INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Grape (Vitis vinifera L.) berries undergo considerable physical and biochemical changes as they develop, particularly during the ripening process. Grapes are considered to be a non-climacteric fruit, and berry development can be divided into three phases on the basis of berry growth. After fruit set, there is an initial phase of cell division (Harris et al., 1968) and cell expansion that results in rapid growth of the berry. This is followed by a lag phase during which berries do not increase in size. Following the lag phase, there is a second phase of berry growth during which ripening occurs (Coombe, 1992). The commencement of ripening is known as "véraison" by viticulturists. Ripening is characterized by a number of changes, including the degradation of chlorophyll, an increase in berry deformability, a rapid increase in the level of hexoses in the berry vacuole, an increase in berry volume, the catabolism of organic acids, the development of skin color (in red grapes), and the formation of compounds important for flavor and aroma. During this first phase of berry expansion, indole-3-acetic acid levels are elevated (Cawthon and Morris, 1982). As the auxin levels are low after véraison and cell division has ended, the second phase of expansion would appear to have a different biochemical basis. Not only does considerable cell expansion occur in this second phase of berry size increase, but this is also the period when berry softening occurs.

A relationship between fruit ripening and changes in mRNA levels has been demonstrated in grape berries by Boss et al. (1996), who showed that the accumulation of transcripts of genes in the flavonoid synthesis pathway was related to anthocyanin production in the berry skin during ripening. In many other fruit, the considerable changes that occur during ripening are also largely the result of changes in gene transcript levels. For example, much is known about the changes in mRNA levels that occur during the ethylene-driven ripening of climacteric fruit, particularly tomato. Screening of cDNA libraries from ripening wild-type and mutant tomato fruit has enabled the identification of a large number of cDNAs associated with the ripening process (Gray et al., 1992; Picton et al., 1993). This has lead to an enhanced understanding of tomato ripening and has allowed the development of transgenic plants with altered ripening characteristics (Gray et al., 1994). Only recently have researchers begun using similar techniques to investigate the molecular biology of ripening in non-climacteric fruit such as strawberry (Medina-Escobar et al., 1997; Manning, 1998) and pepper (Proust et al., 1996). The isolation of ripening-enhanced cDNAs from strawberry (Manning, 1998) and black currant (Woodhead et al., 1998) by differential screening has demonstrated that this technique will be useful in the study of ripening in non-climacteric fruit.

We used the differential screening technique to isolate "Grip" (grape ripening-induced) cDNAs from a ripening grape berry cDNA library. A number of differentially expressed clones were isolated and their sequences and expression patterns in grape tissues analyzed, and their possible function during ripening is discussed.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Tissue Collection, Measurement of Ripening Parameters, and Isolation of RNA

Berries of grape (Vitis vinifera L. cv Shiraz) were sampled at 2-week intervals, beginning at flowering, during the 1995/1996 growing season from 20-year-old vines grown at a commercial vineyard in Willunga, South Australia. Deformability measurements of a number of berries were taken every 2 weeks using a Harpenden skinfold calliper gauge (British Indicators, Burgess Hill, West Sussex, UK) as described by Coombe and Bishop (1980). A randomly chosen sample was weighed and measured for total soluble solids (degrees Brix) with a refractometer (model 10430, Reichert, Vienna). Berries were picked, deseeded (except for the 2 weeks postflowering [wpf] sample), and the flesh (retaining the skins) and seeds were immediately frozen in liquid nitrogen and stored at 80°C until required. For the separation of skin from flesh, the skin was removed from frozen berries that had been partially thawed, and skin and flesh samples were extracted for RNA as described below. Leaf (three stages), root, and flower tissue were similarly frozen and stored.

Total RNA was extracted from the various tissues by the perchlorate method as described by Davies and Robinson (1996). This RNA was used in northern-blot analysis. For cDNA library production, mRNA was purified from total RNA using a mRNA isolation kit (PolyATract, Promega, Madison, WI) according to the manufacturer's instructions.

