Genetic interactions in the control of flowering time and reproductive structure development in tomato (Solanum lycopersicum)

Authors

  • Muriel Quinet,

    1. Unité de Biologie Végétale, Département de Biologie et Institut des Sciences de la Vie, Université catholique de Louvain, Croix du Sud 5, boîte 13, B-1348 Louvain-la-Neuve, Belgium;
    Search for more papers by this author
  • Vincent Dielen,

    1. Unité de Biologie Végétale, Département de Biologie et Institut des Sciences de la Vie, Université catholique de Louvain, Croix du Sud 5, boîte 13, B-1348 Louvain-la-Neuve, Belgium;
    Search for more papers by this author
  • Henri Batoko,

    1. Unité de Biologie Végétale, Département de Biologie et Institut des Sciences de la Vie, Université catholique de Louvain, Croix du Sud 5, boîte 13, B-1348 Louvain-la-Neuve, Belgium;
    Search for more papers by this author
  • Marc Boutry,

    1. Unité de Biochimie Physiologique et Institut des Sciences de la Vie, Université catholique de Louvain, Croix du Sud 5, boîte 15, B-1348 Louvain-la-Neuve, Belgium;
    Search for more papers by this author
  • Andrée Havelange,

    1. Laboratoire de Physiologie végétale, Département des Sciences de la Vie, Université de Liège, Bât. B22-Botanique, Sart Tilman, B-4000 Liège, Belgium
    Search for more papers by this author
  • Jean-Marie Kinet

    1. Unité de Biologie Végétale, Département de Biologie et Institut des Sciences de la Vie, Université catholique de Louvain, Croix du Sud 5, boîte 13, B-1348 Louvain-la-Neuve, Belgium;
    Search for more papers by this author

Author for correspondence: Jean-Marie Kinet Tel:+32 10 472050 Fax:+32 10 473435 Email: kinet@bota.ucl.ac.be

Summary

  • • Different tomato (Solanum lycopersicum) mutants, affected in flowering time, reproductive structure or plant architecture, were crossed to produce double mutants in order to investigate gene interactions in flowering regulation in this autonomous species with a sympodial growth habit.
  • • The compound inflorescence: uniflora, uniflora: self pruning, uniflora: blind, and jointless: uniflora double mutants all produced solitary flowers like their uniflora parent, instead of inflorescences.
  • • All double mutants were late flowering. uniflora: blind and uniflora: self pruning had flowering times intermediate between those of their two parents. jointless: uniflora and compound inflorescence: uniflora flowered later than uniflora, the mutant with the most delayed flowering. All double mutants developed strong lateral shoots at node levels approximately corresponding to the level at which their parent cultivars initiated their first reproductive structure, which is a typical trait of uniflora.
  • • These results suggest that the UNIFLORA gene acts upstream of the other investigated genes in controlling flowering in tomato, and that floral transition of the primary shoot and floral transition of sympodial segments are regulated differently.

Introduction

The genetic control of flowering in tomato (Solanum lycopersicum) is still poorly understood compared with model species such as Arabidopsis thaliana, where the identification of numerous genes regulating flowering time and flower development has resulted in the description of a network of pathways involved in the control of flowering (Boss et al., 2004; Jack, 2004). Several of these genes have been isolated through mutant analyses and, in some cases, the combination of simple mutations through the production of double and triple mutants has allowed the identification of interactions between some of these genes (Reeves & Coupland, 2001).

In contrast to A. thaliana, which is a monopodial quantitative long-day plant, tomato is an autonomously flowering plant with a sympodial growth habit. In tomato, growth of the primary shoot emerging from the seed is terminated by the initiation of the first inflorescence, and growth of the plant continues from the active development of the bud at the axil of the last leaf formed before the reproductive structure. This bud produces a shoot segment bearing some leaves before initiating a new inflorescence, which is once again rejected laterally by the active outgrowth of an axillary bud, and the process is indefinitely reiterated in indeterminate tomatoes which are used for production of fresh fruits in glasshouses and gardens (Kinet & Peet, 1997). The stem portions that are repetitively produced after each new inflorescence are called the sympodial segments. The gene SELF PRUNING (SP) regulates sympodial development by controlling the regularity of the vegetative–reproductive switch of the different sympodial segments (Pnueli et al., 1998). The sp mutant is determinate: the number of vegetative nodes arising on successive sympodial shoots is gradually reduced until the vegetative phase is bypassed completely with the production of two successive inflorescences (Pnueli et al., 1998). Determinate tomatoes are used mainly by the food processing industry. The SP gene has been cloned and is thought to be the orthologue of the A. thaliana TERMINAL FLOWER1 (TFL1) and Anthirrinum majus CENTRORADIALIS (CEN) genes, both acting as inflorescence meristem identity genes (Pnueli et al., 1998).

In tomato, different mutants have been shown to be affected in their flowering response, including uniflora (uf), compound inflorescence (s), blind (bl), and jointless (j) (Dielen et al., 1998, 2004; Quinet, 2005).

