Introduction

Plants are frequently challenged by abiotic stresses such as heat, cold, drought, and high salinity, which cause impaired growth and development of plants. To deal with these stresses, numerous stress-related genes were induced in plants, leading to physiological and metabolic changes that improve stress tolerance (Shinozaki et al. 2003; Zhou et al. 2017b). Among them, late embryogenesis abundant (LEA) proteins, together with osmotin, chaperones, and detoxification enzymes, play vital roles in abiotic stress tolerance (Cao et al. 2017; Chen et al. 2015).

Dehydrins (DHNs) are a group of highly hydrophilic proteins that belong to the group II of LEA with low molecular weights from 10 to 70 kDa, and possess high contents of Gly, Ala and Ser, and large net charge relative to non-polar hydrophobic residues (Charfeddine et al. 2017). Due to their unified and outstanding characteristics, DHNs are universally categorized as a structural family of proteins which are known as intrinsically disordered proteins (IDPs) (Tompa 2002) without a defined 3D structure, and are highly dynamic to form a lot of different structures (Malik et al. 2017). DHNs have three categories of conserved motifs named as K-, Y-, and S-segments. Usually, the K-segment is located near the C-terminus with a typical sequence EKKGIMDIKEKLPG, which was predicted to form an amphipathic helix and to be associated with the functions dehydrin, and it is a common feature of all DHNs (Hara et al. 2017; Koag et al. 2009) except for several members (Perdiguero et al. 2014; Verma et al. 2017). Different DHNs may occasionally have some single nucleotide substitutions or structural modifications in the K-segment (Graether and Boddington 2014). The other two conserved motifs have been termed as the Y-segment and S-segment, which consist of the sequences of DEYGNP and LHRSGSSSSSSSEDD (or related sequences), respectively (Graether and Boddington 2014; Malik et al. 2017). Unlike K-segment, Y-segment and S-segment are not necessarily present in all DHNs. According to the presence and number of Y-, S- and K-segments, DHNs can be divided into five major types: YnSKn, Kn, SKn, YnKn, and KnS (Close 1996; Lv et al. 2017), and an intermediate type SnKS (Abedini et al. 2017).

Many studies have reported that various abiotic stresses can induce the expression of DHN genes, such as cold, drought, and salinity (Abedini et al. 2017; Charfeddine et al. 2017; Jing et al. 2016; Verma et al. 2017). It is worth noting that many DHN genes are only responsive to one or several abiotic stresses, suggesting that different mechanisms have been developed in plants to adapt to low temperature, drought, and salinity (Malik et al. 2017). For example, in apple, MdDHN1, MdDHN2, MdDHN4, MdDHN5, and MdDHN6 were obviously induced under drought conditions, while MdDHN2, MdDHN4, and MdDHN6 were also significantly up-regulated by low temperature (Liang et al. 2012). Many studies showed that the expression levels of SKn-, YnKn-, and KnS-type DHN genes are mainly up-regulated by low temperature, while drought or salt exposure promotes the expression of YnSKn-type DHN genes (Graether and Boddington 2014; Rorat 2006). These findings suggest that DHN proteins and abiotic stress tolerance are positively correlated.

A number of studies have revealed that DHNs can protect cells from dehydration or cryoprotection and participate in responses to various abiotic stresses, and their overexpression can enhance stress tolerance of plants. For example, transgenic Arabidopsis plants overexpressing DHN genes from Arabidopsis or other plant species showed tolerance to various abiotic stresses, such as cold (Aguayo et al. 2016; Peng et al. 2008), salinity (Brini et al. 2007; Saibi et al. 2015), and drought (Chiappetta et al. 2015), or multiple stresses (Cao et al. 2017; Munoz-Mayor et al. 2012). Arabidopsis plants expressing Rab17 from maize (Figueras et al. 2004), DHN-5 from wheat (Brini et al. 2007), HbDHN1 and HbDHN2 from Hevea brasiliensis (Cao et al. 2017) also exhibited higher osmotic stress tolerance. Improved stress tolerance was also observed in cucumber plants with overexpression of DHN24 from Solanum sogarandinum (Yin et al. 2006), strawberry plants with overexpression of wheat Wcor410a (Houde et al. 2004), banana plants with constitutive overexpression of MusaDHN-1 (Shekhawat et al. 2011), rice plants overexpressing OsDhn1 (Kumar et al. 2014), tobacco plants with overexpression of CarDHN (Hill et al. 2016), SiDHN (Guo et al. 2017), CdDHN4 (Lv et al. 2017), and SbDhn1 (Halder et al. 2017), and so on. These results indicate the potential roles of DHNs in stress tolerance.