Preparation of Shiraz Berry cDNA Library

cDNA was prepared using a cDNA synthesis system (SuperScript Choice, Life Technologies/Gibco-BRL, Cleveland) according to the manufacturer's instructions using 5 µg of mRNA from 10 wpf, deseeded Shiraz berries. EcoRI adaptors were ligated to the cDNA, and the resultant fragments were cloned into predigested Lambda ZAP II/EcoRI/CIAP vector and packaged (Gigapack II Gold, Stratagene, La Jolla, CA). Titering, amplification, and screening of the library (with the appropriate probes, see below) were carried out as described in the manufacturer's instruction manual except for the hybridization conditions.

Differential Screening of cDNA Library

Duplicate lifts using Hybond N membrane (Amersham-Pharmacia Biotech, Uppsala) were hybridized with either the pre-véraison probe (made from 6-wpf cDNA, see below) or the post-véraison probe (made from 10-wpf cDNA). The filters were hybridized and washed as described by Davies and Robinson (1996). For the preparation of 6- and 10-wpf probes, 2 µg of mRNA from the 6- and 10-wpf samples was used to produce first-strand cDNA that was then A-tailed using terminal transferase and amplified by PCR using an oligo-dT primer (Chevalier et al., 1995). The mix of fragments thus generated was then labeled by random primer labeling (GIGAprime DNA labeling kit, Bresatec, Adelaide, South Australia), and unincorporated label was removed using a Sephadex G-50 column. Plaques showing hybridization with the 10-wpf but not the 6-wpf probe were isolated as a single species by secondary and (where required) tertiary screening. These single clones were then rescued into pBluescript plasmid vector as described by the manufacturer's instructions (Stratagene). Purified plasmid DNA of the rescued clones was digested with EcoRI endonuclease to determine the insert size. Selected clones were then sequenced by automated sequencing using T3 and T7 oligonucleotide primers and, where appropriate, specifically designed internal primers.

To reduce problems caused by the abundance of Grip 3 and 4 sequences, which predominated the early screening attempts, these sequences were selectively removed from the 10-wpf probe as follows. First-strand cDNA was prepared as described above. After the terminal transferase reaction, the fragments were ethanol precipitated, washed, dried, and taken up in 100 µL of water. A synthetic oligonucleotide with a 5' biotin group attached was designed to a sequence common to both the Grip 3 and 4 sequences (CAAGGCTCCACCACCCATCC). Twenty microliters of 60 pM/µL primer was mixed with 100 µL of washed streptavidin paramagnetic particles (MagnaSphere, Promega) and 80 µL of 10× SSC and incubated at 25°C for 15 min. The beads were then washed three times with 5× SSC and resuspended in 60 µL of 10× SSC. One-hundred microliters of the first-strand cDNA was heated at 95°C for 10 min, snap-cooled on ice, and mixed with 30 µL of the probe/bead suspension. SSC was added to a final concentration of 3.5×. This mixture was incubated with gentle agitation at 38°C for 30 min. The beads were then captured and the cDNA was precipitated from the supernatant by ethanol precipitation, washed, dried, taken up in water, and then amplified and labeled as described above.

Sequence Analysis

Initially BLASTX and BLASTP programs (Genetics Computer Group, Madison, WI) were used to identify the sequences from other plant species and yeasts most closely related to the differentially screened grape clones. The degree of similarity between two sequences was calculated using the GAP program, and the theoretical pI was calculated using the ISOELECTRIC program (both from Genetics Computer Group). Sequence-property-based searches of protein databases were done using the PROPSEARCH program of Hobohn and Sander (1995).

Northern Analysis and Probe Preparation

Northern blotting, hybridization, and washing were done as described by Davies and Robinson (1996). ³²P-Labeled probes to the various genes were prepared by random primer labeling (GIGAprime DNA labeling kit, Bresatec, Adelaide, South Australia) of the appropriate excised fragments.

