Lettuce (Lactuca sativa L.) of GB-30, a variety, which bolts easily and comes from seeding plant protection Co., LTD in Inner Mongolia Bameng Fidelity, were conserved in our laboratory. Seeds were sown in a mixture of sand, soil, and peat (volume ratio of 1:1:1), and grown in the growth chamber under normal conditions (20°C/13°C (day/night), 14 light and 10 h darkness, and 60% relative humidity). Standard pest control and water management practices were used. The seedlings were transplanted into 6 cm × 8 cm pots at the trefoil stage. When the plants developed the sixth true leaf, they were divided into two groups. The control group continued growing under the normal growth conditions as described above. The other group was moved to another growth chamber and treated with high temperatures of 33°C and 25°C during the day and night, respectively. The other environmental factors were unchanged. After high-temperature treatment for 0, 2, 4, 6, 8, 10, 12, 24 h and 0, 8, 16, 24 days, shoot apex of lettuce were harvested for RNA extractions. Root, stem (stalk), stem apex, leaf, flower, fruit, and seed tissue samples from lettuce grown under the normal temperature conditions were also collected. All samples were immediately frozen in liquid nitrogen and stored at −80 ^° C for further measurements on gene expression analysis. Every five plants were used as one replicate for three replicates.

Total RNA was extracted from specific tissues using the Quick RNA isolation Kit (Huayueyang Biotech, Beijing, China), following the user manual from the manufacturer. NanoDrop 2000 Spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, United States) and 1% agarose gel electrophoresis were used to determine RNA quality and concentration. The extracted RNA was digested with EasyScript® One-Step gDNA Removal and cDNA Synthesis SuperMix (Transgen, Beijing, China) was used to synthesize cDNA.

The open reading frame of the cDNA was amplified with gene-specific primers of LsARF3-F/R (Supplementary Table1) based on LsARF3 (AccessionXM_023880987.2) available at NCBI. The whole cDNA of LsARF3 was then amplified with mixed super fidelity polymerase (Vazyme Biotech Co., Ltd.), under thermal cycling conditions of 95°C denaturation for 3 min, followed by 35 cycles of 95°C for 15 s, 60°C for 15 s, 72°C for 2 min, and a final extension of 72°C for 5 min. A single band of expected size 1794 bp was purified from gel and sub-cloned into pEASY vector (TransGen Biotech, Beijing, China).

A homology search of sequences was carried out through BLAST in NCBI, and multiple alignments were analyzed using DNAMAN 9.0. A phylogenetic tree was constructed using the neighbor-joining method (NJ) with MEGA 7.0 software.

The total RNA of fresh lettuce samples was extracted by quick RNA Isolation Kit (Huayueyang, China), and the cDNA was obtained by transcript one step gDNA removal and cDNA synthesis Supermix reagent (Transgen, China). Then RR430b TB green fast qPCR mix (Takara, Japan) preparation system, and the fluorescence data were obtained by Bole bio rad CFX96 real-time fluorescence quantitative PCR instrument. 18 s rRNA (HM047292.1) was selected as the internal reference gene, and the internal reference gene and samples were biologically repeated three times and technically repeated three times (Supplementary Table1). The relative gene expression is based on the 2^{−Δ ΔCt} method.

pRI101 Carrying GFP label after transformation was selected, and pRI101-LsARF3 vector was constructed through homologous recombination. The 35S promoter was used for directing the expression of the fusion gene. Subcellular localization of LsARF3 by fusion with a green fluorescent protein (GFP) in the C-terminal region. It was injected into tobacco with five leaves by Agrobacterium infection. After dark culture for 2 days, it was observed and photographed under a laser confocal microscope.

Using LsARF3 clone plasmid as a template, primer LsARF3-OE-F/R (Supplementary Table1), amplify the target band and add a homologous arm. NdeI restriction endonuclease single enzyme digested the expression vector pRI101-GFP to linearize the plasmid. Clonexpress II one-step cloning kit homologous recombinant PCR product and linearized plasmid were transferred to DH5α competent cell. After colony PCR, the positive clones were sent to sequencing, and the correct sequenced plasmid was transferred into Agrobacterium GV3101. The sequencing primer was pRI101-F/R (Supplementary Table1).

