Triphosphate Tunnel Metalloenzyme 2 Acts as a Downstream Factor of ABI4 in ABA-Mediated Seed Germination

Seed germination is a complex process that is regulated by various exogenous and endogenous factors, in which abscisic acid (ABA) plays a crucial role. The triphosphate tunnel metalloenzyme (TTM) superfamily exists in all living organisms, but research on its biological role is limited. Here, we reveal that TTM2 functions in ABA-mediated seed germination. Our study indicates that TTM2 expression is enhanced but repressed by ABA during seed germination. Promoted TTM2 expression in 35S::TTM2-FLAG rescues ABA-mediated inhibition of seed germination and early seedling development and ttm2 mutants exhibit lower seed germination rate and reduced cotyledon greening compared with the wild type, revealing that the repression of TTM2 expression is required for ABA-mediated inhibition of seed germination and early seedling development. Further, ABA inhibits TTM2 expression by ABA insensitive 4 (ABI4) binding of TTM2 promoter and the ABA-insensitive phenotype of abi4-1 with higher TTM2 expression can be rescued by mutation of TTM2 in abi4-1 ttm2-1 mutant, indicating that TTM2 acts downstream of ABI4. In addition, TTM1, a homolog of TTM2, is not involved in ABA-mediated regulation of seed germination. In summary, our findings reveal that TTM2 acts as a downstream factor of ABI4 in ABA-mediated seed germination and early seedling growth.


Introduction
Seed germination, an early event in plant development, has a significant impact on the later vegetative and reproductive growth of plants [1]. Seeds are induced into primary dormancy after maturation and require prolonged dry storage or low-temperature water absorption (stratification) to release dormancy for germination under favorable environmental conditions [2,3]. Seed germination is regulated by various exogenous and endogenous factors [4][5][6] Among these factors, abscisic acid (ABA) and gibberellin (GA) are the major players associated with seed dormancy and germination, respectively [7,8], and the balance between these two hormones dictates the seeds germination ability [3].
Previous studies displayed that ABA content decreased rapidly after cold stratification of seeds [9] and exogenous application of ABA inhibits seed germination [10,11]. Mutation in CYP707As, key catabolism genes of ABA, results in higher ABA accumulation and leads to delayed seed germination compared with the wild type [12]. Further, ABA modulates the expression of numerous germination genes through its signaling pathways, such as EM1 and EM6 [13,14]. ABA is identified by its receptor PYR/PYL/RCAR family, and the ABA-receptor complex can bind to PP2Cs, thus inhibiting their activity and leading to the activation of SnRK2s and turning on ABA signaling [15][16][17]. ABA insensitive 3 (ABI3), ABI4, and ABI5 are downstream factors of ABA signaling, and all mutants of abi3, abi4, and abi5 exhibit insensitivity to ABA in seed germination [18][19][20].
TTM superfamily exists in all domains of life and hydrolyzes a range of organophosphate substrates, which comprise a group of enzymes [36]. This family has three members in Arabidopsis: TTM1, TTM2, and TTM3. TTM1 and TTM2 contain both the CYTH domain and P-cycle kinase/uric acid kinase domain, but TTM3 consists only CYTH domain [37][38][39]. Although there is a high sequence similarity between TTM1 and TTM2, these two members function differently in plants. For example, TTM1 but not TTM2 plays an active role in regulating plant leaf senescence. When treated with ABA or darkness, TTM1 phosphorylation was increased and the senescence-associated programmed cell death was accelerated [37,40]. TTM2 plays a negative role in plant resistance to oomycete response as ttm2 mutants exhibit enhanced hypersensitive response and increased salicylic acid accumulation, but TTM1 is not involved in plant immunity and has no synergistic effect with TTM2 [37]. Whether and how TTM2 functions in plant growth and development remain unclear.
Here, the function of TTM2 in ABA-dependent seed germination is reported. Our study shows that ABA inhibits TTM2 expression during seed germination. The ttm2 mutants show hypersensitivity to ABA-mediated seed germination, while 35S::TTM2-FLAG plants exhibited less insensitivity to ABA-mediated seed germination. ABA core transcription factor ABI4 represses the transcription level of TTM2 by directly binding to its promoter, and ABA-insensitive germination of abi4-1 is restored by crossing abi4-1 with ttm2-1 ablating TTM2, demonstrating that TTM2 is a downstream target of ABI4. In addition, there is a non-redundant function between TTM2 and TTM1 in regulating seed germination. In conclusion, our results indicate that TTM2 is regulated by ABI4 and contributes to ABA-mediated seed germination and early seedling growth.