The uf mutant produces solitary, normal, fertile flowers instead of inflorescences and always flowers later than the wild type, the late-flowering character being observed in both the initial and the sympodial segments. Flowering time in uf is season-dependent: it is particularly delayed when the daily light energy integral is low. In these conditions, plants develop strong lateral branches at node levels where normally the wild-type plant initiates inflorescences, suggesting that they undergo a partial evocation at this level but are unable to complete the process (Dielen et al., 2004). The UF gene thus regulates time to flowering and inflorescence meristem identity in tomato.

The s inflorescence is highly branched as a result of the conversion of floral meristems into inflorescence meristems; hence, the S gene appears to be a floral meristem identity gene (Quinet et al., 2006).

The j mutant was first characterized as a result of its lack of abscission zone on the flower pedicel (Szymkowiak & Irish, 1999; Mao et al., 2000). It produces inflorescences that contain leaves and revert to vegetative growth (Philouze, 1978; Szymkowiak & Irish, 1999, 2005). These observations indicate that the J gene is involved in the maintenance of the identity of the inflorescence meristem, preventing it from reverting to a vegetative identity. The J gene is a MADS box gene (Mao et al., 2000).

Finally, the bl mutant lacks lateral branches, its inflorescences contain one to three flowers and its flowering time is delayed (Schmitz et al., 2002). This gene encodes a MYB transcription factor (Schmitz et al., 2002).

The aim of this study was to contribute to the elucidation of the potential interactions between these genes through the production of double mutants using, as a common parent, the uf mutant, which we have been investigating for several years. Our previous work has demonstrated that the UF gene is pivotal in regulating flowering in tomato. In this work, we describe the organization and morphogenesis of the reproductive structures and investigate the flowering time of the compound inflorescence: uniflora (s:uf), uniflora: self pruning (uf:sp), uniflora: blind (uf:bl) and jointless: uniflora (j:uf) double mutants.

Materials and Methods

Plant material

Seeds of the Ailsa Craig (AC) tomato (Solanum lycopersicum L.) cultivar and of the uniflora (uf), compound inflorescence (s) and blind (bl) mutants were obtained from the Tomato Genetics Resource Center (University of California, Davis, CA, USA). Heinz (Hz) cultivar and jointless (j) mutant seeds were provided by the Institut National de la Recherche Agronomique (INRA; Montfavet, France). The uf and s mutations were introduced into AC and the j mutation into Hz, but the genetic background of bl is not known. Hz is a determinate cultivar that is mutated for SP. As reported by Philouze (1978) and Schmitz et al. (2002) and shown in Fig. 1, the j and bl mutants used in this study are also mutated for SP.

Figure 1.

Sequencing-derived chromatograms showing nucleotides at the expected position at which SELF PRUNING is mutated. The cDNAs are from tomato (Solanum lycopersicum) Ailsa Craig (AC) and Heinz (Hz) cultivars, uniflora (uf), compound inflorescence (s), blind (bl) and jointless (j) single mutants, and compound inflorescence: uniflora (s:uf), uniflora: blind (uf:bl), jointless: uniflora (j:uf) and uniflora: self pruning (uf:sp) double mutants. The C to T substitution replacing the proline at position 76 by a leucine in sp is shown in bold and underlined.

The double mutants were produced by crossing uf, Hz, s, bl and j as described in the Results section.

Analysis of SELF PRUNING

The presence of the sp mutation in a given genotype was confirmed by sequencing SP cDNA. Total RNA was prepared from leaves of the AC and Hz cultivars and of the uf:sp, s:uf, uf:bl and j:uf double mutant plants using the RNAgents Total RNA Isolation System from Promega (Promega, Madison, WI, USA). The RNA was subsequently used as a template for first-strand cDNA synthesis using the Superscript III First-strand Synthesis for RT-PCR kit (Invitrogen Life Technologies, Groningen, the Netherlands), according to the manufacturer's instructions. For specific amplification of SP, we used GoTaq DNA polymerase (Promega Benelux b.v.), with ATGGCTTCCAAAATGTGTGA and CTGCCGCTAGAAGGCGTTGA as primers, and 1 µl of the total volume of 20 µl from the first-strand reaction as a template. The polymerase chain reaction (PCR) conditions were 2 min at 94°C, 30 cycles each consisting of 30 s at 94°C, 30 s at 55°C and 1 min at 72°C and a final extension at 72°C for 5 min. The PCR products were purified and sequenced using an automatic sequencer (Genetic Analyser 3100; Applied Biosystems, Foster City, CA, USA) and the BigDye Terminator v1.1 Cycle Sequencing Kit (Applied Biosystems, Madison, WI, USA). The resulting nucleotide sequences were aligned using clustalx (Thompson et al., 1997) and checked for the presence of mutation.

