Genome-Wide Identication, Expression and Potential Function Analysis of the ERF and DREB Subfamily Members in Tomato

Background: APETALA2/ethylene responsive factors (AP2/ERFs) are unique regulators found in the plant kingdom that are involved in all life activity processes, including owering, fruit ripening, oral meristem growth, and defense responses. In tomato (Solanum lycopersicum), there are 60 DREB and 80 ERF subfamily members, however, their functionality remains poorly understood. Results: In this work, the AP2 domain conserved amino acid sequences of 68 ERF proteins from 20 plant species were compared and a Multiple Em for Motif Elicitation (MEME) analysis was conducted. Results revealed that the 9 th amino acid of the AP2 domain exhibited marked characteristics during the selection of DRE/CRT and/or GCC boxes as protein binding sites. Moreover, motifs near the AP2 domain may be involved in protein binding to DNA, whereas motifs far away from the AP2 domain may function as a part of the transactivation domain. Furthermore, we compared the expression levels of all ERF genes in 30 tomato organs and under biotic and abiotic stresses. Results indicated that most of 17 ERF and DREB repressor genes were highly expressed in almost all tomato organs and under some biotic and abiotic stress. The transcripts per million (TPM) value ratios of all repressor genes exceeded that of all activator genes in 16 tomato organs. Thus, it can be inferred that these repressor genes play vital roles in balancing the regulatory functions of activator genes and activator genes may also conversely compete with repressor genes to ensure normal growth, development, and defense responses in tomato. Conclusions: This work uncovered the potential functions of all ERF and DREB genes that regulate tomato growth, development, and defense responses, and considers the binding ability of the AP2 domain unique sequences with DRE/CRT and GCC boxes, as well as the relationship of unique motifs with the transactivation domain. These ndings will expand upon our understanding of the functions of ERF and DREB genes in tomato. AP2 ERF GCC TSWV in species SW-7 Fla.8059. SlERF2-6 was down-regulated by the STV, in contrast to the TYLCSV. was down-regulated the TSWV and F. oxysporum. repressor genes, SlERF7-2, SlERF3-16, SlERF9-10, SlERF4-11, SlERF9-1, SlERF2-10, and SlERF4-10, not the STV, TYLCSV, tomato psyllid, TSWV, or F. oxysporum (Fig. 5c; Data S3). These results suggest that some repressor genes are involved in regulating tomato tolerance to abiotic and biotic stress.

Solanum lycopersicum (tomato), as an important fruit vegetable, is widely planted in many countries. Tomato fruit is abundantly nutrition and has a unique avor, and can be eaten raw, boiled, or processed into ketchup or juice. Thus, improving the fruit yield and quality of tomato is the primary goal of tomato production. To achieve this goal, our understanding of the underlying molecular mechanisms of different vital processes must be enhanced, including seed germination, fruit ripening and softening, ower development, and defense responses to biotic and abiotic stresses. Among these processes, ERFs as regulators or repressors play important roles that affect different gene networks. In this study, we identi ed, corrected, and analyzed all ERF and DREB subfamily members based on S. lycopersicum genome database versions 2.0, 3.2, and 4.0. To understand the potential functions of ERF and DREB subfamily members, several RNA sequencing databases were used to analyze gene expression levels during vital processes, including growth, development, and defense responses to biotic and abiotic stresses. These works will help establish the regulatory networks of ERF and DREB subfamilies and uncover effective ways to improve tomato yield and quality.

Identi cation of ERF and DREB subfamily members in tomato
The genome sequences of S. lycopersicum were downloaded from a database of gene annotations, SGN, (Release v4.0 and v3.2, http://solgenomics.net/organism/solanum_lycopersicum/genome). The hidden Markov model (HMM) pro le of the AP2 domain (PF00847) was downloaded from the Pfam database (http://pfam.xfam.org/). HMMER v3.3 was used to search for candidate AP2/ERF genes from the tomato genome database. The default parameters were used and the cutoff value was set to 0.001. All of the candidate AP2/ERF proteins with only a single AP2 domain were selected as candidate ERF proteins. The Pfam, SMART (http://smart.embl-heidelberg.de/), and NCBI CDD databases (https://www.ncbi.nlm.nih.gov/cdd) were used to validate the candidate ERF proteins. Finally, the identi cation results of the 3 genome versions (2.0, 3.2, and 4.0) and NCBI database were compared to determine the nal ERF subfamily members of S. lycopersicum.

