Introduction

Melon (Cucumis melo L.), a cucurbitaceous fruit, is ninth among horticultural crops in terms of worldwide total production (Dasgan and Koc 2009). Since plants are sessile organisms, they are continuously faced with a range of environmental stresses, including diseases such as powdery mildew as well as high salinity, drought, and high and low temperatures in the natural environment. Both field- and greenhouse-cultivated cucurbit crops are threatened by the powdery mildew (PM) fungus Podosphaera xanthii as well as multiple abiotic stresses that hinder plant growth, development and yield in many parts of the world (Sitterly 1978; Jarvis et al. 2002; Kaya et al. 2007; Dasgan and Koc 2009; Lim and Lee 2014). However, plants have evolved physiological, biochemical, and molecular mechanisms to counteract the deleterious effects of biotic and abiotic stresses (Zhu 2002; Yamaguchi-Shinozaki and Shinozaki 2006; Lee et al. 2013). In addition, various conventional agricultural practices involve management approaches to address these stresses. Application of fungicide can reduce powdery mildew disease incidence in melon by up to 70% (Choi et al. 2011). However, PM fungus has a high potential to develop fungicide resistance (McGrath 2001). Therefore, it is desirable to develop stress-resistant plant lines as a control strategy to reduce the use of pesticides in agricultural practices.

The Mildew Locus O (MLO) gene first discovered in the barley belongs to a unique plant-based family of seven transmembrane domain (TM) proteins that are characterized by a C-terminal calmodulin (CaM)-binding domain (CaMBD) and an extracellular N-terminus (Büschges et al. 1997; Devoto et al. 1999; Piffanelli et al. 2002; Kim et al. 2002a). The MLO proteins are found not only in higher plants but also in certain mosses (Devoto et al. 1999, 2003). Although the biological functions of MLO proteins remain mostly unknown, some family members are believed to have a role in modulating host responses to the phytopathogenic powdery mildew fungus (Büschges et al. 1997). Homozygous natural mutant mlo alleles of barley confer durable broad-spectrum resistance to the biotrophic powdery mildew fungus Blumeria graminis f. sp. hordei (Büschges et al. 1997). In tomato and pepper, loss-of function mutations in SlMLO1 and CaMLO2 provide resistance to the powdery mildew pathogens Oidium neolycopersici and Leveillula taurica, respectively (Bai et al. 2008; Zheng et al. 2013a, b). In pea (Pisum sativum), the recessively inherited powdery mildew resistance in er1 plants is reminiscent of that of loss-of-function mutations of the PsMLO1 gene first reported in 1948, which has been used worldwide for breeding purposes as natural resistance source against the powdery mildew pathogen Erysiphe pisi (Harland 1948; Humphry et al. 2011; Pavan et al. 2013). In arabidopsis, homozygous T-DNA insertion into the AtMLO2 gene causes significantly reduced susceptibility to Golovinomyces orontii, although complete powdery mildew resistance requires the closely related genes AtMLO6 and AtMLO12 also to be mutated (Consonni et al. 2006). Hence, these three arabidopsis orthologues show partial functional redundancy. Overall, the experimental evidence suggests that MLO genes are related to powdery mildew susceptibility within the respective species (Jørgensen 1992; Consonni et al. 2006), and are induced in response to powdery mildew inoculation (Piffanelli et al. 2002; Chen et al. 2006; Bai et al. 2008). Indeed, overexpression of barley HvMLO led to powdery mildew super-susceptibility, with almost every attacked barley cell successfully colonised by the pathogen (Kim et al. 2002b).

MLO genes are widely distributed in number of different plant species. A total of 15, 12, 17, 21, 39, 13, 16, 14, 18, and 38 MLO genes were identified in arabidopsis, rice, grape, apple, soybean, cucumber, melon, watermelon, zucchini and cotton, respectively, over the last two decades (Devoto et al. 2003; Liu and Zhu 2008; Feechan et al. 2009; Pessina et al. 2014; Deshmukh et al. 2014; Schouten et al. 2014; Iovieno et al. 2015; Wang et al. 2016; Howlader et al. 2016). Apart from defense response roles, MLO genes also have vital roles in plant development (Piffanelli et al. 2002; Chen et al. 2006; Feechan et al. 2008; Kim and Hwang 2012; Lim and Lee 2014). These genes are involved in biological processes including leaf senescence (Wolter et al. 1993; Kumar et al. 2001; Piffanelli et al. 2002), indicating that MLO genes could be responsive to various environmental stimuli.

MLO genes have been intensively studied in both monocots and dicots such as barley (Piffanelli et al. 2002), arabidopsis (Chen et al. 2006), grapes (Feechan et al. 2008), apple (Pessina et al. 2014), pepper (Lim and Lee 2014; Zheng et al. 2013a, b), rice (Liu and Zhu 2008). Recently, Iovieno et al. (2015) characterized the phylogenic relatedness of MLO gene family in melon, watermelon and zucchini. The authors reported that ‘dicot-specific clade V genes’ associated with PM susceptibility are highly expressed in watermelon (Cucurbita lanatus cv. Sugar Baby) within 24 h after inoculation with Podosphaera xanthii. However, MLO genes outside of clade V have been reported to be associated with PM susceptibility in apple (Pessina et al. 2014). A recent study quantified the expression 14 CmMLO candidate genes at a single time point, 18 days after inoculation, when the severe disease symptoms were observed (Howlader et al. 2016). Despite this progress, systematic analysis of the expression of melon MLO genes from all clades under P. xanthii infection throughout disease development and under abiotic stresses is lacking.

