Genome-Wide Analysis and Abiotic Stress-Responsive Patterns of COBRA-like Gene Family in Liriodendron chinense

The COBRA gene encodes a plant-specific glycosylphosphatidylinositol (GPI)-anchored protein (GAP), which plays an important role in cell wall cellulose deposition. In this study, a total of 7 COBRA-like (COBL) genes were identified in the genome of the rare and endangered woody plant Liriodendron chinense (L. chinense). Phylogenetic analysis showed that these LcCOBL genes can be divided into two subfamilies, i.e., SF I and II. In the conserved motif analysis of two subfamilies, SF I contained 10 predicted motifs, while SF II contained 4–6 motifs. The tissue-specific expression patterns showed that LcCOBL5 was highly expressed in the phloem and xylem, indicating its potential role in cellulose biosynthesis. In addition, the cis-element analysis and abiotic stress transcriptomes showed that three LcCOBLs, LcCOBL3, LcCOBL4 and LcCOBL5, transcriptionally responded to abiotic stresses, including cold, drought and heat stress. In particular, the quantitative reverse-transcription PCR (qRT-PCR) analysis further confirmed that the LcCOBL3 gene was significantly upregulated in response to cold stress and peaked at 24–48 h, hinting at its potential role in the mechanism of cold resistance in L. chinense. Moreover, GFP-fused LcCOBL2, LcCOBL4 and LcCOBL5 were found to be localized in the cytomembrane. In summary, we expect these results to be beneficial for research on both the functions of LcCOBL genes and resistance breeding in L. chinense.


Introduction
Cell walls play a vital role in protecting plants against biotic and abiotic stresses, as well as in plant support and protection [1][2][3]. The primary cell wall, consisting mainly of cellulose, hemicellulose, pectin and proteins, surrounds each plant cell. The secondary cell wall consists mainly of cellulose, hemicellulose and lignin, and, in specific cell types, is deposited between the plasma membrane and the primary cell wall after cell enlargement stops [4]. Cellulose, which is the most abundant biopolymer on Earth, is accumulated in plant cell walls [5]. Cellulose is a linear homopolysaccharide that makes up long and rigid microfibrils and is the load-bearing structure in cell walls [6]. The organization of cellulose is critical for directed plant growth [7,8]. The biosynthesis of cellulose is mediated by cellulose synthase (CesAs), which forms a rosette-like cellulose synthase complex (CSC) on the plasma membrane [5]. A genome-wide association study (GWAS) for culm cellulose content in barley (Hordeum vulgare) identified HvCOBRA, associated with cellulose synthesis [9]. The factors involved in cellulose synthesis were identified by Affymetrix microarray analysis in Arabidopsis thaliana, which indicated that COBRA homologous gene COBL4 was among the top 10 genes co-expressed with CESA4, 7 and 8, indicating a potential role of COBRA gene in protocell wall cellulose deposition [10].

Phylogenetic Analysis of LcCOBL Proteins
In order to explore the phylogenetic relationships between the COBL proteins, we constructed a phylogenetic tree using 7, 12, 11, 11 and 8 COBL proteins from L. chinense, A. thaliana, O. sativa, Vitis vinifera and Amborella trichopoda, respectively ( Figure 1). The results showed that seven LcCOBLs could be divided into two subfamilies, i.e., SF I and SF II, based on the phylogenetic tree ( Figure 1). Four LcCOBLs were grouped in SF I, including LcCOBL1, LcCOBL2, LcCOBL4 and LcCOBL5, while the remaining three LcCOBLs were grouped in SF II. As for A. thaliana, O. sativa, V. vinifera and A. trichopoda, SF I contained seven, eight, seven and six family members, while SF II contained five, three, four and two family members, respectively. The phylogenetic grouping of the LcCOBL proteins was further confirmed with a phylogenetic tree constructed using only LcCOBL proteins ( Figure S2).

