Transcriptomics Reveals the ERF2-bHLH2-CML5 Module Responses to H2S and ROS in Postharvest Calcium Deficiency Apples

Calcium deficiency usually causes accelerated quality deterioration in postharvest fruit, whereas the underlining mechanism is still unclear. Here, we report that calcium deficiency induced the development of bitter pit on the surface of apple peels compared with the healthy appearance in control apples during postharvest storage. Physiological analysis indicates that calcium-deficient peels contained higher levels of superoxide anion (O2•−), malondialdehyde (MDA), total phenol, flavonoid contents and polyphenol oxidase (PPO) activity, and reduced calcium, H2S production, anthocyanin, soluble protein content, and peroxidase (POD) activity compared with those in calcium-sufficient peels. The principal component analysis (PCA) results show that calcium content, ROS, and H2S production were the main factors between calcium-deficient and calcium-sufficient apple peels. Transcriptome data indicated that four calmodulin-like proteins (CMLs), seven AP2/ERFs, and three bHLHs transcripts were significantly differentially expressed in calcium-deficient apple peels. RT-qPCR and correlation analyses further revealed that CML5 expression was significantly positively correlated with the expression of ERF2/17, bHLH2, and H2S production related genes. In addition, transcriptional co-activation of CML5 by ERF2 and bHLH2 was demonstrated by apple transient expression assays and dual-luciferase reporter system experiments. Therefore, these findings provide a basis for studying the molecular mechanism of postharvest quality decline in calcium-deficient apples and the potential interaction between Ca2+ and endogenous H2S.


Introduction
Calcium (Ca) is one of the essential and abundant elements for the growth of plants. It not only determines the yield and quality of agricultural crops, but also plays a pivotal role in maintaining plant cell structure, resistance to adverse stress and signal transduction [1][2][3][4]. Calcium can mitigate stress conditions, for example, by neutralizing reactive oxygen species (ROS) produced in cells [5][6][7]. Calcium deficiency results in ROS accumulation and causes damage to membranes, reduces the cell wall, weakens tissue stiffness, and increases water loss, which in turn leads to leaf wilting and shortens the shelf life of harvested fruit [8][9][10].
scriptome data, and their expression was verified by RT-qPCR. Moreover, the expression of genes related to H 2 S production was also analyzed to reveal the effect of calcium deficiency on H 2 S production. Further, ERF2-bHLH2 complex was found to activate the expression of CML5 by apple transient expression assays and dual-luciferase reporter system experiments, thereby enhancing the understanding of the regulatory mechanism of calcium-deficient apple peels and the potential interaction between Ca 2+ and endogenous H 2 S, contributing to improving the appearance quality of apple fruit.

Calcium Deficiency Significantly Affects the Phenotype of Apple Peels and Reduces H 2 S Production
In this experiment, the phenotypes of calcium-deficient and calcium-sufficient apple fruits stored for different days after harvest were observed. As shown in Figure 1A, the apple fruit did not significantly differ in size, but the color and smoothness of apple peels were significantly different at 0, 7, 14, and 21 days after storage (DAS) between calciumdeficient and calcium-sufficient apples. The appearance of calcium-sufficient apple peels showed no obvious change with increasing days of storage. At 7 DAS, however, bitter pit appeared on calcium-deficient apple peels and increased continuously with increasing days of storage. At the same time, the peel color of calcium-deficient apples gradually deteriorated, and the surface became rough. Then, the calcium contents in calcium-deficient and calcium-sufficient apple peels at different storage times were measured. The calcium content of calcium-deficient apple peels was always lower than that of calcium-sufficient apple peels during storage, but significant difference between the two was observed at 7, 14, and 21 DAS (p < 0.01) ( Figure 1B). It is evident that calcium deficiency in apples affects the phenotype of apple peels and leads to physiological disorders. Hence, it is important to investigate the molecular mechanism of calcium deficiency-induced disorders in apples.
analysis. Then, differentially expressed calcium-regulated genes were screened from tran scriptome data, and their expression was verified by RT-qPCR. Moreover, the expressio of genes related to H2S production was also analyzed to reveal the effect of calcium def ciency on H2S production. Further, ERF2-bHLH2 complex was found to activate the ex pression of CML5 by apple transient expression assays and dual-luciferase reporter sys tem experiments, thereby enhancing the understanding of the regulatory mechanism o calcium-deficient apple peels and the potential interaction between Ca 2+ and endogenou H2S, contributing to improving the appearance quality of apple fruit.

