INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

In general, fleshy fruits are divided in two large groups, climacteric and non-climacteric, based upon the presence or absence of an autocatalytic ethylene burst during ripening (McMurchie et al., 1972). Numerous results have demonstrated the key role of ethylene in the regulation of several ripening processes in climacteric fruit, including ethylene biosynthesis itself (Abeles et al., 1992; Lelièvre et al., 1997; Giovannoni, 2001). However, the regulatory mechanisms governing fruit ripening remain largely unknown.

The ripening of melon (Cucumis melo) fruit of several cultivated varieties and wild ecotypes is climacteric and often associated with fruit detachment (Abeles et al., 1992). Interestingly, there exist several melons that apparently do not abscise and display a long shelf life. It has been shown in some cases that these melon fruits emit little or no ethylene and therefore behave like non-climacteric fruit (Shiomi et al., 1999). Although a number of pleiotropic tomato (Lycopersicon esculentum) mutants resulting in partial or total inhibition of ripening have been described (Thompson et al., 1999; Giovannoni, 2001), studies in different species may lead to identification of other genes or regulatory mechanisms. Cantaloupe is a cultivated species of African origin with a high phenotypic and molecular variation (Naudin, 1859; Stepansky et al., 1999). It is an annual diploid plant, and the size of its genome (450 Mb) is relatively small. The construction of genetic maps with high-density markers was recently achieved (Wang et al., 1997; Périn et al., 2000).

The ethylene biosynthetic pathway is well established (Yang and Hoffman, 1984). The regulation of the expression of the main ethylene biosynthetic genes, 1-aminocyclopropane-1-carboxylic acid (ACC) synthase (ACS) and ACC oxidase (ACO), has been described in many species and tissues (Zarembinski and Theologis, 1994; Giovannoni, 2001), including melon fruit (Yamamoto et al., 1995; Lasserre et al., 1996; Shiomi et al., 1999). This knowledge was used to alter ethylene production in climacteric melon types such as Védrantais (Charentais subgroup of C. melo subsp. melo, cantalupensis). Védrantais displays a high rate of ripening followed also by rapid senescence; both are inhibited in transgenic melon fruit, in which the capacity for ethylene synthesis is inhibited 99% by an anti-sense ACO gene (Ayub et al., 1996). Studying transgenic melon led to the characterization of ripening events that are highly dependent upon ethylene action (Guis et al., 1997; Hadfield et al., 2000). Among these are fruit abscission, degreening of the rind, the duration of shelf-life, and chilling injury (Guis et al., 1997; Ben Amor et al., 1999). The stimulated expression of numerous genes observed during ripening is also strongly dependent on high ethylene levels (Guis et al., 1999; Shiomi et al., 1999; Hadfield et al., 2000).

The non-climacteric phenotype in melon fruit may be attributable to alteration of either upstream developmental processes or any element of the ethylene signal transduction pathway (Giovannoni, 2001). Spectacular progress has been made in the comprehension of the molecular mechanisms of ethylene perception (for review, see Bleecker and Kende, 2000; Stepanova and Ecker, 2000) essentially thanks to Arabidopsis mutants altered in the triple response. Exposure of etiolated seedlings to exogenous ethylene inhibits root and shoot elongation and induces radial swelling of the shoot as well as changes in its orientation to gravity; the whole response is referred to as the triple response (Bleecker and Kende, 2000). Moreover, root hair formation is also regulated by ethylene, at least in Arabidopsis (Kieber et al., 1993; Raz and Ecker, 1999). The tomato Nr mutant, originally identified as a pleiotropic fruit ripening mutant, turned out to display a similar phenotype of global ethylene insensitivity and was identified as an allelic mutation in an ERS-like ethylene receptor gene (Lanahan et al., 1994; Wilkinson et al., 1995). In Arabidopsis and other plants exists a small multigenic family of redundant ethylene receptors (Bleecker and Kende, 2000) expressed in many tissues. All ethylene receptors act as negative regulators of the transduction pathway as shown by elegant genetic analysis in Arabidopsis (Hua and Meyerowitz, 1998) and confirmed in transgenic tomato (Tieman et al., 2000). No receptor gene appears to be specifically expressed during tomato and melon fruit development, although the observed hormonal, temporal, and spatial regulation of the different genes probably play a role in modulating ethylene sensitivity (Wilkinson et al., 1995; Lashbrook et al., 1998; Sato-Nara et al., 1999; Tieman and Klee, 1999). Transcription factors EIN3 and EIN3-like (EIL; Chao et al., 1997) have also been identified. These DNA-binding proteins positively regulate the expression of target genes during the primary response in Arabidopsis (Solano et al., 1998), but little is known about the regulation and role of these elements during fruit ripening.