Phylogenetic analysis
Multiple sequence alignments of the tomato ERF proteins were performed using CLUSTAL W based on the complete sequences. To understand the relationship among the tomato ERF proteins, a phylogenetic tree was inferred using the maximum likelihood method based on the Whelan and Goldman model [23] of MEGA v7.0 with the following parameters: JTT + G model, partial deletion with 80% site coverage cutoff, and 1000 bootstrap replications [24].

Gene structure and conserved motif analyses
According to the cluster analysis results of the tomato ERF gene subfamily, the structural domain analysis of the ERF protein sequences of different groups was conducted using Jalview software [25]. Homologous alignments were compared using T-Coffee software [26]. The protein sequences of non-conservative regions were deleted. The alignment results were preserved in EPS format. Conserved motifs of the tomato ERF subfamily proteins were identi ed using the Multiple Em for Motif Elicitation (MEME) online tool v5.1.1 (http://meme-suite. org/tools/meme) with the following parameters: number of occurrences of a single motif distributed among the sequences within the model, 0 or 1 per sequence; maximum number of motifs, 20; optimum width of each motif, 6-50 residues.
The transcripts per million (TPM) expression values of the transcriptomes of different organs, biotic and abiotic stressors of tomato were obtained using the SRA toolkit and Salmon software [32]. Subsequently, the TPM values were processed to quantify of gene expression levels of the original data. The expression heat map of the ERF genes in different organs, and biotic and abiotic stressors of tomato were drawn using R-pheatmap based on the TPM values.

Sequence correction of ERF and DREB subfamily genes
To ensure the sequence accuracy of all AP2/ERF genes, the Pfam model (pf00847) of the AP2 domain downloaded from the Pfam website was used to search the tomato v4.0 protein database. A total of 166 AP2/ERF proteins with an AP2 domain E-value < 0.001 were obtained. Among these proteins, 20 had ≥ 2 AP2 domains, while 146 proteins had single AP2 domain. Among the latter, 3 proteins with the B3 domain were RAVtype AP2/ERF proteins. Thus, there were 143 ERF subfamily proteins with a single AP2 domain. The 143 protein sequences were submitted to the Pfam, CDD, and smart websites for conservative domain analysis. Subsequently, 140 tomato ERF subunit genes with a single AP2 domain were identi ed. The sequences of these genes were compared in 3 tomato genome sequencing DNA, CDS, cDNA, and protein databases (versions 2.0, 3.2, and 4.0); 26 genes were found to be different (Table S1). The CDS and protein sequences of these 26 genes were compared and con rmed according to the tomato genome and NCBI databases (Tables S2 and S3). Finally, the corrected protein sequences were used for subsequent analyses (Tables S2). Characteristics, polarity, and chemical structure analysis of the 14th and 19th amino acids in the AP2 domain Among the 140 ERF genes with a single AP2 domain, the 14th amino acid of the AP2 domain was V in 57 genes. Among these 57 genes, the 19th amino acid of the AP2 domain was glutamic acid (E) in 30 genes, aspartic acid (D) in 4 genes, ssparagine (N) in 1 gene, glutamine (Q) in 4 genes, histidine (H) in 6 genes, leucine (L) in 10 genes, alanine (A) in 1 gene, and V in 1 gene (Tables 1 and S4). These 57 genes were identi ed as DREB genes. Additionally, the 14th and 19th amino acids of the AP2 domain were isoleucine (I) and D, respectively, in SlERF2-5, SlERF10-6, and SlERF10-8. The codon of I was AUA/AUC, GUA/GUG/GUU/GUC for V, but GCA/GCG/GCU/GCC for A. The characteristics, polarity, and chemical structure of I and V were hydrophobic, nonpolar, and aliphatic, while A was neutral, nonpolar, and aliphatic (Tables 1 and S4). Thus, I can only be a V mutation. Accordingly, the 3 genes were identi ed as DREB genes. In the 19th amino acid