81
Journal of Cell Science 113, 81-89 (2000)
Printed in Great Britain © The Company of Biologists Limited 2000
JCS0929
Plastid tubules of higher plants are tissue-specific and developmentally
regulated
Rainer H. Köhler* and Maureen R. Hanson
Department of Molecular Biology and Genetics, Biotechnology Building, Cornell University, Ithaca NY 14853, USA
*Author for correspondence (e-mail: [email protected])
Accepted 28 October; published on WWW 9 December 1999
SUMMARY
Green fluorescent stroma filled tubules (stromules)
emanating from the plastid surface were observed in
transgenic plants containing plastid-localized green
fluorescent protein (GFP). These transgenic tobacco plants
were further investigated by epifluorescence and confocal
laser scanning microscopy (CSLM) to identify
developmental and/or cell type specific differences in the
abundance and appearance of stromules and of plastids.
Stromules are rarely seen on chlorophyll-containing
plastids in cell types such as trichomes, guard cells or
mesophyll cells of leaves. In contrast, they are abundant in
tissues that contain chlorophyll-free plastids, such as petal
and root. The morphology of plastids in roots and petals is
highly dynamic, and plastids are often elongated and
irregular. The shapes, size, and position of plastids vary in
INTRODUCTION
Chloroplasts, because of their role in transforming the energy
of sunlight into chemical bonds for fixation of carbon, have
been the most intensively studied type of plastid. However,
non-chlorophyll containing plastids, though not involved in
carbon fixation, play key roles in other biochemical pathways
in the plant. Colorless non-chlorophyll containing plastids,
often called leucoplasts, are found in a wide variety of tissues
and cell types such as roots, petals or suspension cells (Kirk
and Tilney-Bassett, 1978). Specialized leucoplasts can be
distinguished by the compounds stored as energy reserves.
Starch storing plastids that are common in roots and some
suspension cells are called amyloplasts, protein storing plastids
are proteinoplasts, lipid storing plastids are elaioplasts and
carotinoid-containing plastids, present in some roots, fruits and
flowers, are called chromoplasts. Most plastid types as well as
proplastids are thought to be able to differentiate into each
other, with exceptions such as gerontoplasts, plastids in
senescing leaves, which are believed to be a non-reversible
terminal differentiation (Strasburger et al., 1991).
Non-chlorophyll containing plastids have been more
difficult to observe by light microscopy than chloroplasts,
restricting their characterization in vivo. Current information
about non-green plastid shapes and distribution is mostly based
particular developmental zones of the root. Furthermore,
suspension cells of tobacco exhibit stromules on virtually
every plastid with two major forms of appearance. The
majority of cells show a novel striking ‘octopus- or
millipede-like’ structure with plastid bodies clustered
around the nucleus and with long thin stromules of up to
at least 40 µm length stretching into distant areas of the
cell. The remaining cells have plastid bodies distributed
throughout the cell with short stromules. Photobleaching
experiments indicated that GFP can flow through
stromules and that the technique can be used to distinguish
interconnected plastids from independent plastids.
Key words: Plastid, Stromule, GFP, Photobleaching, Transgenic
plant
on the investigation of fixed material by electron microscopy
(EM). EM provides very high resolution but fixation artifacts
and poor preservation of structures might result in
misinterpretation or failure to detect sensitive structures.
Furthermore, fixing the tissue for EM studies prevents
examination of dynamic processes; these can be seen only by
labelling of organelles and subcellular structures in living cells.
In order to make all plastid types accessible to observations
in vivo we labeled plastids with green fluorescent protein
(GFP). GFP is a novel marker protein derived from the
jellyfish, Aequorea victoria, that emits green fluorescent light
when excited with blue light (Chalfie et al., 1994). Expression
of GFP fused in frame with the recA-transit peptide (CT-GFP,
Cerutti et al., 1992) regulated by a constitutive 35S CaMV
promoter in transgenic plants permits observation of plastids
by confocal laser scanning microscopy or epifluorescence
microscopy in vivo (Köhler, 1998). The CT-GFP fusion protein
as well as CT-GFP fluorescence are specifically associated with
the soluble stroma fraction of purified plastids and are not
detectable in a membrane fraction that includes broken and
intact thylakoid membranes. Green fluorescent tubular
structures can be observed emanating from the chloroplasts in
CT-GFP-expressing transgenic plants (Köhler et al., 1997).
Because the imported GFP is localized to the stroma, according
to fractionation experiments, we named these tubules
82
R. H. Köhler and M. R. Hanson
stromules, for stroma filled tubules. Stromules are highly
dynamic, variable in length and shape, and they can
interconnect different plastids. Photobleaching experiments
have shown that stromules permit the direct exchange of GFP
between separate plastids providing a novel pathway of
communication between plastids that circumvents passage
through the selective membranes or the selective import and
export machinery (Köhler et al., 1997). Stromules can be seen
on plastids of non-transgenic spinach and tobacco plants by
light microscopy (Spencer and Wildman, 1962; Wildman et al.,
1962; Wildman, 1967; Köhler et al., 1997).
While stromules were observed in leaf tissue of tobacco,
petunia and Arabidopsis (Köhler et al., 1997), most leaf cells
lack detectable stromules, making investigations of the nature
and function of this novel plastid structure cumbersome. By
fluorescence microscopy we examined the appearance of
plastids and abundance of stromules in a number of different
tissue types in transgenic tobacco plants expressing GFP
localized to plastids. Our results indicate that plastid
appearance and stromule number is highly dynamic and
variable between different tissues and cells, with the most
extensive and numerous stromules present in cells containing
non-green plastids.
