SlDELLA interacts with SlPIF4 to regulate arbuscular mycorrhizal symbiosis and phosphate uptake in tomato

Abstract Arbuscular mycorrhizal symbiosis (AMS), a complex and delicate process, is precisely regulated by a multitude of transcription factors. PHYTOCHROME-INTERACTING FACTORS (PIFs) are critical in plant growth and stress responses. However, the involvement of PIFs in AMS and the molecular mechanisms underlying their regulator functions have not been well elucidated. Here, we show that SlPIF4 negatively regulates the arbuscular mycorrhizal fungi (AMF) colonization and AMS-induced phosphate uptake in tomato. Protein–protein interaction studies suggest that SlDELLA interacts with SlPIF4, reducing its protein stability and inhibiting its transcriptional activity towards downstream target genes. This interaction promotes the accumulation of strigolactones (SLs), facilitating AMS development and phosphate uptake. As a transcription factor, SlPIF4 directly transcriptionally regulates genes involved in SLs biosynthesis, including SlCCD7, SlCDD8, and SlMAX1, as well as the AMS-specific phosphate transporter genes PT4 and PT5. Collectively, our findings uncover a molecular mechanism by which the SlDELLA-SlPIF4 module regulates AMS and phosphate uptake in tomato. We clarify a molecular basis for how SlPIF4 interacts with SLs to regulate the AMS and propose a potential strategy to improve phosphate utilization efficiency by targeting the AMS-specific phosphate transporter genes PTs.


Introduction
Phosphorus (P), one of the essential macronutrients, is of indispensable importance in promoting plant growth and development, as well as improving the yield and quality of crops.Although phosphate (Pi) is abundant in agricultural soils, most of it cannot be directly absorbed by plants due to its low mobility and availability [1,2].When faced with an unfavorable environment, plants can establish various mutually beneficial relationships with microorganisms, especially with arbuscular mycorrhizal fungi (AMF) to form symbionts, called arbuscular mycorrhizal symbiosis (AMS) [3,4].Through AMS, the host plant acquires available Pi, water, and other essential nutrients, thereby improving its stress resistance [5][6][7][8].In turn, AMF achieves growth and reproduction by obtaining carbon materials including fatty acids and sugars, provided by the host plant [9,10].Therefore, it is important and highly meaningful to clarify the regulatory mechanism of AMS to improve the utilization efficiency of phosphate and achieve the sustainable development of agriculture.
AMS is a complex and delicate process, including the signaling dialog at the pre-contact stage, the penetration process and arbuscule formation, as well as the arbuscule development and degeneration, which are precisely regulated by phytohormones, especially strigolactones (SLs) [11,12].Root-exudated SLs stimulate the germination, hyphae branching, and elongation of AMF spores existing in soil, thus promoting AMS [13][14][15][16].Previous studies also showed that light signaling inf luenced the colonization of AMF.Compared with low Red/Far-Red light, high Red/Far-Red light significantly promoted the colonization of AMF in tomato and Lotus japonicus [17].While we previously revealed that the main red-light photoreceptor phyB promotes the accumulation of LONG HYPOCOTYL 5 (HY5, a light signal transcriptional factor) protein in tomato roots, which further controls the AMS by systemically regulating SLs biosynthesis [18], the mechanism underlying light-regulated AMS has not yet been fully revealed.It will be interesting to explore more molecular players that integrate light and hormone signals to regulate AMS.
PIFs (PHYTOCHROME-INTERACTING FACTORS), light-regulated bHLH transcription factors, play an important role in plant growth and development [19][20][21].There are eight PIFs in Arabidopsis thaliana and tomato, respectively [22][23][24].All PIFs consist of an APB (active phyB-binding) motif, and a bHLH functional domain including the basic DNA-binding domain and the helix-loophelix (HLH) domain [25].PIF4 is the regulatory center of various environmental signals and endogenous hormone signals involved in regulating plant growth and development as well as stress tolerance.Previous studies indicated that PIF4 not only directly induced or reduced the transcript levels of downstream genes, but also interacted with many proteins to regulate downstream response [26,27].Until now, it has been demonstrated that 25 proteins physically interact with PIF4, including photoreceptors, hormone signaling components, and so on [27][28][29][30].However, the study of PIF4 in regulating AMS remains rudimentary.DELLA, a plant-specific GRAS protein, not only mediates Gibberellins (GAs) signaling but also interacts with diverse transcription factors in many signaling pathways to interfere with or modulate their functions [31,32].Emerging evidence suggests that DELLA protein, as a central hub, is of great importance in controlling AMS development.In pea, DELLA-deficient la crys mutants show reduced arbuscular mycorrhizal colonization [33].Similarly, mycorrhizal colonization is impaired obviously in DELLA mutants in Medicago truncatula and Oryza sativa [34,35].Additionally, DELLA proteins form a complex with certain transcription factors to inf luence the AMS.DELLA interacts with CYCLOPS, a component of the common symbiosis signaling pathway, and together with CCaMK (calcium and calmodulindependent kinase), they form a complex that activates RAM1 expression, thereby regulating arbuscular development [36].DELLA proteins were also revealed to interact with MYB1 to regulate arbuscule degeneration [37].Previous studies demonstrated that DELLA could degrade PIFs protein through the 26S proteasome or sequestrate them from target genes to regulate PIFs transcriptional activity and protein stability [38,39].Although DELLA is a key component in the transcription network to regulate AMS, the relationship between DELLA and PIF4 in AMS development remains unknown and needs further study.
In this study, we show that SlPIF4 negatively regulates the AMS and AMS-induced Pi uptake in tomato, which is contrary to the function of SlDELLA.Moreover, our data demonstrates that SlDELLA physically interacts with SlPIF4 to induce the expression of SlPIF4 targeted genes, including SLs biosynthesis genes and AMS-specific phosphate transporter genes PTs, by reducing the protein stability and attenuating the transcriptional activity of SlPIF4.Together, our findings unravel a novel mechanism regarding the SlDELLA-SlPIF4 module-mediated AMS and Pi uptake via an SLs-dependent pathway in tomato plants.

