Starch granule initiation and growth are altered in barley mutants that lack isoamylase activity

Authors

  • Rachel A. Burton,

    1. 1 Department of Plant Science, University of Adelaide, Waite Campus, Glen Osmond, SA 5064, Australia2 John Innes Centre, Norwich Research Park, Colney, Norfolk NR4 7UH, UK3Institute of Food Research, Norwich Research Park, Colney, Norfolk NR4 7UA, UK
    Search for more papers by this author
  • Helen Jenner,

    1. 1 Department of Plant Science, University of Adelaide, Waite Campus, Glen Osmond, SA 5064, Australia2 John Innes Centre, Norwich Research Park, Colney, Norfolk NR4 7UH, UK3Institute of Food Research, Norwich Research Park, Colney, Norfolk NR4 7UA, UK
    Search for more papers by this author
  • Luke Carrangis,

    1. 1 Department of Plant Science, University of Adelaide, Waite Campus, Glen Osmond, SA 5064, Australia2 John Innes Centre, Norwich Research Park, Colney, Norfolk NR4 7UH, UK3Institute of Food Research, Norwich Research Park, Colney, Norfolk NR4 7UA, UK
    Search for more papers by this author
  • Brendan Fahy,

    1. 1 Department of Plant Science, University of Adelaide, Waite Campus, Glen Osmond, SA 5064, Australia2 John Innes Centre, Norwich Research Park, Colney, Norfolk NR4 7UH, UK3Institute of Food Research, Norwich Research Park, Colney, Norfolk NR4 7UA, UK
    Search for more papers by this author
  • Geoffrey B. Fincher,

    1. 1 Department of Plant Science, University of Adelaide, Waite Campus, Glen Osmond, SA 5064, Australia2 John Innes Centre, Norwich Research Park, Colney, Norfolk NR4 7UH, UK3Institute of Food Research, Norwich Research Park, Colney, Norfolk NR4 7UA, UK
    Search for more papers by this author
  • Chris Hylton,

    1. 1 Department of Plant Science, University of Adelaide, Waite Campus, Glen Osmond, SA 5064, Australia2 John Innes Centre, Norwich Research Park, Colney, Norfolk NR4 7UH, UK3Institute of Food Research, Norwich Research Park, Colney, Norfolk NR4 7UA, UK
    Search for more papers by this author
  • David A. Laurie,

    1. 1 Department of Plant Science, University of Adelaide, Waite Campus, Glen Osmond, SA 5064, Australia2 John Innes Centre, Norwich Research Park, Colney, Norfolk NR4 7UH, UK3Institute of Food Research, Norwich Research Park, Colney, Norfolk NR4 7UA, UK
    Search for more papers by this author
  • Mary Parker,

    1. 1 Department of Plant Science, University of Adelaide, Waite Campus, Glen Osmond, SA 5064, Australia2 John Innes Centre, Norwich Research Park, Colney, Norfolk NR4 7UH, UK3Institute of Food Research, Norwich Research Park, Colney, Norfolk NR4 7UA, UK
    Search for more papers by this author
  • Darren Waite,

    1. 1 Department of Plant Science, University of Adelaide, Waite Campus, Glen Osmond, SA 5064, Australia2 John Innes Centre, Norwich Research Park, Colney, Norfolk NR4 7UH, UK3Institute of Food Research, Norwich Research Park, Colney, Norfolk NR4 7UA, UK
    Search for more papers by this author
  • Sonja Van Wegen,

    1. 1 Department of Plant Science, University of Adelaide, Waite Campus, Glen Osmond, SA 5064, Australia2 John Innes Centre, Norwich Research Park, Colney, Norfolk NR4 7UH, UK3Institute of Food Research, Norwich Research Park, Colney, Norfolk NR4 7UA, UK
    Search for more papers by this author
  • Tamara Verhoeven,

    1. 1 Department of Plant Science, University of Adelaide, Waite Campus, Glen Osmond, SA 5064, Australia2 John Innes Centre, Norwich Research Park, Colney, Norfolk NR4 7UH, UK3Institute of Food Research, Norwich Research Park, Colney, Norfolk NR4 7UA, UK
    Search for more papers by this author
  • Kay Denyer

    Corresponding author
    1. 1 Department of Plant Science, University of Adelaide, Waite Campus, Glen Osmond, SA 5064, Australia2 John Innes Centre, Norwich Research Park, Colney, Norfolk NR4 7UH, UK3Institute of Food Research, Norwich Research Park, Colney, Norfolk NR4 7UA, UK
      *For correspondence (fax +44 1603 450045; e-mail kay.denyer@bbsrc.ac.uk
    Search for more papers by this author

*For correspondence (fax +44 1603 450045; e-mail kay.denyer@bbsrc.ac.uk).

Summary

Two mutant lines of barley, Risø 17 and Notch-2, were found to accumulate phytoglycogen in the grain. Like the sugary mutants of maize and rice, these phytoglycogen-accumulating mutants of barley lack isoamylase activity in the developing endosperm. The mutants were shown to be allelic, and to have lesions in the isoamylase gene, isa1 that account for the absence of this enzyme. As well as causing a reduction in endosperm starch content, the mutations have a profound effect on the structure, number and timing of initiation of starch granules. There are no normal A-type or B-type granules in the mutants. The mutants have a greater number of starch granules per plastid than the wild-type and, particularly in Risø 17, this leads to the appearance of compound starch granules. These results suggest that, as well as suppressing phytoglycogen synthesis, isoamylase in the wild-type endosperm plays a role in determining the number, and hence the form, of starch granules.

Introduction

Normal starch synthesis in plants requires, in addition to starch synthases and starch-branching enzymes, a debranching enzyme (DBE) that cleaves (1,6) α-linkages within amylopectin and related polysaccharides or oligosaccharides. Evidence that DBE is necessary for normal starch synthesis comes from the study of mutant plants and algae lacking enzymes belonging to the isoamylase class of DBE (maize, James et al., 1995; rice, Kubo et al., 1999; Arabidopsis,Zeeman et al., 1998; Chlamydomonas, Mouille et al., 1996). In these mutants, the starch content is lower than normal and there is an accumulation of a soluble (1→4 : 1→6)α-glucan, phytoglycogen. Phytogly cogen does not accumulate in wild-type plants and algae, or in other low-starch mutants.

In the isoamylase (sugary1) mutants of maize and rice, there is also a decrease in the activity of limit dextrinase (LD), another type of DBE (Nakamura et al., 1996; Pan and Nelson, 1984). It has been argued that this decrease, rather than the loss of isoamylase, may be the direct cause of phytoglycogen accumulation. There is an inverse correlation between LD activity and phytoglycogen accumulation in the endosperm of rice mutants carrying sugary1 alleles of different severity (Nakamura et al., 1997). However, LD remains at wild-type levels in the phytoglycogen-accumulating isoamylase mutant of Arabidopsis (dbe1; Zeeman et al., 1998). Thus, it is likely that the isoamylase-type of DBE plays a specific role in the synthesis of starch that cannot be assumed by the LD-type.

Two models have been put forward to explain the role of isoamylase in starch synthesis. The first proposes that isoamylase plays a direct role in the synthesis of amylopectin, the major component of starch granules (Ball et al., 1996; Myers et al., 2000). It is suggested that DBE is required for the synthesis of an amylopectin molecule capable of crystallization from a soluble pre-amylopectin precursor. In the absence of isoamylase, pre-amylopectin is further elaborated by starch synthase and starch-branching enzymes in the stroma to form phytoglycogen. The second model (Zeeman et al., 1998) proposes that isoamylase does not play a direct role in the synthesis of amylopectin. Instead, together with other degradative enzymes, isoamylase degrades soluble α-glucans. In the absence of isoamylase, these accumulate in the form of phytoglycogen along with reduced amounts of normal amylopectin.