Analyses of Locations, Structures and Conserved Motifs of LcCOBL Genes
In order to study the structural characteristics of the LcCOBL genes, we analyzed the conserved motifs and the exon-intron structures of the seven LcCOBL genes ( Figure 2). Among these 10 motifs, only motifs 1 and 3 were possessed by all the LcCOBL proteins. Additionally, motif 4 contained a CCVS-conserved motif of the COBRA protein ( Figure S3). The motif differences between the subfamilies were greater than those within the same subfamily, suggesting functional conservation among LcCOBL proteins within the same subfamily (Figure 2a,b). The LcCOBL genes contained 2-7 exons, and most of the evolutionarily related exon-introns shared similar structures (Figure 2c). In addition, the analysis of the conserved domains of the LcCOBLs showed that all the LcCOBLs possessed a COBRA domain or a COBRA superfamily domain (Figure 2c). These results further validate the reliability of the identified LcCOBL gene family and shed light on its functional evolution. The chromosomal distributions of the LcCOBL genes were determined based on the genome-wide data of L. chinense. The chromosomal localization analysis showed that seven LcCOBL genes were uniformly located on six chromosomes and one contig of L. chinense ( Figure 3). There was only one gene on each chromosome with no tandem gene replication event.
lants 2023, 12, x FOR PEER REVIEW Figure 1. Phylogenetic tree of COBL protein family in L. chinense, A. thaliana, O. sa and A. trichopoda. The full-length COBL protein sequences were aligned using t and the Bayesian method was used to construct the phylogenetic tree in BEAUti. was visualized with the online iToL. The bootstrap values are supported by 1000 are shown beside the branches. According to the phylogenetic tree, the seven LcC divided into two subfamilies (SF I and SF II), where SF I stands for LcCOBL subfa stands for LcCOBL subfamily II.

Analyses of Locations, Structures and Conserved Motifs of LcCOBL G
In order to study the structural characteristics of the LcCOBL genes, w conserved motifs and the exon-intron structures of the seven LcCOBL ge Among these 10 motifs, only motifs 1 and 3 were possessed by all the LcC Additionally, motif 4 contained a CCVS-conserved motif of the COBRA p S3). The motif differences between the subfamilies were greater than th same subfamily, suggesting functional conservation among LcCOBL prote same subfamily (Figure 2a,b). The LcCOBL genes contained 2-7 exons, an evolutionarily related exon-introns shared similar structures (Figure 2c). I analysis of the conserved domains of the LcCOBLs showed that all possessed a COBRA domain or a COBRA superfamily domain (Figure 2c) further validate the reliability of the identified LcCOBL gene family and sh functional evolution. The chromosomal distributions of the LcCOBL determined based on the genome-wide data of L. chinense. The chromosom analysis showed that seven LcCOBL genes were uniformly located on six and one contig of L. chinense (Figure 3). There was only one gene on each with no tandem gene replication event. The full-length COBL protein sequences were aligned using the clustalx tool, and the Bayesian method was used to construct the phylogenetic tree in BEAUti. Finally, the tree was visualized with the online iToL. The bootstrap values are supported by 1000 replications and are shown beside the branches. According to the phylogenetic tree, the seven LcCOBLs could be divided into two subfamilies (SF I and SF II), where SF I stands for LcCOBL subfamily I and SF II stands for LcCOBL subfamily II.

Cis-Element Analysis of LcCOBL Promoters
Cis-elements distributed on gene promoters provide targets that bind to transcription factors, whereby gene expression patterns are activated or inhibited in the processes of plant growth and development and coping with external environmental stresses. We predicted the promoter cis-acting elements through the PlantCARE, which is based on probabilistic sequence models (e.g., Gibbs Sampling) [32]. Based on the conservation of the promoter sequences [33], motifs that have been determined in other species were selected from the screened elements for possible functional analysis in L. chinense. We found four major types of cis-acting elements i.e., related to light, hormones, environmental stress or developmental responsiveness, distributed on the promoter regions of the LcCOBLs (Figure 4). Among the light-response-related cis-elements, G-box was the most abundant type, which can be found within all LcCOBL promoters. Regarding hormones and environmental stress, we found that the ABA-(ABRE) and

Cis-Element Analysis of LcCOBL Promoters
Cis-elements distributed on gene promoters provide targets that bind to transcription factors, whereby gene expression patterns are activated or inhibited in the processes of plant growth and development and coping with external environmental stresses. We predicted the promoter cis-acting elements through the PlantCARE, which is based on probabilistic sequence models (e.g., Gibbs Sampling) [32]. Based on the conservation of the promoter sequences [33], motifs that have been determined in other species were selected from the screened elements for possible functional analysis in L. chinense. We found four major types of cis-acting elements i.e., related to light, hormones, environmental stress or developmental responsiveness, distributed on the promoter regions of the LcCOBLs (Figure 4). Among the light-response-related cis-elements, G-box was the most abundant type, which can be found within all LcCOBL promoters. Regarding hormones and environmental stress, we found that the ABA-(ABRE) and drought-related (MYC and as-1) cis-elements were the most abundant cis-elements, showing that LcCOBLs might play an important role in the response to drought stress. As for the development-related cis-elements, we found that the most abundant type has a function in meristem expression.

