Introduction

Roots are vital for plants to acquire nutrients and water and, thus, for plant growth, survival, and development (Petricka et al. 2012; Vidoz et al. 2010). The study of root development is of great significance to the adaptation of plants to environmental changes and to ensure crop yield under low soil fertility (Kell 2011). The root system is divided into embryo roots and ARs. An AR is defined as a root that forms on the non-root tissue of a plant. ARs share developmental characteristics with lateral roots and have strong plasticity (Verstraeten et al. 2014). They are the post-embryonic roots that generally form on the hypocotyl, and their primordia initiate from pericycle cells (Correa et al. 2012; Falasca and Altamura 2003). ARs enlarge the root system of plants and are widely used in clonal propagation, including growing plant cuttings and for tissue cultures (Ahkami et al. 2009), and they also improve the plant’s stress resistance (Vidoz et al. 2010). The study of ARs is important for agriculture, horticulture, and forestry (Puri and Thompson 2003; Yamauchi et al. 2014).

AR formation is influenced by multiple factors. Wounding, flooding, light, and plant hormones can all induce AR formation (De Klerk et al. 1999; Vidoz et al. 2010). Auxin plays a central role in AR formation. The report indicates that the excised leaves and stem segments in Arabidopsis only rooted after short IBA (auxin analogs) treatment times, which suggests auxin stimulation is a central factor for adventitious rooting (Ludwig-Muller et al. 2005). When free and conjugated IAA (the main endogenous auxin) accumulated in the stem base of the aerial roots (aer) mutant of tomato, a profuse and precocious formation of AR primordia was observed, which suggests auxin induced AR primordial initiation (Mignolli et al. 2017). Silencing of PIN1 in rice plants, which acts as an auxin efflux carrier and transport auxin from the stem apex to the root, led to inhibition of AR emergence and development (Xu et al. 2005). Thus, auxin is necessary for AR formation, and acts as an inducer of AR primordium formation. AR formation is also controlled by related genes. For example, the auxin-response factors ARF7 and ARF19, which are involved in auxin signaling, positively regulate AR formation (Gutierrez et al. 2009); the gene adventitious rootless1 (ARL1), an auxin- and ethylene-responsive gene, whose protein contains a lateral organ boundaries (LOB) domain, is characteristically expressed in lateral and AR primordia in rice and promotes AR formation (Liu et al. 2005); OsAGAP, which encodes an ARF (ADP-ribosylation factor)-GTPase-activating protein, regulates ARs formation by interfering auxin influx in rice (Zhuang et al. 2005, 2006); the gene OsCKI1 encoding a casein kinase I (CKI) in rice is involved in auxin metabolism or that it may affect auxin levels, and promoting AR formation (Liu et al. 2003). However, these related genes were relating to auxin. In addition, AR formation is also regulated by other plant hormones. However, auxin is a central player in the hormone cross-talk that controls adventitious rooting (Pacurar et al. 2014). For example, ethylene is a plant hormone that promotes AR formation by affecting auxin transport (Negi et al. 2010). Thus, auxin determines the formation of AR.

MADS-box transcription factors, such as those related to the development of floral organs, fruits, leaves, and stems have been explored for a long time and have been identified as being involved in the development of plants (Xuhu et al. 2017; Honma and Goto 2001; Kater et al. 2006; Koo et al. 2010; Pelaz et al. 2000; Vrebalov et al. 2002; Zhang and Forde 1998). Recently, MADS-box genes have been reported to regulate lateral root development (Chunyan et al. 2015; Liu et al. 2015; Sun et al. 2017). However, studies on the MADS-box gene in the regulation of AR formation have scarcely been reported. Here, we report a novel gene named SlMADS83 that encodes a member of the tomato MADS-box transcription factors, belonging to a SlMBP9 group, which functions in root development and auxin regulation (Li et al. 2019). Silencing of SlMADS83 enhanced AR formation, and we explored how the SlMADS83 gene negatively regulated ARs, as described in this report.