Eight different P. xanthii races are responsible for PM in melon (Kim et al. 2016); therefore, it is important to understand the race- or isolate-specific behavior of melon genes during disease progression. Such data are essential for developing PM resistance through either reverse genetic engineering or mutation breeding in Cucurbitaceae family crops. Knowledge regarding abiotic stress responsiveness of MLO genes is also important; some arabidopsis MLO genes are differentially expressed under cold, drought, salt and wound stress conditions, further suggesting that MLO genes are involved in the growth and development of plants (Chen et al. 2006). In melon, a previous report showed that one MLO gene is expressed under cadmium abiotic stress (Cheng et al. 2012).

The major objectives of this study were to characterize C. melo MLO proteins and the systemic expression of MLO genes in response to two races (Race 1 and DH487) of P. xanthii during disease development and also in response to four abiotic stress (heat, cold, drought, salt), and one phytohormone (ABA) treatments. We investigated 14 MLO genes in C. melo in silico and then a selected melon line was treated with different biotic and abiotic stresses at the seedling stage to explore changes in relative expression of the selected genes during disease developmental stages and under abiotic stresses. The findings of the study lay the groundwork for possible utilization of selected MLO genes in developing Cucurbitaceae family genotypes that will be PM resistant in various environmental conditions.

Results

In Silico CmMLO Sequence Analysis

Howlader et al. (2016) identified 14 MLO genes and Iovieno et al. (2015) identified 16 MLO genes for C. melo. For further in silico analysis we chose the 14 MLO genes reported by Howlader et al. (2016) and designated as CmMLO1 to CmMLO14. The predicted sizes of the 14 encoded CmMLO proteins ranged from 150 to 582 amino acids (~16.11–66.37 kDa), and their predicted isoelectric points varied from 7.25 to 9.48 (Table 1).

Table 1 In silico analysis of MLO genes in C. melo collected from the Melonomics databasea
Full size table

Analysis of protein domain organization showed that five CmMLO proteins (CmMLO3, CmMLO10, CmMLO12, CmMLO13, and CmMLO14) contained 7 TM (TM1 through TM7) domains (Table S1). SMART software further revealed that CmMLO5 had an additional TM domain between the TM2 and TM3 domains (Table S1). CmMLO4 and CmMLO6 contained all seven TMs domain found in the other proteins, but also had one and two additional TM domains, respectively. The remaining three CmMLO proteins included various degrees of TM domain conservation (Table S1).

Phylogenetic Classification of CmMLO

A phylogenetic tree based on MLO protein sequences showed relatedness between MLO proteins. A total of 59 MLO homologs from different crop species including melon was used to infer phylogenetic distances among them. The MLO proteins were distributed in six phylogenetic clades (Fig. 1), designated clades I to VI based on the position of arabidopsis and monocot MLO homologs (Devoto et al. 2003; Feechan et al. 2008; Liu and Zhu 2008). The CmMLO proteins were distributed in five of the six clades (Fig. 1). Phylogenic clade I comprised 13 MLO homologs, three of which were annotated in Arabidopsis thaliana (AtMLO4, AtMLO11 and AtMLO14), four in Cucumis melo (CmMLO7, CmMLO8, CmMLO9 and CmMLO10), four in Cucumis sativus (CsaMLO5, CsaMLO6, CsaMLO7 and CsaMLO9) and two in Oryza sativa (OsMLO4 and OsMLO11) (Fig. 1). Clade II contained 14 MLO proteins, annotated in arabidopsis (AtMLO1, AtMLO13 and AtMLO15), C. melo (CmMLO1, CmMLO11and CmMLO12), C. sativus (CsaMLO3 and CsaMLO10), O. sativa (OsMLO2, OsMLO5, OsMLO8, OsMLO9 and OsMLO10) and Hordeum vulgare (HvMLO3). Clade III contained 3 MLO homologs (CmMLO4, CmMLO6 and CmMLO14) coupled with five arabidopsis proteins (AtMLO5, AtMLO7, AtMLO8, AtMLO9 and AtMLO10), three C. sativus proteins (CsaMLO4, CsaMLO12, and CsaMLO13) and two O. sativa proteins (OsMLO1 and OsMLO12) (Fig. 1). Clade IV comprised four MLO proteins from monocot species only (OsMLO3, OsMLO6, HvMLO1 and HvMLO2). Clade V had all 10 dicot MLO homologs where three proteins each from C. melo (CmMLO3, CmMLO5 and CmMLO13) and C. sativus (CsaMLO1, CsaMLO8 and CsaMLO11) clustered with A. thaliana (AtMLO2, AtMLO6 and AtMLO12) and Solanum lycopersicum MLO (SlMLO1) proteins that are experimentally established as PM susceptibility factors (Bai et al. 2008; Consonni et al. 2006). Clade VI contained 3 MLO homologs one from each of C. melo (CmMLO3), C. sativus (CsaMLO2) and A. thaliana (AtMLO3) (Fig. 1). A single monophyletic outgroup contained only O. sativa homolog OsMLO7.

Fig. 1
figure 1

Phylogenetic relationships of melon MLO proteins (CmMLO) with MLO proteins of other plant species. Arabidopsis (AtMLO), cucumber (CsaMLO), tomato (SlMLO1), barley (HvMLO), and rice (OsMLO) sequences were included in the unrooted tree, which was generated using the maximum likelihood method in MEGA 6.06 software (Whelan and Goldman 2001). Six evolutionary clades (clades I to clade VI) were named according to the classification of Devoto et al. (2003) and based on the position of arabidopsis and monocot MLO homologs. Reported MLO proteins functionally characterized as powdery mildew susceptibility genes are highlighted in black bold lettering. Numbers at each node represent bootstrap support values (out of 1000 replicates). CmMLO proteins are in red

Full size image

Motif Distribution

There were 15 conserved motifs in melon MLO proteins, ranging between 13 and 50 amino acids in length (Table S2). We further analyzed motif distributions, finding that three CmMLO proteins (CmMLO4, CmMLO6 and CmMLO10) contained all twelve conserved motifs whereas CmMLO1 and CmMLO2 contained only two (Fig. 2). CmMLO14, CmMLO3, CmMLO5 and CmMLO8; CmMLO9, CmMLO7 and CmMLO13, and CmMLO11 contained 11, 10, 9, 8, and 7 motifs, respectively (Fig. 2). In addition, CmMLO proteins of clade I all contained motif 11, which is not present in any other phylogenetic clade (Fig. 2). Similarly, although clade III and clade IV members exclusively contained motif 9, clade III members also had motif 14, which was not present in clade IV CmMLO proteins (Fig. 2). We predict that these clade-specific motifs of CmMLO proteins might have significance in stress environments. Notably, CmMLO1 and CmMLO2 contained the same motif although they are distributed in clade II and clade VI, respectively (Fig. 2).