We initiated physiological and genetic analysis of the non-climacteric phenotype of a nonabscising, long-shelf-life melon fruit. These data provide a starting point to dissect the physiological and molecular differences between climacteric and non-climacteric melons. We report on the genetic control of the climacteric burst, ethylene synthesis, and associated response. This genetic study was conducted on segregating populations between a non-climacteric melon (PI 161375) and a climacteric variety (Védrantais). To exploit natural variation, molecular genetic mapping offers powerful tools to define major genes and quantitative trait loci (QTL) involved in complex traits (Tanksley and McCouch, 1997). Immortalized lines such as recombinant inbred lines (RILs; Burr and Burr, 1991) are choice material because trial replications in different environments and for different characters can be achieved. Here, we show that variation in climacteric versus non-climacteric ripening is due to two loci.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

The PI 161375 Fruit That Does Not Abscise Is Non-Climacteric

As observed previously (Guis et al., 1997), Védrantais melon fruit abscised and displayed a characteristic peak of ethylene production (Fig. 1A) and ACC content (Fig. 1C) 37 d after pollination (DAP). This was associated with a peak of ACS activity 36 DAP (Fig. 1B). In contrast, PI 161375 fruit did not abscise up to 60 to 80 DAP and did not display any peak of ethylene production, ACC content, or ACS activity (Fig. 1, A-C). The PI 161375 fruit remained firm (6 kg cm²) during the same period, whereas firmness had decreased down to 2 kg cm² in Védrantais at 42 DAP (Fig. 1D).

View larger version (28K): [in this window] [in a new window]   Figure 1.   Comparison of internal ethylene concentration (A; in microliters per liter), ACS activity (B), ACC content (C), and flesh firmness (D) in melon fruit of climacteric (Védrantais, ) and non-climacteric (PI 161375, ) genotypes.

Ethylene production and all ethylene-dependent events can be prematurely induced in climacteric fruit by exogenous ethylene or its analog propylene (McMurchie et al., 1972). In Védrantais melon, such treatment can trigger ethylene production and cell separation in the fruit abscission zone (Guis et al., 1997) but not in PI 161375 fruit (not shown). Figure 2A shows that propylene treatment did not induce endogenous ethylene production in PI 161375 fruit. Moreover, neither firmness nor rind color were significantly different between air- or propylene-treated fruit (Fig. 2, B and C). In contrast, a high level of ethylene was produced after wounding in PI 161375 fruit (data not shown) as in Védrantais (Bouquin et al., 1997). Taken together, these results demonstrate that PI 161375 melon fruit can be considered non-climacteric, although they can synthesize high amounts of ethylene after wounding.

View larger version (15K): [in this window] [in a new window]   Figure 2.   Effect of exogenous propylene on PI 161375 melon fruit harvested 45 DAP for 1 or 6 d on ethylene production (A), flesh firmness (B), and color of the peel (C). Dashed line, Fruit in air; solid line, fruit exposed to propylene.

Comparison of the Expression of Genes between Védrantais and PI 161375 Fruit

We examined the expression of ripening- or ethylene-related melon genes previously isolated, such as ACO1, ERS1, PG1, PG2, RM4, RM5, RM7, RM8, RM11, and RM16, or isolated during this study (EIL1 and EIL2). The expression level of PG2, ERS1, EIL1, EIL2, RM7, and RM8 was similar in Védrantais (just before the climacteric crisis) and PI 161375 (Fig. 3A). However, it is noticeable that the decrease in expression of PG2, ERS1, EIL2, RM7, and RM8 observed in Védrantais fruit after d 37 or 38, i.e. after the climacteric ethylene (Fig. 1A), did not occur in PI 161375 fruit (Fig. 3A). Two of these genes (RM7 and RM8) were found to be expressed at lower levels in transgenic (antisense ACO) Védrantais with 99% inhibition of ethylene production but were not up-regulated by exogenous ethylene (Hadfield et al., 2000). We checked their regulation by propylene in PI 161375 fruit harvested 45 DAP (Fig. 3B) and continuously exposed to air supplemented or not with propylene. The expression of none of these genes was different in air- or propylene-treated PI 161375 fruit.