Calcium Deficiency Significantly Affects the Phenotype of Apple Peels and Reduces H2S Pro duction
In this experiment, the phenotypes of calcium-deficient and calcium-sufficient appl fruits stored for different days after harvest were observed. As shown in Figure 1A, th apple fruit did not significantly differ in size, but the color and smoothness of apple peel were significantly different at 0, 7, 14, and 21 days after storage (DAS) between calcium deficient and calcium-sufficient apples. The appearance of calcium-sufficient appl peels showed no obvious change with increasing days of storage. At 7 DAS, howeve bitter pit appeared on calcium-deficient apple peels and increased continuously with in creasing days of storage. At the same time, the peel color of calcium-deficient apples grad ually deteriorated, and the surface became rough. Then, the calcium contents in calcium deficient and calcium-sufficient apple peels at different storage times were measured. Th calcium content of calcium-deficient apple peels was always lower than that of calcium sufficient apple peels during storage, but significant difference between the two was ob served at 7, 14, and 21 DAS (p < 0.01) ( Figure 1B). It is evident that calcium deficiency i apples affects the phenotype of apple peels and leads to physiological disorders. Hence it is important to investigate the molecular mechanism of calcium deficiency-induced dis orders in apples. Comparison of phenotype and endogenous H2S determination of apple peel at 0, 7, 14, and 21 DAS with calcium deficiency and calcium sufficiency. (A) Phenotypic change of apple peel at 0, 7, 14, and 21 DAS. Ca-suf, Ca sufficiency; Cadef, Ca deficiency. Data are presented as means ± SD (n = 5). (B) Calcium contents of calcium-sufficient and calciumdeficient apple peels during storage. Data are presented as means ± SD (n = 3). (C) H2S production capacity from calciumsufficient and calcium-deficient apple peels during storage as detected by the brown precipitate, lead sulfide. Data are Figure 1. Comparison of phenotype and endogenous H 2 S determination of apple peel at 0, 7, 14, and 21 DAS with calcium deficiency and calcium sufficiency. (A) Phenotypic change of apple peel at 0, 7, 14, and 21 DAS. Ca-suf, Ca sufficiency; Ca-def, Ca deficiency. Data are presented as means ± SD (n = 5). (B) Calcium contents of calcium-sufficient and calcium-deficient apple peels during storage. Data are presented as means ± SD (n = 3). (C) H 2 S production capacity from calcium-sufficient and calcium-deficient apple peels during storage as detected by the brown precipitate, lead sulfide. Data are presented as means ± SD (n = 6). (D,E) represent the change of color-parameter L* value and the ratio of a*/b* corresponding to (C). L* represents luminance; a* represents a range from green to magenta; b* represents a range from yellow to blue. Data are presented as means ± SD (n = 6). Asterisks indicate statistical difference of the values at p < 0.05 (*); p < 0.01 (**).
In order to measure H 2 S directly, lead sulfide method was applied to show the changes of H 2 S production in calcium-deficient and sufficient apple peels. As shown in Figure 1C, more H 2 S was produced by calcium-sufficient apple peels than by calcium-deficient apple peels at 0, 7, 14, 21 DAS, evidenced by the darkened precipitate due to lead sulfide formed on the strips. Luminance (L*) and color change (a*/b*) was measured by colorimeter, and they could show the effect of calcium deficiency on H 2 S production. Higher L* values and lower ratio of a*/b* indicated that less H 2 S was absorbed by the lead acetate filter paper. The L* values were higher for calcium-deficient apple peels than for calcium-sufficient apple peels ( Figure 1D), but the ratio of a*/b* was reversed ( Figure 1E). In summary, it is clear that calcium deficiency reduces H 2 S production in apple peels.

Calcium Deficiency Increases the Contents of Flavonoids and Total Phenols but Decreases Anthocyanin Contents in Apple Peels
To further explore the mechanism by which calcium deficiency affects the phenotype of apple peels, the contents of flavonoids, total phenols, and anthocyanins were determined. As shown in Figure 2A, the flavonoid contents of calcium-deficient peels were almost indistinguishable from those of calcium-sufficient peels at 7 DAS, whereas at 0, 14, and 21 DAS, the flavonoid contents of calcium-deficient peels were significantly higher than those of calcium-sufficient peels (p < 0.01). The total phenol contents of calcium-deficient peels were higher than those of calcium-sufficient peels, and the difference was significant at 0, 14, and 21 DAS (p < 0.01 or p < 0.05) ( Figure 2B). At the same time, there was no appreciable difference in anthocyanin contents between calcium-deficient and calcium-sufficient peels at 0 and 7 DAS, but the anthocyanin contents in calcium-deficient peels were significantly lower than those in calcium-sufficient peels at 14 and 21 DAS (p < 0.01) ( Figure 2C).