The phenotypes of the isoamylase mutants so far described do not provide sufficient information to allow these models to be further evaluated. It is not clear which, if either is correct. To investigate further the role of isoamylase in starch synthesis, we identified two allelic isoamylase mutants of barley, Risø 17 and Notch-2, during a screen for altered starch synthesis in the endosperm of high-lysine mutants of barley. Detailed characterization of starch granule structure and the sequence of the isoamylase genes of these mutants have shed light on the possible role of isoamylase during starch synthesis.

Results

Two low-starch barley mutants accumulate large amounts of soluble α-glucan

The starch and soluble α-glucan contents of mature grains of two wild-type barley cultivars (Bomi and Carlsberg II) and several previously identified low-starch mutants (Balaravi et al., 1976; Bansal, 1970; Doll, 1983) were measured (Figure 1). Soluble α-glucan is material that is soluble in aqueous extraction medium but insoluble in > 60% aqueous methanol. Two of the low-starch mutants, Risø 17 and Notch-2, had much higher soluble α-glucan contents than the wild-type lines and the other mutants.

Figure 1.

Starch and soluble α-glucan contents of mature grains.

α-Glucans were extracted from mature grains by homogenization in water. Starch was purified from the water-insoluble fraction and soluble α-glucan by alcohol precipitation of the water-soluble fraction. Soluble α-glucan is material which is soluble in aqueous extraction medium but insoluble in > 60% aqueous methanol. Data are means ± sd of values from 3 to 4 separate extracts. Bomi and Carlsberg II are wild-type with respect to their starch contents. The other cultivars were previously identified as low-starch mutants.

The degree of branching of the α-glucans in the mature grains of wild-type (Bomi) and mutants (Risø 17 and Notch-2) was compared by measuring the wavelengths of maximum absorbance (λmax) of the α-glucan-iodine complexes. All of the starches had λmax of 550–600 nm, which is typical of starches generally. The λmax of the soluble α-glucan from Bomi was also 550–600 nm suggesting that it had a degree of branching similar to starch. The soluble α-glucans from the mutants had λmax values of 420–430 nm, which indicated that they were more highly branched than starch. These values are in the same range as those of phytoglycogen from maize and Arabidopsis (Zeeman et al., 1998), suggesting strongly that the barley mutants accumulate phytoglycogen.

The soluble α-glucan in the mutants is phytoglycogen

To examine the structure of the soluble α-glucan in more detail and to compare the starches from wild-type and mutants, we determined the relative abundance of chains of different lengths in these α-glucans using fluorophore-assisted gel electrophoresis (O'Shea and Morell, 1996). Starch and soluble α-glucan were extracted from endosperms at five different stages of development. For starch, the profiles obtained reflect the distributions of the short- to medium-length chains of the amylopectin component of starch.

There was little alteration in the chain-length profiles of the starches with developmental age, for either the wild-type or the mutants (Figure 2a, i-iii). The mutant starches had chain-length profiles similar to those of the wild-type starches except that mutant starches from older and mature endosperms had relatively more chains of dp 10–13 (Figure 2b, i-ii).

Figure 2.

Figure 2.

Analysis by gel electrophoresis of the short chains of α-glucans.

α-Glucans were extracted from grains of different developmental ages. The stages of development were defined by the FWs of the grains as follows: ○, > 20 mg; □, 20–40 mg; ▵, 40–70 mg; ▿, 70–100 mg; ◊, mature grains. α-Glucan samples were debranched with isoamylase, derivatized with the fluorophore APTS and subjected to electrophoresis in an Applied Biosystems DNA sequencer. Data were analysed using Genescan software.

(a) Analysis of chain-length profiles. The sum of the areas of peaks corresponding to individual chain lengths between 6 and 28 glucose units was calculated and each peak area was expressed as a percentage of the total peak area. Each value is the mean of two replicate measurements. (i-iii) starch; (iv-vi) soluble α-glucan.

(b) Comparison of chain-length profiles. To compare different α-glucans, for each chain length, the difference between values for the percentage total peak area was calculated.

Figure 2.

Figure 2.

Analysis by gel electrophoresis of the short chains of α-glucans.

α-Glucans were extracted from grains of different developmental ages. The stages of development were defined by the FWs of the grains as follows: ○, > 20 mg; □, 20–40 mg; ▵, 40–70 mg; ▿, 70–100 mg; ◊, mature grains. α-Glucan samples were debranched with isoamylase, derivatized with the fluorophore APTS and subjected to electrophoresis in an Applied Biosystems DNA sequencer. Data were analysed using Genescan software.

(a) Analysis of chain-length profiles. The sum of the areas of peaks corresponding to individual chain lengths between 6 and 28 glucose units was calculated and each peak area was expressed as a percentage of the total peak area. Each value is the mean of two replicate measurements. (i-iii) starch; (iv-vi) soluble α-glucan.

(b) Comparison of chain-length profiles. To compare different α-glucans, for each chain length, the difference between values for the percentage total peak area was calculated.

The chain-length profiles of the soluble α-glucan from the barley mutants were similar to those of phytoglycogen from other species (e.g. maize, Dinges et al., 2001; rice, Nakamura et al., 1997; Arabidopsis, Zeeman et al., 1998). There was an increase in the proportion of short chains (dp 6–9) with age (Figure 2a, v-vi and Figure 2b, iii-iv). At all stages of development, the soluble α-glucans from the mutants had a greater proportion of chains of dp 6–8 (Figure 2b, v-vi) and, particularly in Notch-2, fewer chains of dp 10–15 than their respective starches (Figure 2b, v-vi).

Risø 17 and Notch-2 are allelic variants

To determine whether the recessive mutations in Risø 17 and Notch-2 were in the same gene, we crossed them and examined the F1 progeny. All of the F1 grains were shrivelled, indicating that they had a lower than normal starch content, and the starch granules were irregularly shaped and small. Risø 17 and Notch-2 were crossed with other low-starch cultivars (Risø 13, Risø 16, Risø 527, Risø 1508, Notch-1; Doll, 1983). These crosses gave F1 grains that were all, or mostly all, normal in shape. The starch granules in grains resulting from the cross between Risø 17 and Notch-1 were normal in shape. These results indicate firstly, that the mutations in Risø 17 and Notch-2 lie in the same gene and secondly, that none of the other low-starch cultivars carry mutations that are allelic to those in Risø 17 or Notch-2.

Risø 17 and Notch-2 lack isoamylase activity

Isoamylase activity cannot be quantified in crude extracts due to the absence of a unique substrate for this enzyme. However, isoamylase activity in extracts of several plant species is revealed on native, non-denaturing polyacrylamide gels containing amylopectin. The enzyme appears as a blue-staining band with low mobility when the gels are stained with Lugol solution (Dinges et al., 2001; Kubo et al., 1999; Zeeman et al., 1998). Such a band was observed with extracts of developing wild-type barley endosperms of different developmental ages (Figure 3 and data not shown). A second, faint blue-staining band was sometimes observed in wild-type extracts (not visible in Figure 3). Neither blue band was observed with extracts of endosperms of any age for either Risø 17 or Notch-2. To estimate the minimum isoamylase activity that could be detected using this native gel method, we compared a series of dilutions of the wild-type extract on native gels (not shown). From this, we estimated that the isoamylase activity in the mutants was < 4% of that in the wild-type.

Figure 3.

Isoforms of debranching enzymes in developing endosperms.

Crude, soluble extracts of developing endosperms were loaded onto gels containing amylopectin. Each track contained extract (from 1 mg FW of tissue) and loading medium in a ratio of 5 : 1. After electrophoresis, gels were incubated at 37°C for 16 h at pH 6.0 and stained with iodine solution. The position of the blue-staining band that is due to isoamylase activity is indicated with an arrow. Tracks 1 and 2: Bomi, tracks 3 and 4: Risø 17, tracks 5 and 6: NP113, tracks 7 and 8: Notch-2.

