Genome-wide identi cation and expression analysis of autophagy genes in cucumber

Autophagy is an evolutionarily conserved physiological and developmental process in eukaryotes. In this process, damaged proteins in cells are degraded and cytoplasmic materials recycled. When plants are exposure to stress conditions or their growth and development are blocked, autophagy is induced to maintain the cell homeostasis by degrading long-lived proteins in the cells and organelles that function abnormally due to aging or damage. Cell autophagy has multiple functions in plants, it involved in growth and development, senescence, and responses to biotic and abiotic stress. So far, thirty three autophagy genes (ATG) have been found in rice, and more than 30 autophagy-related genes have been found in Arabidopsis, tobacco and corn, respectively. Four autophagy genes induced by salicylic acid were found in cucumber, but a little still unknown about all of autophagy genes in cucumber genome. Our experiment fully explored the ATG gene family of cucumber genome based on bioinformatics methods and identied 20 CsATG genes. We systematically analyzed the structure, conserved motifs, expression and phylogeny relationship of these ATG genes, which lays the foundation for exploring the function of the genes. and explored the expression prole of the 20 CsATG genes in response to SA and MeJA, two hormones related with autophagy, especially in response to three kinds of abiotic stresses (salt stress, drought stress and carbon starvation stress). We found that 20 CsATG genes are specically in cucumber roots, stems leaves, and can respond to high salt, drought and carbon starvation stresses to varying degrees. This study provides important clues for studying the function of cucumber ATG genes. gene MdATG10


Results
A total of 20 putative ATG genes were identi ed in the cucumber genome. Gene duplication analysis showed that both fragmented and tandem duplication played vital roles in the ampli cation of cucumber ATG gene family. Gene expression analysis showed that 16 CsATG genes were induced by the salicylic acid (SA) treatment, and 16 CsATG genes were down-regulated by Methyl jasmonate (MeJA) treatment. Under high salinity stress, 10 CsATG genes were induced in roots. Under drought stress, 16 CsATG genes were induced in roots. Under carbon starvation stress, all of 20 CsATG genes were induced to express in leaves, suggesting that cell autophagy has a potential role in nutritional starvation tolerance.

Conclusion
Our results clearly have deepened our understanding of the characteristics and functions of cucumber ATG gene, and also found some new gene resources that can be used for the future development of cucumber and other crop varieties, which can resist stress.

