INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS AND DISCUSSION CONCLUSION MATERIALS AND METHODS LITERATURE CITED

Bioactive gibberellins (GAs) such as GA₁ are important regulatory factors in the control of shoot growth (Hedden and Proebsting, 1999). Evidence from several species indicates that GAs are also required for the normal development of roots. For example, root development is altered in the maize d5 mutant (Baluska et al., 1993), and excised roots of the gib-1 (tomato) and d1 (maize) mutants grew more slowly than the corresponding wild-type lines (Mertz, 1966; Butcher et al., 1990; Barlow et al., 1991). Shoots of these mutants are known to be GA₁ deficient (Fujioka et al., 1988; Hedden and Lenton, 1988), and it has been assumed that this deficiency extends to the roots. GA biosynthesis inhibitors are also known to inhibit root elongation in wild-type lines of tomato (Butcher et al., 1990), lettuce (Tanimoto, 1990), and maize (Baluska et al., 1993), and this inhibition is overcome by application of bioactive GA, providing further evidence that GAs are important for root growth.

However, in the garden pea (Pisum sativum), the roots of the best known GA₁-deficient dwarf mutant, Mendel's le-1, are usually considered to be normal in appearance (Tanimoto, 1990). Although this has not been confirmed previously using isogenic LE/le-1 lines, the general vigor of le-1 lines, which are favored for agricultural purposes, is certainly consistent with normal root development. Furthermore, le-1 roots do not respond to applied GA (Tanimoto, 1990). Nevertheless, as in wild-type peas, le-1 roots can be shortened by treatment with an inhibitor of GA biosynthesis, an effect reversed by GA application (Tanimoto, 1988, 1990). These observations led to suggestions that roots are more sensitive to GAs than are shoots, and that the small amount of bioactive GA produced by the leaky le-1 mutation is sufficient for normal root growth (Tanimoto, 1990). In shoots, le-1 reduces the content of GA₁ by at least 10-fold (Ross et al., 1992), and if there are similar effects in roots, it would appear that even quite large reductions in bioactive GA do not alter root development. This would suggest that even though GAs are required for normal root development in pea, they may not limit or regulate that process. It might follow that only a very large decrease in GA levels can reduce root growth in pea.

We should not assume, however, that the le-1 mutation actually reduces root GA levels. Smith et al. (1992) found no difference in root GA₁ content between wild-type and le-1 plants, although these plants were too old for root development to be readily compared. At present we are unaware of any other reports regarding root GA levels in GA mutants in any species. The aim of the present study was, therefore, to investigate the effects of the GA mutations le-1, le-2, lh-2, ls-1, and na on endogenous GA levels in, and the development of, pea roots. In shoots, these mutations block GA biosynthesis at the steps shown in Figure 1. The ls and lh mutations block the production and oxidation, respectively, of ent-kaurene (Ait-Ali et al., 1997; Swain et al., 1997). na is thought to block between ent-7-hydroxykaurenoic acid and GA₁₂-aldehyde (Ingram and Reid, 1987). The le alleles block the step GA₂₀ to GA₁ (Ingram et al., 1984; Ross et al., 1989). The le-1 mutation arose from a G to A substitution at position 685, and reduces enzyme activity by about 95%, whereas le-2 involves a base deletion of the le-1 allele at position 376, and results in a severely truncated protein with undetectable enzyme activity (Lester et al., 1997, 1999a; Martin et al., 1997). We also use the sln mutation to examine the effects of increased endogenous GA levels on root growth because this mutation blocks the catabolism of GA₂₀ (Fig. 1; Ross et al., 1995; Lester et al., 1999b), resulting in elevated GA levels at the seedling stage (Reid et al., 1992).

View larger version (15K): [in this window] [in a new window]   Figure 1.   GA biosynthesis pathway in pea, showing the steps affected by the genes studied. GGDP, geranylgeranyldiphosphate; CDP, copalyl diphosphate.