Growth conditions

Seeds were germinated at 25°C in peat compost. When approximately 2 wk old, seedlings were transplanted to 7 cm × 7 cm pots filled with the same compost and, approximately 3 wk later, plants were transferred to pots of 15 cm diameter, and fertilized weekly with a nutrient solution made up of 15 g l−1 of a 16-18-21 N-P-K fertilizer. The F2 populations were grown in phytotronic growth chambers at the University of Liège (Liège, Belgium), where night and day temperatures were maintained at 20–25°C and light was provided exclusively by means of white fluorescent lamps (Osram Cool white, 40 W) at an irradiance of 180 µmol m−2 s−1 over the waveband 400–700 nm. Experiments intended to define the role of the photoperiod in plant development were also performed in the phytotronic growth chambers of the University of Liège, where two day lengths, 16 and 8 h, were compared. All other experiments were carried out in a heated glasshouse in Louvain-la-Neuve (average temperature 20°C). Depending on the objective of the experiment (see Results section), plants were subjected or not to extra lighting provided by Philips HPLR 400-W bulbs giving a 16-h day length at a minimum 100 m−2 s−1 irradiance over the waveband 400–700 nm at the top of the canopy.

Criteria for flowering response

Two criteria were used to evaluate the flowering response: the percentage of plants forming a reproductive structure before the initiation of the 40th leaf, and the flowering time. Flowering time was assessed using two different measurements: (i) the number of days from sowing to macroscopic appearance of the first reproductive structure and (ii) the number of leaves (before the 40th) produced below the first reproductive structure. Depending on the experiment, one or both measurements were taken.

The main difficulty encountered in this study resulted from the fact that some of the double mutants had a mixture of genetic backgrounds, as parental single mutants were only available in different cultivars. Conclusions drawn from comparisons of flowering times of single mutants, double mutants and cultivars used as controls have thus to be treated with caution, especially when the bl mutation is present, as it was introduced into an unknown genetic background. For all the other mutants, which have Ailsa Craig or Heinz as their genetic background, we found no differences in flowering time among the cultivars and the resulting F1 and F2 progenies from crosses between the two cultivars (see Supplementary Material).

Statistical analysis

Normality tests were performed and no further transformation of the raw data was required. Analysis of variance (ANOVA) II (SAS System for Windows, V8; SAS Institute, Cary, NC, USA) was performed to evaluate the effects of genotype and sowing date on the measured parameters. Differences between means were scored for significance according to the Scheffe F-test. Comparison of flowering percentages was performed using the χ2 test.

Results

Double mutant production, identification and verification by backcross and sequencing of the SP coding sequence

The uf mutant was used as female parent in the cross with Hz (sp mutant) and bl and as male parent in the cross with j and s. Pollen from the male parent was manually transferred to the stigma of emasculated female parent flowers. First generation (F1) progeny plants were selfed, and the resulting second generation (F2) populations used to identify double mutants. The F2 populations followed a 9 : 3 : 3 : 1 Mendelian segregation. Unambiguous identification of the double mutants was possible because, on the one hand, they all produced solitary flowers as in uf, and, on the other hand, either they had a flowering time significantly different from the flowering times of both parental mutants, as was the case for the s:uf and uf:sp double mutants (see Supplementary Material), or they exhibited additional traits pertaining to the non-uf parental mutant. This was the case for the uf:bl and j:uf double mutants: in the uf:bl double mutant, lateral shoot production was indeed inhibited as in the bl mutant, while the j:uf double mutant lacked the pedicel abscission zone, which is a j character.

The seeds of the double mutant plants were collected, propogated and pooled to form a stock for further experiments. We never observed variations in the flowering time data for the F3 progeny derived from pooled seeds from F2 double mutants, except for the variability that was clearly caused by the mutations investigated, especially by the uf mutation. This observation strengthens the assumption that the combination of genetic backgrounds in the double mutants had a limited influence on their flowering time.

Backcrosses were performed between two or three plants of each double mutant and their parental genotypes to ascertain the presence of both mutations of interest.

As shown in Table 1, the plants generated by the crosses between each of the double mutants and uf initiated a solitary flower instead of an inflorescence, reminiscent of the uf mutant phenotype, and flowered later than the plants resulting from the other crosses. The variations in flowering time observed among the backcrosses (with the uf parent) were a result of differences in sowing date and cultivation conditions. The results for the other backcrosses (with the non-uf parent) revealed that: (i) the inflorescences of the plants resulting from the cross between s:uf and s were highly branched and contained numerous flowers reminiscent of the s phenotype; (ii) the F1 of the cross between uf:bl and bl produced a reduced number of flowers per inflorescence and had leaf axils deprived of axillary buds, which corresponds to the bl phenotype; (iii) the inflorescences of the plants resulting from the backcross between j:uf and j contained flowers and leaves, and the flower pedicels lacked the abscission zone as in j plants, and finally (iv) the F1 of the cross between uf:sp and sp developed normal tomato inflorescences like the sp plants. As the j and bl mutants used as parents were also mutated for SP and because the sp determinate character was not readily detectable (requiring the analysis of some sympodial segments to assess the reduction in leaf number), the presence or the absence of the sp mutation was determined by sequencing the SP coding sequences of the different genotypes investigated. As shown in Fig. 1, the punctual C-to-T substitution resulting in the replacement of proline 76 by a leucine in the mutant protein was only observed in the Heinz cultivar, the j and bl mutants and the uf:sp double mutant. Hence, the j:uf and uf:bl double mutants were not mutated for SP.

