AUXIN RESPONSE FACTOR 2 mediates repression of strawberry receptacle ripening via auxin-ABA interplay Authors: Bai-Jun Li, Yan-Na Shi, Yan-Ning Xiao, Hao-Ran Jia, Xiao-Fang Yang, Zheng-Rong Dai, Yun-Fan Sun, Jia-Han Shou, Gui-Hua Jiang, Donald Grierson, Kun-Song Chen Published: 2024 ABSTRACT AbstractCultivated strawberry (Fragaria × ananassa) is a popular, economically important fruit. The ripening of the receptacle (pseudocarp), the main edible part, depends on endogenously produced abscisic acid (ABA) and is suppressed by the high level of auxin produced from achenes (true fruit) during early development. However, the mechanism whereby auxin regulates receptacle ripening through inhibiting ABA biosynthesis remains unclear. Here, we identified AUXIN RESPONSE FACTOR 2 (FaARF2), which showed decreased expression with reduced auxin content in the receptacle, leading to increased ABA levels and accelerated ripening. Dual-luciferase, yeast one-hybrid, and electrophoretic mobility shift assays demonstrated that FaARF2 could bind to the AuxRE element in the promoter of 9-CIS-EPOXYCAROT-ENOID DIOXYGENASE 1 (FaNCED1), a key ABA biosynthetic gene, to suppress its transcriptional activity. Transiently overexpressing FaARF2 in the receptacles decreased FaNCED1 expression and ABA levels, resulting in inhibition of receptacle ripening and of development of quality attributes, such as pigmentation, aroma, and sweetness. This inhibition caused by overexpressing FaARF2 was partially recovered by the injection of exogenous ABA; conversely, transient silencing of FaARF2 using RNA interference produced the opposite results. The negative targeting of FaNCED1 by FaARF2 is a key link between auxin-ABA interactions and regulation of strawberry ripening. Introduction Fruit ripening is usually regulated by phytohormone interactions, sometime referred to as hormone crosstalk, which has become an attractive research hotspot (Fenn and Giovannoni 2021; Li et al. 2021). Abscisic acid (ABA) positively regulates nonclimacteric fruit ripening, for example in cultivated strawberry (Fragaria × ananassa), while auxin usually plays a major antagonistic role (Li et al. 2022a). In tomato (Solanum lycopersicum) ripening, there is an analogous interaction between auxin, which inhibits ripening, and ethylene, which promotes ripening in climacteric fruit (Fenn and Giovannoni 2021). In contrast, although many studies indicate that ABA mediated nonclimacteric fruit ripening is suppressed by auxin (Wang et al. 2022a), the specific link between these 2 phytohormones is still unclear yet. Cultivated strawberry is one of the most popular fleshy fruits worldwide. It is a typical nonclimacteric fruit and regarded as a model for studies on the mechanisms of nonclimacteric fruit ripening mediated by ABA (Li et al. 2022a). Numerous studies in strawberries (Fragaria spp.) have demonstrated that high ABA levels can accelerate ripening and development of biochemical quality attributes, such as texture (Molina-Hidalgo et al. 2013; Paniagua et al. 2016), pigmentation (Kadomura-Ishikawa et al. 2015; Zhang et al. 2020a), sweetness (Jia et al. 2016), and aroma (Han et al. 2015; Molina-Hidalgo et al. 2017), via regulating expressions of related structural genes and ripening-related transcription factors (TFs), such as the members belonged to NAC (Martín-Pizarro et al. 2021) and MADS-box families (Daminato et al. 2013; Lu et al. 2018). Moreover, the functions of a series of genes involved in the ABA biosynthesis and signaling pathways in cultivated strawberry fruit ripening have been verified and described (Li et al. 2022a). ABA biosynthesis directly influences ABA accumulation and transiently silencing 9-CIS-EPOXYCAROT-ENOID DIOXYGENASE 1 (FaNCED1), which encodes a limiting enzyme in the ABA biosynthesis pathway, by virus induced gene silencing, can decrease ABA accumulation and inhibit cultivated strawberry (Fragaria × ananassa “Fugilia”, “Fugilia”) ripening (Jia et al. 2011). In ABA signaling pathway, PYRABACTIN RESISTANCE 1 (FaPYR1) and MAGNESIUM-PROTOPORPHYRIN IX CHELATASE H SUBUNIT (FaABAR), 2 ABA receptor genes, have been found to be positive regulators for ripening (Chai et al. 2011; Jia et al. 2011), and several downstream signaling genes have been verified to act as activators, such as ABSCISIC ACID-INSENSITIVE 4 (FaABI4; Chai and Shen 2016), or repressors, including ABSCISIC ACID-INSENSITIVE 1 (FaABI1; Jia et al. 2013) and SUBFAMILY 2 SNF1-RELATED KINASES 2.6 (FaSnRK2.6; Han et al. 2015). Auxin plays a major antagonistic role repressing ABA-promoted strawberry ripening (Li et al. 2022a). The strawberry fruit is a pseudo carp consisting of a receptacle, the main edible part, and many achenes embedded into in the epidermis of the former, and the receptacle development relies on auxin that is produced from achenes (Nitsch 1950; Kang et al. 2013; Guo et al. 2022). Removing the achenes of strawberry fruit reduces the auxin content that reaches the receptacle and leads to a rise in ABA accumulation, which promotes ripening (Li et al. 2013; Gu et al. 2019; Li et al. 2022b, 2023). The expressions of FaNCED genes are downregulated in achene-deficient receptacles and also after treatment of intact fruits with the auxin analogue naphthylacetic acid (NAA) or auxin (3-indoleacetic acid) (Li et al. 2013; Jia et al. 2016; Medina-Puche et al. 2016; Li et al. 2023). This indicates that auxin may repress ABA biosynthesis via negatively regulating FaNCEDs. However, the mechanism whereby the high level of auxin produced from achene suppresses receptacle ripening via restraining ABA biosynthesis remains unclear. AUXIN RESPONSE FACTORs (ARFs) are core auxin signaling pathway genes that usually contain an N-terminal B3-type DNA binding domain (DBD), following an activation (AD) or repression (RD) domain that determines their role in regulating downstream gene expression They also possess a carboxy-terminal dimerization domain (CTD: domain III/IV), which allows protein-protein interactions through dimerization with AUXIN/INDOLE-3-ACETIC ACID (Aux/IAA) family proteins or other ARFs (Li et al. 2016; Roosjen et al. 2018; Li et al. 2024). ARF TFs can bind to the AuxRE elements (TGTCTC/GAGACA) in the promoters via their DBDs to regulate the downstream genes (Mockaitis and Estelle 2008; Boer et al. 2014) and participate in plant growth and development (Guilfoyle and Hagen 2007; Robert et al. 2018; Liang et al. 2020; Truskina et al. 2021) These include genes involved in fruit ripening and quality attributes, and some ARFs also act as a bridge in interactions between auxin and other phytohormones (Fenn and Giovannoni 2021). In tomato (S. lycopersicum “MicroTom”), SlARF2A and SlARF2B are expressed increasingly with fruit development and the former responds to ethylene while the latter response to auxin. Silencing one of them via RNA interference (RNAi) results in fruit ripening defects while simultaneously silencing both of them can strongly inhibit ethylene biosynthesis and ripening (Hao et al. 2015). Additionally, the expression of CpARF2 is induced by ethylene and repressed by auxin, and CpARF2 