Fig. 2
figure 2

Schematic representation of motifs in C. melo MLO proteins. Different motifs are indicated by different colors, and the names of all CmMLO members are shown on the left, along with their phylogenetic relatedness

Full size image

The deduced amino acid sequences of 14 CmMLO shared high sequence similarity with MLO proteins of reference species including Arabidopsis thaliana, Vitis vinifera, Populus trichocarpa and Zea mays (Table 2). The sequence identity ranged from 49 to 84% with the reference species (Table 2). In addition, the nucleotide sequence similarity between the CmMLO genes ranged from 22 to 72% (Table S3).

Table 2 Protein homology analysis of 14 MLO genes in C. melo
Full size table

Tissue-Specific Expression Analysis

We investigated the organ-specific expression patterns of all 14 CmMLO genes using roots, stems, leaves, and flowers of C. melo line ‘SCNU1154’ by RT-qPCR (Fig. 3). The maximum expression (upregulated up to 20-fold) was observed in the root tissues for CmMLO2, 3, 4, 11 and 13 (Fig. 3). CmMLO5, 6, 10 and 14 had the highest expression (upregulated up to 18-fold) in flower tissues (Fig. 3). CmMLO1, 8, 9 and 12 were maximally expressed in leaf tissues, whereas CmMLO7 was expressed most highly in stem tissues (Fig. 3).

Fig. 3
figure 3

Reverse-transcription quantitative PCR expression analysis of CmMLO genes in different tissues of C. melo ‘SCNU1154’ plants. Relative gene expression levels were normalized to actin transcript values. The error bars represent the standard error of the mean of three independent replicates

Full size image

P. xanthii-Induced Disease Development and Gene Expression in C. melo

When we inoculated C. melo leaves with P. xanthii Race 1 or DH487, visible powdery mildew disease symptoms were first observed at 5 d post inoculation (dpi) (Fig. 4). The disease area increased rapidly and reached about one-third of leaf area at 10 dpi (Fig. 4). All 14 CmMLO genes were differentially expressed under both DH487 and Race 1 infection compared to control (Fig. 5). Notably, all 14 genes were markedly downregulated, from 2- to 10-fold, at 12 h by both races compared to control. Race 1-specific up-regulation, from 1.4- to 2.2-fold, of CmMLO 4, 6, 7 and 9 genes were observed at 6 h after inoculation with Race 1 compared to the control (Fig. 5). Upregulation between 1.2- and 3.0-fold of CmMLO1, 2, 3, 11, 12, 13 and 14 by both races was also observed at 10 dpi compared to the control. Notably, CmMLO6 and CmMLO7 were constantly downregulated from 12 h to 10 d after inoculation compared to the control (Fig. 5).

Fig. 4
figure 4

Infection of C. melo ‘SCNU1154’ plants by two different races of Podosphaera xanthii, Race 1 and DH487. Disease symptoms of the 5th inoculated leaves were compared with control at different days post inoculation. The white circle indicates the extent of powdery mass development on the inoculated 5th leaf, numbered from the ground level

Full size image
Fig. 5
figure 5

Reverse-transcription quantitative PCR expression analysis of CmMLO genes in response to two Race 1 and DH487 of P. xanthii fungus inoculation of C. melo ‘SCNU1154’ plants. Relative gene expression levels were normalized to actin transcript values. Error bars represent the standard error of the mean of three independent replicates

Full size image

Gene Expression under Abiotic Stress Conditions

Gene Expression under Drought Stress

Under drought treatment, all 14 CmMLO genes were significantly (p < 0.05) upregulated compared to the control (Fig. 6). Maximal upregulation was observed for CmMLO4, 7, 8 and 12 at 1 h; for CmMLO1, 3 and 11 at 3 h; for CmMLO2, 5, 9 and 13 at 6 h, and for CmMLO6, 10 and 14 at 24 h post treatment (Fig. 6). The relative expression of upregulated genes varied between 1.7- and 12.5- fold compared to the control plants (Fig. 6).

Fig. 6
figure 6

Reverse-transcription quantitative PCR expression analysis of CmMLO genes in response to drought stress treatment of C. melo ‘SCNU1154’ plants. Relative gene expression levels were normalized to actin transcript values. The error bars represent the standard error of the mean of three independent replicates

Full size image

Gene Expression under Salt Stress

Similar to drought, the salt treatment significantly (p < 0.05) upregulated 13 (CmMLO1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 and 13) genes compared to the control (Fig. 7). The exception, CmMLO14, showed no differences in expression under salt stress (data not shown). CmMLO1, 2, 3, 5 and 10 exhibited maximum upregulation at 48 h, CmMLO12 at 24 h, CmMLO4, 6, 7, 8 and 11 at 12 h, CmMLO13 at 6 h and CmMLO9 at 30 min after stress treatment (Fig. 7). The approximate relative expression ranged between 2.1- and 20.5- fold upregulation compared to the control plants (Fig. 7).