Materials and Methods

Plant Materials and Growth Conditions

Seeds of wild-type and transgenic tomato (Solanum lycopersicum Mill. Cv. Ailsa Craig, AC++) plants were sterilized and cultured in 28 °C at 100 rpm for 2 days in the dark. All seeds were sterilized by incubation for 2 min in 75% ethanol, then for 10 min in 10% bleach, and then washed with sterile water. After seeds germinated, they were transferred to Murashige and Skoog (MS) medium (Murashige and Skoog 1962) and cultured in a greenhouse. All the seedlings were grown in a greenhouse with a 16-h-day (26 °C) and 8-h-night (18 °C) cycle, and 70% humidity and 250 μM m−2 s−1 light intensity. Flowers were tagged at anthesis, and the fruits of immature green and mature green were defined as 28 and 35 days post-anthesis (DPA), respectively. The breaker fruits were defined as the fruits’ color changed from green to yellow, and the breaker fruits at 4 days after breaking and 7 days after breaking were also labeled. All the samples were collected at the same time and immediately frozen in liquid nitrogen and stored at − 80 °C until further use.

Total RNA Extraction and Quantitative Real-Time PCR Analysis

Total RNA of tomato was extracted using the RNAiso plus kit (Takara) according to the manufacturer’s instructions. The total RNA (2 μl) was used to synthesize first-strand cDNA using M-MLV reverse transcriptase (Takara) with a tailed oligo-d(T20) primer (Supplementary Table S1). The cDNA was used to clone the specific DNA fragment of SlMADS83 (GenBank accession number: XM_019211854, SGN number: Solyc01g106170) for construction of the RNAi vector and for quantitative real-time PCR (qPCR).

qPCRs were performed using the SYBR PremixEx Taq II Kit (Takara) in a 10-μl total sample volume (5.0 μl of 2 × SYBR PremixEx Taq, 0.5 μl of primers, 1.0 μl of cDNA, and distilled deionized water to a final volume of 10 μl) on the CFX96™ Real-Time System (Bio-Rad). The SlCAC gene (Exposito-Rodriguez et al. 2008) was used as an internal standard. No-template and no-reverse transcription controls were performed for each gene analysis. Three biological repeats were performed in the analyses of gene expression. All primers used in this study are listed in Supplementary Table S1.

Construction of the SlMADS83 RNAi Vector and Plant Transformation

A 480-bp specific DNA fragment of SlMADS83 was amplified with SlMADS83i-F and SlMADS83i-R primers tailed with HindIII/KpnI and XhoI/XbaI restriction sites at the 5′ end, respectively. The purified fragment was digested and linked into the PHANNIBAL (Dong et al. 2013) plasmid at the HindIII/XbaI restriction site in the sense orientation and at the KpnI/XhoI restriction site in the antisense orientation. The double-stranded RNAi-expression unit was digested with XbaI and SacI and linked into the plant binary vector pBIN19. The pBIN19 vector carries the CaMV (cauliflower mosaic virus) 35S promoter, the SlMADS83 fragment in the sense orientation, a PDK intron, the SlMADS83 fragment in the antisense orientation, and an OCS terminator. Finally, the plant binary vector pBIN19 was transferred into the Agrobacterium LBA4404 strain, and then the LBA4404 strain was transferred into wild-type tomato cotyledon explants through the Agrobacterium-mediated transformation method (Guoping et al. 2004).

Phylogenetic Analyses and Gene Expression Prediction

A neighbor-joining tree was constructed using MEGA 5.2 software with a bootstrap analysis of 1000 replicates. All of the proteins used in neighbor-joining tree were acquired from the NCBI database. SlMADS83 gene expression prediction was performed using Tomato eFP Browser (http://bar.utoronto.ca/efp_tomato/cgi-bin/efpWeb.cgi).

Inhibitors Treatment and Chemical Solutions

Germinated wild and transgenic seeds were cultured on medium with different doses of 5-(4-chlorophenyl)-4H-1,2,4–triazole-3–thiol (Yucasin, Sigma, US) (0, 10, 20 and 100 μM) and l-Kynurenine (Kyn, APExBIO, US) (0, 1, 10 and 100 μM), respectively. Nine days later, the number of ARs was counted. Yucasin and Kyn were dissolved in DMSO (Nishimura et al. 2014).

Quantification of ARs

All of the seedlings for the quantification of ARs were grown on the medium for 9 days, including the seedlings with treatment of silver nitrate. The seedlings for adventitious rooting were grown on medium for 8 days. Then the top of these seedlings was removed at a position 5 mm from the stem apex and the roots at a position 5 mm above the hypocotyl/root junction. Finally, the excised hypocotyls were transferred to a new medium for rooting for 9 days, and then the ARs were quantified.