SlPIF4 negatively regulates the arbuscular mycorrhizal symbiosis in tomato
Eight genes encoding proteins in tomato have been identified that are similar to PIFs in A. thaliana [24].Here, we detected the transcript levels of SlPIFs in Rhizophagus intraradices inoculated (AM) and non-inoculated (NM) roots of wild-type (WT) plants that grow under phosphate deficiency at 10 days post-inoculation (dpi).The AMS-marker genes BCP1 (BLUE COPPER-BINDING PRO-TEIN1), PT4, and PT5 (PHOSPHATE TRANSPORTER 4 and 5) were significantly induced in AM roots compared with WT NM roots (Fig. S1a, see online supplementary material).Importantly, the transcript levels of several SlPIFs declined in AM roots, whereas the expression level of SlPIF4 was the most significant (Fig. S1b, see online supplementary material).As the inoculation period extended, the root length colonization (RLC) of hyphae, arbuscules, and vesicles was consistently increased (Fig. S2a, see online supplementary material).However, the transcript level of SlPIF4 in AM roots decreased sharply at the initial stage, which was only 0.4 times of the initial level at 10 dpi, while the transcript level in NM roots showed minimal f luctuation.(Fig. 1a).At the same time, the abundance of SlPIF4 protein in WT AM roots was observably lower than that in NM roots at 10 dpi (Fig. 1b).All these results indicated that SlPIF4 may participate in regulating the AMS in tomato.
To further explore the role of SlPIF4 in mediating AMS, one SlPIF4 overexpression line (PIF4#89) and two SlPIF4 loss-offunction mutants (pif4#12 and pif4#14) using CRISPR/Cas9 were used in this study [24].Compared with WT plants, the AM roots of two pif4 mutants (pif4#12 and pif4#14) exhibited greater RLC and a higher number of mature arbuscule structures, whereas the SlPIF4 overexpression plants (PIF4#89) showed less at 20 dpi (Fig. 1c-d; S3a-b, see online supplementary material).Although both WT, pif4 mutants, and SlPIF4 overexpressing plants could form fully developed arbuscules, the arbuscule in pif4 mutants was larger than those in WT plants, whereas the opposite was observed in SlPIF4 overexpressing plants (Fig. S3d, see online supplementary material).The results were consistent with further observations made from WGA-AF488 stained images (Fig. S3e, see online supplementary material).In AM roots, the transcript level of AMS-marker gene BCP1 in pif4#12 and pif4#14 was much higher than in WT plants, which was opposite in PIF4#89 plants (Fig. 1e; Fig. S3c, see online supplementary material).These data demonstrated that SlPIF4 was a negative regulator of the AMS in tomato plants.

SlDELLA interacts with SlPIF4 in vivo and in vitro and positively regulates AMS in tomato
Considering DELLA-PIF modules integrate GA and light signals in Arabidopsis, we dissected the relationship between SlDELLA and SlPIF4 in tomato [38,39].Yeast two-hybrid (Y2H) analysis showed that only the yeast cells harboring SlPIF4-AD and SlDELLA-BD could grow on the medium (-LWAH) with 1 μM 3-AT, indicating that SlPIF4 could interact with SlDELLA (Fig. 2a).Further, as shown in the results of bimolecular f luorescence complementation (BiFC) assay, there were strong f luorescence signals when SlDELLA-YFP c and SlPIF4-YFP n co-expressed in Nicotiana benthamiana leaves.However, we could not detect f luorescence signals in other leaves co-expressing YFP c and SlPIF4-YFP n , SlDELLA-YFP c and YFP n , as well as YFP n and YFP c (Fig. 2b).Luciferase complementation assay (SLCA) also confirmed this interaction.Strong luciferase activity was detected only in co-expression combination with cLUC-SlPIF4 and nLUC-SlDELLA, reaching nearly 170 times the level observed in the control co-expressed cLUC and nLUC (Fig. 2c-d).Moreover, glutathione-S-transferase (GST) pulldown assay revealed that MBP-SlPIF4 fusion protein was capable of pulling down GST-SlDELLA fusion protein, rather than MBP protein (Fig. 2e).
Since DELLA proteins function as a central node to control the AM development [36,40], we then detected the expression level and protein abundance of SlDELLA in NM and AM roots of WT plants at 10 dpi.Although there was no difference in the transcript level of SlDELLA between the NM roots and AM roots, the protein abundance was significantly increased in WT AM roots at 10 dpi in comparison with that in NM roots (Fig. S4a-c, see online supplementary material).We further used procera (pro) mutant, a putative DELLA mutant [41], and WT plants to test the involvement of SlDELLA in regulating AMS.The results indicated that there were fewer hyphae, arbuscules, and vesicles in pro roots in comparison with WT roots at 20 dpi (Fig. S5b, see online supplementary material).The results were compatible with the WGA-AF488 staining images (Fig. S5a, see online supplementary material).The transcript level of BCP1 was increased in AM roots of WT plants, but significantly decreased in pro roots at 20 dpi (Fig. S5c, see online supplementary material).Collectively, these data demonstrated that SlDELLA physically interacts with SlPIF4 and positively regulates the AMF colonization in tomato.

SlDELLA acts upstream of SlPIF4 to regulate arbuscular mycorrhizal symbiosis
To clarify the genetic relationship of SlDELLA and SlPIF4, we silenced SlDELLA in WT and pif4#12 plants using virus-induced Asterisks indicate a statistically significant difference from the control in the means ( * * P < 0.01; Student's t test).For (a-e), the plants were grown in a sterilized soil: quartz sand: vermiculite mixture (1:1:1) inoculating with (AM) or without (NM) R. intraradices under Pi deficient condition (without KH 2 PO 4 but with the addition of 1 mM KCl).gene silencing (VIGS) technology.The plants with 70-80% silencing efficiency were used in the study (Fig. S6a, see online supplementary material).The deletion of SlPIF4 and reduction of SlDELLA significantly increased and decreased the RLC of hyphae, arbuscules, and vesicles compared with the WT-pTRV plants at 20 dpi, respectively (Fig. 3b), which also be clearly ref lected in the images of WGA-AF488 staining (Fig. 3a).These results indicated that SlPIF4 and SlDELLA played opposite roles in regulating the mycorrhizal colonization.However, silencing SlDELLA in the pif4 mutant (pif4#12) plants did not result in a decrease in mycorrhizal colonization rate compared with pif4#12 plants.The mycorrhizal colonization level of the pif4#12-pTRV-SlDELLA plants was almost the same as that of pif4#12 plants (Fig. 3a-b).Similarly, the transcript levels of AMS-marker genes (BCP1, PT4, and PT5) were lower in pTRV-SlDELLA plants and higher in pif4#12-pTRV than in WT-pTRV plants.Importantly, when silencing SlDELLA in pif4#12-pTRV plants, there was no difference in the transcript levels of these tested AMS-marker genes compared with the pif4#12-pTRV plants (Fig. 3c-e).These results indicated that SlDELLA acts upstream of SlPIF4 to regulate the AMS in tomato.