Fig. 7
figure 7

Reverse-transcription quantitative PCR relative expression analysis of CmMLO genes in response to salt stress treatment of C. melo ‘SCNU1154’ plants. Relative gene expression levels were normalized to actin transcript values. The error bars represent the standard error of the mean of three independent replicates

Full size image

Gene Expression under Heat Stress

Ten genes (CmMLO1, 2, 3, 4, 5, 6, 7, 11, 12 and 14) responded to heat stress (Fig. 8). CmMLO1, 4, 5, 6, 11 and 14 displayed significant (p < 0.05) upregulation compared to the control at 6 h after stress treatment and declined thereafter (Fig. 8). CmMLO2 and CmMLO7 were highly expressed at 1 h; whereas CmMLO12 was highly expressed at 3 h and CmMLO3 was at 12 h post treatment. The approximate relative expression of upregulated genes varied between 1.8- and 25.0- fold compared to control plants (Fig. 8). The other four genes, CmMLO8, 9, 10 and 13 exhibited no significant responses in heat-stressed plants compared to the control (data not shown).

Fig. 8
figure 8

Reverse-transcription quantitative PCR relative expression analysis of CmMLO genes in response to heat stress treatment of C. melo ‘SCNU1154’ plants. Relative gene expression levels were normalized to actin transcript values. The error bars represent the standard error of the mean of three independent replicates

Full size image

Gene Expression under Cold Stress

Notably, under cold treatment, only 4 (CmMLO7, 8, 9, and 12) genes were significantly (p < 0.05) upregulated compared to the control (Fig. 9). CmMLO7, 8, 9, and 12 exhibited maximum upregulation at 48 h of treatment (Fig. 9). The approximate relative expression ranged between 2.0- and 4.0- fold upregulation compared to control plants (Fig. 9). The other ten genes, CmMLO1, 2, 3, 4, 5, 6, 10, 11, 13 and 14, showed no significant response to cold stress treatment (data not shown).

Fig. 9
figure 9

Reverse-transcription quantitative PCR relative expression analysis of CmMLO genes in response to cold stress treatment of C. melo ‘SCNU1154’ plants. Relative gene expression levels were normalized to actin transcript values. The error bars represent the standard error of the mean of three independent replicates

Full size image

Gene Expression under ABA Treatment

Under ABA treatment, 12 (CmMLO2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 13 and 14) genes were significantly (p < 0.05) upregulated (Fig. 10). CmMLO2, 4, 6, 7, 8, 9, 10, 13 and 14 had low expression at early stages of treatment, but eventually peaked at 12 h and afterwards declined. CmMLO3 demonstrated an early maximal response at 30 min, whereas CmMLO11 peaked at 1 d; and CmMLO5 peaked at 6 h (Fig. 10). The approximate relative expressions ranged from 1.7- and 6.7- fold upregulation compared to the control plants (Fig. 10). The remaining two genes, CmMLO1 and CmMLO12 displayed no significant responses in treated plants compared to the controls (data not shown).

Fig. 10
figure 10

Reverse-transcription quantitative PCR relative expression analysis of CmMLO genes in response to exogenous ABA treatment of C. melo ‘SCNU1154’ plants. Relative gene expression levels were normalized to actin transcript values. The error bars represent the standard error of the mean of three independent replicates

Full size image

Analysis of cis-Response Element in the Promoters of CmMLO Genes

The promoter regions comprising about 1.5-kb upstream of the start codons of the 14 CmMLO genes were analyzed to identify putative cis-acting regulatory elements. The identified cis-elements in CmMLO genes included those involved in growth and development, phytohormone signaling, defense and stress responses such as: cis-regulatory elements related to meristem expression (CAT-box), meristem-specific activation (CCGTCC-box/dOCT), endosperm expression (Skn-1-motif/GCN4-motif), differentiation of the palisade mesophyll cells (HD-Zip 1), control of leaf morphology development (HD-Zip 2), light responsiveness (G-box/ACE), circadian control (Circadian), defense and stress responsive elements (TC-rich repeats), fungal elicitor responsive element (Box-W1), elicitor responsive element (ELI-box3), heat-stress element (HSE), low temperature responsive element (LTR), drought inducibility element (MBS), abscisic acid responsive element (ABRE), salicylic acid-responsive element (TCA-element), MeJA-responsive element (CGTCA-motif/ TGACG-motif), and ethylene-responsive element (ERE) (Table S4). Eleven CmMLO genes bearing biotic elicitor responsive cis-acting regulatory elements (CmMLO3, 4, 5, 6, 7, 8, 9, 11, 12, 13 and 14) were either upregulated or downregulated under P. xanthii infection, in this study (Fig. S1). Seven genes bearing drought responsive cis-acting regulatory elements (CmMLO3, 4, 5, 7, 10, 13 and 14), eight genes bearing heat responsive cis-acting regulatory elements (CmMLO2, 4, 5, 6, 7, 11, 12 and 14), one gene bearing cold responsive cis-acting regulatory elements (CmMLO7) and five genes bearing ABA responsive cis-acting regulatory elements (CmMLO3, 5, 6, 7 and 10) were upregulated at any of the points under each stress (Fig. S2a-d).

Discussion

Gene Selection, Molecular Characteristics and Phylogenic Distribution of MLO Homologs

We originally extracted 15 MLO proteins from the melon genome database ‘Melonomics’ using the keyword ‘MLO’. Among the 15 CmMLO proteins, 12 exactly matched those reported in Howlader et al. (2016) and Iovieno et al. (2015) and contained a variable number of the transmembrane domains that are a characteristic feature of barley MLO proteins (Büschges et al. 1997). Two other TM domain-bearing MLO genes CmMLO1 and CmMLO11 matched those reported in Howlader et al. (2016) but were found by Iovieno et al. (2015). Three other CmMLO genes [accessions MELO3C025760 (CmMLO14 in Iovieno et al. 2015), MELO3C005037 (CmMLO15 in Iovieno et al. 2015), and MELO3C007252 (CmMLO16 in Iovieno et al. 2015)] do not encode TM domains; therefore, we did not include them in further analysis. A fourth gene reported by by Iovieno et al. (2015) with accession number ACX55085 does encode TM domains but was not available in ‘Melonomics’ database and was therefore not among our selected genes.