ABA Regulates TTM2 Expression during Seed Germination
While whether and how TTM2 functions in plant growth and development remain unclear, the reported public microarray data showed that TTM2 expression is elevated during seed imbibition [41]. Thus, we used the quantitative real-time PCR (qRT-PCR) to assay transcription levels of TTM2 in dry seeds, imbibed seeds, and germinating seeds on 1/2 Murashige and Skoog (MS) medium (1 day after imbibition, DAI, and 2 DAI). Our data demonstrated that TTM2 expression was markedly increased in imbibed seeds and germinating seeds compared to dry seeds ( Figure 1A). Due to the high homology of TTM2 with TTM1 and TTM3, we investigated their transcription levels during germination and early seedling development. The results showed that TTM1 and TTM3 were not significantly altered during seed germination, which further confirmed the functional diversification of this family ( Figure S1).
Considering that ABA, one of the main hormones regulating seed germination, can inhibit seed germination [3, 7,9], we also investigated whether ABA is associated with TTM2 expression in seed germination. We observed that promoted TTM2 expression in the seeds on 1/2 MS medium was repressed on 1/2 MS medium containing 0.5 µM ABA ( Figures 1B  and S2). In addition, we obtained ProTTM2::GUS transgenic plants, in which GUS is under the control of the TTM2 promoter ( Figure S3). Consistently, GUS staining assays showed that GUS expression in germinating seeds on 1/2 MS medium containing 0.5 µM ABA was much lower than that in untreated germinating seeds ( Figure 1C). The findings imply that TTM2 may be associated with ABA-inhibited seed germination.
Since seed germination is severely regulated with mutual antagonism between ABA and GA [42][43][44]. Therefore, we also tested whether TTM2 expression was regulated by GA during seed germination, and found that, unlike ABA, the transcriptional level of TTM2 was not associated with exogenous GA treatment ( Figure S4). Figure 1. TTM2 expression changes during seed germination in response to ABA. (A) Expression of TTM2 in wild-type dry seeds, imbibed seeds, and germinating seeds (1 DAI and 2 DAI). Analysis of TTM2 transcript levels by qRT-PCR. The wild-type (WT) seeds were soaked at 4 °C for 3 d in water and grown on 1/2 Murashige & Skoog (MS) medium at various times, then harvested the plants at the specified dates. DAI, days after imbibition. The ACT2 gene was applied as an internal control. Expression levels were normalized to dry seeds, which was set at 1. Data are shown as mean ± SD (n = 3). Various letters represent significant differences at p < 0.05 by one-way ANOVA and Tukey's multiple comparison test. (B) The qRT-PCR to determine the expression of TTM2. Imbibed seeds of the wild type were planted in 1/2 MS medium (-ABA) or 1/2 MS medium containing 0.5 µM ABA (+ABA) for 2 days followed by RNA extraction. The ACT2 gene was applied as an internal control. Expression levels were normalized to those of untreated (-ABA) imbibed seeds, which was set at 1. Data are shown as mean ± SD (n = 3). Various letters represent significant differences at p < 0.05 by one-way ANOVA with Tukey's multiple comparison test. (C) GUS staining assay for ProTTM2::GUS #1 lines. Imbibed seeds were planted on 1/2 MS medium (-ABA) or 1/2 MS medium containing 0.5 µM ABA (+ABA) for 2 days, and then germinating seeds were subjected to GUS staining. Bars = 100 µm.