According to the principle of adjacent motif recognition (PAM) of the original spacer sequence of the CRISPR/Cas9 system and the design principle of guide RNAs (gRNAs) of gene knockout target site, CRISPR gRNA is used to design the website¹ and NCBI-Blast potential knockout site detection, and select the two most suitable gRNAs. Four primers were designed according to two gRNA sequences. CRISPR/Cas9 plasmid pKSE401-DT1T2 carries two tandem repeat gRNA expression boxes and Cas9 enzyme. Using pCBC-DT1T2 plasmid as a template, four primers DT1-BsF + DT1-F0/DT1-BsR + DT1-R0 (Supplementary Table1) were amplified to obtain a tandem repeat expression box fragment carrying gRNA1 and gRNA2 with homologous arm. Bsai-HFv2 restriction enzyme linearized pKSE401 vector, ClonExpress II One Step Cloning Kit homologous recombinant PCR product and linearized plasmid, and transferred to DH5α competent cell. After colony PCR, the positive clones were sent for sequencing. The plasmids in which both two gRNAs were successfully constructed were transferred into Agrobacterium GV3101. The colony PCR and sequencing primers were U626-IDF/U629-IDR (Supplementary Table1).

To generate the LsARF3 transgenic lettuce, Agrobacterium-mediated leaf disk transformation of lettuce. Cotyledons of 4-day-old seedlings were pre-cultured in an MS medium supplemented with 0.5 mg/l 6-BA, and 0.05 mg/l NAA for 1 day. The cotyledons were then immersed for 5 min in an Agrobacterium culture (OD600 = 0.1–0.3) with the liquid MS medium. The cotyledons were dried and then co-cultured for 30 h in an MS medium containing 0.5 mg/l Kinetin, and 0.05 mg/l NAA under dark conditions. After co-cultivation, the cotyledons were transformed into a regeneration and selection MS medium (0.5 mg/l 6-BA, 0.05 mg/l NAA, 300 mg/l Timentin, 100 mg/l Kanamycin). After every week, the medium was refreshed until the shooting. Two-to-three-centimeter newly developed shoots were then transformed into a rooting MS medium (0.05 mg/l NAA and 300 mg/l Timentin) until true roots developed. Once the regenerated seedlings had enough and strong roots, they were transplanted to nutrient soil and continued to grow in the greenhouse.

The identification method of overexpression positive seedlings is as follows: T₀ takes the leaves of transgenic seedlings to extract DNA, takes DNA as template, pRI101-F/R as identification primer, amplifies bands, and those with clear bands are positive seedlings. After the seeds of T₀ generation positive seedlings were harvested and sown, the leaves of T₁ generation transgenic seedlings were taken to extract RNA and reverse transcribed to synthesize cDNA. The cDNA was used as the template and GFP-F/R (Supplementary Table1) was used as the identification primer to amplify the bands. If the expression increased significantly, it was a positive seedling. Through the above methods, we have obtained three T₁ lines of overexpressed LsARF3. The identification method of gene editing positive seedlings is as follows: take the leaf DNA of transgenic seedlings as the template, U626-IDF/U629-IDR and Cas9-F/R as the primers, and those with two bright bands in amplification are transgenic positive seedlings. On the genomic DNA sequence of leaf lettuce, the primers CRISPR-DNA-F/R (Supplementary Table1) were designed at about 200 bp before and after the two gRNAs to amplify the target sequence of positive seedlings, and the PCR products were sent to sequencing to identify the editing type. Unedited and homozygous editing types can be directly determined by the sequencing results of PCR products. If the sequencing results of PCR products are bimodal, it is necessary to connect the PCR products to the Blunt-Zero cloning vector, and send at least 10 monoclonal antibodies to each sample for sequencing. The results were consistent and could be regarded as positive seedlings. Through the above methods, three homozygous T₁ lines of LsARF3 gene editing were obtained.

The phenotypic changes of LsARF3-OE and WT plants were observed and recorded at normal temperature (NT), the stem length changes were measured every day. The lsarf3 mutants and WT plants were treated with high temperature (HT) when they had six leaves. After treatment, the changes in phenotype and stem length were observed and recorded every day. In both LsARF3-OE and lsarf3 mutants, when the stem length was significantly different compared to WT plants, flower bud differentiation of stem apexs were observed regularly by paraffin section method (Zhang et al., 2016) to determine bolting time.

When the seedlings of lsarf3 mutants and WT plants grow to six leaves, they are sprayed every other week at a concentration of 40 mg/Kg for four consecutive weeks. The plant phenotype was observed and photographed every day, the stem length was measured every 8 days, the progress of flower bud differentiation of stem apex was determined by the paraffin section method (Zhang et al., 2016), and the transcript level of LsARF3 was detected by qRT-PCR at 24 days after exogenous IAA application.