MATERIALS AND METHODS
Gene construct
The modified GFP4 coding region (Haseloff et al., 1997), which was
further altered by a serine to threonine mutation at codon 65
(S65TmGFP4), was fused in frame with the Arabidopsis recA transit
peptide (CT-GFP; Cerruti et al., 1992). This construct was cloned into
a plant expression vector containing the double 35S CaMV promoter
(Datla et al., 1993; Tobias and Nasrallah, 1996) and the transformation
of tobacco Petit Havana was done as described by Horsch et al.
(1985).
Growth of plant material
Leaf and root tissue were obtained from transgenic tobacco plants
(Nicotiana tabacum c.v. Petit Havana) grown in the green house in
soil or under sterile conditions in Magenta boxes on N-13 medium
(Hosticka and Hanson, 1984) at 25°C under a 16 hour light, 8 hour
dark cycle. There was no difference in plastid appearance between
plants grown in soil or in vitro. Callus growth was induced by
incubation of sterile leaf material from plants kept in Magenta boxes
on NT1 agar medium (MS basal salt mixture (Sigma M 5524, St
Louis, MO), 100 mg/l myo-inositol, 1 mg/l thiamine, 0.2 mg/l 2,4-D,
180 mg/l KH2PO4, 30% sucrose, and 9 g/l Phytagar at a pH of 5.8).
Suspension cells were grown from callus in liquid NT1 medium and
transferred every week to fresh NT1 medium. Cultures were grown at
110 rpm under a 16 hour light and 8 hour dark cycle at 25°C.
Imaging of plant material
Material from plants grown in the greenhouse or in Magenta boxes
was mounted on slides in 20 mM MES pH 5.5, 1 mM CaCl2.
Suspension cells and callus cells were mounted in NT1 medium (s.a.).
Suspension cells were imaged 3 to 4 days after transfer to fresh NT1
medium while they are still in logarithmic growth phase. Imaging was
performed using a Bio-Rad MRC-600/Zeiss Axiovert 10 CLSM
and the COMOS program (Bio-Rad, Hercules, CA). Green GFP
fluorescence and red chlorophyll autofluorescence were collected
simultaneously with a standard K1/K2 filter set (488/568 dual band
pass, Dichroic 560 DCLP, 522DF35 and 585 EFLP emission filter).
The separate images were imported into the red and green RGB
channels of Adobe Photoshop and merged into the final image. Black
and white CLSM images show the green channel only because red
fluorescence was absent. Simultaneous recording of GFP fluorescence
and light microscopic images was done with a BHS filter cube
(488DF10, Dichroic 510, emission 515LP) and fiber optics. Image
stacks taken along the optical z-axis in steps indicated in the figures
were opened in Confocal Assistant 4.02 (Todd Clark Brelje, Bio-Rad,
Hercules, CA) and projected into a single plane using maximum pixel
value. For light micrographs, single images were selected from the
stacks that provided the best representation of the general appearance
of the cell including structures such as the nucleus, cytosolic strands
and, if visible, plastid tubules.
Photobleaching experiments
The areas indicated as boxes in Fig. 5A and 7A were photobleached
by repeated scanning at full power with a 15 mW krypton/argon mixed
gas laser using the K1/K2 filter set. The laser was restricted to this
area by using the zoom function of the COMOS program. A Zeiss
Plan Apochromat ×63 oil immersion objective with NA 1.4 was used
for photobleaching. Images before and after photobleaching were
collected with only 10% laser power to reduce photobleaching and
possible damage to the cells.
RESULTS
Abundance and appearance of plastids and
stromules in different plant tissues
Expression of CT-GFP targeted to the stroma of plastids does
not result in any visible phenotype in the transgenic plants.
Transgenic plants and suspension cultures grow normally, the
plants have normal seed set, and the transgenes segregate as
expected. In order to identify variations in the abundance of
stromules and in the appearance of plastids in different tissues,
we investigated tobacco plants by epifluorescence microscopy
initially (data not shown) and then used confocal laser scanning
microscopy for imaging. The frequency of stromules in
chlorophyll-containing cells is very low (Fig. 1A,B,C). Most
chlorophyll-containing cells of a given cell type do not exhibit
stromules. Nevertheless, stromules are observable in all of
the cell types examined, such as guard cells, trichomes,
parenchyma cells, mesophyll cells and epidermal cells. Even
in cells that have chloroplasts with visible stromules extending
from their surfaces, other chloroplasts in the same cell lack
stromules (Fig. 1A,B,C). Often a single plastid with stromules
is visible in a cell surrounded by other chloroplasts in the same
cell and surrounded by other cells that do not exhibit detectable
stromules. Due to the very low background fluorescence, bright
green fluorescent stromules can be found against a virtually
black background even when present on a single chloroplast in
one out of hundreds of cells.
Mature chloroplasts in vascular plants are generally lensshaped structures of 5-7 µm length and 1-3 µm width and are
bi-convex, plano-convex or concavo-convex in profile (Hoober,
1984; Thomson, 1974). Red chlorophyll autofluorescence that
outlines the distribution of thylakoid membranes and stacks in
chloroplasts appears as a regular shaped round or oval structure
in chloroplasts in transgenic tobacco plants expressing GFP
(Fig. 1A,B,C). If single stroma thylakoids enter stromules, they
are not sufficiently autofluorescent to be detected. Imaging
stroma-targeted GFP reveals that chloroplasts have a more
irregular and dynamic shape than evident from chlorophyll
autofluorescence imaging (Fig. 1A,B,C; Köhler et al., 1997).