A previous research that analyzed cucumber CsLEA gene family suggested that one DHN gene named as CsLEA54 can be induced under drought stress conditions in leaf and root tissues (Altunoglu et al. 2016). Another DHN gene named as CsLEA11 could be induced by heat and cold stress, and heterologous expression of CsLEA11 in E. coli displayed an enhanced tolerance to heat and cold stress (Zhou et al. 2017a). However, these results are not enough for understanding the roles of DHNs in the response to various abiotic stresses in cucumber. In this study, four putative DHN family members were identified from cucumber, and their phylogenetic relationships, protein motifs, gene structures, expression profiles in different tissues and under various abiotic stresses were systematically examined. The results lay a foundation for further functional study of CsDHN genes, and provide new information for the use of them in future breeding of stress tolerance in cucumber.

Materials and methods

Identification of cucumber DHN genes

To predict the DHN gene family members in cucumber, BLAST analysis was conducted using the protein sequences of the DHN genes of Arabidopsis (Hundertmark and Hincha 2008), rice (Verma et al. 2017), and tomato (Cao and Li 2015) as the query sequences against the cucumber genome database from Cucumber Genome Initiative (CuGI, http://cucumber.genomics.org.cn) (Huang et al. 2009). All redundant putative DHN sequences were excluded by aligning to each other, and further confirmation was performed using the Pfam (http://pfam.xfam.org/) (Finn et al. 2016) and Simple Modular Architecture Research Tool (SMART) (http://smart.embl-heidelberg.de) (Letunic and Bork 2018) databases for the existence of a typical dehydrin domain (PF00257) in their protein structures.

Characterization of CsDHN proteins

The disordered regions of the deduced CsDHN proteins were predicted using the protein disorder prediction system (PrDOS) (http://prdos.hgc.jp/cgi-bin/top.cgi) (Ishida and Kinoshita 2007). The theoretical isoelectric point (pI), molecular weight (MW) and the grand average of hydropathy index (GRAVY) of the deduced CsDHN proteins were calculated using the ProtParam tool (http://web.expasy.org/protparam) (Gasteiger et al. 2005). The secondary structure prediction of each CsDHN protein was carried out by SOPMA tool (Geourjon and Deléage 1995).

Multiple sequence alignment and phylogenetic analysis

The multiple sequence alignments of the full-length protein sequences of CsDHNs and DHN proteins from Arabidopsis, tomato, and rice were carried out by ClustalW with default parameters (Larkin et al. 2007). A phylogenetic tree based on the alignment was then constructed using the MEGA 5.0 software by employing the neighbor-joining (NJ) method with 1000 replicates of bootstrap analysis (Tamura et al. 2011).

Chromosomal location, gene duplication, gene structure and promoter region analysis

The chromosomal locations of the CsDHN genes were determined with the MapInspect tool as previously described (Hu et al. 2016). Segmental and tandem duplication events were investigated according to the criteria in previous studies (Kong et al. 2013; Liu et al. 2012). The gene structures of CsDHN genes were analyzed with the Gene Structure Display Server (http://gsds.cbi.pku.edu.cn) (Hu et al. 2015) by comparing the coding sequence (CDS) with their corresponding genomic DNA (gDNA) sequences retrieved from the cucumber genome database. The putative promoter region (1.0-kb sequence upstream of the ATG start codon) of each CsDHN gene was analyzed for the identification of the cis-elements using the PlantCARE database (http://bioinformatics.psb.ugent.be/webtools/plantcare/html) (Lescot et al. 2002).