For the Yeast one-hybrid (Y1H) screening assay, the promoter of LsCO was amplified with specific primers (Supplementary Table 1). LsCO promoter was cloned into the pHIS2 vector, and the full-length coding region of LsARF3 was ligated into the pGADT7 vector, respectively. The combined pHIS2 vector was linearized and transformed to Y187 yeast strain together with an empty AD vector or pGADT7-LsARF3. The transformed yeast cells were screened on SD/−Leu-Trp-media and SD/−Leu-Trp-His + 60 mM 3AT-medium.

Dual-luciferase assay was performed according to the previous study (Ma et al., 2019). The LsCO promoter (2,000 bp) was cloned into the pGreenII 0800-LUC vector at the PstI and BamHI restriction enzyme sites. The LsARF3 coding sequence was cloned into the pGreenII 62-SK vector. For the co-expression analysis, leaves from 5- to 6-week-old tobacco plants were transformed with the recombinant plasmids; the no-effector construct was used as a negative control. The Dual-Luciferase® Reporter Assay System (Promega, Bei-jing, China) was used to analyze firefly luciferase (LUC) and Renilla luciferase (REN) activities. The LUC:REN ratio was used to determine activation or repression.

Three independent biological replicates sampled from different plants were used for each determination. Data were presented as means ± standard errors (SEs) of three replicates. Statistical analysis of the bioassays were performed with SPSS statistical software (version 19.0, SPSS Inc., Chicago, IL, United States). Student’s t-test was used to test for significance at levels of p < 0.05 (∗) or p < 0.01 (∗∗).

To explore the roles of the ARF3 transcription factor on lettuce bolting, LsARF3 was identified through a BLAST search against lettuce genome using Arabidopsps ARF3 (AtARF3) as query sequence. Based on phylogenetic relationships, we found that Lsat_1_v5_gn_4_51680 shared high sequence homology with AtARF3 proteins as well as other ARF3 reported in 11 plant species such as artichoke, populus, asparagus, and etc. Phylogenetic analysis of all these ARF3 proteins showed that they could be divided into two groups (Figure 1A). Among these ARF3s, members from lettuce, Arabidopsis, and artichoke were grouped into the first subclade, suggesting their close relationship; ARFs from poplars, cucumbers, and other plant species consist the second group. Analysis of the conserved domains of LsARF3 protein suggested that the LsARF3 had two conserved domains: B3, and Auxin_resp (Figure 1B). The B3 domain is the motif the is responsible for the binding to target DNA regions, while Auxin_resp motif is the conserved domain specific to the ARFs. Combined with the sequence alignments and phylogenetic analysis, it can be inferred that LsARF3/Lsat_1_v5_gn_4_51680 is the lettuce orthologs of the AtARF3, which might encode the ARF3 transcription factors.

The subcellular localization results suggest that LsARF3 is a nuclear protein (Figure 1C), which is consistent with its roles as transcription factor. In the analysis of tissue/organ-specific expression, LsARF3 had a relatively higher level of expression at lettuce root, stem apex, and seed (Figure 1D), indicating that its functions might be associated with the meristem of the stem apex and root apex. After high-temperature treatment, the expression level of LsARF3 gene decreased in the early stage and increased in late stage with significant difference within 24 h (Figure 1E). Furthermore, within 24 days of high-temperature treatment, the LsARF3 gene in the stem apex was significantly higher, compared to the control group, and the increasing trend was more significant with the extended treatment time (Figure 1F), further suggesting that LsARF3 expression was induced by high temperature. Since our previous study showed that high temperature promotes the bolting of lettuce (Hao et al., 2018), It is reasonable that LsARF3 might be involved in the regulation of accelerated bolting under high temperature.