Introdution
In early single-cell eukaryotes, in order to adapt to nutritional de ciencies or other forms of environmental stress, a mechanism for recovering and reusing nutrients has gradually evolved, which can protect the organism and regulate abiotic stress response (Inoue, et al. 2006). In some nutrient-poor soils, there is often a lack of two nutrients, carbon and nitrogen. This will greatly affect the growth and development of plants. Therefore, the plant evolved an autophagy mechanism to adapt to the plant against nutrient de ciency in the soil. Autophagy has the ability to degrade the damaged proteins and some organelles in the plant, thereby recycling and reusing its own nutrients to ensure normal growth and development of plants (Liu, et al. 2009). There are at least three types of autophagy in eukaryotes: macroautophagy, microautophagy, and chaperone-mediated autophagy. The most widely studied in plants is macroautophagy. The formation of autophagic vesicles (autophagosome) with a double membrane structure is biological feature of macroautophagy (Liu and Bassham 2012;Mizushima, et al. 2008;Mizushima, et al. 2011;Zhang, et al. 2016). Autophagic vesicles are formed by the expansion and growth of phagocytic vesicles (phagophore) (Yoshimoto 2012). The outer membrane of mature autophagic vesicles is fused with the vacuolar membrane and then releasing the autophagosomes that encapsulate the cytoplasmic components into the vacuole cavity and degrading them into amino acids and other small molecular substances by hydrolases. Cell autophagy is involved in many important biological processes such as plant growth, development, and response to biotic and abiotic stress. Also it is closely related to important agronomic traits of crops. Therefore, autophagy is considered to play a vital role in protecting plants against nutritional de ciency and various environmental stresses. Recently, more attention has been paid to the molecular and physiological mechanisms of autophagy.
To our knowledge of autophagy basically came from yeast. So far, 41 ATG genes have been founded in yeast. Many autophagy-defective mutants have been isolated and used to reveal the molecular mechanism of autophagy (Harding, et al. 1995;Kanki, et al. 2009;Klionsky, et al. 2003;Mizushima 2007;Noda, et al. 2002;Okamoto, et al. 2009;Suzuki, et al. 2010;Thumm, et al. 1994;Tsukada and Ohsumi 1993;Wang and Klionsky 2003;Yao, et al. 2015). According to the process of ATG proteins regulating autophagy pathways, these proteins can be divided into different functional groups (Shibutani and Yoshimori 2014;Xie and Klionsky 2007) (1) ATG1-ATG13 kinase complex is responsible for the induction of autophagy and it is negatively regulated by the TOR (target of rapamycin) kinase; (2) class III PI3K (phosphatidylinositol 3-kinase) complex contains ATG6 (called Beclin1 in mammals) and it plays an important role in the nucleation of vesicles; (3) ATG9-ATG2-ATG18 complex transports membrane lipids to phagocytic vesicles, thereby promoting the extension of phagocytic vesicles; (4) ATG12-ATG5 and ATG8-PE (ATG8-phosphatidylethanolamine) linkers generated by two ubiquitin-protein binding systems on the vesicle membrane extension and maturation of autophagic vesicles are essential.
With the development of genome-wide technology, a number of genes homologous to yeast ATG have been identi ed in plants and mammals. Sequence alignment results showed that the sequence homology between the Arabidopsis ATG protein and the corresponding yeast ATG protein is not high and only part of the protein can completely complement the phenotype of yeast mutant. The core autophagy mechanism in plant is similar to yeast. It was found that most of the important amino acid residues of yeast ATG protein were also conserved in Arabidopsis (Yoshimoto, et al. 2010). The discovery of these genes also can prove that the evolution of autophagy is highly conserved. In recent years, autophagy genes have been identi ed in Arabidopsis, rice, and maize genomes by genome sequencing and some of these genes have been functionally analyzed. There were 33 autophagy genes in rice, and over 30 autophagy-related genes were also found in tobacco and corn, respectively, and their functions were veri ed (Chung, et al. 2009;Li, et al. 2015;Su, et al. 2006;Xia, et al. 2011;. The autophagy of Arabidopsis thaliana is particularly well studied in plants. At present, 36 autophagy-related genes in Arabidopsis has been discovered (Avin-Wittenberg, et al. 2012). However, to our knowledge, there are very few studies on cucumber.
Cucumber (Cucumis sativus L.) is an important and popular vegetable crop of the family Cucurbitaceae and is widely cultivated in the temperate and tropical regions. Cucumbers are vulnerable to powdery mildew and downy mildew during their growth process, which affects the growth and development of cucumbers and causes serious economic damage. The aim of this study is to provide a clear role for the mechanism and a better understanding of cucumber autophagy gene for their potential roles in disease resistance. We conducted genome-wide analysis of cucumber ATG on the published cucumber genome, identi ed 20 putative autophagy genes, predicted the chromosomal positions and protein structures of these genes, then detected and analyzed their expression feature using qRT-PCR. Our results deepened our understanding of the characteristics and functions of the cucumber ATG gene, and also found new genetic resources for cucumber and other crop varieties.

Identi cation of CsATG genes family members
A total of 20 putative CsATG genes were identi ed from the cucumber genome. ExPASy Proteomies Server (http://expasy.org/) was used to predict the molecular weight and the isoelectric point of the obtained 20 proteins (Table 1). It can be seen from the table 1 that the amino acid sequence length, predicted molecular weight, and predicted isoelectric point of the 20 CsATG proteins are quite different. The length of the CsATG protein sequence ranges from 87 amino acids (CsATG5a) to 1940 amino acids (CsATG2), and the molecular weight ranges from 9 to 215 kDa, and the predicted isoelectric point varies from 4.81 (CsATG10) to 9.77 (CsATG13a). Based on the name of cucumber ATG on NCBI GenBank, 20 putative CsATG genes were named respectively in reference to the results of homology alignment with Arabidopsis thaliana. Among the CsATGs, CsATG27 and CsATG10 have no corresponding names on NCBI GenBank, so they are named as new names.

Phylogenetic analysis of CsATG genes
To clarify the phylogenetic relationship of the autophagy protein family between cucumber and Arabidopsis thaliana, multiple alignments of ATG protein sequences from C. sativus and A. thaliana were performed using the Clustal X2.1 program, and a phylogenetic tree was constructed by MEGA (Ver. 5.0). Figure 1, the autophagy genes of the two species C. sativus and A. thaliana were divided into four subfamilies. GROUP1 contained CsATG18b, CsATG5a, CsATG2, CsATG10 and CsATG3. GROUP2 contained CsATG9 and CsATG5b. GROUP3 contained CsATG16 and CsATG27. GROUP4 contained CsATG5c, CsATG8a, CsATG8b, CsATG8c, CsATG8d, CsATG8e, CsATG8f,CsATG13a and CsATG13b.