Plastid and stromule morphology
83
Fig. 1. Tissue-specific variations in
stromule abundance and plastid
appearance. CLSM images of different
tobacco cell types and tissues expressing
plastid targeted CT-GFP (Köhler et al.,
1997). The z-axis distance between images
of a stack and the number of images
projected into a single plane is shown in
brackets. Bars, 10 µm. Thin arrows
indicate stromules and arrowheads indicate
plastid bodies. Images (Ag), (Bg), and
(Cg) show GFP fluorescence (green
channel), images (Ar), (Br), and (Cr) red
chlorophyll autofluorescence (red
channel), and images (Am), (Bm), and
(Cm) merged images of the green and red
channels. (A) Leaf mesophyll cells (0.2
µm/28 images); (B) A trichome cell of leaf
(0.2 µm/46 images); (C) Guard cells of
leaf (0.2 µm/36 images); (D) Epidermal
and guard cells in the limb of a petal 5 cm
in size (1 µm/13 images), red fluorescence
has been enhanced relative to GFP
fluorescence to make the autofluorescence
visible; (E) cell of the root hair zone (0.5
µm/13 images).
Green fluorescent tubules, lobes,
knobs, and loops that enclose nonfluorescent areas or particles are
visible. Time-lapse studies show that in
some cells these structures are
amoeboid and highly dynamic, moving, changing their shape
constantly and producing tubules that extend from the plastid
surface, move around or retract (Köhler et al., 1997; see
http://www.mbg.cornell.edu/kohler/kohler.html for movies).
Cells from non-chloroplast containing tissues such as petals
and roots exhibit mostly irregularly-shaped plastids with wide
variation in shape and length and differences in the abundance
and length of stromules (Fig. 1D,E). Because of this
remarkable variation in shape, it is more difficult to distinguish
between the plastid body and stromule in non-green cells than
in chloroplast-containing cells. The name stromule is used for
structures that appear less than 0.8 µm in diameter; larger parts
of the plastid will be referred to as plastid body (Figs 1D,E,
3B,C, 4G). Diameters are estimated from measurements of the
fluorescent area of micrographs taken by CLSM, which usually
results in an overestimation of the actual size. Both stromules
and plastid bodies show green fluorescence as a result of GFP
localized in the stroma. Sometimes stromules can be seen that
appear not be connected to a plastid body at all (data not
shown).
The abundance and appearance of plastids and stromules
were examined in floral tissue. Tobacco flower development
has been divided into 20 stages on the basis of size and
morphological criteria (Drews et al., 1992). Mature stage 12
open tobacco flowers are about 5 cm in size and have a corolla
made of five fused petals. Unusual morphology of GFP-labeled
plastids was visible in the white stalk of the corolla as well as
in the pink limbs. Epidermal cells of stage 12 corolla contain
colorless plastids or leucoplasts and show faint red
autofluorescence in the vacuole in the limb cells, which might
be a result of anthocyanin accumulation (Fig. 1D). Similar to
plastids in root cells (see below), chlorophyll-free leucoplasts
in corolla cells, such as hair cells and epidermal cells, are
irregular and variable in shape and size (Fig. 1D). In contrast,
chlorophyll-containing plastids in guard cells of the corolla are
oval or round and regular in shape. No major difference in the
appearance of leucoplasts within the white or pink corolla parts
at different developmental stages were detected (data not
shown). The plastid bodies in the corolla, which are up to 4-5
µm in diameter, are smaller than chloroplasts in most leaf cells
but larger than plastid bodies of root cells. Tubular plastids are
more frequent in the corolla than in roots and seem to be
present in every cell. Corolla plastids have non-fluorescent
inclusions with diameters of up to at least 2.2 µm.
Developmental changes in plastid appearance and
stromules of roots
The root was selected for more detailed examination of
changes in plastid and stromule appearance because root
development is well defined and cell lineages can be easily
identified (Dolan et al., 1993). For descriptive purposes, the
root can be divided into different overlapping zones, the
meristematic zone, the elongation zone, the differentiation
zone, and the mature root (Benfey and Schiefelbein, 1994;
Dolan et al., 1993). Root plastids differ in their appearance,
length of stromules and intra-cellular distribution depending on
the stage of development, which is defined by their position
within the root (Fig. 2). Most plastids in root cells are smaller
than chloroplasts with a plastid body of about 1.8 to 3 µm in
diameter and stromules of less than 0.6 µm in diameter.
Plastids and stromules in the small cells of the meristematic
zone are densely packed and difficult to separate even by
84
R. H. Köhler and M. R. Hanson
CLSM (Fig. 2A and B). In
most
cells
of
the
meristematic zone, plastids
appear organized in one big
circle around a nonfluorescent area that might be
the nucleus, with stromules
stretching outward (Fig. 2B).
In the elongation zone,
stromules are longer and
plastids become less densely
packed but are still often
around the nucleus (Fig. 2C).
In the differentiation zone,
plastids show elongated
stromules, plastid density
drops and plastids are
occasionally
clustered
around the nucleus (Fig. 2D).
In mature root cells, with
increasing distance from the
root tip, stromules appear
shorter again and plastids
become further dispersed
within the cell with frequent
clustering around the nucleus
(Fig. 2E). Stromules are
visible in most root cells.
Root plastids contain small
non-fluorescent inclusions of
less than 1 µm in diameter
that might be starch grains or
plastoglobuli (Fig. 1E).
Stromules in callus and
suspension cells
The higher abundance of
stromules and the difference
in appearance of plastids in chlorophyll-free cell types, such as
petals and roots, suggests that these differences might be
related to plastid functions other than photosynthesis. To
investigate if these differences can be regulated by a change of
growth conditions that facilitates loss of chlorophyll and
photosynthetic capacity, we initiated callus and suspension
cultures. Chlorophyll content and the transition from
chloroplast to leucoplast or amyloplast is regulated by sucrose
concentration and phytohormone composition of the growth
medium (Kaul and Sabharwal, 1971; Sakai et al., 1996). Callus
growth was induced by placing leaves of transgenic and control
plants on NT1 medium. The phytohormone composition of the
NT1 medium in combination with high sucrose levels results
in a change from autotrophic to heterotrophic growth. Red
autofluorescent chloroplasts are transformed to leucoplasts or
amyloplasts that lack red autofluorescence. Initially, leaf cell
chloroplasts exhibit only a few stromules and chloroplasts are
oval, about 7 µm wide and 1-2 µm deep (data not shown). After
about four weeks on NTI medium, cells vary in their
appearance. Plastids from non-dividing areas of the leaf show
reduced chlorophyll content and some short stromules but lack
the fluidity of shape seen in non-green cells. (Fig. 3A). Nondividing cells contain small round green florescent particles,
Fig. 2. Changes in plastid appearance during root
development. CLSM images of tobacco root cells.
