Introduction

The phytohormone auxin regulates a wide variety of plant developmental and physiological processes. At the cellular level, auxin controls cell division, elongation and differentiation. On the whole-plant level, auxin plays an essential role in many processes, including vascular differentiation and embryogenesis, root elongation, lateral root initiation, stem elongation, fruit set and development, apical dominance, gravitropism and tropisms responses (Friml 2003). This multiplicity of regulatory activities has spurred considerable interests in studying the mechanisms of auxin signalling and response. A major breakthrough in auxin signalling research is the discovery of the F-box protein transport inhibitor response 1 (TIR1), which functions as an auxin receptor (Dharmasiri et al. 2005; Kepinski and Leyser 2005). It has been known that auxin signalling operates by recruiting specific transcription factors, leading to the expression of downstream genes that perform the required responses (Vogler and Kuhlemeier 2003). Three major classes of early auxin response genes have been identified from various plant species: GH3 family, the SAUR family, and the Aux/IAA family (Guilfoyle 1999). Auxin induces many of these genes rapidly, specifically and without the requirement of de novo protein synthesis, and these genes are considered primary response genes (Tian et al. 2002).

Aux/IAA genes encode short-lived and nuclear-localised proteins that contain four highly conserved domains (I, II, III and IV) (Abel et al. 1995). Domain I functions as a transcriptional repressor (Tiwari et al. 2004). Domains II and IV contain functional nuclear localisation signals (Abel and Theoligis 1995). In addition, domain II plays a role in destabilising Aux/IAA proteins and may be a target for ubiquitination (Colon-Carmona et al. 2000). Domain III contains a βαα motif similar to the DNA binding domain found in prokaryotic repressor of the ArcA family (Abel et al. 1994). Evidence indicates that domains III and IV mediate the homodimerisation and heterodimerisation between Aux/IAA proteins and heterodimerisation between Aux/IAA proteins and auxin response factors (ARFs) (Ouellet et al. 2001).

Screening Arabidopsis thaliana mutants has identified mutations in several different Aux/IAA genes, including IAA1, IAA3, IAA6, IAA7, IAA12, IAA14, IAA17, IAA18, IAA19 and IAA28 (Park et al. 2002; Yang et al. 2004; Tian and Reed 1999; Kim et al. 1996; Nagpal et al. 2000; Hamann et al. 2002; Fukaki et al. 2002; Rouse et al. 1998; Reed 2001; Tatematsu et al. 2004; Rogg et al. 2001). Those screened mutants exhibited changes of a variety of auxin-related developmental phenotypes, and all of them were gain-of-function mutations. No obviously visible phenotypes were observed on loss-of-function mutations in most Arabidopsis Aux/IAA genes. Down-regulation of an Aux/IAA gene from tomato, SlIAA9, which has been proved to be involved in tomato fruit development and leaf morphogenesis, was able to convert tomato compound leaves to simple leaves (Wang et al. 2005).

Leaf functions in light capture and photosynthesis, thus it is critical to plant survival. The leaves of higher plants can be divided into two categories: simple leaves and complex leaves. Complex leaves can be pinnate or palmate. The cultivated tomato, Solanum lycopersicum L., has a typical unipinnate compound leaf. However, a large number of tomato mutants, including the entire (e), potato leaf (c), trifoliate (tf), wiry (w), wiry3 (w3), wiry4 (w4), wiry6 (w6), Lanceolate (La), clausa (clau), Mouse ears (Me), Petroselinum (Pts) and tripinnate (tp), that either decrease or increase the degree of leaf complexity, are available. Among these mutants, clau, Me, Pts and tp mutants increase the complexity of tomato leaves, and the others decrease the complexity. Leaf development in most of the above mutants have been studied extensively at the morphology and histology levels (Kessler et al. 2001), whereas there are very limited reports on these mutants at the molecular level.

The tomato e mutation has fewer leaf segments and looks superficially less complex than wild type (Fig. 1, Rick and Butler 1956). Down-regulation of SlIAA9 converted tomato compound leaves to simple leaves and exhibited a similar leaf phenotype of the entire mutant. In addition, the transcription level of the SlIAA9 gene was decreased in the entire mutant (Wang et al. 2005). These results suggested that there were some relationships between the entire mutant and SlIAA9 gene. In this paper, we reported that there is a single-base deletion in the coding region of SlIAA9 from the entire mutant. DNA blot analysis indicated that SlIAA9 was present as a single copy in the genome of tomato. Bin-mapping results showed that SlIAA9 and e genes are located at the same region on tomato chromosome 4. Further investigation also showed that another single-base deletion occurred in the coding region of SlIAA9 from e0880, an entire allelic mutant generated by ethane methyl sulfonate (EMS) mutagenesis.

