Storage protein gene expression has been studied in relation to mitotic activity to ascertain whether these processes are linked during embryo development in pea. Sections from immature pea embryos were probed by in situ hybridisation to show the pattern of vicilin storage protein gene expression. In addition, the location of mitotic cells was identified using fluorescence microscopy. Vicilin mRNA was first localized in the parenchyma cells of the upper adaxial region of the cotyledons. As the embryos increased in fresh weight, gene expression spread from this region, in a wave-like manner, down and across the cotyledons. The gene was only expressed in those regions of the embryo that lacked mitotic activity.

The pattern of growth of legume seeds follows a sigmoidal curve, beginning with a phase of cell division, where there is little increase in absolute mass, followed by a phase of cell expansion (Pate and Flinn, 1971). During the phase of cell expansion, storage protein production occurs, which closely follows the increase in fresh weight (Dure, 1975; Higgins, 1984). There is genetic variation for the timing of the transition from the cell division to the cell expansion phase and for the duration of the phases (Hedley and Smith, 1985). Towards the end of embryo development, storage protein synthesis ceases and the rate of fresh weight increase declines.

The timing of storage protein gene expression during the development of legume seeds has been analysed by ‘Northern’ hybridisation (Meinke et al. 1981; Gatehouse et al. 1982). This method, however, does not allow differences between levels of expression in different cells of the embryo to be seen. Gene expression can be studied in greater detail, at the cellular level, using in situ hybridisation (Angerer and Angerer, 1981; Cox et al. 1984). Such studies have been carried out on the spatial distribution of legume seed storage protein mRNA in Pisum sativum, using cDNA probes (Harris et al. 1989), and in Glycine max, using cRNA probes (Perez-Grau and Goldberg, 1989). In these two reports, the expression of storage protein genes was found to be tissue specific within the developing embryos, and to follow different temporal and spatial patterns during development. However, these patterns were not related to other processes taking place during the development of the embryo.

Smith (1973) reported the spatial distribution of protein production in developing embryos at the cellular level. This was related to the distribution of mitotic activity and levels of RNA across the sections of embryos. Corke et al. (1987) have developed immunocytochemical techniques to look at the deposition of the storage protein, vicilin, and the DNA content in the same cell from developing embryos. It was suggested that there was a correlation between increasing DNA level, resulting from endoreduplication of DNA, and the timing of vicilin production in the parenchyma cells of the embryo (Corke, 1988; Corke et al. 1990). Furthermore, the latter event was also related to cellular expansion.

In the work reported here, we examine the spatial relationship between the production of mRNA for the storage protein, vicilin, using in situ hybridisation with cRNA probes. These data have been related to the distribution of mitotic cells, using fluorescence microscopy, in sections from young Pisum sativum embryos. We have found that the spatial expression of vicilin mRNA differs from both the previous reports on legumes (Harris et al. 1989; Perez and Goldberg, 1989), but it can be related to the distribution of mitotic cells in the cotyledons.

Plant material

Pea plants of genotype JI 181 were grown as described in Corke et al. (1987).

In situ hybridisation

The DNA clone pCD4 (Domoney and Casey, 1983), a gift from Dr C Domoney of the John Innes Institute, was inserted into the transcription vector pGEM 4-Z (Promega), with a T7 and Sp6 transcription promoter, one either side of the inserted DNA.

In situ hybridisation methodology was based on that described in Cox and Goldberg (1988). Pea embryos were removed from the testa, weighed and fixed in paraformaldehyde (4% w/v) in phosphate-buffered saline (0.01M PBS, pH7.2) at 4°C overnight. The tissue was dehydrated in a graded ethanol and Histoclear (Agar Aids) series and embedded in Paramat paraffin wax (BDH). Sections, 10 μm thick, were dried onto poly-L-lysine-coated slides, the paraffin wax removed with Histoclear and rehydrated through a water/ ethanol series. The sections were subsequently pretreated with 125 μgml-1 pronase (Sigma type XIV) in 50 mM Tris-HCl, pH7.5, 5mM ethylenediamine tetraacetic acid (EDTA) at room temperature (RT) for 10min, fixed in 4% w/v paraformaldehyde in PBS, 10 min RT, incubated in 0.5 % v/v acetic acid in 0.1M triethanolamine pH8.0, 10min RT, dehydrated in an ethanol series and air dried.