Calcium Deficiency Induces the Production of ROS in Apple Peels
Excessive production of ROS and oxidative damage are usually observed during fruit storage. To investigate the effect of calcium deficiency on ROS production in apple peels, the accumulation of O2 •− , H2O2, and MDA was determined. Compared to calcium-sufficient peels, the H2O2 contents of calcium-deficient peels were lower during postharvest

Calcium Deficiency Induces the Production of ROS in Apple Peels
Excessive production of ROS and oxidative damage are usually observed during fruit storage. To investigate the effect of calcium deficiency on ROS production in apple peels, the accumulation of O 2 •− , H 2 O 2 , and MDA was determined. Compared to calcium-sufficient peels, the H 2 O 2 contents of calcium-deficient peels were lower during postharvest storage days; in particular, the most remarkable variation was at 0 DAS (p < 0.01) ( Figure 2D). The O 2 •− production rate was significantly higher in calcium-deficient peels than in calciumsufficient peels at 0 and 21 DAS (p < 0.01), but the difference was not significant at 14 DAS ( Figure 2E). Similarly, the MDA contents in calcium-sufficient peels were lower than those in calcium-deficient peels, especially at 14 DAS (p < 0.01) ( Figure 2F).

Changes in POD Activity, PPO Activity Soluble Protein Content between Calcium Deficiency and Sufficiency Conditions
To study the potential mechanism of calcium deficiency on the accumulation of ROS, the peroxidase (POD) and polyphenol oxidase (PPO) activities were determined in calcium deficiency and sufficiency apples. As shown in Figure 2G, the POD activity in calciumsufficient peels was maintained at a higher level than that in calcium-deficient peels at 7 and 14 DAS (p < 0.05 or p < 0.01). The POD activity gradually increased with increasing days of storage and then decreased. The PPO activity in calcium-deficient peels was significantly higher than that in calcium-sufficient peels at 14 DAS (p < 0.01) ( Figure 2H). To compare the effects of calcium deficiency on apple peel quality, soluble protein content was measured. The soluble protein content sustained at a lower level in calcium-deficient peels than that in the calcium-sufficient peels at 0, 7, 14, and 21 DAS ( Figure 2I).

PCA Analysis of the Bioactive Substance Changes in Apple Peels
The principal component analysis (PCA), a statistical method of dimensionality reduction, was carried out to investigate the main factors of the bioactive substances in apple peels that were most affected by calcium deficiency, and O 2 •− production rate, ratio of a*/b*, L* value, POD and PPO activity, H 2 O 2 , calcium, soluble protein, flavonoid, total phenolics, anthocyanins, MDA contents were used in PCA ( Figure 3). The contribution rates of PC1 and PC2 were 84.7% and 15.3%, respectively. There was a significant difference between calcium-deficient apple peels and calcium-sufficient apple peels. In PC1, H 2 O 2 , calcium content, O 2 •− production rate, and ratio of a*/b* were the main factors, while POD activity, L* value, and O 2 •− production rate were the main factors in PC2 ( Table 1). The above results indicated that calcium deficiency caused physiological disorders and changes in postharvest storage quality in apple peels, which severely affected the ROS, H 2 S production, and redox processes in apple peels. rates of PC1 and PC2 were 84.7% and 15.3%, respectively. There was a significant difference between calcium-deficient apple peels and calcium-sufficient apple peels. In PC1, H2O2, calcium content, O2 •− production rate, and ratio of a*/b* were the main factors, while POD activity, L* value, and O2 •− production rate were the main factors in PC2 ( Table 1). The above results indicated that calcium deficiency caused physiological disorders and changes in postharvest storage quality in apple peels, which severely affected the ROS, H2S production, and redox processes in apple peels.