Plant materials and treatments

The seeds of a cucumber inbred line (Cucumis sativus L. cv. Chinese long No. 9930) were germinated and grown in the field of Jiangxi Agricultural University as previously described (Zhou et al. 2017c). Different tissues including root, stem, leaf, flower, and fruit were sampled and snap-frozen in liquid nitrogen for storage at − 80 °C before isolation of total RNA.

Two-week-old cucumber seedlings were subjected to different abiotic stresses including heat, salt, cold, abscisic acid (ABA), H2O2 and PEG treatments, which were described previously (Zhou et al. 2017b, c). The leaf tissues were harvested at 0, 3 and 12 h after treatment, frozen immediately in liquid nitrogen, and stored at − 80 °C before isolation of total RNA.

RNA extraction, RT-PCR and quantitative RT-PCR (qRT-PCR)

Total RNA of the harvested samples was extracted using Trizol reagent (Tiangen Biotech Co., Ltd. Beijing, China) according to the manufacturer’s instruction, and RNA integrity was visualized on a 1% agarose gel. The first-strand cDNA was synthesized according to the manufacturer’s instructions (Tiangen Biotech Co., Ltd. Beijing, China). RT-PCR was performed with Taq DNA Polymerase (TaKaRa, Japan), and the thermal cycling procedure was conducted under the following conditions: 94 °C for 5 min, followed by 28–30 cycles of 58 °C for 30 s, 72 °C for 30 s, and 72 °C for 5 min. qRT-PCR was performed in triplicate with SYBR Green Master Mix (Tiangen Biotech Co., Ltd. Beijing, China) on the Light Cycler LC480 system as described previously (Zhou et al. 2017b). The relative expression level of each target gene was determined based on the 2−ΔΔCT method (Livak and Schmittgen 2001), using an endogenous actin gene (CsActin) as the reference gene (Zhou et al. 2017a). The primers used for RT-PCR and qRT-PCR are designed and provided in Table S1.

Results

Genome-wide identification of CsDHNs in cucumber

To identify the DHN members in cucumber, the sequences of DHN proteins from other species including Arabidopsis, tomato, and rice were used to conduct an extensive search of the cucumber database (http://cucumber.genomics.org.cn). The resulting candidate members were further checked for the presence of dehydrin domain using Pfam and SMART tools. As a result, a total of four candidate DHN genes were identified (Table 1), which were named as CsDHN1–CsDHN4 according to their chromosome locations (Table 1). Among them, CsDHN2, CsDHN3 and CsDHN4 genes were previously designated as CsLEA31, CsLEA54, and CsLEA11, respectively (Altunoglu et al. 2016).

Table 1 Basic information of four CsDHN genes and their encoded proteins in cucumber
Full size table

Characterization and comparison of deduced DHN proteins in cucumber

CsDHN genes exhibited gDNA lengths varying from 836 bp (CsDHN2) to 3175 bp (CsDHN1), and encode putative polypeptides with lengths of 162–255 amino acids. CsDHN4 and CsDHN1 showed the smallest (17.47 kD) and largest (29.32 kD) molecular weight of the encoded proteins. GRAVY analysis revealed that the predicted CsDHN proteins were highly hydrophilic, with GRAVY values ranging from − 1.525 (CsDHN3) to − 0.982 (CsDHN1). Among the four proteins, CsDHN1, CsDHN2 and CsDHN4 were basic proteins, and CsDHN3 was an acidic protein (Table 1).