Repression of TTM2 Expression Is Required for ABA-Mediated Inhibition of Seed Germination
Then, we investigated whether changes in TTM2 expression can affect ABA-mediated seed germination. Therefore, we generated 35S::TTM2-FLAG transgenic lines ( Figure  S3) and assessed their seed germination when treated with ABA. Under normal conditions, the 35S::TTM2-FLAG transgenic plants had similar germination rates to the wild type. However, when germinated on 1/2 MS medium containing 0.5 or 1 µM ABA, compared with the wild type, the germination rate of 35S::TTM2-FLAG was higher ( Figure  2A-C). Consistent with the germination rate, in the absence of ABA, cotyledon greening of 35S::TTM2-FLAG plants had no difference from the wild type. In the presence of ABA, 35S::TTM2-FLAG plants were hyposensitive to ABA by displaying increased cotyledongreening percentages compared with the wild type ( Figure 2D, E). Combined with the data above that ABA suppresses the expression of TTM2, these results reveal that the repression of TTM2 expression is needed for ABA-mediated inhibition of seed germination and early seedling development.
This conclusion was further supported by assaying seed germination rate and cotyledon greening of ttm2-1 (SALK_145897) and ttm2-2 (SALK_114669), two previously reported T-DNA insertion knockout mutants [38]. When germinated and grown on 1/2 MS medium without ABA, the germination rates and cotyledon greening of these two mutants had no obvious difference from the wild type. However, when 1/2 MS medium containing 0.5 or 1 µM ABA, both ttm2-1 and ttm2-2 exhibited lower seed germination rate and reduced cotyledon greening in comparison to the wild type. And the ttm2 mutants, Figure 1. TTM2 expression changes during seed germination in response to ABA. (A) Expression of TTM2 in wild-type dry seeds, imbibed seeds, and germinating seeds (1 DAI and 2 DAI). Analysis of TTM2 transcript levels by qRT-PCR. The wild-type (WT) seeds were soaked at 4 • C for 3 d in water and grown on 1/2 Murashige & Skoog (MS) medium at various times, then harvested the plants at the specified dates. DAI, days after imbibition. The ACT2 gene was applied as an internal control. Expression levels were normalized to dry seeds, which was set at 1. Data are shown as mean ± SD (n = 3). Various letters represent significant differences at p < 0.05 by one-way ANOVA and Tukey's multiple comparison test. (B) The qRT-PCR to determine the expression of TTM2. Imbibed seeds of the wild type were planted in 1/2 MS medium (-ABA) or 1/2 MS medium containing 0.5 µM ABA (+ABA) for 2 days followed by RNA extraction. The ACT2 gene was applied as an internal control. Expression levels were normalized to those of untreated (-ABA) imbibed seeds, which was set at 1. Data are shown as mean ± SD (n = 3). Various letters represent significant differences at p < 0.05 by one-way ANOVA with Tukey's multiple comparison test. (C) GUS staining assay for ProTTM2::GUS #1 lines. Imbibed seeds were planted on 1/2 MS medium (-ABA) or 1/2 MS medium containing 0.5 µM ABA (+ABA) for 2 days, and then germinating seeds were subjected to GUS staining. Bars = 100 µm.
Since seed germination is severely regulated with mutual antagonism between ABA and GA [42][43][44]. Therefore, we also tested whether TTM2 expression was regulated by GA during seed germination, and found that, unlike ABA, the transcriptional level of TTM2 was not associated with exogenous GA treatment ( Figure S4).

Repression of TTM2 Expression Is Required for ABA-Mediated Inhibition of Seed Germination
Then, we investigated whether changes in TTM2 expression can affect ABA-mediated seed germination. Therefore, we generated 35S::TTM2-FLAG transgenic lines ( Figure S3) and assessed their seed germination when treated with ABA. Under normal conditions, the 35S::TTM2-FLAG transgenic plants had similar germination rates to the wild type. However, when germinated on 1/2 MS medium containing 0.5 or 1 µM ABA, compared with the wild type, the germination rate of 35S::TTM2-FLAG was higher (Figure 2A-C). Consistent with the germination rate, in the absence of ABA, cotyledon greening of 35S::TTM2-FLAG plants had no difference from the wild type. In the presence of ABA, 35S::TTM2-FLAG plants were hyposensitive to ABA by displaying increased cotyledon-greening percentages compared with the wild type ( Figure 2D,E). Combined with the data above that ABA suppresses the expression of TTM2, these results reveal that the repression of TTM2 expression is needed for ABA-mediated inhibition of seed germination and early seedling development.
This conclusion was further supported by assaying seed germination rate and cotyledon greening of ttm2-1 (SALK_145897) and ttm2-2 (SALK_114669), two previously reported T-DNA insertion knockout mutants [38]. When germinated and grown on 1/2 MS medium without ABA, the germination rates and cotyledon greening of these two mutants had no obvious difference from the wild type. However, when 1/2 MS medium containing 0.5 or 1 µM ABA, both ttm2-1 and ttm2-2 exhibited lower seed germination rate and reduced cotyledon greening in comparison to the wild type. And the ttm2 mutants, which were hypersensitive to ABA, had almost no green cotyledons at 8 DAI when treated with 1 µM ABA ( Figure 2). In our further experiments, 0.5 µM ABA was used for treatment.
which were hypersensitive to ABA, had almost no green cotyledons at 8 DAI when treated with 1 µM ABA ( Figure 2). In our further experiments, 0.5 µM ABA was used for treatment. Cotyledon greening rate. The green cotyledon rate was measured as (D). Each replicate had at least 100 seeds counted in it. Data are shown as mean ± SD (n = 3). Asterisk denotes a statistically significant difference compared with their wild type at indicated times. *, p < 0.05; **, p < 0.01; ***, p < 0.001 (Student's t-test).