The activities of several other enzymes are affected in the mutants

The activities of many of the enzymes involved in the conversion of sucrose to starch were determined in endosperms from Risø 17 and its parent variety, Bomi and in Notch-2 and the variety from which it was derived, NP113. When the activities in grains of 45–55 mg FW in the mutants and their corresponding wild-types were compared, the activities of many enzymes were not statistically significantly different (Table 1). Some enzymes (Table 1) did show differences but these were not consistently different in both mutants. For example, the activity of soluble starch synthase was higher in Risø 17 than in Bomi but there was no significant difference in the activity of this enzyme between NP113 and Notch-2. Most of the enzymes that differed in activity showed a higher activity in the mutant than in the wild-type. These increases in activities are unlikely to cause or contribute to the decrease in total α-glucan synthesis observed in the mutants. Two enzymes, SBE and alkaline pyrophosphatase, showed lower activity in Risø 17 than in Bomi. These data suggest that the mutations cause pleiotropic effects on the activities of other enzymes in the pathway of starch synthesis. Such pleiotropic effects are common in starch mutants of cereals (maize, Singletary et al., 1997; barley, Schulman and Ahokas, 1990).

Table 1.  Comparison of the maximum catalytic activities of enzymes in crude extracts of developing endosperms
EnzymeActivity (µmol min−1 g−1FW)
BomiRisø 17NP113Notch-2
  1. Developing endosperms from grains of 45–55 mg FW were extracted as described in Experimental procedures and assayed for enzyme activities. Values are means ± se of measurements made on a number of independent extracts (shown in parentheses). Comparison of the activities of each enzyme in Bomi and Risø 17 and in NP113 and Notch-2 was done using Microsoft Excel software (t-test, 2-tailed distribution, 2-sample equal variance). This statistical analysis showed that the activities of soluble starch synthase and ADPG pyrophosphorylase were higher in Risø 17 than in Bomi (P < 0.05), the activities of fructokinase and glucokinase were higher in Notch-2 than in NP113 (P < 0.05) and the activities of SBE and alkaline pyrophosphatase were lower in Risø 17 than in Bomi (P = 0.05). All other pair-wise comparisons showed no statistically significant difference (P > 0.05). ND = not determined.

Sucrose synthase3.39 ± 0.16 [8]3.79 ± 0.53 [5]2.81 ± 0.72 [4]3.80 ± 0.59 [5]
UDPG pyrophosphorylase54.13 ± 4.68 [7]68.87 ± 8.51 [5]41.04 ± 5.10 [5]55.53 ± 11.91 [5]
Fructokinase0.245 ± 0.022 [8]0.311 ± 0.039 [5]0.182 ± 0.021 [5]0.291 ± 0.026 [5]
Glucokinase0.435 ± 0.050 [8]0.523 ± 0.056 [5]0.386 ± 0.034 [5]0.612 ± 0.054 [5]
Phosphoglucomutase19.02 ± 3.92 [8]38.35 ± 10.40 [6]27.51 ± 7.87 [6]28.37 ± 6.94 [6]
Phosphoglucose isomerase19.61 ± 2.26 [8]21.40 ± 1.71 [4]18.68 ± 0.79 [5]22.83 ± 2.88 [5]
ADPG pyrophosphorylase4.44 ± 0.37 [8]6.88 ± 0.97 [6]3.72 ± 0.28 [5]5.25 ± 0.78 [6]
Soluble starch synthase0.216 ± 0.020 [8]0.399 ± 0.026 [4]0.131 ± 0.025 [5]0.118 ± 0.037 [4]
Granule-bound starch synthase0.387 ± 0.048 [8]0.373 ± 0.069 [4]0.252 ± 0.051 [4]0.222 ± 0.074 [3]
Alkaline pyrophosphatase7.34 ± 0.94 [8]4.52 ± 0.64 [5]3.68 ± 0.57 [5]3.58 ± 0.64 [5]
Starch-branching enzyme42.93 ± 2.62 [6]31.54 ± 0.78 [5]NDND

Attempts to assay limit dextrinase (LD) activity in crude extracts of developing barley endosperm were unsuccessful due to the presence of protein inhibitors (Macri et al., 1993). The activity of LD measured in mixed extracts of developing pea embryo and developing barley endosperms was much less than that expected from measurements of the LD activity in these tissues extracted separately (data not shown). Inclusion of a chemical modifier of the LD inhibitor, phenyl glyoxal (MacGregor et al., 2000) did not increase the measurable LD activity in extracts of barley. As an alternative approach to estimate the LD activity, we used native gels similar to those described above to identify isoamylase but containing red-pullulan rather than amylopectin. Red-pullulan is a unique substrate for LD, and is not degraded by isoamylase or any other enzymes. Two bands of LD activity were observed (data not shown). Comparison of the LD activity in extracts of mutant and wild-type developing endosperms did not reveal any consistent differences. This suggested that in the barley isoamylase mutants, unlike those of rice and maize, the LD activity is not decreased relative to that in the wild-type. However, this conclusion must be treated with some caution as we do not know to what extent the LD inhibitors affect the LD activity revealed in these gels.

Cloning and sequencing the isoamylase cDNA

To discover whether the lack of isoamylase activity was due to mutations within the isoamylase gene, PCR was used to amplify the cDNAs encoding isoamylase in the wild-types and mutants. A nested PCR strategy, using the high-fidelity Taq polymerase Elongase, was used to obtain full-length isoamylase cDNAs. The cDNA sequences have been submitted to GenBank under accession numbers AF490375 (Bomi), AF490376 (Notch-2) and AF490377 (Risø 17), and the predicted amino acid sequences are shown in Figure 4a. This is the first published report of the full-length sequence of barley isoamylase cDNA; the sequence previously reported by Sun et al. (1999) is truncated at both the 5′ and 3′ ends. We will refer to this cDNA as isa1.

Figure 4.

Figure 4.

Mutations in the isoamylase genes of Risø 17 and Notch-2.

(a) Predicted amino acid sequences of cDNAs isolated from developing barley endosperms of wild-type (wtype), Bomi, Notch-2 and Risø 17. The putative translation start point has not been determined experimentally and may be any of the methionines located at positions 1, 3, 4 or 6. The deletion in Risø 17 is indicated by the dotted line. All deviations from the wild-type sequence are underlined and in bold. The sequence for Notch-2 shows the entire intron sequence although the protein would be expected to terminate at the first stop codon shown. Notch-2 A and Notch-2B represent additional cDNAs resulting from mis-splicing of the intron beyond the normal 5′ boundary. Amino acids numbers are indicated on the right hand margin and amino acids encoded by the mutant transcripts that are different from those in the wild-type are indicated in bold. * = stop codon.

(b) Partial cDNA sequence of isoamylase from Notch-2 compared with the genomic and cDNA sequences from Aegilops tauchii (At; GenBank AX031278 and AX031277, respectively). The intron sequence is shown in lower case, bold. The putative single base change in Notch-2 giving rise to mis-splicing of the intron is shown in white on black. Predicted amino acid sequences are shown below the cDNA sequences. Stop codons are indicated by stars.

(c) Comparison of transcript abundance by Northern blot. Tracks are: B = Bomi, R = Riso17, NP = NP113 and N2 = Notch 2. RNA was extracted from endosperms of grains of 12–20 mg FW and separated on a denaturing agarose gel. Upper panels: Northern blot hybridized with a full-length isoamylase cDNA. The approximate size of the wild-type isoamylase transcript is indicated. Lower panels: ethidium-bromide-stained gel photographed before blotting.

(d )Genetic map location of the isoamylase gene in a Chebec × Harrington cross in relation to RFLP markers (ABC, ABG, BCD, CDO, KSU, pTAG and PSR prefix), AFLP markers (AA/CAC) and SSR markers (HV and AWB). The full-length wild-type cDNA was used as a probe. Figures to the left of the map are genetic distances between markers (cM). Loci were positioned using the ‘find best location’ function of Map Manager QT (version b29ppc; Manly and Cudmore, 1997).