can interact with CpEIL1 (EIN3/ETHYLENE INSENSITIVE3-LIKE1), an important ethylene signal transcription factor (TF), to promote papaya (Carica papaya) fruit ripening via activating a series of ripening associated genes (Zhang et al. 2020b). MdARF2 and MdARF13 can repress anthocyanin biosynthesis by directly inhibiting the expression of the anthocyanin biosynthesis regulator gene MdMYB10 in apple (Malus domestica) fruit (Wang et al. 2018, 2020). Recently, the FveARF2 isolated from Fragaria vesca “Ruegen”, a diploid strawberry, has been found to repress pigmentation and sugar accumulation (Yi et al. 2022). SlARF4, an ARF that positively responds to auxin, can reduce chloroplast formation, repress sugar biosynthesis, and promote softening, in tomato fruit (Jong et al. 2009; Sagar et al. 2013; Wu et al. 2020). Further study showed that ABA and auxin antagonistically controlled ascorbic acid (AsA) biosynthesis via the SlMAPK8-SlARF4-SlMYB11 complex, and ABA can induce MITOGEN-ACTIVATED PROTEIN KINASE 8 (SlMAPK8) expression to inhibit the transcriptional activity of SlARF4 and activate SlMYB11 expression to promote AsA accumulation (Xu et al. 2022). In apple, auxin-activated MdARF5 positively regulates fruit ripening initiation via activating the expression of genes participating in ethylene biosynthesis (Yue et al. 2020). SlARF6A can directly downregulate S-ADENOSYLMETHIONINE SYNTHETASE 1 (SlSAM1), a biosynthetic gene in the ethylene biosynthesis pathway, and this reduces ethylene production and represses tomato fruit ripening (Yuan et al. 2019), and SlARF10 can promote chlorophyll and sugar accumulation during fruit development (Yuan et al. 2018). A recent study has shown that exogenous ABA treatments of sweet cherry (Prunus avium) can decrease the expression of PavARF8, which may negatively regulate fruit softening via activating DNA-BINDING WITH ONE FINGER 2 and −15 (PavDof2/15) and suppressing PavFaNCED1 expressions (Zhai et al. 2022). However, there is a lack of understanding about whether ARFs act as a link between auxin and ABA to play a role in fruit ripening regulation, especially in nonclimacteric fruit. In this study, we functionally characterized a strawberry ARF, FaARF2, which is upregulated by the high level of auxin in unripe fruit and shows decreasing expression during receptacle ripening. Dual-luciferase, yeast one-hybrid (Y1H), and electrophoretic mobility shift assays (EMSA) demonstrated that FaARF2 could directly bind to the promoter of FaNCED1 (proFaNCED1) to negatively regulate its expression. Moreover, combining transient gene overexpression and silencing assays with expression analyses and quality assessments showed that FaARF2 could inhibit ABA biosynthesis by repressing FaNCED1 expression to negatively regulate strawberry receptacle ripening. Based on these results, we proposed a model whereby the strawberry receptacle ripening is mediated by an auxin-FaARF2-ABA interplay mechanism, and high auxin level induces FaARF2 to inhibit strawberry receptacle ripening by repressing ABA biosynthesis prior to ripening. Results Some FaARFs may negatively regulate FaNCED1 expression and ABA production Removing achenes from the Fragaria × ananassa “Yuexin” (“Yuexin”) receptacle injected with water (DA) promoted receptacle coloration and ABA accumulation and decreased the auxin content compared with the control receptacle with achenes injected with water (Achened) (Fig. 1, A to C). In contrast, receptacles without achenes injected with NAA (DA + NAA) or exogenous ABA (DA + ABA) could, respectively, suppress or accelerate ripening phenotypes (Fig. 1, A and C). The above results indicated that the high level of auxin could repress receptacle ripening through inhibiting ABA biosynthesis. Only FaNCED1, a limiting enzyme in the ABA biosynthetic pathway, has been verified to directly influence the ABA level and ripening in the cultivated strawberry (Fragaria × ananassa “Fujilia”) receptacle (Jia et al. 2011). The homolog of FaNCED1 from “Yuexin” was identified in this study and the transiently overexpressing FaNCED1 in “Yuexin” receptacle could promote ABA accumulation (Supplementary Fig. S1) while the transiently silencing assays showed the opposite results (Supplementary Fig. S2), suggesting that FaNCED1 as a key role also participated in ABA biosynthesis in “Yuexin” receptacle. Based on the weighted gene coexpression network analysis (WGCNA) from our previous study (Li et al. 2023), FaNCED1 belongs to the ripening-related coexpression modules and is highly positively associated with ABA content changes in the “Yuexin” strawberry receptacle. Reverse transcription quantitative PCR results showed that the expressional change of FaNCED1 in the receptacles after treatments corresponded to the changes in ABA levels (Fig. 1C). These results suggest that the high level of auxin enters the receptacle represses ABA biosynthesis, possibly by downregulating FaNCED1 expression. To test this, we cloned the −2049bp sequence of proFaNCED1 from “Yuexin” strawberry and found an AuxRE element at −1325 to 1330 bp (Fig. 1D), indicating that there might be FaARFs that responded positively to auxin that were involved in negatively regulating FaNCED1. Furthermore, the result of WGCNA indicated that seven FaARFs transcripts were highly negatively correlated with the ABA content, and the heatmap analysis showed that they were decreasingly expressed as the receptacles ripened and were induced and downregulated by NAA and ABA, respectively (Li et al. 2023). The expressional profile of FaNCED1 (Fig. 1C) showed an opposite pattern, suggesting that these seven FaARFs (annotated as FaARF1, FaARF2, FaARF3, FaARF6, FaARF19a, FaARF19b by NCBI blast analysis) might negatively regulate FaNCED1 expression. Of these FaARFs, the transcriptome sequencing showed that there was a synonymous mutation between Isoform0005225 and Isoform0007925, coding sequences (CDS; Supplementary Figs. S3 and S4) which were both annotated as FaARF2, and we only cloned and obtained the Isoform0005225 CDS by PCR. The domain analysis showed that FaARF1, FaARF2, FaARF6, and FaARF19a had 3 intact ARF family domains, for B3 DNA binding, ARF, and Aux/IAA family, whereas FaARF3 lacked the Aux/IAA family domain and FaARF19b lacked the B3 DNA binding and ARF domains (Fig. 1E). ARFs regulate downstream genes via interacting between their B3 DBD and promoters of target genes (Mockaitis and Estelle 2008; Boer et al. 2014), therefore it is likely that only five FaARFs (FaARF1, FaARF2, FaARF3, FaARF6, and FaARF19a) with complete B3 DBDs, may have regulatory effects on the proFaNCED1. Screening FaARF candidates potentially involved in inhibiting FaNCED1 expression. A) Receptacles with achenes removed and injected with either ABA or NAA. White bar = 1 cm. B) The auxin (indole-3-acetic acid, IAA) content of receptacles 12 days after the treatments. Achened indicates receptacles with intact achenes injected with water; DA, de-achened, receptacles with achenes removed and injected with water. C) The ABA contents and relative expressions of FaNCED1 in receptacles 12 days after the treatments. DA + NAA and DA + ABA, receptacles without achenes injected with NAA and ABA, respectively. D) A AuxRE element was found in the promoter of FaNCED1 cloned from “Yuexin”. E) The amino acid sequences of seven FaARFs, which are associated with receptacle ripening based on the WGCNA from the previous study (Li et al. 2023), are analyzed. Domain analysis of FaARF protein sequences was carried out using the batch search tools in the Pfam database. There were 2 FaARF with the same amino acid sequences, which were annotated as FaARF2. Domain analysis identified 3 domains: B3 DNA binding, ARF, and Aux/IAA family. The data of B and C are the mean ± SE from 3 biological replicates, analyzed by Student's t-test and the different letters indicate significant differences with P < 0.05. FaARF2 is a negative regulator of FaNCED1 expression Further screening, using dual-luciferase assay, showed that the relative value of fluorescence signals after overexpressing 4 FaARF candidates, including FaARF1, FaARF3, FaARF6, and FaARF19a, were close to or higher than the control value, (i.e. signal value = 1), suggesting that they did not have a negatively regulatory effects on the proFaNCED1, while overexpressing FaARF2 caused the lowest relative value of fluorescence signals, approximately 40% decrease compared with the control value (Fig. 2A). FaARF2 was the homolog of tomato SlARF2A and SlARF2B (Supplementary Fig. S5), 2 SlARF2s which promote tomato ripening and enhance ethylene biosynthesis (Hao et al. 2015), indicating that members of the ARF2 clade of the ARF family may be involved in regulation of phytohormone biosynthesis in fruits. Moreover, the reduction in FaARF2 transcripts evaluated by qRT-PCR (Fig. 2B) followed the decrease in auxin content during receptacle ripening (Fig. 2C), and the FaARF2 expression were significantly increased and decreased in the receptacles for short term after the NAA and ABA treatments compared with the control, respectively (Supplementary Fig. S6). Additionally, combining the expression pattern of FaARF2 estimated by qRT-PCR in the receptacles under different treatments supported the conclusions from the transcriptome analysis result that FaARF2 was downregulated with receptacle ripening induced by exogenous ABA application and upregulated by NAA treatment (Li et al. 2023), further confirming that FaARF2 expression positively and negatively responded to auxin and ABA, respectively. As expected, subcellular localization investigation indicated that FaARF2 was localized to the cell nucleus (Fig. 2E). Taken together these results indicate that FaARF2 probably plays a role in negatively manipulating ABA biosynthesis via directly downregulating FaNCED1 expression, and FaARF2 was therefore selected for further characterization. Regulatory effects of FaARF candidates on the promoter of FaNCED1. A) The regulatory effects of 5 FaARF candidates on the proFaNCED1, estimated by dual-luciferase assay in tobacco (Nicotiana tabacum) leaves. B) The expression pattern of FaARF2 in receptacles during development was estimated using qRT-PCR. The different letters indicate significant differences by Fisher's Lsd with P < 0.05. R, fully ripening stage. C) The IAA content of receptacle during development was analyzed by Fisher's LSD, and the different letters indicate significant differences with P < 0.05. D) The expression patterns of FaARF2 in receptacles 12 days after the treatments were estimated using qRT-PCR, and the different letters indicate significant differences by Student's t-test with P < 0.05. E) Subcellular localization of FaARF2-GFP (green color) in stably transgenic N. benthamiana leaves with a red nuclear localization marker (NLS-mCherry). The fluorescence was imaged using a laser scanning microscope 780 microscope under 488 nm. White bar = 50 μm. The data of A, B, C, D are the means ± SE from 3 biological replicates. FaARF2 directly binds to the promoter of FaNCED1 FaARF2 had a negative regulatory effect on the transcriptional activity of proFaNCED1 containing AuxRE elements (Fig. 2A), suggesting that FaARF2 might negatively control FaNCED1 transcript levels, thereby inhibiting ABA production. Accordingly, we investigated the physical interaction between FaARF2 and proFaNCED1 to confirm whether FaARF2 could directly regulate FaNCED1. The Y1H assay autoactivation test showed that yeast with the recombination pAbAi vector containing the proFaNCED1 AuxRE element (proFaNCED1-pAbAi) could grow in SD/-Ura (SD medium lacking Ura) but not in SD/-Ura with 400 ng/mL aureobasidin A (AbA) (Fig. 3A). Moreover, yeast cells containing proFaNCED1-pAbAi and the empty pGADT7 vector could grow in the SD/-Leu (SD medium lacking Leu) but not in the SD/-Leu with 400 ng/mL AbA, while yeast cells containing proFaNCED1-pAbAi and the pGADT7 vector containing FaARF2 could clearly grow in the SD/-Leu with 400 ng/mL AbA (Fig. 3A). This indicated that FaARF2 could physically interact with the proFaNCED1 containing the AuxRE element. Furthermore, the EMSA result showed that the Probe-Biotin (i.e. dsDNA probes constructed with 3′-biotin-label based on the sequence at the −1307 to 1348 bp position of proFaNCED1) could be bound by the FaARF2-GST but not to the GST-tag protein alone (Fig. 3B). Additionally, the cold probe (i.e. dsDNA probes without 3′-biotin-label) could effectively compete with Probe-Biotin and interaction with FaARF2-GST whereas the probe with a mutant AuxRE element could not (Fig. 3B), demonstrating that FaARF2 could physically bind to the AuxRE element of proFaNCED1. Moreover, the result of dual-luciferase assay showed that the suppression of transcriptional activity for proFaNCED1 mediated by FaARF2 (Fig. 2A) was lost after mutating the sequence of AuxRE element in proFaNCED1 (Supplementary Fig. S7). The EMSA result combing with the results of Y1H (Fig. 3A) and dual-luciferase assays (Fig. 2A; Supplementary Fig. S7) indicated that FaARF2 could directly bind to the AuxRE element of proFaNCED1 to negatively regulate FaNCED1 transcription. The interaction between FaARF2 and promoter of FaNCED1. A) The Y1H assay. The auto activation test of pAbAi vector containing proFaNCED1 was performed in SD/-Ura (SD medium lacking Ura) with and without 400 ng/mL aureobasidin A (AbA). The test of physical interaction between FaARF2 and proFaNCED1 was performed in SD/-Leu (SD medium lacking Leu) with and without 400 ng/mL AbA. The empty pGADT7 vector was used as a negative control. B) The physical interaction between FaARF2 and proFaNCED1 was further verified by EMSA. The green star represents AuxRE element in the proFaNCED1. Probe-biotin and cold probe dsDNA probes with and without 3′-biotin-labelled, respectively, were designed based on the sequence containing the AuxRE element (marked in red) at −1307 to 1348 bp of proFaNCED1. Moreover, the cold probe with a mutant AuxRE element (the sequence marked in blue) was utilized in the mutant dsDNA probe, marked as Mutant probe. GST protein was used as a negative control for the purified FaARF2 fused GST protein (FaARF2-GST). The presence or absence of specific probes and proteins is indicated