SlPIF4-mediated AMS is partly dependent on SLs biosynthesis, accumulation, and exudation in tomato roots.
Under Pi deficiency, the host plants will synthesize and secrete SLs into the soil [42,43].Secreted SLs stimulate the AMF spore germination, hyphal branching, chitin oligomer production, and soon, thus facilitating symbiosis [16,44,45].Based on this, we used two SlPIF4 mutants and WT plants to explore the relationship between SlPIF4 and SLs.Under sufficient Pi conditions, the deletion of SlPIF4 apparently increased the relative content of root-exudated SLs (Fig. S7a, see online supplementary material).Importantly, the SLs relative content and the transcript levels of SlCCD7, SlCCD8, and SlMAX1 in AM roots of pif4#12 and pif4#14 plants (e) Pull-down assay showing that MBP-tagged SlPIF4, but not MBP alone, could pull down GST-tagged SlDELLA in vitro.Recombinant MBP-SlPIF4 and GST-SlDELLA were detected with anti-MBP and anti-GST, respectively.The red arrow represents MBP-SlPIF4.
were higher than WT plants at 10 dpi (Fig. 4a-d).In addition, Pi starvation also induced the biosynthesis and root exudation of SLs (Fig. S7a-d, see online supplementary material).On the contrary, the PIF4#89 plants displayed a decrease in SLs content and the transcript levels of SlCCD7, SlCCD8, and SlMAX1 both in NM and AM roots, and the downtrend was more significant in AM roots (Fig. S8a-d).
To demonstrate that SLs were regulated by SlPIF4 in AMS development, the synthetic SL rac-GR24 and SlPIF4 overexpression plants were used in the study.At 20 dpi, rac-GR24 application significantly increased the colonization of AMF in WT plants showing more hyphae, arbuscules, and vesicles (Fig. 4e-f).Similarly, the transcript levels of BCP1, PT4, and PT5 were also increased by rac-GR24 (Fig. 4g; Fig. S8e and f, see online supplementary material).In comparison with the WT plants, the AMF colonization was decreased in PIF4#89 plants.Significantly, the defects of AMF colonization and transcription of related genes in PIF4#89 plants can restore partially by the application of rac-GR24 (Fig. 4e-g; Fig. S8e and f, see online supplementary material).Taken together, SlPIF4 negatively regulated the AMS in tomato by modulating the biosynthesis and secretion of SLs.

SlPIF4 directly binds to and inhibits the promoters of strigolactones biosynthesis genes
It has been suggested that PIF4 could bind to the G/E-box motifs (CANNTG) to modulate downstream gene expression [46].There are two E-box motifs in the promoters of SlCCD7, SlCCD8, and SlMAX1, respectively, while there is only one G-box motif in SlCCD8 and SlMAX1 promoters, respectively (Fig. S9, see online supplementary material).Therefore, recombinant MBP-SlPIF4 protein and SLs biosynthesis genes-related probes were used in EMSA assay.MBP-SlPIF4 could bind to the SlCCD7 and SlCCD8 probes with E-box motif and the SlMAX1 probe with G-box motif, respectively (Fig. 5a).However, the bound probes decreased gradually with the increase of competitors, and there were no bound probes when the probes lacked E/G-box motif (Fig. 5a).Then, in chromatin immunoprecipitation (ChIP) assay, AMF-inoculated SlPIF4 overexpressing (PIF4-OE) and WT plants at 10 dpi were used to examine the binding ability of SlPIF4 in vivo.After being immunoprecipitated with anti-HA antibody, the promoter fragments of SlCCD7, SlCCD8, and SlMAX1 in OE-SlPIF4 lines were 4.4, 3.9, and 3.5 times more abundant than those in WT plants, respectively.On the contrary, there were no combinations of SlPIF4-HA and DNA fragments that could be immunoprecipitated after immunoprecipitation with IgG control antibody (Fig. 5b).Finally, dual-luciferase transient expression (LUC) assay demonstrated that SlPIF4 significantly inhibited the transcripts of SlCCD7, SlCCD8, and SlMAX1 (Fig. 5c).Taken together, SlPIF4 could directly bind to the SlCCD7, SlCCD8, and SlMAX1 promoters to suppress their transcription.

SlPIF4 negatively regulates the AMS-induced phosphate uptake by inhibiting the transcription of AMS-specific PTs
AMS is an adaptive strategy for plants to improve phosphate uptake under phosphate-deficient conditions.PT4 and PT5, the AMS-inducible phosphate transporters, were proposed to import phosphate on the symbiotic interface [10,47].Compared with WT NM plants, the content of phosphorus increased in AM plants at 30 dpi.Surprisingly, AMS-induced phosphorus accumulation was drastically promoted in pif4#12 and pif4#14 plants and restrained in PIF4#89 plants (Fig. 6a and b).Consistently, the transcripts of the AMS-specific phosphate transporter genes PT4 and PT5 were lower in the PIF4#89 plants, but higher in the roots of pif4#12 and pif4#14 plants than in WT plants (Fig. 6c and d).These results suggested that SlPIF4 deletion could promote AMS-specific phosphate transporter genes expression, and thus the phosphate uptake in tomato.
To dissect whether SlPIF4 inf luence the transcript levels of AMS-specific PTs by binding to their promoters, we analysed the promoters of PTs, and found two and three E-box motifs were present in the promoters of PT4 and PT5, respectively (Fig. S10, see online supplementary material).EMSA assay was performed to confirm that SlPIF4 could respectively bind to the E-box motif of PT4 and PT5, but unable to bind to mutant probes (Fig. 6e).The bound probes decreased gradually with the increase of competitors (Fig. 6e).Furthermore, ChIP-qPCR analysis substantiated that SlPIF4 binds to the PT4 and PT5 promoters (Fig. 6f).We finally carried out a dual-luciferase assay to show that SlPIF4 could inhibit the transcription of PT4 and PT5 (Fig. 6g).Taken together, SlPIF4 could directly bind to the G/E box motifs of AMS-specific PTs promoters to reduce genes expression and suppress their transcription.