Organspecific Expression Patterns of LcCOBLs
To further explore the potential functions of the LcCOBL family in tissue development, we investigated the expression patterns of the LcCOBLs across different organs of L. chinense ( Figure 5a). According to the transcriptome data regarding eight tissues, namely bark, bud, phloem, sepal, stamen, stigma and xylem. We found that five LcCOBL genes were expressed in these selected organs, except for two genes (LcCOBL6 and LcCOBL7). Moreover, these five LcCOBL genes were mainly expressed in the phloem, stamen, stigma and xylem tissues, with almost no expression in the bark, bud or sepal tissues. Interestingly, LcCOBL5 was mainly expressed in the phloem and xylem tissues, showing its potential role in stem growth. Furthermore, a quantitative reverse-transcription PCR (qRT-PCR) analysis confirmed the higher expression of LcCOBL5 in the stems than in the roots and leaves ( Figure 5b). drought-related (MYC and as-1) cis-elements were the most abundant cis-elements, showing that LcCOBLs might play an important role in the response to drought stress. As for the development-related cis-elements, we found that the most abundant type has a function in meristem expression.

Organspecific Expression Patterns of LcCOBLs
To further explore the potential functions of the LcCOBL family in tissue development, we investigated the expression patterns of the LcCOBLs across different organs of L. chinense (Figure 5a). According to the transcriptome data regarding eight tissues, namely bark, bud, phloem, sepal, stamen, stigma and xylem. We found that five LcCOBL genes were expressed in these selected organs, except for two genes (LcCOBL6 and LcCOBL7). Moreover, these five LcCOBL genes were mainly expressed in the phloem, stamen, stigma and xylem tissues, with almost no expression in the bark, bud or

Expression Patterns of LcCOBLs in Response to Abiotic Stresses
To analyze the responses of the LcCOBL genes to abiotic stresses, we examined the expression levels of these LcCOBLs under cold, drought and heat stress based on the available RNA-seq data (Figure 6a-c). The results show that three LcCOBLs, i.e., LcCOBL3, LcCOBL4 and LcCOBL5, transcriptionally responded to these abiotic stresses. Specifically, LcCOBL3 was highly up-regulated in response to these abiotic stresses and peaked at 12-24 h, 1 d and 3 d under cold, drought and heat stress, respectively. In contrast, LcCOBL5 was dramatically down-regulated as soon as it was exposed to these stresses. Compared with LcCOBL5 and LcCOBL3, the expression dynamic of LcCOBL4 was less violent, although it first increased and then decreased. sepal tissues. Interestingly, LcCOBL5 was mainly expressed in the phloem and xylem tissues, showing its potential role in stem growth. Furthermore, a quantitative reversetranscription PCR (qRT-PCR) analysis confirmed the higher expression of LcCOBL5 in the stems than in the roots and leaves (Figure 5b). Expression profiles of LcCOBL genes in bark, bud, phloem, sepal, stamen, stigma and xylem tissues. The heatmap data were averaged and plotted using TBtools, with red representing a high expression level and blue representing a low expression level. (b) qRT-PCR analysis of LcCOBL genes in different organs including leaves, roots and stems. The expression levels of related genes were calculated with 2 −ΔΔCt using a leaf as control. Mean values ± SE are shown for the 3 replicates, and the levels of significance relative to the control are ns: no significant difference, * p < 0.05 and ** p < 0.01.

