*3.1. Identification of BAG Homologs in Plants*

Previous studies have demonstrated that the BAG protein family was evolutionarily conserved and highly similar in structure and function in eukaryotes [26]. Phylogenetic analysis of the *BAG* gene family across species was significant for understanding the differences in function or predicting similarities between tomato and other species. We identified 10 *BAG* genes in the tomato genome using the SGN database (https://solgenomics.net/ (accessed on 19 April 2020)) and named *BAG1-10* based on homology and evolutionary analysis with the *Arabidopsis* protein sequences (Figure 1A, Table S1). In light of the function of BAGs, we performed a phylogenetic analysis of BAG proteins from three dicot plants, *Arabidopsis*, tomato, and tobacco (*Nicotiana tabacum*), and a monocot plant, rice. Based on the resultant phylogenetic tree, the BAG proteins of the four species were divided into three subfamilies (Figure 1A). BAG5, BAG6, BAG8, and BAG9 belonged to the first group, BAG7 belonged to the second group, and BAG1, BAG2, BAG3, BAG4, and BAG10 belonged to the third group.

Then, we further analyzed the structural domains of the BAG proteins (Figure 1B). Results showed that all BAG proteins contained a conservative BAG domain. Furthermore, BAG1-4 and BAG10 contained extra ubiquitin-like (UBL) structural domains at the N-terminus, while BAG6, BAG8, and BAG9 each comprised an extra CaM-binding motif. In addition, BAG7 protein was distinguished since it had no other kinds of motifs but triple BAG domains. In terms of the length of BAG proteins, BAG5 was the shortest, while BAG6 had the longest sequence length.

**Figure 1.** Phylogenetic tree construction of BAGs from different plants and protein structures of BAGs. (**A**) Phylogenetic tree of BAG proteins. The different colored circles on the outside of the protein names represented the types of structural domains possessed by amino-acid sequences (blue for BAG domains, red for UBL domains, and green for calmodulin-binding (CaM) motifs). The symbols on the inside of the protein names represented different species (purple stars for *Arabidopsis thaliana*, green triangles for *Nicotiana tabacum*, blue squares for *Oryza sativa*, and red stars for *Solanum lycopersicum*). (**B**) Schematic diagram of the domains of BAG proteins in tomato. The protein lengths were shown in grey.

#### *3.2. Involvement of BAG9 in Tomato Thermotolerance*

Transcript analysis of 10 *SlBAGs* under heat stress was conducted to determine whether heat stress induced BAG gene expression.

Figure 2 displayed that exposure to heat within 1 h can quickly induced *BAG6*, *BAG8*, and *BAG9* and whose transcript levels subsequently reached a maximum after 3 h. Nevertheless, the expression levels of other *BAG*s were not changed or decreased after heat stress (Figure 2). These results suggested that *BAG6*, *BAG8*, and *BAG9* may be important in regulating tomato response to heat stress.

Then, we analyzed the cis-elements in promoters of *BAG* genes and found that only the *BAG9* promoter contained the heat shock element (HSE), which was transcriptionally regulated by heat shock factors under heat stress (Figure S2). To investigate whether BAG9 was involved in the regulation of plant thermotolerance, we generated the *bag9* mutants and *BAG9* overexpressing plants as described in the "Materials and Methods" section (Figure S1). As shown in Figure 3A, the phenotypes of *bag9* mutants and *BAG9* overexpressing plants were similar to WT plants, when they were grown under normal conditions (Figure 3A).

To examine how BAG9 functions in tomato under heat, *bag9* mutants, WT plants, and *BAG9* overexpressing plants grown for about 5 weeks were kept in a 45 ◦C growth chamber for 10 h. The exposure of tomato plants to heat stress resulted in plant withering and decreased *Fv/Fm* value, more significantly in *bag9* mutants compared with WT plants (Figure 3). In contrast, thermotolerance was significantly increased in *BAG9* overexpressing plants with higher *Fv/Fm* value (Figure 3). Moreover, heat stress inhibited photosynthesis in tomato plants. Net photosynthetic rate (*P*n) was decreased by 42.4% in *bag9* mutants

but was increased by 100.1% in *BAG9* overexpressing plants compared with WT plants (Figure 3D). Additionally, MDA accumulation was aggravated in *bag9* mutants, while alleviated in *BAG9* overexpressing plants compared with WT plants (Figure 3E). Thus, these results suggested that BAG9 played a positive role in tomato response to heat stress.

