Columella Cells Revisited: Novel Structures, Novel Properties, and a Novel Gravisensing Model
L. Andrew Staehelin 1 *, Hui Qiong Zheng1 , Thomas L. Yoder2,4 , Jeffrey D. Smith2,5 and Paul Todd3
1
Departments of Molecular, Cellular and Developmental Biology; 2 Aerospace Engineering Science and
3
Chemical Engineering, University of Colorado, Boulder CO; 4 present address: USAF Academy, Department of
Aeronautical Engineering, USAFA, CO; 5 present address: NASA Ames Research Center, Moffett Field CA
ABSTRACT
A hundred years of research has not produced a clear
understanding of the mechanism that transduces the energy
associated with the sedimentation of starch-filled amyloplast
statoliths in root cap columella cells into a growth response.
Most models postulate that the statoliths interact with
microfilaments (MF) to transmit signals to the plasma
membrane (or ER), or that sedimentation onto these organelles
produces the signals. However, no direct evidence for statolithMF links has been reported, and no asymmetric structures of
columella cells have been identified that might explain how a
root turned by 90o knows which side is up. To address these and
other questions, we have (1) quantitatively examined the effects
of microgravity on the size, number, and spatial distribution of
statoliths; (2) re-evaluated the ultrastructure of columella cells in
high-pressure frozen/freeze-substituted roots; and (3) followed
the sedimentation dynamics of statolith movements in reoriented
root tips. The findings have led to the formulation of a new
model for the gravity-sensing apparatus of roots, which
envisages the cytoplasm pervaded by an actin-based cytoskeletal
network. This network is denser in the ER-devoid central region
of the cell than in the ER-rich cell cortex and is coupled to
receptors in the plasma membrane. Statolith sedimentation is
postulated to disrupt the network and its links to receptors in
some regions of the cell cortex, while allowing them to reform
in other regions and thereby produce a directional signal.
INTRODUCTION
Columella cells have occupied a central position in
gravitropism research ever since Nemec (1900) and
Haberlandt (1900) reported on the gravity-dependent
location of their starch-rich amyloplasts. Since then,
numerous researchers have reported on columella cell
structure, development, physiological properties and their
responses to cytoskelton-disrupting drugs, and on how
these features are related to columella cell gravity-sensing
functions (reviewed in Björkman, 1988; Sievers et al.,
1991; Sack, 1991,1994, and 1997; Konings, 1995;
Baluska and Hasenstein, 1997; Chen et al., 1999). These
studies have shown that
• columella cells arise from the calyptrogen cells
of the root apical meristem;
• columella cells possess a polar organization;
• amyloplasts do indeed act as statoliths;
• the distribution and the mobility of
amyloplasts is dependent on the cytoskeleton;
• sedimentation of the statoliths onto sheets of ER
cisternae could play a role in gravity perception
and/or the gravity signaling response.
*Correspondence to: L. Andrew Staehelin: e-mail:
[email protected]
Despite these advances, we are still far from
understanding in mechanistic terms how the physical
process of statolith sedimentation is converted into a
biochemical signal that can be used to bring about a
growth response. While reviewing this state of affairs, we
have become aware that many of the critical observations
underlying current assumptions about columella cell
structure and function are based on experiments that are
no longer state-of-the-art. To this end, we have
undertaken a series of studies to reevaluate the structural
and physiological foundations upon which current
theories of columella cell function are built. These studies
have included
• microgravity experiments designed to determine if plants do indeed “sense” the absence of
gravity, as evidenced by measurable compensatory growth responses (Smith et al., 1997);
• experiments in which we have observed and
quantitatively analyzed the dynamics and
trajectories of sedimenting statoliths in living
cells (Yoder, 1999; Yoder et al., personal
communication);
• experiments in which we have reassessed the
architecture of columella cells and how this
architecture is affected by cytoskeletondisrupting drugs (Zheng and Staehelin, submitted).
These experimental studies, as well as a series of physical
modeling experiments, have resulted in a number of
unexpected findings that cannot be reconciled with
current models of columella cell function. This, in turn,
has led us to formulate the new model of the gravisensory
apparatus of columella cells described in this report.