Fig. 1
figure 1

Leaves from tomato cultivar Ailsa Craig (AC) and entire AC mutant. a Seedling of tomato cultivar AC; b seedling of entire AC mutant (LA2922); c compound leaf of tomato cultivar AC; d complexity-reduced leaves of entire AC mutant

Full size image

Materials and methods

Plant materials and growth conditions

Tomato cultivar LA3475 (M82) and the wild species LA716 (S. pennellii) were used for genomic DNA analysis. Seventy-five introgression lines (ILs) generated by Eshed and Zamir (1995) and their parents LA716 and LA3475 were used for genetic mapping of SlIAA9. Tomato cultivar Ailsa Craig (AC) and entire mutants including LA2922 (entire AC), LA0159, LA0885, LA0902, LA0917, LA0990, LA0991, and LA2360 were used for cDNA sequence analysis, as well as tomato cultivar LA3475 (M82) and entire allelic mutant e0880. The immature leaves of tomato Ailsa Craig and tomato entire mutants, LA2922, LA0159, LA0885, LA0902, LA0917, LA0990, LA0991 and LA2360 were used for expression analysis of SlIAA9. All plants were grown in a naturally illuminated greenhouse.

Cloning and sequence analysis of SlIAA9

We had isolated the full-length cDNA of SlIAA9 from tomato cultivar Zhongshu No. 5 and submitted the sequences to GenBank (accession number: DQ115325). Because the SlIAA9 cDNA clone was found to correspond to a previously isolated 301-bp partial tomato IAA4 clone (GenBank accession number AF022015; Nebenfuhr et al. 2000), initially we named this gene SlIAA4. In order to comply with the numbering retained for the Arabidopsis Aux/IAA genes, Wang et al. (2005) renamed this gene SlIAA9. So we use this designation throughout our paper. According to the full-length cDNA sequences, the primers 5′-GTTACTGTCTGCCAAATGGAGGGT-3′ (forward) and 5′- GCACACGCATACTTAATCAAACGACA-3′ (reverse) were used to amplify the cDNA fragment of SlIAA9 from tomato cultivars Ailsa Craig and LA3475 (M82), entire mutants and entire allelic mutant. Polymeraise chain reaction (PCR) products were gel purified and cloned into a plasmid vector, pMD18-T (TaKaRa, Japan), and four independent clones were sequenced to determine the nucleotide acid sequences. The sequence comparisons were performed using the Clustal W programme with standard parameters (http://www.ebi.ac.uk/clustalw/).

DNA isolation, DNA gel-blot analysis and bin mapping of SlIAA9 gene

Tomato cultivars M82 and LA716 were used for genomic DNA analysis. Total genomic DNA was isolated from these two accessions using the method described in Fulton et al (1995). Genomic DNA (15 μg) was digested with EcoR I , EcoR V, Dra I, and Hae III, respectively, then fractionated in 1% agarose gel, transferred onto a nylon membrane and hybridised with 32P-labelled SlIAA9 full-length cDNA. The hybridisation was done at 65°C in a phosphous buffer (0.5 M) containing 7% sodium dodecyl sulfate (SDS), 1% bovine serum albumin (BSA) and 1 mM ethylenediaminetetraacetate (EDTA). The membrane was washed with a solution of 2× sodium saline citrate (SSC) and 0.5% SDS at room temperature for 10 min, followed by a solution of 0.2× SSC and 0.1% SDS under 65°C for 5 min.

Seventy-five LA716-derived ILs were used to bin map the SlIAA9 gene. Each IL differs in a single defined LA716 chromosome fragment introgressed into the M82 background, and collection of 75 ILs provides overlapping coverage of the entire LA716 genome (Eshed and Zamir 1995). A DNA gel blot containing genomic DNA from all 75 ILs and their parents (M82 and LA716) digested with Hae III was generated and hybridised with SlIAA9 full-length cDNA under high-stringency conditions as described in genomic DNA analysis.