SlDELLA improves the expression of SlPIF4 target genes by reducing the protein stability and attenuating the transcriptional activity of SlPIF4
Having ascertained that SlDELLA interacts with SlPIF4 genetically and physically, we sought to determine their functional interplay.
To test this, we first detected the SLs content of pro plants and found that the relative SLs content was lower than that in WT AM plants at 10 dpi (Fig. S11a, see online supplementary material).Additionally, the transcript levels of SlCCD7, SlCCD8, and SlMAX1 followed the same trend as the relative content of SLs (Fig. 7a; Fig. S11c and d, see online supplementary material).In WT plants, the content of phosphorus increased in AM plants in comparison with NM plants at 30 dpi.However, the extent of AMS-induced increase in P content was less pronounced in the pro mutant than in WT plants at 30 dpi (Fig. S11b, see online supplementary material), which corresponds with the expression levels of PT4 and PT5 (Fig. 7b).These results demonstrate that the biosynthesis of root SLs and Pi uptake are inhibited by SlPIF4 but promoted by SlDELLA.
To further explore the molecular mechanism underlying the inf luence of SlDELLA on SlPIF4 protein at a molecular level, we first asked whether SlDELLA might regulate the transcript level of SlPIF4 by using the WT and pro plants.The results indicated that the SlPIF4 transcript level between WT and pro plants remained almost unchanged (Fig. S11e, see online supplementary material).In contrast, the SlPIF4 protein abundance was sharply increased in pro plants in comparison with WT plants according to the immunoblot analysis (Fig. 7c), indicating that SlDELLA affected SlPIF4 protein accumulation by reducing its stability.Cell-free degradation assay was carried out to further confirm the result.The total protein extracted from WT and pro plants was incubated with equal amounts of purified recombinant MBP-SlPIF4 protein, respectively.The results showed that MBP-SlPIF4 was gradually degraded with the extension of the ATP application time.However, the MBP-SlPIF4 protein degradation rate was slower with pro-extracted protein than that with WT-extracted protein (Fig. 7d; Fig. S11f, see online supplementary material).Additionally, we performed dual-luciferase assay to verify that SlDELLA inf luences the transcriptional activity of SlPIF4.When SlPIF4 and SlDELLA co-expressed with the gene promoters, there was no change in compared with the control (Fig. 7e and f; Fig. S12a-c, see online supplementary material).To further substantiate the results, we carried out EMSA assay using recombinant protein and probes.The results indicated that the bound probes were reduced constantly along with the increase GST-SlDELLA, but not the GST (Fig. 7g and h; Fig. S12d-f, see online supplementary material).Collectively, these data indicated that SlDELLA reduced the SlPIF4 protein stability and attenuated its transcriptional activity.