of the AP2 domain, the hydrophilic amino acids included E, D, N, Q, and H, the hydrophobic amino acids included L and V, and the neutral amino acids included A. The negative charged amino acids (E and D), uncharged amino acids (N and Q), and positively charged amino acids (H) were polar; the nonpolar amino acids included L, A, and V. Additionally, H had a heterocycle chemical structure, while the others were aliphatic (Table 1). These differences may affect the functionality of DREB protein interactions with DRE and GCC boxes. Among the 80 ERF subfamily members, the 14th and 19th amino acids of the AP2 domain were A and D in 70 genes. Additionally, there was an A and tyrosine (Y) in 1 gene, A and N in 1 gene, threonine (T) and D in 1 gene, serine (S) and D in 4 genes, E and D in 1 gene, glycine (G) and N in 2 genes, and I and V in 1 gene (Tables 2 and S5). In the 14th amino acid of the AP2 domain, the neutral amino acids included A, T, S, and G, and the hydrophilic amino acid included E. The nonpolar amino acid was A, the polar amino acids without charges were T, S, and G, and the chemical structure of these amino acids is aliphatic. In the 19th amino acid of the AP2 domain, D, Y, and N comprised the hydrophilic amino acids, the negative charged amino acids (D), the uncharged amino acids (Y and N) were polar, and the chemical structure of these amino acids was aliphatic (Table 2). Thus, the 80 genes with a single AP2 domain were identi ed as ERF subfamily members. These differences may affect the functionality of ERF protein interactions with GCC boxes.
Unique amino acids affected the ability of protein to bind with DRE and GCC boxes Previous studies found that some DREB and ERF subfamily proteins only bound to DRE or GCC boxes, but most of these proteins can also interact with these boxes. However, the correlation between the characteristics and binding ability of DREB and/or ERF subfamily proteins remains unclear. To distinguish the difference between DREB and ERF proteins during binding with DRE or GCC boxes, the AP2 domain amino acid sequences of 49 Arabidopsis and 19 other species ERF proteins, including 8 tomato ERF proteins, were compared. The binding assays of the 68 ERF proteins with DRE and GCC boxes were completed through an electrophoretic mobility shift assay (EMSA), yeast one-hybrid, or proteome chip assays. Among these proteins, there were 42 protein AP2 domains that included P9, 5 included H9, 5 included S9, 6 included N9, 3 included Q9, 2 included K9, 2 included T9, and 1 included I9 (Fig. 3 (Fig. 3). These results suggest that almost all ERFs with P9 and H9 can interact with GCC box, and most can also bind with DRE. All DREB with S9 can only interact with DRE, but other DREBs with N9, K9, Q9, T9, and I9 may only bind with DRE or with DRE and GCC boxes. The A14 and A15 amino acids of ERF AP2 domain were conserved, but the 13th amino acid may be Y, F, or W. The W13 and V14 amino acids of the DREB AP2 domain were conserved, but the 15th amino acid may be S, A, or C (Fig. 3). These characteristics of ERFs and DREBs may affect the ability of proteins to bind with DRE and GCC boxes.

Expression analysis of CBF genes in tomato
To understand the function of different ERF and DREB genes in tomato growth and development, the expression levels of ERF and DREB subfamily members were analyzed according to tomato RNA-Seq data abtained from NCBI SRA library. Seven tomato CBF proteins had the PKKPAGR motif in the N-terminal and the DSAWR motif in the C-terminal of the AP2 domain, but these highly similar proteins to CBF did not have the same motif (Fig. 4a). CBF proteins had N/D9W13V14C15 in β2, while other proteins had N/S/A9W13V14S15 (Fig. 4a). This analysis suggests that CBF and its highly similar proteins may process different abilities with GCC and/or DRE boxes, which leads them to play different roles in regulating tomato growth, development, and defense responses to abiotic stress.