Figure 4.

Figure 4.

Mutations in the isoamylase genes of Risø 17 and Notch-2.

(a) Predicted amino acid sequences of cDNAs isolated from developing barley endosperms of wild-type (wtype), Bomi, Notch-2 and Risø 17. The putative translation start point has not been determined experimentally and may be any of the methionines located at positions 1, 3, 4 or 6. The deletion in Risø 17 is indicated by the dotted line. All deviations from the wild-type sequence are underlined and in bold. The sequence for Notch-2 shows the entire intron sequence although the protein would be expected to terminate at the first stop codon shown. Notch-2 A and Notch-2B represent additional cDNAs resulting from mis-splicing of the intron beyond the normal 5′ boundary. Amino acids numbers are indicated on the right hand margin and amino acids encoded by the mutant transcripts that are different from those in the wild-type are indicated in bold. * = stop codon.

(b) Partial cDNA sequence of isoamylase from Notch-2 compared with the genomic and cDNA sequences from Aegilops tauchii (At; GenBank AX031278 and AX031277, respectively). The intron sequence is shown in lower case, bold. The putative single base change in Notch-2 giving rise to mis-splicing of the intron is shown in white on black. Predicted amino acid sequences are shown below the cDNA sequences. Stop codons are indicated by stars.

(c) Comparison of transcript abundance by Northern blot. Tracks are: B = Bomi, R = Riso17, NP = NP113 and N2 = Notch 2. RNA was extracted from endosperms of grains of 12–20 mg FW and separated on a denaturing agarose gel. Upper panels: Northern blot hybridized with a full-length isoamylase cDNA. The approximate size of the wild-type isoamylase transcript is indicated. Lower panels: ethidium-bromide-stained gel photographed before blotting.

(d )Genetic map location of the isoamylase gene in a Chebec × Harrington cross in relation to RFLP markers (ABC, ABG, BCD, CDO, KSU, pTAG and PSR prefix), AFLP markers (AA/CAC) and SSR markers (HV and AWB). The full-length wild-type cDNA was used as a probe. Figures to the left of the map are genetic distances between markers (cM). Loci were positioned using the ‘find best location’ function of Map Manager QT (version b29ppc; Manly and Cudmore, 1997).

Figure 4.

Figure 4.

Mutations in the isoamylase genes of Risø 17 and Notch-2.

(a) Predicted amino acid sequences of cDNAs isolated from developing barley endosperms of wild-type (wtype), Bomi, Notch-2 and Risø 17. The putative translation start point has not been determined experimentally and may be any of the methionines located at positions 1, 3, 4 or 6. The deletion in Risø 17 is indicated by the dotted line. All deviations from the wild-type sequence are underlined and in bold. The sequence for Notch-2 shows the entire intron sequence although the protein would be expected to terminate at the first stop codon shown. Notch-2 A and Notch-2B represent additional cDNAs resulting from mis-splicing of the intron beyond the normal 5′ boundary. Amino acids numbers are indicated on the right hand margin and amino acids encoded by the mutant transcripts that are different from those in the wild-type are indicated in bold. * = stop codon.

(b) Partial cDNA sequence of isoamylase from Notch-2 compared with the genomic and cDNA sequences from Aegilops tauchii (At; GenBank AX031278 and AX031277, respectively). The intron sequence is shown in lower case, bold. The putative single base change in Notch-2 giving rise to mis-splicing of the intron is shown in white on black. Predicted amino acid sequences are shown below the cDNA sequences. Stop codons are indicated by stars.

(c) Comparison of transcript abundance by Northern blot. Tracks are: B = Bomi, R = Riso17, NP = NP113 and N2 = Notch 2. RNA was extracted from endosperms of grains of 12–20 mg FW and separated on a denaturing agarose gel. Upper panels: Northern blot hybridized with a full-length isoamylase cDNA. The approximate size of the wild-type isoamylase transcript is indicated. Lower panels: ethidium-bromide-stained gel photographed before blotting.

(d )Genetic map location of the isoamylase gene in a Chebec × Harrington cross in relation to RFLP markers (ABC, ABG, BCD, CDO, KSU, pTAG and PSR prefix), AFLP markers (AA/CAC) and SSR markers (HV and AWB). The full-length wild-type cDNA was used as a probe. Figures to the left of the map are genetic distances between markers (cM). Loci were positioned using the ‘find best location’ function of Map Manager QT (version b29ppc; Manly and Cudmore, 1997).

Figure 4.

Figure 4.

Mutations in the isoamylase genes of Risø 17 and Notch-2.

(a) Predicted amino acid sequences of cDNAs isolated from developing barley endosperms of wild-type (wtype), Bomi, Notch-2 and Risø 17. The putative translation start point has not been determined experimentally and may be any of the methionines located at positions 1, 3, 4 or 6. The deletion in Risø 17 is indicated by the dotted line. All deviations from the wild-type sequence are underlined and in bold. The sequence for Notch-2 shows the entire intron sequence although the protein would be expected to terminate at the first stop codon shown. Notch-2 A and Notch-2B represent additional cDNAs resulting from mis-splicing of the intron beyond the normal 5′ boundary. Amino acids numbers are indicated on the right hand margin and amino acids encoded by the mutant transcripts that are different from those in the wild-type are indicated in bold. * = stop codon.

(b) Partial cDNA sequence of isoamylase from Notch-2 compared with the genomic and cDNA sequences from Aegilops tauchii (At; GenBank AX031278 and AX031277, respectively). The intron sequence is shown in lower case, bold. The putative single base change in Notch-2 giving rise to mis-splicing of the intron is shown in white on black. Predicted amino acid sequences are shown below the cDNA sequences. Stop codons are indicated by stars.

(c) Comparison of transcript abundance by Northern blot. Tracks are: B = Bomi, R = Riso17, NP = NP113 and N2 = Notch 2. RNA was extracted from endosperms of grains of 12–20 mg FW and separated on a denaturing agarose gel. Upper panels: Northern blot hybridized with a full-length isoamylase cDNA. The approximate size of the wild-type isoamylase transcript is indicated. Lower panels: ethidium-bromide-stained gel photographed before blotting.

(d )Genetic map location of the isoamylase gene in a Chebec × Harrington cross in relation to RFLP markers (ABC, ABG, BCD, CDO, KSU, pTAG and PSR prefix), AFLP markers (AA/CAC) and SSR markers (HV and AWB). The full-length wild-type cDNA was used as a probe. Figures to the left of the map are genetic distances between markers (cM). Loci were positioned using the ‘find best location’ function of Map Manager QT (version b29ppc; Manly and Cudmore, 1997).

The isa1 cDNAs from the wild-types, Bomi and NP113, are almost identical and very similar to isoamylase cDNAs from other cereals (data not shown). The isa1 cDNAs from both mutants differ from the wild-type isa1 cDNAs. That from Risø 17 contains an 872-bp deletion, starting at amino acid 338 (Figure 4a), whilst that from Notch-2 contains a 72-bp insertion. The insertion in the Notch-2 cDNA includes two in-frame stop codons (Figure 4a). A BLASTN search with the 72 bp sequence reveals that it is in the same position and has 90% sequence identity with intron 9 of a DBE gene from Aegilops tauchii (Figure 4b). This suggests that the insertion in Notch-2 is an intron that is either not removed or is incorrectly spliced out of the mRNA, possibly due to a single base change (G to A) at the 5′ intron splice junction (Figure 4b). A thorough investigation of the Notch-2 cDNA population revealed a number of other isoamylase cDNAs, for example, Notch-2A and Notch-2B (Figure 4a) in which the intron appears to have been mis-spliced. The mis-splicing in both Notch-2A and B results in the removal of two amino acids and causes frameshifts that result in the introduction of downstream stop codons. No full-length, wild-type cDNAs were found in Notch-2 or Risø 17.

