The Arabidopsis minD mutation causes aberrant FtsZ1 ring placement and moderate heterogeneity of chloroplasts in the leaf epidermis

ABSTRACT

Symmetric division of leaf mesophyll chloroplasts requires MinD and MinE, which work together to suppress division other than at the mid-chloroplast. arc11 is a MinD loss-of-function mutant of Arabidopsis thaliana. In arc11 plants, asymmetric chloroplast division, as well as its delay or arrest, results in extreme size polymorphism of chloroplasts in mature mesophyll cells. The current study examined chloroplast phenotypes in the epidermis of arc11 leaves. Fluorescence microscopy analysis revealed that epidermal chloroplasts in mature leaves exhibited moderate heterogeneity in size. This probably resulted from completion of many of the previous non-equatorial or multiple division events in expanding leaves. Additionally, analyses of plastids found that epidermal chloroplasts in arc11 mutants showed several phenotypes that have not previously been described.

Chloroplast division is a complex process involving the determination of the division site, typically at the equatorial plane only, and the constriction and scission of the chloroplast membranes (inner and outer envelopes).¹ In the early stages of chloroplast division, FtsZ proteins assemble to form a ring on the stromal side of the inner envelope at the division plane.^2,3 Together with FtsZ, MinE is one of the most important proteins for chloroplast division in mesophyll cells of higher plants,^4-6 working cooperatively with MinD,^7-9 MCD1,¹⁰ and ARC3¹¹ to regulate the division (FtsZ ring-formation) site.¹² Previously, we showed that the proper balance of activity between MinD and MinE is crucial for symmetric and binary division of mesophyll chloroplasts; excess MinD inhibits division, while excess MinE induces asymmetric and multiple divisions.¹³ More recently, we observed leaf epidermal chloroplasts (plastids)^14-17 in the minE mutant¹³ (i.e., under conditions of MinD excess) of Arabidopsis thaliana (which we designated as atminE1), using stroma-targeted fluorescent proteins.¹⁸ While the number of chloroplasts per cell in the atminE1 mesophyll was greatly reduced and their size enlarged in a uniform manner,¹³ chloroplasts in the atminE1 leaf epidermis exhibited a wide spectrum of aberrant plastid morphology. We observed (a) the formation of heterogeneous plastid populations, even within a single cell, showing a varied mixture of sizes from giant to mini-plastids; (b) increased formation of stromules (stroma-filled tubules),^19,20 similar to that observed in the epidermal cells of hypocotyls, stamens, and roots of the Arabidopsis arc6 mutant,²¹ another mutant severely defective in mesophyll chloroplast division²²; and (c) other novel plastid morphologies (or subcellular distribution), such as grape-like clusters. These results demonstrated a significant difference between mesophyll and epidermal chloroplasts in the regulation of division. To further elucidate this issue, the present study was designed to clarify whether chloroplast phenotypes in arc11, an Arabidopsis minD mutant^1,23 which produces a decreased level of dysfunctional AtMinD1(A296G) proteins from a wild-type (WT) level of mRNAs^10,23 (i.e., under conditions of MinE excess), were conserved between the mesophyll and the epidermis within a single leaf. We therefore examined the morphology of plastids in the leaf epidermis of arc11 plants, especially with respect to their size distribution, stromule formation, and division-associated (FtsZ ring-based) constrictions, using the same methodology as our previous study of epidermal chloroplasts in atminE1.¹⁸

Leaf mesophyll cells isolated from WT Arabidopsis plants contained a relatively homogenous population of chloroplasts with regard to size and shape (Fig. 1A), consistent with previous observations.²⁴ In arc11 plants, by contrast, a single mesophyll cell contained chloroplasts of heterogeneous sizes, including giant (> 10 µm in diameter), normal, and miniature (< 2 µm in diameter) chloroplasts (Fig. 1B), as shown earlier.¹ We propose that this phenotype results from a combination of 3 processes: asymmetric and/or multiple divisions; uncontrolled, but symmetric, divisions that occurred stochastically; and inhibition of division by chloroplast expansion that is capable of overcoming the force of division constriction.¹³

Figure 1.

