Arc6, an extreme chloroplast division mutant of Arabidopsis also alters proplastid proliferation and morphology in shoot and root apices

Several nuclear ARC loci which control chloroplast development in leaf mesophyll cells have been identified by mutant analysis inArabidopsis thaliana (Pyke and Leech, 1992, 1994; Pyke et al., 1994). These mutants were identified by their reduction in or increase in chloroplast number per leaf mesophyll cell compared to wild type. The most extreme of these mutants, arc6 (Pyke et al., 1994), has only two enormously enlarged chloroplasts per mesophyll cell, which do not increase in number during mesophyll cell expansion, inferring that chloroplast division has been severely perturbed in arc6 mesophyll cells. Consequently arc6 is the most radically altered chloroplast number mutant in higher plants. Despite the abnormal chloroplast morphology, arc6 plants grow normally with no dramatic change in whole plant phenotype and are fully fertile (Pyke et al., 1994). The mutant phenotype of arc6 mesophyll cells is highly stable and segregates as a single nuclear locus.

Since the number of chloroplasts in arc6 cells is considerably lower than the estimated proplastid number per mitotic cell in Arabidopsis, of 14 (Pyke and Leech, 1992, 1994), it is possible that proplastids in young arc6 seedlings may also be perturbed in a radical way (Pyke et al., 1994). Proplastids are the progenitor organelles from which specific types of plastid can differentiate and are characteristic of meristematic cells in shoot and root apices (Kirk and Tilney-Bassett, 1978; Thomson and Whatley, 1980). Chloroplasts in mature leaf cells develop from proplastids in younger cells (Possingham, 1980), whereas in roots plastid differentiation leads to the formation of amyloplasts, in particular statoliths in the columella cells and other small root plastids (Whatley, 1983). Although the shoot and root meristems of Arabidopsis have been relatively well characterised morphologically (Medford et al., 1992; Dolan et al., 1993), proplastid development in Arabidopsis meristems has not been examined. Reports of the few studies that have been made of other higher plants indicate that meristematic cells at the shoot apex contain between 10 and 20 proplastids (Cran and Possingham, 1972; Lyndon and Robertson, 1976; Fujie et al., 1994). In order to maintain the continuity of plastids through cell lineages in the plant, proplastids must divide during mitosis and segregate into daughter cells at cell division in an ordered manner (Birky, 1982; Possingham and Lawrence, 1983) and it is generally assumed that cells lacking plastids as a consequence of abnormal proplastid segregation would not be viable. The rate of proplastid division in relation to cell division is also critical to competent cell development, for the number of proplastids in the young post-mitotic cell will influence the subsequent size of the plastid cell population arising by differentiation and plastid division.

The availability of the Arabidopsis arc6 mutant with it’s normal fertile plant phenotype gives us a unique opportunity to examine the control of plastid development in the tissues of a higher plant. To gain an insight into the role of ARC6 in the control of early plastid development we have examined cells from the shoot meristem, root meristem and stomatal guard cells of the wild-type (ARC6) and the arc6 mutant of Arabidopsis to determine how early in plastid development the arc6 mutant phenotype can be recognised.

Wild-type Arabidopsis plants of the ecotype Wassilewskija (WS) and the arc6 mutant were grown in controlled conditions as described previously (Pyke and Leech, 1991). The arc6 mutant was isolated from a T-DNA mutagenised Arabidopsis population in the WS background (Pyke et al., 1994).

For ultrastructural analysis, shoot and root apices were harvested from 5-day-old seedlings. The shoot apex and the first leaf primordium of young Arabidopsis seedlings were examined after fixation and embedding in Spurr’s resin, as previously described for Arabidopsis leaf tissue (Pyke et al., 1994). To examine the shoot apical meristem, whole seedlings were embedded intact and sectioned longitudinally. Root tips were dissected from 5-day-old seedlings and fixed and embedded using the same protocols. Measurements of plastid profiles from electron micrographs were made using a Seescan image analysis system (Cambridge, UK) (Pyke and Leech, 1991). Sectioning of young A rabidopsis seedlings embedded in polyethylene glycol (PEG) and staining with 4′,6′-diamidino-2-phenylindole (DAPI) was carried out using methods described previously (Marrison and Leech, 1992).