The relative amounts of isa1 transcript in developing endosperms of the wild-types and mutants were compared by Northern analysis (Figure 4c). Using this technique, the transcripts in the mutants were apparently absent. However, as transcripts were detected in the mutants using the more sensitive PCR technique, we assume that a very low level of isa1 transcript is present in the mutants but that it is below the level of detection in the Northern analysis.

The isoamylase gene is located on chromosome 7H

The mutation responsible for the low-starch phenotype of Risø 17 was previously assigned to barley chromosome 7H using a set of translocation lines (Jensen, 1979). We mapped the isoamylase gene, isa1 by RFLP in a population of 86 doubled haploid (DH) lines from a Chebec × Harrington cross. A single polymorphism between restriction enzyme digests of the parental lines was identified. Scoring this polymorphism in the mapping population confirmed the chromosome 7H location (Figure 4d) and reference to markers listed in Langridge et al. (1995) placed the gene genetically close to the centromere on the short arm. The gene was also mapped in a population of 90 DH lines from a Galleon × Haruna Nijo cross (Langridge et al., 1995) which gave an equivalent map location (data not shown). The limit dextrinase gene is also located on chromosome 7H (Figure 4d; R.A. Burton and G.B. Fincher, unpublished data).

The mutants have radically altered starch-granule architecture

Scanning electron microscopy of mature grains (Figure 5a) showed that the endosperm of the wild-type barley, Bomi, contained the A- and B-type starch granules typical of the Triticeae family (Jane et al., 1994). As observed previously (Burgess et al., 1982; Gautam et al., 1994; Sood et al., 1992; Tester et al., 1993), the granules in the mutants were completely different in appearance to those in the wild-type. They were irregular in shape and intermediate in size compared with the A- and B-type granules in Bomi.

Figure 5.

Scanning electron microscopy of barley grains.

Mature (a) and frozen immature (b) grains were fractured to reveal the starch granules within endosperm cells and viewed in a SEM. The magnification is indicated by the scale bar. Note that panels are at different magnifications. Labels are A-type starch granules (A), B-type starch granules (B), compound starch granule (cs), mutant starch granules (s), plastid stroma (ps).

Some of the granules in Risø 17 resembled the compound granules of rice and oat endosperm. In these Risø 17 granules, a number of irregular granulae were closely oppressed to form a smooth compound granule (Figure 5a). Compound granules were previously observed in the developing endosperms of Notch-2 (Sood et al., 1992). These observations and our own of the young developing endosperm of Risø 17 (Figure 5b), suggested that the individual components of the compound granules in these mutants were initiated separately within the plastid. Later, these component granules grew together to fill the available space and therefore became irregular in shape. The granules in these mutants do not appear to arise as single granules which later fracture into multiple, irregular pieces as in developing embryos of the starch-branching enzyme mutant of pea (rr, Lloyd, 1995).

Endosperm cells contain different amounts of phytoglycogen

To investigate the distribution of starch and phytoglycogen within endosperms of barley, slices of endosperm were cut from developing grains, fixed and embedded in resin. Thin sections were cut, stained and examined with either a light microscope (Figure 6, left panels) or a transmission electron microscope (TEM; Figure 6, right panels).

Figure 6.

Light microscopy and TEM of thin of sections of developing endosperm.

Developing grains (40–50 mg FW) were fixed in glutaraldehyde and osmium tetroxide, embedded in epoxy resin and sectioned. Sections for light microscopy (left panels) and TEM (right panels) were stained with toluidine blue. Labels are protein bodies (P), starch granules (S) and phytoglycogen (PG). The magnification is indicated by the scale bars.

As well as the small, irregular granules in the mutants (Figure 6b-f), some plastids in cells of the mutant endosperms contained a diffuse material that stained lightly with toluidine blue (Figure 6b-d). The TEM pictures showed that this material was particulate (Figure 6e,g,h). This is likely to be phytoglycogen as it resembled similar material seen in the leaves of isoamylase mutants of Arabidopsis (Zeeman et al., 1998) and in sections of developing maize endosperm from sugary1 mutant plants prepared using the same procedures used for the barley endosperm sections shown in Figure 6 (M. James and M. Parker, Iowa State University and IFR Norwich, respectively, personal communication).

In the endosperm of older Risø 17 grains and in both young and older Notch-2 endosperm, most cells contained phytoglycogen (not shown). In many plastids, the starch granules had sharp, well-defined edges (e.g. Figure 6f). However, plastids containing large amounts of phytoglycogen as well as starch often had granules with irregular, diffuse edges (Figure 6c,g). At high magnification (Figure 6h), it was evident that the boundary between the granules and phytoglycogen was not well defined.

The phytoglycogen content of adjacent cells was variable, particularly in young endosperm of Risø 17 (Figure 6b,c). Many cells contained little or no phytoglycogen. The phytoglycogen-containing cells were randomly dispersed throughout the endosperm. Within a cell, there was variation between plastids in phytoglycogen content. Some plastids contained large amounts of phytoglycogen and small amounts of starch whilst other plastids contained more starch and relatively little phytoglycogen (Figure 6e). To investigate whether this could have been due to loss of phytoglycogen during fixation or artefacts due to the staining procedure, we compared endosperm sections of the maize mutant sugary1 (kindly provided by M. James and M. Parker) with barley sections that were prepared and stained in the same way. In the maize sections, there was more phytoglycogen and less starch than in the barley sections. This is consistent with measurements of the phytoglycogen and starch content of sugary1 maize endosperm (55% and 14% (w/w), respectively; Dinges et al., 2001). There was also less heterogeneity between cells and between plastids within cells in the amounts of these materials in maize compared to barley. This suggested that the lack of phytoglycogen in some of the cells and plastids of the barley endosperm was not likely to be due to loss during preparation but to real cell-to-cell variation in phytoglycogen and starch content. The underlying cause of this variation is not understood.

There is a single wave of granule initiation in the mutants

To discover whether the total number of granules in the endosperm as a whole was altered in the isoamylase mutants, starch was extracted from endosperms of different ages and the numbers of granules in samples of these were estimated using a haemocytometer (Shannon et al., 1996). In wild-type barley endosperms, A-type granules initiate at approximately 5–10 days after anthesis (DAA) followed approximately 10 days later by a second wave of granule initiation that gives rise to the B-type granules. Our measurements of granule number in Bomi show these two waves of initiation (Figure 7). Before 20 DAA, almost all of the granules were large and disc-shaped, which is typical of A-type granules. After 20 DAA, there was a sudden increase in granule number and we observed small, spherical B-type granules as well as the larger, A-type granules. In the mutant Risø 17, the number of granules per endosperm was higher than in wild-type endosperms before 20 DAA, showing that many more granules initiate in the young endosperm of the mutant than in the wild-type. The mean of the values shown in Figure 7 for endosperms less than 20 DAA was 109.6 ± 15.3 million (mean ± se, n = 11) for Risø 17 and 19.4 ± 3.8 million (mean ± se, n = 11) for Bomi. However, in the mutant, there was no second wave of initiation. Thus, the total number of granules per endosperm in the mutant in the later half of development (after 20 DAA) was similar to that in the endosperm of the wild-type. A second, independent experiment on a separately grown batch of plants gave results that showed the same trends as those shown in Figure 7.

Figure 7.

The number of granules in developing endosperm.

Starch granules were purified from developing endosperms of Bomi (◆) and Risø 17 (◊) and suspended in water. The number of granules in replicate aliquots of the suspension was estimated using a haemacytometer. The age of the grains from which the endosperms were dissected is expressed as days after anthesis (DAA). Values are the means of measurements from three replicate aliquots of a suspension. Each aliquot was sampled 10 times. Each sample contained approximately 50–250 starch granules per 0.00625 mm3.