CHANGES IN STATOLITH MASS AND
GROUPING IN COLUMELLA CELLS OF PLANTS
GROWN IN MICROGRAVITY
The signaling capability of amyloplast statoliths is
based on the principle that their buoyant weight must be
sufficiently large to overcome other interfering forces,
such as thermal motion, cytoskeletal tension, and
cytoplasmic streaming (Björkman, 1988). At the same
time, it is obvious that—with amyloplasts occupying only
a small volume of the gravity-sensing cells—these cells
do not produce the largest possible amyloplast statoliths.
Thus, plants appear to regulate the buoyant weight of their
statoliths for optimal sensory perception. This hypothesis
can be tested by changing the magnitude and direction of
gravitational stimulation and determining if the cells
respond by altering the mass or number of their statoliths
We have performed such experiments by comparing the
Gravitational and Space Biology Bulletin 13(2), June 2000
95
COLUMELLA CELLS REVISITED
differences were seen in plants grown the same
length of time under different simulated gravity
conditions.
Figure 1. Three-dimensional Reconstruction of Contiguous
Clover Root Columella Cells Grown under (a) 1 -g Control,
(b) Microgravity, and (c) 2 Days of Growth in 1-g, Followed
by a Third Day of Clinorotation. The light spheres correspond
to the amyloplast statoliths; the dark spheres to the nuclei. Note
that in the microgravity-grown cells, the statoliths are clustered
towards the cell center and individual statoliths appear bigger
than those in the 1-g control cells. (From Smith et al., 1997)
statoliths and their cellular distribution in clover seedlings
grown in 1 g, microgravity, and on the clinostat (Smith et
al., 1997; Smith et al., 1999).
We germinated and grew seedlings of white clover
(Trifolium repens)—contained in the Fluid Processing
Apparatus developed by BioServe Space Technologies
and described by Smith et al. (1999)—between watermoistened filter paper for 24, 40 and 72 hours. The Fluid
Processing Apparatus let us hydrate the seedlings with
distilled water and grow them in orbit under sterile
conditions, then chemically fix them with 1%
glutaraldehyde at the end of the growth period. For
clinostat experiments, we rotated the seedlings only
during the final 24 hours of their growth (since longer
exposure to the clinostat has led to severe perturbations in
the architecture of the columella cells). When the
glutaraldehyde-fixed samples were removed from the
apparatus, the root tips were excised, fixed in osmium
tetroxide, dehydrated, and embedded in Spurr's resin. We
serially sectioned the root tips (0.5 µm thick), digitized
the light microscopic images of the columella cells, and
reconstructed groups of adjacent cells using the
Reconstruction of Serial Sections (ROSS) software
developed at the Biocomputation Center (NASA Ames
Research Center, Moffett Field, CA) (Figure 1). Some
root tips were also sectioned for electron microscopic
analysis to determine the starch content and the
ultrastructural features of the statolith amyloplasts.
The principal findings of these studies were as
follows:
1. Statolith amyloplasts differed in their
architecture from starch-storage amyloplasts by
having a 0.1- to 0.3-µm-thick boundary layer
around the starch granules.
2. The boundary layer appeared to link the different
starch granules into a single structural unit
within the amyloplasts.
3. Plants grown 40 hours contained more starch per
amyloplast than those grown 72 hours; but no
96
Gravitational and Space Biology Bulletin 13(2), June 2000
In addition, serial section reconstruction experiments provided quantitative information about the size,
shape, volume, and spatial relationships of groups of
columella cells, as well as the size and spatial distribution
of statoliths and the size and position of the nuclei
(Figure. 1). The most striking and potentially most
important finding was that, in plants of equivalent age, the
size of the amyloplasts increased in microgravity relative
to the amyloplasts of ground-grown plants, but remained
constant in seedlings grown on the clinostat. The number
of amyloplasts per cell was found to be proportional to
cell volume in both ground and microgravity-grown
plants, and to be decreased in clinorelated specimens.
Moreover, under none of the conditions tested did the
amyloplasts become randomly organized. Thus, in
microgravity, they were grouped near the cell centers; in
clinostat-treated cells, they appeared more dispersed but
still retained some grouping.
