Analysis of the Golgi Apparatus in Arabidopsis Seed Coat Cells
during Polarized Secretion of Pectin-Rich Mucilage
Robin E. Young,aHeather E. McFarlane,bMichael G. Hahn,cTamara L. Western,bGeorge W. Haughn,a
and A. Lacey Samuelsa,1
aDepartment of Botany, University of British Columbia, Vancouver, Canada V6T 1Z4
bDepartment of Biology, McGill University, Montreal, Canada H3A 1B1
cComplex Carbohydrate Research Center and Department of Plant Biology, University of Georgia, Athens, Georgia 30602
Differentiation of the Arabidopsis thaliana seed coat cells includes a secretory phase where large amounts of pectinaceous
mucilage are deposited to a specific domain of the cell wall. During this phase, Golgi stacks had cisternae with swollen
margins and trans-Golgi networks consisting of interconnected vesicular clusters. The proportion of Golgi stacks producing
mucilage was determined by immunogold labeling and transmission electron microscopy using an antimucilage antibody,
CCRC-M36. The large percentage of stacks found to contain mucilage supports a model where all Golgi stacks produce
mucilage synchronously, rather than having a subset of specialist Golgi producing pectin product. Initiation of mucilage
biosynthesis was also correlated with an increase in the number of Golgi stacks per cell. Interestingly, though the morphol-
ogy of individual Golgi stacks was dependent on the volume of mucilage produced, the number was not, suggesting that
proliferation of Golgi stacks is developmentally programmed. Mapping the position of mucilage-producing Golgi stacks
within developing seed coat cells and live-cell imaging of cells labeled with a trans-Golgi marker showed that stacks were
randomly distributed throughout the cytoplasm rather than clustered at the site of secretion. These data indicate that the
destination of cargo has little effect on the location of the Golgi stack within the cell.
The secretory system of plants is integral to the synthesis and
deposition of many cell wall components, including pectin (Doblin
et al., 2003). The plant Golgi apparatus is comprised of large
numbers of individual Golgi stacks, which are widely distributed
mediated by the actin cytoskeleton (Griffing, 1991; Boevink et al.,
1998; Nebenfu ¨hr et al., 1999). During the late stages of mitosis,
the Golgi apparatus proliferates and plays an integral role in the
synthesis of the cell plate between daughter cells (Hirose and
Komamine, 1989; Ju ¨rgens, 2005; Segui-Simarro and Staehelin,
2006). The capacity for Golgi streaming seen in plants and many
cargo to and from the Golgi that are not available to species in
which the Golgi is maintained in a single, centralized position
(Nebenfu ¨hr and Staehelin, 2001). Not only do plant Golgi stacks
have the capacity to travel to the site of deposition if needed, but
also targeting of vesicles to and from the Golgi is complicated by
the fact that Golgi stacks are in motion.
Much of the cargo that passes through the plant secretory
pathway to the plasma membrane is comprised of cell wall
polysaccharides synthesized in the Golgi. The plant cell wall is a
dynamic extracellular matrix, and its organization is integral to its
proper function (Somerville et al., 2004; Cosgrove, 2005). The
current model for primary cell wall structure proposes a series of
macromolecular networks that are intertwined and intercon-
1993; Carpita and McCann, 2000) necessary for the plant to
withstand both tension and compression, while at the same time
having the ability to grow in a precisely controlled and directed
to consist of cellulose and hemicellulose. In the Arabidopsis
thaliana primary cell wall, the predominant hemicellulose is
xyloglucan (XG). Unlike the plasma membrane–localized syn-
thesis of cellulose, XG has been shown to be synthesized inside
the Golgi and then secreted to the outside of the plasma
membrane via vesicular transport (Moore et al., 1991). The
hemicellulose-cellulose framework described above is thought
gel matrix composed primarily of pectic polysaccharides, such
as homogalacturonans, rhamnogalacturonan I (RGI), and rham-
nogalacturonan II. Pectin biosynthesis is also a Golgi-mediated
process, as first shown by autoradiography (Northcote and
Theuse of antibodies has furtherexpanded our understanding
of the flexibility of the Golgi apparatus in the synthesis of cell wall
polysaccharide/glycoprotein components, though there is still
1Address correspondence to firstname.lastname@example.org.
The author responsible for distribution of materials integral to the
findings presented in this article in accordance with the policy described
in the Instructions for Authors (www.plantcell.org) is: A. Lacey Samuels
WOnline version contains Web-only data.