Almost all of the predicted melon CmMLO proteins had amino acid lengths comparable to those of arabidopsis AtMLO homologs, ranging from 460 to 593 residues (Devoto et al. 2003). The exceptions were CmMLO1 and CmMLO2, which might represent incomplete sequences as these two proteins were much shorter in length, at 150 and 195 bp, respectively. Therefore, obtaining full length cDNA sequences of these two genes through cloning and sequencing are required for further characterization.

Inference of phylogenetic relatedness between MLO proteins revealed that the dicot-specific clade V, which includes three melon homologs CmMLO3, CmMLO5 and CmMLO13, and the monocot-specific clade IV contain MLO homologs experimentally shown to be involved in PM susceptibility (Büschges et al. 1997; Devoto et al. 2003; Consonni et al. 2006; Bai et al. 2008; Zheng et al. 2013a, b; Humphry et al. 2011). Interestingly, our phylogenetic analysis also uncovered the presence of one additional clade (clade VI) not reported in earlier MLO studies in melon (Fig. 1) (Iovieno et al. 2015). Along with CmMLO2, MLO homologs from Rosaceae crops viz., peach (Prunus persica), strawberry (Fragaria vesca), and apple (Malus domestica); and grapevines (Vitis vinifera) clustered together with an arabidopsis (AtMLO3) homolog in clade VI, phylogenic reconstruction is required for clade distribution in melon (Feechan et al. 2008; Pessina et al. 2014). Clearly, the existence of such additional clade in melon deserves further investigation including a larger dataset of MLO proteins.

Expression Analysis of CmMLO Genes in Different Organs

To carry out expression profiling of the CmMLO gene family, we performed RT-qPCR analysis of samples treated with biotic and abiotic stresses. The differential expression of the 14 CmMLO genes in the tissues of different organs tested here indicated that although CmMLO proteins are constitutively expressed, overlapping expression or functional redundancy likely exists among CmMLO members (Fig. 3). The promoters in almost all CmMLO genes except CmMLO1 and CmMLO10 contained various growth and development responsive cis-elements, consistent with CmMLO proteins being expressed in a range of tissues (Table S4). Therefore, CmMLO proteins might play differential roles in growth and development processes of C. melo. These results are consistent with those for MLO proteins in arabidopsis (Chen et al. 2006), grape (Feechan et al. 2008), apple (Pessina et al. 2014), melon (Cheng et al. 2012), and cucumber (Wang et al. 2016). However, we did not find any clade-specific expression pattern among the CmMLO proteins in different organs (Fig. 3).

Expression Analysis of CmMLO Genes under Biotic and Abiotic Stresses

Plants exposed to various environmental stresses develop characteristic physiological, biochemical, and molecular mechanisms to overcome the harmful effects of stress conditions (Lim and Lee 2014). In response to biotic and abiotic stresses, some specific stress-inducible genes are transcriptionally activated in the plants (Cheong et al. 2007; Kilian et al. 2007). Biotic and abiotic stress-responsive genes often contain stress-responsive cis-elements in their promoters corresponding to each stress factor. Here, we analyzed 14 CmMLO genes from C. melo, most of which except CmMLO1, 2, and 10 had a predicted defense and stress-responsive cis-acting element (TC-rich repeats) (Table S4). CmMLO3, 4, 5, and 13 had a fungal elicitor-responsive element (Box-W1); CmMLO8 contained elicitor-responsive element (ELI-box3); CmMLO2, 4, 5, 6, 7, 8, 9, 11, 12, 13, and 14 contained heat-stress element (HSE); CmMLO1, 2, 5, 7, 13, and 14 contained low temperature responsive element (LTR); CmMLO3, 4, 5, 7, 10, 13, and 14 contained a drought inducing element (MBS); CmMLO3, 5, 6, 7, 10, and 12 contained abscisic acid responsive element (ABRE) (Table S4). Conversely, the corresponding putative cis-elements detected were not closely related to CmMLO8, 9, and 13 genes under heat stress; CmMLO1, 2, 5, 13, and 14 under cold stress; CmMLO12 under ABA stress conditions (Figs. 8, 9 and 10 and Table S4).

Expression Analysis of CmMLO Genes under P. xanthii Stresses

Powdery mildew fungus negatively regulates plant antifungal defense reactions by using MLO proteins as an ‘invasion door’ for successful penetration before haustorium formation (Panstruga 2005). Susceptible MLO genes are upregulated at early stages of PM fungus inoculation in barley (Piffanelli et al. 2002), tomato (Bai et al. 2008), grape (Feechan et al. 2008; Winterhagen et al. 2008) and pepper (Zheng et al. 2013a, b). A previous report detailed the expression patterns of all 14 CmMLO genes in response to seven different isolates of powdery mildew (PM) fungus P. xanthii at the later stages (18 dpi) of disease development but did not explore the expression behavior of the CmMLO genes at the early stages of fungal invasion (Howlader et al. 2016). Therefore, in this study, we systematically investigated the expression patterns of all 14 CmMLO genes at different time points in response to Race 1 and DH487 of P. xanthii. The observed upregulation of CmMLO8 and CmMLO10 (in response to both races); CmMLO7 and CmMLO9 (Race 1-specific) genes of clade I; and CmMLO14 (in response to both races), and CmMLO4 and CmMLO6 (Race 1-specific) of clade III at 6 h after inoculation suggested that the associated proteins might be utilized by PM fungi to invade into the host cell at the initial stage of inoculation (Fig. 5). All CmMLO genes, except CmMLO4 (which contains a fungal elicitor Box-W1 cis-element) and CmMLO10, contained defense and stress-responsive cis-acting elements (TC-rich repeats). It is possible that the PM fungus negatively modulates these elements and might use those MLO proteins to invade in to the host cell as an initial step of disease progress.