DHNs are rich of glycine (Gly) and polar amino acids, but are lack of cysteine (Cys) and tryptophan (Trp) (Jing et al. 2016). Similarly, four CsDHNs also show such characteristics, except for CsDHN1 and CsDHN3. CsDHN1 possesses 1 Cys (0.4%), 1 Trp (0.4%), 16 Gly (6.3%), 37 negatively charged residues including Asp (6.3%) and Glu (8.2%), 39 positively charged residues including Arg (7.8%) and Lys (7.5%). CsDHN2 harbors no Cys or Trp, with much higher Gly (20.3%), and high percentages of charged and polar amino acids such as Asp (4.5%), Glu (8.5%), Arg (6.2%), Lys (6.8%), Thr (9.6%), and Ser (6.8%). One Cys (0.4%) was also found in the CsDHN4 protein sequence, and the CsDHN3 protein was rich of polar amino acids including Glu (24.2%), Lys (18.2%), Ala (7.2%), Gly (6.8%), Ser (5.9%) and Asp (4.2%). No Cys or Trp was detected in the CsDHN4 protein sequence, and this protein possessed high percentages of polar amino acids such as Thr (13.0%), Gly (10.5%), Ser (9.3%), and Glu (8.6%). CsDHNs had a large number of disorder-promoting amino acids such as Glu, Gly, Lys, Ala, Thr and Ser (Fig. 1), suggesting that they are IDPs.

Fig. 1
figure 1

Prediction of intrinsically disordered residues of CsDHN protein sequences via PrDOS program. Disorder probability above 0.5 represents the disordered amino acid residues. Red: disordered residues. Black: ordered residues

Full size image

The DHN proteins occurring as IDPs in aqueous solutions could form randomly coiled unordered structures, which can maintain the large amount of hydrogen bonds with water molecules and the small percentage of intramolecular hydrogen bonds (Banerjee and Roychoudhury 2016; Tompa et al. 2005). In addition, K-segments have been proposed to resemble a class “A” amphipathic alpha helix, which thus enhances their amphipathic character in the interactions between protein and protein or protein and biomembrane (Abedini et al. 2017; Hanin et al. 2011; Koag et al. 2009). We then analyzed the secondary structure of CsDHN proteins by SOPMA tool. The CsDHN1 protein was composed by 40.39% of alpha helix, 12.94% of extended strand, 7.84% of beta turn, and 38.82% of random coil (Table 1; Fig. 2). Compared with CsDHN1, CsDHN3 had a larger amount of alpha helices (54.24%) and a relatively smaller amount of random coils (36.86%). The amounts of random coils in CsDHN2 and CsDHN4 were significantly larger than those in CsDHN1 and CsDHN3, while the amount of alpha helices was much smaller.

Fig. 2
figure 2

Protein secondary structural analysis of CsDHN proteins. The blue longest vertical lines represents alpha helical regions, the red vertical lines represent extended strand regions, the green vertical lines represent beta turn regions, and the purple shortest vertical lines represent random coil regions

Full size image

To further characterize the CsDHNs, the full-length deduced amino acid sequences were comparatively analyzed by ClustalW. The results revealed that the four DHN proteins belong to three subgroups: YnKn (CsDHN1), SKn (CsDHN3) and YnSKn (CsDHN2 and CsDHN4) (Fig. 3). The RRKK motif, which is considered as the nuclear localization signal (NLS), was present in 2 DHN proteins (CsDHN2 and CsDHN4). In addition, CsDHN2 and CsDHN4 contained a novel conserved motif EDDGXGG(R)1−3 or GXGGRRKK between the S-segment and the first K-segment (Fig. 2) as described previously (Abedini et al. 2017; Malik et al. 2017). Moreover, CsDHN2 harbored an additional dehydrin conserved motif (LXRXXS) which is phosphorylated by an Snf1-related kinase (SnRK2-10) (Vlad et al. 2008). This phenomenon was also found in some DHNs in other plant species including grape (Yang et al. 2012), pepper (Jing et al. 2016), and barley (Abedini et al. 2017).