Table 1.  Flowering response of tomato (Solanum lycopersicum) plants derived from the backcrossing of the double mutants compound inflorescence: uniflora (s:uf), uniflora: blind (uf:bl), jointless: uniflora (j:uf) and uniflora: self pruning (uf: sp) with their parental mutants
Backcross (n)Flowering percentageNo. days before MA (mean ± SD)No. leaves below first inflorescence (mean ± SD)No. flowers in first inflorescence (mean ± SD)No. leaves in first inflorescence (mean ± SD)
  1. MA, macroscopic appearance of the first inflorescence; n, number of plants per batch; SD, standard deviation.

s:uf × s (18)10040.38 ± 3.110.6 ± 0.845.1 ± 11.70.6 ± 1.0
s:uf × uf (16) 60 56.6 ± 6.218.8 ± 1.1 1.0 ± 0.00.0 ± 0.0
uf:bl × bl (9)100 42.7 ± 2.214.5 ± 1.8 1.5 ± 0.50.0 ± 0.0
uf:bl × uf (10) 90 45.2 ± 5.114.5 ± 4.5 1.0 ± 0.00.0 ± 0.0
j:uf × j (10)100 54.2 ± 1.811.7 ± 1.2 3.1 ± 0.72.5 ± 1.3
j:uf × uf (8)100 58.7 ± 4.115.2 ± 1.2 1.0 ± 0.00.0 ± 0.0
uf:sp × sp (10)100 58.1 ± 1.810.4 ± 0.9 6.1 ± 1.30.0 ± 0.0
uf:sp × uf (10) 40 72.5 ± 6.517.7 ± 2.1 1.0 ± 0.00.0 ± 0.0

Flowering responses – reproductive structures

As shown in Table 2, all double mutants produced solitary, normal, fertile flowers – instead of inflorescences – like the uf mutant plant. Whereas the s phenotype was characterized by highly ramified inflorescences which could sometimes contain leaves, the j mutant plants developed inflorescences containing both flowers and leaves, the bl mutant plants developed inflorescences containing usually one or two flowers which could be partially fused, and in the AC and Hz (sp mutant) cultivars five to seven flowers were visible in the inflorescences (Table 2).

Table 2.  Genotype-dependent development of flowers and leaves in inflorescences of tomato (Solanum lycopersicum) Ailsa Craig (AC) and Heinz (Hz) cultivars, uniflora (uf), compound inflorescence (s), blind (bl) and jointless (j) mutants and compound inflorescence: uniflora (s:uf), uniflora: blind (uf:bl), jointless: uniflora (j:uf) and uniflora: self pruning (uf:sp) double mutants
Genotype (n)No. of flowers (mean ± SD)No. of leaves (mean ± SD)
  1. n, number of plants per batch; ND, not determined; SD, standard deviation.

AC (20) 6.85 ± 0.9  0 ± 0
Hz (20)  5.4 ± 1.1  0 ± 0
uf (20)    1 ± 0  0 ± 0
s (20)103.5 ± 65.6ND
bl (20)  1.4 ± 0.5  0 ± 0
j (20)  6.5 ± 2.57.4 ± 3.5
s:uf (20)    1 ± 0  0 ± 0
uf:bl (20)    1 ± 0  0 ± 0
j:uf (20)    1 ± 0  0 ± 0
uf:sp (20)    1 ± 0  0 ± 0

The reproductive structures of these single mutants, double mutants and cultivars were similar in the different sympodial segments.

Flowering responses – flowering time

The compound inflorescence:uniflora double mutant  Four sowings were carried out in the glasshouse without extra lighting: for two of these sowings (on 2 and 26 September), the resulting plants grew under autumn conditions, and for the other two (23 May and 20 July), plants were allowed to grow under summer conditions. As shown in Table 3, all the AC and s plants flowered regardless of the sowing date. The flowering percentage for uf was maximal only when the plants were grown in summer, while the double mutant s:uf never flowered before the initiation of the 40th leaf in the glasshouse. The uf mutant always flowered later than the AC cultivar, whereas the s mutant usually had the same behaviour as AC.

Table 3.  Flowering percentage and flowering time for the double mutant compound inflorescence: uniflora (s:uf)
Sowing dateGenotype1 (n)Flowering percentage2No. days before MA (mean ± SD)2No. leaves below first inflorescence (mean ± SD)2
  • 1

    AC, Ailsa Craig cultivar; s, compound inflorescence mutant; uf, uniflora mutant; s:uf, compound inflorescence: uniflora double mutant.

  • 2

    Values followed by the same letter in a column for the same sowing date are not statistically different at the 5% level.

  • MA, macroscopic appearance of the first inflorescence; n, number of plants per batch; ND, not determined.

  • –, no data as plants did not flower before the 40th leaf.