To further verify the LsARF3 function, the CRISPR/Cas9 system was used to knock out LsARF3 in lettuce. Two single-guide RNAs were designed for knockout of the LsARF3 genes, and they were precisely targeted to the first or the second exons of LsARF3. Twelve T₀ transgenic lines were identified by PCR using Cas9 gene-specific primers. Sequencing of LsARF3 gene of T₁ plants identified three lines of homozygous mutants, lsarf3-4, lsarf3-6, lsarf3-9, for phenotypic analysis. As shown in the Figure 2A, the lsarf3-4 carries a C to G mutation at the first exon, the lsarf3-6 harbors a 82-bp deletion at the first exon, and lsarf3-9 brings insertions and deletions in the second exon. All those changes in the amino acid sequence could result in the frameshift mutations in the LsARF3 proteins. As shown in the Figure 2A, three lines of lsarf3 mutants delayed bolting upon high temperatures. After grown in normal conditions, WT lettuce started anthesis at 161 days post germination (DPG), while lsarf3 mutants did not finished the floral transition yet (Figure 2A), suggesting that LsARF3 might play vital role in lettuce bolting and flowering. In previous studies, we have demonstrated that the GB-30 lettuce initiates stem elongation after 8-day heat treatments (Hao et al., 2018). In this study, the stem lengths of lsarf3 mutants were significant lower than that of WT from 16 days after high-temperature treatment (Figure 2B). Microscopic observation suggested that lsarf3 mutants entered the initial stage of flower bud differentiation after high-temperature treatment for 32 days (Figure 2C), suggesting that lsarf3 mutants had started bolting approximately after 32-day heat treatments. Measurements of bolting time, budding time, and anthesis time of the first flower indicated that lsarf3 mutants were about 20, 32, and 48 days later than WT plants, respectively (Figures 2D–F). At the same time, we observed the budding time of mutant plants at normal temperature, and found that it began to bud about 161 days, about 20 days later than WT (Figure 2A). Together, the results indicated that absence of functional LsARF3 postponed floral transition in the knockout lines, resulting in delayed bolting and flowering under normal or heat conditions.

To further investigate that the LsARF3 roles in lettuce bolting, we generated the LsARF3-overexpressing (OE) transgenic lettuce plants. The LsARF3-OE plants displayed accelerated bolting and flowering (Figure 3A). The three OE lines (OE-1, OE-7, OE-6) revealed that the expression levels of the LsARF3 gene were nearly 4 times, 8 times, and 16 times higher than that of wild-type plants (WT), respectively (Figure 3B). Microscopic analysis showed that LsARF3-OE lines began the initiation stage of flower bud differentiation at 60 days, whereas the growth cone of WT control was prominent and wrapped in the middle of the leaf primordium that showed a semicircular shape, indicating they were still in the vegetative growth stages (Figure 3C). In addition, the stem length of LsARF3-OE lines was higher than that of WT with a significant or extremely significant difference from growing 40 days (Figure 3D). From phenotype and microscopic observation, it also supported the observations that the LsARF3-OE lines had entered the bolting stage after growing for 60 days. Similarly, the bolting time, budding time, and first flower anthesis time of LsARF3-OE lines were 48, 33, and 40 days earlier than that of WT, respectively (Figures 3E–G), reflecting that the LsARF3 overexpression could accelerate bolting and flowering in lettuce.

Our previous studies have shown that exogenous IAA promoted bolting, and it occurred begin 8 days after exogenous application (Wang et al., 2022b). To explore whether LsARF3 was involved in auxin-induced bolting of lettuce, the exogenous IAA was applicated to WT and lsarf3 mutants. After the application of exogenous IAA, the plants showed obvious differences (Figure 4A). It was found that the expression levels of LsARF3 in lsarf3 mutants was extremely significant lower than that of the WT(Figure 4B). Miscroscopy experiments demonstrated that, after exogenous IAA application for 24 days, WT plants entered the stage of flower bud differentiation, but knockout mutants were insensitive to auxin applications (Figure 4C). The stem of WT lettuce elongated rapidly after IAA application, while it was not observed in the lsarf3 mutants. After 24 days after IAA application, the stem lengths of lsarf3 mutants were shorter by 58.3% to that of the WT (Figure 4D). Thus, it could be concluded that WT plants started bolting on 24th Day after IAA application, but at that time, lsarf3 mutants were still in the vegetative stage. Afterward, the stems become visible and elongated above the rosette leaves in WT plants; in contrast, the stem of lsarf3 mutants is still embedded in the rosette leaves of transgenic lettuces. It took 90 days to entered budding stage for WT lettuces, whereas lsarf3 lines kept at vegetative stages (Figure 4A). In other words, the defects in bolting time of lsarf3 mutants could not fixed by exogenous IAA applications (Figures 2A,E, 4A). Together, the results suggested that exogenous auxin accelerated bolting in WT, but it cannot recover the delayed bolting phenotype of lsarf3 mutants.