View larger version (37K): [in this window] [in a new window]   Figure 3.   RNA analysis of genes whose expression in PI 161375 fruit is at the same level as in Védrantais fruit before the climacteric crisis. Phosphor imager data of RNA gel-blot analysis of PG2, ERS1, RM7, RM8, EIL1, and EIL2 in developing Védrantais and PI 161375 melon fruit expressed for each individual blot in terms of percentage of the maximum signal detected. A, Comparison of gene expression between Védrantais and PI 161375 melon fruit during ripening on the vine. Tissues were from seven stages of fruit development (30, 35, 36, 37, 38, 39, and 41 DAP) for Védrantais climacteric line or from 6 stages (30, 35, 40, 45, 50, and 60 DAP) for the non-climacteric PI 161375 line. B, Comparison of gene expression in harvested PI 161375 fruit (45 DAP) treated with air or 2,000 µL L1 of propylene for 1 to 6 d.

A second group of genes were expressed at low levels in PI 161375 fruit (Fig. 4A). Most genes of this group had a maximum expression level at the climacteric crisis in Védrantais fruit (Fig. 4A). This group includes the ethylene biosynthetic genes ACO1 and ACS1. The latter was slightly detected in Védrantais and not at all in PI 161375 fruit (data not shown). RM2 and PG3 transcripts were not detected in PI 161375. Again, regulation by ethylene of the expression of these genes was examined in PI 161375 fruit and no propylene-induced changes were observed (Fig. 4B). Propylene treatment performed on older PI 161375 fruit (50 DAP) yielded identical results (not shown). Wound-induced ethylene production, which is largely independent of ethylene action, correlated with strong stimulation of ACO1 and ACS1 expression in both PI 161375 and Védrantais fruit (data not shown; Bouquin et al., 1997).

View larger version (22K): [in this window] [in a new window]   Figure 4.   RNA analysis of genes whose expression in PI 161375 fruit is lower than in Védrantais fruit before the climacteric crisis. Legend is identical to Figure 3.

Seedling Tissues of the PI 161375 Display the Triple Response Induced by Ethylene

Both PI 161375 and Védrantais seedlings displayed the triple response (Fig. 5). In both genotypes, ethylene inhibited root and shoot elongation and induced radial swelling. However, some differences were noted: (a) root elongation was partially inhibited by ethylene in PI 161375, and root hair formation was completely inhibited; and (b) the apical hook in ethylene treated Védrantais seedlings did not display the exaggerated curvature that is seen in PI 161375 (Fig. 5). There was almost no hook curvature at any stage in Védrantais seedlings grown in air (data not shown), contrary to what is observed in Arabidopsis plants.

View larger version (106K): [in this window] [in a new window]   Figure 5.   Comparison of ethylene-induced triple response between Védrantais (A) and PI 161375 (B) etiolated seedlings. An enlargement of the apical region in ethylene-treated seedlings is shown in the right corners of A and B.

Two Loci Control the Fruit Abscission Phenotype and Are Localized in Two Different Linkage Groups (LG)

F₁ (Védrantais × PI 161375) fruits displayed the Védrantais phenotype, i.e. they abscised and displayed a peak of ethylene during ripening (data not shown). This indicates that both phenotypes are dominant over the phenotypes of nonabscission and absence of peak of ethylene. On a population of 111 RILs derived from the Védrantais × PI 161375 cross, the observed distribution of the fruit abscission phenotype was consistent with a 3:1 distribution (Table I), which corresponded to the segregation of two independent genes coding for the same function. We propose the names abscission layer-3 and abscission layer-4 (Al-3 and Al-4) for these genes.

A logarithm-of-odds (LOD) score mapping method developed to localize Al-3 and Al-4 on the melon genome (Périn, 2000) led to the identification of only two genomic regions with an LOD score above the threshold 2.0. Al-3 and Al-4 genes were mapped close to the amplified fragment-length polymorphism (AFLP) markers H33/M43_21 (LOD 3.79) on LGVIII and H36/M37_11a (LOD 3.09) on LGIX, respectively (Fig. 6).

View larger version (39K): [in this window] [in a new window]   Figure 6.   Mapping of the genes ech, Al-3, and Al-4; the QTLs eth1.1, eth2.1, eth3.1, and eth11.1; and some candidate genes (in bold on the left side of each LG) on the composite map of melon. All of them were mapped on a population of RILs generated between Védrantais, a climacteric line, and PI 161375, a non-climacteric line, with the exception of the ERS1 locus, which was mapped on a RIL population derived between Védrantais and PI 414723.