23 MayAC (15)100 aND  8.5 ± 0.82 a
s (42)100 aND 8.49 ± 0.93 a
uf (15)100 aND11.73 ± 2.19 b
s:uf (14)  0 b
20 JulyAC (20)100 aND 10.1 ± 1.2 a
s (46)100 a 38.3 ± 7.02 a10.25 ± 6.44 a
uf (20)100 a 48.3 ± 7.95 b 16.4 ± 3.00 b
s:uf (13)  0 b
2 SeptemberAC (20)100 aND  8.8 ± 1.11 a
s (13)100 aND 8.77 ± 0.92 a
uf (20) 40 bND18.75 ± 1.28 b
s:uf (20)  0 c
26 SeptemberAC (15)100 a48.13 ± 0.52 a10.27 ± 0.88 a
s (15)100 a44.60 ± 3.46 b 9.93 ± 1.16 a
uf (15)  0 b
s:uf (21)  0 b

In an attempt to determine whether the absence of flowering in the double mutant reflected an intrinsic inability to flower or a much delayed flowering response, another set of plants were cultivated in phytrotonic growth rooms under two different day lengths, 16-h long days and 8-h short days, at a light intensity of 180 µmol m−2 s−1 over the waveband 400–700 nm. All the AC and s plants flowered under these conditions (Fig. 2), whereas flowering percentage in uf was higher under long days. Interestingly, one s:uf plant succeeded in flowering before the production of the 40th leaf in both 16- and 8-h day lengths. The s mutant flowered almost at the same time as AC in both day lengths; for both genotypes, the number of days before flowering was slightly higher under the 8-h photoperiod compared with the 16-h photoperiod. In any case, the number of leaves under the first reproductive structure was not affected by the photoperiod. The flowering time of uf and s:uf was strongly delayed in comparison to AC, with the double mutant flowering later than uf. The number of leaves under the first reproductive structure for uf as well as the number of days before flowering for both uf and uf:s was slightly higher under the 8-h photoperiod compared with values recorded under the 16-h photoperiod. The particular effect of the genotype was significant for the number of days before flowering (F = 314.23; P < 0.0001) and the number of leaves under the first reproductive structure (F = 418.69; P < 0.0001), while the variable ‘photoperiod’ affected only the flowering time (F = 5.10; P = 0.0255).

Figure 2.

Day-length effect on flowering parameters in the tomato (Solanum lycopersicum) Ailsa Craig cultivar (black bars), the uniflora (dark grey bars) and compound inflorescence (white bars) mutants, and the compound inflorescence: uniflora double mutant (light grey bars). The percentage of flowering plants, the number of days between sowing and the macroscopic appearance of the first reproductive structure and the number of leaves under this reproductive structure were scored. Number of plants per batch = 20.

Interestingly, for plants that failed to flower, we observed the development of lateral shoots at node 10.11 ± 1.52 (mean ± SD) in uf and 11.55 ± 1.46 in s:uf, approximately at the position at which the wild type produced its first inflorescence.

The uniflora:self pruning double mutant  Three sowings were carried out in the glasshouse and extra lighting was provided for the sowing of 13 October, making a comparison of the season effect unreasonable. The three sowings are therefore considered as repetitions. The number of plants per batch varied from 10 to 20 depending on the sowing date. Under these conditions, every plant of the AC and Hz cultivars flowered, whereas the flowering percentage of the uf plants varied between 40 and 100%, and most of the uf:sp double mutants produced flowers (Table 4). AC and Hz flowered at the same time and after production of the same number of leaves, while in uf and uf:sp flowering time was delayed in comparison to the parental cultivars. The uf mutant flowered significantly later than the uf:sp double mutant.

Table 4.  Flowering percentage and flowering time for the double mutant uniflora: self pruning (uf:sp)
Sowing dateGenotype1 (n)Flowering percentage2No. days before MA (mean ± SD)2No. leaves below first inflorescence (mean ± SD)2
  • 1

    AC, Ailsa Craig cultivar; Hz, Heinz cultivar; uf, uniflora mutant; uf:sp, uniflora: self pruning double mutant.

  • 2

    Values followed by the same letter in a column for the same sowing date are not statistically different at the 5% level.

  • MA, macroscopic appearance of the first inflorescence; n, number of plants per batch; ND, not determined.

5 MayAC (10)100 a42.4 ± 0.9 a 9.9 ± 0.8 a
Hz (10)100 a37.6 ± 4.8 a 9.7 ± 1.6 a
uf (10) 90 a71.6 ± 15.3 b19.3 ± 5.2 b
uf:sp (10)100 a66.6 ± 6.5 b15.4 ± 1.2 c
2 SeptemberAC (20)100 aND 8.8 ± 1.1 a
Hz (16)100 aND 8.4 ± 0.7 a
uf (20) 40 bND18.7 ± 1.3 b
uf:sp (20) 95 aND17.3 ± 1.3 c
13 OctoberAC (10)100 a  52 ± 1.1 a11.7 ± 1.5 a
Hz (10)100 a53.5 ± 2.6 a10.8 ± 0.6 a
uf (10) 90 a73.1 ± 1.0 b18.3 ± 1.1 b
uf:sp (10) 90 a59.1 ± 4.1 c15.4 ± 1.5 c

Once again, the development of lateral shoots was recorded in plants that did not flower in both the uf and uf:sp sets of plants. The first actively growing lateral bud occurred at node 10.35 ± 0.63 in uf and 11 ± 2.40 in uf:sp.