To explore the molecular mechanism underlying LsARF3-mediated bolting upon thermal treatments, the expression profiles of selected flowering-related genes in LsARF3-OE and mutant lines were analyzed in the shoot apex samples from three lsarf3 mutant lines and wild type plants. The results showed that the expression levels of LsCO, LsFT, LsLFY, LsGI, and LsSOC1 were significantly upregulated in these three LsARF3-OE lines, while their expression levels of LsFLC and LsFLM were much lower compared to that of WT lettuce plants (Figure 5A). Conversely, the expression of LsCO, LsFT, LsLFY, LsGI, and LsSOC1 were attenuated in lsarf3 mutants, and accordingly, the transcript levels of LsFLC and LsFLM were significantly increased compared to WT plants (Figure 5B). The results suggested that LsCO, LsFT, LsLFY, LsGI, and LsSOC1 exhibited similar expression pattern in the LsARF3-OE and knockout lines. Therefore, it is possible that those flowering-related genes could be the potential downstream genes of LsARF3, thereby controlling the floral transition and the bolting of lettuce.

Given that LsARF3 could co-expressed with a number of flowering-related genes such as LsCO, LsFT, and LsGI, we expect that the transcription of these genes might be directly regulated by LsARF3. Based on the analysis of the 2 kb promoter sequences of LsCO, LsFT, and LsGI, an AuxRE cis-elements in the LsCO promoter was identified, suggesting LsCO was a possible LsARF3 direct target (Figure 6A). To test the hypothesis, a yeast one-hybrid (Y1H) assay was conducted to determine whether LsARF3 can bind directly to the LsCO promoter. As shown in Figure 6B, the Y1H results showed that the yeast cells containing the bait vector combined with the promoter region of LsCO could grow on the SD/−Leu-Trp-His medium supplemented with 60 mM 3AT, when co-transformed with pGADT7-LsARF3 construct. However, the yeast cells transformed with the empty pGADT7 vector did not grow on the same selective medium. Hence, the results suggested that LsARF3 could in vivo bind to the promoter of LsCO and activate the reporter gene in yeast. This finding was further confirmed by a dual luciferase assay in tobacco leaves with LsCO promoter activity being monitored with a ProLsCO::LUC reporter construct. Overexpression of LsARF3 significantly increased the activity of LsCO promoter (Figure 6C), further indicating that LsARF3 could promote LsCO expression via directly binding to its promoter.

The transition from vegetative to reproductive growth is critical to plant development because the proper timing of flowering ensures seed production required for the survival of species (Takeno, 2016). Given that auxin regulates diverse developmental processes from phototropism, root/shoot tropisms, to apical dominance in plants (Enders and Strader, 2015), it is perhaps not surprising that auxin is involved in the transition to bolting and flowering in lettuce. It was demonstrated that, under normal growth conditions, the IAA content in lettuce stems was significantly increased upon bolting, and exogenous application of IAA could promote bolting of some lettuce cultivars after 4-day treatments (Wang et al., 2022b). Recently, another study also showed that overexpression of LsRGL1, encoding a functional DELLA repressor, resulted in the reduction of IAA biosynthesis, while auxin feeding restored the bolting time of LsRGL1 overexpression lines (Wang et al., 2022a). Their observations suggested that auxin was closely involved in the lettuce bolting, but whether auxin is involved in the accelerated bolting under thermal stresses remains to be examined. In our analysis, we found that IAA treatments indeed promoted the elongation of inflorescence stems under heat stresses. Accordingly, we observed that the IAA content in lettuce stems kept increasing with the progression of heat treatments (Hao et al., 2018). Therefore, our results demonstrated that auxin, at least in part, participated the thermal stress-induced bolting in lettuce, linking abiotic auxin signaling pathways with inflorescence development under abiotic stresses.