Al-3 and Al-4 Control the Presence of (Climacteric) Autocatalytic Ethylene Production

Sixty-six RILs were chosen at random, and the ethylene production during fruit ripening was measured. Fruit abscission was always found to be associated with the peak of ethylene in fruit (internal ethylene = 31 ± 20 µL L¹), whereas the absence of abscission was always correlated with the lack of ethylene burst (mean value = 2 ± 2 µL L¹). Moreover, in all nonabscising lines, propylene treatment failed to induce fruit abscission and ethylene production (data not shown).

QTLs Controlling the Level of Climacteric Ethylene Production Are Not Linked to Al-3 and Al-4

We used ethylene measurements observed among 43 climacteric RILs during fruit ripening for quantitative trait analysis of the internal ethylene maximum in fruit. Four QTLs were detected at an LOD score threshold of 2.0 (Table II) and were localized on LGI, LGII, LGIII, and LGXI of the composite map (Fig. 6). Each of them explained the same phenotypical variance explained (PVE) of about 9 µL L¹ maximum of fruit internal ethylene. PI 161375 alleles eth1.1 and eth11.1 increased internal ethylene even if PI 161375 fruits did not produce ethylene during ripening.

The Difference for the Apical Hook Curvature Is under Monogenic Control

We further analyzed the genetic control of the apical hook curvature. The F₁ (Védrantais × PI 161375) seedlings displayed the Védrantais phenotype, i.e. absence of hook curvature in air-grown seedlings and simple curvature in ethylene-treated seedlings. We examined the distribution of the hook phenotype under ethylene on the RIL and back-cross (BC₁; Védrantais × PI 161375) × PI 161375 populations and found that it was fully compatible with a monogenic control (Table I). On the BC₁ population, one-half of the population tested displayed the exaggerated hook curvature, whereas the other half displayed the phenotype of the Védrantais parent. We propose the name exaggerated curvature of the hook (ech) for this locus, which was localized on the LGI on the melon reference map (Périn et al., 2000) close to the E40/M34_1 AFLP marker. The same genomic region comprised the QTL eth1.1.

An Ethylene Receptor Gene ERS1 Was Tightly Linked to eth1.1 and ech

We tried to map several melon sequences that are homologs to Arabidopsis genes encoding elements of the ethylene signal transduction pathway, such as ETR1 or ERS1 (Bleecker et al., 1998; Hua and Meyerowitz, 1998), CTR1, related to MAP kinase genes (Kieber et al., 1993), and EIN3 (Chao et al., 1997). Unfortunately, the available polymorphism in the two populations allowed mapping of only melon ERS1 to LGI at a single locus near the AFLP marker E42/M35_16a. This region also contained ech and eth1.1. Among ethylene biosynthetic genes, ACS5 was mapped to LGIII, ACO1 and ACS1 to LGV, and ACO2 to LGVIII (Fig. 6). None of these genes colocalizes with Al-3 or Al-4.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

The Non-Climacteric Phenotype of PI 161375 Is Associated with an Alteration of Ethylene Perception, Limited to Fruit Tissue

Non-climacteric fruit are defined by the absence of ethylene burst, which may correspond to the inability to synthesize ethylene autocatalytically either because synthesis, as in anti-sense tomato (Oeller et al., 1991; Picton, 1993) and melon (Ayub et al., 1996) fruit, or perception, as in spontaneous Nr tomato mutant (Wilkinson et al., 1995), is blocked. As an alternative, the non-climacteric phenotype may be due to the alteration of early steps of the ripening process as in rin, nor, and Cnr tomato mutants (Giovannoni, 2001).

In melon fruit, ethylene-dependent events during fruit ripening have been defined, in particular using antisense ACC oxidase transgenic lines (Ayub et al., 1996; Guis et al., 1999; Hadfield et al., 2000). Events that are dependent on high levels of ethylene production have been defined, whereas others are either ethylene independent or extremely sensitive to low levels of ethylene (Hadfield et al., 2000). All events of the first category are absent in PI 161375 fruit. The non-climacteric phenotype of PI 161375 was confirmed by treating fruit with exogenous ethylene or propylene; fruit firmness, rind color, ethylene synthesis, ACC content, and expression of known targeted genes were not stimulated. Hence, ethylene-regulated genes (RM5, RM11, PG1, PG3, and ACO1) were all expressed at low levels or undetected and were never induced by exogenous ethylene/propylene.