Based on these findings, it appears that plants do
sense the absence of gravity stimulation when grown
under microgravity conditions, as evidenced by their
attempt to compensate for this loss by increasing the mass
of the statoliths. These findings support the hypothesis
that the size of amyloplasts in columella cells is regulated
by a feedback control system that is integrated into the
gravity-sensing/signaling pathway. The observed grouping of statoliths suggests that plants might increase the
signal-to-noise ratio of their amyloplast statolith gravitysensing system by maintaining the statoliths in a grouped
configuration. The mechanism that promotes this
grouping has yet to be elucidated, but could involve some
kind of tethering between the statoliths or their exclusion
Figure 2. Electron Micrograph of a Cryofixed/FreezeSubstituted Tobacco Columella Cell. (Membranous organelles
are traced to highlight their distribution.) The tubular ER is
confined to a tight band underlying the plasma membrane. With
the exception of the mitochondria (M), all of the other
organelles are limited to the central, ER-devoid region of the
cell. Am=amyloplast statolith; G=Golgi; V=vacuole.
COLUMELLA CELLS REVISITED
Figure 3. Electron Micrograph of Two Interconnected Nodal
ER Membrane Domains in a Cryofixed/Freeze-substituted
Tobacco Columella Cell. Sheetlike rough ER cisternae appear
attached along their edges to the ~100 nm-in-dia. nodal rods.
Am=amyloplast statolith; G=Golgi stack; M=mitochondrion.
from a gel-like cytosolic cytokeletal matrix. We postulate
that the latter type of system is responsible for the
grouping behavior (discussion follows). Our data also
provide further evidence that clinostats constitute a less
than ideal system for mimicking microgravity, since the
chronic gravity signaling the overload to which plants
grown on clinostats are subjected leads to harmful
changes in columella cell architecture, as already reported
by Hensel and Sievers (1980) and Guikema et al. (1993).
NODAL ER, A NOVEL FORM OF ER FOUND
EXCLUSIVELY IN COLUMELLA CELLS
Root cap columella cells exhibit a distinct
architectural polarity, postulated to be related to their
gravity-sensing function. This polarity manifests itself in
vertically oriented roots with the nuclei located adjacent
to the top (proximal side) and the sedimented amyloplast
statoliths concentrated close to the bottom (distal end) of
the cells. Electron micrographs of columella cells have
also demonstrated that most of the ER is located in the
cell periphery (Sack and Kiss, 1989). The observed sedimentation of statoliths onto the surface of sheets of distal
ER membranes has led to the hypothesis that statolith-ER
interactions could mediate the gravity response.
Chemical fixation is known to frequently produce
artifactual images of cellular structures (Gilkey and
Staehelin, 1986). Therefore, because virtually all of our
structural knowledge of columella cells has come from
studies of chemically fixed cells, we have reexamined the
ultrastructure of columella cells of tobacco (Nicotiana
tabaccum) seedlings preserved by high-pressure freezing/
freeze-substitution techniques (Kiss and Staehelin, 1995).
These studies have not only produced a number of new
insights into the spatial distribution of organelles and the
nature of the cytosolic matrix, but they have also led to
the discovery of a new form of ER that is unique to
columella cells and has not been reported previously.
Figure 2 depicts an electron micrograph of a
longitudinal
section
through
a
cryofixed/freeze-
substituted columella cell in which all of the membranous
organelles have been traced to highlight their spatial
distribution. Most striking is the presence of a tubular ER
network in the cell periphery that exhibits a very sharp
transition to the ER-free central region of the cell. Moreover, due to the compactness of this tubular ER network,
all of the major membranous organelles—such as
amyloplasts, Golgi, and vacuoles—are excluded from the
ER-rich cell cortex and confined to the central region.
Only mitochondria are observed in both of these regions.
The cytosol of the central region is also unusual in that it
occupies more of the cytoplasmic volume than other root
tip cells and is devoid of actin filament bundles. Instead, it
appears to be composed of randomly oriented single actin
filaments that, together with other molecules, form a
network-like cytoskeletal matrix.
A novel form of ER discovered in our cryofixed
specimens is illustrated in Figure 3. Based on its
architecture, we have termed this "nodal ER." These
unique domains consist of a central "nodal rod" element
to which three to eight (usually seven) rough ER cisternae
are attached. The edge-on attachment of the ER
membranes to the rod appears to be partly responsible for
the sheetlike organization of the ER membranes, an
organization that stimulates the binding of polysomes. We
have mapped the spatial distribution of these nodal ER
domains within central and flanking columella cells and
found that, in both cell types, most of these domains are
localized into cortical patches near the equator of the
cells. In flanking cells, they also occur in the basal region.