**Figure 2.** Transcripts of *BAG* genes in response to heat stress. Cluster analysis of expression patterns of *BAGs* under heat stress at 0 h, 1 h, 3 h, 5 h, and 7 h. The heat map was manufactured using log2 logarithmic-transformed expression values. The color transition from red to green on behalf of high to low expression levels. According to the expression, the *BAGs* were clustered in the figure. The data represented the means ± SD of three biological replicates.

**Figure 3.** Influence of BAG9 on tomato thermotolerance. (**A**) Representative images of *bag9* mutants, wild type (WT), and *BAG9* overexpressing (*BAG9*-OE) plants without or with heat stress. Bar = 10 cm. The plants were subjected to normal temperature (25 ◦C) or high temperature (45 ◦C) treatment for 10 h, photographs of plants were then taken. (**B**,**C**) After undergoing different temperature treatments for 7 h, images of representative leaves showed the maximum photochemical efficiency of photosystem II (*Fv/Fm*). At the bottom, a color gradient showed the strength of the fluorescence signal depicted by each color. (**D**) Net photosynthetic (*P*n) efficiency at 7 h under heat. (**E**) MDA content at 7 h under heat or without heat stress. Data were the means ± SD of three biological replicates. Different letters represented significant differences (*p* < 0.05) according to Tukey's test.

#### *3.3. BAG9 Alleviates Heat-Stress-Induced ROS Accumulation*

ROS production and scavenging keep homeostasis balanced in plants under normal conditions [42]. However, this homeostasis will be disturbed after heat-stress exposure [2]. To verify the effect of BAG9 on heat-induced oxidative stress, we first detected H2O2 and O2 •− accumulation. Tomato leaves were stained with DAB dye for histochemical detection of H2O2 and with NBT dye for O2 •− detection. As shown in Figure 4, heat stress induced H2O2 and O2 •− production in the leaves of WT plants. Interestingly, H2O2 and O2 •− production was significantly induced in *bag9* mutants, whereas it was reduced in *BAG9* overexpressing plants (Figure 4A). Similarly, the H2O2 content was quantitatively analyzed in support of the observation that H2O2 was more accumulated in *bag9* mutants, but significantly reduced in *BAG9* overexpressing plants compared with WT plants (Figure 4B).

**Figure 4.** The accumulation of reactive oxygen species (ROS) and oxidative proteins in tomato plants under heat stress. (**A**) Representative images of H2O2 and O2 •− accumulation were detected by DAB and NBT staining, respectively. Bar = 5 cm. (**B**) Quantification of H2O2 at 7 h under heat. (**C**) Oxidative proteins. An anti-DNP antibody was used to detect total proteins on SDS-PAGE. Coomassie Blue staining (CBB) was applied to indicate the protein input, and on the top of the image was the relative intensity of oxidative proteins. Three independent experiments were performed with similar results. Data were the means ± SD of three biological replicates. Different letters represented significant differences (*p* < 0.05) according to Tukey's test. WT, wild type; *BAG9*-OE, *BAG9* overexpressing plants; RLS, Rubisco large subunit.

To further investigate whether heat-induced oxidative stress caused the oxidation of functional proteins, SDS-PAGE was used to analyze protein oxidation among proteins isolated from total proteins. Figure 4C illustrated that the accumulation of oxidative proteins was similar in *bag9* mutants, WT, and *BAG9* overexpressing plants under normal conditions. Mutants *bag9* and plants overexpressing *BAG9*, however, had increased and decreased levels of oxidative proteins, respectively, compared to wild-type plants. Thus, these results suggested that BAG9 reduced the accumulation of ROS and the oxidation of protein caused by heat.

#### *3.4. BAG9 Enhances Antioxidant Capacity under Heat Stress*

Antioxidant defense mechanisms contain antioxidant enzymes such as SOD, APX, GR, CAT, DHAR, POD, and antioxidants such as ASA and GSH to trap and scavenge free radicals and ROS, thereby protecting plant cells and organelles from destruction and increasing stress resistance [47]. As shown in Figure 5, heat stress increased all six antioxidant enzyme activities in WT and *BAG9* overexpressing plants. However, in *bag9* mutants, POD, APX, GR, DHAR, and CAT activities between control and heat treatment showed no significant difference (Figure 5). The enzyme activities in *BAG9* overexpressing tomato were higher than those in WT. According to these results, BAG9 promoted the activities of antioxidant enzymes under heat stress.