As shown in
Gene structure analysis of CsATG genes Exon-intron structural divergences within gene families have played a critical role in the evolution of multigene families (Zhang, et al. 2012). To further reveal the structural features of CsATG genes, their structures, including CDS and UTR, were analyzed ( Figure 2). We found that their UTR number ranges from 0 to 3. CsATG5a, CsATG5b, CsATG8a, CsATG11 and CsATG27 do not contain UTR, and CsATG18b contains 3 UTR, other CsATGs contain 1 or 2 UTR. CDS number was from 1 to 13, CsATG13a contains only 1 CDS, CsATG2 contains 13 CDS, CsATG8b, CsATG8c, CsATG8d, CsATG8f have the same number of CDS, The number of CDS members in other groups is different, and their number ranges from 2-11. Analysis of the CDS-UTR structure shows that some members of the ATG subfamily have the same or similar CDS-UTR structure. So these genes vary widely.

Synteny analysis of ATG genes between cucumber and Arabidopsis
Based on the analysis of the syntenic relationship between the cucumber ATGs and the Arabidopsis thaliana ATGs (Figure 3), the ATG homologous genes on chromosome 1 of cucumber can be found on chromosome 1, 2, 3, 4, 5 of Arabidopsis, and the ATG homologous genes on chromosome 2 of cucumber can correspond to chromosome 4 of Arabidopsis, and the ATG homologous genes on chromosome 3 of cucumber can correspond to chromosome 2 and 3 of Arabidopsis, and ATG homologous genes on chromosome 4 of cucumber can correspond to chromosome 5 of Arabidopsis, and the ATG of homologous genes on chromosome 5 of cucumber can correspond to chromosome 2, 3 and 4 of Arabidopsis, and the ATG homologous genes on chromosome 6 of cucumber can correspond to chromosome 1, 3, 4, and 5 of Arabidopsis. Therefore, the ATG genes of cucumber have a high colinearity with ATG genes of Arabidopsis.

Analysis of conserved motifs of ATG genes in cucumber
We used the MEME web server to search for conserved sequence motifs and to predict their composition and diversity of the motifs present in the CsATG proteins. A total of 20 different conserved motifs were found in these CsATG proteins. It is clear that some genes (CsATG8a, CsATG8b, CsATG8c, CsATG8d, CsATG8e, and CsATG8f) share the same motif ( Figure 4), which shows that members of the same subfamily usually contain very similar or the same conserved structural motif. Some CsATGs, such as CsATG5a, CsATG5b, and CsATG13a, have no motif structure, but CsATG9 contains a maximum of 11 motifs.

Chromosomal location of ATG genes in cucumber
The cucumber genome contains seven chromosomes. The physical location information of 20 cucumber autophagy genes was obtained from the PHYTOZOME (www.phytozome.net/) database. 20 CsATG genes were unevenly located on seven chromosomes of cucumber (Fig. 5). Chromosomes 3 and 6 each contained 6 autophagy genes, Chromosome 2 contained 3 autophagy genes, Chromosomes 1, 5, and 7 each contained 2 autophagy genes, and Chromosome 1 contained only 1 gene. The CsATG8a, CsATG8e, CsATG8b, and CsATG18b genes were segmental duplication, the CsATG11, CsATG18b genes were tandem duplication, the CsATG5a, CsATG5b, and CsATG5c genes were tandem duplication, and the CsATG13a and CsATG13b genes were tandem duplication.