Expression Patterns of LcCOBLs in Response to Abiotic Stresses
To analyze the responses of the LcCOBL genes to abiotic stresses, we examined the expression levels of these LcCOBLs under cold, drought and heat stress based on the available RNA-seq data (Figure 6a-c). The results show that three LcCOBLs, i.e., LcCOBL3, LcCOBL4 and LcCOBL5, transcriptionally responded to these abiotic stresses. Specifically, LcCOBL3 was highly up-regulated in response to these abiotic stresses and peaked at 12-24 h, 1 d and 3 d under cold, drought and heat stress, respectively. In contrast, LcCOBL5 was dramatically down-regulated as soon as it was exposed to these stresses. Compared with LcCOBL5 and LcCOBL3, the expression dynamic of LcCOBL4 was less violent, although it first increased and then decreased. Expression profiles of LcCOBL genes in bark, bud, phloem, sepal, stamen, stigma and xylem tissues. The heatmap data were averaged and plotted using TBtools, with red representing a high expression level and blue representing a low expression level. (b) qRT-PCR analysis of LcCOBL genes in different organs including leaves, roots and stems. The expression levels of related genes were calculated with 2 −∆∆Ct using a leaf as control. Mean values ± SE are shown for the 3 replicates, and the levels of significance relative to the control are ns: no significant difference, * p < 0.05 and ** p < 0.01.
To further determine whether the expression levels of the LcCOBL genes were influenced by abiotic stresses, Liriodendron seedlings treated with low-temperature conditions (4 • C) were collected to quantify the expression levels of the LcCOBLs using qRT-PCR analysis. The results show that each gene has a different expression pattern after cold treatment (Figure 6d). The expression level of LcCOBL3 in the experimental group was significantly increased compared with the control group, while the expression level of LcCOBL2 in the treatment group was decreased. LcCOBL4 and LcCOBL5 showed opposing expression patterns. LcCOBL4 expression was downregulated at 6 h and then upregulated. LcCOBL5 was upregulated at 24 h and then downregulated. This suggests that the LcCOBLs exhibit different expression trends under stress treatment, indicating that these genes respond to the regulation of abiotic stress to varying degrees.

Subcellular Localization of LcCOBL Proteins
To further explore the potential roles of these LcCOBL proteins, we cloned the fulllength CDSs of three LcCOBLs, i.e., LcCOBL2, LcCOBL4 and LcCOBL5 (Supplementary  Table S2). Then, these LcCOBL genes without terminating codons were inserted into pJIT166 vectors to obtain LcCOBL-GFP fusion vectors, which were driven by a 35S promoter. Then, these three vectors were separately transferred into the protoplasts of L. chinense calli via PEG-mediated protoplast transformation and transferred into onion epidermal cells by gene gun. At the same time, the 35S:H2B-mCherry vector was transferred into callus protoplasts as the control for the nuclear localization. The results show that the protoplasts transferred with the 35S:GFP vector expressed the GFP signal in the whole cell, which was a constitutive expression, while in 35S:H2B-mCherry lines, the mCherry signal was only observed in the nucleus. Furthermore, for the LcCOBL5-GFP fusion vectors, the GFP signal was found in the cytomembrane, and the weak fluorescence signal was also observed in the cytoplasm (Figure 7). However, we cannot determine the expression of LcCOBL2, To further determine whether the expression levels of the LcCOBL genes were influenced by abiotic stresses, Liriodendron seedlings treated with low-temperature conditions (4 °C) were collected to quantify the expression levels of the LcCOBLs using qRT-PCR analysis. The results show that each gene has a different expression pattern after cold treatment (Figure 6d). The expression level of LcCOBL3 in the experimental group was significantly increased compared with the control group, while the expression level of LcCOBL2 in the treatment group was decreased. LcCOBL4 and LcCOBL5 showed opposing expression patterns. LcCOBL4 expression was downregulated at 6 h and then upregulated. LcCOBL5 was upregulated at 24 h and then downregulated. This suggests that the LcCOBLs exhibit different expression trends under stress treatment, indicating that these genes respond to the regulation of abiotic stress to varying degrees.

Subcellular Localization of LcCOBL Proteins
To further explore the potential roles of these LcCOBL proteins, we cloned the fulllength CDSs of three LcCOBLs, i.e., LcCOBL2, LcCOBL4 and LcCOBL5 (Supplementary  Table S2). Then, these LcCOBL genes without terminating codons were inserted into pJIT166 vectors to obtain LcCOBL-GFP fusion vectors, which were driven by a 35S promoter. Then, these three vectors were separately transferred into the protoplasts of L. show that the protoplasts transferred with the 35S:GFP vector expressed the GFP signal in the whole cell, which was a constitutive expression, while in 35S:H2B-mCherry lines, the mCherry signal was only observed in the nucleus. Furthermore, for the LcCOBL5-GFP fusion vectors, the GFP signal was found in the cytomembrane, and the weak fluorescence signal was also observed in the cytoplasm (Figure 7). However, we cannot determine the expression of LcCOBL2, 4 proteins in other organelles except cytomembrane ( Figure S7). This requires further experiments to verify.