Cloning of a Putative Cytochrome P450

A cDNA fragment encoding part of a putative cytochrome P450 enzyme was generated by chance by PCR using degenerate oligonucleotide primers designed to amplify plant glucanases (A. Jacobs, personal communication). This clone was used as a probe to isolate a full-length clone (gfh2) from the 10-wpf Shiraz cDNA library in Lambda Zap II. This cDNA clone was used to probe northern blots.

Cloning of the Thaumatin-Like VvTL2 cDNA

A partial clone of VvTL2 (homologous to the thaumatin-like clone from Sultanina [Loulakakis, 1997]) was cloned serendipitously by PCR from Shiraz 10-wpf cDNA (see above for the preparation of cDNA). The primers used were WHP5F: 5'-TGATTCAGGTAGCGGCAGT-3' and B25 (Frohman et al., 1988). A 670-bp subclone was isolated and used to probe northern blots.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Differential Screening of a Ripening Shiraz Berry Library

The initial differential screening was conducted using a cDNA library made from poly(A⁺) RNA from Shiraz berries harvested 10 wpf (2 weeks after véraison). Duplicate lifts of this library were screened using either cDNA made from 6-wpf (2 weeks before véraison) berry RNA as a probe or cDNA made from RNA extracted from 10-wpf berry tissue. Autoradiographs of these filters showed many plaques that hybridized strongly with the 10-wpf, but not with the 6-wpf, probe (data not shown). The pattern of hybridization indicated that there is a change in the accumulation of transcripts of a number of genes during ripening. A number of these plaques were isolated as single species and rescued into pBluescript. DNA endonuclease digestion and DNA sequencing showed that the majority of these clones (11 of 19) comprised two very closely related sequences, Grip 3 and 4 (Table I). Although these cDNAs were apparently ripening related and therefore of interest to us, their dominance made screening for other ripening-related clones more difficult. To circumvent this problem, a biotinylated synthetic oligonucleotide designed to the Grip 3 and 4 sequences was used to subtract these sequences from the 10-wpf probe. Subsequent differential screening demonstrated that this approach was successful in reducing the number of Grip 3 and 4 clones isolated. Selected clones were sequenced and unique sequences were compared with the database sequences to identify homologs in other species and thereby determine possible functions.

Transcript Accumulation Patterns of Ripening-Associated Genes

The cDNA clones described above were used to probe northern blots prepared with RNA from a number of tissues, i.e. roots, leaves at three stages of development, seeds, flowers, and a developmental series of berry samples. In the Shiraz samples used for RNA extraction the first signs of ripening, as judged by an increase in berry deformablity and hexose accumulation, occurred between 8 and 10 wpf (Davies et al., 1997). A number of cDNAs whose corresponding mRNAs accumulated during berry development were identified. Some of these had predominantly berry- and ripening-specific patterns of accumulation. The steady-state mRNA levels of others increased during ripening but were also present in other tissues.

The clones that exhibited predominantly berry- and ripening-specific mRNA accumulation patterns included Grip 3, 4, 13, 15, 22, 28, 32, 51 (VvTL1), VvTL2, and Grip 61 (Fig. 1A). These clones can be further divided into two groups on the basis of the patterns of mRNA accumulation. The transcript levels for Grip 3, 4, 13, and 15 appear to be maximal early in ripening, because they are highest in the 10-wpf sample and then begin to decrease significantly. In contrast, the remaining Grip sequences shown in Figure 1A (Grip 22, 28, 32, 51, 61, and VvTL2) increase in transcript level later during ripening (12 wpf).

View larger version (46K): [in this window] [in a new window]   Figure 1.   Northern-blot analysis of the expression of Grip clones (and the gfh2 and VvTL2 clones) in various grape tissues. RNA from root (R), young leaf (Y), mid leaf (M), old leaf (O), seed from berries 4 wpf (S), flower (F), and a series of samples taken from developing berries at two weekly intervals (commencing at 2 wpf) were probed with the cDNAs as indicated. The dashed line indicates véraison. A, Genes exhibiting berry- and ripening-specific expression. B, Genes with up-regulated expression during ripening but also expressed in other tissues. CW, Clones encoding putative cell wall proteins. The bottom panel is a photograph of an ethidium-bromide-stained gel to show the intactness and relative loadings of the RNA samples used in the northern analysis.