Selection pressure analysis of homologous gene pairs
Ks represents the background base replacement rate of the evolutionary process. Ks can be used to infer the time of the event, such as the whole genome doubling time which has an important role in exploring the origin of species. Through BLASTn sequence similarity analysis, we identi ed two putative homologous gene pairs (CsATG8a/CsATG8eand CsATG8b/CsATG8e) in the cucumber genome. To estimate the divergence time of the ATG gene, we calculated the Ka/Ks ratio of these two gene pairs. We also conducted a statistical analysis of the Ka/Ks ratio and Ks value. The Ks values of the CsATG8a/CsATG8e, and CsATG8b/CsATG8e paralogous pairs were 1.3233 and 2.0606. The Ka values of the CsATG8a/CsATG8e, and CsATG8b/CsATG8e paralogous pairs were 0.0956 and 0.1416. Then the large-scale duplication activity was about 108.46 and 168.9 MYA. The Ka/Ks distribution in the cucumber genome ranges from 0.068 to 0.072. When Ka/Ks <1, this indicates that a gene has undergone CED negative selection and puri cation selection. Therefore, in the cucumber genome, there is a strong puri cation selection in the ATG gene.
Expression pro les of ATG genes in cucumber under phytohormone, abiotic stress, and carbon starvation treatments In order to investigate the CsATG genes response to salicylic acid (SA) and Methyl jasmonate (MeJA), qRT-PCR of 20 CsATG genes in leaves were performed. The result was showed in Figure 6. Under SA treatment, the expression of CsATG16, CsATG11, CsATG18b, and CsATG5c were down-regulated, while the other CsATG genes were up-regulated. The CsATG2, CsATG3, CsATG9, CsATG101, CsATG27, CsATG8d, CsATG8f, CsATG13a, and CsATG13b were up-regulated by more than 5 times. The time of the highest expression levels were at 6h for CsATG8f and CsATG101, at 9h for CsATG8a, CsATG8b, CsATG8c, CsATG8d, CsATG8e, CsATG5a, CsATG5b, and at 24h for CsATG2, CsATG9, CsATG3, CsATG27 and CsATG13a after two hormones treatments. The ATG genes in the same subfamily exhibited the same expression patterns. Unlike SA treatment, under MeJA treatment (Figure 7) Most of ATG genes showed down-regulated expression (CsATG5a, CsATG5b, CsATG5c, CsATG10, CsATG11, CsATG18b, and CsATG16) or had no change in their expression level (CsATG3, CsATG9, CsATG27, CsATG101, CsATG8d, CsATG13a, and CsATG13b). While CsATG8a, CsATG8b, CsATG8c, CsATG8e, CsATG8f, and CsATG2 were up-regulatory expression and their highest expression occurred at 9h. Interestingly, these up-regulated genes were also up-regulated under SA treatment. The genes CsATG5c, CsATG11, CsATG18b, and CsATG16 were down-regulated genes under both SA and MeJA treatment. From the expression patterns of 20 CsATG genes, we can see that these ATG genes displayed different expression pro les after the treatment with the two hormones, which may be related to the two hormones participated different autophagy pathways. In the roots (as shown in Fig. 7), under salt stress, the number of up-regulated CsATG genes (CsATG101, CsATG11, CsATG10, CsATG10b, CsATG2,CsATG8a, CsATG8b, CsATG8d, CsATG8f, CsATG9 and CsATG18b) were about the same as the number of CsATG genes whose expression level did not change much (CsATG3, CsATG27, CsATG5a, CsATG5b, CsATG8c, CsATG8e, CsATG13a, CsATG13b and CsATG16). Under drought stress, only CsATG10, CsATG101 and CsATG8f had no much change in the expression level. The other 17 CsATG genes were all up-regulated, in which the expression level of CsATG2, CsATG3, CsATG5b, CsATG5c CsATG8a, CsATG9 and CsATG13a were almost more than ten times that of the control group. Under the dark treatment, unlike salt and drought stress, all CsATG genes were up-regulated, in which nine CsATG genes (CsATG3, CsATG27, CsATG5a, CsATG5b, CsATG5c, CsATG8f, CsATG13a, CsATG9 and CsATG16) had signi cant up-regulated expression, while these CsATG genes were interestingly not up-or down-regulated under salt stress. The expression levels of 20 CsATG genes after dark treatment were signi cantly higher than those under drought and salt stress.

Discussion
Identi cation of 20 CsATGs in the cucumber genome In recent years, autophagy-related genes in multiple plant species have been identi ed. 41 ATG genes have been identi ed in yeast, and more than 30 ATG genes have been identi ed in the genomes of Arabidopsis, corn, rice and tobacco, respectively, which showed the evolutionary conservation of core autophagy mechanisms in different species. In our study, 20 ATG genes were found in the cucumber genome, which is much less than the number of autophagy genes in rice, tobacco and corn. Unlike yeast, which has only a single copy, autophagy genes present in other species have multiple copies, such as ATG8. There are 5, 9, and 11 copies of ATG8 in tobacco, Arabidopsis and soybean, respectively (Chung, et al. 2009). In cucumber, the autophagy gene ATG8 also has multiple copies, (ATG8a-ATG8f), which is similar to that in Arabidopsis, tobacco and so on.

Phylogenetic analysis of cucumber ATGs
The phylogenetic tree indicated that the ATG gene family of cucumber was divided into 4 groups (G1-G4), and each group contained ATG genes from both cucumber and Arabidopsis. Among these groups, CsATG family members are closely related to the corresponding members in Arabidopsis, such as: GROUP1 (AtATG3 and CsATG3, AtATG2 and CsATG2, AtATG10 and CsATG10, AtATG18b and CsATG18b), GROUP2 (AtATG9 and CsATG9), GROUP3 (AtATG16 and CsATG16), GROUP4 (AtATG8a and CsATG8a, AtATG8b and CsATG8b, AtATG8c and CsATG8c, AtATG8d and CsATG8d, AtATG8e and CsATG8e, AtATG8f and CsATG8f, AtATG11 and CsATG11, AtATG101 and CsATG101, AtATG13a and CsATG13a, AtATG13b and CsATG13b). Each CsATG protein sequence is highly similar to a homologous protein in Arabidopsis, which indicating that their ATG genes are largely functionally similar, and these results are consistent with current understanding of plant evolutionary history (Yoshimoto, et al. 2004). So far, the expression or function of most autophagy genes in Arabidopsis has been reported, which lays the foundation for predicting the function of cucumber autophagy genes.

