Physical Restriction of Pods Causes Seed Size Reduction of a Brassinosteroid-deficient Faba Bean (Vicia faba)

Abstract

• Background and Aims A brassinosteroid-deficient mutant faba bean (Vicia faba ‘Rinrei’) shows dwarfism in many organs including pods and seeds. ‘Rinrei’ has normal-sized seeds together with dwarf seeds, suggesting that dwarfism in the seed may be indirectly caused by brassinosteroid deficiency. The mechanism of seed size reduction in this mutant was investigated.

• Methods The associations between seed orientation in the pod, seed numbers per pod and pod lengths with seed sizes were analysed in ‘Rinrei’ and the wild-type plant.

• Key Results ‘Rinrei’ seeds are tightly arranged in pods containing two or three seeds. Seed size decreased as the number of seeds per pod increased or as the length of the pod decreased. Where no physical restriction occurred between seeds in a pod, the wild-type faba bean seeds had a nearly constant size regardless of seed number per pod or pod length. ‘Rinrei’ seeds in pods containing single seeds were the same size as wild-type seeds. Brassinolide treatment increased the seed size and the length of pods containing three seeds in ‘Rinrei’.

• Conclusion Seed size of ‘Rinrei’ is mainly regulated through a reduction of pod length due to brassinosteroid deficiency; physical restriction within pods causes a reduction in seed size. These results suggest a possible mechanism for increasing faba bean yields to optimal levels.

INTRODUCTION

Dwarfism in many plants is often derived from deficiency in, or insensitivity to, growth-promoting hormones such as gibberellins and brassinosteroid. Faba bean (Vicia faba) ‘Rinrei’ is a brassinosteroid-deficient semi-dwarf mutant generated by gamma-ray irradiation (Fukuta et al., 2004). ‘Rinrei’ is defective in the conversion of 24-methylenecholesterol to campesterol (Fukuta et al., 2004) and the mutant gene is probably a homologue of lkb of pea (Pisum sativum) (Nomura et al., 1999; Schulz et al., 2001) and dwf1 (dim, cbb1) of Arabidopsis thaliana (Takahashi et al., 1995; Kauschmann et al., 1996; Klahre et al., 1998; Choe et al., 1999a). Loss of this conversion ability results in a decrease in endogenous levels of active brassinosteroids.

Brassinosteroids are steroidal plant hormones that participate in many growth processes, including cell elongation, cell division, vascular differentiation, seed germination and photomorphogenesis (Mandava, 1988; Clouse and Sasse, 1998; Bishop and Yokota, 2001). ‘Rinrei’ is characterized by reductions in plant height, internode length and petiole length (Fukuta and Yukawa, 1998). These characteristics are restored to those of wild-type plants by brassinolide application (Fukuta et al., 2004). The dwarfed phenotype of ‘Rinrei’ is also expressed in the size of the seed, as observation also reported in brassinosteroid-deficient dwarf mutants of Arabidopsis dwarf5-1 (Choe et al., 2000) and rice (Oryza sativa) brd1 (Mori et al., 2002). Investigations of these diverse plant species indicate that brassinosteroids seem to regulate seed growth as well as plant height.

In ‘Rinrei’, normal-sized seeds occur together with dwarf seeds, suggesting that dwarfism of the seed may not be directly caused by brassinosteroid deficiency. This study examined seed number per pod and pod growth as potential factors that might regulate seed size. Faba bean has a relatively large seed size compared with other crops and usually each pod contains only a few seeds. This property provides an advantage for investigating the relationship between seed number and size over rice, which has only a single seed per hull, and Arabidopsis, which has very many small seeds in each silique. The influence of seed number per pod on seed size was also studied. Seed size regulation through pod growth in faba bean is discussed.

MATERIALS AND METHODS

Varieties

‘Rinrei’ was created by gamma-ray irradiation of ‘Niigata-zairai’ seedlings, a native variety of faba bean (Vicia faba) from Niigata Prefecture, Japan. Cultivar ‘Mild Green’ was selected from the same native variety and is regarded as the original wild-type progenitor of ‘Rinrei’. Seeds of both varieties were purchased from Hokuetsu-nouzi Co. (Niigata, Japan) (Fukuta et al., 2004).

Plant growth

Seeds of ‘Rinrei’ and ‘Mild Green’ were imbibed for 5 d by floating on tap water in October, 2001. Germinated seeds were sown in vinyl pots 12 cm in diameter containing horticultural soil (Kureha Chemical, Tokyo, Japan) and grown in a growth chamber at 25 °C under natural daylight for 1 month. Seedlings were transplanted to the field at the National Institute of Floricultural Science during November, 2001. Flowering of these plants started at the beginning of April to the end of May, 2002. Three wild-type plants and 13 ‘Rinrei’ plants were obtained, five of which were used for the brassinolide-treatment experiment. Approximately 45 and 20 pods per plant were produced on wild-type and ‘Rinrei’ plants, respectively. For each experimental treatment, 100–150 pods were analysed.