The z-axis distance between images of each stack is
0.5 µm. Root hairs (h), nuclei (n) and stromules
(arrows) are indicated. Bars, 10 µm. The number of
images projected into a single plane is indicated in
brackets. (A) Overview of the root tip zone (42
images); (B) Meristematic zone (6 images); (C)
Elongation zone (17 images); (D) Differentiation
zone (27 images); (E) Mature root near the root tip
(29 images).
likely to be proplastids, which are about 1 µm in diameter and
do not show red chlorophyll autofluorescence. In contrast,
plastids in dividing callus cells are dynamic in shape and
slowly lose all their chlorophyll, observable by the loss of red
autofluorescence (Fig. 3B and C). The majority of these
plastids consist of an enlarged portion, the plastid body, and a
tubular component extending from the enlarged portion. Both
parts are highly variable and dynamic. In some cells
chlorophyll autofluorescence has vanished completely and
most plastids have an elongated shape with enlargements in
one or more places. Non-fluorescent inclusions that show wide
variability between cells might be starch grains (amyloplasts,
Fig. 3B). Often several plastid bodies are connected to form
small branched networks of plastid bodies and stromules (Fig.
3C). It is not evident whether these networks originate from a
single plastid or are fusions of originally separate plastids.
Placing callus cells into liquid NT1 suspension culture
promotes further striking changes in plastid structure and
intracellular distribution. A minor fraction of the plastids in
suspension cells exhibit structures comparable to that of callus
cells. They have elongated plastids with short stromules that
are distributed throughout the cell and sometimes contain nonfluorescent inclusions, likely starch grains (Fig. 4C,D,E,F). In
Plastid and stromule morphology
85
contrast, the majority of cells shows striking ‘octopus-’ or
‘millipede-like’ plastid structures (Fig. 4A,B,G). Most of the
plastid bodies in these cells are clustered around the nucleus
and often seem to be virtually absent from other parts of the
cell. From some plastid bodies, long stromules of up to at least
40 µm reach into distant areas of the cell (Fig. 4G). Stromules
in suspension cells are variable in shape, can be branched or
even encircle larger areas (Fig. 6A). They are highly dynamic,
producing new branches, frequently retracting branches or
moving about (data not shown; movies available at
http://www.mbg.cornell.edu/kohler/kohler_JCS.html).
extend from the plastid bodies into distant areas of suspension
cells. Stromules that are close by but are obviously not
connected do not change in fluorescence (Fig. 7 stromule b).
Some stromules appear to be connected, but there is no flow
of GFP between bleached and unbleached regions (Fig. 7,
stromule c). It is likely that these stromules, though in close
proximity, are actually not fused to one another, preventing
exchange of their contents.
Movement of GFP through stromules
Mitochondria of yeast cells are present in the form of a single
extended network during certain stages of development
(Hermann and Shaw, 1998). Although a single network of
plastids has not been reported in vascular plants, it is known
that plant organelles at some stages of development fuse to
larger units. For example, mitochondria of tobacco protoplasts
can fuse to long vermiform and branched structures (Stickens
and Verbelen, 1996; R. H. Köhler, unpublished observations)
and plastids of vascular plants and Acetabularia can be joined
by tubular connections (Wildman et al., 1962; Menzel, 1994;
Köhler et al., 1997). In order to investigate whether the plastid
bodies that are concentrated around the nucleus in suspension
cells are connected with each other to form a single extended
network, photobleaching experiments were performed (Fig. 5).
Plastids close to the nucleus and parts of the stromules that
originate from them were photobleached by repeated scanning
with the laser (Fig. 5A). Images taken after photobleaching
over a period of 12 minutes do not show any major flow of
fluorescent GFP from the unbleached plastids into the bleached
plastids (Fig. 5B and C) indicating that the plastids around the
nucleus do not form a single communicating network. The
sensitivity and resolution of this experiment is not sufficient to
exclude that some of the plastids might be connected and could
exchange GFP through stromules. Exposure to the intense laser
light during photobleaching does not damage the cells.
Bleached cells show unaltered movement of stromules within
the bleached and unbleached area as observable by comparing
Fig. 5A and B. The movements continue for at least several
hours after bleaching (data not shown).
The independence of most plastids is also supported by the
observations of single images taken along the optical z-axis.
Single CLSM images that were collected by taking images
about 0.5 µm apart from each other along the optical z-axis
indicate that most plastids around the nucleus are separate units
and are not connected through fluorescent tubules (Fig. 6).
Plastids in root cells are sometimes connected by stromules,
and previously we demonstrated that GFP can move through
the root stromules from one plastid body to another (Köhler et
al., 1997). To investigate whether GFP can move freely through
the extremely long stromules extending from plastids near the
nucleus into distant areas of suspension cells, additional
photobleaching experiments were performed (Fig. 7).
Photobleaching of a part of a stromule by scanning a box
shaped area repeatedly with the laser led to the loss of GFP
fluorescence in the exposed area and in areas of the connected
stromule that are not exposed to the laser (Fig. 7, stromule a).