Analysis of ATG gene structure in cucumber
In general, the pattern of introns and exons in the genome provides important evidence for their evolutionary relationship. In this study, we comprehensively analyzed the exons and introns structure, distribution and length of ATG gene family members in cucumber. We found the conservation in the exonintron structure of the ATG genes in cucumber, which was similar to that of Arabidopsis and our results revealed that they have a highly conserved conlinearity. It supports a close evolutionary relationship between the plants.

Analysis of conserved motifs of ATG genes in cucumber
This study found that CsATG8a, CsATG8b, CsATG8c, CsATG8d, CsATG8e, CsATG8f share the same motif. It can be seen that members of the same subfamily usually contain very similar or identical conservative structural motifs, which indicates that different members in the same group may have functional redundancy. However, the motif in CsATG16, CsATG18b, CsATG5a, and CsATG5b is different, which indicates the different members of the same group have different functions and different responses to different environmental signals. Our results were similar to that of previous studies on tobacco and Arabidopsis (Xia, et al. 2011).

Chromosomal mapping of autophagy gene family in cucumber
Gene duplication (including tandem duplication, segmental duplication, and whole-genome duplication) plays an important role not only in the process of genome shu ing and genome expansion, but also in the diversi cation of gene functions and the large-scale generation of gene families (Vision, et al. 2000). In this study, it was found that the physical positions of 20 CsATG homologues were unevenly assigned to 7 chromosomes in cucumber. This is consistent with the uneven distribution of 33 OsATG genes on 12 chromosomes in rice (Xia, et al. 2011), and 37 SiATG genes on 9 chromosomes in millet (Liu and Bassham 2012). Both segmental duplication and tandem duplication of genes lead to multiple copies of the CsATG gene, which play an important role in the expansion of the CsATG gene family. In addition, some repetitive genes also play important roles in the evolution of new traits and in speciation. Some genes may play additional important role in long-term evolution.

Selection pressure analysis of homologous gene pairs
In genetics, the Ka/Ks ratio is an important parameter that determines whether positive Darwinian selection is related to gene differentiation (Juretic, et al. 2005). In this study, the Ka/Ks ratio of 2 pairs of duplicate genes in the CsATG gene family is <1. Compared with both the Ka/Ks ratio of 4 pairs of duplicate genes in the OsATG gene family is <1 (Xia, et al. 2011) and the Ka/Ks ratio of 6 pairs of duplicate genes in the SiATG gene family is <1 (Li, et al. 2016), the CsATG gene family has less pressure for puri cation selection than the SiATG and OsATG gene families. In all organisms, most nonsynonymous substitutions are deleterious mutations; only a few are neutral or advantageous mutations. The positive Darwinian selection will retain the advantages of non-synonymous mutations, and puri cation selection will gradually remove deleterious non-synonymous mutations (Juretic, et al. 2005).