Antisense and sense (control) RNA probes were made by transcribing the pGEM 4-Z clones. The RNA probes were radioactively labelled with [3H]CTP (1mCiml-1, NEN) and hydrolysed to 150bp, at 60°C, with carbonate hydrolysis buffer (120mM Na2CO3/80mM NaHCO3, pH 10.2). Sections were hybridised with 0.2ng μl-1 cRNA probe in 50% v/v formamide, 10% w/v dextran sulphate, 300mM NaCl, 10 μm Tris, ImM EDTA, 1 × Denhardt’s, 75units μl-1 RNAsin (Amersham), 1 μg μl-1E. coli tRNA, for 16h at 50°C. After washing 3 times at 50°C in wash buffer (50 % v/v formamide, 2 × standard saline citrate [2 × SSC; 0.06M sodium citrate, 0.6M sodium chloride]), the slides were treated with 20 × gml-1 RNase A (Sigma) in 0.5M NaCl, I0mM Tris-HCl pH7.5, 1mM EDTA, for 30min at 30°C. The final wash consisted of 50% v/v formamide, 2 × SSC 50°C then 1 × SSC at RT. The sections were dehydrated through an ethanol series (each with 300 mM ammonium acetate) and 100% ethanol.

Following drying, the slides were coated with Ilford K5 emulsion diluted 1:1 with 0.1% v/v glycerol in water and exposed for 4 – 5 weeks at 4 °C. They were developed in Kodak D-19 developer for 2 min, stopped with 1% v/v acetic acid, 1 % v/v glycerol, and fixed in 30 % w/v sodium thiosulphate, all at 14°C. Sections were rinsed and stained with 0.025 % w/v toluidine blue and mounted under Canada balsam before viewing using a Zeiss axiophot microscope under dark-field and bright-field illumination. Photographs were taken with Kodak EPT 160 film.

Nuclear DNA staining

Rehydrated wax sections were incubated in a solution of 4’6-diamidino-2-phenylindole (DAPI; 1 μgml-1 in distilled water) for 5 min RT. Slides were drained and coverslips mounted on antifade Citifluor glycerol mountant (AF2 Citifluor, City University, London). Sections were viewed under epifluorescence with Plan-neofluar objectives (×10 NA=0.3; × 40 NA=0.75). DAPI-stained nuclei were photographed at the lower magnification and photographs compiled to show the entire section. These montages were then used to pinpoint mitotic cells.

1. Gene expression is temporally and spatially regulated in the cotyledon

Pea embryos were weighed, fixed and embedded in wax. Ten embryos over the weight range of 8 – 18 mg were sectioned, yielding 150 – 300 sections each. In situ hybridisation was performed on these sections, using a tritium-labelled RNA probe complementary to pCD4, a vicilin pea storage protein cDNA (Domoney and Casey, 1983). The sections described here are representative of the changes in the overall pattern of vicilin mRNA production which were seen to occur through this period of embryo development. Hybridisation was first detected in embryos of 10 mg fresh weight, when mRNA was localised in the parenchyma cells of the adaxial (inner) region of the cotyledons (Fig. 1). There was a high level of hybridisation (white grains) to the cytoplasm of these cells (Fig. 2). At this stage, however, all the vicilin mRNA was confined to this small group of cells, with no labelling of the axis, the epidermal cells or even the remainder of the parenchyma. In the lower (distal) region of the cotyledons, there was also little labelling of the parenchyma cells (Fig. 3).

Fig. 1.