Thus, there is an exchange and movement of unbleached and
bleached GFP molecules through the tubular structures that
Long protuberances extending from vascular plant chloroplasts
in living cells have been reported in the literature a number of
times (e.g. Heitz, 1936; Wildman et al., 1962; Köhler et al.,
1997). The most impressive studies prior to GFP technology
were movies of phase-contrast images of hand-sectioned live
leaf cells performed by Wildman and his colleagues (Spencer
and Wildman, 1962; Wildman et al., 1962; Wildman, 1967).
Unfortunately, these studies have largely been forgotten or
ignored. There are several reports of tubular membrane bridges
connecting chloroplasts in Acetabularia (Menzel, 1994) and,
during a particular state of the cell cycle, in Euglena (Ehara et
al., 1990), but these interconnections have been largely been
considered to be unusual traits of these unicellular organisms
rather than general features of chloroplasts. When stromules
were observed occasionally in vascular plants, they were
usually dismissed as artifacts or abnormalities resulting from
viral infection, chloroplast breakdown, or cell death.
Now that transgenic plants can be produced with GFP
localized in plastids, many plastid types can be observed in
vivo and stromules can readily be detected. It is unlikely that
plastid localization of GFP causes any significant disturbance
of plastid morphology. The transgenic plants grow normally
and our images of plastids in transgenic plants are consistent
with observations made by us and others on plastids of nontransgenic plants. Observations of wild-type plastids in situ
were previously limited to naturally colored plastids such as
chloroplasts or chromoplasts, or required electron microscopic
examination of fixed tissue. Green leaf tissue, widely used for
chloroplast studies, exhibits a very low frequency of stromules,
perhaps accounting for the rarity of their detection by light
microscopy of wild-type plants. Furthermore, stromules lack
detectable chlorophyll, making visualization by chlorophyll
autofluorescence impossible. With plastid-localized GFP, the
only limitation on tissue type that can be examined is the level
of GFP expression that can be obtained with the gene
regulatory sequences used to express the chimeric gfp gene and
the depth of penetration possible by the microscope. The
double 35S CaMV promoter (Datla et al., 1993), a highly active
promoter expressed in many, but not all tissues, was utilized in
our studies. GFP expression was too low to observe plastids in
sporogenous tissue, the tapetum, pollen or the inner cell layers
of the stem, but plastids in most other cell types and tissues
could be observed. Because stromules are often very thin, a
higher level of GFP expression is needed to observe them than
to observe GFP-labelled plastid bodies. Thus it is possible that
some stromules may have gone undetected even in our
transgenic plants.
Because of the highly dynamic nature of the plastid envelope
and protuberances, light and fluorescence microscopy of living
DISCUSSION
86
R. H. Köhler and M. R. Hanson
Fig. 3. Regulation of stromule abundance and of plastid morphology
by growth conditions. Callus cells were imaged by CLSM four weeks
after transgenic tobacco leaf explants were placed on NTI agarose
medium under 16 hour light/ 8 hour dark cycle for induction of callus
growth. Yellow and orange color result from a combination of red
chlorophyll autofluorescence and green GFP fluorescence; change
from yellow or orange to green indicates decreasing chlorophyll.
Bars, 10 µm. Images were taken at 1 µm intervals along the optical zaxis. The number of images projected into a single plane is indicated
in brackets. (A) Non-dividing cell from the leaf area (25 images);
(B) Cell from pale green callus (14 images); (C) Cell from white
chlorophyll-less callus (12 images). Large arrowheads, plastid bodies;
small arrowheads, proplastids; arrows, stromules.
wild-type and transgenic cells has previously demonstrated
their existence most convincingly. However, some electron
microscopic studies of plastids have captured images of plastid
protuberances, often reported in papers describing plants
stressed in one way or another (e.g. Shalla, 1962; Newcomb,
1967; Nichols et al., 1967), as the accepted paradigm for
normal plants did not include the existence of tubular plastid
extensions. Though some of the previously reports of
extensions may be artifact, a number of the images appear to
be high quality and are consistent with our observations at the
light microscopy level. Furthermore, chloroplast protuberance
in rice leaves were observed when Bourett et al. (1999)
preserved cells by high-pressure freezing and freeze
substitution. Two factors have probably contributed to the
paucity of electron micrographs of stromules. First, most
investigators have utilized leaf tissue, in which these structures
are not common. Second, stromules are not well preserved by
standard fixation methods for electron microscopy (S. Wilson
and M. R. Hanson, unpublished data).
Fig. 4. Pleomorphic plastids in
transgenic tobacco suspension
cells. (A,C,E) Single light
microscopic images of suspension
cells expressing CT-GFP; (B,D,F)
CLSM images of the same
suspension cells. (B) Projection of
83 images collected at 0.5 µm
intervals. (D) Projection of 25
images imaged at 1 µm intervals.
(F) Single image. (G) ‘Millipedeor octopus-like’ plastid structure in
two tobacco suspension cells; one
cell shows a side view of the
nucleus, the other cell is a view
from above; projection of 64
images taken at 0.5 µm intervals.
Stromules are indicated by
arrowheads and plastid bodies by
arrows. Bars, 10 µm. A movie of a
complete z-stack and of a rotating
3-D reconstruction of a cell with
‘millipede- or octopus-like’
plastids can be viewed at
http://www.mbg.cornell.edu/kohler/
kohler_JCS.html.
Plastid and stromule morphology
87
Fig. 5. GFP photobleaching to assess degree of autonomy
of plastids clustered around the nucleus. CLSM images
of plastids in a tobacco suspension cell expressing CTGFP. (A and B) Projections of 36 images taken at 1 µm
intervals along the optical z-axis. (A) Pre-bleach image;
(B) Post-bleach image collected after the time-lapse;
(C) Single images of a time-lapse series collected after
photobleaching of the area indicated by a box in A. The
area was photobleached with 100 scans at full laser
power. The first image was taken 80 seconds after the
beginning of photobleaching. The time-lapse images in C
were taken at 20 second intervals over a period of 12
minutes. Shown are selected images with the time after
the first image indicated in seconds.