Discussion
Arbuscular mycorrhizal symbiosis is precisely controlled by many transcription factors from the pre-signals exchange to the final arbuscule formation and degradation [48,49].Although the role of many transcription factors has been revealed, such as GRAS family and AP2/ERF family, the underlying molecular mechanisms of other TFs regulating AMS are not well elucidated [50,51].Previous studies indicated that light, especially high R/FR light significantly promoted AMF colonization in tomato and L. japonicas [17].Our recent study also pointed out that a phyB-HY5-SLs cascade signaling facilitated the AMS [18].Additionally, the transcript level of PIF3 decreased in roots and leaves of trifoliate orange inoculated with AMF (Funneliformis mosseae) [52,53].Here, we revealed the transcript and protein abundance of SlPIF4 significantly decreased in mycorrhizal tomato roots (Fig. 1a and b).Further genetic analysis showed that SlPIF4 negatively regulates AMF colonization (Fig. 1c-e; Fig. S3, see online supplementary material).
The DELLA proteins have central functions in modulating AM development.On the one hand, the deficiency of DELLA resulted in a significant decrease in mycorrhizal colonization in M. truncatula, O. sativa, and pea [34,35].Consistent with this, we found that SlDELLA positively regulated the AMS in tomato.The hyphae, arbuscules, and vesicles in pro roots were less than in WT plants, as well as the transcript levels of BCP1 (Fig. S5a-c, see online supplementary material).On the other hand, DELLA can also interact with other transcription factors to inf luence the AMS, such as CYCLOPS, NSP2, and MYB1 [36,37,54].Additionally, DELLA interacted with PIFs to promote its degradation or sequestrating it from targeted genes in Arabidopsis [38,39].Here, we also found SlDELLA interacted with SlPIF4 physically by Y2H, BiFC, SLCA, and pull-down assays (Fig. 2a-e).Surprisingly, both SlPIF4 and SlDELLA regulated the SLs biosynthesis and AMS-induced Pi uptake, and also the transcript levels of SLs biosynthesis genes and AMS-specific PTs, but acted opposite roles (Figs 4, 6, 7a and b; Figs S8a-d and S11a-d, see online supplementary material).NSP1 and NSP2, the GRAS-type transcriptional regulators, not only regulated SLs biosynthesis controlling the expression of D27 and MAX1 in M. truncatula and rice, but also were required for AMF infection, and the SLs and mycorrhization levels decreased in double nsp1/nsp2 mutants [55].Additionally, DELLAs could form a protein complex with NSP2-NSP1 and other proteins to inf luence mycorrhizal symbiosis [35,37,54].Here, we also showed that SlDELLA interacted with SlPIF4 to regulate the SLs biosynthesis to control the AMS in tomato; however, the relationship between NSP2-NSP1 and DELLA-PIF4 in the regulation of AMS in tomato needs further study.
Previous studies have shown that PIFs bind to G/E-box motifs (CANNTG) to regulate target gene transcriptionally [25].Our previous study indicated that PIF4 directly regulated the expression of DELLA in response to cold stress [24].PIF4 and PIF5 also regulated the expression of IAA19 and IAA29 by binding to the G-box (CACGTG) motifs in their promoters to participate in auxin-mediated hypocotyl phototropic growth [56].PIF1 directly regulated PORC in a G-box-dependent manner to control chlorophyll biosynthesis in Arabidopsis [57].Consistently, our study showed that SlPIF4 suppressed the transcription of SLs biosynthesis genes and AMS-specific PTs by binding to G/E-box Figure 8.A proposed model showing SlDELLA-SlPIF4-SLs/PTs signaling pathway regulates AMS and phosphate uptake in tomato.SlDELLA interacts with SlPIF4 to reduce SlPIF4 protein stability and its transcriptional activity toward downstream target genes, including SLs biosynthesis genes and AMS-specific phosphate transporter genes, thus promoting SLs accumulation, AMS development, and Pi uptake in tomato.motifs of their promoters using dual-luciferase, ChIP-qPCR, and EMSA assays (Figs.5a-c and 6e-g).Additionally, SlPIF4 activity is tightly controlled by SlDELLA.There was a significant difference in SlPIF4 protein abundance rather than SlPIF4 transcript between WT and pro plants (Fig. 7c; Fig. S11e, see online supplementary material).We also found that SlDELLA enhanced the degradation rate of SlPIF4 by cell-free protein degradation assay (Fig. 7d).Besides, SlDELLA inhibited the transcription of SlPIF4 to its downstream genes (Fig. 7e-h; Fig. S12, see online supplementary material).Consistent with this, the sequestration and degradation of PIF3 by DELLAs contribute to a reduction of PIF3 binding ability to its target genes in Arabidopsis [38].
AMS are well known for their role in promoting mineral nutrition acquisition of plants, especially phosphorus, and the symbiotic route is responsible for about 70% of the overall Pi delivery [58,59].Obviously, in the study, SlPIF4 negatively regulates the SLs biosynthesis to control AMS, thus inf luencing the phosphate uptake (Fig. 8).Except that, SlPIF4 also inhibited the transcription of AMS-specific PTs.In M. truncatula, PHT1;4 (PT4), the AMS-specific Pi transporter gene, was mutated, which not only impaired the development of the interaction but also inhibited the symbiotic Pi uptake [60].Consistent with this, mutations in either PT11 or PT13 in O. sativa affected the development of symbiosis [59,61].In tomato, we found that the PT11 ortholog PT4 and PT5 in two pif4 mutants were increased significantly compared with WT plants, which was the opposite in PIF4#89 plants (Fig. 6c and d).More importantly, the deletion of SlPIF4 significantly improved the AMS-induced Pi uptake (Fig. 6a), and the AMS-induced P content was lower in PIF4#89 plants than in WT plants (Fig. 6b).Except for symbiotic Pi uptake, plants can also absorb Pi directly, in which the phosphate transporters PHT1 family genes are most relevant for Pi uptake from soil [62].Previous studies reported that PIF4 and PIF5 inhibited the Pi uptake and accumulation directly by negatively modulating the expression of Pi uptake and Pi starvation-responsive genes in A. thaliana [63,64].Here, we also found that the total P contents in pif4#12 and pif4#14 plants were higher than in WT plants inoculated without R. intraradices (Fig. 6a), which indicated that SlPIF4 may regulate Pi uptake under Pi deficient conditions in an AMS independent pathway.However, whether and which PHT1 family genes are involved remains to be further verified in tomato.
In conclusion, our findings revealed a novel SlDELLA-SlPIF4-SLs/PTs signaling pathway that regulates AMF colonization and Pi uptake in tomato.SlDELLA interacted with SlPIF4 and positively regulated AMF colonization, which was contrary to the regulation of SlPIF4 to AMS.As a transcription factor, SlPIF4 regulates SLs biosynthesis and AMS-induced phosphate uptake by binding to the G/E box motifs of the SLs biosynthesis genes and AMS-specific PTs promoters directly.SlDELLA improves the transcript levels of SlPIF4 target genes by reducing protein stability and attenuating the SlPIF4 transcriptional activity (Fig. 8).Our finding reveals a regulatory mechanism of SlDELLA-SlPIF4 module in AMS.We clarify a molecular basis that SlPIF4 interacts with SLs signaling to regulate the AMS and provide a potential way to improve phosphate utilization efficiency by targeting the AMS-specific phosphate transporter genes PTs.

Plant materials, growth conditions, and chemical treatments
The wild-type (WT, Solanum lycopersicum cv Ailsa Craig) and procera (pro) mutant, a putative DELLA mutant, were used in this study.The SlPIF4-overexpressing and CRISPR/Cas9 SlPIF4 mutants plants were generated as the previous description [24].The VIGS technique was used to generate SlDELLA silencing tomato plants.The target cDNA of SlDELLA was amplified using the specific primers presented in Table S1 (see online supplementary material), and then inserted into TRV2.VIGS was conducted as the previous description [65].
The VIGS plants were grown in the chambers at 21 • C (day) and 20 • C (night), with a photoperiod of 12 h/12 h (day/night) and a light intensity 200 μmol m −2 s −1 .Other tomato plants were grown in the chambers at 25 • C (day) and 20 • C (night) with the same photoperiod and light intensity.The mycorrhizal inoculation method was conducted as the previous description [66], and the AMF species R. intraradices UT126a [15] with a mixture of quartz sand and vermiculite (1:1, v/v) was used in this study.4-leaf stage seedlings were grown in a 600 mL pot.One plant per pot inoculated with ∼400 R. intraradices spores.Using Hoagland's nutrient solution (1 mM KCl replaces KH 2 PO 4 ) the pot plants were watered three times a week.For the chemical treatments, tomato seedlings were treated with l μM rac-GR24 (CR9420; Coolabel (Beijing, China)) by application to the soil three times a week.Seedlings treated with ddH 2 O served as control.The first application started at 3 dpi (days postinoculation).

Morphological analyses and microscopy
Mycorrhizal colonization was visualized using f luorescence imaging, following established procedures [67].Brief ly, segments of mycorrhizal roots were immersed in 50% ethanol (v/v) for over 4 hours, and subsequently acidified with 0.1 M HCl for approximately 2 hours following incubation in 20% KOH (w/v) for about 3 d.The roots were then rinsed and submerged in a staining solution of WGA-Alexa Fluor 488 (1 μg mL −1 in PBS) in the dark for 12 h.Imaging was observed using a laser scanning confocal microscope (Leica TCS SL, Leica Microsystems (Heidelberg, Germany)) with standard settings for WGA-Alexa Fluor 488.
For quantification of root length colonization (RLC), the roots inoculated with R. intraradices were treated under 95 • C as follows: 10% KOH (w/v) for 35 min, 2% HCl for 5 min; ddH 2 O for rinsing three times, trypan blue (0.05%) for staining at 95 • C for 10 min, and lactic acid mixed with glycerin (1:1, v/v) used for bleaching for 3 days.Images were taken using Leica DM4000B microscope (Leica Microsystems (Heidelberg, Germany)).Per treatment contains 6-12 plants and about 300 root segments were placed on a slide with a 0.5 cm gridline to assess the RLC according to the previous description [68].The RLC was calculated according to the following formula: (hyphae, arbuscules, vesicles, and hyphopodia)/intersections of root segment and grid.The size of the grid is 0.5 cm [15].