Analyses

Pod length was monitored during seed development, beginning 7–8 d after flowering. Seed size (length in the longitudinal dimension of the oblong-shaped seed), pod length, seed numbers per pod and seed orientation were analysed using pods harvested 30 d after flowering. The number of ovules was calculated by adding the numbers of aborted seeds to the number of mature seeds in each pod.

Brassinolide application

Brassinolide was purchased from Wako Pure Chemical Industry, Ltd (Osaka, Japan) and was dissolved in water to a final concentration of 2 µm, a concentration close to saturated. The aqueous saturated brassinolide solution was applied directly to each pod on five ‘Rinrei’ plants daily for 2 d from the end of April to the beginning of May, 2001. In total about 0·2 nmol of brassinolide was applied to the surface of a pod.

RESULTS

Pod and seed morphology

Pod lengths of the wild-type ‘Mild Green’ and a brassinosteroid-deficient dwarf mutant, ‘Rinrei’, were measured after flowering. Figure 1 shows the length of a single-seeded pod from both varieties as representative data for pod growth. Pods produced by the wild-type plant and ‘Rinrei’ reached their maximum size at 23 and 25 d, respectively. Pods and seeds harvested 30 d after flowering were used for all analyses as representative material for both varieties. The morphology of seeds and pods of the wild-type and mutant is shown in Fig. 2. There was enough space in the wild-type pods for each seed to develop separately. By contrast, the internal pod space of the ‘Rinrei’ pods was limited and the seeds grew in close contact.

Fig. 1.

Growth of single-seeded pods during seed development and monitored after flowering in wild-type plants and ‘Rinrei’. Early flowering individuals were chosen for the monitoring.

Fig. 2.

Pod morphology and seed proximity of a wild-type plant (left) and ‘Rinrei’ (right). Single-seeded pods (bottom), double-seeded pods (middle) and triple-seeded pods (top).

Seeds and ovule numbers in a pod or an ovary

The majority of wild-type pods contained two seeds, whereas the majority of ‘Rinrei’ mutants had three seeds per pod (Fig. 3A). Wild-type plants had no pods containing four seeds, whereas in ‘Rinrei’ the second highest proportion of pods had four seeds. Average seed numbers in a pod were 2·2 and 3·1 in the wild-type and ‘Rinrei’ plants, respectively. The majority of wild-type pods contained three ovules and the second most numerous pods contained two ovules (Fig. 3B). There were more ovules per pod in ‘Rinrei’, with the highest number of pods containing four ovules, followed by three ovules per pod. Average ovule numbers in a pod were 2·9 and 3·8 in the wild-type plant and ‘Rinrei’, respectively. Brassinolide application on the pod surface did not have a significant effect on seed number per pod in ‘Rinrei’, with the majority still producing three seeds per pod (data not shown).

Fig. 3.

Percentage of seed numbers (A) and ovule numbers (B) in pods of the wild-type plant (left) and ‘Rinrei’ (right).

Pod length and seed size

Wild-type seeds were similar in size regardless of pod length and seed numbers per pod (Fig. 4A–C). ‘Rinrei’ seeds from pods containing multiple seeds tended to be reduced in size (Fig. 4B and C), while seeds occurring in pods containing a single seed were similar in size to wild-type seeds (Fig. 4A). The length distribution of single-seeded pods of ‘Rinrei’ overlapped with that of the wild-type plants, with values for the former ranging from 65 to 85 mm. Wild-type pods containing single seeds were up to 95 mm long. All seeds from pods containing a single seed were about 26 mm long for both varieties regardless of pod size.

Fig. 4.

Seed sizes in pods containing (A) single seeds, (B) two seeds and (C) three seeds of wild-type (open circles), ‘Rinrei’ (closed squares) and brassinolide-treated ‘Rinrei’ plants (closed triangles). Vertical bars indicate s.e.

The length distribution of pods containing two seeds also partially overlapped between the wild type and the mutant (Fig. 4B) and the seed sizes of ‘Rinrei’ increased from 18 to 24 mm as pod length increased. Seed size also increased slightly with increased pod length in the wild type, but many of the seeds were about 26 mm in length. The two varieties had similar sized seeds in the same sized pods.

Pod length distributions of the wild type and ‘Rinrei’ mutant did not overlap in pods containing three seeds per pod (Fig. 4C). Pod lengths ranged from 80 to 100 mm in ‘Rinrei’ and from 105 to 135 mm in the wild type. The seed sizes of ‘Rinrei’ increased from 17 to 21 mm as pod length increased. The seed length distribution of the brassinosteroid-treated ‘Rinrei’ ranged from 20·5 to 24·5 mm, with the distribution of pod lengths partially overlapping both that of ‘Rinrei’ and the wild type. Regardless of pod length, seed sizes were about 26 mm in three-seeded pods of wild-type plants.