Our observations of remarkable variations in
shape and sizes of plastids in non-green transgenic
plant tissues and cell cultures are consistent with
reports of pleomorphic plastids in wild-type nongreen cells, observed both by light and electron
microscopy. By phase-contrast microscopy, Suzuki
et al. (1992) were able to observe ‘elongated and
string-like’ plastids similar to those shown in Fig. 4.
Nishibayashi and Kuoriwa (1982) found that living
onion epidermal cells contain leucoplasts with
irregular and dynamic shapes, describing them as
‘oval, dumbbell, beads, rod frypan.’ Steffen (1964)
reported the presence of filamentous and ‘ameboid’
chromoplasts in lily petals, and Whatley and
Whatley (1987) described pleomorphic chromoplasts in petals
of a number of species. In elegant studies of plastid
development in bean roots, consistent with our observations in
tobacco roots, Whatley (1983) documented a progression of
shape changes from spherical to ameboid non-green plastids to
the classic disc morphology of the mature chloroplast. Threedimensional reconstruction of electron microscopic sections of
pine secretory cell leucoplasts by Charon et al. (1987) showed
they had ‘complex shape
with many arms.’ In
contrast to these reports
of pleiomorphic plastids,
leucoplasts and chloroplasts
in petals of Arabidopsis
were found to exhibit a
Fig. 6. Individual CLSM
images of GFP-labeled plastids
clustered around a nucleus in a
suspension cell compared to the
composite projection. (A)
Projection of 31 images taken
along the optical z-axis at 0.5
µm intervals; (B,C,D) Single
images number 8, 9, and 10 of
the stack in A. Small arrows
indicate plastids that seem not
to be connected to other
plastids in the particular plane
of focus, the large arrow points
to a non-fluorescent inclusion
in plastids, and the arrowhead
indicates an area enclosed by a
branched stromule. Bar, 10 µm.
rather uniform regular morphology throughout petal
development (Pyke and Page, 1998). It is evident that GFP
labelling of plastids will allow further studies of developmental
changes in shape, number, and position in non-green cells that
have previously been possible only in certain amenable cell
types in particular species.
We do not know how much communication occurs between
plastids within the same cell. Our studies have conclusively
88
R. H. Köhler and M. R. Hanson
Fig. 7. GFP photobleaching to assess flow of GFP through individual
stromules. A selected area indicated by a white box in image A was
bleached by scanning 30 times at full laser power. A portion of
stromule (a) exhibits loss of fluorescence though outside the zone of
bleaching, indicating flow of GFP between the bleached and
unbleached region. Stromule (b) does not lose fluorescence in any
part. Stromule (c) does not lose fluorescence in the outer part,
indicating it is not connected to a stromule in the bleached zone. Bar,
10 µm. (A) Pre-bleach image; (B) Cell after bleaching.
shown that one plastid can be connected by a tubular extension
to one or a few other plastid bodies within the same cell at a
single moment in time (e.g. Fig. 3C and Fig. 6A). We have
shown that GFP can flow through a stromule that connects two
root plastids (Köhler et al., 1997) and that GFP can flow through
stromules of suspension cells (Fig. 7) but we do not know about
flow of other proteins, small molecules, or RNA. We do not
know how often these connections break and reconnect to other
plastid bodies. In projections of CLSM images of suspension
cells, plastids appear to form one big interconnected unit
because of the clustering of plastid bodies around the nucleus
and occasional branching of plastid stromules. However, a flow
of GFP from plastids localized in one half of the cell into the
other half could not be observed after photobleaching indicating
that the majority of plastids was not interconnected during the
time of the experiments. Until longer time-lapse studies are
performed, we will not know whether there is a point during
development or the cell cycle when all plastids of a particular
cell type are connected simultaneously. In Euglena, connections
between plastids were cell cycle stage-specific (Ehara et al.,
1990). In Acetabularia, tubular connections are found on
virtually all chloroplasts in non-synchronized cultures, and only
very few chloroplasts are not connected (Menzel, 1994).
Though plastid tubules in tobacco cell culture cells often
appeared to be continuous, some of our photobleaching
experiments revealed that GFP does not always flow from an
unbleached tubule into a bleached tubule that appeared to be
connected, indicating that the plastid tubules were independent
rather than interconnected. Therefore, caution must be used
in inferring interconnectedness from images of fluorescent
plastids and stromules without confirmation by photobleaching.
Conversely, sometimes plastids apparently not connected may
actually be bridged by a stromule that lacks sufficient GFP to
be detectable.
The genetic independence of plastids has previously been
shown by the low frequency of recombination between different
plastid genomes introduced into the same cell by protoplast
fusion. Most somatic hybrid plants contain one or the other
plastid genome, evidently as a result of segregation of different
plastid types during regeneration (Hanson et al., 1985). By
imposing selection for drug resistances carried by each of the
parental plastid genomes, Medgyesy et al. (1985) were able to
obtain somatic hybrids carrying both resistance genes, resulting
from recombination of plastid genomes originally separate in
different organelles. However, the frequency of recombinant
plastids was extremely low. This suggests that the entire plastid
genome is rarely, if at all, transmitted through stromules. Plastid
DNA in vascular plants has been reported to be membranebound, but nucleoid organization and attachment is still not well
characterized (reviewed by Gillham, 1994). Perhaps chloroplast
DNA, organized in nucleoids, is restricted from movement
through the narrow stromules that sometimes connect different
vascular plant plastids. In contrast, Osafune et al. (1993)
reported immunolocalization experiments indicating that
chloroplast DNA is present in the cell-cycle-specific bridges
that form between Euglena gracilis chloroplasts.