Measurements of strigolactones content
The measurement of SLs in the roots was performed as the previous description [66] with some modifications.Homogenized 0.5 g roots samples with 40% acetone (v/v) and the supernatant removed after centrifugation at 4 • C. Next, extracted with 500 mL 50% acetone (v/v) and collected the supernatant after centrifugation.Subsequently, samples were evaporated under a vacuum to 1 mL and then extracted with ethylacetate and vortexed centrifugation.A total of 750 mL organic phase was evaporated under vacuum to dry, and redissolved with 25% acetonitrile (1:3 v/v).The SLs content was analysed by LC-MS/MS.
The extraction of root exudates SLs was consulted [69] and made some necessary changes.Four-leaf stage seedlings were cultured under P deficient (-P; 0.95 mM KCl replaces 0.05 mM KH 2 PO 4 ) condition for 10 days.The root exudates were collected after 6 hours.The exudates (50 mL) were passed through the CNWBOND HC-C18 SPE cartridge (SBEQ-CA0854; CNW (Shanghai, China)) pre-equilibrated with methanol.Subsequently, root exudates were collected with 100% acetone after washing with ddH 2 O.Samples were evaporated under a vacuum to 1 mL, then extracted with ethylacetate and subjected to vortex centrifugation.A total of 750 mL organic phase was evaporated under a vacuum to dry and redissolved with 25% acetonitrile (1:3 v/v).The root exudates SLs were analysed by LC-MS/MS.

Measurements of P content
The P measurement was conducted according to the previous description [70].Brief ly, the total plants were harvested and heated or steamed to de-enzyme at 105 • C for 30 min and dried at 65 • C for 3 days.The 0.1 g samples were weighed and digested with H 2 SO 4 at 150 • C. For redox reaction, 30% H 2 O 2 was added and the homogenates diluted to 50 mL (V 2 ) using ddH 2 O after filtration.The 5 mL homogenates were made up to 45 mL with ddH 2 O and adjusted the pH (7 ∼ 8) using 6 M NaOH and 2 M H 2 SO 4 , then mixed with 5 mL assay solution (V 1 , 4.5 M H 2 SO 4 , 0.5% antimony potassium tartrate, 85 mM L-ascorbic acid and 1% NH 4 MoO 4 ), and reacted at 28 • C for 30 min (V 3 = 50 mL).The P content was calculated based on the absorbance at 700 nm:

RNA isolation and quantitative real-time PCR (RT-qPCR) analysis
RNA extraction, reverse transcription, and real-time PCR assay were performed according to the previous description [15].The samples were extracted from tomato roots.The relative expression was calculated by the 2 -ΔΔCT method [71].The ACTIN2 was a reference gene for normalization.Primers are presented in Table S2 (see online supplementary material).

Recombinant proteins and electrophoretic mobility shift (EMSA) assay
The MBP-SlPIF4 was constructed by ligating the CDS region of SlPIF4 into the MBP vector by homologous recombination and heat shocked into Escherichia coli strain BL21 (DE3).The 0.1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) was used to induce the MBP-SlPIF4 protein for 12 h.Amylose Resin High Flow was used to purify the MBP-SlPIF4 protein.The SlCCD7, SlCCD8, SlMAX1, PT4, and PT5 promoter probes containing a G/E-box motif and the mutant probes were labeled by biotin using the Biotin 3 End DNA Labeling Kit (catalog.No. 89818; Pierce (Rockford, IL, USA)).The labeled and unlabeled oligonucleotides were annealed to double-stranded DNA probes.The annealed and unlabeled oligonucleotides were used as competitor probes.EMSA assay was carried out by the LightShift Chemiluminescent EMSA kit (catalog.no.20148; Thermo Fisher Scientific (Waltham, MA, USA)).The primers are listed in Table S3 (see online supplementary material).

Bimolecular fluorescence complementation (BiFC) assay
The YFP n -SlPIF4 and YFP c -SlDELLA were constructed by ligating the CDS regions of SlPIF4 and SlDELLA into the N-terminal of YFP (YFP n ) and the C-terminal of YFP (YFP c ) vectors using the primer sequences listed in Table S6 (see online supplementary material).A. tumefaciens strain GV3101 containing YFP n / YFP c -SlDELLA, YFP c / YFP n -SlPIF4, YFP n -SlPIF4/ YFP c -SlDELLA and YFP n /YFP c was rinsed with infiltration buffer (10 mM MES, 150 μM acetosyringone, and 10 mM MgCl 2 (pH 5.6)) and adjusted OD 600 = 0.75 and the bacterium was infiltrated into N. benthamiana leaves, respectively.Fluorescence was observed using an A1 confocal laser scanning microscope (Nikon, Tokyo, Japan) after 48 hours.

Yeast two-hybrid (Y2H) assay
The AD-SlPIF4 was constructed by ligating the CDS region of SlPIF4 into the pGADT7 vector by homologous recombination.The BD-SlDELLA was constructed by ligating the CDS region of SlDELLA into the pGBKT7 vectors by homologous recombination.The primers are presented in Table S8 (see online supplementary material).The recombined constructs were transformed into the yeast strain Y2HGold and then cultured on SD medium (−Leu/−Trp) for 2 days or SD medium (−Leu/−Trp/−Ade/-His) medium for 5 days at 30 • C.