Identification and GO Classification and KEGG Enrichment Analysis of DEGs between Calcium-Deficient and Calcium-Sufficient Apple Peels
To investigate the molecular mechanisms of calcium regulation in apple peels, differentially expressed genes (DEGs) associated with calcium regulation between calciumsufficient and calcium-deficient apple peels were obtained from published transcriptome data. Pairwise comparison of the three sequenced samples resulted in three sets of contrasts with T02/T01, T03/T01, and T03/T02 (T01, diseased peel of calcium-deficient apples; T02, healthy peel of calcium-deficient apples; and T03, peel of calcium-sufficient apples). In the T02/T01, T03/T01, and T03/T02 comparisons, 1323, 2880, and 2182 DEGs were upregulated, respectively, while 1573, 1606, and 269 DEGs were downregulated, respectively (Figure S1A-C). In total, 1031 DEGs were common to all three comparisons ( Figure S1D).
GO classification analysis was used to investigate the gene expression profiles of calcium-deficient apple peels, and the T03/T01 comparison with the highest number of DEGs was selected for gene annotation analysis. The DEGs were classified into 51 functional groups based on their biological processes ( Figure S1E). The major subcategories were as follows: 20 subcategories for biological process, 15 subcategories for cellular component, and 16 subcategories for molecular function. The DEGs in 'signaling' and 'antioxidant activity' played important roles during the calcium regulated processes. These results provided a comprehensive perspective for screening candidate genes involved in calcium regulation.
KEGG analysis provided information and further understanding of secondary metabolites induced by calcium signaling. As shown in Figure S1F, 20 pathways were enriched with over 20 DEGs. Plant hormone signal transduction was the most abundant DEG found in metabolism category followed by plant-pathogen interaction, biosynthesis of amino acids, and phenylpropanoid biosynthesis. These results suggested that the DEGs in the T03/T01 comparison were mainly enriched in metabolic processes related to plant signal transduction and plant pathogen infestation.

Correlation Analysis between Genes Expression and Physiological Parameters in Apple Peels
The result of the correlation analysis between genes expression and physiological parameters in apple peels is shown in Figure 7. Calcium content was negatively correlated with MDA, O2 •− , H2O2 contents, and ratio of a*/b*, positively correlated with POD activity, H2S production on filter papers, DCD1, DCD2, DCD3, DCD4, LCD1, LCD2, LCD3, SiR1, and OASTL1 expression levels, indicating that calcium deficiency activated ROS production in apples but reduced H2S production. The expression levels of ERF2, ERF17, bHLH2, and CML5 were positively correlated with calcium content, PPO activity, POD activity, H2S production on filter papers, DCD1, DCD2, DCD3, DCD4, LCD1, LCD2, LCD3, SiR1, and OASTL1 expression levels, but negatively correlated with MDA, O2 •− , H2O2 contents, and ratio of a*/b*. Moreover, H2S production showed positive correlation with MDA, O2 •− , H2O2 contents in apple peels. Based on Figure 7 and the above results, it could be concluded that there was a strong positive correlation between the expression of presumably ERF2, ERF17, bHLH2, and CML5, and they were highly expressed and differ significantly between calcium-deficient and calcium-sufficient apple peels at storage time. It was therefore hypothesized that ERF2, ERF17, and bHLH2 may be involved in the regulation of CML5 and H2S production during postharvest apple fruit storage.

Correlation Analysis between Genes Expression and Physiological Parameters in Apple Peels
The result of the correlation analysis between genes expression and physiological parameters in apple peels is shown in Figure 7. Calcium content was negatively correlated with MDA, O 2 •− , H 2 O 2 contents, and ratio of a*/b*, positively correlated with POD activity, H 2 S production on filter papers, DCD1, DCD2, DCD3, DCD4, LCD1, LCD2, LCD3, SiR1, and OASTL1 expression levels, indicating that calcium deficiency activated ROS production in apples but reduced H 2 S production. The expression levels of ERF2, ERF17, bHLH2, and CML5 were positively correlated with calcium content, PPO activity, POD activity, H 2 S production on filter papers, DCD1, DCD2, DCD3, DCD4, LCD1, LCD2, LCD3, SiR1, and OASTL1 expression levels, but negatively correlated with MDA, O 2 •− , H 2 O 2 contents, and ratio of a*/b*. Moreover, H 2 S production showed positive correlation with MDA, O 2 •− , H 2 O 2 contents in apple peels. Based on Figure 7 and the above results, it could be concluded that there was a strong positive correlation between the expression of presumably ERF2, ERF17, bHLH2, and CML5, and they were highly expressed and differ significantly between calcium-deficient and calcium-sufficient apple peels at storage time. It was therefore hypothesized that ERF2, ERF17, and bHLH2 may be involved in the regulation of CML5 and H 2 S production during postharvest apple fruit storage.

ERF2-bHLH2 Coregulation Promotes CML5 Expression in Apple Peels
To further investigate whether and how ERF2/17 and bHLH2 regulate CML5 in apple peels, a transient transformation experiment was performed using 'Honeycrisp' apples at 60 days after flower blooming, and ERF2/ERF17 and bHLH2 were individually or cotransformed in apples via Agroinfiltration. Transient expression assays in apples showed that individual transformation of ERF2 and bHLH2 resulted in a minor deposition of anthocyanin and a slight loss of water in the peel at the injection site. In contrast, when ERF17 was individually transformed, anthocyanin deposition was evident, and there was no water loss-induced wrinkling of the peel at the injection site. When ERF2 and bHLH2 were cotransformed, there was also minor anthocyanin deposition and significant water loss-induced wrinkling of the peel at the injection site, whereas when ERF17 and bHLH2 were Figure 7. Correlation analysis between gene expression and identified indices in apple peels. R scripts were used to analyze Pearson's correlation coefficients. The color of the scale label from blue to red represents the change in Pearson's correlation coefficients from '1' to '−1' with the following indicators: '+' represents a positive correlation, '−' represents a negative correlation, 0.8-1 represents a highly strong correlation, 0.6-0.8 represents a strong correlation, 0.4-0.6 represents a moderate correlation, 0.2-0.4 represents a weak correlation, and 0-0.2 represents a very weak correlation or no correlation.