The functionally upregulated MLO homologs of barley, pea, pepper and tomato in clade V appear to be involved in PM and we therefore refer to those candidates as PM susceptibility factors (black bold letter in Fig. 1) (Büschges et al. 1997; Devoto et al. 2003; Consonni et al. 2006; Chen et al. 2006; Bai et al. 2008; Zheng et al. 2013a, b; Humphry et al. 2011). Three mutant MLOs in this clade of A. thaliana are necessary to achieve a fully resistant phenotype (Consonni et al. 2006; Reinstädler et al. 2010; Pavan et al. 2011; Zheng et al. 2013a, b), which highlights that some of MLOs clustered with functional MLOs in clade V might play a role in PM resistance due to loss-of-function mutation. In addition, accumulating evidence also indicates that MLO genes outside of clade V are also upregulated under PM fungus challenge (Pessina et al. 2014). Therefore, additional attempts to identify PM susceptibility genes might be essential. An upregulation of two pathogen-dependent homologs, CmMLO3 and CmMLO13, in clade V against Race 1 and DH487 at 10 dpi indicated that higher expression of those genes might be associated with disease appearance for both races and might be candidates for PM susceptibility factors (Fig. 5). Notably, CmMLO homologs in clade V contained the fungal elicitor-responsive cis-acting Box-W1 elements in their promoter region (Table S4). However, the absence of the fungal elicitor Box-W1 cis-acting element in the upregulated CmMLO2 homolog in clade VI; CmMLO1, CmMLO 11 and CmMLO12 in clade II; and CmMLO14 in clade III indicated that upregulation of these genes might be elicited via other unknown cis-elements and hence PM susceptibility factor(s) might be distributed to clades other than V. Thus, PM susceptibility factors may not be confined to the dicot-specific clade V, and simple clustering might not be sufficient for recognition of a gene as a susceptibility factor.

Due to the loss of function, mutants of SlMLO1 in tomato (Bai et al. 2008), CaMLO2 in pepper (Zheng et al. 2013a, b), PsMLO1 in pea (Humphry et al. 2011), and triple mutants of AtMLO2, AtMLO6 and AtMLO12 in arabidopsis (Consonni et al. 2006) in clade V and natural mutant of HvMLO in barley in clade IV (Büschges et al. 1997) showed downregulation that provided durable broad-spectrum PM disease resistance against all known isolates of PM fungus. Lack of upregulation has also been observed in clade V MLO genes under PM fungus inoculation in apple (Pessina et al. 2014) and grapevine (Feechan et al. 2008). Consistent with a previous report (Howlader et al. 2016), we observed downregulated expression of CmMLO5 in clade V in response to both P. xanthii races at 12 h and 5 dpi, (Fig. 5). It is possible that CmMLO5 shows natural variation that serves to resist invading and developing PM in melon. The molecular functions of MLO proteins remain unknown, but mutation of conserved residues in MLO genes might be involved in inactivation of potential cis-acting fungal elicitor Box-W1 element (Piffanelli et al. 2002). In addition, mutated CmMLO genes might also be related to generating rapid cell wall-localized H2O2 bursts at the sites of attempted penetration by PM fungi beneath the epidermis of cells, which might lead to PM resistance (Piffanelli et al. 2002).

Expression Analysis of CmMLO Genes under Abiotic Stresses

Environmental stresses such as high and low temperatures, drought and salinity limit crop productivity worldwide. Understanding plant responses to these stresses is necessary for rational engineering of hardier crop plants. Accumulating evidence has revealed that apart from biotic stress responses, MLOs gene have additional functions in relation to various abiotic stress and exogenous phytohormone stimuli (Piffanelli et al. 2002; Feechan et al. 2008; Konishi et al. 2010; Shen et al. 2012; Ablazov and Tombuloglu 2016). In addition, some MLO genes associated with PM susceptibility in barley, arabidopsis and grapes are also induced by abiotic stresses including low temperatures, drought, salinity and wounding, which indicates that there are differential response patterns for these genes under different environmental stress conditions (Piffanelli et al. 2002; Chen et al. 2006; Feechan et al. 2008). Some AtMLO genes are upregulated, although downregulation was also noticed at 24 h under drought stress conditions in arabidopsis (Chen et al. 2006).

Since all 14 CmMLO genes were expressed in all organs including leaf samples, we carried out expression analysis under various stress conditions only for the leaf samples. All 14 CmMLO genes were variably upregulated in response to drought stress, and many of them contained cis-acting elements such as Myb binding site (MBS) involved in drought-inducibility as well as defense and stress-responsive elements (TC-rich repeats) in their promoters (Fig. 6 and Table S4). We speculate that the induced members of CmMLO genes were drought stress-responsive in a temporal fashion. In addition, those induced genes might be related to ABA signaling as ABA concentrations increase under drought stress conditions in leaves to regulate stomatal closure and minimize water loss from leaves (Lim and Lee 2014; Cutler et al. 2010; Hubbard et al. 2010; Lee and Hwang 2006, 2009; Fujita et al. 2005; Nakashima et al. 2009). Our speculation is further supported by the variable and significantly different transcript levels produced by the majority of those 12 CmMLO genes under ABA treatment (Fig. 10). These expressed CmMLO genes might also be related to pathogenicity reactions, as under drought stress conditions many pathogenesis-related genes are upregulated in plants (van Loon et al. 2006).