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS AND DISCUSSION CONCLUSION MATERIALS AND METHODS LITERATURE CITED

Distribution of GAs in Roots

The level of GA₁ was approximately 15-fold greater in the apical 25 mm ("tip") of the taproot than in the remainder of the taproot, and was also relatively high in the lateral roots (Table I). This is consistent with the proposition that GA₁ is the important bioactive GA in pea because the root tips analyzed would have included the zone of maximal taproot growth (Heyes, 1977). Levels of the GA₁ catabolite, GA₈, were also higher in tips and laterals than in the remaining portions. GA₁₉ was relatively abundant in all portions, and its level consistently exceeded that of its product, GA₂₀. This contrasts with the shoots, where GA₂₀ levels are often greater than those of GA₁₉, especially in leafy apical portions (for example, see Ross, 1998). Except for GA₁₉, the level of GAs in the root, even in the tips, was typically at least an order of magnitude lower than in typical shoot apical portions (Ross, 1998). In roots, GA₃ and GA₄ were not detected, although the corresponding internal standards, added at a rate of 500 pg g¹ fresh weight, were recovered.

View this table: [in this window] [in a new window]   Table I.   Distribution of GAs in wild-type roots Line 107 roots were partitioned into taproot tips (25 mm), lateral roots, and the top and intermediate portions of the remainder. Data are in pg g1 fresh wt, and are the means and SEs of two replicate harvests. Plants were grown in a growth chamber at 18°C and were 12 d old at harvest.

The le Mutations Are Not Expressed in Roots

Comparison of isogenic lines (205+/205) confirmed that Mendel's le-1 mutation did not affect root development. The overall phenotype of le-1 roots was indistinguishable from that of the wild type (Fig. 2A), and le-1 and LE taproots were of similar length (Table II). Furthermore, in le-1 (205) root tips GA₁ and GA₂₀ levels were similar to those in the isogenic wild type (205+), in contrast with the shoots, where le-1 dramatically reduces GA₁ content but elevates that of GA₂₀ (Ingram et al., 1984; Ross et al., 1992). The normal root GA content explains why le-1 roots elongated to the same extent as wild-type roots.

View larger version (13K): [in this window] [in a new window]   Figure 2.   Photograph showing root phenotypes. A, Left, wild type (205+); right, le-1 (205). B, Left, wild type; right, na. C, From left to right, wild type, na SLN, na sln, and NA sln. Plants were 8 (C) or 9 (A and B) d old and were grown at 21°C in a growth cabinet.

View this table: [in this window] [in a new window]   Table II.   Morphology of, and root GA levels in, LE, le-1, and le-2 seedlings There were two pairs of isolines: LE and le-1 (205+/205), and le-1 and le-2 (Dippes gelbe Viktoria/M66A). Morphological data are means ± SE of 12 replicates; plants were grown in a growth chamber at 21°C. GAs were quantified from 40-mm taproot tips and plants were grown in a heated greenhouse. GA levels are shown as means ± SE of two replicate harvests. All plants were 12 d old.

It was formally possible that the "leakiness" of the le-1 allele in some way enabled le-1 roots to maintain normal levels of GA₁. This possibility was investigated by observing root development in mutant le-2. The le-2 mutation appears to be biochemically null, because when the le-2 cDNA was expressed in Escherichia coli, the expression product failed to convert GA₂₀ to GA₁ (Martin et al., 1997; Lester et al., 1999a). A direct comparison of LE and le-2 roots was not possible in the present study, because LE/le-2 isolines were not available. Nevertheless, a valid comparison could be made using the two pairs of isolines LE/le-1 (205+/205) and le-1/le-2 (Dippes gelbe Viktoria/M66A), because lines 205 and Dippes gelbe Viktoria have an identical le-1 cDNA sequence (Lester et al., 1997, 1999a). In the le-2 mutant, root development and root GA levels were similar to those in its le-1 isoline (Dippes gelbe Viktoria; Table II). Because le-1 itself did not affect root development or root GA levels (see above), we can conclude that root development and root GA levels were normal (i.e. similar to wild type) in the le-2 mutant.

The level of GA₁ did vary between the two le-1 lines, 205 and Dippes Gelbe Viktoria (Table II), which are derived from quite distinct pea varieties. This demonstrates the effect that genetic background may exert on hormone levels and reinforces the need for using isogenic lines when examining the effects of major genes; such lines are still not always used.

The lack of effect of le-1 and le-2 on root GA₁ levels strongly supports the idea that GA₁ can be synthesized in root tissues because it is difficult to envisage how le-1 and le-2 roots could contain normal GA₁ levels if they were dependent on the shoot for this GA (because le-1 and le-2 shoots contain less than 10% and 2%, respectively, of the GA₁ content of wild-type shoots; Ross et al., 1989, 1992).