TTM2 Acts Downstream of ABI4 to Regulate Seed Germination
ABI genes are the critical components of the ABA signaling response. Among them, ABI3, ABI4, and ABI5 have been identified as core transcription factors in seed germination [4,18,45]. To investigate how ABA regulates TTM2, we examined TTM2 transcriptional levels in abi mutants. We found that TTM2 expression can be still repressed by ABA in either abi3-4 or abi5-1 compared to the wild type, while the inhibition of TTM2 transcription by ABA in the wild type was abolished in abi4-1 (Figures 3 and S5), indicating ABI4 is essential for ABA-mediated inhibition of TTM2 expression. (E) Cotyledon greening rate. The green cotyledon rate was measured as (D). Each replicate had at least 100 seeds counted in it. Data are shown as mean ± SD (n = 3). Asterisk denotes a statistically significant difference compared with their wild type at indicated times. *, p < 0.05; **, p < 0.01; ***, p < 0.001 (Student's t-test).

TTM2 Acts Downstream of ABI4 to Regulate Seed Germination
ABI genes are the critical components of the ABA signaling response. Among them, ABI3, ABI4, and ABI5 have been identified as core transcription factors in seed germination [4,18,45]. To investigate how ABA regulates TTM2, we examined TTM2 transcriptional levels in abi mutants. We found that TTM2 expression can be still repressed by ABA in either abi3-4 or abi5-1 compared to the wild type, while the inhibition of TTM2 transcription by ABA in the wild type was abolished in abi4-1 (Figure 3 and Figure S5), indicating ABI4 is essential for ABA-mediated inhibition of TTM2 expression.
To explore how ABI4 modulates TTM2 expression, we first searched for the CCAC box, an important element in ABI4 target gene promoters [18], in the TTM2 promoter and found five CCAC boxes ( Figure 4A). Then, we generated 35S::ABI4-GFP transgenic plants ( Figure S6) and used them for chromatin immunoprecipitation (ChIP). Our findings demonstrated that the P3 region of the TTM2 promoter was substantially enriched by anti-GFP in chromatin immunoprecipitated with 35S::ABI4-GFP, indicating ABI4 binding of the P3 region ( Figure 4B). Our DNA electrophoretic mobility shift assays (EMSAs) utilizing ABI4 protein produced and purified from E. coli further confirmed this binding. Our results found that the ABI4 protein bound to the P3 probe that was biotin-labeled, while this binding could be completed by an unlabeled version of the DNA probe ( Figure 4C). To further substantiate the impact of ABI4 on the modulation of TTM2 expression, we also detected the expression of TTM2 with dual luciferase assays in Nicotiana benthamiana leaves. In these experiments, the reporter vector consists of a firefly luciferase (LUC) gene driven by the TTM2 promoter, the ABI4 gene driven by the CaMV 35S was generated for the effector vector ( Figure 4D) and LUC/REN ratio in the dual-LUC assay indicated relative LUC activity. The analysis displayed that the co-expression of 35S::ABI4-GFP and ProTTM2::LUC markedly reduced LUC activity in comparison with the 35S:: GFP and ProTTM2::LUC co-expression ( Figure 4E). Taken together, the tests indicated that ABI4 is a transcriptional repressor of TTM2. To explore how ABI4 modulates TTM2 expression, we first searched for the CCAC box, an important element in ABI4 target gene promoters [18], in the TTM2 promoter and found five CCAC boxes ( Figure 4A). Then, we generated 35S::ABI4-GFP transgenic plants ( Figure S6) and used them for chromatin immunoprecipitation (ChIP). Our findings demonstrated that the P3 region of the TTM2 promoter was substantially enriched by anti-GFP in chromatin immunoprecipitated