Longitudinal wax section of a 10 mg JI 181 pea embryo, hybridised with tritium-labelled pCD4 cRNA. Composite of bright-field illumination (A), to show nuclear staining, and the dark-field mirror image (B), where the hybridisation signal is seen as white grains. Vicilin mRNA is localised in the cells in the inner region of the cotyledon (c), but not in the axis (a). Bar=350 μm.

Fig. 1.

Longitudinal wax section of a 10 mg JI 181 pea embryo, hybridised with tritium-labelled pCD4 cRNA. Composite of bright-field illumination (A), to show nuclear staining, and the dark-field mirror image (B), where the hybridisation signal is seen as white grains. Vicilin mRNA is localised in the cells in the inner region of the cotyledon (c), but not in the axis (a). Bar=350 μm.

Fig. 2.

Same section as Fig. 1, proximal region of the cotyledon, at higher magnification. Vicilin mRNA is localised in the adaxial region (ad) of the cotyledon, with absence from the axis (a) and epidermal ceils (e). Bright-field illumination (A), dark-field (B). Bar=110 μm.

Fig. 2.

Same section as Fig. 1, proximal region of the cotyledon, at higher magnification. Vicilin mRNA is localised in the adaxial region (ad) of the cotyledon, with absence from the axis (a) and epidermal ceils (e). Bright-field illumination (A), dark-field (B). Bar=110 μm.

Fig. 3.

Same section as Fig. 1, distal region of the cotyledon, at higher magnification. Vicilin mRNA is localised in some, but not all, of the adaxial (ad) parenchyma cells. Those towards the bottom of the cotyledon are not labelled for the mRNA at this stage of development. Bright-held illumination (A), dark-field (B). Bar=110 μm.

Fig. 3.

Same section as Fig. 1, distal region of the cotyledon, at higher magnification. Vicilin mRNA is localised in some, but not all, of the adaxial (ad) parenchyma cells. Those towards the bottom of the cotyledon are not labelled for the mRNA at this stage of development. Bright-held illumination (A), dark-field (B). Bar=110 μm.

In transverse sections of a 14 mg embryo, the distri-bution of vicilin mRNA labelling was, again, nonuniform (Fig. 4). Labelling for mRNA was restricted to one sector of the cotyledon. This sector was towards the adaxial surface and to one face of the cotyledons, which imparted an asymmetric distribution of the mRNA across the cotyledons. Not all cells within this sector were labelled - there was a group of small cells that were devoid of labelling (Fig. 4B). These were identified as provascular tissue (Craig et al. 1979). Furthermore, the remainder of the cotyledon was unlabelled, including the abaxial cells, as also seen in Fig. 1.

Fig. 4.

Transverse section of a 14 mg JI 181 embryo, from about one-third down the embryo from the axis, hybridised as for Fig. 1. Vicilin mRNA is localised in the sector towards the adaxial surface. There is a small group of cells within this sector that are not labelled, provascular cells (pv). Bright-field illumination (A), dark-field (B). Bar =110 μm.

Fig. 4.

Transverse section of a 14 mg JI 181 embryo, from about one-third down the embryo from the axis, hybridised as for Fig. 1. Vicilin mRNA is localised in the sector towards the adaxial surface. There is a small group of cells within this sector that are not labelled, provascular cells (pv). Bright-field illumination (A), dark-field (B). Bar =110 μm.

With an increase in fresh weight of embryo, the labelling for vicilin mRNA increased to cover most of the cotyledon, as seen in the section from an 18 mg embryo shown in Fig. 5. There was still no detection of the message in the axis, epidermis or the provascular tissue, which could be seen at this stage as a strip of cells with no labelling above them (Fig. 5). The parenchyma cells adjacent to the adaxial surface and at the bottom of the cotyledons were not labelled.

Fig. 5.