The level of LE mRNA is reported to be quite high in roots (at least in dark-grown plants; Martin et al., 1997), which suggests a role for LE in these organs. However, the normal GA₁ level in le-2 roots indicates that at least one other 3-hydroxylase gene also operates in roots, and that this second gene can completely compensate for the loss of LE activity in le-2 roots. Therefore, the mere presence of mRNA for a gene should not be interpreted to mean that the tissue depends on that gene for a certain step to proceed.

Mutations na, lh-2, and ls-1 Are Expressed in Roots

In contrast with the le mutations, the mutation na markedly affected root growth (Fig. 2B). The length of the taproot, measured 8 (Fig. 2C), 9 (Fig. 2B), and 12 d (Table III) after sowing, was consistently reduced in na plants, compared with wild-type plants. Lateral roots were also substantially shorter in na plants (Table III). It is important that the na mutation markedly reduced GA₁ content in root tips (Tables III and IV). This appears to be the first time that reduced root growth has been directly correlated with reductions in endogenous root GA levels. Furthermore, the diameter of na roots (20 mm from the root tip) was significantly greater than that of wild-type roots (Table III), consistent with previous reports linking GA deficiency with root thickening (Butcher et al., 1990; Tanimoto, 1994). The GA levels in Tables III and IV are from 12.5-mm root tips (na) and 25-mm tips (NA), which were harvested this way to avoid including too much mature root tissue in the na harvest. Both sections would have included the entire growth zone (Heyes, 1977). Very similar results were obtained when root tips of the same length (25 mm) from both genotypes were harvested (data not shown). These results support the theory that GAs are important for root development. Consistent with that theory, roots of the na mutant responded to exogenous GA₁ application (Table V).

However, the most striking rescue of the na root phenotype was observed after recombining na with the sln mutation (Table IV and Fig. 2C). It was reported previously that sln results in elevated levels of GA₁ in the shoots of seedlings (Reid et al., 1992), and here we show that the same is true for roots (Table IV). The observation that na sln roots were essentially wild type in appearance (Fig. 2C) confirms that GA deficiency is the primary reason for the shorter roots of na SLN plants. The shoots of na sln plants were elongated at 8 (Fig. 2C) and 9 d (Table IV), but their internodes shortened dramatically as the plants matured (Ross et al., 1995). Roots of sln plants contained high levels of other GAs derived (directly or indirectly) from GA₂₀, namely GA₂₉ and GA₈ (Table IV). The GA₂₀ that gave rise to these elevated levels originated in the seed (Reid et al., 1992; Ross et al., 1995). Seeds of genotype na sln would have contained high GA₂₀ levels because the na mutation does not block GA biosynthesis in seeds (Potts and Reid, 1983). It is also probable that even in na SLN plants some GA₂₀ moved from the seed into the root. This would explain why in Table IV there were relatively high levels of GA₈ and GA₂₉ in na SLN roots.

Mutations lh-2 and ls-1 also affected both root development and endogenous GA levels (Table VI). The phenotypic effects of these mutations were subtle, but both significantly (P < 0.05) reduced taproot length at 12 d, the number of lateral roots longer than 10 mm, and the length of the longest lateral roots (Table VI). Mutation lh-2 reduced GA₁ content substantially but reduced root elongation less than na did (Tables III and VI). The reason for this is not clear, but possibly on the LH LS/lh-2/ls-1 genetic background the roots are more sensitive to GA than on the NA/na background.

In both lh-2 and ls-1 roots the reduction in GA₁₉ content was greater than the drop in GA₁. This indicates that in roots, as in shoots (Hedden and Croker, 1992), GA₁ negatively regulates its own biosynthesis, with the metabolism of GA₁₉ occurring at a faster rate in GA₁-deficient roots. This suggestion also explains why, on an NA background, the level of GA₁₉ was higher in sln roots (where the GA₁ level was elevated) than in SLN roots (Table IV).

The shorter roots of na plants are most likely not a consequence of reduced photosynthesis by na shoots (which are very small; Fig. 2), because in dark-grown plants na caused a similar reduction to that observed in the light (data not shown). In general, there was no evidence from the present study that dwarfism in the shoot is necessarily associated with shorter roots. For example, the roots of the le-1 and le-2 mutants were normal even though shoot height was markedly reduced by le-1 and particularly by le-2 (Table II).