Discussion

In many respects, the Risø 17 and Notch-2 mutants are similar to other cereal isoamylase mutants. They lack isoamylase activity and have a reduced starch content, an increased sugar content, an altered storage-protein composition and shrivelled grains at maturity, and they accumulate phytoglycogen (our data and those of Bansal, 1970; Doll, 1983; Sood et al., 1992).

The cDNA sequences confirmed that Risø 17 and Notch-2 have mutations in the isoamylase gene, isa1 that abolish function and are very likely to be responsible for the phenotype. The gene from Risø 17 has a 872-bp deletion in the coding region. In Notch-2, there is a single-base substitution in a 5′-intron splice consensus sequence (converting GT to AT) that appears to interfere with the normal splicing of an intron from the primary mRNA transcript. As a result, the cDNA from Notch-2 has either a 72-bp insertion containing two in-frame stop codons or a downstream stop codon introduced by a mis-splicing event. In addition, the abundance of isa1 transcripts is severely reduced in the mutants compared to the wild-types. This could be due to the fact that the mutant transcripts are unstable or are recognized by the plant as aberrant and are therefore rapidly turned over.

As well as causing a decrease in starch content and an accumulation of phytoglycogen, the mutations have a profound effect on the structure, number and timing of initiation of starch granules. As observed previously, in both mutants, granules are smaller than A-type granules from the wild-type and are irregularly shaped (Burgess et al., 1982; Gautam et al., 1994; Sood et al., 1992; Tester et al., 1993). In Risø 17, a single wave of granule initiation occurs at the time in endosperm development when A-type granules initiate in the wild-type. More than one granule initiates per plastid in the mutant and these pack together to form compound granules resembling those found in normal rice and oat grains. In the early stages of endosperm development (< 20 DAA), the total number of granules in the endosperm of Risø 17 is more than five times greater than in the wild-type, suggesting that the mutation conditions an increase in granule initiations. However, a second wave of granule initiations, giving rise to the B-type granules, occurs in the wild-type. Mature grains of Risø 17 thus contain a similar (our experiments) or slightly reduced (71%; Tester et al., 1993) total number of granules compared with Bomi. The lack of a second wave of granule initiations in the mutants may be an indirect effect due to the disruption to starch metabolism in early endosperm development.

Phytoglycogen synthesis was previously thought to be an inevitable consequence of the lack of isoamylase and models to explain the function of isoamylase during starch synthesis have, at their core, ideas to explain the production of this polymer. However, the present study of the barley isoamylase mutants reveals a different picture. Firstly, the fact that phytoglycogen does not accumulate in all endosperm cells shows that loss of isoamylase activity does not necessarily result in phytoglycogen synthesis. Secondly, the initiation of abnormally large numbers of granules early in endosperm development in all cells of the mutants suggests that isoamylase plays a fundamental role in granule initiation.

There is other evidence to suggest that the lack of isoamylase affects starch synthesis without necessarily leading to the production of phytoglycogen. Starch and phytoglycogen accumulation was studied in developing endosperms of maize containing 0–3 doses of the sugary1 mutant allele (Singletary et al., 1997). A statistically significant decrease in starch content was measured in endosperms with 2 and 3 doses of the mutant allele. At maturity, these endosperms had 71% and 57%, respectively, of the starch content per endosperm of the wild-type. However, phytoglycogen accumulation was only observed in endosperms containing three doses of the mutant allele.

An effect of the lack of isoamylase activity on starch granule number and/or shape has been reported for sugary1 mutants of maize and rice. Unlike normal maize endosperm which has simple granules, the granules in the sugary1 mutant of maize are compound (Boyer et al., 1977). In the sugary1 mutant of rice, numerous small granules not observed in the wild-type are present in addition to the normal compound granules (Kubo et al., 1999). This implies that in maize and rice, as in barley, there is an increase in the number of granule initiations per plastid in mutants that lack isoamylase.

We have two suggestions for possible mechanisms through which isoamylase activity might suppress granule initiation. Firstly, the number of granules that are initiated in a plastid might be determined by the concentration and/or structure of soluble α-glucan in the stroma at the onset of starch synthesis. High concentrations of soluble α-glucans with a chain-length profile conducive to crystallization may favour more nucleation events leading to a greater number of granules. Zeeman et al. (1998) suggested that isoamylase, together with other starch-degrading enzymes, might act to reduce the synthesis of soluble α-glucans during the growth of the granule. Similarly, isoamylase might contribute to the destruction of soluble α-glucans at the time of granule initiation, thus limiting the number of nucleation events leading to granules as well as later, inhibiting the synthesis of phytoglycogen. In the absence of isoamylase, more soluble α-glucans capable of crystallization may accumulate in the stroma and this may favour more nucleation events and hence, initiation of more than the usual number of granules.

Secondly, isoamylase may destroy a specific primer required for granule initiation. In animals and yeast, a self-glucosylated protein primer, glycogenin, is required for the synthesis of glycogen particles. Glycogen synthase cannot elongate small oligosaccharides unless these are attached to glycogenin (Alonso et al., 1995). Isoamylase might act on some glycogenin-like protein required for granule or polymer initiation in plants, preventing the initiation of large numbers of granules and phytoglycogen particles by cleaving off the associated α-glucan chains. Isoamylase from Pseudomonas is able, in vitro, to cleave the α-glucan chain from a primed glycogenin molecule (Lomarko et al., 1992). Limit dextrinase, the other form of starch debranching enzyme found in plants, cannot cleave this protein-glucan complex (Lomarko et al., 1992).

The role of glycogenin-like proteins in starch synthesis is not established. Starch synthases can elongate small malto-oligosaccharides and so it is not necessary to postulate the presence of a priming protein for polymer synthesis. However, plant genes encoding glycogenin-like proteins have been reported (e.g. Roach and Skurat, 1997) and various self-glycosylating proteins have been identified biochemically (Dhugga et al., 1997). The possibility that granule initiation is controlled via the action of isoamylase on a protein-glucan primer requires further investigation.

Experimental procedures

Plants

Grains of NP113 were obtained from the National Small Grains Research Facility, Idaho, USA, Risø 17 from the Nordic Gene Bank, Alnarp, Sweden and all other grains from the John Innes Centre Germplasm Collection. Plants were grown in a greenhouse in individual pots at a minimum temperature of 12°C, with supplementary lighting in winter.

Extraction of starch and soluble α-glucan

Glucans were extracted from individual mature grains by grinding to a fine powder in a pestle and mortar. The powder was suspended in 5 ml ice-cold H2O, ground further and centrifuged at 2500 g for 5 min at 4°C. The supernatant was removed and the pellet was resuspended in 2 ml ice-cold H2O, centrifuged again and the supernatant pooled with the previous supernatant. The pellet was resuspended in 1 ml ice-cold H2O followed by 4 ml ethanol and incubated on ice for 15–30 min. After centrifugation as before, the supernatant was discarded and the pellet was resuspended in 12 ml H2O. Duplicate aliquots of the suspended starch were diluted 5-fold with H2O, autoclaved and assayed for α-glucan. The pooled supernatants containing the soluble α-glucan were diluted 6-fold with aqueous methanol/KCl (75% (v/v) methanol, 1% (w/v) (KCl), incubated at 4°C for 12 h and centrifuged as before. The supernatant was discarded and the pellet was resuspended in 12 ml H2O. Duplicate aliquots of the supernatant were diluted 6-fold with H2O, autoclaved and assayed for α-glucan.

The wavelength of maximum absorption of the α-glucan-iodine complex was determined as follows. Purified α-glucans were dissolved by suspension at 20 mg ml−1 in 1 m NaOH and diluted with an equal volume of water. An aliquot (20–50 µl) of the dissolved α-glucan was added to a cuvette containing, in a final volume of 1 ml, 100 mm NaOH, 100 mm acetic acid and 100 µl Lugol solution. The maximum absorbance (400–800 nm) was determined relative to that of a sample identical except that it contained no α-glucan.