In different tomato organs, SlERF8-2 was mainly expressed in mature petals and sepals, cotyledons, hypocotyl, and leaf lamina and veins; SlERF8-3 was only expressed in mature petals and sepals, and hardly in other organs. SlERF12-11 and SlERF1-3 were hardly expressed in all organs, except SlERF1-3 in young stamen. SlERF3-6 was mainly expressed in mature sepals, hypocotyl, leaf lamina and veins, and internodes. SlERF3-7 was expressed in mature anthers and petals, young and mature sepals, meristems after transitioning to owering, cotyledons, hypocotyl, leaf lamina and veins, root apexes and roots without apexes, and internodes, especially in cotyledons, hypocotyl, and leaf lamina. SlERF3-22 was expressed in mature anthers, petals and sepals, meristems after transitioning to owering, cotyledons, hypocotyl, leaf lamina, and internodes ( Fig. 4b; Data S1). These results indicate that 6 CBF genes, except SlERF12-11, may be involved in regulating the development of some oral organs, meristems after transitioning to owering, cotyledons, hypocotyl, leaves, roots, and internodes. Although other highly similar DREB proteins to CBF do not have PKKPAGR and DSAWR motifs, most of these proteins play roles in some tomato organs. For example, SlERF6-1 was expressed in mature petals and sepals, meristems after transitioning to owering, cotyledons, hypocotyl, and leaf veins. SlERF8-12 and SlERF8-11 were highly expressed in mature anthers and young sepals, respectively, but were lowly expressed or undetectable in other organs. SlERF3-20 was highly expressed in some organs, including mature anthers, oral meristems, meristems after transitioning to owering, vegetative meristems, hypocotyls, leaf veins, young leaves, and internodes; SlERF6-8 was highly expressed in hypocotyls and internodes. These results indicate that CBFs and some DREB genes that are highly similar to CBF regulate the development of the same organs, such as the expression of SlERF8-2, SlERF3-7, SlERF3-22, SlERF6-1, and SlERF3-20 in meristems after transitioning to owering ( Fig. 4b; Data S1).
Among 7 CBF genes, SlERF12-11 and SlERF1-3 did not respond to abiotic stress, including drought, heat, cold, strong light, and red light. SlERF8-2 and SlERF8-3 positively responded to cold and red light, but not to drought, heat, or strong light. SlERF3-6 was positively induced by heat and red light, negatively regulated by drought, and did not respond to cold or strong light. SlERF3-7 was positively induced by heat, cold, and strong and red lights, but did not change due to drought. SlERF3-22 responded to cold, and strong and red lights, but did not change due to drought or heat ( Fig. 4c; Data S2). These results suggest that different CBF genes may play different roles in response to different abiotic stressors. However, some DREB genes that are highly similar to CBF also responded to abiotic stress, including SlERF6-1, SlERF1-13, SlERF12-9, and SlERF6-8, which up-regulated by drought, and SlERF6-1, SlERF8-4, SlERF8-12, and SlERF12-2, which up-regulated by cold. SlERF9-1 and SlERF10-9 were not expressed or induced by abiotic stress, while SlERF8-10 and SlERF8-11 were lowly expressed under abiotic stress ( Fig. 4c; Data S2).
Among the 17 repressor genes, at least 13 were expressed in every organ, especially 16 in green large seeds ( Fig. S1a; Data S1). Most activator genes were expressed in every organ, including a maximum of 94 genes in mature petals and minimum of 65 in red pulp ( Fig. S1a; Data S1). However, the total TPM values of 17 repressor genes had a very high ratio among all ERF genes in every organ and exceeded the activator genes in most organs, including 56.14% in mature owers, 57.64% in oral meristems, 56.58% in vegetative meristems, and 60.05% in cotyledons ( Fig. S1b; Data S1). These results suggested that the 17 repressor genes play important roles in balancing the regulatory functions of other ERF and DREB subfamily genes and their downstream target genes during tomato growth and development.