Anatomical Analysis of the Segment of Hypocotyl Base

The segments of the hypocotyl base from 5 mm above the hypocotyl/root junction to the hypocotyl/root junction of the 5-day-old seedlings were collected and immediately fixed in FAA liquid (38% formaldehyde/acetic acid/70% ethanol = 1:1:18, by volume). Subsequently, the samples underwent dehydration, fixation, sectioning, and dewaxing. Longitudinal sections were made by cutting the samples along the middle of the segment. The entire longitudinal sections were visualized under a microscope and photographed.

Application of 1-Aminocyclopropane-1-Carboxylate (ACC) and Silver Nitrate

The germinated wild-type seeds were respectively sowed on medium containing 0, 1, 5, and 10 μM ACC. Subsequently, the entire stem (containing the shoot apex) of wild-type seedlings was collected for total RNA extraction and qPCR analysis after 9 days. Both transgenic and wild-type seeds were grown on medium containing 50 μM silver nitrate (AgNO3) for 9 days, and then the number of ARs was counted and the entire stems (containing the stem apex) were collected for qPCR.

Quantification of Free Indole-3-Acetic Acid (IAA) in Hypocotyl Base

The samples for free IAA quantification were the segments of the 9-day seedlings of the three Ri lines (Ri-4, Ri-5, and Ri-9) and three wild type excised from 5 mm above the hypocotyl/root junction to the hypocotyl/root junction where ARs formed. Free IAA was extracted using an IAA kit (IAA-4-C Comin Biotechnology Co., Ltd., China). Finally, the free IAA content was quantified by high-performance liquid chromatography.

Results

Analysis of the Expression Pattern of SlMADS83 in Wild-Type Tomato Plants

In this study, we cloned and isolated a new tomato MADS-box transcription factor gene, SlMADS83, whose sequence was searched from the NCBI database. Phylogenetic tree analysis showed that SlMADS83 belongs to a SlMBP9 clade, which includes the other two MADS-box proteins, AtAGL21 and AtAGL17 (Fig. S1a). All of the proteins of the SlMBP9 clade play important roles in root development (Li et al. 2019), and it is speculated that SlMADS83 may play a role in root development. To explore the putative function of SlMADS83 in tomato development, qPCR was performed to analyze its transcript accumulation in different tissues and organs. SlMADS83 transcripts were detectable at very low levels in senescent leaves and ripe fruits, but its transcripts accumulated mostly in root tissue, sepals, and floral organs (Fig. 1). This result was consistent with our prediction expression profile of the SlMADS83 gene using Tomato eFP browser (Fig. S1b, c). Transcripts also accumulated in stems, young leaves, mature leaves, and immature green fruits. Notably, during leaf senescence and fruit ripening, the SlMAD83 transcript level displayed a decreasing trend (Fig. 1). Ethylene acts as a positive regulator of leaf senescence and fruit ripening, and ethylene biosynthesis improves with the processes of leaf senescence and fruit ripening (Aharoni and Lieberman 1979; Alexander and Grierson 2002; Barry and Giovannoni 2007; Bleecker and Patterson 1997; John et al. 2010; Picton et al. 1993). During leaf senescence and fruit ripening, the expression of SlMADS83 showed an opposite trend to ethylene biosynthesis. We thus inferred that SlMADS83 may mediate the regulation of ethylene.

Fig. 1
figure 1

Expression pattern analysis of SlMADS83 in wild-type tomato plants. RT, root; ST, stem; YL, young leaf; ML, mature leaf; SL, senescent leaf; SE, sepal; F, flower; IMG, immature green fruit; MG, mature green fruit; B, breaker fruit; B4, breaker fruit after 4 days; B7, breaker fruit after 7 days

Full size image

Anatomical and Macroscopic Analysis of Phenotype of Enhanced ARs

We selected three RNAi transgenic tomato lines (Ri-4, Ri-5, and Ri-9) with low transcript accumulation in roots tissue for further exploration of SlMADS83 function (Fig. 2a). Seeds of transgenic lines and wild-type tomato were grown on the MS medium for 9 days. A phenotype of enhanced ARs on the basal region from 5 mm above the hypocotyl/root junction to the hypocotyl/root junction was observed in the RNAi transgenic lines (Fig. 2b). Statistical analysis showed that the number of ARs was enhanced by 2.5- to 3.7-fold in the 9-day transgenic lines compared to the wild-type plants (Fig. 2c). Additionally, we anatomically analyzed the segment of the hypocotyl base (from 5 mm above the hypocotyl/root junction to the hypocotyl/root junction) of 5-day-old seedlings, which showed no emergence of macroscopic ARs (Fig. 2d). At ×40 magnification, several early AR primordia were found in the hypocotyl base segment of the transgenic line (Fig. 2f), whereas no early AR primordium had initiated in the wild type (Fig. 2e).