Chloroplast populations in mature leaf cells of WT and arc11 Arabidopsis plants. (A,B) Isolated leaf mesophyll cells of WT (A) and arc11 (B) plants. (C–F) Leaf petioles of WT (C,E) and arc11 (D,F) plants expressing stroma-targeted GFP. The epidermis and cortex (C,D) or the epidermis (E,F) of petioles from leaves 3 to 4 from 4-week-old seedlings were observed using CLSM. “Maximal intensity” projection images are shown. Arrow: site of epidermal chloroplasts; single arrowhead: chloroplast constrictions; double arrowhead: stromule; triple arrowhead: stomata. Scale bar: 10 µm.

The chloroplast morphology phenotypes of WT and arc11 plants were readily examined in situ in the cortex of the leaf petiole in lines expressing stroma-targeted green fluorescent protein (GFP; Fig. 1C,D). Chloroplasts in the cortex of WT petioles were homogeneous (Fig. 1C), as described previously.²⁵ The number of chloroplasts per cell was reduced in arc11 plants, and individual plastids were mostly enlarged and occasionally formed division constrictions at random sites (Fig. 1D, single arrowheads), indicative of deregulated positioning of division sites.²³ Unlike canonical chloroplasts in the leaf mesophyll and petiole cortex, those in the petiole epidermis were only slightly pigmented with chlorophyll, and thus considered immature. Chloroplasts in the petiole epidermis of mature WT leaves were generally oval in shape and smaller than those in the mesophyll (Fig. 1E).¹⁸ In addition, dumbbell-shaped chloroplasts, probably in the process of binary division, were often detected, even in mature leaves. The formation of stromules was also observed (Fig. 1E, double arrowhead),¹⁸ possibly reflecting dynamic remodeling of the envelope membranes of epidermal plastids. In contrast to those in cortical (mesophyll) cells, chloroplasts in the epidermis (pavement and guard cells) of arc11 leaves appeared largely similar to those in WT plants in terms of their oval or dumbbell shapes, size, and subcellular distribution (Fig. 1F). We did not detect any giant chloroplasts or grape-like chloroplast clusters in the arc11 epidermis, although both of these were striking features of epidermal plastids in the atminE1 mutant.¹⁸ Despite the absence of the strong phenotypes seen in atminE1, careful examination revealed a moderate level of size polymorphism in the epidermal chloroplasts of arc11 plants (Fig. 1D,F). Furthermore, chloroplasts with 2 constrictions, or with a single constriction at a non-central position, were detected in the arc11 epidermis, indicating disturbances in the positioning of division sites.

To investigate the subtle phenotypes of epidermal chloroplasts in mature arc11 leaves in more detail, we used arc11 plants expressing stroma-targeted cyan fluorescent protein (CFP),²⁶ an effective fluorophore for labeling epidermal chloroplasts with lower background signals from the mesophyll.¹⁸ We observed both CFP-labeled chloroplasts and chlorophyll-derived autofluorescence, and confirmed 3 features of arc11 epidermal chloroplasts not previously reported. Firstly, chloroplasts in arc11 ranged more widely in size than those in the WT, from mini-chloroplasts (< 2 µm in diameter) to slightly enlarged ones, and assumed an oval or vesicle-like shape (Fig. 2A). This size polymorphism might result from asymmetric chloroplast division, as suggested by the presence of dumbbell-shaped chloroplasts with a non-central constriction (Fig. 2A). Secondly, in addition to chloroplasts showing a chlorophyll signal (Fig. 2B), vesicular mini-plastids with no detectable chlorophyll fluorescence signal were observed at low frequency in arc11 plants (Fig. 2C, asterisks). Lastly, stromules formed in a similar manner in both arc11 and WT plants. One to 3 stromules per plastid body were observed. These took various forms including short projections and elongated filaments; the latter often branched or formed blobs (Fig. 2D,E). Although the degree of stromule formation varied in epidermal cells of both genotypes, the most developed stromules in arc11 appeared longer and more stable than those in WT plants. Interestingly, mini-plastids were able to form stromules (Fig. 2E, asterisk) and constriction isthmi, irrespective of the presence or absence of chlorophyll signals. While bacterial “minicells” formed by the minD mutation are nucleoid-free and thus unable to proliferate,²⁷ at least a subset of mini-plastids in the arc11 (minD) epidermis might retain some or most plastid activities, including replication. It is currently unknown how the chlorophyll-deficient mini-plastids form. They may result from an extremely asymmetric division of chloroplasts, from degeneration of mini-chloroplasts due to secondary effects of asymmetric division (because, for example, they have received an insufficient amount of chloroplast DNA), or from fission of stromules, which inherently lack chlorophyll.