Semiquantitative RT-PCR analysis

Total RNA was isolated from young leaves with Trizol reagent (Invitrogen, USA) according to the supplier’s instruction. Reverse transcription (RT) reactions were carried out as recommended by the manufacturer (Takara, Japan). The products of RT were diluted with diethylpyrocarbonate (DEPC)-treated water. The resulting dilutions were used for gene expression analysis by semiquantitative RT-PCR. To examine the transcription level of SlIAA9, PCR reactions were carried out for 26 cycles of 94°C 45 s/57°C 1 min/72°C 1 min in a volume of 20 μl containing cDNA template dilution, 200 μM each dNTP, 4 mM MgCl2, 4 mM forward and reverse primer pairs, 1 U Taq polymerase (recombinant, MBI, USA). The primers used for SlIAA9 were 5′-GTTACTGTCTGCCAAATGGAGGGT-3′ (forward) and 5′-CTCGCCTACTAGAAATGCTGGA-3′ (reverse). SlUBI3, a tomato ubiqitin 3 gene (Hoffman et al. 1991), was used as an internal control. The primers used for SlUBI3 were 5′-CTGGGGATGGTGTCAGCCACAC-3′ (forward) and 5′-CACCGAACTTTCTCTCGGAAGGTG-3′ (reverse). Ten microlitre PCR products were detected by agarose gel electrophoresis. The PCR reactions were repeated three times using independent samples.

Results

A deletion mutation occurs in the tomato entire mutant

The SlIAA9 gene was involved in tomato fruit development and leaf morphogenesis, and down-regulation of SlIAA9 would convert tomato compound leaves to simple leaves, a phenomenon also observed in the entire mutant. Transgenic lines with repressed expression of SlIAA9, the same as the entire mutant, carried a lower frequency, asymmetrical sepals and a multifusion phenotype affecting leaves, flowers and fruits mutant. Molecular analysis also indicated that the SlIAA9 transcription level was decreased in the entire mutant (Wang et al. 2005). So it was logical to think that there might be some relationships between SlIAA9 gene and the entire mutant. In order to examine these relationships, the complete coding sequence of SlIAA9 from Ailsa Craig and LA2922 (entire AC) were analysed. Sequence analysis revealed that a single-base cytosine deletion of SlIAA9 mRNA occurs at the nucleotide position 626 in the entire AC (Fig. 2). Amino acid sequences of SlIAA9 were identical between Ailsa Craig and the entire AC for the first 208 bases. Then, the deletion changed the subsequent reading frame in the entire AC altered the following two amino acids, and the polypeptide was terminated immediately. The truncated polypeptide consisted of only 210 amino acids and lacked domains III and IV (Fig. 3). Previous studies indicated that domains III and IV of Aux/IAA proteins were involved in mediating the homodimerisation and heterodimerisation between Aux/IAA proteins and ARFs (Ouellet et al. 2001). The deletion of domain III and IV regions should make the SlIAA9 protein functionally defective.

Fig. 2
figure 2

A single-base cytosine deletion of SlIAA9 mRNA occurs in the entire AC mutant. AC-SlIAA9:SlIAA9 from tomato cultivar Ailsa Craig; eAC-SlIAA9:SlIAA9 from the entire AC mutant. Arrow indicates the position of cytosine deletion of eAC-SlIAA9

Full size image
Fig. 3
figure 3

A single-base deletion of SlIAA9 from the entire AC mutant changed the subsequent reading frame and prematurely terminated the translation. Conserved domains I and II of Aux/IAA proteins are framed, the cytosine deletion site is shown by the arrow and the stop codon is indicated by asterisks

Full size image

SlIAA9 gene expressed at a lower level in entire mutants

Besides LA2922 (entire AC), the e gene had been interlined into tomato materials with different backgrounds by hybridising and self crossing, including 3-616, LA0159, LA0281, LA0648, LA0774, LA0781, LA0784, LA0885, LA0886, LA0888, LA0902, LA0917, LA0920, LA0982, LA0990, LA0991, LA1038, LA1075, LA2360 and LA2492. We collected some of these entire mutants from the Tomato Genetics Resource Center (TGRC), including LA0159, LA0885, LA0902, LA0917, LA0990, LA0991 and LA2360. The transcription levels and the complete coding sequences of SlIAA9 from these collected materials were investigated. RT-PCR analysis indicated that the transcription of SlIAA9 in young leaves was significantly lower in these entire mutants, as well as entire AC, when compared to that in wild-type cultivar Ailsa Craig (Fig. 4). Sequence analysis indicated that the coding sequences of SlIAA9 gene from all the investigated entire mutants were the same as that of entire AC, with a single-base cytosine deletion at nucleotide position 626.