by (+) or (−). Transient overexpression of FaARF2 in receptacles tissue represses ABA biosynthesis and ripening Transient overexpression of FaARF2 showed that after 12 days there was an obvious difference in the coloration of receptacles injected with the recombinant pGreen II 0029 62-SK containing the FaARF2 coding sequence (OE- FaARF2), which was driven by 35S promoter, and the empty vector (OE-Control) (Fig. 4A). At 17 days after infection, OE-Control receptacles were at the full red stage while OE-FaARF2 was mottled red/light green (Fig. 4A). The results of qRT-PCR showed that the expression level of FaARF2 in OE-FaARF2 receptacles was significantly higher than that in OE-Control tissue (Fig. 4B), while the FaNCED1 expression was significantly lower in OE-FaARF2 compared with the control (Fig. 4C), indicating that overexpressing FaARF2 could repress FaNCED1 expression and transcript accumulation. Further investigation of ABA levels showed that the ABA content in OE-FaARF2 was significantly lower than that in OE-Control tissue (Fig. 4D), which was in accordance with the difference in the FaNCED1 expressional levels between OE-FaARF2 and the control. Effect of altering FaARF2 expression by transient overexpression in the receptacle on ripening. A) Transient overexpression of either empty vector and the recombination vector containing the coding sequence of FaARF2 in the strawberry receptacles, named as OE-Control and OE-FaARF2, respectively. The significant difference in coloration between OE-Control and OE-FaARF2 were measured 17 days after the start of the overexpression. White bar = 1 cm. B, C) The relative expression levels of FaARF2 and FaNCED1 in the samples were evaluated using qRT-PCR. D–G) The contents of ABA, anthocyanin, HDMF (i.e. 4-hydroxy-2,5-dimethyl-3(2)H-furanone, a characteristic aroma compound in strawberry), and sugar contents in the samples. H) The relative expressions of genes related to ripening and quality characteristics were estimated by qRT-PCR. The gene abbreviations in this figure are: SHP, SHATTERPROOF-like; RIF, Ripening Inducing Factor; ASR, ABA-stress-ripening; ABI1, ABSCISIC ACID-INSENSITIVE 1; SnRK2.6, subfamily 2 SNF1-related kinases 2.6; QR, quinone oxidoreductase; SUT1, sucrose transporter gene 1; SPS, sucrose-phosphate synthase; SPP, sucrose-6-phosphatase; INV, beta-fructofuranosidase/invertase. All of the data in this panel are the means ± SE from 3 biological replicates, respectively, analyzed by Student's t-test (*P < 0.05). Further, quality metrics, including anthocyanin, HDMF (i.e. 4-hydroxy-2,5-dimethyl-3(2)H-furanone, a characteristic aroma compound in strawberry), and sugar contents were measured in receptacles to estimate the degree of ripening between OE-Control and -FaARF2. The results showed that the anthocyanin content, the main pigment in OE-FaARF2, was significantly higher than that in OE-Control (Fig. 4E), which was in keeping with the observation that control tissue was redder than OE-FaARF2 tissue (Fig. 4A). The HDMF was detected in OE-Control but not in OE-FaARF2 (Fig. 4F). Additionally, the total content of sugar in OE-FaARF2 was decreased approximately 15% compared with the control (Fig. 4G). To further understand these changes in ripening quality metrics, the expression levels of a series of genes participating in strawberry receptacle ripening and quality development which has also been shown previously to be influenced by ABA level in strawberry fruit (Li et al. 2022a) were investigated (Fig. 4H). FaSHP (SHATTERPROOF-like, a MADS-box TF gene; Daminato et al. 2013), FaRIF (Ripening Inducing Factor, a NAC TF gene; Martín-Pizarro et al. 2021), and ABA-stress-ripening (FaASR; Jia et al. 2016) as positive-acting TF regulators of ripening were significantly suppressed in the OE-FaARF2 tissue compared with the control, while the expression of negative regulators, including FaMADS1a (Lu et al. 2018), and 2 ABA signaling genes, FaABI1 and FaSnRK2.6, was significantly upregulated. Moreover, the expression levels of FaMYB10, the main activator promoting expression of a series of anthocyanin biosynthetic genes (Medina-Puche et al. 2014), was significantly lower in OE-FaARF2 than in the OE-Control. For aroma, the crucial gene FaQR (quinone oxidoreductase) in the HDMF biosynthesis pathway (Raab et al. 2006) was significantly downregulated in OE-FaARF2 tissue compared with the control. Additionally, the expressions of several genes, including FaSUT1 (sucrose transporter gene 1; Jia et al. 2013), FaSPS (sucrose-phosphate synthase), and FaINV (beta-fructofuranosidase/invertase), related to sugar accumulation and biosynthesis in OE-control tissue were significantly higher than those in OE-FaARF2 tissue. The expression profiles of these genes were consistent with the faster development of ripening quality attributes in OE-Control tissue compared with OE-FaARF2 receptacles. Suppression of receptacle ripening by transient overexpression of FaARF2 could be partially reversed by exogenous ABA treatment Since transient overexpression indicated that FaARF2 could inhibit receptacle ripening via suppression of ABA biosynthesis (Fig. 4), we injected exogenous ABA into the receptacles 3 days after overexpressing FaARF2 to test whether this inhibition of receptacle ripening is caused by the low level of ABA induced by FaARF2. The result showed that the OE control receptacles became red whereas the OE-FaARF2 receptacles remained green. At least two-thirds of the area of OE-FaARF2 receptacles injected with exogenous ABA (OE-FaARF2 + ABA) developed a red color (Fig. 5A). Furthermore, the qRT-PCR results showed that the expression levels of FaARF2 were higher in OE-FaARF2 and -FaARF2 + ABA tissue than that in OE-Control (Fig. 5B), whereas the change in FaNCED1 transcript levels showed the opposite trend (Fig. 5C). The ABA content in OE-FaARF2 was significantly lower than that in the OE-Control (Fig. 5D), which was consistent with the previous assay result (Fig. 4D), while OE-FaARF2 + ABA tissue had the highest ABA level, due to the injection of exogenous ABA. Measurement of the ripening quality attributes indicated that in the OE-FaARF2 + ABA samples they were in general higher than in OE-FaARF2 tissue and lower than the OE-Control, respectively (Fig. 5, E to G). These results showed that exogenous ABA treatment could partially reverse the suppression of receptacle ripening caused by the low ABA level induced by FaARF2. Additionally, the gene expression levels of FaSHP, FaRIF, FaASR, FaMYB10, FaQR, FaSUT1, and FaINV, were significantly lower in OE-FaARF2 than in OE-Control and OE-FaARF2 + ABA, whereas the expression levels of FaMADS1, FaABI1, and FaSnRK2.6 were the highest in OE-FaARF2 tissue (Fig. 5H). Exogenous ABA treatments partially restore ripening of receptacles inhibited by transiently overexpressing FaARF2. A) The phenotypes of receptacles were analyzed following different combinations of treatments, including injecting water into receptacles 3 days after the overexpression of empty vector (OE-Control) and OE-FaARF2 (OE-FaARF2), respectively, and injecting exogenous abscisic