Careful analysis of the spatial relationship among nodal
ER domains, the tubular ER in the cell cortex, and the
amyloplasts has shown that these specialized ER domains
(1) are located at the interface between the cortical tubular
ER and the central domain, and (2) constitute a physical
barrier to the amyloplasts, which prevents them from
approaching the tubular ER and the plasma membrane.
On a more global scale, the highest density of nodal ER
domains is seen along the outer walls of the flanking columella cells (Figure 4). This type of asymmetric arrangement is what one might expect of a structural system that
Figure 4. Distribution of Nodal ER Domains (Stars) in a
Reconstructed 0.2 -µM-Thick Cross-section through a
Tobacco Root Tip at the Level of the Second-Tier Columella
Cells. Note that the majority of the nodal ER domains in the
columella cells (central white cells) are found in the outer cortex
region of the flanking cells.
Gravitational and Space Biology Bulletin 13(2), June 2000
97
COLUMELLA CELLS REVISITED
Figure 5. Tensegrity-based Model of the
Gravisensing Apparatus of Columella Cells.
The model shows an actin-based cytoskeletal
network (crosshatched lines) that
•
•
•
pervades the entire cytoplasm;
is denser in the cell center than in
the cell cortex;
is coupled to stretch-sensitive
receptors in the plasma membrane.
We postulate that the model’s numbered
statoliths are not linked to the cytoskeletal
network. They function to activate/inactivate the
receptors in the plasma membrane by locally
disrupting the network as they sediment to a new
location, allowing new connections to be formed
at sites they vacate. The asymmetrically
distributed nodal ER domains shield local
plasma membrane sites from approaching
statoliths, and may thereby provide a
directionality vector to the sensing system. The
interface between the ER-rich cortical and the
ER-devoid central regions of the cell constitutes
a plane of weakness in the cytoskeletal network,
and the statoliths travel preferentially within this
region. In side-to-side sedimentation experiments
(e, f, and g), the statoliths first move horizontally
towards forming "channels" before they pass
through the channels to the lower side of the cell.
enables the columella cells to produce an asymmetric
gravisensing signal, leading to expansive growth on the
upper side and to a reduction in cell expansion on the
lower side of a root turned by 90o . This role has been
incorporated into the new gravisensing model depicted in
Figure 5.
AMYLOPLAST SEDIMENTATION
MAIZE COLUMELLA CELLS
DYNAMICS
IN
We have used videomicroscopy to analyze the
sedimentation behavior of amyloplast statoliths in living
columella cells of Zea mays seedlings (Yoder, 1999;
Yoder et al., personal communication). Control cells
displayed no cytoplasmic streaming, and their statoliths
exhibited only random movements characteristic of
Brownian motion. In roots rotated by 90o , the statoliths
moved downward along the distal wall at an average
velocity of 1.7 µm/min and then spread out along the new
cell bottom (Figure 6A). After this settling, the roots were
rotated along their longitudinal axis by 180o to position
the statoliths along the top wall. When forced to traverse
the complete width of the cells, the statoliths displayed an
unexpected behavior. Instead of settling as individual
units across the cytoplasm, the majority initially moved
horizontally towards "channels" and then translocated
98
Gravitational and Space Biology Bulletin 13(2), June 2000
vertically
through
these
channels
(Figure
6B).
Furthermore,
unlike
the
distal-to-side
sedimenting
statoliths that displayed their highest velocity at the
beginning and then slowed down monotonically as they
approached the lower plasma membrane, statoliths that
crossed the central region were slowest at the beginning
and then accelerated once they reached a channel. In the
presence of the actin-disrupting drug cytochalasin D at a
concentration of 100 µM (which does not completely
disrupt all actin filaments), the channeling effect was less
pronounced, suggesting that the channeling is related to
the presence of an actin-based cytoskeletal network in the
central region of the cell. In such cells, the statoliths were
also seen to accelerate sooner than in the control cells,
suggesting that the actin-based network resists penetration
by the statoliths and that the slow initial acceleration was
related to the formation of the channels. Ultrastructural
studies of high-pressure frozen/freeze-substituted cells
revealed that maize columella cells possess the same type
of tubular cortical ER system as the tobacco cells
discussed earlier. This finding suggests that the lowresistance statolith pathway in the cell periphery, which
gives rise to the horizontal motion in side-to-side
sedimentation experiments, could be associated with the
interface between the ER-rich cortical and the ER-devoid
central region of these cells.