All II-A subgroup genes did not respond to drought or strong light stress, but some of these genes were up-regulated by heat, cold, and red light. For example, the expression of SlERF4-2, SlERF1-16, and SlERF6-6 increased under cold and red light stress, and SlERF9-6, SlERF12-10, and SlERF6-6 increased under heat stress ( Fig. 6c; Data S2). These results indicate that some of the II-A subgroup genes improve heat, cold, and red light tolerance. However, all II-A subgroup genes did not respond to biotic stress and some genes were even down-regulated, including SlERF1-16 and SlERF10-3 under F. oxysporum treatment ( Fig. 6c; Data S3). These results suggest that the II-A subgroup genes were not involved in enhancing the tolerance to biotic stress.
Among the 14 DREB genes, SlERF4-6 and SlERF4-9 were highly expressed in all tomato organs, but other genes were speci cally expressed in some organs, including SlERF3-14 in young stamens, young sepals, and senescent leaves, SlERF6-4 in young sepals and seeds, and SlERF3-15 in red and dry seeds. SlERF12-5 was not expressed in any tomato organs ( Fig. 7b; Data S1). These results suggest that the I-B subgroup genes, except SlERF12-5, are involved in regulating different tomato growth and development processes.
Under abiotic stress, the expression of all I-B subgroup genes did not increase and SlERF8-5 was down-regulated by drought stress. Under heat stress, the expression of SlERF6-4 and SlERF8-14 decreased, SlERF4-9 and SlERF7-4 increased, and other genes did not change. SlERF8-14, SlERF4-9, and SlERF7-4 were up-regulated by cold stress, while other genes did not change. Almost all I-B subgroup genes did not respond to strong or red light stress, but SlERF3-14 and SlERF6-7 were positively regulated and SlERF6-4 was negatively regulated by red light (Fig. 7c; Data S2). All I-B subgroup genes did not positively respond to biotic stress ( Fig. 7c; Data S3). These results indicate that a few I-B subgroup genes play roles in abiotic stress responses and all I-B subgroup genes are not involved in biotic stress responses.

Discussion
Conserved sequence characteristics of the AP2 domain affected the ability of ERF proteins to bind with DRE and GCC boxes DRE/CRT and GCC boxes are two cis-acting elements that ERF and DREB subfamily members bind to in higher plants. In previous reports, the 14th and 19th amino acids of the AP2 domain were used to distinguish DREB and ERF proteins, including V14E19, and A14D19, which were respectively named DREB and ERF proteins. Additionally, DRE/CRT and GCC boxes only interact with DREB and ERF proteins, respectively [11]. However, an increasing number of reports have proved that some DREB and ERF proteins can interact with both DRE/CRT and GCC boxes [34,37,38,44,53,54], while a few DREB proteins can only bind with GCC box and a few ERF proteins can only bind with DRE/CRT element [34]. These ndings suggest that the 14th and 19th amino acids of the AP2 domain play certain roles in interacting with DRE/CRT and GCC boxes, but some other amino acids also have important functions.
In this study, 140 AP2/ERF proteins with a single AP2 domain in tomato were classi ed into 6 groups (Fig. 1). Among these proteins, 60 and 80 members were distinguished as DREB-and ERF-type, respectively. Among the DREB-type proteins, there were 30 proteins with V14E19, 27 with V14 and D/N/G/H/L/A/V19, and 3 proteins with I14 and D/V19. E and D have the same characteristics, but others exhibit different characteristics (Table 1). Thus, AP2 domains with either E19 or D19 are not affected by the ability to bind with DRE/CRT or GCC boxes. For example, At5g19790 with V14D19 and At1g75490 with V14E19 can interact with DRE/CRT and GCC boxes [34]. Among the ERF-type proteins, there were 70 proteins with A14D19, 4 with A14Y19, A14N19, T14D19, and E14D19, 4 with S14D19, and 2 with G14N19 ( Table 2). In tomato, SlERF3-12 and SlERF5-8 with A14D19 only bind to GCC box [33], but SlERF9-9, SlERF3-21, SlERF6-6, and SlERF9-6 with A14D19 can interact with DRE/CRT and GCC boxes [39,[41][42][43]. In addition, SlERF5-7 with S14D19 also bound to DRE/CRT and GCC boxes (Fig. 3) [36]. These results suggest that the 14th and 19th amino acids of the AP2 domain cannot be the only standard by which to evaluate the ability of ERF and DREB subfamily members binding with DRE/CRT and GCC boxes.