Fig. 4
figure 4

Transcript accumulation of SlIAA9 in tomato leaves. Reverse transcription polymerase chain reaction (RT-PCR) was performed using RNA extracted from young leaves of tomato cultivar Ailsa Craig (lane 1) and different tomato entire mutants LA2922 (entire AC, lane 2), LA0159 (lane 3), LA0885 (lane 4), LA0902 (lane 5), LA0917 (lane 6), LA0990 (lane 7), LA0991 (lane 8) and LA2360 (lane 9). PCR reactions were performed for 26 cycles for SlIAA9 and 28 cycles for an internal control ubi3

Full size image

SlIAA9 and e gene locate on the same region of tomato chromosome 4

Genetically, the e gene had been mapped on tomato chromosome 4 (Tanksley et al. 1992). In order to define SlIAA9 gene linkage to the e gene, the SlIAA9 gene was bin mapped in an effort to check whether it also located on the same region of tomato chromosome 4. As plant Aux/IAA genes exist as a multigene family, DNA gel blot was first performed to investigate the copy number of SlIAA9 in M82 and the wild relative LA716. The hybridisation pattern is shown in Fig. 5. One prominent band was revealed in the DNA sample digested with EcoRV. Two prominent bands were detected in the DNA samples digested with EcoRI, Dra I or HaeIII. The SlIAA9 cDNA sequence used to probe the blots has restriction sites for EcoR I (one site), Dra I (two sites) and HaeIII (five sites), whereas no site was identified for EcoRV. Except that some small fragments could not be detected and there might be other sites in intron sequence for these restriction enzymes, the number of hybridising fragments detected in each case suggested that the gene corresponding to SlIAA9 should be present as a single copy in the genome of M82 and the wild relative LA716. The result also indicated that the DNA sequence of SlIAA9 was considerably diverged from other members of the tomato Aux/IAA family.

Fig. 5
figure 5

Southern-blot analysis of SlIAA9 in tomato genome. Each lane contains 15 μg genomic DNA extracted from leaves of tomato M82 and LA716 following digestion with the indicated enzymes. Genomic DNA was size fractionated on an agarose gel, blotted and hybridised with the 32P-labelled SlIAA9 full-length cDNA

Full size image

Figure 5 also shows the restriction fragment length polymorphisms (RFLPs) for SlIAA9 between M82 and LA716 when the genomic DNA was digested with the four enzymes. The RFLPs of digestion with EcoRV were employed to map the SlIAA9 gene using the population of 75 ILs whose genomic DNA was also hybridised with the probe of SlIAA9 (full-length cDNA). IL 4-3 displayed the LA716 allele (Fig. 6), whereas IL 4-3-2 and IL 4-4 did not, indicating that chromosome segment from LA716 in IL 4-3 contained the SlIAA9 locus, but that chromosome segments from LA716 in IL 4-3-2 and IL 4-4 did not. In accordance with these findings, the SlIAA9 was mapped to the region designated as 4-G (Fig. 7). Interestingly, the e gene was mapped in the same region with the SlIAA9 gene on the tomato chromosome 4 (Tanksley et al. 1992).

Fig. 6
figure 6

Bin mapping of SlIAA9 on tomato genome. Genomic DNA from 75 LA716-derived introgression lines (ILs) and their parents, M82 and LA716, were digested with EcoR V, size-fractionated on an agarose gel, blotted and hybridised with the 32P-labelled SlIAA9 cDNA. Lanes 4-2, 4-3 and 4-3-2 represent IL4-2, IL4-3 and IL4-3-2, respectively

Full size image
Fig. 7
figure 7

The position of SlIAA9 on the tomato genome. The chromosome is depicted as wide bar, and the LA716 introgressed segments appear as thin bars to the left of the chromosome. Arrow indicates the bin location of SlIAA9 on the tomato genome

Full size image

A single-base deletion of SlIAA9 occurs in the mutant with a recessive entire allele

A mutant population of 13,000 tomato M2 families of the M82 variety had been generated using EMS and fast-neutron mutagenesis by Zamir lab at the Hebrew University in Jerusalem, Israel (Menda et al. 2004). The plants from this population were visually phenotyped in the field, and some of them had been cataloged according to the previously described phenotypes from the monogenic mutant collection at TGRC (http://www.tgrc.ucdavis.edu/). Several entire allelic mutants from this population, including e0444, e0880, e2978, e2986, e3335 and n0741, have been determined by crossing to a known entire mutant and testing the progeny for the phenotype. The corresponding information of these mutants is available and can be searched on the Web site http://www.zamir.sgn.cornell.edu/mutants/. In an effort to confirm that the SlIAA9 gene was the e gene, the complete coding region of the SlIAA9 gene from e0880, a mutant generated using EMS mutagenesis, was sequenced. Sequence analysis revealed that a single-base guanine deletion of SlIAA9 occurs at the nucleotide position 583 in e0880 (Fig. 8a). Amino acid sequences were identical between M82 and e0880 for the first 194 bases. Then, the deletion changed the subsequent reading frame of SlIAA9 in e0880 by altering the following 16 amino acids and terminating the polypeptide thereafter (Fig. 8b). As with the SlIAA9, in the entire mutant, the truncated polypeptide also consisted of only 210 amino acids. However, besides domain III and IV, this truncated polypeptide also lacked domain II, which had been confirmed to play a role in destabilising Aux/IAA proteins.