acid (ABA; 500 μM) into the receptacle 3 days after the overexpression of OE-FaARF2 (OE-FaARF2+ABA). Significant differences in coloration between the receptacles under different treatments were observed at 13 days after the transient overexpression assays. White bar = 1 cm. B, C) The relative expression levels of FaARF2 and FaNCED1 in the samples were evaluated using qRT-PCR. D–G) The ABA, anthocyanin, HDMF, and sugar contents in the samples. H) The relative expressions of genes related to ripening and quality formation were estimated by qRT-PCR. All of the data presented are the means ± SE of 3 biological replicates, analyzed by Student's t-test and the different letters indicate significant differences with P < 0.05. Transiently silencing FaARF2 by RNAi promotes ABA biosynthesis and ripening of receptacles Transiently silencing FaARF2 by overexpressing a recombinant RNAi vector containing the FaARF2 transcript fragment (RNAi-FaARF2) driven by the 35S promoter, led to faster ripening of receptacles, judged by the more rapid coloration. Ripening was slower in receptacles injected with the RNAi empty vector (RNAi-Control). RNAi-FaARF2 receptacles became fully red 13 days after silencing FaARF2 while only the apical part of the receptacle became red in the OE-Control (Fig. 6A). The expression level of FaARF2 was significantly lower in RNAi-FaARF2 than the control (Fig. 6B), while the difference in FaNCED1 transcript levels showed the opposite trend (Fig. 6C), and ABA content in RNAi-FaARF2 receptacle tissue was significantly higher than that in the control (Fig. 6D). These results were opposite to those observed in receptacles overexpressing FaARF2 (Fig. 4 and 5) and indicated that silencing FaARF2 could enhance FaNCED1 expression to promote ABA biosynthesis and ripening. Effect of transient RNAi-silencing of FaARF2 on receptacle ripening. A) Transient overexpression of RNAi-FaARF2 and RNAi-empty in the strawberry receptacles. Significant differences in coloration between RNAi-Control and RNAi-FaARF2 were observed at 13 days after silencing, when the samples were collected. White bar = 1 cm. B, C) The relative expression levels of FaARF2 and FaNCED1 in the receptacle samples using qRT-PCR. D, G) The ABA, anthocyanin, HDMF, and sugar contents in the samples. H) The relative expressions of genes related to ripening and quality formation were estimated by qRT-PCR. All of the data presented are the means ± SE of 3 biological replicates, analyzed by Student's t-test (*P < 0.05). The anthocyanin content in RNAi-FaARF2 receptacles was obviously higher than that in the control (Fig. 6E), which was in accordance with the difference in the color phenotypes between RNAi-FaARF2 and RNAi-Control (Fig. 6A), and a similar difference was detected for HDMF content (Fig. 6F). Additionally, the total sugar content in RNAi-FaARF2 receptacles was raised approximately 20% compared with the control (Fig. 6G). The expression levels of genes increased by ABA included FaSHP, FaRIF, and FaASR, and negatively regulated genes included FaMADS1a, FaABI1, and FaSnRK2.6, were significantly altered in RNAi-FaARF2 receptacles compared with the control (Fig. 6H). Additionally, genes such as FaMYB10, FaQR, FaSUT1, FaSPS, and FaINV, related to anthocyanin, HDMF, and sugar accumulation in RNAi-FaARF2 were significantly upregulated. These results showed that the ripening and expression of genes involved in development of quality attributes was faster in RNAi-FaARF2 than RNAi-Control receptacles, which was opposite to the result observed when FaARF2 was overexpressed in the receptacle (Figs. 4 and 5). Therefore, all of these transient transgenic assays support the conclusion that FaARF2 plays a key regulatory role in inhibiting strawberry receptacle ripening via downregulation of FaNCED1 to limit ABA biosynthesis and the consequential slowing or inhibition of ripening. Discussion ABA has been well established as the main hormonal regulator of strawberry receptacle ripening acting by up or downregulating the expression of a wide range of genes, while the high level of auxin content can repress receptacle ripening by inhibiting ABA biosynthesis (Gu et al. 2019; Li et al. 2022a, 2023). Auxin repression of ripening has also has been found in other climacteric fruits (Wang et al. 2022a), but the mechanism is still unclear. In this study, by analyzing transcriptome data we identified an auxin response factor (FaARF2), which is positively and negatively regulated by auxin and ABA during receptacle ripening, respectively (Figs. 1 and 2). Dual-luciferase, Y1H and EMSAs indicated that FaARF2 could directly bind to the AuxRE element of ProFaNCED1 to negatively control FaNCED1 expression and inhibit ABA biosynthesis (Figs. 2A and 3). Moreover, phenotypic and molecular analyses of receptacles transiently overexpressing or inhibiting FaARF2 by RNAi confirmed that FaARF2 is a key regulator that suppresses strawberry receptacle ripening through inhibiting ABA biosynthesis to influence the accumulation of transcripts from ripening-related TFs, ABA signaling and quality-related genes, and improvements in quality metrics, including anthocyanin, characteristic aroma, and sugars (Figs 4 to 6). The development of the strawberry receptacle depends on auxin produced from the achene (Guo et al. 2022), much of which is transported from the achene to the receptacle at early developmental stages while the auxin levels in the achene are high and decreases in the achenes, and the receptacle during ripening (Symons et al. 2012; Gu et al. 2019; Li et al. 2022b). Many studies have shown that there is an phytohormone interaction in strawberry fruit, where the high level of auxin at early developmental stages to inhibit the ABA biosynthesis and ripening of the receptacle (Nitsch 1950; Kang et al. 2013; Guo et al. 2022; Li et al. 2022b, 2023). Our results showed that the auxin level in the receptacle was decreased with ripening (Fig. 2C), and removing achenes from receptacle could decrease auxin content (Fig. 1B) while promote the ABA biosynthesis and receptacle ripening, and this could be either suppressed or accelerated, respectively, by supplying an auxin analogue (NAA) or exogenous ABA (Fig. 1, A and C), indicating that the high level of auxin that reached the receptacle could suppress ABA biosynthesis in the receptacle. The limiting enzyme in the ABA biosynthesis pathway, FaNCED1, has been verified to influence the ABA levels in cultivated strawberry (Jia et al. 2011). In our previous WGCNA analysis (Li et al. 2023), we showed that FaNCED1 was closely related to ripening and ABA levels, consistent with playing a key role in ABA accumulation (Jia et al. 2011). Additionally, the transiently overexpressing and silencing assays verified that FaNCED1 cloned form “Yuexin” acted as a positive dominator in the ABA biosynthesis of receptacle, which supported to the FaNCED1 cloned from “Fugilia” as a key role of ABA biosynthesis in the receptacle (Jia et al. 2011). Moreover, the qRT-PCR confirmed that FaNCED1 expression could be increased by removing achenes, and suppressed and promoted by NAA