In this study, we compared the AP2 domain sequences of 68 ERF and DREB proteins from 20 species. The ability of these proteins to bind with DRE/CRT and GCC boxes were demonstrated by EMSA, yeast one-hybrid, and proteome chip assays. We found that 46 proteins with P/H9 in the AP2 domain, except At4g23750, can interact with GCC box, 28 of which can also bind with DRE/CRT box (Fig. 3). However, 5 DERB proteins with S9 in the AP2 domain only interacted with DRE/CRT box. Other site sequences of these 5 proteins did not display some typical differences compared to other DREB proteins that bind to DRE/CRT and GCC boxes (Fig. 3). These results indicate that the 9th amino acid of the AP2 domain plays an important role in protein selection at its binding sites for DRE/CRT and/or GCC boxes. In tomato, 11 DREB proteins had S9, and 70 ERF and 19 DREB had P/H9 at the 9th amino acid of the AP2 domain. These results suggest that DREB proteins with S9 may only interact with DRE/CRT box, and ERF and DREB proteins with P/H9 may bind to GCC box or to both DRE/CRT and GCC boxes. However, the ability of these proteins without S/P/H9 binding with DRE/CRT and GCC boxes is poorly understood. For example, the 9th amino acid of 6 CBF proteins in tomato is N or D, and some Arabidopsis DERB proteins with N9 can interact with only DRE/CRT or both DRE/CRT and GCC boxes [34,55,56]. The abilities of DREB proteins with D9 were not demonstrated by EMSA, yeast one-hybrid, or proteome chip assays.
Functionally, ERF and DREB subfamily proteins work as activators or repressors in plant growth, development, and defense responses. Proteins, including EAR motif, can repress the expression of downstream target genes by recruiting a histone deacetylase complex to affect chromatin structures [57][58][59]. However, although some reports have veri ed that EDLL motif is an activating domain [60][61][62], the activating domain of ERF and DREB subfamily proteins is complex and remains elusive. In tomato, EDLL motif is only displayed in some the II-B group members (Fig. S2), but this motif was included in motif 6. Some members of the I-A group including CBF and a few members of the II-B group also had motif 6 (Fig. 2). These ndings suggest that motif 6 instead of the EDLL motif may be a more reliable transactivation domain. The MEME analysis results suggest that motifs 1, 2, 3, and 4 are located in the AP2 domain, while others are distributed outside the AP2 domain. Thus, motifs relatively distant from the AP2 domain, including motifs 7-9, 11, 12, 14, 17-19, and 22-24 may also act as a part of the transactivation domain. Motifs near the AP2 domain, including motifs 5, 10, 13, 15, 20, 24, and 25 may help the AP2 domain bind to DRE/CRT and GCC boxes (Fig. 2). For example, the PKKPAGR signature sequence of Arabidopsis CBF1 is located near the left of the AP2 domain and mutations within this motif reduce the ability of CBF1 to bind to DRE/CRT element and decrease the expression levels of the COR gene [63]. In this study, motif 10 included the PKKPAGR signature sequence and was located near the left of the AP2 domain in 7 tomato CBF proteins (Fig. 4a). These results indicate that these motifs near the AP2 domain may play important roles in helping the AP2 domain to bind to DRE/CRT and GCC boxes, but the functions of motifs far away from the AP2 domain remain unclear during the regulation of transcription of downstream genes.