OAOpen Access articles can be viewed online without a subscription.
The Plant Cell, Vol. 20: 1623–1638, June 2008, www.plantcell.org ª 2008 American Society of Plant Biologists
production of cell wall polysaccharides in suspension-cultured
cells, which are constitutively secreting cell wall components to
all regions of the plasma membrane, is a stepwise process
occurring from cis to trans and that different products can be
synthesized in the same Golgi stack. A separate study showed
that transport vesicles in clover root tips, another diffusely
secreting cell type, can carry both XG and RGI epitopes at the
same time (Lynch and Staehelin, 1992). In the Arabidopsis root
cap, the ultrastructure of the Golgi apparatus was observed to
undergo changes that were consistent with the increased levels
of secretion that occuras the meristematic cells differentiate into
mucilage producing cells (Staehelin et al., 1990). However, in all
in a given cell, as a possible mechanism for accommodating
increased levels of secretion, were not determined.
Arabidopsis seeds synthesize and secrete large amounts of
time period of seed maturation (between 6 and 8 d postanthesis
[DPA]). The mucilage is secreted into a donut-shaped pocket at
Windsor etal.,2000).Followingmucilagesecretion,the epidermal
results in the formation of a volcano-shaped protrusion through
the center of the mucilage pocket at seed maturity. A number
of mutants with defects in the capacity of these cells to produce
and/or secrete mucilage have been studied, including mucilage
modified4 (mum4) (Western et al., 2001, 2004), in which the pro-
duction of mucilage in the seed coat has been disrupted. The
MUM4 gene encodes a putative rhamnose synthase that is
upregulated during the mucilage-producing stage of seed coat
development (Usadel et al., 2004; Western et al., 2004).
In this article, we have investigated how the secretory system
can accommodate striking changes in both the amount and type
The onset of enhanced polysaccharide production was accom-
stacks in the seed coat cells. The ultrastructural changes in the
as mutants lacking mucilage did not have the characteristic
observed in the cell was not dependent on mucilage production,
as the Golgi number in mum4 mutants still increased in a devel-
opmentally appropriate manner. The mucilage-containing Golgi
stacks were not clustered near the site of secretion but were
instead randomly distributed throughout the cell. These data
cues and surges in production as a coordinated unit.
Seed Coat Cells Undergo Distinct Ultrastructural
Rearrangements during Differentiation
An overview of seed coat cell differentiation was undertaken to
examine the organellar rearrangements that occur in the whole
cell and to put polysaccharide production and secretion into this
context. Arabidopsis seeds were excised from siliques and high
pressure frozen before (at 4 DPA), during (at 7 DPA), and after (at
9 DPA) the mucilage-producing stages of seed coat develop-
ment and the ultrastructure, specifically in the Golgi apparatus,
Prior to the mucilage-secreting phase (4 DPA), epidermal cells
had a large central vacuole, a nucleus typically located at the
basolateral side of the cell, and amyloplasts (containing starch
were seen in the narrow cytosolic region surrounding the vacu-
ole. Cisternae were long and relatively thin, and the trans-Golgi
network (TGN) was small, compared with other stages of devel-
opment (Figure 1B). In many cases, the Golgi appeared cup-
shaped, with the TGN on the concave surface.
During mucilage secretion (7 DPA), seed coat epidermal cells
had well-developed, ring-shaped mucilage pockets between the
plasma membrane and primary cell wall, with the vacuole limited
granules were positioned in the upper portion of the cytoplasmic
column. Interestingly, microtubules were seen lining the mem-
brane on the cytosolic side of the mucilage pocket but not in
other regions of the cell (McFarlane et al., 2008). The Golgi stack
morphology at 7DPA differed from the 4 DPA morphology in that
the stacks had a shorter diameter than those observed at 4
DPA, and the cis-trans polarity across the stack was more
pronounced. The lumen of the cis cisternae was wider, while the
medial and trans cisternae had compressed lumens, like those
at 4 DPA, except with swollen margins (cf. Figures 1B and 1D). In
addition, there appeared to be more TGN and more large ves-
icles associated with the TGN (Figure 1D).