Fig. 2
figure 2

Anatomical and macroscopic analysis of phenotypes. a Expression level of SlMADS83 in the root tissue of transgenic lines and wild type of 9-day-old seedlings. b Phenotype of enhanced ARs of 9-day-old seedlings. Orange arrows indicate ARs in the basal region of the hypocotyl. c Statistical analysis of the number of ARs of 9-day-old seedlings. Means ± SE with n > 20 per group. d Segments for longitudinal sections excised at a position 5 mm above the hypocotyl/root junction and at the site of the hypocotyl/root junction. Red boxes indicate the basal region of segments that were excised for anatomical analysis. e, f Anatomical analysis of the segment of hypocotyl base from wild-type (e) and transgenic line Ri-4 (f) of 5-day-old seedlings. Red arrows show the early adventitious root primordia. Asterisks indicate a significant difference (P < 0.05)

Full size image

Adventitious Rooting Analysis

Based on the enhanced ARs forming on the hypocotyl base, we determined the free IAA (the main endogenous auxin) content in the hypocotyl base segment (from 5 mm above the hypocotyl/root junction to the hypocotyl/root junction). The free IAA level in the hypocotyl base of the RNAi lines was elevated over twofold compared to the wild type (Fig. 3a).

Fig. 3
figure 3

Analysis of adventitious rooting and free IAA content. a Quantification of free IAA in the basal region of the hypocotyl (from 5 mm above the hypocotyl/root junction to the hypocotyl/root junction) where the enhanced AR formation occurs. b The pictures above show ARs of seedlings grown on medium for 8 days and de-rooted for subsequent adventitious rooting for 9 days. The pictures in the bottom show the ARs of seedlings grown on medium for 8 days and then capitated and de-rooted (excised at a position 5 mm above the hypocotyl/root junction) for adventitious rooting for 9 days. Bars = 10 mm. c Statistical analysis of ARs in decapitated and de-rooted seedlings and de-rooted only seedlings. Means ± SE with n > 15 per group. *P < 0.05

Full size image

Next, we investigated whether silencing SlMADS83 enhanced auxin signaling to promote AR formation. After seeds were cultured on MS medium for 8 days, the top of the shoots (where auxin is produced) and the hypocotyl base (where auxin accumulated) of the seedlings were removed. The decapitated and de-rooted hypocotyls and the de-rooted seedlings were transferred to new MS medium for 9 days. Only the de-rooted seedlings produced more ARs than the wild-type plants, while the decapitated and de-rooted seedlings produced a similar number of ARs in both the transgenic and wild-type tomato (Fig. 3b, c).

Silencing of SlMADS83 Upregulated Auxin Biosynthesis and Altered Auxin Transport Capacity

Figure 4a shows a tryptophan-dependent IAA biosynthesis pathway in Arabidopsis to better understand how IAA is synthesized in stem apex. The genes of YUC (YUCCA) encode proteins of flavin monooxygenase that catalyze the rate-limiting step of the conversion of IPA to IAA, which is the main IAA biosynthesis pathway (Mashiguchi et al. 2011). The ToFZY tomato genes are the orthologs or potential orthologs of YUC genes and encode proteins with the same enzymatic activity as YUC and are involved in the IAA biosynthesis pathway (Exposito-Rodriguez et al. 2007, 2011). By qPCR, the transcripts of ToFZY1, ToFZY4, and ToFZY5 were significantly increased—among which, ToFZY5 was upregulated by fourfold to tenfold—while neither ToFZY2 nor ToFZY3 were significantly changed in the stem apex of the RNAi plants (Fig. 4b). Polar transport and distribution of auxin are dependent on the auxin efflux carrier PINs (Naramoto 2017). Tomato PIN1, which mediates basipetal polar auxin transport in the stem (Shi, et al. 2017), and PIN4 and PIN7, which are named after AtPIN4 and AtPIN7 of Arabidopsis (Pattison and Catala 2012), were unaffected by the silencing of SlMADS83 (Fig. 4c). However, the transcript level of tomato auxin transporter gene PIN8, which is an ortholog of AtPIN8 (Pattison and Catala 2012), was upregulated by 50% to 80% (Fig. 4c), which indicates that the silencing of SlMADS83 may positively alter auxin transport capacity.