ERF2-bHLH2 Coregulation Promotes CML5 Expression in Apple Peels
To further investigate whether and how ERF2/17 and bHLH2 regulate CML5 in apple peels, a transient transformation experiment was performed using 'Honeycrisp' apples at 60 days after flower blooming, and ERF2/ERF17 and bHLH2 were individually or cotransformed in apples via Agroinfiltration. Transient expression assays in apples showed that individual transformation of ERF2 and bHLH2 resulted in a minor deposition of anthocyanin and a slight loss of water in the peel at the injection site. In contrast, when ERF17 was individually transformed, anthocyanin deposition was evident, and there was no water loss-induced wrinkling of the peel at the injection site. When ERF2 and bHLH2 were cotransformed, there was also minor anthocyanin deposition and significant water loss-induced wrinkling of the peel at the injection site, whereas when ERF17 and bHLH2 were cotransformed, there was significant anthocyanin deposition but no water loss-induced wrinkling ( Figure 8A). The trends of the L* and a*/b* values in the injection regions shown in Figure 8B,C were consistent with the phenotypic results. cotransformed, there was significant anthocyanin deposition but no water loss-induced wrinkling ( Figure 8A). The trends of the L* and a*/b* values in the injection regions shown in Figure 8B,C were consistent with the phenotypic results.  RT-qPCR was then used to analyze the expression of ERF2, ERF17, bHLH2, and calcium regulatory genes (CML1, CML2, CML3, CML4, and CML5). The expression of ERF2, ERF17 and bHLH2 was significantly increased when the apple peels were injected individually or cotransformed with bHLH2 (p < 0.01, Figure 8D-F). Interestingly, bHLH2 expression progressively increased with individual transformation of ERF2, ERF17, and bHLH2 or when cotransformed with bHLH2. The expression of CML1, CML2, CML3, and CML4 increased significantly (p < 0.01) when ERF2 was individually transformed. There was a slight increase in CML1, CML2, CML3, and CML4 expression when bHLH2 was individually transformed. The expression of CML1, CML2, CML3, and CML4 decreased when ERF2, ERF17, and bHLH2 were cotransformed ( Figure 8G-J). The expression of CML5 increased significantly (p < 0.01) when bHLH2 was individually transformed and with ERF2 cotransformation, and the expression of CML5 was slightly increased when ERF2 was individually transformed ( Figure 8K). In addition, the expression level of CML5 were significantly higher than those of CML1, CML2, CML3, and CML4 when ERF2 and bHLH2 were cotransformed Thus, these results demonstrated that individual transformation of bHLH2 or cotransformation with ERF2 significantly activates CML5 transcription, suggesting that ERF2 and bHLH2 cooperate to activate CML5 to regulate calcium signaling and postharvest apple storage quality.
Promoter cis-acting elements were first predicted for CML5, ERF2/17, and bHLH2 and the results are shown in Figure 8L. The 2 kb region upstream of the CML5 promoter contains one or more cis-acting elements of the DRE core/DRE1 of AP2/ERFs and the G-Box of bHLHs. Then, a dual-luciferase reporter system was used to verify the mode of regulation between ERF2/17 and bHLH2 and CML5. The results showed that ERF2 and bHLH2 co-injection significantly activated CML5 transactivation ( Figure 8M). From those it can be concluded that ERF2-bHLH2 coregulation promotes CML5 expression in apple peels, thereby regulating calcium signaling and postharvest storage quality.