Fig. 8
figure 8

A single-base guanine deletion of SlIAA9 occurs in the entire allelic mutant e0880. a M82-SlIAA9 represents the SlIAA9 coding sequence (CDS) from tomato cultivar M82. The 0880-SlIAA9 represents the SlIAA9 gene fragment from the entire allelic mutant e0880. The guanine deletion of the 0880-SlIAA9 gene is indicated by the arrow. b Deletion mutation altered the subsequent reading frame and prematurely terminated the translation of the SlIAA9 gene from the entire allelic mutant e0880. The cytosine deletion site is indicated by an arrow. and the stop codon is indicated by asterisks

Full size image

Discussion

SlIAA9 was involved in fruit development and leaf morphogenesis. Transgenic lines with decreased level of SlIAA9 showed nearly identical phenotypes as the tomato entire mutant in various aspects, including converting compound leaves to entire leaves, carrying a lower frequency, asymmetrical sepals and a multifusion phenotype affecting leaves, flowers and fruits (Wang et al. 2005). SlIAA9 located on the same region as the e gene. One single-base deletion in the coding region of SlIAA9 was observed in both entire and entire allelic mutants. All of these results illustrate that the entire mutant was produced by the deletion mutation of the SlIAA9 gene.

The tomato is a powerful system in which to study complex leaf morphogenesis due to the fact that many leaf morphology mutants are available (Kessler et al. 2001). The tomato mutants, including entire (e), potato leaf (c), trifoliate (tf), wiry (w), wiry3 (w3), wiry4 (w4) and wiry6 (w6), reduce the compound nature of tomato leaves. Semidominant mutant Lanceolate (La) even converts its leaves to a fully simple architecture (Hareven et al. 1996). Most genes controlling these traits have been mapped. The w, w4 and e genes were all mapped on chromosome 4. The tf gene was mapped on chromosome 5. The c gene was mapped on chromosome 6. The La gene was mapped on chromosome 7. The w6 was mapped on chromosome 10 (Tanksley et al. 1992; Kim et al. 2003). However, to date none of these genes has been cloned. The e gene is the first to be cloned and analysed. Further analysis of the SlIAA9 gene and entire mutant will illustrate the mechanism of a compound leaf formation.

The function of several different Arabidopsis Aux/IAA genes had been characterised by screening various different mutants. It is striking that all of these mutants were caused by gain-of-function mutations and the mutations were all found in the highly conserved domain II of the canonical Aux/IAA proteins, which are responsible for protein degradation (Kim et al. 1996; Rouse et al. 1998; Tian and Reed 1999; Nagpal et al. 2000; Rogg et al. 2001; Fukaki et al. 2002; Hamann et al. 2002). The insertion mutants of 12 Arabidopsis Aux/IAA genes, including the AtIAA9 gene, showed no visible phenotype alterations, indicating the functional redundancy among the various members of the Aux/IAA family (Overvoorde et al. 2005). However, the transgenic plants with a decreased level of SlIAA9 and mutants with defective mutation of SlIAA9 all exhibited dramatic phenotype alterations, indicating that there was no obvious functional redundancy of the SlIAA9 gene in tomato. Furthermore, Wang et al. (2005) revealed that down-regulation of several other tomato Aux/IAA genes, including SlIAA1, SlIAA3 and SlIAA8, also displayed specific, reproducible phenotypes. Our group recently generated several independent RNAi transgenic lines of SlIAA14, which showed a decreased level of the target gene and dramatic auxin-related phenotype alternations (J. Zhang and Z.Ye, unpublished data). All these findings suggest that functional redundancy of Aux/IAA genes, which was common in Arabidopsis thaliana, was not obvious in tomato. These observations sustain the idea that evolution endowed the unique functions to at least several tomato Aux/IAA genes.