For α-glucan chain-length analysis, α-glucans were extracted from grains (0.5–3.0 g FW) ground to a fine powder in liquid nitrogen in a pestle and mortar. The powder was suspended in 2 ml 10% (w/v) perchloric acid, transferred to a 50-ml tube and shaken on ice for 30 min. After centrifugation at 18000 g for 30 min at 4°C, the supernatant (containing the soluble α-glucan) was retained and the pellet (containing the starch) was resuspended in 10 ml H2O, filtered through muslin (washing through with additional H2O) and centrifuged at 2500 g for 10 min. The grey layer on top of the starch was removed and the starch was washed successively in 20 g l−1 SDS (thrice) and H2O (twice). After proteinase K treatment to remove surface proteins (Rahman et al., 1995), the starch was washed successively in 20 g l−1 SDS (once), H2O (twice) and ice-cold acetone (twice), and freeze-dried.

The soluble α-glucans were precipitated by the addition of 3.5 vols methanol to the perchloric acid-soluble material, incubated on ice overnight and the precipitate was collected by centrifugation at 2500 g for 15 min at 4°C. The supernatant was discarded and the pellet was resuspended in 4 ml H2O, and subjected to re-precipitation with methanol as above. The supernatant was discarded and the soluble-glucan pellets were freeze-dried.

Glucan assays

Glucans were assayed as glucose released after digestion with specific glucosylases. Control reactions, in which the glucosylases were omitted, were also performed. Duplicate samples (0.5 ml) were each incubated with 0.5 ml 50 mm Na acetate pH 5.2, 2 U α-amylase and 10 U amyloglucosidase (enzymes from Roche Diagnostics, Lewes, UK) at 25°C for 12–24 h. The samples were heated to 100°C for 2 min, centrifuged at 14000 g for 2 min and the supernatants were assayed spectrophotometrically for glucose according to Lowry (1972).

Determination of chain-length distribution of α-glucans

The methods used were based on those of O'Shea and Morell (1996) as described in Edwards et al. (1999).

Scanning electron microscopy

Mature barley grains were fractured transversely using a razor blade to initiate the fracture. The pieces were mounted on aluminium stubs using silver conducting paint. Samples were sputter-coated (Emitech Ltd, Ashford, UK) with a layer of gold approximately 25 nm thick and examined and photographed in a Leica Stereoscan 360 SEM (LEO, Cambridge, UK).

Light microscopy and transmission electron microscopy

Tissue slices approximately 1.5 mm thick were cut from the endosperm of developing barley and fixed for at least 4 h in 3% glutaraldehyde in 0.1 m cacodylate buffer, pH 7.4. The slices were washed 3 times in buffer, cut into smaller pieces and fixed overnight in 1% aqueous osmium tetroxide. The pieces were dehydrated in an ethanol series with 3 changes in 100% ethanol and transferred to acetone. Tissue was infiltrated and embedded in epoxy resin (Spurr's, Agar Scientific, Stansted, UK).

Sections 1 or 1.5 µm thick were stained with 1% (w/v) toluidine blue in 1% (w/v) borax, pH 11 and examined with an Olympus BX60 microscope (Olympus Optical, Japan) and recorded digitally with AcQuis Bio software (Synoptics, Cambridge, UK).

Sections showing silver-gold interference colours were cut from embedded material with a diamond knife and collected on copper grids. They were stained sequentially in uranyl acetate and Reynold's lead citrate and examined and photographed in an electron microscope (JEOL 1200EX/B).

Native gels

All procedures were carried out at 4–6°C. Developing endosperms were homogenized in approximately 10 vols extraction medium (100 mm MOPS pH 7.2, 10 mm EDTA, 50 ml l−1 ethanediol, 1 mm DTT) in a pestle and mortar. The extract was centrifuged at 28000 g for 5 min at 4°C. The resulting supernatant was added to sample loading medium (600 ml l−1 glycerol, 2 mg ml−1 bromophenol blue, 20 mm DTT) in a volume ratio of 1 : 5 (loading buffer: to extract).

Native gel electrophoresis was carried out according to Laemmli (1970), except that SDS was omitted from all solutions and the separating gel contained 0.1% (w/v) acarbose (Glucobay 100, Bayer plc, Berkshire, UK) to inhibit α-amylase activity. The separating gels consisted of 7.5% acrylamide, were 1 mm thick and contained potato amylopectin (1 mg ml−1, Sigma, Poole, UK) or red pullulan (10 mg ml−1, Megazyme International, County Wicklow, Ireland).

After electrophoresis at 15 mA per gel and 4°C, native gels containing amylopectin were rinsed in medium A (100 mm MES pH 6.0, 5 mm DTT, 50 ml l−1 ethanediol, 0.1% (w/v) acarbose), incubated in this medium for 16 h at 37°C, rinsed briefly in water and then stained with Lugol's solution. Native gels containing red pullulan were rinsed in medium A and then incubated in this medium for 6–16 h at 37°C until bands were visible. After incubation, gels were soaked in 50 ml l−1 aqueous ethanol to enhance the contrast between the bands and the background.

Enzyme activities

All procedures were carried out at 4–6°C. Developing endosperms were dissected from grains of 45–55 mg FW and homogenized in approximately 5 vols extraction medium (100 mm MOPS pH 7.2, 5 mm DTT, 5 mm MgCl2, 5% (v/v) glycerol, 1% (w/v) BSA, 1% (w/v) (PVP) in a pestle and mortar followed by an all-glass homogenizer. The extract was centrifuged at 28000 g for 10 min at 4°C and the resulting supernatant was assayed. For total starch synthase assays, the extract was not centrifuged.

For each enzyme, the activity reported was dependent upon the presence in the assay of all of the appropriate substrates and cofactors and also upon extract concentration within the range used to make the measurements. The concentrations of components of each of the assays and their pH values were optimized to give the maximum rate using extracts of Bomi. The rate of the reaction was linear with respect to time for at least 4 min in spectrophotometric assays and for at least 10 and 30 min in assays of sucrose synthase and starch synthase, respectively. For starch-branching enzyme, activity was calculated from the rate of reaction during the phase of the assay in which it was linear with respect to time. Reaction mixtures were as follows:

Sucrose synthase. As in Craig et al. (1999) except that the buffer was 82 mm AMPSO (3-[(1,1-Dimethyl-2-hydroxyethyl)amino]-2-hydroxy-propanesulphonic acid) pH 9.0.

UDPG pyrophosphorylase. The assay contained, in 1 ml, 100 mm HEPES pH 8.1, 2 mm MgCl2, 0.8 mm NAD, 0.8 mm UDP-glucose, 1 mm NaPPi, 2 U phosphoglucomutase (PGM), 5 U G6PDH and 10–50 µl of a 10-fold dilution of extract in extraction medium. The reaction was initiated with sodium pyrophosphate (NaPPi) and monitored spectrophotometrically at 340 nm.

Fructokinase. As in Craig et al. (1999) except that 2.5 mm NAD, 3 mm MgCl2 and 100 µl extract were used.

Glucokinase. As in Craig et al. (1999) except that 100 mm Bicine pH 8.5, 2 mm NAD, 2.5 U G6PDH, 100 µl extract were used and PGM was omitted.

Phosphoglucomutase. The assay contained, in 1 ml, 50 mm Bicine pH 8.0, 0.6 mm NAD, 6 mm glucose-1-phosphate, 2 U G6PDH, 5–10 µl extract.

Phosphoglucose isomerase. The assay contained, in 1 ml, 100 mm glycyl glycine pH 8.4, 1 mm NAD, 10 mm fructose-6-phosphate, 4 U G6PDH, 5–10 µl extract.

ADPG pyrophosphorylase. As in Smith et al. (1989; assay 2b), except that 100 mm HEPES pH 7.9, 0.4 mm NAD, 1 mm ADP-glucose, 1.5 mm NaPPi and 5 U PGM were used.