and ABA treatments, respectively (Fig. 1C), suggesting that FaNCED1 could be the target for the high level of auxin that inhibited ABA biosynthesis and slowed ripening in the receptacle. ARFs are normally associated with auxin signaling and regulate downstream genes via binding to the AuxRE elements of their target promoters to participate in plant growth and development (Robert et al. 2018; Liang et al. 2020; Truskina et al. 2021). They can also act as a bridge in interactions between auxin and other phytohormones during fleshy fruits ripening and development of quality attributes (Fenn and Giovannoni 2021). Recently, PavARF8, which belongs to the ARF3 clade (Supplementary Fig. S5), has been found to directly bind to the AuxRE element of proPavNCED1 to suppress the transcriptional activity, using EMSA and dual-luciferase assays, indicating it might be involved in the regulation of sweet cherry softening by PavDof2/15 (Zhai et al. 2022). This evidence suggests that there may be ARFs that positively respond to auxin and negatively inhibit ABA biosynthesis in fruits. Analysis of ProFaNCED1 cloned from cultivated strawberry identified an AuxRE element at −1325 to 1330 bp (Fig. 1D), and the expressions of several FaARFs showed significantly negative correlation with ABA accumulation based on the WGCNA analysis (Li et al. 2023). Among the FaARF candidates identified, FaARF2 could significantly repress the transcriptional activity of ProFaNCED1 (Fig. 2A) and directly bind to the AuxRE element of ProFaNCED1 (Fig. 3). This gene was decreasingly expressed during receptacle development and negatively responded to ABA, as has been found for PavARF8 (Zhai et al. 2022), while expression is promoted by auxin (Fig. 2, B and D). Although they belong to different ARF subfamilies (Supplementary Fig. S5), both FaARF2 and PavARF8 could downregulate the expression of NCED1s by binding to the AuxRE elements of their promoters in strawberry and sweet cherry, respectively. This indicates that diverse subfamily members may participate in the ARF-mediated regulation of ABA biosynthetic genes in fruits. It has been unclear how the high level of auxin represses the ABA biosynthesis and ripening in strawberry of receptacle tissue (Li et al. 2013; Gu et al. 2019; Li et al. 2022b, 2023). Here, removing achenes could decrease auxin content (Fig. 1B) and downregulate FaARF2 expression in the receptacle, which was reversed by auxin treatment (Fig. 2D), suggesting that the regulation of FaARF2 targeting for FaNCED1 was a bridge between auxin and ABA in the strawberry receptacle. In this study, transiently overexpressing FaARF2 in the receptacle caused an obviously delay in coloration (Fig. 4, A and B) and significantly suppressed FaNCED1 expression and ABA biosynthesis (Fig. 4, C and D), resulting in the inhibition of the appearance of ripening-related attributes, including coloration, aroma, and sweetness (Fig. 4, E to G). ABA is the major promoter that accelerates strawberry receptacle ripening via influencing the expressions of a number of ripening-related TFs and quality-related biosynthetic genes (Sánchez-Gómez et al. 2022; Li et al. 2022a). With decreasing ABA content in the OE-FaARF2 receptacle, the ripening-related negative regulators, such as FaMADS1a (Lu et al. 2018) and two ABA signaling genes, FaABI1 (Jia et al. 2013) and FaSnRK2.6 (Han et al. 2015), were significantly upregulated in OE-FaARF2 tissue compared with OE-control while the expressions of positive TFs, including FaRIF (Martín-Pizarro et al. 2021), and FaASR (Jia et al. 2016), and genes related to ripe fruit quality development, such as FaMYB10 (Medina-Puche et al. 2014), FaQR (Raab et al. 2006), and FaSUT1 (Jia et al. 2013), were significantly decreased (Fig. 4H). These observations support the suggestion that they are involved in the inhibition of quality development and ripening in OE-FaARF2 (Fig. 4, E to G). Additionally, injecting exogenous ABA into the receptacle 3 days after the overexpression of OE-FaARF2 could partially reverse the suppression of ripening caused by the decrease in ABA content induced by overexpressing FaARF2 (Fig. 5), indicating that the reduction in ABA biosynthesis caused by downregulating FaNCED1 was the reason for the inhibition of receptacle ripening after overexpressing FaARF2 (Figs. 4 and 5). Compared with the transient overexpressing assays, transiently silencing FaARF2 using RNAi gave the completely opposite results and the RNAi-FaARF2 receptacle had higher FaNCED1 expression and a lower ABA level and was stimulated to ripen compared with the RNAi-Control (Fig. 6), confirming that FaARF2 suppressed FaNCED1 expression to inhibit ABA biosynthesis, resulting in the repression of receptacle ripening. Recently, it has been reported that FveARF2 can inhibit anthocyanin and sugar accumulation in the diploid strawberry fruit through directly suppressing anthocyanin biosynthetic and sucrose transporter genes (Yi et al. 2022). Besides formation quality traits of strawberry receptacle mediated by ARF2 directly, there is another pathway by which FaARF2 negatively regulates fruit ripening through FaNCED1 ABA-dependently. These findings manifest that FaARF2 will be an important target for shortening or lengthening the ripening period of the strawberry receptacle, to improve the timing of strawberry ripening through stable transgenic methods to meet market demands and improve shelf-life and quality. The phylogenetic tree showed that FaARF2 belongs to the ARF2 clade and has the closest relationship with SlARF2A and SlARF2B (Supplementary Fig. S5), which have been shown to promote tomato fruit ripening by enhancing ethylene biosynthesis (Hao et al. 2015). Another homologous gene, CpARF2, is up and downregulated by ethylene and auxin, respectively, and its protein interacts with an ethylene signaling TF, CpEIL1, to promote transcription of ripening-related genes and papaya fruit ripening (Zhang et al. 2020b). Tomato and papaya are climacteric fruits whereas strawberry is a typical nonclimacteric fruit. Interestingly, although they all regulate ripening by manipulating the major phytohormones, including ethylene or ABA, the ARF2s have opposite expression patterns during development and different roles in the climacteric and nonclimacteric fruit ripening, suggesting there is a different in the mechanisms whereby ARF2s mediate ripening in these two types of fleshy fruits. The opposing expression patterns of ARF2s may be caused by the difference in their promoters and upregulation mechanisms, such as DNA and RNA epigenetic modifications (Lang et al. 2017; Cheng et al. 2018; Zhou et al. 2019, 2021), and there are several potential reasons that could explain different functions between these ARF2s: (i) the Aux/IAA (AUXIN/INDOLE-3-ACETIC ACID) proteins, one of important classes of auxin signaling proteins, usually interact with ARFs to inhibit the regulation of downstream genes (Li et al. 2016; Roosjen et al. 2018) and there may be Aux/IAA proteins with different functions between these two types of fleshy fruit. (ii) Many TFs, such as the basic helix–loop helix group (Varaud et al. 2011; Oh et al. 2014), phytochrome-interacting factor 4 (Oh et al. 2014), MADS-box TF FRUITFULL and AP1 (Smaczniak et al. 2012; Ripoll et al. 2015), has been found to interact with ARFs to participate in regulating plant growth and development and there may be different interacting TFs functioning in climacteric and nonclimacteric fruits. (iii) Protein modifications, including phosphorylation (Cho et al. 2014; Simonini et al. 2016; Zhao et al. 2016) and histone modifications (Fukaki et al. 2006; Weiste and Dröge-Laser 2014; Wu et al. 2015), has been found to be involved in ARF-mediated processes and these mechanisms may vary in different fruits. Further functional studies in the climacteric and nonclimacteric fruits, especially their model fruit tomato and strawberry, are required in order to distinguish between these possibilities. Additionally, different ARFs can respond to phytohormones other than auxin, and this also needs to be checked when investigating the role of the ARF family members. Our results have revealed that FaARF2 negatively targets FaNCED1 and is a key bridge in the mechanism of the strawberry receptacle ripening regulated by auxin-ABA pathway (Fig. 7) besides the formation quality traits of strawberry receptacle mediated by ARF2 directly (Yi et al. 2022). At the early developmental stage, the high level of auxin in receptacle enhances FaARF2 expression, resulting in inhibition of receptacle ripening via downregulation of FaNCED1 to suppress ABA biosynthesis. With strawberry ripening, the low auxin level and high ABA concentration mediates a negative feedback loop in the receptacle decreased FaARF2 expression, which gave rise to ripening caused by promoting FaNCED1 expression and ABA biosynthesis. These findings enrich understandings about fruit ripening mediated by auxin-ABA pathway. The regulatory mechanism whereby auxin regulates receptacle ripening through ABA-dependent pathway mediated by FaARF2 in strawberry. At an early developmental stage, the high level of auxin enhances the FaARF2 level, which binds to the AuxRE element of proFaNCED1 to inhibit FaNCED1 expression and ABA biosynthesis, resulting in the suppression of receptacle ripening. During later development stage, the reduction in auxin level and the increase in ABA level form a negative feedback loop causing a decrease in FaARF2 expression in the receptacle, which upregulates FaNCED1 expression and ABA biosynthesis to promote ripening. Materials and methods Plant materials and sampling The cultivated strawberry (Fragaria × ananassa “Yuexin”) plants used in this study was grown in a Zhejiang Academy of Agricultural Sciences plastic greenhouse (Zhejiang, China) under natural light with daytime and nighttime temperatures of 10 to 24 °C. Receptacle samples were harvested at the G, T, HR, and R stages, for the NAA treatment, and used in transient overexpression and RNAi assays and subsequent analyses. All of the fruits used in this study were sampled after removal of achenes from the receptacle, immediately frozen in liquid nitrogen and then ground into powders and stored at −80 °C for later analysis. Removing achene and exogenous phytohormone treatments The removal of achene and exogenous phytohormone treatments were carried out as previously described (Li et al. 2023). Briefly, all of the achenes were removed from the strawberry receptacle at the G stage and injected with water, ABA (500 μM), and NAA (500 μM), respectively, while the receptacle samples retaining achenes were injected with water as negative control. Transcriptome analyses All of transcriptome data including the transcript sequences, expression patterns, and the result of WGNCA, were obtained from the accessibly public data sets from our previous study also with Fragaria × ananassa “Yuexin”, and the methods of transcriptome analyses has been described clearly (Li et al. 2023). In this study, heatmap analysis was performed via Omiscshare tools (http://omicshare.com/tools/). Domain analysis of protein sequence and visualization were conducted using the batch search tools in Pfam database (http://pfam-legacy.xfam.org/). Phylogenetic analysis The phylogenetic tree was constructed based on the neighbor-joining (NJ) method with bootstrap values evaluated from 1,000 replicate runs via Omiscshare tools (http://omicshare.com/tools/) and used to analyze the phylogenetic relationships between FaARF2 and ARF members from Arabidopsis thaliana and other plants. The protein sequence alignment was carried out using MUSCLE (Edgar 2004). The gene accession numbers of all reported ARF members are shown in the Supplementary Fig. S5 legend. Genomic DNA and total RNA extraction The genomic DNA from the Fragaria × ananassa “Yuexin” leaves was extracted using a CTAB method and used to clone proFaNCED1. The total RNA from Fragaria × ananassa “Yuexin” fruit was extracted using a CTAB method as previously described (Zhang et al. 2020a), and used to clone the full-length CDSs and qRT-PCR. Quantitative real-time PCR analysis The cDNA template was fully mixed with primers and HiScript III RT SuperMix (Vazyme, China) and used for qRT- PCR in a CFX96 Real-Time System (Bio-Rad, USA). The qRT-PCR protocol was performed according to the manufacturer's recommendations. The expressional level of each gene was calculated using 2−▵▵Ct and the FaGADPH (Castillejo et al. 2020) as housekeeping gene was used for normalization. qRT-PCR analysis of each gene was performed using three biological replicates. All primers used in qRT-PCR analysis were designed using NCBI Primer-BLAST (http://www.ncbi.nlm.nih.gov) and are shown in the Supplementary Table S1. Cloning gene promoters and full-length CDS The total RNA samples were used to synthesize cDNAs using Vazyme accordingly to the manufacturer's protocol (Vazyme, China). The proFaNCED1 and the full-length CDSs of FaARFs and FaNCED1 were cloned from the genomic DNA and cDNA using Phanta Max Super-Fidelity DNA Polymerase (Vazyme, China) accordingly to the manufacturer's recommendations. The cloning primers of the full-length CDS were designed according to our public accessibly data (Li et al. 2023). The cultivated strawberry “Camarosa” (Fragaria × ananassa “Camarosa”) was used in the breeding process of “Yuexin”, and thus the cloning primer for the FaNCED1 promoter was designed based on the promoter sequence of the FaNCED1 homolog (ID: augustus_masked-Fvb3-4-processed-gene-264.5-mRNA-1 in GDR: https://www.rosaceae.org/) in the genome of “Camarosa”. Dual-luciferase assay Dual-luciferase assay was performed as previously described (Zhang et al. 2020a) to explore regulatory effects of FaARFs on proFaNCED1. Briefly, eight full-length CDSs of FaARF candidates were recombined into pGreen II 0029 62-SK vector, respectively, while the sequence of proFaNCED1 and mutated proFaNCED1 (i.e. the sequence of AuxRE element in the proFaNCED1 was mutated) from −1 to −2049bp (Supplementary Fig. S8) was reconstructed into pGreen II 0800-LUC vector using ClonExpress II One Step Cloning Kit (Vazyme, China). The cloning primers used in this assay are shown in Supplementary Table