Functions of LcCOBLs in Stem Development and Stress Responses
In phylogenetic trees, genes with similar clusters may have similar functions. Therefore, LcCOBLs are likely to have similar biological functions to COBL proteins known to be in other species in this group. It was stated that genes with less intron number may be expressed faster than other genes by upstream signals [34]. Combining the gene expression patterns of members of the COBL gene families helps to predict the gene functions of the LcCOBL gene family members. At present, many genes related to brittle traits, including OsBCIL4 and OsBC1, that have been resolved have been reported in rice, indicating that the bc mutant gene in rice controls the culm mechanical strength mainly by influencing cellulose metabolic enzymes [14,35]. A phylogenetic analysis showed that LcCOBL1 and LcCOBL3 are homologous with OsBC1Lp1, OsBC1L1 and OsBC1L8 in rice, respectively, and may be involved in the metabolic synthesis of cellulose. Studies on Atcob-1 have shown that AtCOB is an important factor in the highly anisotropic expansion of plant morphogenesis by participating in the directional growth of cellulose microfibrils [36]. LcCOBL4 has high homology with AtCOB and may be involved in the directional expansion of cellulose microfibrils.
The expression patterns of the LcCOBLs in different tissues showed that all LcCOBLs except LcCOBL1 were highly expressed in stems, followed by roots. In the root, stem and leaf tissues of the plant, the cellulose contents in the stems, roots and leaves ranged from the highest to lowest, respectively, which further proves that LcCOBLs are associated with cellulose biosynthesis. In Arabidopsis, COBL proteins are key regulators of the direction of cell expansion in roots. In the identification of the expression sequence markers of potential homologous genes in other plants, it has been shown that COBL-gene-related functions may be necessary for all vascular plants [16]. The content of cellulose in the root tips of Arabidopsis mutants was significantly reduced, and the root cells were laterally expanded, suggesting that the regulation of cell directional expansion by COB is related to cellulose deposition [11]. In rapeseed (Brassica napus), RNA-seq analysis showed that BnaCOBL9, BnaCOBL35 and BnaCOBL41 were highly expressed in stems with high breaking resistance and may be involved in the stem development and stem-breaking resistance of rapeseed [37]. Combining these findings with those of previous studies, LcCOBL genes may be involved in plant cellulose synthesis and enhance stem fracture resistance.
To investigate the functions of the LcCOBLs in cold stress, we used qRT-PCR to analyze the expression patterns of the LcCOBLs in stems under low-temperature (4 • C) treatment. Over different processing times, different LcCOBLs showed different expression trends. Moreover, in the cis-elements analysis of LcCOBL promoters, each LcCOBL gene had elements related to low temperature. These results suggest that these LcCOBL genes may help improve the cold tolerance of plants. Similarly, previous studies have shown that COBL genes play important roles in drought and salt tolerance in other species [19,26,27].

Subcellular Localization of LcCOBLs and Their Associations with Potential Functions
Plant extracellular pH is acidic (pH 5.7), and many physiological and external environmental factors can cause changes in extracellular pH [38]. It was found that GFP was pH sensitive and its fluorescence decreased significantly under acidic conditions [39]. Although the shape of the spectrum does not change significantly with pH, the intensity gradually decreases with lowered pH, decreasing to 50% of maximal intensity at a pH of 6 [40,41]. TMHMM results showed that only LcCOBL2 had 1 transmembrane domain. However, a glycosylphosphatidylinositol (GPI) anchor is an alternative means of attaching a protein to the membrane [42]. As LcCOBL proteins are anchored to the outer layer of the cytoplasmic membrane and in an acidic environment, the GFP signal is weakened, resulting in an unclear fluorescence signal in the result images. The current experimental results show that, for the LcCOBL5-GFP fusion vector, the GFP signal was found in the cytomembrane, the weak fluorescence signal was also observed in the cytoplasm (Figure 7). However, we cannot determine the expression of LcCOBL2, 4 proteins in other organelles except cytomembrane ( Figure S7). This requires further experiments to verify.
The result of subcellular localization indicated that the LcCOBL5 in L. chinense calli and onion epidermal cells are localized in the cytomembrane and cytoplasm, which is consistent with the results of other species. By imaging living cells in Arabidopsis, it was found that COBL was located in particles in the plasma membrane [1,43]. In onion epidermal cells bombarded with an Ubi::OsBC1L4:GFP vector, GFP signals were found in the cell wall and plasma membrane [35]. This is consistent with the results of COB being observed in the elongation region of Arabidopsis roots using transmission electron microscopy. COB exists in the plasma membrane and some cell walls [36]. However, in Cunninghamia lanceolata, the fused ClCOBL1-RFP protein is transformed into tobacco cells. RFP signals mainly exist in the cell wall and membrane, but there is also a weak signal in the cytoplasm [43]. The COBRA gene encodes a plant-specific glycosylphosphatidylinositol (GPI)-anchored protein (GAP). GPI biosynthesis and transfer to proteins are carried out on the endoplasmic reticulum [13,44]. It may be that the process of GPI biosynthesis and transfer leads to the detection of fluorescence signals in the cytoplasm. Based on the above analysis, there are differences in the expression positions of COBL genes in different species, but most of the members are mainly expressed in the plasma membrane, a few in the cytoplasm, indicating that LcCOBLs participate in a variety of biological regulation processes.