Fig. 3
figure 3

Protein sequence alignment of CsDHN protein sequences. Y-, S- and K-segments are boxed by green, red and black, respectively. The conserved motif EDDGXGG(R)1−3 or GXGGRRKK and SnRK-10 sites are highlighted in red and blue, respectively. The NLS motifs (RRKK) are boxed by blue

Full size image

Phylogenetic analysis and gene structure of DHN gene family between cucumber and other plant species

To investigate the phylogenetic relationships of DHN genes in plants, a neighbor-joining phylogenetic tree was generated according to the alignment of DHN protein sequences from cucumber, Arabidopsis, tomato, and rice. These DHN proteins could be classified into five groups that correspond to the types of YnSKn, Kn, SKn, YnKn, and KnS (Fig. 4a). CsDHN1 protein clustered with AtLEA45 in YnKn group, while tomato and rice were lack of this type DHNs. CsDHN4 was closely related to AtLEA45, but it clustered together with CsDHN2 in YnSKn group. CsDHN3 shared a 51.7% identity in deduced amino acid sequence with SlLEA10, and clustered with the KnS-type DHN proteins (Fig. 4a). Interestingly, no Kn- and KnS-type DHN proteins were detected in cucumber and tomato. It could be observed that CsDHN2 and CsDHN3 in cucumber had close relationships with SlLEA7 and SlLEA10 in tomato, respectively, suggesting that these two pairs of proteins may be functionally similar.

Fig. 4
figure 4

Phylogenetic relationships and gene structure of DHN gene family from cucumber, Arabidopsis, tomato, and rice. a The phylogenetic tree of cucumber, Arabidopsis, tomato, and rice DHN proteins was constructed from a complete alignment of DHN proteins using MEGA 5.0 by the neighbor-joining method with 1000 bootstrap replicates. The groups are marked by colorful backgrounds. The protein names and corresponding accession numbers used to build the phylogenic tree are listed in Table S2. b Gene structure analyses were carried out by GSDS software. The blue boxes indicate upstream/downstream, the green boxes indicate CDSs, and the black lines indicate introns

Full size image

To study the gene structures of these DHN genes, we used GSDS software to compare the CDS sequences with the corresponding gDNA sequences of the DHN genes in cucumber, Arabidopsis, tomato, and rice. Most of these genes had only one intron (23 out of 28) (Fig. 4b). Among these genes, three had two introns (SlLEA7), one had ten introns (OsLEA22), and three had no intron (AtLEA33, OsLEA24, AtLEA18). The intronless genes were clustered in the Kn and KnS groups. In addition, some closely related DHN genes showed similar gene structures, such as OsLEA27 and OsLEA28, OsLEA24 and AtLEA18.

Chromosomal location and duplication of CsDHN genes

These four DHN genes were distributed on two cucumber chromosomes, except for one, which was found on an unassembled sequence scaffold000176 (Table 1; Fig. 5). Segmental and tandem duplication events are recognized as the primary reasons for DHN gene family expansion, and have been found in some plant species, such as Arabidopsis (Hundertmark and Hincha 2008), poplar (Liu et al. 2012), apple (Liang et al. 2012), barley (Abedini et al. 2017), and rice (Verma et al. 2017; Wang et al. 2007). In the current study, no segmental duplication event was identified. However, a pair of genes (CsDHN1 and CsDHN2) were arranged in tandem repeats on chromosome 4, implying that they may be tandemly duplicated genes.

Fig. 5
figure 5

Positions of CsDHN genes on cucumber chromosomes. Scale represents a 10 Mb chromosomal distance

Full size image

Analysis of CsDHNs promoters

To study the potential gene function, 1.0-kb promoter region of each CsDHN gene was downloaded from CuGI and analyzed by PlantCARE. The results showed that the cis-acting regulatory elements in CsDHNs promoters could be mainly categorized into hormone-responsive and stress-responsive elements (Table S3; Table 2). Five kinds of stress-responsive elements were found, and all of the CsDHN promoter sequences contained LTR motif (low temperature) and anaerobic induction element (ARE), which reflect plant responses to low temperature and anaerobic induction, respectively. Three other stress-related elements, including MBS and TC-rich repeats, were also present in a series of members. Meanwhile, we observed the wide presence of nine kinds of hormone-related cis-elements in the promoter regions of CsDHNs (Table S3; Table 2). These results suggested that CsDHNs are related to the responses to various abiotic stresses and hormones.