Considering that a large, convoluted TGN might be difficult to
capture in thin sections, as connections between portions of the
TGN might be outside of the plane of section, electron tomog-
raphy was used to examine the morphology of the 7-DPA Golgi
stack in greater detail. Models that had been built from three
different tomograms showed that much of what appeared to be
free vesicles near the TGN in thin sections were interconnected
vesicular clusters (Figure 2). During the intense secretion of
mucilage that is characteristic of 7-DPA seed coat cells, the
swollen vesicular regions of the TGN lacked obvious coats and
were large (;200 nm). This is morphologically distinct from the
partially coated reticulum, an interconnected cluster of coated
vesicles, with dilations that are approximately half the size of
those described here (Pesacreta and Lucas, 1985). The clusters
in this study appeared to be budding from flattened cisternal
Golgi stack (Figures 2C and 2D). The TGN-interconnected
vesicular clusters were immediately adjacent to the trans-most
cisterna, in agreement with the results of Sequi-Simarro and
Staehelin (2006), where all TGN and Golgi stacks in meristematic
cells were tracked, demonstrating the proximal arrangement of
TGN to Golgi stacks.
After the secretion of mucilage ended (9 DPA), seed coat cells
had a distinguishable layer of secondary cell wall lining the apical
side of the cell, between the mucilage pocket and the plasma
membrane and along the top of the cytoplasmic column (Figure
1E). The vacuole was still located at the basal side of the cell but
1624The Plant Cell
appeared smaller in volume. Microtubules around the mucilage
pocketwerenotprominent atthisstage.Golgi stacksweresimilar
to those seen in 7-DPA cells, exhibiting short, compressed cis-
ternae with swollen margins and extensive TGN (Figure 1F).
Antibodies to Mucilage Components Demonstrate That
Plant Golgi Stacks Produce Mucilage Synchronously
during Seed Coat Epidermal Differentiation
Previous work hasshown thatdifferent antipectin antibodies can
beusedto labelhydrated,extrudedmucilage frommature seeds
follow mucilage secretion within the cell, we required an antibody
thatreactedstronglywithmucilage incryofixed, resin-embedded,
sectioned material. We initially tested a range of antibodies for
their ability to bind extruded mucilage (see Methods for compre-
hensive list). Though there was a number of antibodies from this
screen that were capable of binding hydrated mucilage in mature
seeds (see Supplemental Figure 1 online), only two of those were
capable of binding to mucilage in cryofixed, sectioned material
(CCRC-M36 and antixyloglucan [a-XG]; Figures 3, 4, and 6A).
Thus, CCRC-M36 and a-XG were chosen to investigate changes
Figure 1. The Stages of Arabidopsis Seed Coat Development in Cryofixed Samples.
(A) Cells (4 DPA) are relatively undifferentiated, with a large central vacuole.
(B) Detail of a 4-DPA Golgi stack.
(C) Cells (7 DPA) show the characteristic mucilage pocket, forming a ring around a central cytoplasmic column.
(D) Detail of a 7-DPA Golgi stack.
(E) Cells (9 DPA) have secondary cell wall on the apical side of the cell.
(F) Detail of a 9-DPA Golgi stack.
a, amyloplast; c, cis face of Golgi stack; 1cw, primary cell wall; 2cw, secondary cell wall; m, mucilage; n, nucleus; t, trans face of Golgi stack; V, vacuole.
Bars ¼ 5 mm in (A), (C), and (E) and 250 nm in (B), (D), and (F).
Dispersed Plant Golgi during Polarized Secretion1625
in polysaccharide product by immunolabeling sections of seeds
before, during, and after mucilage production.
To study pectinaceous mucilage, an ideal probe would react
strongly with the pectin of the mucilage but only weakly or not at
all with the pectins of the cell wall. CCRC-M36 is a monoclonal
antibody raised against Arabidopsis seed coat mucilage, and it
binds strongly to RGIs from Arabidopsis, tomato (Solanum
lycopersicum), lettuce (Lactuca sativa), and soybean (Glycine
max) as well as to Arabidopsis and mustard (Sinapis alba) seed
coat mucilages (T. Bootten, Z. Popper, A.G. Swennes, and M.G.
Hahn, unpublished data). Its epitope appears to reside on the
unbranched backbone of RGI, and it does not bind strongly with
uronans (C. Deng, R. Jia, A. Albert, W.S. York, M.A. O’Neill, and
M.G. Hahn, unpublished data). At 4 DPA (Figure 3A), prior to the
onset of mucilage production, CCRC-M36 showed no reactivity
to any cell walls in cryofixed, sectioned seeds, including specif-
ically the seed coat cells (Figure 3D). At 7 DPA (Figures 3B and
4A), at the height of mucilage production, the mucilage pockets
of seed coat cells between the primary cell wall and plasma
membrane were brightly labeled, but the cell walls of these and
other cell types were not (Figures 3E and 4C). Punctate struc-
tures were observed in the cytosol of 7-DPA seed coat cells but
not at 9 DPA (Figures 4C and 4D), when polysaccharide produc-
tion had shifted from the production of mucilage to the produc-
tion of secondary cell wall (Figures 3C and 4B). Labeling of the
mature mucilage pockets was strong in 9-DPA seed coat cells
(Figures 3F and 4D).