COLUMELLA CELLS REVISITED
A NEW MODEL OF THE GRAVITY-SENSING
APPARATUS OF COLUMELLA CELLS
The sedimentation behavior of statoliths in maize
columella cells described in the preceding section is
difficult to reconcile with current hypotheses for the
gravisensing
mechanism
of
columella
cells—most
notably with models in which the statoliths are postulated
to be tethered to receptors in the ER or the plasma
membrane through actin filaments (Björkman, 1988;
Sievers et al., 1991; Sack, 1994; Baluska and Hasenstein,
1997). Our findings lead us to formulate a new hypothesis
to explain how the sedimentation of statoliths in
reoriented columella cells may locally stimulate or inhibit
stretch-activated receptors in the plasma membrane to
produce a gravity signal.
The model in Figure 5 envisions the entire
cytoplasm pervaded by an actin-based cytoskeletal
network that is denser in the ER-devoid central region of
the cell than in the ER-rich peripheral cytoplasm. This
network is postulated to be attached to stretch-sensitive
receptors in the plasma membrane and to be associated
with microtubules and membranous organelles to form a
tensegrity-based force interaction system (Ingber, 1993
and 1998). In this scheme, the sedimentable amyloplast
statoliths are not bound to any of the components of the
cytoskeletal network; instead, they are postulated to
function by locally disrupting the network. Thus,
displacement of the statoliths is envisioned to produce a
signal by disrupting links from the cytoskeletal network to
the plasma membrane in some places and simultaneously
allowing such links to form elsewhere (depicted in Figure
5). Since all of the tensional elements are mechanically
coupled to each other, this redistribution of cytoskeletalto-plasma membrane links also simultaneously produces
more global tensional changes throughout the network,
which may aid in the integration of the signals of
individual receptors. Consistent with this idea of the
importance of actin filaments for gravitropic signaling,
Guikenia and Gallegos (1992) have reported that, when
cytochalasin D is applied in agar blocks at greater than
40ìg/ml to maize root caps, the roots lose their ability to
directionally reorient themselves in response to a change
in the gravitational field. Furthermore, since the primary
changes would involve redistribution of links to the
plasma membrane, the directionality of the signaling
could be refined by the distribution of nodal ER domains
that may locally shield some links from the statoliths’
effects on the cortical cytoskeleton. This model makes a
number of predictions that can be experimentally tested.
For example, the observed clustering of amyloplasts near
the center of columella cells of microgravity-grown
seedlings is predicted to result from the exclusion of the
statoliths from the surrounding cytoskeletal network and
not from the tethering of the statoliths to each other via
actin-filament links. This and other predictions will be
tested by using optical tweezers to measure the forces
needed to displace individual and grouped statoliths in
different regions of central and flanking columella cells.
Figure 6. Tracings of Sedimenting Amyloplast Statoliths in
Maize Columella Cells. (A) Typical distal-to-side movement
profile of 7 statoliths with 15 seconds between frames. (B)
Typical side-to-side movement profile with 9 statoliths falling
the entire width of the columella cell with 30 seconds between
frames. Note that most but not all statoliths pass through a
common vertical channel.
CONCLUSIONS
While confirming many classical ideas about how
columella cells use amyloplast statoliths to sense gravity,
the studies described in this paper have also yielded
findings that challenge several proposals concerning the
actual energy transduction mechanism. In particular, our
data are difficult to reconcile with the concept of tethered
statoliths, including how such statoliths may transmit the
physical force of sedimentation to stretch receptors in the
plasma membrane. Most notably, our data suggest that
1. the statoliths produce a signal by locally
disrupting a tensegrity-based cytoskeletal
network;
2. specialized ER domains could modify the
network disruption signals to add a directionality component to the sensing system.