Calcium Deficiency Increases ROS Generation, Reduces H 2 S Production and Postharvest Storage Quality in Apple Peels
Calcium is one of the essential elements for plant growth, and it has important functions in maintaining plant cell structure, resistance to stress and signal transduction [1][2][3][4]. In the present study, the peel of calcium-deficient apples became rough during postharvest storage due to water loss, and they developed distinct bitter pit compared to calciumsufficient apples ( Figure 1A,B). These phenotypic changes were consistent with those reported in previous studies [6,9,10]. In addition, compared to calcium-sufficient peels, calcium-deficient peels produced less H 2 S ( Figure 1C-E), significantly increased O 2 •− , MDA, total phenol and flavonoid contents, and PPO activity, and decreased POD activity ( Figure 2). The increase in ROS and antioxidants such as total phenol and flavonoid in calcium-deficient peels may be caused by bitter pit disorder due to calcium deficiency, which was also consistent with previous reports [5][6][7]. The result of PCA shows that calcium content, ROS, and H 2 S production were main factors between calcium-deficient and calcium-sufficient apple peels. The above results indicated that calcium deficiency in apples reduced H 2 S production and increased ROS production, thereby reducing postharvest fruit storage quality. These results are in agreement with those reported previously [38]. However, the specific regulatory network that exists between calcium signaling and the production of H 2 S and ROS is currently unknown and rarely reported, and further research is needed.

The ERF2-bHLH2 Coactivate the CML5 Expression and Coregulate Downstream Responses in Calcium-Deficient Apple Peels
Calmodulin-like (CML) protein in plant tissues is an important calcium receptor protein in the process of signal transmission. In the present study, the expression of CML5 was negatively correlated with ROS contents, but positively correlated with POD activity, flavonoids, total phenols, and calcium content (Figure 7), suggesting the potential impor-tance of CML5 in regulating ROS metabolism. There are numerous studies reporting that CMLs are involved in regulating a variety of abiotic stress processes. For example, SlCML37 enhances cold tolerance in tomato fruit [39]; CmCML13 improves drought resistance in Arabidopsis [40]; CML21 is involved in abiotic stress response in grapevine [41]. Thus, it can be hypothesized that CML5 may be involved in regulating postharvest storage quality and antioxidant-related processes in calcium-deficient apple peels.
Furthermore, the expression levels of ERF2, ERF17, and bHLH2 were screened based on their significant positive correlation with CML5 and they were significantly positively correlated with the H 2 S production related genes (DCD1, DCD2, DCD3, DCD4, LCD1, LCD2, LCD3, SiR1, and OASTL1). The results of the apple transient transformation experiment and the dual-luciferase reporter system experiment showed that the co-transformation of ERF2 and bHLH2 significantly activated the transcription of CML5 (Figure 8). Thus, the ERF2-bHLH2-CML5 module may be not only involved in the regulation of calcium signaling, but also in the regulation of H 2 S production in postharvest calcium-deficient apples. It was reported that H 2 S could reduce ROS production and increase antioxidant capacity, thereby extending postharvest life of banana [42] and inhibiting enzymatic browning of fresh-cut Chinese water chestnuts [43]. Meanwhile, in Arabidopsis thaliana roots, H 2 S induces ROS accumulation, thus inducing the appearance of Ca 2+ signal [44,45]. In the present study, calcium deficiency caused an increase in ROS production, a decrease in H 2 S production in apple peels and a reduction in post-harvest fruit quality, while further research was needed on how H 2 S was involved in calcium regulation and affected post-harvest fruit quality. Recent studies have also shown that H 2 S emission induced by chromium (Cr 6+ ) stress could be modulated by the Ca 2+ level [46]. However, the molecular mechanism of Ca 2+ -induced endogenous H 2 S emission has rarely been reported. Therefore, the mechanism of how the ERF2-bHLH2-CML5 module regulates calcium signaling and H 2 S production in calcium-deficient apples, leading to reduced quality in postharvest storage apples, needs further investigation.
The anthocyanin content of calcium-deficient peels did not differ significantly from calcium-sufficient peels at 0 and 7 DAS, but was significantly lower at 14 and 21 DAS ( Figure 2C), indicating that calcium deficiency reduced anthocyanin synthesis in postharvest stored apple peels. Moreover, in the apple transient expression assays experiment, anthocyanin deposition was observed in the peel at the injection site when ERF2, ERF17, and bHLH2 were individually transformed or cotransformed ( Figure 8A), suggesting that ERF2, ERF17, and bHLH2 may not only be involved in calcium regulation in postharvest stored apples, but also in regulating the anthocyanin biosynthesis pathway in the peel of postharvest stored apples. The mechanism of ERF2, ERF17, and bHLH2 in regulating anthocyanin biosynthesis still needs further investigation.