To contrast with the binding of CCRC-M36 to mucilage, an
antibody that reacted with the cell wall was required. Though a
number of typical cell wall antibodies were tested against pri-
a-XG (Figures 4E and 4F). This antibody was raised against
sycamore maple (Acer pseudoplatanus) XG and does not cross-
react with pectic components, like RGI (Moore and Staehelin,
1988). In addition to the labeling of developing secondary cell
walls in the Arabidopsis seed coat, this antibody also labeled
primary cell walls of the seed throughout development, and the
mucilage pockets, but at a less intense level than CCRC-M36
(Figures 3G to 3I and 6A). Due to the nature of the cellulose-
hemicellulose network, these results suggested that cellulose
confirmed by labeling hydrated, extruded mucilage with cellu-
lose binding domain, conjugated to the fluorescent molecule
Oregon Green (CBD-OG; see Supplemental Figure 1D online), a
result which is consistent with the study of Macquet et al. (2007).
The changes in the CCRC-M36 labeling patterns observed in
7-DPA seed coat cells indicated that a closer examination of the
secretory apparatus might yield new information about the reg-
ulation of the secretory pathway during seed coat development.
Since CCRC-M36 labels only mucilage and not other primary cell
wall components in this cell type, this antibody was used to
the production of mucilage during seed coat development. To
resolve the puncta observed in fluorescence microscopy and to
quantify individual Golgi stacks producing mucilage, sections of
7- and9-DPA seeds were immunolabeled witheither CCRC-M36
or a-XG together with gold-conjugated secondary antibodies for
examination by transmission electron microscopy (TEM) (Figure
5). CCRC-M36 labeled 70% of all Golgi stacks visible in 7-DPA
seed coat sections, most commonly labeling the medial and/or
trans cisternae as well as the TGN. By contrast, 0% of Golgi
stacks were labeled by CCRC-M36 in 9-DPA cells. These data
indicate that the pectin epitope is indeed present in the Golgi
stages during synthesis of the secondary cell wall. Thus, the
short-term, specialized production of mucilage is reflected in the
16% of Golgi stacks in 7-DPA cells, and 22% of Golgi stacks in
9-DPA cells, suggesting that XG is secreted during the synthesis
of both the mucilage and columella.
Despite the high percentage of Golgi stacks that label with
CCRC-M36 at 7 DPA, it is conceivable that the epitopes recog-
nized by the two different antibodies are being carried by
different groups of Golgi stacks. To address this question,
double labeling experiments, using both CCRC-M36 and a-XG,
were performed on 7-DPA seed coat cells (Figure 6). Results of
doublelabeling wereconsistentwithsinglelabeling assays;77%
of Golgi stacks labeled with CCRC-M36 in the double labeling
experiment, whereas 11% of Golgi stacks labeled with a-XG. Of
the six Golgi stacks observed to have a-XG label, only one of
Figure 2. Tomographic Reconstruction of a Single Golgi Stack during
the Mucilage-Secreting Phase of an Arabidopsis Seed Coat Cell (7 DPA).
(A) and (B) Slices (2 nm) from a tomogram, showing how the TGN
appears as a series of vesicles.
(C) and (D) En face (C) and side (D) view of a three-dimensional model of
a Golgi stack, based on tomographic reconstruction shown in (A) and
(B). The TGN (bright pink) is an extensive interconnected network of large
c, cis face of Golgi stack; t, trans face of Golgi stack. Bars ¼ 250 nm.
1626 The Plant Cell
them did not also have CCRC-M36 label, suggesting that the
within the cell.