Table 2 Putative stress-responsive and hormone-responsive cis-elements in the promoter regions of CsDHN genes
Full size table

Additionally, several cis-acting regulatory elements involved in developmental processes, such as flavonoid biosynthetic genes regulation (MBSI), seed-specific regulation (RY-element), zein metabolism regulation (O2-site), circadian control (circadian), and endosperm-specific element (Skn-1 motif), were also found in several CsDHNs promoters (Table S3), implying that CsDHNs may have tissue-specific expression.

Tissue-specific expression of CsDHN genes

To understand the functions of CsDHNs, their expression profiles in cucumber tissues were determined by RT-PCR. As shown in Fig. 6, CsDHN2 and CsDHN3 had similar expression patterns, and their expression was detected in all five tissues, with higher expression levels in stem, flower and fruit, and significantly lower expression levels in root and leaf. Interestingly, CsDHN4 was only expressed in leaf at a relatively low level, and the expression of CsDHN1 was not observed in all five tissues (Fig. 6). The overlapping but different expression patterns of the CsDHN genes indicate that they may play various important roles in cucumber.

Fig. 6
figure 6

RT-PCR analysis of CsDHN genes in different tissues of cucumber. RNA was isolated from different tissues including root, stem, leaf, flower, and fruit

Full size image

Expression profiles of CsDHN gene under various abiotic stresses and ABA treatment

To determine the effect of environmental stresses on the expression of CsDHN genes, the transcripts of CsDHN genes under various abiotic stresses and ABA treatment were examined by qRT-PCR. Most CsDHN genes showed altered expression under these treatments, except for CsDHN1, whose expression was not detected (Fig. 7), implying that it is a pseudogene. All of the remaining three CsDHN genes were gradually up-regulated under heat stress, and their maximum transcripts were found at 12 h (10.99- to 757.74-fold) (Fig. 7a). Similar results were also observed under cold stress, except for CsDHN4, whose expression was down-regulated at 3 h, and subsequently reached to the maximum (12.72-fold) at 12 h (Fig. 7b). Upon NaCl treatment, the CsDHN genes showed diverse expression profiles (Fig. 7c). The transcription of CsDHN2 sharply increased at 3 h (527.44-fold), and continued to increase after 12 h treatment (898.08-fold). The expression of CsDHN3 was induced at 3 h (46.05-fold), followed by a decrease at 12 h (37.31-fold). The transcription of CsDHN4 was slightly induced at 3 h and became the highest (11.50-fold) at 12 h. All of the three CsDHN genes were induced at 3 h by PEG treatment, and the expression of CsDHN2 dramatically increased at 12 h (335.29-fold), while that of CsDHN3 and CsDHN4 slightly decreased at 12 h (Fig. 7d). Under H2O2 treatment, the expression of CsDHN2 was induced at 3 h and dramatically increased at 12 h (260.46-fold) (Fig. 7e). The transcription of CsDHN3 and CsDHN4 also reached the highest level at 12 h, while that of CsDHN3 was unchanged at 6 h, and CsDHN4 was sharply induced at 6 h (Fig. 7e). Under ABA treatment, all of the three CsDHN genes were notably induced at 3 h, followed by an observable decrease at 12 h (Fig. 7f).

Fig. 7
figure 7

qRT-PCR analysis of cucumber DHN genes in response to various abiotic stresses including heat (a), cold (b), PEG (c), NaCl (d), H2O2 (e) and ABA (f). Error bars indicate SD based on three biological replicates