Eleven candidate CmMLO (CmMLO1, 2, 3, 4, 5, 6, 8, 10, 11 and 13) genes temporally-induced under salt stress might be responsive to salt tolerance in melon (Fig. 7). This finding is consistent with the expression of AtMLO11 in arabidopsis and the PM-susceptible CaMLO2 gene in pepper, which were also induced by salt stress (Chen et al. 2006; Kim and Hwang 2012). Salt-induced CmMLO genes might also be regulated by the ABA signaling pathway, since ABA can convert the initial stress signal of high salinity into a cellular response (Fujita et al. 2005; Nakashima et al. 2009).

The ten candidate genes (CmMLO1, 2, 3, 4, 5, 6, 7, 11, 12 and 13) induced in response to heat stress within 12 h after stress began contained the heat stress responsive cis-acting HSE element (Fig. 8 and Table S4). Hence, induced CmMLO genes might be related to heat tolerance mechanisms in melon through a heat signal transduction pathway (Liu et al. 2003). A MLO protein containing C-terminal calmodulin (CaM) activated by cytosolic Ca2+ under heat stress initiates the transcription and translation of heat shock proteins (HSPs) to maintain cellular homeostasis (Liu et al. 2003). Although four induced candidates (CmMLO7, 8, 9, and 12) of CmMLO genes expressed at 48 h contained defense and stress-responsive (TC-rich repeats) cis-acting elements; however, only CmMLO7 contained low-temperature responsive (LTR) element (Fig. 9 and Table S4). Therefore, we conclude that these genes might be strong candidates to be related to cold tolerance at the later stage of the stress period (Fig. 9). By contrast, an earlier report showed that some AtMLO genes are induced at the early stage (12 h) by cold treatment (Chen et al. 2006). Together, these findings indicate that MLO genes might be cold responsive in a temporal fashion in diverse crops.

Since the 12 CmMLO genes were differentially induced in response to exogenous ABA treatment and most of them contained abscisic acid responsive element (ABRE) and defense and stress-responsive (TC-rich repeats) cis-acting elements, we predict that the induced CmMLO genes were ABA-responsive in a temporal fashion (Fig. 10). In addition, exogenous ABA-induced genes might be related to ABA-dependent signaling pathway in melon plants (Shinozaki and Yamaguchi-Shinozaki 2000; Ramanjulu and Bartels 2002; Lim and Lee 2014), which supports the possibility of their function in stress tolerance under abiotic stress. For further investigation, we propose simultaneous inoculation of PM fungus followed by different abiotic stress treatments to explore the interactive function of MLO proteins in C. melo.

In conclusion, we extracted 14 CmMLO genes from the Melonomics database. Sequence analysis revealed a variable number of TM domains among CmMLO proteins, which showed a high degree of similarity to MLO proteins of other species. Phylogenetic analysis revealed that 14 CmMLO proteins were distributed into five out of six distinct clades. RT-qPCR analysis of 14 CmMLO genes revealed differential expression patterns among four different organs that indicated their roles in growth and development in melon plants. We systemically studied the expression behavior of all 14 CmMLO genes in response to two races of the PM fungus P. xanthii, four different abiotic stress treatments and one exogenous phytohormone treatment. Upregulation of CmMLO3 and CmMLO13 homologs located in clade V; CmMLO2 of clade VI; CmMLO1, CmMLO11 and CmMLO12 of clade II; and CmMLO14 of clade III might be associated with disease appearance with both races. A high transcript abundance of 14 CmMLO genes in response to drought, 13 in response to salt, 10 in response to heat, 4 in response to cold and 12 in response to ABA treatments indicated that some family members of CmMLO genes are responsive to multiple stresses. The data reported in this study will help to facilitate selection of appropriate candidate CmMLO genes for developing stress-resistant cultivars in C. melo.

Materials and Methods

In Silico Analysis of CmMLO Genes

Protein and CDS sequences for melon MLO were retrieved by from the Melon Genome Database (https://melonomics.net/), using keyword ‘MLO’. The 14 extracted MLO genes of C. melo were designated as CmMLO1 to CmMLO14. The ‘SMART’ software from EMBL, a web-based tool, (http://smart.emblheidelberg.de/smart/set_mode.cgi?Genomic=1) was used to identify the TM domains, which were annotated manually. The Basic Local Alignment Search Tool (BLAST) that is available in the National Center for Biotechnology Information (NCBI) database (http://www.ncbi.nlm.nih.gov/BLAST/) was used for protein homology analysis. ProtParam (http://expasy.org/tools/protparam.html) software was used to study the primary structures of the genes. Nucleotide sequence similarity was analyzed for the 14 CmMLO genes using ClustalW (http://www.genome.jp/tools/clustalw/). MEME software (Multiple Em for Motif Elicitation, V4.10.0) (http://meme-suite.org/tools/meme) was used to analyze the conserved motifs of the MLO protein sequences (Bailey et al. 2006). The MEME search settings were (1) optimum motif width ≥ 6 and ≤200 and (2) maximum number of motifs for identification = 15. To identify the putative cis-acting regulatory DNA elements in the promoters of CmMLOs genes, about 1500-bp CDS sequences upstream from the translation initiation codon (ATG) sequence was used for analysis using PlantCARE (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) (Lescot et al. 2002).

Phylogenetic Relationships among MLO Proteins

We retrieved the arabidopsis, rice, barley, and tomato MLO proteins from NCBI (http://www.ncbi.nlm.nih.gov/) and cucumber MLO proteins from the Cucurbit Genomics Database (http://www.icugi.org/cgi-bin/ICuGI/genome/search.cgi). The ClustalW program (Thompson et al. 1997) was used to align the CmMLO proteins with the MLO proteins of the above crops. A phylogenetic tree was constructed using the maximum likelihood method based on the Whelan and Goldman model with 1000 bootstrap replicates, using MEGA6.06 software (http://www.megasoftware.net) (Whelan and Goldman 2001; Tamura et al. 2007). Six distinct clades (I through VI) were identified, and these clades were named according to the classification of Devoto et al. (2003).