Longitudinal wax section from an 18 mg embryo, hybridised as for Fig. 1. Labelling for vicilin mRNA is found throughout the section, except in the axis (a), provascular tissue (pv) and region next to the abaxial surface (ab). Composite of bright-field illumination (A) and the dark-field mirror image (B). Bar=335 μm.

Fig. 5.

Longitudinal wax section from an 18 mg embryo, hybridised as for Fig. 1. Labelling for vicilin mRNA is found throughout the section, except in the axis (a), provascular tissue (pv) and region next to the abaxial surface (ab). Composite of bright-field illumination (A) and the dark-field mirror image (B). Bar=335 μm.

Sections from embryos at all stages of development, hybridised with tritium-labelled control (mRNA) probe showed no labelling over the sections (results not shown).

2. Gene expression correlates to areas of reduced mitotic activity in the developing cotyledon

Sections (150 – 300 per embryo) from four embryos between 10 – 20 mg fresh weight were treated with a DNA stain and cells in mitosis identified by fluorescence microscopy. Fig. 6 shows the location of mitotic cells in similar sections from embryos of the same fresh weight range labelled for vicilin mRNA and shown in Figs 1 – 5. In Figs 1 – 3, the region in which vicilin mRNA was located was related to the region that lacked mitotic activity in equivalent weight embryos (Fig. 6A). Mitotic cells were seen in the axis, epidermis and abaxial region. The same was found in transverse sections (Fig. 6B). Mitotic cells were located mainly in the abaxial region, which was the sector where vicilin mRNA localisation was reduced in Fig. 4. There was a small group of mitotic cells within this sector. These were located in the region of provascular cells (Fig. 4A) which were not labelled for vicilin mRNA. The provascular tissue in Fig. 5 (18 mg) corresponded to a similar region in a 20 mg embryo which was found to contain mitotic cells (Fig. 6C). The sections shown in Fig. 6 illustrate a zone of mitotic activity towards the abaxial surface and distal end of the cotyledons with small areas delineating the provascular tissue. Fig. 7 shows a diagrammatical representation of the general pattern of mitotic activity which was seen in sections from embryos of this developmental range. The region where in situ hybridisation localised vicilin expression, therefore, showed little, if any mitotic activity.

Fig. 6.

Line diagrams to show the location of mitotic cells in DAPI-stained sections from JI 181 embryos of various fresh weights. Longitudinal (6A) and transverse (6B) sections of 10mg embryo and longitudinal section of 20mg embryo (6C). Bar=350 μm.

Fig. 6.

Line diagrams to show the location of mitotic cells in DAPI-stained sections from JI 181 embryos of various fresh weights. Longitudinal (6A) and transverse (6B) sections of 10mg embryo and longitudinal section of 20mg embryo (6C). Bar=350 μm.

Fig. 7.

Diagrammatic representation of the overall pattern of mitotic activity in embryos of the same developmental range as used for mRNA studies. (A) longitudinal section, (B) transverse section.

Fig. 7.

Diagrammatic representation of the overall pattern of mitotic activity in embryos of the same developmental range as used for mRNA studies. (A) longitudinal section, (B) transverse section.

Vicilin gene expression has been localised in regions of reduced mitotic activity, such as the adaxial region, where the message can first be detected by in situ hybridisation at an early stage of embryo development. As the embryo develops and grows, the regions in the cotyledons where mitosis is occurring are reduced and, correspondingly, the region over which vicilin gene expression is localised, expands in a wave-like manner through the embryo. This wave-like pattern of expression, beginning at the upper adaxial region and spreading down and across the cotyledons, has been seen also for the initiation of legumin mRNA synthesis (results not shown). The two storage proteins show similar spatial patterns of labelling, but legumin has a different temporal pattern since it is detectable slightly later in development.