The uniflora:blind double mutant  Four cultures were performed. The plants sown on 14 February were cultivated in a growth cabinet and the other experiments were carried out in the glasshouse, with extra lighting for the sowing of 4 December The different sowings are thus considered as repetitions of the same experiment. As shown in Table 5, all the AC and bl plants flowered. This was also the case for the uf:bl double mutant, but one plant, derived from seeds sown on 13 October, did not produce reproductive structures before the initiation of the 40th leaf. In contrast, the flowering percentage for uf varied from one repetition to another. The bl and uf mutants flowered later than the AC cultivar. The number of days before flowering for uf:bl did not significantly differ from that for bl but was reduced compared with uf, which exhibited the most delayed flowering. Similar tendencies were noted when the number of leaves below the first reproductive structure was used as a measurement of flowering time, with the exception of the sowing in September.

Table 5.  Flowering percentage and flowering time for the double mutant uniflora: blind (uf:bl)
Sowing dateGenotype1 (n)Flowering percentage2No. days before MA (mean ± SD)2No. leaves below first inflorescence (mean ± SD)2
  • 1

    AC, Ailsa Craig cultivar; uf, uniflora mutant; bl, blind mutant; uf:bl, uniflora: blind double mutant.

  • 2

    Values followed by the same letter in a column for the same sowing date are not statistically different at the 5% level.

  • MA, macroscopic appearance of the first inflorescence; n, number of plants per batch; ND, not determined.

4 DecemberAC (25)100 aND 8.8 ± 1.1 a
bl (23)100 aND13.3 ± 1.2 b
uf (25) 84 bND14.7 ± 2 b
uf:bl (36)100 aND13.9 ± 2.4 b
14 FebruaryAC (16)100 a44.1 ± 7.2 a 9.0 ± 3.3 a
bl (20)100 a52.2 ± 2.5 b12.8 ± 0.9 b
uf (20)100 a58.3 ± 8.5 c17.2 ± 3.6 c
uf:bl (19)100 a51.8 ± 3.8 b14.5 ± 1.7 b
2 SeptemberAC (20)100 aND 8.8 ± 1.1 a
bl (19)100 aND14.4 ± 0.9 b
uf (20) 40 bND18.7 ± 1.3 c
uf:bl (13)100 aND16.4 ± 1.8 d
13 OctoberAC (10)100 a  52 ± 1.1 a11.7 ± 1.5 a
bl (10)100 a58.8 ± 1.9 b13.4 ± 0.7 ab
uf (10) 90 a73.1 ± 1.0 c18.3 ± 1.1 c
uf:bl (10) 90 a  62 ± 3.2 b15.4 ± 1.8 b

It is worth noting that lateral shoot production was inhibited in bl: only the two or three nodes around the third node and the nodes just below the reproductive structure developed axillary buds. In contrast, axillary buds developed at the axil of all leaves in AC and uf. Lateral shoot production in the uf:bl double mutant was limited, as in bl, but strong axillary development was evident around the tenth node (9.3 ± 1.41) in uf:bl while the first strong axillary shoot developed at node 10.52 ± 0.65 in uf.

The jointless:uniflora double mutant  Four sowings were carried out in the glasshouse and extra lighting was provided for the sowing of 13 October, so the different sowings are considered as repetitions. All the AC, Hz and j plants flowered and the flowering percentage for the uf mutants and j:uf double mutants varied according to the sowing date (Table 6). The AC and Hz cultivars as well as the j mutant did not show significant differences in number of days before flowering and number of leaves under the first inflorescence, except for the sowing of 5 August, for which j flowered later than AC and Hz. The uf mutant and the j:uf double mutant plants flowered later and flowering time in j:uf was usually delayed compared with uf.

Table 6.  Flowering percentage and flowering time for the double mutant jointless: uniflora (j:uf)
Sowing dateGenotype1 (n)Flowering percentage2No. days before MA (mean ± SD)2No. leaves below first inflorescence (mean ± SD)2
  • 1

    AC, Ailsa Craig cultivar; Hz, Heinz cultivar; uf, uniflora mutant; j, jointless mutant; j:uf, jointless: uniflora double mutant.

  • 2

    Values followed by the same letter in a column for the same sowing date are not statistically different at the 5% level.

  • MA, macroscopic appearance of the first inflorescence; n, number of plants per batch; ND, not determined.