with 35S::ABI4-GFP, indicating ABI4 binding of the P3 region ( Figure 4B). Our DNA electrophoretic mobility shift assays (EMSAs) utilizing ABI4 protein produced and purified from E. coli further confirmed this binding. Our results found that the ABI4 protein bound to the P3 probe that was biotin-labeled, while this binding could be completed by an unlabeled version of the DNA probe ( Figure 4C). To further substantiate the impact of ABI4 on the modulation of TTM2 expression, we also detected the expression of TTM2 with dual luciferase assays in Nicotiana benthamiana leaves. In these experiments, the reporter vector consists of a firefly luciferase (LUC) gene driven by the TTM2 promoter, the ABI4 gene driven by the CaMV 35S was generated for the effector vector ( Figure 4D) and LUC/REN ratio in the dual-LUC assay indicated relative LUC activity. The analysis displayed that the co-expression of 35S::ABI4-GFP and ProTTM2::LUC markedly reduced LUC activity in comparison with the 35S:: GFP and ProTTM2::LUC co-expression ( Figure 4E). Taken together, the tests indicated that ABI4 is a transcriptional repressor of TTM2.   To characterize the genetic relationship between ABI4 and TTM2, we obtained the abi4-1 ttm2-1 double mutant by crossing abi4-1 with ttm2-1. When germinated on 1/2 MS medium, all genotypes including the wild type, abi4-1, ttm2-1, and abi4-1 ttm2-1 showed similar germination rates ( Figure 5A). When subjected to ABA treatment, the seed germination rate of the abi4-1 mutant was higher than that of the wild type as previously reported [46], whereas the abi4-1 ttm2-1 mutant, similar to the ttm2-1 mutant, exhibited a lower germination rate compared with the wild type ( Figure 5B). Similarly, both ttm2-1 and abi4-1 ttm2-1 mutants showed reduced cotyledon greening compared with the wild type when grown on 1/2 MS medium containing 0.5 µM ABA, whereas the cotyledon greening was higher in the abi4-1 mutant than the wild type ( Figure 5C,D). Combined with our data above that ABI4 modulates TTM2 expression by directly binding to its promoter, these findings indicate that TTM2 is located genetically downstream of ABI4 in ABA-mediated inhibition of seed germination and early seedling development.

TTM1 Is Not Involved in ABA-Mediated Seed Germination and Early Seedling Development
TTM1 and TTM2 belong to the TTM family with a high degree of amino acid sequence similarity, however, these two members function differently in plants [38]. Here, we employed the ttm1-1 mutant (SALK_079237), a previously reported T-DNA insertion knockout line, to explore whether TTM1 acts in ABA-mediated seed germination and early seedling development [37]. Our outcomes indicated that the ttm1-1 mutant showed similar germination rates and cotyledon-greening percentages to the wild type when germinated and grown on 1/2 MS medium in the presence or absence of ABA ( Figure 6).  -1. (A, B) Seed germination curves. Germination rates are for imbibed Arabidopsis WT (Col-0), abi4-1 mutants, ttm2-1 mutants, and abi4-1 ttm2-1 double mutant planted on 1/2 MS medium (A) or 1/2 MS medium with 0.5 µM ABA (B) for germination experiments and then calculated at the indicated times. (C) Seed germination phenotypes. Imbibed seeds were incubated on 1/2 MS medium containing 0 or 0.5 µM ABA for 8 days. (D) Cotyledon greening rate. The green cotyledon rate was measured as (C). Each replicate had at least 100 seeds counted in it. Data are shown as mean ± SD (n = 3). Asterisk denotes a statistically significant difference compared with their wild type at indicated times. *, p < 0.05; **, p < 0.01; ***, p < 0.001 (Student's t-test).