This makes PI 161375 different from the non-climacteric Earl's Favourite melon in which ethylene treatment increases ACO1 gene expression (Shiomi et al., 1999). Both ACO1 and ACS1 gene expression were strongly stimulated in all genotypes by wounding (Shiomi et al., 1999; this report). In the case of ACO1, promoter analysis led to the identification of two clearly separated cis-regulatory regions, one necessary for wounding response and the other for ethylene response (Bouquin et al., 1997). Thus, wound-induced signaling proceeds independently of ethylene-induced signaling. The fruit-specific inhibition of ethylene perception in PI 161375 is mediated by a different mechanism than that present in Nr tomato because the dominant mutation in the NR (ERS1) ethylene receptor ethylene response blocks ethylene response in all Nr tissues (Lanahan et al., 1994; Wilkinson et al., 1995; Yen et al., 1995). Besides the Nr tomato mutant already mentioned, other known non-climacteric mutant fruit are different. For instance, the rin tomato mutant fruit retains the capacity to display ethylene-stimulated ACO gene expression (Knapp et al., 1989; Shiomi et al., 1999) among other responses. In this report, we have shown that in the non-climacteric chinensis melon, ethylene perception is highlyif not completelyinhibited.

Al-3 and Al-4 Proteins Are Most Likely Fruit-Specific Elements Necessary for Climacteric Fruit Phenotype and Complete Ethylene Response

The redundant genes Al-3 and Al-4 identified by genetic analysis are controlling ethylene-dependent abscission and ethylene production in fruit. Two unlinked genes named Al-1 and Al-2 were reported previously to control fruit abscission in another population of melon (Takada et al., 1975). We have not performed an allelism test, and the map location of Al-1 and Al-2 and their involvement in fruit ethylene production and sensitivity remain unknown.

Two other features observed in PI 161375 and the RILs provide additional clues for Al-3 and Al-4. First, the PI 161375 allelic forms are recessive, and one Védrantais allele at any of these loci confers normal ethylene perception. The Al-3 or Al-4 loci had no major effect on the level of ethylene production during fruit ripening. The high level of variation found among the set of climacteric RI lines is under a distinct genetic control with four different QTLs modulating fruit ethylene production. Second, both recessive alleles confer ethylene insensitivity in fruit tissues only. A partial ethylene insensitivity in PI 161375 roots was observed in abscising (one Védrantais allele at any locus) and nonabscising (al-3/al-3 and al-4/al-4) RILs (data not shown).

These results are in favor of the hypothesis that ethylene perception is faulty in PI 161375 and in non-climacteric RIL fruits but could be also attributable to the absence of some ethylene-independent developmental regulation in PI 161375 fruit (McCourt, 1999). Thus, distinction between climacteric/non-climacteric and ripening/non-ripening fruit is not significant to characterize a ripening mutant in melon as for other climacteric species. For instance, the tomato ripening mutants rin and nor are non-climacteric but are part of a developmental control upstream ethylene signaling (Giovannoni, 2001).

Cosegregation of QTL eth1.1 with ech Raises the Possibility That Regulatory Networks for Ethylene Biosynthesis or Response Are Operating in Multiple Tissues

The presence of the ethylene burst in melon fruit is controlled by two independent loci. However, at least four modifier loci (QTL) localized on other genomic regions are controlling its maximum intensity. None of the ethylene biosynthetic melon genes studied (ACO1, ACO2, ACS1, and ACS5) were linked to any QTL. On the other hand, the locus controlling the curvature of the hook was localized close to the QTL eth1.1 and to the gene encoding a member of the ethylene receptor family ERS1. This colocalization is still approximate and does not exclude the examination of other candidate genes for eth1.1 or ech (Raz and Ecker, 1999; Bleecker and Kende, 2000). Nevertheless, the hypothesis that ERS1, eth1.1, and/or ech are the same gene fits well with what we know of ethylene receptors if one assumes that PI 161375 ERS1 is a loss-of-function allele. One way to test the role of ERS1 in the melon hook formation and fruit ethylene production could be to compare the effect of the dominant allele (Védrantais in the hook, PI 161375 in fruit) in transgenic plants.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Plant Material

All physiological studies were performed on 144 RILs resulting from the cross between melon (Cucumis melo) types Védrantais and PI 161375 (Périn, 2000). A second population of 64 RILs issued from the cross between Védrantais and PI 414723 was used to map some candidate genes. The three parents belong to very different subspecies and cultigroup (Pitrat et al., 2000): Védrantais (released by Vilmorin S.A.) is a commercial variety of the Charentais type (C. melo subsp. melo, cantalupensis group); PI 161375 (C. melo subsp. agrestis, chinensis group) and PI 414723 (C. melo subsp. agrestis, momordica group) corresponded originally to two cultivars collected in Korea and India, respectively, and that have been maintained by self-pollination. To check the monogenic inheritance for hook phenotype in etiolated seedlings, we also used 200 individuals of the BC₁ population (Védrantais × PI 161375) × PI 161375.