In vitro pull-down assay
The MBP-SlPIF4 and GST-SlDELLA were constructed by ligating the CDS regions of SlPIF4 or SlDELLA into the MBP or GST vectors by homologous recombination using the primer sequences presented in Table S9 (see online supplementary material), respectively.MBP and MBP-SlPIF4 fusion proteins were extracted with extraction buffer [Tris-HCl (1 M, pH = 8.0), EDTA (0.5 M, pH = 8.0), NaCl (5 M), and Triton X-100] and kept immobilized on Amylose Resin (NEB).GST-SlDELLA was purified using Pierce Glutathione Agarose (Thermo Fisher Scientific (Waltham, MA, USA)).Amylose Resin (NEB) containing MBP or MBP-SlPIF4 was incubated with GST-SlDELLA at 4 • C for 4 hours using the SDS-PAGE to separate the pulled-down proteins after washing.Using anti-GST antibody, anti-MBP antibody (Cell Signaling Technology (Boston, MA, USA)), and anti-mouse antibody (AP124P; Millipore (Billerica, MA, USA)) to determine the protein.

Statistical analysis
GraphPad Prism 9.0 Software was used for analysing the data.Protein quantification was fulfilled with ImageJ.SPSS statistical software was used to conduct the statistical analysis.One-way ANOVA followed by Tukey's test or two-tailed Student's t-test was performed.For each experiment, at least three biological replicates were used.The samples were taken from two different plant roots as a biological replicate.

Figure 1 .
Figure 1.SlPIF4 negatively regulates the AMF colonization in tomato.(a) Temporal expression pattern of SlPIF4 in roots of WT plants.Dpi: days-post-inoculation.Data are presented as the means of three biological replicates (±SD).Asterisks indicate a statistically significant difference from the control in the means ( * * P < 0.01; Student's t test).(b) Immunoblots showing the SlPIF4 protein level in roots of WT plants at 10 dpi.The same membrane was split into two sheets and incubated with anti-PIF4 and anti-HSP70 antibodies, respectively.Representative pictures are shown on the left.The relative protein levels are shown on the right of the blots, and the relative protein levels of WT in NM roots were set to 1.00.Data are presented as the means of three biological replicates (±SD).Asterisks indicate a statistically significant difference from the control in the means ( * * P < 0.01; Student's t test).(c) Representative images of WGA-AF488-stained roots of WT and pif4 mutant plants at 20 dpi.BF, bright-field image.Merge, merged WGA-AF488 and BF image.Scale bar = 100 μm.(d) The root length colonization of hyphae, arbuscules, vesicles, and hyphopodia in the roots of WT and pif4 mutant plants at 20 dpi.Data are presented as the means of four biological replicates (±SD).Each replication contains 80 root segments of two plants.Asterisks indicate a statistically significant difference from the control in the means ( * * P < 0.01; Student's t test).(e) Transcript of AMS-marker gene BCP1 in roots of WT and pif4 mutant plants at 20 dpi.Data are presented as the means of four biological replicates (±SD).Asterisks indicate a statistically significant difference from the control in the means ( * * P < 0.01; Student's t test).For (a-e), the plants were grown in a sterilized soil: quartz sand: vermiculite mixture (1:1:1) inoculating with (AM) or without (NM) R. intraradices under Pi deficient condition (without KH 2 PO 4 but with the addition of 1 mM KCl).

Figure 2 .
Figure 2. SlDELLA physically interacts with SlPIF4 in vivo and in vitro.(a) Yeast two-hybrid assay to confirm the interaction between SlDELLA and SlPIF4.Yeast cells were grown on SD/−Leu/−Trp (-LW) for 2 days or SD-Leu/−Trp/−Ade/-His (-LWAH) medium with 1 μM 3-AT for 5 days.(b) BiFC assay showing the interaction of SlDELLA with SlPIF4 in Nicotiana benthamiana.Full-length SlPIF4 and SlDELLA were fused to YFP n or YFP c , respectively.Scale bar = 25 μm.Images were taken using a confocal microscope.H2B-mCherry was used as a nuclear marker.(c)-(d) Split luciferase complementation assay (SLCA) showing the interaction of SlPIF4 with SlDELLA.The cLUC-SlPIF4 and nLUC-SlDELLA constructs were co-transformed into N. benthamiana leaves, and the luminescence intensity was detected after 48 h.The luminescence intensity detected in N. benthamiana leaves co-expressed cLUC and nLUC as the control.Representative photographs are shown in (c), and luciferase activity is shown in (d).Data are presented as the means of six biological replicates (±SD).Asterisks indicate a statistically significant difference from the control in the means ( * * P < 0.01; Student's t test).(e) Pull-down assay showing that MBP-tagged SlPIF4, but not MBP alone, could pull down GST-tagged SlDELLA in vitro.Recombinant MBP-SlPIF4 and GST-SlDELLA were detected with anti-MBP and anti-GST, respectively.The red arrow represents MBP-SlPIF4.

Figure 4 .
Figure 4. SlPIF4-mediated AMS is partly dependent on SLs biosynthesis, accumulation and exudation in tomato.(a) Relative contents of SLs from WT and pif4 mutant plants roots at 10 dpi.Data are presented as the means of three biological replicates (±SD).Different letters indicate significant differences by One-way ANOVA followed by post hoc Tukey test (P < 0.05).(b)-(d) The expression of SlCCD7, SlCCD8, and SlMAX1 in roots of WT and pif4 mutant plants at 10 dpi.Data are presented as the means of four biological replicates (±SD).Different letters indicate significant differences by One-way ANOVA followed by post hoc Tukey test (P < 0.05).(e) Representative images of WGA-AF488-stained roots of in WT and SlPIF4-overexpressing (PIF4#89) plants applied with or without 1 μM rac-GR24 at 20 dpi.BF, bright-field image.Merge, merged WGA-AF488 and BF image.Scale bar =100 μm.(f) The root length colonization of hyphae, arbuscules, and vesicles in the roots of WT and SlPIF4-overexpressing (PIF4#89) plants applied with or without 1 μM rac-GR24 at 20 dpi.Data are presented as the means of four biological replicates (±SD).Each replication contains 80 root segments of two plants.Different letters indicate significant differences by one-way ANOVA followed by post hoc Tukey test (P < 0.05).(g) Transcripts of AMS-marker gene BCP1 in roots of WT and SlPIF4-overexpressing (PIF4#89) plants applied with or without 1 μM rac-GR24 at 20 dpi.Data are presented as the means of four biological replicates (±SD).Different letters indicate significant differences by One-way ANOVA followed by post hoc Tukey test (P < 0.05).For (a)-(d), the plants inoculated with (AM) or without (NM) Rhizophagus intraradices were grown in a sterilized soil: quartz sand: vermiculite mixture (1:1:1) under Pi deficient condition (without KH 2 PO 4 but with the addition of 1 mM KCl).For (e)-(g), the WT and SlPIF4-overexpressing (PIF4#89) plants inoculated with R. intraradices were treated with ddH 2 O (Control) or + rac-GR24 (1 μM) three times a week.