TTM1 Is Not Involved in ABA-Mediated Seed Germination and Early Seedling Development
TTM1 and TTM2 belong to the TTM family with a high degree of amino acid sequence similarity, however, these two members function differently in plants [38]. Here, we employed the ttm1-1 mutant (SALK_079237), a previously reported T-DNA insertion knockout line, to explore whether TTM1 acts in ABA-mediated seed germination and early seedling development [37]. Our outcomes indicated that the ttm1-1 mutant showed similar germination rates and cotyledon-greening percentages to the wild type when germinated and grown on 1/2 MS medium in the presence or absence of ABA ( Figure 6). Then, we tried to test whether TTM1 has a possible synergistic effect with TTM2. Thus, we obtained the ttm1-1 ttm2-1 mutant ( Figure S7) and examined germination genotypes. All tested genotypes including the wild type, ttm1-1, ttm2-1, and ttm1-1 ttm2 showed similar germination and cotyledon-greening rates when germinated and grown on 1/2 MS medium without ABA ( Figure 6A). When ABA was applied, the ttm2-1 and ttm1-1 ttm2-1 mutants exhibited similar but lower seed germination rates and reduced cotyledon greening in comparison to the wild type ( Figure 6B-D), suggesting that TTM2 but not TTM1 has a role in ABA-regulated seed germination and early seedling development. Asterisk denotes a statistically significant difference compared with their wild type at indicated times. **, p < 0.01; ***, p < 0.001 (Student's t-test).

Discussion
Seed germination is one of the most important stages of plant growth and development, which is affected by multiple exogenous and endogenous factors [8,47,48]. Here, we focused on TTM2, a previously reported gene involved in plant biotic stress, whose function in plant growth and development is unclear [38]. Our data found that ABI4 represses TTM2 expression during ABA-mediated seed germination. We first explored the expression of TTM2 at various stages of seed germination based on public microarray data. The results displayed that the TTM2 expression increased rapidly after the imbibition and germination stage ( Figure 1A). We speculated that TTM2 may be involved in the seed germination process. Since many events of germination are regulated by the incorporation of ABA and GA, two hormones that have antagonistic effects on the germination process. We examined the transcript level of TTM2 under ABA and GA treatment, respectively. The experimental results revealed that ABA significantly represses the transcription of TTM2, but GA does not (Figures 1B and S4). Further analysis showed that germination of the ttm2 knockout mutant was hypersensitive to ABA, whereas TTM2-overexpressing plants were less insensitive to ABA (Figure 2). The results suggested that TTM2 is involved in ABA-mediated seed germination. Figure 6. The ttm2-1 and ttm1-1 ttm2-1 mutants exhibited similar responses to ABA during seed germination. (A,B) Seed germination curves. Germination rates are for imbibed Arabidopsis WT (Col-0), knockout mutants (ttm1-1, ttm2-1), and double mutant ttm1-1 ttm2-1 planted to 1/2 MS (A) or 1/2 MS medium with 0.5 µM ABA (B) for germination experiments and then calculated at the indicated times. (C) Seed germination phenotypes. Imbibed seeds were incubated on 1/2 MS medium containing 0 or 0.5 µM ABA for 8 days. (D) Cotyledon greening rate. The green cotyledon rate was measured as (C). Each replicate had at least 100 seeds counted in it. Data are shown as mean ± SD (n = 3). Asterisk denotes a statistically significant difference compared with their wild type at indicated times. **, p < 0.01; ***, p < 0.001 (Student's t-test).