While the vascular plant plastid genome may normally be
immobilized, unable to venture through stromules, it is possible
that smaller DNA molecules might be able to be transmitted
from one to another plastid. Following bombardment with
plasmids containing an antibiotic resistance gene controlled by
plastid gene regulatory sequences, tissue can be grown on
selective medium, and cells can be obtained that contain many
plastids carrying the resistance gene, though presumably only
one chloroplast originally incorporated the bombarded plasmid
DNA (Svab and Maliga, 1993). In addition to multiplication of
the resistant chloroplast perhaps the transforming plasmids,
much smaller than the chloroplast genome, can move through
stromules to plastids within the same cell. Movement of
plasmids could explain in part the recent observations of
Knoblauch et al. (1999), who were able to inject a single
chloroplast with a GFP gene controlled by plastid gene
regulatory sequences. Green fluorescence first appeared in the
injected chloroplast, but then became evident in other plastids
within the same cell. Perhaps plasmid, RNA or protein – or
possibly all three types of molecules – was transmitted from
one chloroplast to another.
The role of stromules in the functioning of the plastid remains
unknown. In Acetabularia, a role of the long plastid tubules in
chloroplast motility has been proposed (Menzel, 1994).
Alternatively, extensions of the envelope membrane could serve
to facilitate transport of substances in and out of the plastid by
increasing surface area. The extensive network of plastids and
stromules in cultured cells could allow plastid-synthesized
molecules to be delivered to the entire cell or could assist uptake
of cytoplasmic molecules by the plastid from diverse regions of
the cell. GFP-labeled stromules frequently surround nonfluorescent areas or particles that might be mitochondria or
peroxisomes (data not shown; Köhler et al., 1997). Amoeboid
plastids in Phaseolus vulgaris root tips frequently encircle
cytosol with ER, ribosomes and mitochondria (Newcomb,
1967). Perhaps close association of stromules with other
organelles or subcellular structures enhances the transfer of
molecules from one organelle to another. If so, the frequent
occurrence of plastids and stromules in close proximity to
nuclei could be an indication of one or two-way exchange
between them. In future studies. labelling plastids with a
Plastid and stromule morphology
particular fluorescent protein in conjunction with labelling other
parts of the cell with other fluorescent proteins or dyes should
allow clarification of any such interactions.
This work was supported by grants to M.R.H. from the Growth and
Development Program of the United States Department of Agriculture
National Research Initiative (96-35304-3865) and the Energy
Biosciences Program of the Department of Energy (DE FG0289ER14030).
REFERENCES
Benfey, P. N. and Schiefelbein, J. W. (1994). Getting to the root of plant
development: The genetics of Arabidopsis root formation. Trends Genet. 10,
84-88.
Bourett, T. M., Czymmek, K. J. and Howard, R. J. (1999). Ultrastructure of
chloroplast protuberances in rice leaves preserved by high-pressure freezing.
Planta 208, 472-479.
Cerutti, H., Osman, M., Grandoni, P. and Jagendorf, A. T. (1992). A
homolog of Escherichia coli Reca protein in plastids of higher plants. Proc.
Nat. Acad. Sci. USA 89, 8068-8072.
Chalfie, M., Tu, Y., Euskirchen, G., Ward, W. W. and Prasher, D. C. (1994).
Green fluorescent protein as a marker for gene expression. Science 263, 802805.
Charon, J., Launay, J. and Carde, J. P. (1987). Spatial organization and
volume density of leucoplasts in pine secretory cells. Protoplasma 138, 4553.
Datla, R. S. S., Bekkaoui, F., Hammerlindl, J. K., Pilate, G., Dunstan, D. I.
and Crosby, W. L. (1993). Improved high-level constitutive foreign gene
expression in plants using an AMV RNA4 untranslated leader sequence.
Plant Sci. 94, 139-149.
Dolan, L., Janmaat, K., Willemsen, V., Linstead, P., Poethig, S., Roberts,
K. and Scheres, B. (1993). Cellular organisation of the Arabidopsis thaliana
root. Development 119, 71-84.
Drews, G. N., Beals, T. P., Bui, A. Q. and Goldberg, R., B. (1992). Regional
and cell-specific gene expression patterns during petal development. Plant
Cell 4, 1383-1404.
Ehara, T., Osafune, T. and Hase, E. (1990). Interactions between the nucleus
and cytoplasmic organelles during the cell cycle of Euglena gracilis in
synchronized cultures: IV. An aggregate form of chloroplasts in association
with the nucleus appearing prior to chloroplast division. Exp. Cell Res. 190,
104-112.
Gillham, N. W. (1994). Organelle genes and genomes. New York, Oxford:
Oxford University Press, Inc.
Hanson, R. M., Rothenberg, M., Boeshore, M. L. and Nivison, H. T. (1985).
Organelle segregation and recombination following protoplast fusion:
analysis of sterile cytoplasms. In Biotechnology in Plant Science: Relevance
to Agriculture in the Eighties (ed. M. Zaitlin, P. Day and A. Hollaender), pp.
129-144. Orlando: Academic Press.
Haseloff, J., Siemering, K. R., Prasher, D. and Hodge, S. (1997). Removal
of a cryptic intron and subcellular localization of green fluorescent protein
are required to mark transgenic Arabidopsis plants brightly. Proc. Nat. Acad.
Sci. USA 94, 2122-2127.
Heitz, E. (1936). Untersuchungen über den Bau der Plastiden I. Die gerichteten
Chlorophyllscheiben der Chloroplasten. Planta 26, 134-163.
Hermann, G. J. and Shaw, J. M. (1998). Mitochondrial dynamics in yeast.
Annu. Rev. Cell Dev. Biol. 14, 265-303.
Hoober, K. J. (1984). Chloroplasts (series ed. P. Siekivitz). Plenum Press, New
York, London.