Soluble starch synthase. As in Jenner et al. (1994), the resin method except that 0.5 mg potato amylopectin, 2 mm ADP[U-14C] glucose at 2.3 GBq mol−1 and 10 µl of extract were used.

Granule-bound starch synthase. As above for soluble starch synthase except that extracts which had not been centrifuged to remove insoluble material (total extracts) as well as soluble extracts were assayed. The granule-bound activity was calculated as the difference between the activities in the total and soluble extracts.

Alkaline inorganic pyrophosphatase. As in Gross and ap Rees (1986) except that 50 mm Bicine pH 8.9, 20 mm MgCl2, 1.25 mm NaPPi and 50 µl extract were used.

Starch-branching enzyme. As in Smith (1990), the phosphorylase-stimulation assay using MES buffer except that assays were processed using DOWEX rather than methanol/KCl essentially according to Jenner et al. (1994), the resin method. 0–10 µl of a 10-fold dilution of extract was used and activity was expressed in each case relative to that in assays with no extract.

RNA extraction and cDNA synthesis

Total RNA was extracted from barley tissues using a commercially available phenol-guanidine isothiocyanate preparation (Trizol; Gibco BRL, Gaithersburg, MD, USA). Single stranded cDNA was prepared from 1 µg of total RNA with Thermoscript reverse transcriptase (Gibco BRL) and either an oligo(dT)20 or the TRACE (Frohman et al., 1988) primer according to the manufacturers instructions. The cDNA was treated with RNase H for 20 min at 37°C prior to the PCR reaction.

Isolation of the 5′ and 3′ ends of the barley isoamylase cDNA

A cDNA sequence for barley isoamylase was published by Sun et al. (1999) but comparison with published sequences from other cereals (Zea mays, Beatty et al., 1997; AF030882 and Triticum aestivum,Luetticke et al., 2000; AX010486) indicated that it was not full length. The missing 5′ end of the isoamylase transcript was obtained using nested PCR on cDNA prepared from total RNA from developing grain (Bomi) 10–13 days after fertilization. The primer sequences were 5′-CCGATAAATAATCCCACCTCGC-3′ and 5′-ATCACTGCCTTAGCATAAGGATCC-3′ for the first round of PCR and 5′-GGCTGCAGGGCATGAAGATGATGGCCAT-3′ with 5′-GAACCTCCTCGCTCACCCTAT-3′ for the second, nested round, with PCR conditions of 4 min at 94°C, followed by 35 cycles of 1 min at 94°C, 1 min at 53°C, 1 min at 72°C and a final extension time of 10 min at 72°C. The 3′ end of the cDNA sequence was obtained using a 3′ RACE protocol where the first round of PCR was carried out with a gene-specific isoamylase primer 5′-CCACTTATTGACATGATCAGC-3′ and the RACE 3′ primer 5′-GACTC GAGTCGACATCG-3′ (Frohman et al., 1988). A second PCR reaction was carried out with a nested gene-specific primer 5′-CGTCAAGCTCATTGCTGAAGC-3′ and RACE 3′ with PCR conditions of 4 min at 94°C, followed by 35 cycles of 30 sec at 94°C, 30 sec at 50°C, 2 min at 72°C and a final extension time of 10 min at 72°C. PCR products to be analysed were cloned into the pGEM-T Easy vector (Promega, Madison WI, USA) and sequenced using an ABI 3700 capillary sequencer.

Isolation of full-length isoamylase cDNAs

Primers for the isolation of full-length isa1 cDNAs from wild type and mutant barley lines were designed to the 5′ untranslated region upstream from the predicted translation start site and to the 3′ untranslated region at the 3′ end of the transcript. A first-round PCR reaction was carried out with the primers 5′-CCGATAAATATCCCACCTCGC-3′ and 5′-CCGCCGAACGACTACATATAC-3′ using the high-fidelity Taq polymerase Elongase (Gibco, BRL) and PCR conditions of 45 sec at 94°C, followed by 35 cycles of 45 sec at 94°C, 45 sec at 54°C, 3 min 30 sec at 72°C and a final extension time of 7 min at 72°C. A fully nested, second round of PCR was carried out with 5′-GGCTGCAGGGCATGAAGATGATGGCCATGG-3′ and 5′-TCAAACATCAGGGCGTGATACAA-3′ under the same PCR conditions. Putative full-length transcripts were cloned into the pGEM-T Easy vector (Promega). The cDNA transcripts were fully sequenced in both directions using a set of overlapping primers on an ABI 3700 capillary sequencer. The resultant chromatograms were edited using Chromas (Technelysium, Helensvale, Queensland, Australia) software and analysed with Genetic Computer Group (Madison, WI, USA) software in the ANGIS suite of programs at the University of Sydney.

Northern analysis

Approximately 10 µg of total RNA was separated on a 1% agarose denaturing gel with size standards (Promega, Madison, USA). RNA was visualized with ethidium bromide under ultraviolet light to ensure equal loading. RNA was transferred to Duralon nylon membrane (Stratagene, La Jolla, USA) by capillary transfer and crosslinked using UV light. A [32P]-labelled full-length isoamylase cDNA was synthesized by random priming, essentially as described by Feinberg and Vogelstein (1983) and the membrane was probed as described by Banik et al., 1996). Autoradiography was performed for 5 days at − 80°C with X-ray film and an intensifying screen.

Mapping the isoamylase gene, isa1

The probe DNA for Southern hybridization was radioactively labelled using standard methods. Hybridization methods were as described in Rogowsky et al., 1991), except that both pre-hybridization and hybridization were carried out in the same solution (0.9 m NaCl, 30 mm Pipes pH 6.8, 0.75 mm EDTA, 7.5% (w/v) dextran sulphate, 0.6% (w/v) BSA, 0.6% (w/v) Ficoll 400, 0.6% (w/v) polyvinyl-pyrollidone, 250 mg ml−1 denatured salmon sperm).

Counting starch granules

All procedures were carried out at 4–6°C. Developing endosperms were dissected from immature grains and homogenized in approximately 50 vols extraction medium (50 mm HEPES pH 7.8, 10 mm EDTA, 10 mm DTT, 0.1 mg ml−1 and 0.1 mg ml−1 Proteinase K) in a pestle and mortar. The homogenate was incubated at 37°C for 1 h, centrifuged at 28000 g for 5 min and the supernatant discarded. The pellet was washed successively in 1 ml aliquots of 20 g l−1 SDS (twice), water (twice), 0.5 m NaCl, water, and the resulting starch preparation was resuspended in 1–10 ml of water. Three 200-µl or 1-ml aliquots of the suspension were removed for analysis. For each aliquot, the number of granules per ml was estimated using a haemocytometer slide with a unit volume of 0.00625 mm3. Starch granules stained with Lugol solution, were diluted to approximately 50 granules per unit volume and viewed under a light microscope. The precise number of granules per unit volume was determined 5 times each for two replicate dilutions of each sample of resuspended starch. These results were used to calculate the number of granules per endosperm.

Acknowledgements

The John Innes Centre is supported by a competitive strategic grant from the Biotechnology and Biological Sciences Research Council (BBSRC), UK. The authors are extremely grateful to Alison M. Smith for support, encouragement and useful discussions throughout the course of this work and for constructive criticism of the manuscript. Nicola Patron and Margaret Pallota are thanked for DNA and RNA preparation, and chromosome mapping, respectively. Tamara Verhoeven thanks the BBSRC for a research studentship and Syngenta for additional financial support. Work at the University of Adelaide was supported by grants (to Geoff Fincher) from the Grains Research and Development Corporation of Australia and (to Kay Denyer) from the JIC/CSIRO/Waite fund.

Accession numbers for GenBank database: AF490375, AF490376, AF490377

Ancillary

Advertisement