Epidermal strips for visualising stomatal guard cells were peeled from the lower surface of first leaves between 20 and 25 days after sowing. Chloroplasts in stomatal guard cells were visualised by mounting epidermal strips in 1%(w/v) AgNO3. Sizes of guard cell chloroplasts visualised by AgNO3 staining were determined directly from the microscope by image analysis. Images of guard cell chloroplast autofluorescence in epidermal strips were obtained using a Nikon FXA fluorescence microscope with a filter combination of dichroic mirror 510, excitation filter 450–490 nm and barrier filter 515F.

The shoot apex of the arc6 mutant of Arabidopsis consists of a central meristematic region surrounded by leaf primordia which develop on the flanks of the apical dome (Fig. 1). The morphology of the arc6 shoot apex (Fig. 1) does not differ from that of the wild-type shoot apex, which has been well characterised (Medford et al., 1992; Fujie et al., 1994). The apical dome itself consists of three cell layers L1, L2 and L3 (Fig. 1) and it’s ultrastructure can be clearly observed by electron microscopy (Fig. 2A,B). Transects through the three layers of the wild-type apical meristem cells reveal populations of proplastids recognisable by their dense staining, two envelope membranes and limited internal membranes occasionally connected with the internal envelope membrane or appressed (Fig. 2A,B). The profile shape of wild-type proplastids is very uniform (Fig. 2B,D), typically ellipsoidal with average length/breadth dimensions of 1.2 m μm 0.7 μm and a cross-sectional area which is very uniform throughout the proplastid population (Table 1). Serial sectioning in the wild-type apical meristem confirmed that each proplastid profile in a cell transect represents an individual proplastid. Proplastid profiles in wild-type cell transects average four per cell (s.e.=0.2, n=61), e.g. see Fig. 2D. Proplastids in both the wild-type and arc6 apical meristem are normally evenly distributed throughout the cytoplasm and in wild-type cells often form a distinct ring around the nucleus (Fig. 2D).

In arc6, the extremely large proplastids appear to be wrapped around the nucleus (Fig. 2C). In very marked contrast to the ellipsoidal wild-type proplastids the profiles of arc6 proplastids in the three layers of the apical meristem are highly variable in size and shape (Fig. 2A,C) and the electron micrographs reveal that the arc6 mutation has had a radical effect on proplastid morphology in cells throughout the apical meristem. Transects through arc6 apical meristem cells contain on average only two (s.e.=0.16, n=52) proplastid profiles per cell, and these profiles are highly variable in size with an average twofold increase in area compared to wild type (Table 1). Although the arc6 proplastids are larger and more irregular in shape, their internal structure is similar to wild type and a small amount of internal membrane is present. Because of their enlarged size, a complete three-dimensional reconstruction of an arc6 proplastid requires a series of sequential sections. In Fig. 2E sequential sections are shown through a representative cell from the arc6 apical meristem (Fig. 2C) and clearly show how small individual arc6 proplastid profiles contribute to a single large proplastid with a complex three-dimensional morphology. arc6 proplastids have many surface protuberances producing an amoeba-like structure, similar to but smaller than the mature arc6 chloroplasts in mesophyll cells (Pyke et al., 1994). Because of these differences in size and shape, comparison of numbers of proplastid profiles per cell transect will underestimate the reduction in proplastid number in arc6 cells compared to wild type. We can conclude from the electron micrographs that there is a very significant reduction in proplastid number per cell in the arc6 apical meristem associated with an average doubling in proplastid size and an increase in the complexity of plastid morphology.