Fig. 4
figure 4

Analysis of the expression levels of auxin biosynthesis and transport genes. a A tryptophan-dependent IAA biosynthesis pathway in Arabidopsis. b The transcript levels of auxin biosynthesis genes in the stem apex. c The transcript levels of auxin transport genes in stem. The samples used for the examination of gene expression were 9-day-old seedlings. Each value represents the mean ± SE of three replicates. *P < 0.05

Full size image

Silencing of SlMADS83 Increased the Transcript of Ethylene Biosynthesis Genes

As mentioned above, SlMADS83 may mediate the regulation of ethylene, and we thus examined the expression of ethylene biosynthesis genes, including ACS2 [1-aminocyclopropane-1-carboxylate (ACC) synthase 2], ACS4 (ACC synthase 4), ACO1 (ACC oxidase 1), and ACO3 (ACC oxidase 3) (Blume et al. 1997; Eum et al. 2009), which are expressed in root tissue. By qPCR analysis, the expression levels of ACO1 and ACS6 in transgenic lines were all increased by 70–130% and ACS2 was increased by 140–260% in root tissue compared to the wild type (Fig. 5).

Fig. 5
figure 5

Relative mRNA transcription levels of tomato ethylene biosynthesis-related genes. Each value represents the mean ± SE of three replicates. Asterisks indicate a significant difference (P < 0.05) between the wild-type and transgenic lines

Full size image

Treatment with 1-Aminocyclopropane-1-Carboxylate (ACC) and Silver Nitrate

Based on the significant upregulation of IAA biosynthesis and ethylene biosynthesis genes, we treated wild-type tomatoes with the ethylene precursor ACC to simulate the increased endogenous ethylene biosynthesis of RNAi plants. After the wild-type seeds were respectively cultured on MS medium with different doses (0, 1, 5, and 10 μM) of ACC for 9 days, the expression of ToFZY1, ToFZY4, and ToFZY5 was examined in the stem apex. Of these, only ToFZY5 was upregulated (onefold–tenfold) at high concentrations of ACC treatment (Fig. 6a). In addition, when transgenic and wild-type tomatoes were treated with 50 μM silver nitrate, which is an ethylene signaling antagonist, for 9 days, the expression level of ToFZY5 in RNAi plants was restored and even downregulated below the original expression level of the wild type (Fig. 6b). In contrast to untreated seedlings, the number of ARs of seedlings after 50 μM silver nitrate treatment was decreased (Fig. 6c, d).

Fig. 6
figure 6

Analysis of treatment with ACC and silver nitrate. a The expression levels of auxin biosynthesis- and transport-related genes in the stems (containing the stem apex) under different doses (0, 1, 5, and 10 μM) of ACC treatment. b Relative expression of ToFZY5 in the stem apex of seedlings treated with 50 μM silver nitrate and untreated seedlings. Each value represents the mean ± SE of three replicates. c ARs in 9-day-old seedlings treated with 50 μM silver nitrate. Bars = 10 mm. d Statistical analysis of ARs of the seedlings treated with 50 μM silver nitrate and untreated seedlings. The seedlings were 9 days old. Means ± SE with n > 20. The asterisk (*) indicates statistically significant differences between silver nitrate-treated wild-type and silenced lines as determined by a Student’s t-test. The hash (#) indicates statistically significant differences between untreated wild-type and silenced lines as determined by Student’s t-test

Full size image

Treatment with Auxin Biosynthesis Inhibitors Yucasin and Kyn

To further confirm whether SIMADS83-mediated elevated auxin biosynthesis is involved in the increase in AR formation, we cultured wild and transgenic seedlings grown on medium with different doses of the auxin biosynthesis inhibitors Yucasin (0, 10, 20, and 100 μM) and Kyn (0, 1, 10, and 100 μM). Nine days later, we counted the AR number of these seedlings. Statistical analysis suggests that under treatment with high concentration of Yucasin (100 μM), the AR numbers of both WT and transgenic seedlings were significantly decreased compared to that of the control. The wild-type tomato hardly had ARs, whereas a small number of ARs were formed in the transgenic seedlings that were similar to the wild type under normal growth conditions (Fig. 7a, e). Although the ARs of transgenic seedlings were obviously suppressed under the treatment of 20 μM Yucasin, the number of ARs was still higher than that of the untreated wild-type seedlings (Fig. 7a, e).