Harris et al. (1989) studied the expression of legumin and vicilin genes in pea, but did not make comparisons with other processes taking place in the developing embryo. The same pattern of tissue specificity for labelling within the embryo was seen, with no labelling of the provascular tissue, the epidermis or the axis at the early stages of development. However, they suggested that the onset of expression of these genes was synchronous through the parenchyma cells of the cotyledons and hence very different to the pattern reported here. We suggest that the growth conditions that Harris et al. employed contributed to these differences, since they used growth conditions 13 °C above those used here, which would have produced rapid and condensed growth. This, together with the lower frequency of sampling, would have meant that the ‘wave’ of transcriptional activity probably would not have been resolved. The fact that two different genotypes of pea were used in the two studies may also have contributed to the different results.

Using in situ hybridisation, a similar investigation of gene expression in soybean embryos was reported by Perez-Grau and Goldberg (1989). In mid-maturation soybean embryos, a ‘wave’ of transcriptional activity of several different genes, including storage protein genes, was detected from the outer (abaxial) surface to the inner (adaxial) surface of the cotyledons. Different proteins were expressed at different times in development, but followed the same tissue specificity as we found for vicilin and legumin expression. However, the ‘wave’ was in the opposite direction to that reported here for pea. The abaxial localisation of protein mRNAs in soybean was maintained for a longer time than the adaxial localisation in pea, but in mid-maturation (60 mg) pea embryos, the vicilin mRNA was distributed evenly across the sections (unpublished data). Nevertheless, the direction of the ‘wave’ of activity observed by Perez-Grau and Goldberg cannot be a general phenomenon as held by the authors, since in another legume, pea, the ‘wave’ is different. The difference in transcriptional activity is more likely to be due to differences in the ontogeny of the two legume embryos. Moreover, soybean storage protein genes are expressed at a later stage in development than in pea (Meinke et al. 1981; Gatehouse et al. 1982).

In pea, the pattern of vicilin mRNA production, when first localised in 10mg embryo sections, correlates with a similar pattern for reduced mitotic activity in embryos of the same fresh weight. The same gene expression pattern was seen for older embryos, but covering a greater number of the cells. The ‘wave’ of transcriptional activity reported here, spreading through the embryo sections with increasing fresh weight, takes place over a very short time, in developmental terms. The genotype II 181 was chosen for the study of protein production at the cellular level because the switch between an embryo containing mostly cells in division to one with cells mostly in expansion is quite distinct (Hedley and Smith, 1985). Corke et al. (1990) suggested that the switch from the phase of division to expansion may be important in determining the timing of the initiation of storage protein synthesis. This is reinforced by the fact that storage protein has never been detected in mitotic cells in previous studies (Corke, 1988; Corke et al. 1987, 1990).

Smith (1973) studied the pattern of mitosis and level of total RNA across sections of Pisum arvense embryos. Cessation of mitosis was found to occur first in the adaxial cells, which agrees with results reported here. However, the level of RNA was found to be higher in the abaxial region, during mid-maturation, than in the adaxial. This study involved staining for all RNA and was suggested to be mostly indicative of ribosomal RNA levels, which will have masked differences in the levels of other types of RNA. The results from in situ hybridisations to a specific mRNA, reported here, have given better resolution of the changes in the level of the RNA, taking place during the development of the embryo.

The results from this work highlight the importance of combining information on processes taking place concurrently during development. It is clear that the pattern of transcriptional activity of seed protein genes is different between plant species. It may be possible to relate these differences to the ontogeny of the embryos and as such, the cellular development, especially mitotic activity, may be a useful indicator.

This work was supported by the Agricultural and Food Research Council via a grant-in-aid to the John Innes Institute. Support was also received in the form of a studentship for AJ Hauxwell from the John Innes Foundation. We would like to thank Dr Claire Domoney for the gift of the pCD4 probe and Peter Scott, Andrew Davies and Nigel Hannant for photographic assistance.