The transcripts of the remainder of the Grip clones isolated (Fig. 1B) do not accumulate in a fruit- or ripening-specific manner; however, the levels of transcript are significantly increased during ripening. There is a diverse range of expression patterns displayed. For example, the Grip 21 message is present in all tissues at varying levels, while the Grip 24 sequence is only readily detectable in berries, seeds, and older leaves.

Properties of the Putative Ripening-Related cDNAs

General features of the grape clones, including their nearest match with other sequences in the databases, are given in Tables I and II. The values for percentage identity between the grape sequences and their homologs in other species vary considerably, so caution must be used in interpreting matches at the lower end of the range.

Putative Cell Wall Proteins

Grip 3 and 4

Grip 13 and 15

Grip 28

Grip 31 and 68

Putative Stress-Induced Proteins

Grip 21

Grip 22

Grip 24

Grip 32

Thaumatin-Like Proteins

Grip 51 (VvTL1) and VvTL2

Grip 55

Grip 58

Grip 61

Gfh2

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

It is apparent from the differential screening analysis presented in this work that there is a dramatic change in the mRNA population in grape berries as they enter ripening. The transcript levels of a range of genes increase at véraison, including, in addition to the cDNAs described in this paper, genes encoding alcohol dehydrogenase, chitinase, thaumatin-like proteins, and anthocyanin synthesis pathway enzymes (Boss et al., 1996; Robinson et al., 1997; Sarni-Manchado et al., 1997; Tattersall et al., 1997). There are also genes whose transcript levels decrease at around the time of véraison (e.g. the genes encoding the grape putative vacuolar invertases [Davies and Robinson, 1996]). These changes in mRNAs levels are likely to be involved in the physical and metabolic changes that occur during ripening. Similar differences in steady-state mRNA levels have been observed between pre- and post-ripening states in other fruit. An example of this in nonclimacteric fruit is provided by strawberries, in which ripening commences when the auxin concentration decreases below a threshold level (Given et al., 1988). In strawberry, auxins reduce the levels of some mRNAs and proteins and increase the levels of others (Veluthambi and Poovaiah, 1984; Reddy and Poovaiah, 1990; Manning, 1994; Manning, 1998).

The different patterns of mRNA accumulation observed by northern analysis (Fig. 1) suggest that these genes are under a range of regulatory controls. Many of the sequences isolated by the differential screening were present at high levels in ripening fruit (as shown by northern-blot analysis and the frequency of cloning during library screening). The ripening-associated P/HPRGs Grip 3 and 4 sequences, for example, are major components of the profile of mRNAs present during berry ripening. Transcript levels for some cDNAs were much lower (as assayed by northern analysis) and were isolated only once during the differential screening procedure (e.g. Grip 55 and 58). This suggests that although the methodology used was more likely to isolate more populous ripening-associated species, it did not exclude the isolation of less-prevalent cDNAs. Further studies using techniques such as differential display may identify ripening-induced cDNAs present at lower levels.

The putative Grip proteins described in this paper were divided into two general groups based on their proposed functions in the berry. One group consists of proteins that may be involved in cell wall structure and includes members of the P/HRGP family, a diverse group of proteins, many of which do not have proven functions but are thought to be involved in providing additional support to the polysaccharide network in the cell wall by the formation of intermolecular cross-links (Sommer-Knudsen et al., 1998). The expression of genes encoding these proteins in other species is variously induced by wounding, by pathogen attack, during normal developmental processes, and during root nodule formation. The proposal that the putative Pro-rich cell wall proteins described in this paper are expressed to high levels after véraison gives further credence to the observations of Nunan et al. (1998). These authors have shown that the protein content of berry mesocarp cell walls increased from approximately 8% by weight early in development to nearly 12% after véraison. Amino acid analysis showed that the levels of Hyp increased dramatically over the same period. A similar increase in the Hyp content of the Shiraz berries used in this study was found using a colorimetric assay (data not shown).

