergent expression patterns, that is, one gene is induced, whereas the other is suppressed by the treatment. The results reveal that the BrrCIPK paralogs exhibit distinct expression patterns under different stress conditions. This occurrence also suggests the functional divergence of the BrrCIPK-duplicated genes.

According to previous reports, CBL–CIPK complexes play highly important roles in response to external stimuli (Manik et al., 2015; Sun et al., 2015). We wish to determine whether the same stress responses of BrrCBL and BrrCIPK transcripts can interact with each other. To investigate this aspect in the turnip expression profiles, we performed coexpression analysis, which can predict the interaction of the BrrCBL–BrrCIPK complex in response to stress. A number of BrrCBLs and BrrCIPKs were upregulated and downregulated in response to one or more stress treatments (Supplementary Figure 7). For example, the transcript abundances of BrrCBL1.1, -8, -9.2, and -10.4 as well as BrrCIPK2.3, -8, -14, and -24 were increased by cold. We therefore speculate that potential interactions may exist between these BrrCBL and BrrCIPK genes.

To further elucidate the interaction preferences between CBLs and CIPKs in turnip, we used a yeast two-hybrid (Y2H) system. Up to 15 BrrCBL and 47 BrrCIPK genes were cloned and used in Y2H assays. In the current study, we demonstrated 166 interactions between BrrCBLs and BrrCIPKs; the interactions displayed varying strengths and specificities (Figure 7). Notably, BrrCBL9.1 and -9.2 were paralog proteins and showed strong interactions with 22 and 25 BrrCIPKs, respectively. Some specific BrrCIPKs were identified, and these genes interacted specifically with either of the others. BrrCBL9.1 specifically interacted with BrrCIPK6.1, -25, and -26.2, whereas BrrCBL9.2 interacted specifically with BrrCIPK6.2, -6.4, -7.1, -18, -22.1, and -23.3. Thus, BrrCBL9.1 and -9.2 were involved in more interactions with BrrCIPKs than those of the other BrrCBLs. However, the two genes displayed a fairly different interaction with its paralogs, and hence may be appropriate for further in-depth research. Overall, this type of interaction analysis enabled the deduction of potential gene functions to guide future functional studies. We then selected 10 of the interactions (BrrCBL3.2-BrrCIPK16, BrrCBL4.2-BrrCIPK7.1, BrrCBL9.1-BrrCIPK16, -20.1 and -23.2, BrrCBL9.2-BrrCIPK7.1, -16, -20.1, -23.2, -23.3) detected between BrrCBL and BrrCIPK proteins using the Y2H assay for further analyses using the BiFC method. We found that green fluorescence signals were observed when BrrCBLs and BrrCIPKs were co-expressed in epidermal cells of tobacco leaves (Figure 8). It seemed that the fluorescent signal were consistent with the localization signal of corresponding BrrCBL proteins (Supplementary Figure 1). In contrast, in controls in which BrrCBL3.2, -4.2, -9.1, or -9.2 was expressed together with YFP_N only, as well BrrCIPK7.1, -16, -20.1, -23.2, or -23.3 with YFP_C only, no signal of reconstructed YFP appeared.

The AtCBL9 gene in Arabidopsis possesses two counterparts, namely, BrrCBL9.1, and -9.2, in turnip. The differences between these genes were discovered through expression and interaction profiles. Previous studies revealed that the Arabidopsis cbl9 mutant is hypersensitive to ABA, salt, drought, and glucose stress (Pandey et al., 2004). To dissect the biological function of turnip CBL9.1 and -9.2, we overexpressed each gene in Arabidopsis using a CaMV 35S promoter (OE-BrrCBL9.1 and OE-BrrCBL9.2, respectively). To examine whether BrrCBL9.1 and -9.2 can rescue the cbl9 mutant phenotype, we transformed a cbl9 mutant plant with BrrCBL9.1 and -9.2 using the CaMV 35S promoter (cbl9/BrrCBL9.1 and cbl9/BrrCBL9.2, respectively). Seeds from wild-type (WT), cbl9 mutant, complemented, and overexpressing plants were germinated on 1/2 Murashige and Skoog (MS) and 1/2 MS with 100 mM NaCl (Figures 9A,B). On the 1/2 MS medium plates, the WT, cbl9 mutant, complemented, and overexpressing lines displayed relatively similar growth phenotypes. The latter three plants showed no significant difference in primary root length from the WT. After 100 mM NaCl treatment, no significance difference in the primary roots was observed between cbl9 and cbl9/BrrCBL9.1, as well as between the OE-BrrCBL9.1 and WT under salt-stress conditions. However, the primary roots of cbl9/BrrCBL9.2 were restored to WT background levels, and the primary roots of OE-BrrCBL9.2 were much longer than those of WT plants. Therefore, our data indicate that the BrrCBL9.2 gene can promote root elongation in both transgenic and complementary Arabidopsis under salt stress. We also detected the expression level of corresponding BrrCIPK genes specially interacted with BrrCBL9.1 and -9.2 in complemented and overexpressing transgenic plants (Figure 9C). As a result, BrrCIPK6.4 and -18 showed the most significant increase in complemented and overexpressing transgenic Arabidopsis of BrrCBL9.2 under NaCl treatment, but BrrCBL9.1 does not affect. This indicated that the interaction between BrrCBL9.2 and BrrCIPK6.4 and -18 may play a role in salt responses.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.