Angerer
,
L. M.
and
Angerer
,
R. C.
(
1981
).
Detection of poly A+ RNA in sea urchin eggs and embryos by quantitative in situ hybridisation
.
Nucl. Acid Res
.
9
,
2819
2840
.
Corke
,
F. M. K.
(
1988
).
Immunocytochemical investigation of pea seed development
.
Ph.D. Thesis, University of East Anglia, UK
.
Corke
,
F. M. K.
,
Hedley
,
C. L.
,
Shaw
,
P. J.
and
Wang
,
T. L.
(
1987
).
An analysis of seed development in Pisum sativum V. Fluorescence triple staining for investigating cotyledon cell development
.
Protoplasma
140
,
164
172
.
Corke
,
F. M. K.
,
Hedley
,
C. L.
and
Wang
,
T. L.
(
1990
).
An analysis of seed development in Pisum sativum XI. Cellular development and the position of storage protein in immature embryos grown in vivo and in vitro
.
Protoplasma (in press)
.
Cox
,
K. H.
,
De Leon
,
D. V.
,
Angerer
,
L. M.
and
Angerer
,
R. C.
(
1984
).
Detection of mRNAs in sea urchin embryos by in situ hybridisation using asymmetric RNA probes
.
Devl Biol
.
101
,
485
502
.
Cox
,
K. H.
and
Goldberg
,
R. B.
(
1988
).
An analysis of plant gene expression
.
In Plant Molecular Biology, a Practical approach
(ed.
C. H.
Shaw
), pp.
1
35
.
IRL Press
:
Oxford
.
Craig
,
S.
,
Goodchild
,
D. J.
and
Hardham
,
A. R.
(
1979
).
Structural aspects of protein accumulation in developing pea cotyledons. (I) Qualitative and quantitative changes in parenchyma cells
.
Aust. J. Pl. Physiol.
6
,
81
98
.
Domoney
,
C.
and
Casey
,
R.
(
1983
).
Cloning and characterisation of complementary cDNA for convicilin, a major seed storage protein in Pisum sativum L
.
Planta
159
,
446
453
.
Dure
,
L. S.
(
1975
).
Seed formation
.
A. Rev. Pl. Physiol
.
26
,
259
278
.
Gatehouse
,
J. A.
,
Evans
,
I. M.
,
Croy
,
R.R.D.
,
Boulter
,
D.
(
1982
).
Differential expression of genes during legume seed development
.
Phil. Trans. R. Soc. Land. B
.
314
,
367
384
.
Harris
,
N.
,
Grindley
,
H.
,
Mulchrone
,
J.
and
Croy
,
R. R. D.
(
1989
).
Correlated in situ hybridisation and immunocytochemical studies of legumin storage protein deposition in pea (Pisum sativum L)
.
Cell Biol. Int. Reps
.
13
,
23
35
.
Hedley
,
C. L.
and
Smith
,
C. M.
(
1985
).
Genetic variation for peaseed development
.
In The Pea Crop
(ed.
P. D.
Hebbletwaite
,
M. C.
Heath
,
T.C.K.
Dawkins
), pp.
329
338
.
Butterworths
:
London
.
Higgins
,
T. J. V.
(
1984
).
Synthesis and regulation of major proteins in seeds
.
A. Rev. Pl. Physiol
.
35
,
191
221
.
Meinke
,
D. W.
,
Chen
,
J.
and
Beachy
,
R. N.
(
1981
).
Expression of storage protein genes during soybean seed development
.
Planta
153
,
130
139
.
Pate
,
J. S.
and
Flinn
,
A. M.
(
1971
).
Fruit and seed development
.
In The Physiology of the Garden Pea
(ed.
J. F.
Sutcliffe
and
J. S.
Pate
), pp.
431
468
.
Academic Press
:
London
.
Perez-Grau
,
L.
and
Goldberg
,
R. B.
(
1989
).
Soybean seed protein genes are regulated spatially during embryogenesis
.
The Plant Cell
1
,
1095
1109
.
Smith
,
D. L.
(
1973
).
Nucleic acid, protein and starch synthesis in developing cotyledons of Pisum arvense L
.
Ann. Bot
.
37
,
795
804
.