**Figure 5.** Activities of SOD, POD, APX, GR, CAT, and DHAR with or without heat stress in tomato leaves. Data were the means ± SD of three biological replicates. Different letters represented significant differences (*p* < 0.05) according to Tukey's test. WT, wild type; *BAG9*-OE, *BAG9* overexpressing plants.

To determine whether BAG9-induced thermotolerance was related to the state of cellular redox, the variation of contents and ratios of AsA/DHA and GSH/GSSG were examined (Figure 6). Heat stress had little effect on the AsA and GSH levels but significantly increased the DHA and GSSG contents, leading to significant declines in the AsA/DHA and GSH/GSSG ratios in all plants compared with control. Under heat stress, the DHA and GSSG contents were considerably increased in *bag9* mutants but reduced in *BAG9* overexpressing plants compared with WT plants. Meanwhile, ratios of AsA/DHA and GSH/GSSG were lower in *bag9* mutants but higher in *BAG9* overexpressing plants compared with WT plants (Figure 6).

**Figure 6.** Effects of heat stress on AsA and GSH pools in tomato leaves. Data were the means ± SD of three biological replicates. Different letters represented significant differences (*p* < 0.05) according to Tukey's test. WT, wild type; *BAG9*-OE, *BAG9* overexpressing plants.

#### *3.5. BAG9 Interacts with Hsp20s and Maintains Hsps Stability under Heat Stress*

We next identified BAG9-interacting proteins by applying yeast two-hybrid screens. Choosing the fused BAG9 protein as baits, we screened 6 × <sup>10</sup><sup>6</sup> independent transformants of a tomato cDNA prey library and identified more than twenty clones. The proteins encoded by these positive clones included four Hsp20s (Hsp17.7A, Solyc06g076520; Hsp17.7B, Solyc09g015020; Hsp17.6B, Solyc06g076560; Hsp17.6C, Solyc06g076570). Then, we performed yeast two-hybrid assays to explore whether BAG9 interacted with Hsp20s. By co-transforming the bait and prey vectors, we found that BAG9 interacted with four Hsp20 proteins in yeast (Figure 7A).

To determine whether BAG9 and Hsp20s interact in vivo, we performed a BiFC assay in *A. tumefaciens-infiltrated* tobacco. BAG9 was fused to the C-YFP vector (BAG9-C-YFP) and Hsp20s were fused to the N-YFP vectors (Hsp 17.7A, Solyc06g076520; Hsp17.7B, Solyc09g015020; Hsp17.6B, Solyc06g076560; Hsp17.6C, Solyc06g076570). When the BAG9- C-YFP was co-expressed with four Hsp-N-YFP in tobacco leaves, YFP signals were observed

in tobacco cells that had been transformed (Figure 7B). All these experiments revealed that BAG9 interacted with four Hsp proteins.

**Figure 7.** BAG9 interacted with Hsp20s. (**A**) Yeast two-hybrid assay showed interactions between BAG9 and Hsp17.7A, Hsp17.7B, Hsp17.6B, and Hsp17.6C. By growing yeast cells at different concentrations lacking Trp (T), Leu (L), Ade (A), and His (H), the interaction of proteins has been evaluated. (**B**) BiFC analysis showed that the interaction between BAG9 and Hsp20s took place in the cytoplasm. Spliced YFP fusion constructs were transiently coexpressed in *N. benthamiana* leaves for 2 d. The YFP fluorescence signals were obtained by confocal microscopy.

BAG9 and Hsps are both chaperones. To investigate whether BAG9 affects the stability of Hsps under heat stress, we examined the accumulation of Hsps by Western blotting. As shown in Figure 8, there was almost no difference in the accumulation under normal conditions. While heat stress induced a great accumulation of Hsp20, Hsp70, Hsp90, and Hsp101 in all genotypes. However, compared with WT, the accumulation of these four Hsps was still lower in *bag9* mutants, while higher in *BAG9* overexpressing plants (Figure 8). Thus, BAG9 promoted the stability of Hsps under heat stress.

**Figure 8.** The accumulation of Hsps with or without heat stress in tomato leaves. Hsp17.6, Hsp70, Hsp90, and Hsp101 were detected by immunoblot analysis. After exposing to heat for 7 h, the leaf samples were obtained for experiments. The protein input was indicated by Coomassie Blue staining (CBB). Three independent experiments were performed with similar results. WT, wild type; *BAG9*-OE, *BAG9* overexpressing plants; RLS, Rubisco large subunit.