Conclusions
In summary, it can be concluded that the production of ROS, total phenols and flavonoids was increased in calcium-deficient peels compared to calcium-sufficient peels, but calcium content, H 2 S production, anthocyanin content, and POD activity were reduced. Calcium content, ROS, and H 2 S production were main factors between calcium-deficient and calcium-sufficient apple peels by PCA analysis. Four CMLs, seven AP2/ERFs, and three bHLHs were screened from transcriptome data analysis. In addition, ERF2-bHLH2 co-activated CML5 transcription and ERF2, bHLH2, and CML5 expression levels were significantly positively correlated with H 2 S production genes. Thus, the ERF2-bHLH2-CML5 module is not only involved in the regulation of calcium signaling, but also in the regulation of postharvest quality of fruit and H 2 S production under calcium deficiency in apples. These findings improved the understanding of the molecular basis of postharvest quality decline in calcium-deficient fruit, and the relationship between calcium and endogenous H 2 S.

Plant Materials and Treatment
The calcium-sufficient and calcium-deficient 'Honeycrisp' apples used in this study were provided by the Shandong Academy of Agricultural Sciences (Tai'an, China). Postharvest calcium-sufficient and calcium-deficient apples of similar size and free from pathogen infection were selected for sealed storage in glass containers. In total, 20 calcium-sufficient apples and 20 calcium-deficient apples were placed in individual sealed containers, fumigated with distilled water to maintain a relative humidity of approximately 85% and stored at 25 • C for treatment. After 0, 7, 14, and 21 days after storage (DAS), five calciumsufficient and calcium-deficient apples were randomly selected for peel sampling. Immediately after stripping, the peels were chilled in liquid nitrogen and stored at −80 • C. Some samples were utilized for the determination of physiological parameters, and some samples were utilized for RNA extraction. Three biological replicates were prepared for all samples. In addition, the 'Honeycrisp' apples at 60 days after full blooming used in this study were provided by Shandong Academy of Agricultural Sciences (Tai'an, Shandong Province, China) for transient transformation experiments, and were observed 5 days after injection. Apple peels from the injection sites were obtained to measure the related gene expression levels.

Determination of Calcium Content
A sample of freeze-dried peel (1 g) was dissolved in 1 mol·L −1 HNO 3 . Three replicates of each sample were examined. The samples were then diluted with a 5% solution of LaCl 3 . The calcium content of the apple peel samples was measured by atomic absorption spectrometer (Hitachi Z2000, Tokyo, Japan) with an air-acetylene flame [47]. The final calcium content in the peel was expressed as milligrams per kilogram dry weight.

Determination of H 2 S Production in Apple Peels
Lead sulfide method [48] was used to measure H 2 S production capacity. In the presence of excess substrate Cys and cofactor pyridoxal-5 -phosphate (PLP), a specific reaction between H 2 S and lead acetate was used to form a black precipitate (lead sulfide) which can be captured and observed on filter paper containing lead acetate. A sample of fresh apple peel (1 g) was dissolved in PBS supplemented with 10 mM Cys and 10 mM PLP. Six lead acetate H 2 S detection papers (Sigma, Darmstadt, Germany) were placed above the liquid phase in a closed container (covered 96-well plate) and incubated 2-5 h at 37 • C until lead sulfide darkening of the paper occurred.
The changes in H 2 S production capacity were assayed according to Hu's method [49]. In brief, a lead acetate H 2 S detection paper color was assayed by a colorimeter (model WSC-100, Konica Minolta, Tokyo, Japan) of each paper, and the values of L*, a*, and b* were obtained by averaging the data of three sites.

Determination of Total Phenol, Flavonoid, and Anthocyanin Contents
The method of Piri and Mullins [50] was used to determine the total phenol content of apple peels using a spectrophotometer at 280 nm. The calibration curve was generated using gallic acid as a reference standard.
By measuring the absorbance at 510 nm, the content of flavonoids was determined by aluminum chloride colorimetry [51]. Rutin was applied as a calibration standard.
The method proposed by Lee and Wicker [52] was used to obtain and measure the anthocyanin content. Apple peels (2 g) were ground with 10 mL of 1% hydrochloric acid methanol solution. The absorptions were measured at 530, 620, and 650 nm, and the anthocyanin content is presented as milligrams per gram fresh weight.

Superoxide anion (O 2
•− ), hydrogen peroxide (H 2 O 2 ), and malondialdehyde (MDA) production were determined based on methods previously published by Hu [50]. The MDA content of apple peels was measured using the thiobarbituric acid (TBA) reaction with values expressed as µmol g −1 . The content of H 2 O 2 and the production of O 2 •− were presented as µmol g −1 and nmol g −1 min −1 , respectively.