Full size image

Discussion

In previous studies, genome-wide identification and characterization of DHN family have been performed in various plant species, such as Arabidopsis (10 members) (Hundertmark and Hincha 2008), Populus trichocarpa (11 members) (Liu et al. 2012), Vitis vinifera (4 members) (Yang et al. 2012), Malus domestica (12 members) (Liang et al. 2012), Hordeum vulgare (13 members) (Abedini et al. 2017; Karami et al. 2013), Solanum lycopersicum (6 members) (Cao and Li 2015), Pyrus pyrifolia (7 members) (Hussain et al. 2015), Capsicum annuum (7 members) (Jing et al. 2016), Oryza sativa ssp. japonica (8 members) (Verma et al. 2017), and Solanum tuberosum (5 members) (Charfeddine et al. 2017). In the current study, a total of four DHN family genes were identified from cucumber (Table 1). The number of DHN genes is not proportional to the size of genomes in the above mentioned plant species. This phenomenon might be due to gene duplication, which plays an important role in the evolution of gene families in different plants (Flagel and Wendel 2009). In addition, one tandem duplication event (CsDHN1 and CsDHN2) was found on chromosome 4 (Fig. 5), which might have made contribution to the expansion of CsDHN gene family in cucumber.

The four CsDHN proteins, respectively, consisted of 255, 177, 236, and 162 amino acids, and showed common features of DHN proteins. For example, they were rich of glycine (Gly) and polar amino acids, and were strongly hydrophilic (Table 1). In addition, the distribution of disorder residues in CsDHN protein sequences implied that they are IDPs (Fig. 1). Protein secondary structure analysis by SOPMA tool revealed that CsDHN proteins possess a high percentage of random coils (Table 1; Fig. 2), which can maintain a large amount of hydrogen bonds with water molecules and a small amount of intramolecular hydrogen bonds. These results further confirm that CsDHNs are IDPs, which is in accordance with DHN protein characteristics (Charfeddine et al. 2017).

Among the five types of DHNs, only three types were found in cucumber, including two YnSKn-type (CsDHN2 and CsDHN4), one YnKn-type (CsDHN1), and one SKn-type (CsDHN3) DHNs, while KnS- and Kn-type DHNs were not detected (Figs. 3, 4a). CsDHN2 and CsDHN4 are YnSKn-type DHNs, and contain the RRKK motif and a novel conserved motif between the S-segment and the first K-segment, which was previously identified as a phosphorylation site (Lv et al. 2014). In addition, CsDHN2 harbors an extra conserved motif (LXRXXS) that can be phosphorylated by an Snf1-related kinase (Vlad et al. 2008). Phosphorylation of DHNs was suggested to play an important role in functionally regulating stressed plant cells and modulating the membrane binding of DHNs (Eriksson et al. 2011; Yang et al. 2012). Moreover, the YnSKn-type CsDHNs are basic in character, and were suggested to combine to negatively charged membrane with a high affinity (Jing et al. 2016; Koag et al. 2009).

The four CsDHNs (except for CsDHN1) show closer phylogenetic relationships with the DHNs in Arabidopsis and tomato than with those in rice (Fig. 4a). Gene structure analysis revealed that each of the four CsDHNs harbors one intron (Fig. 4b), which is in agreement with the previous reports in pepper (Jing et al. 2016). The DHN genes contain very few introns, with most of them harboring one intron. This phenomenon may be caused by a conservative evolution pattern, and the intron length can affect the functional divergence in plants (Guo et al. 2014). A previous study has also reported that plants tend to retain more genes with no introns or short introns (Mattick and Gagen 2001). In addition, some closely related DHN genes exhibit similar gene organization patterns and demonstrate a conserved evolutionary pattern, implying that they may have similar functions.

Gene expression patterns are important for understanding gene functions, and the expression of DHN genes has been reported to be at high levels in all vegetative tissues of many plant species, including apple (Liang et al. 2012), pepper (Jing et al. 2016), and rice (Verma et al. 2017). In this study, CsDHN2 and CsDHN3 were expressed in all five detected tissues (Fig. 6), which is consistent with previous studies of other plant species, indicating that they may be involved in the growth and development of cucumber. Notably, CsDHN4 was only expressed in leaf, implying that it may play a particular role in leaf. However, the expression of CsDHN1 was not detectable in any tissue under the experimental conditions of this study (Fig. 6), suggesting that it may be a pseudogene. Such results were also obtained in some other plant species, such as VvDHN3 (Yang et al. 2012) and CaDHN6 (Jing et al. 2016). The spatial variations in the expression levels of CsDHN genes in different organs suggest that CsDHNs may participate in various processes of cucumber growth and development.