The P/HRGPs produced in the berry at véraison could have two not necessarily exclusive functions. One would be to provide additional strength to the cell walls during rapid expansion of the cells; the other would be to restrict the invasion of pathogens. The berries from which the 10-wpf cDNA library was constructed did not have any visible signs of pathogenic attack, so a direct response to infection seems unlikely. Also, the expression of Grip 3, 4, 13, and 15 in berries only after ripening had commenced suggests that their expression is under developmental control rather than being a response to environmental factors or pathogenic infection. It is possible that the expression of these putative cell wall genes in ripening berries is a developmentally regulated prophylactic measure against pathogenic attack. The existence of such a program in grapes is indicated by the ripening-associated induction of expression of pathogenesis-related genes and the expression of the corresponding proteins in uninfected berries after véraison (Robinson et al., 1997; Tattersall et al., 1997; Jacobs et al., 1999).

The considerable increase in cell volume after véraison (Coombe, 1992) requires the berry cell walls to expand rapidly. Pericarp cell walls do not appear to thicken appreciably during ripening (Hardie et al., 1996) and there are no major changes to the composition of the cell wall polysaccharides during ripening (Nunan et al., 1998). Thus, the expression of the putative P/HPGPs during ripening may be a developmentally controlled part of the normal ripening process designed to hold the cell walls together as they undergo the second phase of rapid expansion and softening.

The remaining Grip putative proteins appear to be non-cell wall proteins likely to be involved in stress response. This is suggested by the proposed functions of their homologs in other plants and by the stimuli that enhance the expression of these homologous genes. Stress due to changes in osmotic potential may occur during low-temperature treatments, pathogen infection, drought, salinity (Bray, 1997), and, as argued below, due to the storage of large amounts of osmotically active substances in storage tissues such as fruit. During berry ripening, cells expand rapidly as they accumulate large amounts of the hexoses, Glc, Fru, and water. This results in considerable changes in osmotic pressure and water potential. Fruit such as grapes, which store particularly high levels of sugars (approximately 20% [w/w] hexoses when fully ripe [Lott and Barrett, 1967]), may be especially vulnerable to the reduction in water activity caused by the high vacuolar levels of sugars. Part of the adjustment to the rapid increase in vacuolar sugar levels may be the synthesis of proteins involved in stress management. A number of cDNAs cloned in this study have homologs implicated in the response to water deficit in other plants.

Some of the putative Grip proteins also have homologs that are expressed during morphogensis in other plants. Neale et al. (1990) concluded that some genes normally associated with the response to pathogens (including a chitinase, a -1,3-glucanse, and an extensin) are involved in normal developmental processes in healthy tobacco plants. Metallothionein-like proteins are expressed in a developmental manner during such processes as leaf senescence (Buchanan-Wollaston, 1994), the ethylene-promoted abscission of leaflets (Coupe et al., 1995), and embryogenesis (Chatthai et al., 1997).

The changes observed in the mRNA population at véraison may be influenced by changes in the levels of plant growth regulators. In grape berries indole acetic acid levels are highest early in fruit development and decrease to low levels during the period leading up to véraison (Cawthon and Morris, 1982). When a synthetic, auxin-like compound is applied to grape berries just prior to véraison, the transcript levels of genes normally expressed prior to véraison are maintained and the accumulation of transcripts for ripening-associated genes is suppressed (Davies et al., 1997). This coincides with a delay in the commencement of ripening and in the usual post-véraison increase in ABA levels. ABA is often associated with the control of expression of stress-related genes (Bray, 1997; Zhu et al., 1997), and many of the cDNAs isolated by differential screening have homologs implicated in the stress response.

In summary, the putative Grip proteins may function to protect tissues that are more at risk of pathogenic infection due to changes in cell wall properties during ripening (and possibly other morphogenic events). Equally important may be the role of some of the putative Grip proteins in protecting ripening berry tissues from the consequences of an important part of the ripening process itself, i.e. the storage of large amounts of hexoses (and water) in the cell vacuole.