Determination of POD, PPO Activity, and Soluble Protein Content
Guaiacol peroxidase (POD) activity was determined based on our previously published method [49] with some modifications. Apple peels (2 g) were homogenized in 2.5 mL of extraction buffer and centrifuged at 13,400× g for 30 min at 4 • C. The supernatant obtained was used to measure the activity of POD.
The methods of Benjamin and Montgomery [53] and Beaudoin-Eagan and Thorpe [54] were used to measure the activity of polyphenol oxidase (PPO). Apple peels (2 g) were used for the preparation of PPO enzyme by homogenization in 2 mL of sodium phosphate buffer (50 mM, pH 6.8). POD and PPO activities were expressed as active unit (U) per milligram of protein, and 1 U means 1 µmol product per min.
Bradford's [55] method was applied to measure the soluble protein content, and the absorbance value of soluble protein was recorded at 595 nm.

Transcriptome Data Analysis
Transcriptome data of calcium sufficient and calcium deficient apple peels were published in the NCBI database (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA733 599, accessed on 15 September 2021) by Qiu et al. [56]. The three sequenced conditions in this transcriptome were T01, diseased peel of calcium-deficient apples; T02, healthy peel of calcium-deficient apples; and T03, peel of calcium-sufficient apples, and each condition had two replicates. The sample peels used for transcriptome were taken from ripe apples. Differentially expressed genes (DEGs) were screened from this transcriptome data (the screening criteria for DEGs are generally: Fold Change ≥ 2 and FDR < 0.01.).
Gene Ontology (GO) is an officially standardized international classification system for the functional classification of genes, that describes the attributes of genes and their outputs in any organism. Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis provided information and further understanding of secondary metabolites induced by calcium signaling. Then, DEGs were subjected to GO classification and KEGG enrichment analysis.

RNA Extraction and Real-Time Quantitative PCR Analysis
Total RNA from diseased peels of calcium-deficient apples and calcium-sufficient apple peels at 0, 14, and 21 DAS as well as transiently infected apple peel samples was extracted using the Plant RNA Isolation Kit (Foregene, Chengdu, China) and first strand ribonucleic acid was synthesized using a Primer Reverse Transcription Master Mix Kit (Takara, Tokyo, Japan). The RT-qPCR primers are shown in Table S1. The expression levels of MdTUB (TUB, accession number GO562615) and MdUBQ (UBQ, accession number MDU74358) were used as normalization genes, and the expression of relative genes was determined with the 2 −∆∆CT method. Three biological replicates were used for all analyses.

Prediction of Cis-Acting Elements in Gene Promoters
The upstream 2 kb promoter sequence of the target gene was taken from the NCBI (https://www.ncbi.nlm.nih.gov/, accessed on 24 September 2021), and the sequence was then input into the PlantPAN 3.0 (http://plantpan.itps.ncku.edu.tw/index.html, accessed on 24 September 2021) for cis-acting element prediction.

Transient Expression Assays in Apple Peels
In the transient transformation experiments, PCR amplification was performed using Phanta ® Super-Fidelity DNA Polymerase (Vazyme, Nanjing, China), and the full-length coding sequences of ERF2, ERF17, and bHLH2 were introduced into the pSAK277 vector using EcoRI and XbaI with the 35S promoter as a control. The primer sequences are listed in Table S2. The integration constructs were chemically transformed into the Agrobacterium tumefaciens GV3101 strain, and the cells were cultivated at 28 • C for 2 d. The detailed approach for the infiltration experiment has been previously reported by Voinnet et al. [57]. Five apples of similar size and growth were selected for infiltration in each combination, with four sites infiltrated in each apple. At 5 days after infiltration, apple peels were collected for RNA extraction. The negative controls were empty vector infiltrations (pSAK277).

Dual-Luciferase Reporter Assay of Tobacco Leaves
To construct the dual-luciferase reporter vector, the 2 kb upstream promoter region of CML5 (from the ATG start codon) was amplified from the genomic DNA of peels from 'Honeycrisp' apples and inserted into a pGreen II 0800-LUC binary vector. In addition, the injection solutions were prepared as in Section 5.10. Agrobacterium cells harboring the pGreen II 0800-LUC recombinant vector and pSAK277 vector, ERF2/17, bHLH2 were mixed at a 1:9 ratio. The mixture of Agrobacterium cells was injected into young N. tabacum leaves that were 2 weeks old. At 48-72 h after infiltration, the LUC and Ren activity were measured with an E1910 Dual-Luciferase ® Reporter Assay System (Promega, Madison, WI, USA).

Statistical Analysis
For the statistical analyses, Student's t-test and one-way ANOVA were performed. Significance was indicated by asterisks * (p< 0.05) or ** (p< 0.01) or different letters. The main factors were assessed by reducing dimensionality when conducting the PCA. R studio software was used for correlation analysis and heatmap analysis. All samples were assessed at least three times independently, and all data are represented as the mean ± SD.