Figure 2.

Plastid morphology in mature leaf epidermis of arc11. (A–E) Images of stroma-targeted CFP (black and white) and chlorophyll (magenta) in petioles of the first leaf of 3-week-old seedlings collected using fluorescence microscopy. Asterisks indicate mini-plastids without a detectable level of chlorophyll signal. Scale bar: 10 µm.

The “terminal” phenotypes of epidermal chloroplasts in mature arc11 leaves described above implied that unequal replication of chloroplasts had occurred during earlier stages of development, although in a weaker manner than in the mesophyll (cortex). To reveal the modes of chloroplast replication in developing epidermal cells, we used a dual detection method to observe stroma-targeted CFP²⁶ and FtsZ1-GFP,¹³ as previously applied to the study of epidermal chloroplasts in atminE1.¹⁸ Unexpectedly, given the terminal phenotypes, chloroplasts with multiple division sites were frequently observed (Fig. 3A,B). Those chloroplasts exhibited filamentous (one-dimensional) growth and bore up to 8 division constrictions. A variety of size distribution patterns of plastid bodies within a single, multiple-dividing chloroplast were seen; some comprised similarly sized plastid bodies and others variably sized ones. A single cell could contain chloroplasts with a wide range of sizes and shapes, from mini-chloroplasts to axially elongated ones resembling beads on a string. Those morphological patterns were also observed in the epidermis of arc11 cotyledons. The single CFP image shown in Fig. 3C contains an enlarged chloroplast in the mesophyll (asterisk) and an elongated chloroplast with multiple constrictions in the epidermis, emphasizing the differences in chloroplast phenotypes between arc11 cotyledon epidermis and mesophyll, and also the shared phenotype seen in the epidermis of the cotyledons and true leaves. Next, we used the FtsZ1-GFP reporter to monitor FtsZ1 within chloroplasts in the developing epidermis of arc11 leaves (Fig. 3D–I). FtsZ1 was localized in ring-like structures at the isthmi (Fig. 3E,H,I, arrowheads) and the future division sites in constricting or elongated chloroplasts. Additionally, a faint diffusion signal of FtsZ1-GFP was detected across the stroma, which resembled that observed in atminE1.¹⁸ It was noted that the appearance of ‘budding-like’ chloroplasts, which could also occur in mesophyll or cortex chloroplasts of arc11 leaves,^1,13,23 was observed to possess FtsZ1 rings at their constricting neck (Fig. 3I). These chloroplasts were considered to represent an extreme type of asymmetric division with aberrant FtsZ1 ring placement, rather than temporary protrusion of a ‘chloroplast bud’.

Figure 3.

Plastid morphology and FtsZ1 localization in late-expanding or maturing leaf epidermis of arc11 plants. (A–C) Plastids from lines of arc11 plants expressing TP_FtsZ1–1-CFP. (D–I) Plastids from lines of arc11 plants expressing both TP_FtsZ1–1-CFP and FtsZ1-GFP. Fluorescence microscopy images showing CFP (black and white or cyan), GFP (orange), chlorophyll (magenta), and merged fluorescence produced by petioles from leaves 3 to 8 of 3-week-old (A,B,D–I) or cotyledons from 11-day-old (C) Arabidopsis seedlings. Arrow: mini-plastid; arrowheads: plastid constrictions; asterisk: enlarged chloroplast in the mesophyll. Scale bars: 10 µm (single line) or 5 µm (double line).