Fig. 7
figure 7

Inhibition treatment. a The phenotype of the 9-day-old seedings treated with different doses of Yucasin (0, 10, 20, and 100 μM) and 100 μM Kyn. b The number of ARs in the 9-day-old seedlings treated with Yucasin. Means ± SE with n > 15. *P < 0.05, **P < 0.01

Full size image

Kyn is a co-inhibitor of ethylene response and auxin biosynthesis (He et al. 2011). In contrast to the Yucasin treatment, a high concentration (100 μM) of Kyn treatment resulted in an increase in the number of ARs in both wild-type and transgenic seedlings (Fig. 7a). However, low doses of Kyn (1 and 10 μM) did not significantly reduce the number of ARs (data not shown).

Discussion

Recently, we reported a MADS-box gene SlMBP9, which suppresses lateral root formation by regulating local auxin accumulation (Li et al. 2019). In this study, a novel tomato MADS-box gene, SlMADS83, which is homologous to SlMBP9 (Fig. S1a) and may play an important role in root development, was isolated from wild-type tomato roots based on a cDNA clone. Silencing of SlMADS83 enhanced AR formation. Because the enhancement of AR formation could be used in breeding programmes to improve plant survival, we examined how the silencing of SlMADS83 positively regulates AR formation.

SlMADS83 was expressed mostly in roots, sepals, and flowers, and presented a trend of downregulation during leaf senescence and fruit ripening, which indicates that SlMADS83 may function in root regulation and ethylene-mediated progress as mentioned above. Subsequently, we constructed a SlMADS83-RNAi vector and produced SlMADS83-RNAi transgenic plants, and three RNAi transgenic lines (Ri-4, Ri-5, and Ri-9) were selected for further experiments (Fig. 2a). Based on the statistical analysis, 2.5- to 3.7-fold more ARs were produced in the region of the hypocotyl base of the transgenic plants than in that of the wild-type tomato plants (Fig. 2b, c). AR development is a normal part of plant development, but generally, only external stimulation such as wounding can induce the formation of additional ARs (Ahkami et al. 2009). Silencing of SlMADS83 resulted in an increase in ARs in the transgenic seedlings grown under normal conditions without external stimuli. Thus, we inferred that SlMADS83 is a key factor that negatively regulates AR formation.

Segments of the hypocotyl base of 5-day-old seedlings, which had no macroscopic AR emergence, were excised from a position 5 mm above the hypocotyl/root junction to the hypocotyl/root junction (Fig. 2d) and used for anatomical analysis. Several early AR primordia were initiated in the hypocotyl base of the transgenic line (Ri-4), while no AR primordia were observed in the wild-type plants (Fig. 2e, f). Auxin is an inducer of AR formation, and early AR primordium formation is strictly dependent on auxin (Correa et al. 2012). Thus, these results suggest that silencing SlMADS83 accelerates early AR primordium formation, which may be ultimately caused by the enhanced auxin or auxin signaling that we predicted.

To explore whether silencing of SlMADS83 leads to increased auxin or auxin signaling, we determined the free IAA content in the region of the hypocotyl base where the increased number of ARs formed. The free IAA content in RNAi lines was twofold greater than that in the wild type (Fig. 3a). After adventitious rooting for 9 days, the de-rooted transgenic seedlings produced more ARs than wild-type de-rooted seedlings (Fig. 3b, c). However, the AR formation in decapitated and de-rooted hypocotyls, which may be induced by residual auxin, did not significantly increase in RNAi lines. These data indicate that the silencing of SlMADS83 induced the accumulation of auxin in the hypocotyl base but did not enhance auxin signaling.

The expression levels of ToFZY1, ToFZY4, and ToFZY5 were upregulated, which suggests that the silencing of SlMADS83 may increase auxin biosynthesis in the shoot apex. The transcript of the auxin transporter gene PIN8 was also increased, which indicates that the silencing of SlMADS83 may increase auxin transport capacity in the stem (Fig. 4c). However, PIN1, the main basipetal polar auxin transporter gene (Shi et al. 2017), as well as PIN4 and PIN7, was unaffected (Fig. 4c). Although only PIN8 of the auxin transport genes was upregulated, we cannot rule out the possibility that basipetal polar auxin transport was enhanced. The aerial roots (aer) mutant of tomato showed that PIN1 and PIN4 were more expressed in shoot apices than in the hypocotyl (Mignolli et al. 2017). In our study, the samples for the detection of PINs were the entire shoots (containing the shoot apex), and thus, this may have led to the expression of PINs being apparently unchanged. The detailed pathway of polar auxin basipetal transport in the stem requires further experimental exploration.