Thus, when these statoliths are induced to sediment to a
new location, the translocation causes a disruption of the
cytoskeletal network-to-plasma membrane-receptor links
in some places and to the reformation of such links in
others. The combination of these two sets of stimuli then
Gravitational and Space Biology Bulletin 13(2), June 2000
99
COLUMELLA CELLS REVISITED
would form the basis of an integrated gravisensing
response.
Nemec, B. 1900. Über die Art der Wahrnehmung des
Schwerkraftes bei den Pflanzen. Berichte der Deutschen
Botanischen Gesellschaft 18:241-245.
Acknowledgements
This work was supported by grants NAG 5-3967 and
NCC 8-131 from the National Aeronautics and Space
Administration
REFERENCES
Baluska, F. and Hasenstein, K.H. 1997. Root
cytoskeleton: its role in perception of and response to
gravity. Planta 203:S69-S78.
Björkman, T. 1988. Perception of gravity by plants.
Advances in Space Research) 15:1-41.
Chen, R., Rosen, E. and Masson, P.H. 1999. Gravitropism
in higher plants. Plant Physiology 120:343-350.
Gilkey, J.C. and Staehelin, L.A. 1986. Advances in
ultrarapid freezing for the preservation of cellular
ultrastructure.
Journal
of
Electron
Microscopy
Technology 3:177-210.
Guikema, J.A., and Gallegos, G.L. 1992. Plastids:
Dynamic components of plant cell development.
Transactions of the Kansas Academy of Sciences 90:5054.
Guikema, J.A., Hilaire, E. and Odom, W.R. 1993.
Cytoskeleton-amyloplast interactions in sweet clover.
Transactions of the Kansas Academy of Science 96:13-19.
Haberlandt, G. 1900. Über die
geotropischen
Reizes.
Berichte
Botanischen Gesellschaft 18:261-272.
Perzeption des
der
Deutschen
Hensel, W., and Sievers, A. 1980. Effects of prolonged
omnilateral gravistimulation on the ultrastructure of
statocytes and on the graviresponse of roots. Planta
150:338-346.
Ingber, D. 1993. Cellular tensegrity: defining new rules of
biological design that govern the cytoskeleton. Journal of
Cell Science 104:613-627.
Ingber, D. 1998. The
American 278:30-39.
architecture
of
life. Scientific
Kiss, J.Z. and Staehelin, L.A. 1995. High pressure
freezing. In: Rapid Freezing, Freeze-Fracture and Deep
Etching: The State of the Art (Severs, N. J. and Shotton,
D. M., Eds.) New York: John Wiley-Liss, Inc..
Konings, H. 1995. Gravitropism of roots: an evaluation of
progress during the last three decades. ACTA Botanica
Neerlandica 44(3):195-223.
100
Gravitational and Space Biology Bulletin 13(2), June 2000
Sack, F.D. 1991. Plant gravity sensing. International
Review of Cytology 127:193-252.
Sack, F.D. 1994. Cell biology of plant cell gravity
sensing. Advances in Space Research 14:117-119.
Sack, F.D. 1997. Plastids and gravitropic sensing. Planta
203:S63-S68.
Sack, F.D. and Kiss, J.Z. 1989. Root cap structure in wild
type and in a starchless mutant of Arabidopsis. American
Journal of Botany 76:454-464.
Sievers, A., Buchen, B., Volkmann, D. and Hejnowicz, Z.
1991. Role of the cytoskeleton in gravity perception. In:
The Cytoskeletal Basis of Plant Growth and Form (Lloyd,
C., Ed.) London: Academic Press.
Smith, J.D., Todd, P. and Staehelin, L.A. 1997.
Modulation of statolith mass and grouping in white clover
(Trifolium repens) grown in 1-g, microgravity and on a
clinostat. Plant Journal 12:1361-1373.
Smith, J.D., Staehelin, L.A. and Todd, P.W. 1999. Early
root cap development and graviresponse in white clover
(Trifolium repens) when grown in space and on a two axis
clinostat. Journal of Plant Physiology 155:543-550.
Smith, J.D., Todd, P. and Staehelin, L.A. 1997.
Modulation of statolith mass and grouping in white clover
(Trifolium repens) grown in 1-g, microgravity and on a
clinostat. Plant Journal 12:1361-1373.
Yoder, T.L. 1999. Characterizing the plant root
gravitropic mechanism: a systems modeling approach.
ASEN. Boulder, University of Colorado; 445.