In order to ensure that we compared exactly similar regions of the mutant and wild-type meristem, cell samples were always taken from the central region of the first leaf primordia. Dramatic differences in proplastid morphology and number in the apical meristem of wild-type and arc6 Arabidopsis are maintained and enhanced in the young leaf primordia (Fig. 3). Wild-type leaf primordial cells contain between four and eight plastids undergoing differentiation from proplastids to chloroplasts with small amounts of internal thylakoid membrane and a simple ellipsoid profile (Fig. 3B,D). In arc6 leaf primordial cells (Fig. 3A,C) the plastid profiles are significantly enlarged compared to wild type (Table 1), often with a more complex profile shape (Fig. 3C). In arc6 plastids the internal membrane system is well developed, with organised thylakoid membranes orientated parallel with the long axis of the plastid. Chloroplast profiles contain an average of six to ten stacks of three or four appressed membranes. This is in marked contrast to wild-type plastids in which the thylakoid membranes are much less well developed with little obvious organisation (Fig. 3B,D). The increased size and more complex internal membranes in the arc6 leaf primordial plastids suggest that these plastids are at a more advanced stage of development than in wild-type tissue of the same age.

In some wild-type leaf primordial cells, plastids undergoing division can be identified by a central constriction separating two approximately equal sized daughter plastids (Fig. 3F). arc6 plastids must also be capable of division within the meristematic tissues because segregation into both daughter cells at cell division is clearly ensured, since all the cells in the arc6 apical meristems, in leaf primordia and in the mature leaf contain plastids. Occasionally we have observed arc6 proplastid profiles reminiscent of wild-type plastid division with a thin isthmus forming two equal sized daughter plastids (Fig. 3E).

In order to determine whether the effect of the arc6 mutation is restricted to the shoot, we analysed plastids in the root tip of wild-type and of arc6 Arabidopsis seedlings. In cells of the wild-type Arabidopsis root meristem, the plastids are an ellipsoid with few internal membranes (Fig. 4B) and similar in structure to plastids from the wild-type apical meristem. In contrast arc6 root plastids have a more elongated morphology (Fig. 4A), which in general is more pleiomorphic than the ellipsoidal wild-type plastids, but wild-type and arc6 root plastids are similar in size. In contrast, statoliths, the specialised root plastids found in the columella cells are greatly enlarged (Fig. 4C) compared to wild-type plants (Fig. 4D), and contain more starch. These observations show that the arc6 mutation also significantly perturbs root plastid development in addition to shoot and leaf plastid development and that ARC6 functions in cells throughout the Arabidopsis seedling.

The stomatal guard cells in the leaf epidermis provide an ideal tissue in which to follow changes in plastid population size and segregation during development, since all the plastids of the cell can be viewed simultaneously by light microscopy within whole unsectioned guard cells in epidermal strips. The L1 layer of cells in the shoot apex gives rise to the epidermis from which stomata differentiate. Plastids in the arc6 apical L1 cells are abnormal compared to wild type and there are very few of them (Fig. 2A).

If the rate of proplastid division relative to cell division is radically affected by arc6 the number of proplastids in the guard mother cell will be significantly reduced and so will the number in the guard cells. arc6 guard cells consistently contain fewer plastids per guard cell than wild type, so clearly there is a large difference between the plastid segregation pattern in arc6 and wild-type stomatal guard cells (Fig. 5). In both wildtype and arc6 stomata there is a predominance of guard cell pairs with equal plastid numbers, in 20% of wild-type stomata each guard cell has four chloroplasts (Fig. 6A,B) and in 20% of arc6 stomata each guard cell has only one chloroplast. We made the very important observation that a significant proportion of guard cells in arc6 stomata contain no plastids: a characteristic which is never observed in wild-type stomata. Indeed 4% of arc6 stomata lack plastids in both guard cells (Fig. 6G,H). In the arc6 cell population sampled, 21% of stomata show 1/0 plastid segregation in the guard cell pair (Fig. 6E,F) and 4% show a 2/0 plastid segregation. In those arc6 guard cells containing only a single chloroplast, the chloroplast is enlarged (Fig. 6C,E) compared to chloroplasts in wild-type guard cells containing four or more chloroplasts (Fig. 6A,B). Examination of the ultrastructure of guard cell chloroplasts (Fig. 4E,F) confirmed that arc6 guard cell chloroplasts are four to five times larger than wild type but are similar in form and internal structure.