	CONCLUSION

TOP ABSTRACT INTRODUCTION RESULTS AND DISCUSSION CONCLUSION MATERIALS AND METHODS LITERATURE CITED

These results show, for the first time, that some mutants with GA-deficient shoots also possess GA-deficient roots. This GA deficiency was associated with impaired root elongation (Tables III, IV, and VI). Mutant na roots responded to exogenous GA (Table V), and were rescued by mutation sln (Fig. 2C), which increases root GA content (Table IV). Therefore, our results support the theory that GAs are important for normal root growth. It is interesting that on an NA (wild-type) background, sln did not increase taproot length (in contrast with stem length; Table IV; Fig. 2C). This result is consistent with the inability of applied GA to promote elongation in wild-type pea roots, unless those roots are pretreated with a GA biosynthesis inhibitor (Tanimoto, 1988). These findings suggest that the GA₁ level in wild-type roots is sufficient for maximal elongation. Because this is not the case in shoots, even though they contain more GA₁ than roots, the present results support the previous theory (Tanimoto, 1990) that roots are more GA sensitive than shoots. However, there is no longer any evidence from the le-1 mutation for this theory (Tanimoto, 1990). It now appears that mutant le-1 roots are normal not because moderate reductions in root GA₁ level have no effect, but because le-1 does not reduce root GA₁ levels (Table II).

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS AND DISCUSSION CONCLUSION MATERIALS AND METHODS LITERATURE CITED

Plant Material and Growing Conditions

Three pairs of isogenic lines, and a fourth group of three isogenic lines, were used in this study. These were: NGB1769 (NA) and NGB1766 (na); 205+ (LE) and 205 (le-1); Dippes gelbe Viktoria (DGV, le-1) and M66A (le-2, previously le^d; Ross and Reid 1987); and 107 (LH LS, selected from cv Torsdag), NGB5843 (lh-2), and 181 (ls-1) (Reid and Potts, 1986; Swain and Reid, 1992). Each pair or triplet possesses a different genetic background. Genotypes na sln, na SLN, NA sln, and NA SLN originated from a cross between line NGB6074 (sln) and an na segregate from cross NGB1766 × NGB1769.

The growth medium was a 1:1 (v/v) mixture of dolerite chips:vermiculite, topped with 20 to 30 mm of sterilized potting mix. Plants were grown either in a heated greenhouse (Beveridge and Murfet, 1996) with an 18-h photoperiod, provided by extending the natural photoperiod at its beginning and end with a mixture of fluorescent (40-W cool-white, Osram, Munich) and incandescent (100-W, Thorn, Melbourne, Australia) light (25 µmol m² s¹ at pot top); or in a growth cabinet in an 18-h photoperiod provided by a mixture of fluorescent and incandescent light (36 and 60 W, respectively; Thorn, Australia; 200 µmol m² s¹ at pot top).

Extraction and Quantification of GAs

Root tips were harvested by excising 12.5, 25, or 40 mm from the taproot apex; typically, approximately 1 g of material from 30 to 60 plants was harvested. The distribution of GAs in wild-type roots was investigated using line 107. In this case 25-mm taproot tips were excised; these constituted about 10% of the total root fresh weight. Lateral roots were also excised (about 10% of total root fresh weight). The remainder of the root was partitioned into "top" and "intermediate" sections, consisting of approximately 55% and 25%, respectively, of the total root fresh weight.

Harvested material was immediately immersed in cold (20°C) methanol and placed in a freezer. The tissue was macerated using a blender and GAs were extracted at 2°C to 4°C for 24 h, prior to filtering. The internal standards were [²H₂]GA₁₉, [²H₂]GA₂₀, [²H₂]GA₂₉, [²H₂]GA₁, [²H₂]GA₈, [²H₂]GA₃, and [²H₂]GA₄. Extracts were purified as before (Ross, 1998), except that (in some cases) after the Sep-Pak step GAs were extracted into ethyl acetate at pH 2.9, prior to HPLC, performed as before (Ross, 1998). In other cases, GAs were further purified at the (post-HPLC) methyl ester stage by taking up the extract in 1 mL of distilled water and partitioning three times against 400 µL of diethyl ether. Gas chromatography-mass spectrometry was performed as described previously (Ross, 1998), except that in some cases extracts were derivatized twice, first in 10 µL of dry pyridine and 40 µL of N,O-bistrimethylsilyltrifluoroacetamide + 1% (v/v) trimethylchlorosilane, and then, (after drying), in 15 µL of bistrimethylsilyltrifluoroacetamide + 1% (v/v) trimethylchlorosilane. Peak areas were corrected for contributions from naturally occurring isotopes and for small amounts of unlabeled material in the internal standards. The limit of detection (for GA₁) was approximately 30 pg gFW¹ for a harvest of 1 g of tissue.