ARF3 and its orthologues are found in various angiosperms, including monocots and dicots, but so far, its roles in plant development were mainly obtained from the model plant Arabidopsis (Hobecker et al., 2017; Brackmann et al., 2018; Zhang et al., 2018). To our knowledge, the functions of ARF3 orthologues in monocots such as lettuce remain largely intangible. Previous extensive evidence demonstrated that ARF3/ETTIN was involved in the regulation of organ polarity determination and pattering, gynoecium morphogenesis, self-incompatibility, flower development, and other auxin-responsive behaviors in Arabidopsis (Nemhauser et al., 2000; Liu et al., 2014b; Zhang et al., 2018). For example, ettin mutants exhibited severe defects in leaf polarity specification and gynoecium patterning (Nemhauser et al., 2000). Zhang and his collogues also reported that ARF3 regulates floral meristem determinacy by directly inhibiting the expression of ISOPENTENYLTRANSFERASE responsible for the rate-limiting step of cytokinin biosynthesis (Zhang et al., 2018). A recent study suggested that Medicago truncatula plants overexpressing MtARF3 exhibited curling leaf margins with more serrations which resemble the phenotypes of the palm1 mutant to a large extent; further experiments confirmed that MtARF3 restricted spatiotemporal expression of PALMATE-LIKE PENTAFOLIATA1 (PALM1) through selective interaction with the promoter of PALM1 (Peng et al., 2017). Expression of OsARF3 construct with altered tasi-RNA targeting sites resulted in dwarf and tufty phenotypes, irregular phyllotaxis, with rolled-up or thread-like leaves in transgenic rice lines, indicating its function and regulation are conserved in higher plants (Song et al., 2012). Taken at face value, most studies supported that ARF3s serves as transcriptional repressors in the auxin-regulated pathways. Nevertheless, our results indicated that overexpression of LsARF3 promoted plant growth and early bolting and flowering under normal or heat-stress conditions. Surprisingly, no developmental defects in leaf or flower organs were observed in the LsARF3 overexpression or knockout lines, despite the slightly reduced flower fertility, which was not reported, to our knowledge, in other ARF3 transgenic plants yet. Furthermore, the expressions of several floral genes including LsCO, LsFT, LsSOC1, and LsLFY were positively correlated with that of LsARF3 in the overexpression or loss-of-function lines, implying their transcription could be induced by LsARF3 during the lettuce bolting process. Thus, it is likely that LsARF3 was able to activate the expression of those floral integrator genes at high temperatures. Importantly, it is also worth noting that LsARF3 might function as a transcriptional activator in lettuce, whereas other ARF3 homologs reported in Arabidopsis, legume, or rice, act as transcriptional repressors in the auxin signaling pathways. The discrepancy between LsARF3 and its orthologues in other plant species implies that it might participate in some auxin signaling pathways in ways different from its orthologues in other plant species.

In the last decade, several genes or quantitative trait loci (QTLs) that affect the timing of floral transition have been reported to be associated with heat-induced bolting and flowering (Hartman et al., 2013; Han et al., 2021). In Arabidopsis, FT, SOC1, and CO are major floral integrators because diverse floral induction pathways are mainly converged upon these developmental regulators (Yoo et al., 2005; Valverde, 2011; Pin and Nilsson, 2012). Accordingly, their lettuce orthologues of these floral genes were functionally characterized to unveil their potential roles in floral transition, in particular under heat stresses (Chen et al., 2017; Liu et al., 2018). Overexpression of LsFT and LsSOC1 can promote flowering, whereas their knockdown by RNA interference in lettuce resulted in significantly delayed bolting and insensitivity to high temperatures (Fukuda et al., 2011; Chen et al., 2017). Although lettuce COs have not been characterized, at least one CO-like gene was found to be located within flowering QTL, qFLT5.3, which implies its possible contributions to implicated in bolting initiation in lettuce (Seki et al., 2020). In the present study, AuxRE element was identified in the promoter region of LsCO gene, suggesting that CO expression might be regulated by ARF transcription factors. Indeed, LsCO was constitutively activated in transgenic lettuce overexpressing the LsARF3; by contrast, deletion of LsARF3 dramatically attenuated the expression level in lettuce, implying that LsCO might be the target of LsARF3. Consequently, upregulation of LsARF3 promoted lettuce bolting, while its knockout lines failed to accelerate bolting despite auxin feeding plus thermal treatments. Furthermore, yeast one-hybridization analysis verified that LsARF3 could physically interact with AuxRE cis-elements of the LsCO promoter. Therefore, our findings reinforced the idea that LsARF3 probably directly operated the transcription of LsCO under flowering-inducible conditions. However, the question of whether LsARF3 directly activates the expression of LsFT and LsSOC1 to facilitate bolting will be examined in future experiments.

Based on our experiments, one possible scenario for interpreting lettuce early bolting upon heat stress is to postulate an auxin-ARF3 mediated signaling pathway: thermal treatment could activate the auxin biosynthesis and/or transport in inflorescence stem, resulting in the increase of local auxin levels and subsequent LsARF3 transcription. Subsequently, LsARF3 might directly activate the transcription of LsCO and indirectly induce the expression of LsFT and LsSOC1 in an auxin-dependent manner, eventually leading to the accelerated bolting and flowering in lettuce. Overall, our study provides new insights into lettuce bolting in the context of ARFs signaling and heat response. However, we cannot exclude the possibility that other ARFs might be also involved in lettuce floral transition, and further experiments are still needed to investigate the functions of ARF genes in this important vegetable.