S2, and the proFaNCED1 was directly synthesized by Zhejiang Sunya Biotechnology Co., Ltd (Hangzhou, China). The recombinant vectors were transformed into Agrobacterium tumefaciens GV3101 via electroporation, and then the A. tumefaciens was suspended by adding liquid infection medium (sterile ultrapure water containing 10 mm MES, 10 mm MgCl2, 150 mm acetosyringone, pH = 5.6) to an OD600 = 1. Subsequently, a mixture of infection liquid including 1 mL A. tumefaciens containing FaARFs and 10 μL A. tumefaciens containing proFaNCED1 was injected into one side of tobacco leaves using a needle-less syringe, while the mixture of A. tumefaciens containing 1 mL empty pGreen II 0029 62-SK vector and 10 μL proFaNCED1 was injected into the other side of leaves as the control. Three days after injecting, the discs from the tobacco leaves were sampled to measure enzyme activity of firefly and renilla luciferases using dual-luciferase reagents (Promega, America). Each combination between FaARF candidates and proFaNCED1 was independently repeated 3 times. Subcellular localization analysis Subcellular localization analysis in this study was performed as previously described (Zhu et al. 2019). Briefly, the full-length CDS of FaARF2 were recombined into pCAMBIA1300-eGFP using ClonExpress II One Step Cloning Kit (Vazyme, China), and then the recombined and empty vectors were electroporated into A. tumefaciens GV1301, and transiently overexpressed in transgenic N. benthamiana leaves that were stably transformed with the nuclear location signal NLS-mCherry (Li et al. 2017) using the same method and infiltration buffer used with the dual-luciferase assay. Imaging of fluorescence from N. benthamiana leaves was conducted with a Zeiss LSM780NLO confocal laser scanning microscope. The primers used for subcellular localization analysis are shown in Supplementary Table S3. Transient overexpression and RNAi assays in strawberry receptacle tissues The transient overexpression and RNAi assays in strawberry receptacle were performed as previous described (Zhang et al. 2020a; Li et al. 2023). All the vectors used in transient transgenic assays were driven by the CaMV 35S promoter. The full-length CDSs and RNAi fragments of FaARF2 (Supplementary Fig. S9) and FaNCED1 (Supplementary Fig. S10) were recombined into the pGreen II 0029 62-SK vector (35Spro: FaARF2-OE and 35Spro: FaNCED1-OE) using ClonExpress II One Step Cloning Kit (Vazyme, China) and pHELLSGATE 2 (35Spro: FaARF2-RNAi and 35Spro: FaNCED1-RNAi) vectors using BP Clonase (Invitrogen, MA, USA), respectively, and were then transferred into A. tumefaciens GV3101. The A. tumefaciens containing the overexpression and RNAi vectors were suspended by adding liquid infection medium (the same as for dual-luciferase assay) to OD an OD600 = 1. Then, A. tumefaciens cells carrying the 35Spro:FaARF2-OE, 35Spro:FaARF2-RNAi, empty pGreen II 0029 62-SK and pHELLSGATE 2 vectors were injected into the strawberry receptacle at the G stage (Li et al. 2023). Receptacles injected with the empty vectors were used as negative controls. Additionally, some receptacle samples were injected with exogenous ABA (500 μM) 3 days after injection of gene constructs, while the OE-Control and OE-ARF2 receptacles were injected with water as the negative and positive controls, respectively. The receptacles were collected when the significant difference between OE-Control and OE-FaARF2 and OE-FaFaNCED1 or RNAi-Control and RNAi-FaARF2 and RNAi-FaNCED1 were evident. Three biological replicates in each assay were used for subsequently analyses. The primer pair used for cloning the FaARF2 and FaNCED1 RNAi fragments is shown in Supplementary Table S4. Determination of ABA and auxin contents and quality attributes related to receptacle ripening The ABA in receptacle was extracted and quantified using high performance liquid chromatography (HPLC, Waters e2695 Separations Module, Waters, USA) as described in our previously study (Li et al. 2023). The auxin in receptacle was extracted and determined using HPLC-mass spectrometry (HPLC/MS, Agilent 1,290 infinity HPLC system coupled to an Agilent 6,460 Triple Quad liquid chromatography-mass spectrometry device, Agilent Technologies Inc., MA, USA) as described previously (Wang et al. 2016). The furanone and sugar in receptacles were extracted and measured using gas chromatography mass spectrometry (GC/MS, Agilent 7890A GC System, Agilent Technologies Inc., MA, USA) as described in our previous studies (Zhang et al. 2018; Li et al. 2022c). The anthocyanin content of receptacles was extracted and estimated using an ultraviolet spectrophotometer (UV-2600, SHIMADZU, Japan) as previously described (Wang et al. 2022b). Y1H assay The Y1H was used to test the interaction between FaARF2 and proFaNCED1 using Matchmaker Gold Y1H Library Screening System (Clontech, Japan) according to the manufacturer's recommended procedure. Briefly, the proFaNCED1 from −423 to −1461 bp and the full-length CDS of FaARF2 were cloned and recombined into pAbAi and pGADT7 vectors (ProFaNCED1-pAbAi, FaARF2-pGADT7), respectively. The ProFaNCED1-pAbAi was linearized and transformed into the Y1HGold yeast strain, which was used to test the promoter auto activation according to the manufacturer's procedure. The FaARF2-pGADT7 and empty pGADT7 were transformed into the Y1HGold yeast strain carrying ProFaNCED1-pAbAi, respectively, and used to identify the interaction between FaARF2 and proFaNCED1 and the empty pGADT7 plasmid as a negative control separately. The primer pairs used for this assay are shown in Supplementary Table S5. EMSA The FaARF2 CDS from 1 to 1200 bp was cloned and recombined into the pGEX vector carrying a GST-tag and then transferred into Escherichia coli strain Rosetta. The GST protein and FaARF2-GST fusion proteins were induced with 1 mm isopropyl β-D-1-thiogalactopyranoside at 16 °C for 18 h in E. coli strain Rosetta and then the induced proteins were purified with a GST-tag Protein Purification Kit (Beyotime Biotechnology, Shanghai, China). According to the sequence of proFaNCED1, the biotin-labeled and cold oligonucleotide probes including the AuxRE element were designed and synthesized by Sunya Biotech (Hangzhou, China). The EMSAs were performed using a LightShift Chemiluminescent EMSA Kit (Thermo Scientific, Waltham, MA, USA) according to the manufacturer's protocol. Based on the biotin-labeled and cold oligonucleotide probes, the unlabeled cold probes and mutant probes were designed and synthesized to use for competition experiment and the GST protein was used as a negative control in EMSA. The primer pairs used for this assay are shown in Supplementary Table S6. Histogram and statistical analyses Histogram analysis and statistical significance was estimated with Student's paired t-test and Fisher's Lsd was carried out using GraphPad Prism 8 (GraphPad Software, USA). Accession numbers Sequence data from this article can be found in the China National Center for Bioinformation (https://ngdc.cncb.ac.cn/genbase/?lang=en) under accession numbers FaARF2 (GB0004639) and FaNCED1 (GB0004642). Supplementary Material