It is well known that there is cross-talk between ethylene and auxin. Ethylene promotes auxin biosynthesis and regulates auxin transport and enhances AR formation (Barry et al. 2001; Negi et al. 2010; Swarup et al. 2007). Enhancing ethylene synthesis or signaling may elevate auxin biosynthesis. For example, the evidence proved that exogenous ACC addition improved auxin synthesis (Swarup et al. 2007). The mutant ctr1, which showed enhanced ethylene signaling, significantly increased auxin biosynthesis (Muday et al. 2012). Ethylene upregulated the tryptophan (Trp, the substrate for auxin biosynthesis) biosynthesis genes WEAK ETHYLENE INSENSITIVE2 (WEI2) and WEI7, resulted in an increase in auxin (Stepanova et al. 2005). In addition, a molecular mechanism by which ethylene promotes auxin biosynthesis is via activation of the transcription of TAA1 and its close homologs TARs (Stepanova et al. 2008), which are involved in Trp into IPA pathway of IAA biosynthesis pathway (Fig. 4a). Furthermore, ethylene increases the capacity for auxin transport by upregulating the genes of auxin carriers such as AUX1 in roots (Ruzicka et al. 2007). Ethylene also promotes auxin transport in stems (Ivanchenko et al. 2010), e.g., the mutant epi, which overproduces ethylene, significantly increased basipetal auxin transport in stems and formed more ARs (Negi et al. 2010). In contrast, auxin also elevates ethylene biosynthesis by upregulating ACS genes encoding ACC synthase (Stepanova et al. 2007). Moreover, ethylene affects root development, including root elongation, lateral roots, and AR formation. However, the combination of the description above and other reports (Muday et al. 2012) suggests that ethylene regulates root development by modulating auxin biosynthesis and transport. Mutations that impair ethylene perception (etr1), or signaling (ein2 and ein3) reduced root growth inhibition regulated by elevated ethylene (Muday et al. 2012), suggesting that ethylene-mediated auxin regulation occurred in the downstream of the ethylene pathway.

During leaf senescence and fruit ripening, the expression of SlMADS83 presented a trend of downregulation, which contrasted with the trend of ethylene biosynthesis (Fig. 1). We suspected that silencing SlMADS83 would improve ethylene biosynthesis. To determine whether the silencing of SlMADS83 mediates the progress of ethylene biosynthesis, we examined the genes involved in ethylene biosynthesis in root tissue. The results showed that ACS2, ACS6, and ACO1 were upregulated in the root tissue of transgenic plants compared to that of wild-type plants (Fig. 5), which suggests that the silencing of SlMADS83 may promote ethylene biosynthesis.

Subsequently, we examined the expression levels of ToFZY1, ToFZY4, ToFZY5, and PIN8 in wild-type tomato after ACC treatment, which simulates endogenous ethylene biosynthesis in transgenic plants, to explore whether these genes in transgenic lines were regulated by enhanced ethylene biosynthesis. Only the transcript of ToFZY5, whose level of gene upregulation was similar to that of RNAi plants grown on untreated medium, was increased (onefold to tenfold) under high concentrations of ACC (Fig. 6a). In addition, the transcript abundance of ToFZY5 in transgenic lines treated with a high concentration (50 μM) of silver nitrate (an ethylene signaling antagonist) was downregulated to below the original expression level of the wild type, which is in contrast to that in the untreated transgenic plants (Fig. 6b). Silver ions replace the copper cofactor binding site of ethylene receptors and interact with ethylene to resist the ethylene response, but silver ions also have potential off-target effects (Strader et al. 2009; Schaller and Binder 2017). However, the high concentration (50 μM) of silver nitrate still suppressed the expression of ToFZY5 in transgenic lines without making a larger impact. These results suggest that ethylene may affect not only the upstream processes of IAA biosynthesis but also the transcription of the YUC-like gene ToFZY5. The increased ToFZY5 may be regulated by enhanced ethylene biosynthesis in transgenic plants but not directly by SlMADS83. The upregulation of ToFZY1 and ToFZY4 may be regulated by another pathway via the silencing of SlMADS83. After treatment with 50 μM silver nitrate, the expression of ToFZY5 returned to the wild-type level, and it was inferred that the ethylene signaling of transgenic lines probably also returned to the wild-type level. This high-concentration silver nitrate treatment may lead to normal levels of ethylene-stimulated auxin biosynthesis and transport. Simultaneously, the number of ARs in transgenic plants was decreased compared to the number in untreated transgenic plants, but was still much higher than the original AR number of the wild-type tomato plants (Fig. 6d). These data suggest that the increase in ARs by SlMADS83 might be mainly achieved through a non-ethylene dependent pathway, and the improved transcript level of ethylene biosynthesis genes may increase ethylene biosynthesis and partially increase AR formation by regulating auxin biosynthesis and transport.