Horsch, R. B., Fry, J. E., Hoffmann, N. L., Eichholtz, D., Rogers, S. G. and
Fraley, R. T. (1985). A simple and general method for transferring genes
into plants. Science 227, 1229-1231.
Hosticka, L. P. and Hanson, M. R. (1984). Induction of plastid mutations in
tomatoes by nitrosomethylurea. J. Hered. 75, 242-246.
Kaul, K. and Sabharwal, P. S. (1971). Effects of sucrose and kinetin on growth
and chlorophyll synthesis in tobacco-D tissue ultures. Plant Physiol. 47, 691695.
Kirk, J. T. O. and Tilney-Bassett, R. A. E. (1978). The Plastids: their
89
Chemistry, Structure, Growth and Inheritance. Amsterdam, New York:
Elsevier/North Holland Biomedical Press.
Knoblauch, M., Hibberd, J. M., Gray, J. C. and van Bel, A. J. E. (1999).
The galinstan expansion femtosyringe allows microinjection of eukaryotic
organelles and prokaryotes. Nature Biotechnol. 17, 906-909.
Köhler, R. H., Cao, J., Zipfel, W. R., Webb, W. W. and Hanson, M. R.
(1997). Exchange of protein molecules through connections between higher
plant plastids. Science 276, 2039-2042.
Köhler, R. H. (1998). GFP for in vivo imaging of subcellular structures in plant
cells. Trends Plant Sci. 3, 317-320.
Medgyesy, P., Fejes, E. and Maliga, P. (1985). Interspecific chloroplast
recombination in a Nicotiana somatic hybrid. Proc. Nat. Acad. Sci. USA 82,
6960-6964.
Menzel, D. (1994). An interconnected plastidom in Acetabularia: Implications
for the mechanism of chloroplast motility. Protoplasma 179, 166-171.
Newcomb, E. H. (1967). Fine structure of protein-storing plastids in bean root
tips. J. Cell Biol. 33, 143-163.
Nichols, B. W., Stubbs, J. M. and James, A. T. (1967). The lipid composition
and ultrastructure of normal developing and degenerating chloroplasts. In
Biochemistry of Chloroplasts, vol. 2 (ed. T. W. Goodwin), pp. 667-690. New
York, London: Academic press.
Nishibayashi, S. and Kuroiwa, T. (1982). Behavior of leucoplast nucleoids
in the epidermal cell of onion (Allium cepa) bulb. Protoplasma 110, 177184.
Osafune, T., Ehara, T., Hase, E. and Schiff Jerome, A. (1993). Stagedependent association of nuclear and chloroplast DNA molecules through
bridges transiently formed between the two organelles in synchronized cells
of Euglena gracilis. J. Electron Microsc. 42, 197-201.
Pyke, K. A. and Page, A. M. (1998). Plastid ontogeny during petal
development in Arabidopsis. Plant Physiol. 116, 797-803.
Sakai, A., Yashiro, K., Kawano, S. and Kuroiwa, T. (1996). Amyloplast
formation in cultured tobacco cells; effects of plant hormones on
multiplication, size, and starch content. Plant Cell Rep. 15, 601-605.
Shalla, T. A. (1962). Assembly and aggregation of tobacco mosaic virus in
tomato leaflets. J. Cell Biol. 21, 253-264.
Spencer, D. and Wildman, S. G. (1962). Observations on the structure of
grana-containing chloroplasts and a proposed model of chloroplast structure.
Aust. J. Biol. Sci. 15, 599-610.
Steffen, K. (1964). Chromoplastenstudien: I. Der amöboide Chromoplastentyp.
Planta 60, 506-522.
Stickens, D. and Verbelen, J. P. (1996). Spatial structure of mitochondria and
ER denotes changes in cell physiology of cultured tobacco protoplasts. Plant
J. 9, 85-92.
Strasburger, E., Noll, F., Scheck, H. and Schimper, A. F. W. (1991). Lehrbuch
der Botanik für Hochschulen (ed. P. Sitte, H. Ziegler, F. Ehrendorfer and A.
Bresinsky), Stuttgart, Jena, New York: Gustave Fischer Verlag.
Suzuki, T., Kawano, S., Sakai, A., Fujie, M., Kuroiwa, H., Nakamura, H.
and Kuroiwa, T. (1992). Preferential mitochondrial and plastid DNA
synthesis before multiple cell divisions in Nicotiana tabacum. J. Cell Sci.
103, 831-837.
Svab, Z. and Maliga P. (1993) High-frequency plastid transformation in
tobacco by selection for a chimeric aadA gene. Proc. Nat. Acad. Sci. USA
90, 913-917.
Thomson, W. W. (1974). Ultrastructure of mature chloroplasts. In Dynamic
Aspects of Plant Ultrastructure (ed. A. W. Robarts), pp. 138-177. London,
New York: McGraw-Hill Book Company (UK) Limited.
Tobias, C. M. and Nasrallah, J. B. (1996). A S-locus-related gene in
Arabidopsis encodes a functional kinase and produces two classes of
transcripts. Plant J. 10, 523-531.
Whatley, J. M. (1983). The ultrastructure of plastids in roots. Int. Rev. Cytol.
85, 175-220.
Whatley, J. M. and Whatley, F. R. (1987). When is a chromoplast? New
Phytol. 106, 667-678.
Wildman, S. G., Hongladarom, T. and Honda, S. I. (1962). Chloroplast and
mitochondria in living plant cells: cinematographic studies. Science 138, 434436.
Wildman, S. G. (1967). The organization of grana-containing chloroplasts in
relation to location of some enzymatic systems concerned with
photosynthesis, protein synthesis, and ribonucleic acid synthesis. In
Biochemistry of chloroplasts, vol. 2 (ed. T. W. Goodwin), pp. 295-319.
London, New York: Academic press.