Expression pro les of ATG genes in cucumber under phytohormone, abiotic stress, and carbon starvation treatments
Plants are often exposed multiple stresses, such as salt, cold and drought in nature, and have evolved a set of defense mechanisms to adapt to the environmental stresses. Plant hormones, such as auxin (IAA), gibberellin (GA), abscisic acid (ABA), salicylic acid (SA), and methyl jasmonate (MeJA), all participate in and regulate the stress response. Autophagy also plays an important role in plant stress response. In fact, plant hormones are closely related to cellular autophagy (Cao, et al. 2020;Porta and Jiménez-Nopala 2019). Although most of the Arabidopsis autophagy-de cient mutants may show signs of premature aging due to excessive accumulation of salicylic acid, these mutants can complete their life cycle under normal growth conditions. Tests on mutants of Arabidopsis thaliana with de cient in SA biosynthesis or its signaling have further demonstrated that the atg2 mutant requires the SA pathway for enhanced resistance to powdery mildew . Some studies have shown that SA and its analogs can activate autophagy in Arabidopsis cells, while autophagy negatively regulates SA synthesis and its signal transduction (Zhang, et al. 2017). Many studies have shown that there are complex networks of various hormones and their signaling pathways in plant growth and development and stress response (Koornneef, et al. 2008). In addition, a large number of gene chip data con rm that many genes are coregulated by different hormones, which shows that hormone signals have a key function in coordinately regulating the same biological process in plants (De Paepe, et al. 2004).
In cucumber leaves after treatment with SA and MeJA, 20 autophagy genes showed different expression levels. Most ATG genes showed a downward trend under MeJA treatment, while most ATG genes showed an upward trend under SA treatment. There are six genes (CsATG8a, CsATG8b, CsATG8c, CsATG8e, CsATG8f, and CsATG2) that were up-regulated in both phytohormone treatment, and there are four genes (CsATG5c, CsATG11, CsATG18b, and CsATG16) that were down-regulated in both phytohormone treatments. This result is inconsistent with many previous studies. In millet, whether treated with SA or MeJA, only a few ATG genes were down-regulated, while the rest of the ATG genes were either upregulated or remained unchanged (Li, et al. 2016). Of 30 ATG genes in tobacco, there were only 4 ATG genes and 11 ATG genes up-regulated by SA and JA, respectively, and there were only 1 ATG gene upregulated by SA, while other ATG genes remained unchanged (Klionsky, et al. 2003). However, a recent study found that most of the ATG genes were up-regulated by LCSA when it is revealed that low concentration SA could inhibit MeJA-induced senescence in Arabidopsis thaliana leaves (Yin, et al. 2020). These results may indicate that there is a certain relationship among salicylic acid, methyl jasmonate and autophagy. Both the two hormones can induce the occurrence of autophagy, but the expression patterns of the genes caused by the two hormones are different, which can explain the different ways of SA and MeJA-induced autophagy in cucumber leaves.
High concentration of salts can cause ionic stress, and a large amount of oxidatively inactivated proteins will be produced in plant cells, which will cause greater damage to the plant. This is one of the reasons why stresses often induce plant autophagy. Studies have shown that some autophagy genes, such as OsATG10b in rice (Shin, et al. 2009), AtATG18a in Arabidopsis (Liu, et al. 2009) and MdATG10 in apple (Huo, et al. 2020), were involved in salt stress. In this paper, we found that most of ATG genes response to salt stress, their expression decreased or increased under salt stress, except CsATG16, CsATG5a, CsATG5b and CsATG27. In general, when salt treatment of cucumber roots, stems and leaves, the highest expression levels of autophagy-related genes in response to salt stress were found in the leaves, followed by roots, and nally stems. This shows that autophagy responds to salt stress in different ways and in different degrees.
Drought stress can activate autophagy pathway in plants to favor their environmental adaptations. Many genes have been revealed to play an important role in drought resistance in plants and their overexpression can simultaneously improve plant resistance to drought stresses (Bao 2020;Kuzuoglu-Ozturk, et al. 2012;Li, et al. 2015;Luo, et al. 2020;Sun, et al. 2018;Tsugane, et al. 1999). In wheat, ATG8 expression was strongly up-regulated under drought stress, especially in the roots when compared to leaves (Kuzuoglu-Ozturk, et al. 2012). In foxtail millet, SiATG8a expression was dramatically induced by drought stress treatments (Li, et al. 2015). In this paper, we also found that most of ATG genes response to drought stress. In contrast to salt stress, the number of drought stresses responsive ATG genes in roots and stems was higher than in leaves, which imply that autophagy plays an important role in roots in response to drought stress.
Carbon is the highest content element in plants and plays an important role in the growth and development of plants. The roles of carbohydrates in plants include the energy consumed to maintain metabolism, the construction of carbon skeleton, the energy consumed during defense, and the maintenance of turgor pressure (Izumi, et al. 2013). Carbon starvation means that the organic carbon xed by photosynthesis in plants cannot satisfy the consumption of respiration. This phenomenon often occurs during the growth and development of plants (McDowell and Sevanto 2010). In plants, in order to adapt to the nutrient de ciency, a mechanism for nutrient recovery and recycling has evolved. This mechanism is often called autophagy (Rodríguez, et al. 2014;Toyooka, et al. 2006). Studies have shown that autophagy-related genes will undergo a coordinated transient up-regulation during starvation, which provides a basis for plant autophagy to play an important role in the nutrient cycle (Chung, et al. 2009). The Arabidopsis atg11 mutant appears more sensitive under nutrient-restricted conditions (Li, et al. 2014;Li and Vierstra 2014). In this paper, we found that the expression of 20 ATG genes in cucumber roots, stems, and leaves increased signi cantly under dark treatment compared to under other stress treatments (high salt and drought). Furthermore, the expression level of ATG genes in cucumber under dark treatment, like under drought stress, was higher in roots than in leaves, and the expression level was the lowest in stems. This indicates that there are differences in spatiotemporal expression in these three tissues, and autophagy responds to abiotic stress through different pathways.