Discussion
Seed germination is one of the most important stages of plant growth and development, which is affected by multiple exogenous and endogenous factors [8,47,48]. Here, we focused on TTM2, a previously reported gene involved in plant biotic stress, whose function in plant growth and development is unclear [38]. Our data found that ABI4 represses TTM2 expression during ABA-mediated seed germination. We first explored the expression of TTM2 at various stages of seed germination based on public microarray data. The results displayed that the TTM2 expression increased rapidly after the imbibition and germination stage ( Figure 1A). We speculated that TTM2 may be involved in the seed germination process. Since many events of germination are regulated by the incorporation of ABA and GA, two hormones that have antagonistic effects on the germination process. We examined the transcript level of TTM2 under ABA and GA treatment, respectively. The experimental results revealed that ABA significantly represses the transcription of TTM2, but GA does not (Figures 1B and S4). Further analysis showed that germination of the ttm2 knockout mutant was hypersensitive to ABA, whereas TTM2-overexpressing plants were less insensitive to ABA (Figure 2). The results suggested that TTM2 is involved in ABA-mediated seed germination.
ABI4 acts as a key player in ABA-mediated germination [49,50]. We discovered that exogenous ABA significantly repressed TTM2 expression in the wild type. However, the repression was significantly attenuated in the abi4-1 mutant (Figure 3), suggesting that TTM2 may act downstream of ABI4. ABI4, an active ABA signaling response transcription factor, has both transcriptional repression and activation effects [34]. It can bind to gene promoters CE1-liKE motif (CACCG) to regulate the transcription of some genes such as ABI5, NCED6, Starch Branching Enzyme, Acyl-coenzyme A: Diacylglycerol Acyltransferase1, and CHOTTO1 responding to ABA signaling [33,35,[51][52][53]. Further studies revealed that binding capacity still exists in degenerate bases of the CE1-liKE motif (CCAC) in the promoters of some genes [54], such as CYCP707A1, CYCP707A2, and Vitamin C Defective 2 [18,23]. We identified the CCAC box in the TTM2 promoter ( Figure 4A) and our results revealed that ABI4 binds to the TTM2 promoter and inhibits the expression of TTM2 ( Figure 4B-E). Consistently, a recent study showed that ABI4 can bind thousands of gene promoters, including TTM2 by using the ChIP-seq assay [55]. Furthermore, in germination experiments, the ttm2-1 mutant rescued the ABA-insensitive phenotype of the abi4-1 mutant ( Figure 5), further supporting that ABI4 directly inhibited TTM2 expression in response to ABA-dependent seed germination. In addition, we found that the transcription level of TTM2 can be still repressed by ABA in either abi3-4 or abi5-1 (Figures 3 and S5). Additionally, the TTM2 promoter does not contain the core binding motifs of ABI3 and ABI5, which further validates that TTM2 may not be regulated by ABI3/5 in ABA signaling. However, ABA-dependent transcription factors also contain many members, such as the MYB superfamily and WRKY superfamily [56][57][58]. And we found a large number of WRKY core motif W-boxes (TGAC) and MYB core sequences (TTGTTA) in the TTM2 promoter. It means that TTM2 may act as an essential target regulated by multiple aspects of ABA signaling, which deserves further consideration in the future.
The TTM family is an evolutionarily well-conserved protein family in eukaryotes. Until now, only a few reports have described the function of this family in plants [37][38][39][40]59]. Here, we found that ttm1-1 mutants don ' t show sensitivity to ABA during seed germination, and the sensitivity of the ttm1-1 ttm2-1 double mutant to ABA was similar to that of the ttm2-1 mutant in seed germination ( Figure 6), which means there was non-redundant function in TTM1 and TTM2. Interestingly, researchers have recently found that TTM1 is implicated in ABA-mediated leaf senescence [40]. Multiple mitogen-activated protein kinases (MPK1/3/4/6/7) phosphorylate the three main TTM1 sites (Ser10, Ser437, and Ser490), and the phosphorylation of Ser437 is essential for the role of TTM1 in leaf senescence [40]. TTM2 has extremely high sequence identity (65.7%) and similarity (92.3%) to TTM1 [37]. Amino acid sequence alignment showed that Ser437 and Ser490 are fully conserved in TTM1 and TTM2. Whether the regulation of MPK phosphorylation is also present in TTM2 protein in ABA-mediated seed germination also deserves further investigation.
In conclusion, our results indicate that TTM2 acts as a downstream factor of ABI4 in ABA-mediated seed germination and early seedling growth.

Production of Transgenic Plants
For generating the ProTTM2::GUS vectors, the TTM2 promoter sequence covering the 2 kb upstream of the ATG start codon was amplified and cloned into the pCAMBIA1300-GUS vector. To construct 35S::TTM2-FLAG vectors, the amplified CDS of TTM2 was cloned into the pCAMBIA1300-FLAG vector controlled by the 35S promoter. To construct 35S::ABI4-GFP vectors, the amplified CDS of ABI4 was cloned into the pCAMBIA1300-GFP vector controlled by the 35S promoter. Then the vectors of ProTTM2::GUS, 35S::TTM2-FLAG, and 35S::ABI4-GFP were transformed into Arabidopsis Col-0,respectively, using Agrobacterium-mediated flower immersion [62]. The homozygous T3 plants were selected on a medium containing kanamycin and then verified by PCR and qRT-PCR. Primer sequences are listed in Table S1.