Genetic Mapping for Monogenic Traits and Candidate Genes

A saturated map with molecular markers was developed by merging the two maps obtained on the two RIL populations (Védrantais × PI 161375) and (Védrantais × PI 414723; Périn, 2000).

Genetic linkage analysis was performed using Mapmaker software 3.0 (Lander et al., 1987). Monogenic traits were mapped through the Mapmaker commands "build" and "try" using the framework map. Distances (in centiMorgans) were calculated with the Kosambi (1944) function. Candidate genes were mapped by RFLP or PCR on the two populations used for composite map construction.

Mapping of Al-3 and Al-4

We derived a maximum likelihood equation for linkage between duplicated genes and molecular markers on a RIL population according to Fisher procedure. A maximum likelihood ratio was evaluated for each marker of the framework map published under the hypothesis of linkage versus independence (Périn, 2000). A threshold value of 2.0 was chosen to declare linkage of an abscission layer (Al-3 or Al-4) gene and a molecular marker. The highest values were used to localize Al-3 and Al-4 on the composite map.

QTL Detection

QTL search for ethylene production by the fruits was performed using three methods: single-factor analysis of variance, simple interval mapping (Lander et al., 1987) and composite interval mapping (CIM) with QTLCartographer software (Zeng, 1994). A threshold value of 2.0 was used for QTL detection. Only CIM data were used to estimate the genetic effect of the QTLs detected and to localize them on the composite map.

Ethylene and Propylene Treatments

Seeds were surface-sterilized in 30% (v/v) ethanol for 2 min then in 5% (w/v) hypochlorite solution for 5 to 8 min, rinsed abundantly in sterile water, and placed on half-concentrated Murashige and Skoog macronutrients medium supplemented with 0.8% (w/v) agar in boxes (Magenta, Chicago). Growth was performed in small compartments placed in darkness with an air flux containing 10 µL L¹ ethylene or not. After 5 to 10 d at 28°C, seedlings were observed. Fruits were treated at room temperature under a similar gas flow, except that ethylene was eventually replaced by 1,000 µL L¹ propylene. After harvest at 40, 50, and 60 DAP, fruits were placed in jars with air or air plus ethylene (propylene) flow. Firmness, ethylene production, and flesh and peel color were measured as previously described (Guis et al., 1997).

Molecular Techniques

DNA (10 µg) was digested with EcoRI, EcoRV, and HindIII enzymes and blotted. Southern and northern hybridization procedures were performed using standard protocols (Sambrook et al., 1989) and ³²P-labeled probes. Total RNA was extracted from fruit essentially as described by Ayub et al. (1996) and separated by electrophoresis on formaldehyde agarose gels (Sambrook et al., 1989). RNA blots were probed with either the entire sequence of melon ACO1 (Lasserre et al., 1996), ERS1 (Sato-Nara et al., 1999), PG1, PG2, PG3 (Hadfield et al., 1998), RM2, RM4, RM5, RM7, RM8, RM11, RM16 (Hadfield et al., 2000), ACS5 (J.-M. Lelièvre, unpublished data), or partial sequence for ACS1 (Yamamoto et al., 1995). Total fruit cDNA was used for a PCR-based amplification of partial sequences of CTR1-like and EIN3-like cDNAs using degenerate primers coding for conserved regions of the proteins. The CTR1-like and the EIN3-like sequences have been deposited under the GenBank code numbers AF387794 and AF387795, respectively. This EIN3-like sequence detected three transcripts in fruits of both genotypes (data not shown) with two expression patterns that were named EIL1 and EIL2.

Quantitative dosage of radioactive signals hybridizing with RNA was achieved with an Ambis 100. Signals were normalized according to the signal obtained with the DNA probe from the 18S rRNA gene of squash (Cucurbita pepo; Torres-Ruiz and Hemleben, 1994).