5 MayAC (10)100 a42.4 ± 0.9 a  9.9 ± 0.8 a
Hz (10)100 a37.6 ± 4.8 a  9.7 ± 1.6 a
uf (10) 90 a71.6 ± 15.3 b 19.3 ± 5.2 b
j (10)100 a  43 ± 1.4 a 10.3 ± 1.2 a
j:uf (10)100 a88.6 ± 3.5 c 26.2 ± 2.1 c
5 AugustAC (10)100 a34.4 ± 2.5 a  9.8 ± 0.8 ac
Hz (10)100 a34.5 ± 2.2 a  9.0 ± 0.8 a
uf (10)100 a52.0 ± 5.5 b 18.6 ± 2.2 b
j (10)100 a41.3 ± 3.9 c 11.6 ± 1.9 c
j:uf (5) 80 a59.5 ± 2.6 d 21.5 ± 2.6 b
2 SeptemberAC (20)100 aND  8.8 ± 1.1 a
Hz (16)100 aND  8.4 ± 0.7 a
uf (20) 40 bND 18.7 ± 1.3 b
j (14)100 aND    9 ± 1.2 a
j:uf (5) 40 bNDND
13 OctoberAC (10)100 a  52 ± 1.1 a11.75 ± 1.5 a
Hz (10)100 a53.5 ± 2.6 a 10.8 ± 0.6 a
uf (10) 90 a73.1 ± 1.0 b 18.3 ± 1.3 b
j (10)100 a53.9 ± 2.3 a 11.6 ± 0.7 a
j:uf (10)100 aND 19.1 ± 1.1 b

Strong development of lateral shoots on node 10.3 ± 0.48 of uf mutants and 12.25 ± 1.25 of j:uf double mutants was also observed in nonflowering plants.

Discussion

Gene interactions in the control of reproductive structure development

The s:uf, uf:sp, uf:bl and j:uf double mutants all initiated solitary normal fertile flowers, like the uf mutant, indicating that UF is epistatic to S, SP, BL and J in regulating morphogenesis of the reproductive structure of tomato. Dielen et al. (1998, 2004) postulated that UF is a pivotal gene with a dual role, regulating flowering time and inflorescence meristem identity (Dielen et al., 1998, 2004): at floral transition, the uf shoot apical meristem (SAM) is directly converted to a floral meristem without production of an inflorescence meristem, so that a solitary flower is formed instead of an inflorescence. That UF acts upstream of J and S is consistent with the view that J is involved in the maintenance of the inflorescence meristem identity and that S is a floral meristem identity gene (Quinet, 2005), in which case both will necessarily act downstream of a gene that regulates the production of an inflorescence meristem. The phenotype of the bl mutant, which has been reported to produce one to four flowers per inflorescence (Schmitz et al., 2002), suggests that the BL gene is also involved in the maintenance of the inflorescence meristem, and thus acts downstream of the gene controlling the inflorescence identity. The SP gene has been reported not to be involved in the regulation of reproductive structure development (Pnueli et al., 1998), thus accounting for the dominant role of the uf mutation in the double mutant uf:sp. Several double mutants having sp as a parent have been described and, in agreement with the view that SP is not involved in the morphogenesis of the reproductive structures, their phenotype was reminiscent of that of the non-sp parent. This was reported by Schmitz et al. (2002), who showed that the reproductive structures of bl and of blind2 (bl2) and torosa (to), which are both single mutants allelic to bl, are alike, although bl is a bl:sp double mutant. Other studies indicated that the single flower truss: self pruning (sft:sp) double mutant has the same phenotype as single flower truss (sft) (Molinero-Rosales et al., 2004) and that the anantha: self pruning (an:sp) double mutant produces anantha (an) inflorescences (Pnueli et al., 1998). The only effect of the sp mutation on the reproductive structure of tomato was reported by Rick (1955) and Rick & Butler (1956), who observed that sp reduces the formation of leaves in the j:sp inflorescence. We nevertheless observed, as did Philouze (1978), the presence of leaves in the inflorescences of the j mutant investigated in this work. Philouze (1978) reported that this discrepancy could be a result of the differences in the genetic backgrounds into which the j mutation was introduced. Although double mutants containing the sp mutation usually exhibit the same reproductive structure as their other parent, they all display the determinate sp character (Pnueli et al., 1998; Schmitz et al., 2002; Molinero-Rosales et al., 2004; Quinet et al., 2006). The only exception is the j:sp double mutant, which can be indeterminate, the j mutation partly masking the sp mutation (Emery & Munger, 1970; Philouze, 1978). Interestingly, overexpression of SP in wild-type tomato and sp mutant plants resulted in a tendency for the transgenics to promote the development of extra leaves in the inflorescences (Pnueli et al., 1998). In the same way, an and an:sp plants overexpressing SP have a phenotype similar to that of an:j or falsiflora (fa), which produce indeterminate leafy inflorescences (Allen & Sussex, 1996; Pnueli et al., 1998, 2001). A role for SP in the regulation of the reproductive structure of tomato cannot therefore be excluded.

In A. thaliana, the existence of numerous genes specifying the identity of floral and inflorescence meristems has been reported, and tentative schemes of genetic regulation of reproductive structure morphogenesis have been proposed (Jack, 2004). In tomato, however, such background knowledge is still insufficient to allow elaboration of comparative schemes.