Many studies have reported the induction of DHN gene expression by various abiotic stresses in plants, suggesting that DHNs may play a positive role in plant response and adaptation to abiotic stresses (Banerjee and Roychoudhury 2016; Charfeddine et al. 2017; Jing et al. 2016; Rodziewicz et al. 2014; Verma et al. 2017). In this study, the expression of all four CsDHN genes except for CsDHN1 was induced in response to stresses such as heat, PEG, salt and cold (Fig. 7), which is in agreement with the findings in previous studies. Phylogenetic analysis showed that CsDHN3 and SlLEA10 were in the same cluster of the SKn subgroup (Fig. 4a), and were both up-regulated under salt and drought treatments (Fig. 7) (Cao and Li 2015). In addition, CsDHN genes can be induced in response to stresses, but they may have different expression patterns. For example, under cold treatment, the transcription of CsDHN4 sharply decreased at 3 h, followed by an increase at 24 h, while that of CsDHN2 and CsDHN3 gradually increased (Fig. 7b). Similar expression patterns were also observed under H2O2 treatment (Fig. 7e). Upon NaCl treatment, the transcription of CsDHN4 had no significant change at 3 h, while that of CsDHN2 and CsDHN3 was significantly up-regulated; the expression of CsDHN2 and CsDHN4 remarkably increased at 12 h, while that of CsDHN3 decreased (Fig. 7d). Under PEG treatment, the expression of CsDHN2 was continuously up-regulated at 12 h, while that of CsDHN3 and CsDHN4 had no significant change (Fig. 7c). These findings indicate that they play various important roles in cucumber, and different mechanisms might have been formed in cucumber to adapt to various abiotic stresses. The diverse expression patterns could be owing to the presence of cis-elements related to stress in CsDHN promoters. Obvious differences were found in the abundance and distribution of stress-responsive cis-elements in the promoters of four CsDHNs (Table 2). For example, all the CsDHN promoters had up to two cold-responsive elements (LTR), which are related to low-temperature response. CsDHN2 and CsDHN4 possessed four and two TC-rich repeats, respectively (Table 2). Each of CsDHN1, CsDHN2 and CsDHN4 promoters contains one MBS, while CsDHN3 promoter does not contain any MBS. CsDHN1 promoter has two HSE, and CsDHN1 promoter includes one HSE (Table 2); however, CsDHN2 and CsDHN4 were remarkably induced by heat stress (Fig. 7a), implying that some other heat-responsive elements are present in their promoter regions. Further research is required to determine whether and how the cis-elements of CsDHNs regulate the response to various abiotic stresses.

Previous studies have revealed that most of DHN genes are induced under ABA treatment, and several members were proven to play important roles in plant resistance to stress via ABA-dependent pathway (Cao et al. 2017; Lv et al. 2017). In this study, the expression of CsDHN2, CsDHN3 and CsDHN4 was up-regulated by ABA, and they harbored 2–6 ABA-responsive elements (ABREs) (Table 2). The number of ABREs was correlated with the transcripts of DHN genes in response to ABA (Fig. 7f). In addition, SnRK2-10, which positively regulates ABA-mediated signaling, may be related to the phosphorylation of S-segments in DHNs (Vlad et al. 2008). These results reveal that CsDHNs are regulated through an ABA-dependent signal pathway.

Conclusions

In this study, a complete analysis of the DHN genes in cucumber was carried out, including genome organization, gene structure, phylogenetic relationship, and expression pattern analyses. Our results would help the functional characterization of CsDHN genes, and facilitate future study of their response to various abiotic stresses.

Author contribution statement

YZ, LH, and SL conceived and designed the experiments. YZ, LH, SX, LJ, and SL carried out the experiments. YZ, LH, and SL analyzed the data. YZ and SL wrote the paper. SL edited the manuscript, secured the funds to support this research.