Conclusion
In this paper, we identi ed a total of 20 cucumber ATG genes by genome-wide analysis, and analyzed the chromosomal distribution, intron-exon structures, duplication and divergence rates, et al. We explored the expression pro le of the 20 CsATG genes in response to SA and MeJA, two hormones related with autophagy, especially in response to three kinds of abiotic stresses (salt stress, drought stress and carbon starvation stress). We found that 20 CsATG genes are speci cally expressed in cucumber roots, stems and leaves, and can respond to high salt, drought and carbon starvation stresses to varying degrees. This study provides important clues for studying the function of cucumber ATG genes.

Plant materials and treatments
The experiment was conducted in the Key Laboratory of Plant Biology, Harbin Normal University from March to June, 2019.
Cucumber line 9930 was donated by Prof. Huang Sanwen (Institute of Vegetables and Flowers, Chinese Academy of Agricultural Sciences), and was used for cloning of CsATG genes. Cucumber variety "Changchunmici" was used to expression analysis of CsATG genes. The seedlings were cultivated in a light incubator at a temperature of 28°C for 14 hours under light and a temperature of 24°C for 10 hours in the dark, and were used in following treatments when the seedlings grow to the fourth true leaf. For hormones treatment, 50 μL of 10 mM SA or 0.2% MeJA was added dropwise at 3 cm intervals between leaf veins. The seedlings were exposed to 200 mM NaCl solution for salt stress, or to 20% (w/v) polyethylene glycol (PEG6000) for drought stress. The seedlings were covered with black plastic bags for dark treatment (as carbon starvation stress). Samples from the roots, stems and leaves of the treated plants were harvested at 0 h, 6 h, 9 h and 24 h after the treatment, and frozen in liquid nitrogen and stored at -80 °C for RNA extraction.

Database searches for cucumber ATG genes
To identify the ATG gene members in cucumber, the Hidden Markov Model (HMM) pro les of all sequences containing an ATG domain (PF03987) were used to search the Cucurbit Genomics database (http://www.cucurbitgenomics.org/organism/2). All redundant sequences were removed using the 'decrease redundancy' tool at http://web.expasy.org/decrease_redundancy/. In addition, in order to verify the identity of these putative CsATGs, the non-redundant candidate ATG genes in cucumber were identi ed using the InterproScan program to con rm the existence of the conserved ATG domain, and all of the ATG genes without an ATG domain were discarded. Information for the CsATG genes and predicted proteins, including CDS lengths, the predicted number of amino acids, and physicochemical parameters were obtained from the CuGenDB (http://cucurbitgenomics.org/).

Phylogenetic analyses and intron-exon structure determination
To examine the domain organization of the predicted ATG proteins in cucumber in detail, multiple sequence alignments of ATG domain-containing sequences were performed using Clustal W software, and we constructed a phylogenetic tree based on the complete CsATG sequences using the N-J method as implemented in MEGA software (v5.1) with a bootstrap analysis of 1000 replicates. The combined phylogenetic tree of AtATG and CsATG protein sequences was generated using the same method. In addition, the predicted exon-intron structures of the CsATG genes were visualized using the online Gene Structures Display Server (http://gsds.cbi.pku.edu.cn) by comparing the cDNA with the corresponding genomic DNA sequences.

Chromosome location and
All genes were mapped to 7 cucumber chromosomes in ascending order of CsATG gene physical location (bp). From short arm telomeres to long arm telomeres, they were visualized using MapChart.
Identi cation of conserved protein sequence motifs Conserved motifs present in the CsATG proteins were identi ed with the online MEME tool (http://meme.sdsc.edu/meme/intro.html) using the default parameter settings: maximum number of motifs = 20; optimum motif length range between 6 and 200. We also used the Pfam (http://pfam.sanger.ac.uk/search) and SMART (http:// smart.embl-heidelberg.de/) databases to annotate the structural motifs.

Identi cation of paralogs and Arabidopsis orthologs in cucumber
We used BLASTn to perform all-against-all nucleotide sequence similarity searches of the transcribed ATG sequences to identify paralogous sequences as shown by Blanc and Wolfe. Sequences that showed at least 40% identity with aligned regions >300 bp were de ned as paralogs. Putative Arabidopsis orthologs were identi ed by using each sequence as a query to search against all sequences from cucumber. If the ATG gene sequences gave the best hit, and >300 bp of the two sequences aligned, the two genes were then de ned as being orthologous.