Gene interactions in the control of flowering time

All double mutants were late flowering. uf:bl and uf:sp had flowering times intermediate between those of their two parents and flowered almost at the same time, while j:uf and s:uf flowered later than the uf mutant. The percentage of plants that produced a flower before the 40th leaf was similar in j:uf and uf in all experimental repetitions, and the flowering of the s:uf double mutant was dramatically delayed. Note that, in the glasshouse, none of the s:uf plants flowered before initiation of the 40th leaf, and it was only after several months of growth that a couple of flowers were formed during the summer on old, bushy plants; in the phytotrons, only one s:uf plant succeeded in flowering before the 40th leaf, under each of the two day lengths tested. These last observations demonstrate that the s:uf double mutant is not unable to flower. To date, complete inhibition of flowering in tomato has only been reported by Molinero-Rosales et al. (2004), who produced the sft:fa double mutant which did not undergo the floral transition even after the production of more than 100 leaves.

Interestingly, all double mutants developed strong lateral shoots at node levels approximately corresponding to the level at which their parent cultivars initiate their first reproductive structure under the same growth conditions. Dielen et al. (2004) indicated that this is a typical trait of the uf mutant and they suggested that, as release of apical dominance is known to be an early event associated with floral evocation (Bernier et al., 1981), the uf plants underwent a partial evocation at approximately the same time as the wild type, but that they were unable to complete the process. That uf and all double mutants having uf as a common parent exhibit this same phenotype could be indicative of the occurrence of processes upstream of UF that direct the SAM of the primary shoot into reproductive growth, suggesting that the genes investigated in the present work would have little influence on the timing of events that actually initiate the floral transition of the primary shoot.

Dielen et al. (2004) also reported that, after its partial evocation, the uf SAM returned to vegetative functioning, and apical dominance seemed to be re-established as axillary outgrowth was inhibited at the upper nodes. Our results clearly suggest that, from that time-point, the different genes investigated in the present work were interacting to regulate flowering time. It is therefore possible that, in the uf mutant and the described double mutants, after partial evocation, the plants were in a condition comparable to that of the wild type during its sympodial growth, which results in the regular alternation of vegetative and reproductive phases (Dielen et al., 2004). This condition is different from that of plants during their initial growth from germination to floral transition of the primary segment because, as demonstrated by Pnueli et al. (1998), the flowering times of the sympodial segments and of the primary shoot in tomato are not regulated in the same way, floral transition being regulated by SP in the former but not in the latter. In the uf mutant and the double mutants, genes affecting the flowering time of the sympodial segments could thus be activated after the first partial floral evocation, a process that could be reiterated, even if the plants do not initiate flowers. This hypothesis could account for the flowering times recorded for the double mutants. Indeed, uf:sp and uf:bl flowered before uf plants, which is consistent with the assumption that the sp and bl mutations activate the flowering of the sympodial segments: in the sp mutant, the sympodial segments develop progressively fewer nodes until the shoot is terminated by two consecutive inflorescences (Pnueli et al., 1998), while bl mutants show a tendency to terminate shoot growth after formation of an inflorescence (Schmitz et al., 2002). In contrast, s:uf and j:uf flower later than uf plants. Philouze (1978) reported that the j mutant produces on average two extra leaves under the first reproductive structure compared with the wild type. We also observed a slight and consistent delay in flowering time in j in comparison with Hz, but this difference was not always statistically significant. This delay could apparently enhance the effect of the uf mutation on flowering time. In the case of the s:uf double mutant, its difficulty in undergoing floral transition could be a result of the fact that two critical processes that occur in sequence, namely specification of the inflorescence meristem and specification of the floral meristem, are impaired, increasing the chance of flowering inhibition during each successive sympodial phase.

Unravelling the genetic mechanisms involved in the regulation of floral transition in tomato thus appears to be complicated, compared with A. thaliana, because tomato has a sympodial growth habit and floral transition of the primary shoot is apparently not regulated in the same way as floral transition of the sympodial segments. Schemes intended to summarize the genetic control of flowering of tomato have to take into account this particular and challenging feature. To date, the number of tomato mutants reported to be affected in flowering time is limited, unlike the situation in A. thaliana (Boss et al., 2004; Bernier & Périlleux, 2005). This could be a consequence of (i) the fact that tomato has been far less extensively investigated than A. thaliana with respect to flowering time, and/or (ii) domestication, during which potential mechanisms controlling flowering time may unwittingly have been eliminated in order to extend the environmental conditions under which tomato, the model of autonomous flowering plants, is able to undergo floral transition.

Acknowledgements

This work was supported by funding from the Belgium ‘Fonds National de la Recherche Scientifique (FNRS)’ (through a grant from its Fonds de la Recherche Fondamentale et Collective, 2001-4). MQ is grateful to the FNRS for the award of a research fellowship. HB is a Research Associate of the FNRS.

Ancillary