Figure 5 .
Figure 5. SlPIF4 directly binds to promoters of SLs biosynthesis genes and inhibits its transcription.(a) EMSA assay showing the binding of SlPIF4 to the promoters of SlCCD7, SlCCD8, and SlMAX1.Excess amounts (100×, 300×, 500×) of non-labeled or mutant oligo (mut) were set as the competitors.The biotin non-labeled oligo was used as a competitor, −: absence; +: presence.(b) ChIP-qPCR showing the binding of SlPIF4 to the SlCCD7, SlCCD8, and SlMAX1 promoters containing the G/E-box motifs in vivo.HA, hemagglutinin; OE, overexpressing.Values are percentages of DNA fragments that coimmunoprecipitated with anti-HA antibodies or anti-IgG relative to the input DNA.Data are presented as the means of three biological replicates (±SD).Different letters indicate significant differences by one-way ANOVA followed by post hoc Tukey test (P < 0.05).(c) Dual-luciferase assay showing the inhibition of SlPIF4 to the SlCCD7, SlCCD8 and SlMAX1 promoters in Nicotiana benthamiana.EV: empty vector.The LUC/REN ratios of the empty vector (EV) plus promoter were set at '1'.Data are presented as the means of six biological replicates (±SD).Asterisks indicate a statistically significant difference from the control in the means ( * * P < 0.01; Student's t test).

Figure 6 .
Figure 6.SlPIF4 negatively regulates the AMS-induced phosphate uptake by inhibiting the transcription of AMS-specific PTs.(a)-(b) Total P content of WT, SlPIF4 overexpressing and pif4 mutant plants at 30 dpi.Data are presented as the means of four biological replicates (±SD).Different letters indicate significant differences by one-way ANOVA followed by post hoc Tukey test (P < 0.05).(c)-(d) Transcripts of PT4 and PT5 in roots of WT, SlPIF4 overexpressing (PIF4#89) and pif4 mutant plants at 20 dpi.Data are presented as the means of four biological replicates (±SD).Different letters indicate significant differences by one-way ANOVA followed by post hoc Tukey test (P < 0.05).For (a)-(d), the plants inoculated with (AM) or without (NM) R. intraradices were grown in a sterilized soil: quartz sand: vermiculite mixture (1:1:1) under Pi deficient condition (without KH 2 PO 4 but with the addition of 1 mM KCl).(e) EMSA assay showing the binding of SlPIF4 to the promoters of AMS-specific PTs.Excess amounts (100×, 300×, 500×) of non-labeled or mutant oligo (mut) were set as the competitors.The biotin non-labeled oligo was used as a competitor, −: absence; +: presence.(f) ChIP-qPCR showing the binding of SlPIF4 to the AMS-specific PTs promoters containing the E-box motifs in vivo.HA, hemagglutinin; OE, overexpressing.Values are percentages of DNA fragments that coimmunoprecipitated with anti-HA antibodies or anti-IgG relative to the input DNA.Data are presented as the means of three biological replicates (±SD).Different letters indicate significant differences by one-way ANOVA followed by post hoc Tukey test (P < 0.05).(g) Dual-luciferase assay showing the inhibition of SlPIF4 to the AMS-specific PTs promoters in Nicotiana benthamiana.EV: empty vector.The LUC/REN ratios of the empty vector (EV) plus promoter were set at '1'.Data are presented as the means of six biological replicates (±SD).Asterisks indicate a statistically significant difference from the control in the means ( * * P < 0.01; Student's t test).

Figure 7 .
Figure 7. SlDELLA regulates the expression of SlPIF4 target genes by reducing the protein stability and attenuating the transcriptional activity of SlPIF4.(a) Transcript of SlCCD7 in roots of WT and pro plants at 10 dpi.Data are presented as the means of four biological replicates (±SD).Different letters indicate significant differences by one-way ANOVA followed by post hoc Tukey test (P < 0.05).(b) Transcripts of AMS-specific PTs in roots of WT and pro plants at 20 dpi.Data are presented as the means of four biological replicates (±SD).Asterisks indicate a statistically significant difference from the control in the means ( * * P < 0.01; Student's t test).(c) Immunoblots showing the SlPIF4 protein level in WT and pro plants at 10 dpi.A same membrane was split into two sheets and incubated with anti-PIF4 and anti-HSP70 antibodies, respectively.Anti-HSP70 was used as a loading control for the western blot analysis.Representative pictures are shown.Relative protein levels are shown on the right side of the blots.Data are presented as the means of three biological replicates (±SD).Asterisks indicate a statistically significant difference from the control in the means ( * * P < 0.01; Student's t test).(d) SlPIF4 degradation in cell-free degradation assay.Total proteins extracted from the roots of WT and pro plants inoculated with R. intraradices were incubated with or without MG132 (a 26S proteasome inhibitor) over the indicated time course, and the protein levels of SlPIF4 were detected using an anti-MBP antibody.Anti-HSP70 was used as a loading control for the western blot analysis.For (a)-(d), the plants inoculated with (AM) or without (NM) R. intraradices were grown in a sterilized soil: quartz sand: vermiculite mixture (1:1:1) under Pi deficient condition (without KH 2 PO 4 but with the addition of 1 mM KCl).(e)-(f) Dual-luciferase assay showing the regulatory effect of SlPIF4 inf luenced by SlDELLA on the promoters of SlCCD7 and AMS-specific PT4.The ratio of firef ly luciferase (LUC) and renilla luciferase (REN) of the empty vector (EV) plus promoter was set at '1'.Data are presented as the means of six biological replicates (±SD).Different letters indicate significant differences by one-way ANOVA followed by post hoc Tukey test (P < 0.05).(g)-(h) Electrophoresis mobility shift (EMSA) assay.The biotin-labeled SlCCD7 and AMS-specific PT4 oligos was used as SlPIF4-targeted DNA sequence.1×, 3×, and 5× indicated the intensity of the GST-SlDELLA protein, and 1×, 2×, and 3× indicated the intensity of the GST protein.The protein purified from the empty vector was used as a negative control.