To further confirm whether SlMADS83-induced AR formation was mainly regulated by the auxin pathway, we performed Yucasin and Kyn treatments. Yucasin is a potent inhibitor of YUC protein activity that is involved in the key step of IAA biosynthesis (Fig. 4a), and Yucasin significantly reduced IAA levels. Simultaneously, high concentration ( > 25 μM) of Yucasin restored the high-auxin phenotype of YUC-overexpression seedlings (Nishimura et al. 2014). Kyn is also an inhibitor of IAA biosynthesis, through the repression of EIN3 nuclear accumulation which involved ethylene signaling and then targeted to inhibit TAA1 activity (He et al. 2011). In our results, treatment with 20 μM Yucasin decreased the AR number in transgenic lines, but the AR number is still higher than that of the untreated wild-type. This suggests that this dose of Yucasin is not enough to abolish the auxin accumulation in the hypocotyl base by SlMADS83, but it still proved that Yucasin suppressed the effects of SlMAS83-induced AR formation. Furthermore, a high concentration (100 μM) of Yucasin restored the AR phenotype, demonstrating that SlMADS83 may mainly regulate auxin biosynthesis and transport, resulting in the accumulation of auxin in the hypocotyl base and finally enhancing AR formation. In addition, 10 μM Kyn inhibitor treatment did not affect AR formation; it was speculated that the deficiency in auxin biosynthesis by Kyn may be compensated by increased auxin transport regulated by SlMADS83. Surprisingly, 100 μM Kyn was found to increase the ARs in both transgenic and wild-type tomato plants. This is the opposite of what we would expect if Kyn reduced ethylene signaling and auxin biosynthesis. We speculate that the high concentration of Kyn may cause damage to the primary root and thus increase AR formation.

To date, we have not been able to determine whether SlMADS83 directly or indirectly regulates these upregulated genes, including the genes involved in auxin biosynthesis and transport, and ethylene biosynthesis. The MADS domain of the MADS-box gene could recognize and bind to a DNA sequence called a GArG-box motif (Tang and Perry 2003; Aerts et al. 2018). A recent report demonstrated that the chrysanthemum MADS-box gene CmANR1 directly binds to the recognition site GArG motif of CmPIN2 through ChIP-PCR. In vivo, CmANR1 directly activates the transcription of CmPIN2 and regulates auxin transport (Sun et al. 2018). Our phylogenetic analysis suggests that SlMADS83 is homologous to SlMBP9. SlMBP9 was a member of the AtANR1 clade (Hileman et al. 2006). SlMADS83 may also recognize and bind some GArG-box motifs of auxin biosynthesis and transport genes. We thus explored whether some putative GArG motifs exist in the promoter sequence of the genes upregulated by SlMADS83. We found that all of these genes have one or several potential GArG-box motifs in the 1000 bp sequence upstream of the initiation codon ATG (Supplementary Table S2). ToFZY4, ToFZY5, and ACS2 have potential GArG motifs 500 bp upstream of the ATG (Supplementary Table S2). This suggests that these genes may be directly regulated by some MADS-box genes, including the MADS-box gene SlMADS83. However, further investigation is required in the future to determine whether increased ethylene biosynthesis is induced by increased auxin or directly activated by SlMADS83.

In conclusion, the silencing of SlMADS83 enhances AR formation. Our study suggests that silencing SlMADS83 leads to more free IAA accumulation in the hypocotyl base. The high concentration of endogenous IAA accumulation may promote early AR primordia initiation, thereby enhancing AR formation. The increased auxin in transgenic plants is regulated by the silencing of SlMADS83, and the increased transcript of ethylene biosynthesis genes may enhance endogenous ethylene biosynthesis, positively regulating auxin biosynthesis and transport, thus leading to auxin accumulation in the hypocotyl base. Together, the silencing of SlMADS83 might, in cooperation with the increased transcript levels of ethylene biosynthesis genes, lead to auxin accumulation in the hypocotyl base and increased AR formation (Fig. 8).

Fig. 8
figure 8

Schematic of AR formation regulated by SlMADS83

Full size image