Calculation of Ka/Ks values
Pairwise alignments of the paralogous and orthologous ATG gene sequence pairs were performed with ClustalX 2.11, and the results were further analyzed using MEGA 6.0. A synonymous substitution (Ks) is de ned as a mutation in which a nucleotide base is replaced by a different base in a protein-coding region of a gene that does not result in an amino acid change in the encoded protein, while a nonsynonymous substitution (Ka) results in a change in the amino acid sequence of a protein. The nonsynonymous and synonymous substitution rates were then calculated using DnaSP 5 to analyze gene duplication events. Ks can be used as a proxy for time when dating large-scale duplication events (Ina 1995). Therefore, the date of duplication events was subsequently converted into divergence time (T) using the formula T = Ks/ 2λ MYA for each gene pair (λ=6.1×10 −9 ).
Total RNA isolation and quantitative real-time PCR The total RNA of the samples was extracted with the OminiPlant RNA Kit (DNase I) from Beijing CoWin Biotech Co., Ltd., and cDNA synthesis was nished using HiFiScript cDNA Synthesis Kit from Beijing CoWin Biotech Co., Ltd. Quantitative RT-PCR (qRT-PCR) was performed to detect the relative expression level of the CsATG genes using ABI 7500 Fast Real-Time PCR System.
We then designed 20 pairs of gene-speci c primers using Primer Express 3.0. qRT-PCR ampli cations were performed on an ABI 7300 Real-Time system (Applied Biosystems) in 20 μL reactions containing 1 μL of each gene-speci c primer, 1 μL of cDNA sample, 7 μLddH 2 O, and 10 μL SYBR Green Master Mix reagent (Applied Biosystems). All primers for ampli cation of CsATG genes are given in Table 2. The qRT-PCR ampli cation conditions were: 95 °C for 30 s, followed by 40 cycles of 95 °C for 10 s, 55 ° C for 15 s, and 72 °C for 10 s. A melting curve analysis was performed for each sample to verify the speci city of the reactions. There were three biological and three technical replicates performed for each sample. The relative expression levels were evaluated using the ΔΔCT method. It is worth noting that for the stress treatments, relative gene expression [2−ΔΔCT, CK (0 h)] for each gene in the control plants was normalized to 1 as described previously (Schmittgen and Livak 2008 Tables   Table 1 Amino acid  Phylogenetic tree of ATG proteins in Arabidopsis and C. sativus. The ATG genes of C. sativus are divided into four subfamilies (Groups 1-4, with blue, purple, green, and red colors, respectively). The tree was constructed from an analysis with MEGA 5.0 software using a maximum likelihood method.

Figure 2
Exon-intron structure of the CsATG genes in C. sativus. The exon-intron structure coding sequence of the CsATG genes is compared with the corresponding genomic sequence using the GSDS program of the genetic structure. The green and yellow boxes represent untranslated upstream/downstream regions and coding region, respectively. Lines indicate introns.

Figure 3
Synteny analysis of the ATG genes between C. sativus and Arabidopsis. The yellow and green bars represent the C. sativus and Arabidopsis chromosomes, respectively.

Figure 4
Conserved motifs of ATG proteins in C. sativus. Different color boxes represent different types of motifs, and the length of each box represents the actual motif size labeled by the number of nucleotides.

Figure 5
Distribution of 20 ATG genes on the seven chromosomes in C. sativus.

Figure 6
Expression of 20 CsATG genes in C. sativus under SA and MeJA treatment. The relative expression of each ATG in C. sativus seedlings under different hormone levels was normalized to Actin mRNA. 10 mM SA and 0.2% MeJA were added to the leaves for hormone induction treatment, and the expressions of each ATG were calculated at 6h, 9h, and 24h at the dropwise locations, and compared with the ATG of seedlings under normal growth conditions.

Figure 7
Expression of 20 CsATG genes in Cucumis sativus roots under three abiotic stresses. The relative expression of each ATG gene is normalized to Actin mRNA. The roots of the four-leaf stage seedlings were treated with 200 mM NaCl, 20% PEG6000, or in dark for 6h and 24h, respectively. The expression levels of each ATG were calculated and compared with the ATG expression levels of the seedlings under normal growth conditions.

Figure 8
Expression of 20 CsATG genes in C. sativus stems under three abiotic stress. The relative expression of each ATG gene was normalized to Actin mRNA in stems of C. sativus seedlings under drought, high salt, and carbon starvation stress. Four-leaf stage C. sativus seedlings were treated with 200 mM NaCl, 20% PEG6000, or in dark for 6h and 24h, calculate the expression of each ATG, and compare with the ATG expression of seedlings under normal growth conditions.