ERF repressors balance the regulatory functions of other ERF activating genes during tomato growth, development, and defense responses AP2/ERF TFs play important roles in regulating plant growth, development, and defense responses. For example, AtERF115 can repress adventitious rooting in Arabidopsis through the JA and CTK signaling pathways [64], but PUCHI (At5g18560) can positively regulate oral meristems, organ initiation, and lateral root development in Arabidopsis [12,13]. Rice OsERF101 regulates leaf senescence and responses to drought stress in reproductive tissues [65,66]. These ndings suggest that some ERF proteins often regulate several plant life processes. However, an ERF protein can only act as an activator or repressor. In this study, we found that there were 11 ERF and 6 DREB proteins including the EAR repression domain, in tomato. Among these proteins, SlERF7-3 and SlERF10-2 had 2 LxLxL and 1 DLNxxP, and SlERF3-4 had 1 LxLxL and 1 DLNxxP at the C-terminal. These results suggest that SlERF7-3, SlERF10-2, and SlERF3-4 may recruit more histone deacetylase complexes to repress the expression of their downstream target genes. The expression levels of SlERF10-2 (LeERF3b) markedly increased in low ET tomato fruit containing an ACC oxidase sense-suppression transgene and in the ET insensitive mutant never ripe (Nr), in contrast to SlERF5-10 (Pti4) without the EAR motif [67]. SlERF5-10 regulated fruit ripening, seed germination, and responses to biotic and abiotic stress [68][69][70][71]. In this study, SlERF7-3, SlERF10-2, and SlERF3-4 were expressed in several organs, especially SlERF7-3 and SlERF10-2, which were highly expressed in tomato all owers, fruits, meristems, seeds, leaves, and roots (Fig. 5b). The 3 genes also responded to some biotic and abiotic stressors (Fig. 5c). These results indicate that SlERF7-3, SlERF10-2, and SlERF3-4 as repressors are involved in several metabolic pathways that affect tomato growth, development, and defense responses. Other genes, including SlERF10-1, SlERF7-5, SlERF12-1, SlERF7-2, SlERF2-6, SlERF2-10, SlERF4-10, SlERF5-8, and SlERF4-1, were also highly expressed in several tomato organs. Overexpression of SlERF10-1 (SlERF36) caused early owering and plant senescence, and affected stomatal density, photosynthesis, and plant growth [72]. These ndings suggest that they may play the same or similar roles, like SlERF7-3, SlERF10-2, and SlERF3-4. However, SlERF9-1 was not expressed in any tomato organ or under biotic and abiotic stress. Moreover, its EAR motif was located in the AP2 domain terminal. Thus, SlERF9-1 may be an invalid gene. SlERF3-16, SlERF9-10, SlERF8-14, and SlERF4-11 were expressed some tomato organs and under biotic and abiotic stress (Fig. 5b and c). SlERF3-16 (ENO) was involved in the CLAVATA-WUSCHEL signaling pathway that regulates oral meristem development [73]. These results indicate that these genes play repression roles in some organs and under certain environmental conditions.
Although there are 60 DREB and 80 ERF subfamily members in tomato, only 11 ERF and 5 DREB genes, except SlERF9-1, as repressors were expressed in all or some tomato organs or under biotic and abiotic stress. In this study, we analyzed the expressed gene numbers of all repressors and activators in every tomato organ and found that at least 13 repressor genes were expressed in several oral organs, red pulp, cotyledons, and senescent leaves, and, at most, 16 repressor genes were expressed in green large seeds (Fig. S1a). However, most activator genes were expressed in every organ, including the maximum of 94 genes in mature petals and the minimum of 65 in red pulp (Fig. S1a). The TPM value ratios of all expressed repressors and activators were compared. Results indicated that the ratio of repressors exceeded that of activators in 16 tomato organs, especially in cotyledons (Fig. S1b), suggesting that the 16 repressor genes control the activation function of other ERF and DREB genes by competing with the binding sites of their target genes, and nally balancing the expression of their target genes to ensure normal growth and development.

Conclusions
This work highlights that much remains to be understood about the relationship between repressors and activators in tomato ERF and DREB subfamily members. By performing Pfam model (pf00847) search and sequences comparing in 3 tomato genome sequencing databases (versions 2.0, 3.2, and 4.0), 140 AP2/ERF genes were identi ed. These genes included 60 DREB and 80 ERF subfamily members and were classi ed into six subgroups. The expression pro les of DREB and ERF genes in 30 tomato organs as well as under biotic and abiotic stresses were investigated and compared, which could considered as the candidates for further study of their function in plant growth, development, and defense response.