This study showed that, in the epidermis of expanding arc11 leaves, FtsZ1 rings assembled at a non-central site or, often non-synchronously, at multiple sites, resulting in unequal or multiple divisions of epidermal chloroplasts. In some cases, a single chloroplast gave rise to 2 or more, similarly sized (sometimes WT-sized), daughter chloroplasts. This led us to conclude that the mode of formation of FtsZ1 rings and constricting isthmi in the epidermal chloroplasts of arc11 resembled that in the mesophyll (cortical) chloroplasts.^13,23 Although we did not detect giant chloroplasts, as observed in arc6²⁸ and atminE1,¹⁸ in the epidermal cells of arc11, such giant chloroplasts were generated in arc11 mesophyll (cortical) cells. We therefore concluded that, following FtsZ1 ring formation, the process of chloroplast replication and morphogenesis in arc11 differed between epidermal and mesophyll cells. Based on our hypothesis¹³ that extreme expansion of chloroplasts inhibits FtsZ-mediated division constriction, we predict, from the relatively underdeveloped status of leaf epidermal chloroplasts, that this inhibitory effect is attenuated in these plastids and occurs only slightly, if at all. Our results from arc11 implied that at least some of the dumbbell-like chloroplasts (those undergoing binary fission with a single constriction) observed in the epidermis of mature leaves were derived from chloroplasts that had harbored multiple constrictions during the leaf expansion stage. This would suggest that the division process was completed at many of the potential division sites, marked by FtsZ1 rings, in epidermal chloroplasts in the expanding leaves of arc11. Our results therefore provide experimental support for the above hypothesis, or are at least consistent with it.

The current model of chloroplast division in higher plants is implicitly postulated to be universal. The present study, however, implies the possibility of a certain level of variation in the mechanisms of chloroplast division. Moreover, these results may offer useful insights into the diversity of division and morphogenesis mechanisms of non-green plastids, which was extensively discussed in recent reviews.^24,29,30 Suggestively in our previous study,³¹ the terminal phenotype of plastids in mature pollen grains of arc11 was essentially identical to the WT phenotype. Moreover the giant chloroplasts in pericarp cells of tomato suffulta were reported to undergo structural remodeling and multiplication during chromoplast differentiation,³² hinting at the existence of, as yet undescribed, diversity of those mechanisms. These lines of research will shed more light on the morphological aspects of plastid differentiation.

Seeds of arc11 (arc11–1 allele; Ler background)¹ were obtained from the Arabidopsis Biological Resource Center (Ohio State University, Columbus, OH, USA). The stable transgenic Arabidopsis lines, ptA5–3 (Col background),³³ FC1–7 (Col background),¹⁸ and FL4–4 (Col background),²⁶ which express stroma-targeted GFP (TP_RBCS3A-GFP), CFP (TP_FtsZ1–1-CFP), and the TP_FtsZ1–1-CFP and matrix-targeted YFP (Pre_mtHSP60-YFP), under the control of the CaMV35S promoter, were used to visualize the stroma in living tissues. Another line, FC1–7 × Z1g11, which simultaneously expresses TP_FtsZ1–1-CFP and FtsZ1-GFP,¹³ was constructed previously,¹⁸ and crossed with arc11 plants to enable the monitoring of plastids or FtsZ1 protein in the mutant background. The F₂ generation and later progenies were used for microscopy, as described below. Plants were germinated and grown according to standard methods.¹⁸ All plants used in this study exhibited healthy growth and reproduction under laboratory conditions.

Whole plant organs were mounted in water under a glass coverslip and observed using confocal laser scanning microscopy (CLSM) and epifluorescence microscopy. As in our previous studies,^{13,18,23,25,26} only the adaxial side of leaf petioles was analyzed. A Zeiss LSM700 system (Carl Zeiss, Jena, Germany) was used for CLSM analyses. Maximal intensity projection images were constructed using the LSM software. An Olympus IX70 microscope equipped with CCD digital cameras (model ORCA-ER, Hamamatsu Photonics, Hamamatsu, Japan; model DP26, Olympus, Tokyo, Japan) was used for epifluorescence microscopy. Fluorescence signals from CFP, GFP, and chlorophyll were obtained using appropriate filter sets, as described previously.¹⁸ In both microscopy manipulations, bright field images were taken with differential interference contrast optics. Digital images were imported into RGB channels of Adobe Photoshop CS6 (Adobe Systems Inc., San Jose, CA, USA) to obtain the final merged images.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgments

The authors thank Arabidopsis Biological Resource Center for plant seeds, Yusuke Kazama, Kei H. Kojo and Hiroki Ishikawa for discussion, and Sumie Ohbu for excellent technical support.

Funding

This work was supported by the Ministry of Education, Culture, Sports, Science, and Technology of Japan under grants 25450136 (to MTF) and 26440152 (to RDI).