Effect of brassinolide

Sufficient numbers of samples were obtained only from pods containing three seeds per pod; therefore, brassinolide treatment of the mutant was limited to this pod class. The average mature pod length was 112 mm for wild-type plants and 88 mm for ‘Rinrei’. Application of brassinosteroid partially compensated for the effect of the mutation, resulting in an average mature pod length of 100 mm (Fig. 4C). Brassinolide application also increased the average ‘Rinrei’ seed size to 21 mm, a value overlapping that for the mutant (19 mm) and for the wild type (25 mm).

DISCUSSION

Average pod length and seed size of the brassinosteroid-deficient mutant faba bean ‘Rinrei’ were reduced to about 80 % of the values for the wild-type ‘Mild Green’. Application of brassinolide partially restored seed length in the mutant, suggesting that seed size, in addition to plant height (Fukuta et al., 2004), is regulated by brassinosteroid. It is noteworthy that seed length was not reduced for seeds that developed individually in pods containing single seeds (Figs 2 and 4). This observation suggests that seed size regulation by brassinosteroid is indirect.

Seed size reduction of ‘Rinrei’ was found in pods containing two or three seeds where the seeds were in contact with neighbouring seeds (Figs 2 and 4). The tightly packed arrangement of ‘Rinrei’ seeds in pods indicates that physical restriction of each seed may have occurred. Reduction of seed size did not appear in ‘Rinrei’ without the restriction when only one seed developed per pod. As the number of seeds per pod increased in ‘Rinrei’, smaller seeds developed in pods of a similar length. The correlation between seed size and pod length was found also in seeds from brassinolide-treated pods. It therefore appears that brassinosteroid deficiency causes dwarfing of faba bean seeds mainly by physical restriction within reduced length of pods.

The increased number of ovules per pod found in ‘Rinrei’ probably leads to the increased seed number per pod compared with wild-type plants (Fig. 3), leading to additional effects on the reduction of average seed size. This indicates that brassinosteroids may be involved with determining ovule and seed number per pod in faba bean. In contrast to ‘Rinrei’, some Arabidopsis brassinosteroid-deficient mutants (dwarf1-1, Choe et al., 1999a; dwarf4, Azpiroz et al., 1998; dwarf5-2, 3, 4, 5 and 6, Choe et al., 2000; dwarf7-1, Choe et al., 1999b) and receptor mutants (BRI1-4, 5 and 7, Noguchi et al., 1999) were reported to have reduced seed numbers per silique. The reduced vegetative growth of ‘Rinrei’ might promote the supply of assimilates to the reproductive organs, resulting in increased numbers of ovules and seeds per pod. This follows the same argument used to explain the increase in production yield in dwarf rice and wheat in the ‘green revolution’ (Gale and Youssefian, 1985; Evans, 1993; Peng et al., 1999).

Seed size and silique length reductions have also been reported in an Arabidopsis brassinosteroid-deficient mutant dwf5-1 (Choe et al., 1999a), in which seed numbers per pod were similar to those of the wild type. Dwarfism of seeds and hulls was also reported in a brassinosteroid-deficient mutant of rice (brd1, Mori et al., 2002). Seed size reduction of these mutants may be due to physical restriction by the matrix as in ‘Rinrei’. Each organ must be separately analysed in order to clarify the effect of the matrix on seed growth; however, this type of analysis presents challenges for single (rice) or multiple (Arabidopsis) number seeds in the matrix. Among Leguminosae plants, a brassinosteroid-deficient mutant of pea lkb has been described (Nomura et al., 1997, 1999) but the morphology of the seed and pod of this mutant has not been discussed (Reid and Ross, 1989).

‘Rinrei’ has a higher snow tolerance than other taller faba bean varieties as a result of its dwarf phenotype (Fukuta and Yukawa, 1998; Fukuta et al., 1999, 2001; Fukuta, 2000). Although ‘Rinrei’ is regarded as a suitable variety for cultivation in snowy districts, the small size of its seeds reduces its commercial value. Brassinolide application to ‘Rinrei’ in spring after the snowmelt may be a promising technique to increase seed size without reducing the benefit of its greater tolerance.

The data presented in Fig. 4 indicate that seed size of the ‘Mild Green’–‘Rinrei’ line faba bean increases with increasing pod length. The minimum pod length for single-seeded pods was 65 mm, whereas those for double- or triple-seeded pods were 90 and 105 mm, respectively, to give a maximum seed size of 26 mm. Pod growth over the length described above cannot increase seed size. Sufficient pod growth to support the growth of full-sized seeds may gain optimal yields at the expense of pod biomass. These results suggest a new strategy for increased yields of faba bean via restriction of pod growth.