Phylogenetic Analysis and Classification of the LcCOBL Family
According to the classification of the COBL proteins of Arabidopsis, 7 COBL proteins of L. chinense were divided into 2 groups. The COBL family protein sequences of A. thaliana, O. sativa, V. vinifera and A. trichopoda were downloaded from the Phytozome database (https://phytozome-next.jgi.doe.gov/) (accessed on 28 March 2022). The full-length COBL protein sequences were aligned using the clustalx tool. TrimAl ('automated1' mode) was used to trim the aligned sequences to generate the trimmed MSA file, which was then used to build the COBL phylogenetic tree. Using BEAUti software, the clipped FASTA file was output into an XML format. After the BEAUti program was completed, the TreeAnnotator program was used to construct the Bayesian phylogenetic tree. The TreeAnnotator program set the posterior probability limit to 1.0, the Burnin percentage to 90, the target tree type to maximum branch confidence tree, and the node height to common ancestor heights. Finally, the tree was visualized with the online iToL (https://itol.embl.de/) (accessed on 1 March 2023).

Analyses of the Chromosomal Locations, Structures and Conserved Motifs of LcCOBL Genes
The phylogenetic relationships between the LcCOBL genes were analyzed with MEGA X using the maximum likelihood (ML) method with 1000 bootstrap replicates. The position information of the LcCOBL genes on the chromosome was picked up from the L. chinense annotations using TBtools [49]. MEME (https://meme-suite.org/meme/doc/meme.html) (accessed on 31 March 2022) was further carried out to investigate their conserved motifs [50]. The GFF annotation files and the conserved structure sequences were imported into TBtools to visualize the exon-intron structures and COBRA domains. The GFF annotated files were obtained from the L. chinense genome file, and the conservative structural sequences were obtained from the NCBI (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi) (accessed on 30 October 2022). In addition, the software PlantCARE (https://bioinformati cs.psb.ugent.be/webtools/plantcare/html/) (accessed on 7 April 2022) was used to predict the cis-acting elements of the 7 LcCOBL genes in the upstream 2000 bp range. The Liriodendron hybrids plants used in this study were obtained from the Key Laboratory of Forest Genetics and Biotechnology of the Ministry of Education at Nanjing Forestry University. Liriodendron hybrids seedlings generated via somatic embryogenesis were cultured in a greenhouse under light 16 h/dark 8 h conditions (temperature, 25 • C; humidity, 70%).
In the tissue-specific expression experiments, leaf, root and stem tissues were taken from 3-month-old tissue culture bottle seedlings. Under abiotic stress, 3-month-old tissue culture bottle seedlings were cultured in a cold (4 • C) growth chamber. Stem tissues from the treatment group and control group were collected at 0, 6, 24 and 48 h after the experiment and stored at −80 • C.
Total RNA was isolated from each sample using the Eastep ® Super Total RNA Extraction Kit (Promega), and first-strand cDNA was synthesized from the proposed RNA using a PrimeScript RT Master Mix (Takare). A qPCR SYBR Green Master Mix (Vazyme) was used for real-time quantitative PCR, and the GAPDH gene was used as the internal control gene [51]. The real-time PCR cycling parameters were 95 • C for 30 s, followed by 45 cycles at 95 • C for 5 s and 60 • C for 30 s, with a melting curve analysis. All reactions were performed in triplicate to ensure the repeatability of the results. Gene expression levels were calculated using 2 −∆∆Ct .
The seedlings generated through somatic embryogenesis grown at 22 • C, light 16 h/dark 8 h, 75% relative humidity were treated at 4 • C, 40 • C or 15% PEG 6000 for 1 h, 3 h, 6 h, 12 h, 1 day and 3 days, respectively, and were treated with cold, heat and drought. Each treatment consisted of five replicates for each sampling time, taking leaf tissue samples for RNA-seq analysis. The transcriptome data about cold and heat stress (PRJNA679089) and drought stress (PRJNA679101) could be downloaded from NCBI [52]. The transcriptome data are in regards to eight organs from the author of this paper [30].