Plant Materials

Overnight-soaked seeds of C. melo line ‘SCNU1154’, susceptible to all known races of powdery mildew fungus, were sown in plastic pots containing sterilized soil mixture for germination, and seedlings were produced in growth chambers at 22 °C under 16:8 h light/darkness with a photon flux density of 140 μmolm−2 s−1. The relative humidity was maintained between 65% and 75%. Fresh roots, stems, leaves at three growth stages, and flowers were sampled. The selected plant organs were sampled, frozen immediately in liquid nitrogen and stored at −80 °C. These samples were used for organ-specific expression analysis of CmMLO genes by reverse-transcription quantitative PCR (RT-qPCR) to investigate their possible role in growth and development of the melon plants.

Biotic Stress Treatments

Four-week-old seedlings of C. melo ‘SNCU1154’ were treated separately with two different races of P. xanthii. Race 1 and isolate DH487 (Race N) (Kim et al. 2016) of P. xanthii were cultured on detached leaves of the SCNU1154 melon line in Petri dishes containing fungal growth medium (mixture of 5 g agar powder, 20 g D-mannitol, and 10 g saccharose in 1000 ml distilled water) in a growth chamber at 22 °C with a 16:8 h day: night photoperiodic cycle and a photon flux density of 140 μmolm−2 s−1 (Howlader et al. 2016). Relative humidity of 70–80% was maintained during the culture period. Active fungal inocula of the two races were dusted manually onto the adaxial part of the 5th leaf from the ground using sterilized pencil brushes to induce biotic stress (Howlader et al. 2016). Control and mock-treated plants were rubbed with a pathogen-free sterilized pencil brush. P. xanthii-treated and mock-treated plants were covered by separate perforated and transparent polythene bags in a growth chamber and provided with mist to maintain high relative humidity. Samples were collected from the 5th leaf position considering the youngest leaf as a reference point. Samples were harvested from P. xanthii-inoculated and mock-treated plants at 0 h, 6 h, 1 d, 2 d, 5 d, and 10 d post inoculation (dpi). The harvested samples were snap-frozen immediately in liquid nitrogen and stored at −80 °C until RNA was extracted.

Abiotic Stress Treatments

For different abiotic stress treatments, seedlings of C. melo ‘SNCU1154’ at the three true leaf stage were subjected to different abiotic stress treatments. The seedlings were separately exposed to 4 °C for 48 h and 42 °C for 24 h to induce cold and heat shock, respectively (Wan et al. 2010). The roots of seedlings were soaked in 300 mM NaCl solution for salt treatment at 25 °C up to 48 h (Wan et al. 2010). To impose drought treatment, seedlings were carefully removed from the soil to avoid injury and then subjected to dehydration up to 48 h on Whatmann 3 mm filter sheets (Advantec®, Tokyo, Japan) (Ahmed et al. 2013). For the phytohormone treatment, seedlings were exogenously sprayed with a 100 μM solution of abscisic acid (ABA) at their adaxial position (Wan et al. 2010). The plants were wrapped in polyvinyl bags for 48 h, leaving perforations for air exchange after the treatment. A non-treated plant at the commencement of treatment imposition was regarded as the ‘control’. The non-treated ‘mock’ samples were also collected along with treated samples at each time point. Both ‘control’ and ‘mock’ plants were treated with water. For all treatments, including mock treatments, leaf samples (1st and 2nd leaves) were collected at 0 h, 30 min, 1 h, 3 h, 6 h, 12 h, 24 h and 48 h (except heat treatment). The collected samples were snap-frozen immediately in liquid nitrogen and kept at −80 °C for RNA isolation.

RNA Extraction and cDNA Preparation

Total RNA from different organs, as well as the control and stress-treated samples was extracted using an RNeasy mini kit (Qiagen, Hilden, Germany). RNase-free DNase treatment was performed (Promega, Madision, USA) to eliminate the contaminants of genomic DNA. A Superscript® III First-Strand synthesis kit (Invitrogen, California, USA) was used for cDNA synthesis according to the manufacturer’s instructions. The transcript level of roots tissue is considered as control to compare the relative expression levels among different tissues (Khatun et al. 2016, 2017; Howlader et al. 2017).

Reverse-Transcription Quantitative PCR Analysis

The gene-specific primers developed by Howlader et al. (2016) and the housekeeping Cm-Actin primers from C. melo (Gene Bank Acc. AY859055) (Cheng et al. 2012) were used as the internal control in all analyses (Table S5). Efficiency of the primers were tested following Robin et al. (2016). RT-qPCR was conducted using 1 μL of 50 ng cDNA as template separately from all four tissues including roots, stems, leaves, and flowers of control; and mock and stress-treated C. melo line ‘SCNU1154’ in a 20-μL reaction volume employing a 2× qPCR BIO SyGreen Mix Lo-Rox SYBR® Green Super-mix with ROX (PCR Biosystems Ltd., London, UK). The RT-qPCR conditions were set at pre-incubation for 5 min at 95 °C, followed by 3-step amplifications at 95 °C for 15 s, 58 °C for 15 s, and 72 °C for 20 s for 40 cycles. The melting temperature was 95 °C for 10 s, 65 °C for 60 s, and 97 °C for 1 s as a default setting. The fluorescence reading was recorded at the last step of each cycle. There were three biological replicates and three technical replications per biological replicate. Light Cycler 96 (Roche, Germany) was used to detect amplification detection and also to analyze data. The calculation of relative gene expression levels were carried out following the 2 -∆∆ct method (Livak and Schmittgen 2001).

Statistical Analyses

Analysis of variance was conducted following a generalized linear model using Minitab statistical software version 17 (Minitab Inc., State College, PA, USA) to determine the significance of variation between sampling points under each treatment.