We have clearly demonstrated the importance of the ARC6 gene for normal proplastid development in Arabidopsis. ARC6 is the first gene to be identified in Arabidopsis which has a global effect on plastid development in cells arising from both the shoot and root meristems. Although other genes have been identified which affect chloroplasts in Arabidopsis, i.e. pale cress (Reiter et al., 1994), proplastids in the meristems of this mutant have normal morphology and the mutant gene specifically affects the chloroplast differentiation process. Our results show that the ARC6 gene functions very early in proplastid development prior to chloroplast differentiation, since proplastids in both shoot and root meristems are already grossly affected by the arc6 mutation. In addition our studies have shown that arc6 does not appear to have a significant effect on the ability of plastids in arc6 plants to function normally, since the growth of the arc6 plant is not severely limited by the arc6 mutation (Pyke et al., 1994). We have also established that arc6 causes severe disruption to both proplastid and chloroplast division. It remains to be determined whether perturbation of chloroplast division in arc6 leaves is a direct consequence of the arc6 mutation or a perturbation of proplastid development in the shoot meristem. Many morphological similarities have been shown between the proplastid and chloroplast division processes (Chaly and Possingham, 1981; Whatley, 1988) so it is possible that the ARC6 gene is involved in a process common to both proplastid and chloroplast division. There is considerable potential to use other arc mutants, which specifically affect different stages of chloroplast division (Pyke and Leech, 1992, 1994) to establish how different ARC genes interact in the control of proplastid and chloroplast division.

Since apoplastidy in higher plant cells is extremely rare (Sears, 1980) a control must exist in normal meristematic cells which prevents the formation of apoplastid daughter cells. In arc6 cells, with only two chloroplasts, the mechanism controlling segregation must be especially stringent. Characterisation of the arc6 mesophyll cell phenotype suggests that arc6 chloroplasts in expanding mesophyll cells do not divide (Pyke et al., 1994), but arc6 proplastids in meristem cells are apparently capable of limited division; no cells in arc6 meristems or leaf mesophyll cells lack plastids. The putative division profiles of enlarged arc6 proplastids in leaf primordia cells have been observed, so division of arc6 plastids in dividing cells may be a rare but consistent event. The close association of proplastids with the nucleus in meristematic cells may be important in determining their segregation into daughter cells at division by virtue of their spatial arrangement. This type of plastid distribution has been reported previously in dividing cells of higher plants (Whatley, 1986; Baluska et al., 1993) and could ensure partitioning of some plastids into both daughter cells at division (Birky, 1982). arc6 chloroplasts appear to wrap around the nucleus and this may facilitate the approximately equal partitioning of plastids into the two daughter cells. Compared to the apical meristem, the control of plastid segregation during stomatal formation appears to be less stringent, since guard cells lacking chloroplasts do exist in arc6. Stomatal guard cells lacking plastids have only been reported once previously (Nelson and Mayo, 1975) in an orchid species, and the absence of plastids in the guard cell of the arc6 mutant of Arabidopsis could prove very useful in further examination of the role of guard cell plastids in stomatal function.

The increase in size and internal membrane organisation of arc6 plastids reflects an advanced state of development of arc6 plastids. It is possible that mutation in the ARC6 gene allows the premature differentiation of proplastids into either chloroplasts or statoliths, resulting in prematurely large organelles and a disruption of the normal balance between plastid differentiation and cell development.

In this study we have shown the ARC6 gene of Arabidopsis, identified originally by its abnormal chloroplast number in the arc6 mutant, to be crucial to the early development of plastids in both shoot and root cells. The global role of this gene in Arabidopsis suggests that it is of major importance in the nuclear control of plastid differentiation in higher plants.

We thank Keith Partridge for assistance in growing the plants and the Agriculture and Food Research Council (UK) for a Plant Molecular Biology II grant (LR87/528) to R.M.L.