Subcellular Localization
In order to verify the prediction results of the subcellular localization of the LcCOBL proteins, we cloned the full-length CDSs of three LcCOBLs, i.e., LcCOBL2, LcCOBL4 and LcCOBL5. Then, these LcCOBL genes without terminating codons were inserted into pJIT166 vectors to obtain LcCOBL-GFP fusion vectors, which were driven by a 35S promoter. Then, these three vectors were separately transferred into the protoplasts of L. chinense calli via PEG-mediated protoplast transformation and transferred into onion epidermal cells by gene gun. At the same time, the 35S:H2B-mCherry vector was transferred into callus protoplasts as the control for the nuclear localization. The plasmids of the corresponding vectors were extracted, and the final concentration reached 1 ug/uL for use. Appropriate amounts of calli were put into the culture dishes, and an enzymatic solution was added to separate the cell walls from the protoplasts. They were then placed in a shaker at 27 • C at 40 rpm and enzymolyzed in the dark for 3-6 h. The carrier plasmid was added to the isolated and purified protoplast solutions, and a PEG solution of equal volume was added. The vectors were transferred into protoplasts via PEG mediation. Finally, the positions of GFP and mCherry were observed and photographed with a confocal laser microscope.
In order to further verify the subcellular localization of LcCOBL proteins, onion epidermal cells were transformed with the 35S: LcCOBL2, 4 5-GFP vectors by gene gun, with 35S: GFP vector as a control. The bombarded cells were incubated in the dark at 22-24 • C for 12-24 h to allow transient expression of the proteins. The GFP location was observed and photographed with a fluorescent microscope.

Conclusions
In this study, a genome-wide analysis of the LcCOBL gene family was conducted by focusing on their genetic structures and responses to stress treatments, and a total of seven LcCOBL genes were identified. The chromosomal distributions, gene structures and motifs, cis-regulatory elements in the promoter region, and expression patterns of the LcCOBL genes under different stress treatments and subcellular localizations were analyzed. The subcellular localization analysis and experimental results show that the LcCOBL genes are localized in the cytoplasm. The tissue expression pattern analysis of the LcCOBL genes shows that the LCCOBL genes are highly expressed in plant stems and may be involved in cellulose biosynthesis. The LcCOBL genes showed different response trends under cold, heat and drought stress treatments, indicating that they play a role in coping with environmental stresses. These results provide basic information on the COBL gene family and a good platform for exploring the specific roles of these genes in stress tolerance and the development of L. chinense.

Supplementary Materials:
The following supporting information can be downloaded at https://ww w.mdpi.com/article/10.3390/plants12081616/s1. Figure S1: Hydrophobicity analysis of amino acid sequences of LcCOBLs; Figure S2: Prediction results of GPI modification sites of protein sequences of LcCOBL; Figure S3: Prediction results of signal peptides of protein sequences of LcCOBL; Figure S4: Prediction results of transmembrane domains of LcCOBL proteins; Figure S5: ClustalW was used for multi-sequence alignment of LcCOBL protein sequences; Figure S6: Amino acid sequences of each motif; Figure S7: Subcellular localization of LcCOBL2, 4 proteins in L. chinense calli and in onion epidermal cells; Table S1: Protein names and sequences of COBL; Table S2: Primer sequences of LcCOBL proteins for subcellular localization; Table S3: Primer sequences of LcCOBLs and reference genes in qRT-PCR.

Data Availability Statement:
The original contributions presented in this study are publicly available. The L. chinense protein sequences can be found at https://hardwoodgenomics.org/Genome-asse mbly/2630420 (accessed on 3 March 2022). The RNA sequences can be obtained from the NCBI, accession numbers PRJNA679089 and PRJNA679101.