Determination of Germination Rate and Green Cotyledon
Germination rate and green cotyledons were determined following the methods of previous studies [63]. Briefly, the seeds were sterilized and imbibed at 4 • C for 3 days and then they were planted on 1/2 MS medium with or without ABA/GA 3 (Solarbio Science Technology Co., Ltd., Beijing, China). Seed germination was followed for 8 days and the germination rate was measured by the appearance of radicles through the seed coat; green cotyledons were counted by the emergence of green cotyledons in seedlings at 8 d. Three biological duplicates of each experiment were run, and equivalent outcomes were achieved. For each treatment, more than 100 seeds were counted.

RNA Extraction and Quantitative Real-Time PCR
Total RNA was isolated from seeds or seedlings using TRIzol Reagent (TransGen Biotech, Beijing, China). The cDNA was synthesized by reverse transcription after the removal of DNA contamination (TransGen Biotech, Beijing, China). The qRT-PCR was performed in 96-well plates by adding SYBR Green I dye (Monad, Suzhou, China) and then placed on a Bio-Rad CFX96 device (Bio-Rad, Hercules, CA, USA) and programmed as follows: 95 • C/3 min for pre-denaturation, followed by 95 • C/15 s and 60 • C/30 s for 45 cycles. The ACT2 gene was used as a reference gene. The experiments were duplicated at least three times biologically. The primer sequences are listed in Table S1.

Electrophoresis Mobility Shift Assay (EMSA)
The CDS of ABI4 was integrated into the pET28a and transformed into E. coli BL21 strain in order to produce the ABI4 protein. The biotin-labeled oligonucleotide probe and purified ABI4 protein were utilized for EMSA determination. According to the manufacturer's protocol, experiments were conducted using the Light Shift chemiluminescent EMSA kit (Thermo Scientific, Waltham, MA, USA). The probes used in the experiments are listed in Table S1.

Chromatin Immunoprecipitation (ChIP)-qPCR Assay
Details of ChIP assay are reported in previous studies [64]. The 35S::ABI4-GFP transgenic plants germinating for 3 days were fixed with formaldehyde, then the tissues were crushed, and the chromatin was separated and ultra-sonicated. Anti-GFP antibody was used for immunoprecipitation. ChIP-qPCR was performed on DNA fragments enriched with Anti-GFP, while non-immunoprecipitated sonicated chromatin was utilized as input control. The experiment used UBQ5 as an internal control. The results of the experiment were similar in three independent biological replicates. The primers are listed in Table S1.

Dual Luciferase Assays
In a dual luciferase (dual-LUC) assay, the ABI4 CDS was inserted into the pCAMBIA1300-GFP vector under the control of a 35S promoter to generate an effector construct and transformed the recombinant construct into Agrobacterium tumefaciens strain GV3101. To generate the reporter construct, the TTM2 promoter region was cloned separately into the pGreenII 0800-LUC and the recombinant vectors were transformed into Agrobacterium rhizogenes strain C58C1. Pseudomonas hybrids were infected with Nicotiana benthamiana leaves. Following the manufacturer's instructions, the fluorescence values were read on a LUC fluorescence detector (Promega, Madison, WI, USA) by using a dual fluorogenic enzyme reporter gene assay kit (Beyotime Biotechnology, Shanghai, China). The primers are listed in Table S1.

GUS Staining
GUS assays were carried out using a staining kit (Coolaber, Beijing, China). The staining buffer was used to incubate ProTTM2::GUS transgenic plant tissues at 37 • C for 8-12 h. After termination of the reaction, the tissue was cleared with 70% ethanol and imaged.

Conclusions
In summary, our work uncovers a novel function of TTM2, which is involved in ABA-mediated seed germination and early seedling development and is transcriptionally repressed by ABI4. In contrast, TTM1, a homologous protein of TTM2, is found to have no apparent function in this process, further suggesting that they are involved in independent biological processes. Our study provides new insights into further understanding of seed germination regulatory mechanisms.