The Number of Golgi Stacks in the Cell Increases during
To determine whether the onset of mucilage production resulted
in changes in the number of individual stacks in the Golgi
apparatus, the number of Golgi stacks in 4-, 7-, and 9-DPA cells
was quantified. This was done using a twofold approach. First,
high-resolution TEM of seed coat cells at each developmental
stage was undertaken to count Golgi stacks in seed coat cells
directly. When thin sections of seed coat cells were examined in
TEM, an increase in the number of Golgi stacks was observed
(Figure 7A). However, during the same period (between 4 and 7
DPA), seed coat cells underwent dramatic cytosolic rearrange-
ments (cf. Figures 1A and 1C), which manifested as an increase
and 7 DPA (Figure 7B). Thus, the surface area of cytoplasm
available for quantification at 4 DPA was significantly less than at
7 DPA, which could have resulted in an underrepresentation of
the true number of Golgi stacks present in 4-DPA cells. For this
reason, we determined the density of Golgi stacks (number of
Golgi stacks/mm2) visible in our sections by dividing the absolute
number of stacks by the cytoplasmic surface area (Figure 7C).
Three-way analysis of Golgi density at 4, 7, and 9 DPA showed
statistically significant differences [Kruskal-Wallis test; H(2) ¼
66.94, P < 0.05]. Post hoc analysis using the Mann-Whitney test
showed a significant increase in the density of Golgi stacks at 7
The number of Golgi stacks/mm2approximately doubled be-
tween 4 and 7 DPA (0.08 stacks/mm2versus 0.14 stacks/mm2,
respectively). By contrast, when the density of Golgi stacks in
7- and 9-DPA seed coat cells were compared, no significant
differences were found. This demonstrates that an increase in
the number of Golgi stacks found in seed coat cells is correlated
with the increase in polysaccharide production at the onset of
The second method that was used to quantify Golgi stacks
was live imaging of Golgi stacks in whole seed coat cells via
fluorescent molecular markers for the Golgi. Developing seeds
were dissected from siliques of transgenic Arabidopsis plants
bearing the trans-Golgi marker sialyl transferase conjugated to
green fluorescent protein (ST-GFP), driven by the constitutive
35S promoter (Boevink et al., 1998). Despite reports that the 35S
promoter results in gene expression in all tissues (Odell et al.,
1985), the expression of ST-GFP in the Arabidopsis seed coat
was patchy, with only a few cells in each seed coat displaying
fluorescence (Figure 8A). In cells that exhibited fluorescence,
were observed (Figure 8). These puncta were mobile but did not
display the rapid long-distance streaming that has been ob-
served in fluorescent Golgi stacks in Nicotiana (Boevink et al.,
1998; Nebenfu ¨hr et al., 1999). This may reflect the different cell
Figure 3. Distribution of Seed Mucilage (CCRC-M36) or XG (a-XG) Epitopes in Developing Arabidopsis Seeds.
(A) to (C) Toluidine blue–stained sections show seed anatomy at 4, 7, and 9 DPA.
(D) to (F) Sections of 4-, 7-, and 9-DPA seeds fluorescently immunolabeled with CCRC-M36. CCRC-M36 labels only 7 (E) and 9 (F) DPA seed coat cells.
(G) to (I) Sections of 4-, 7-, and 9-DPA seeds immunolabeled with a-XG. a-XG labels primary cell walls of seed coats at all time points, in addition to the
label observed in mucilage at 7 (H) and 9 (I) DPA.
Bars ¼ 100 mm.
Dispersed Plant Golgi during Polarized Secretion 1627
a seed coat cell and the cortical cytoplasm of epidermal or
suspension-cultured cells. The number of Golgi stacks approx-
imately doubled from 4 DPA (24 6 9 Golgi stacks/cell, n ¼ 33
cells; Figure 8B) to the mucilage secretion phase at 7 DPA (49 6
14 Golgi stacks/cell, n ¼ 42 cells; Figure 8C). This was consid-
ered significant (Mann-Whitney U ¼ 65.0, P < 0.025).
Together, these two techniques provide complementary in-
formation: analysis by TEM allows Golgi stacks to be positively
identified, and the cytosol is sampled in square microns, provid-
ing a statistical sampling of the Golgi density. The fluorescence
data can directly report the number of ST-GFP positive puncta in
a whole cell. As a further comparison of two-dimensional data
obtained from sections viewed by TEM and whole-cell data
viewed by fluorescence microscopy, stereological techniques
were used to extrapolate TEM results to whole cells (Elias and
Hyde, 1983). Extrapolating on TEM results, there were ;25 and
56 Golgi stacks per cell in 4- and 7-DPA seed coat cells,
respectively, which are numbers consistent with what was found
using live-cell imaging.
Golgi Stacks Do Not Cluster Near the Site of Secretion
Since mucilage secretion is targeted to the outer tangential side
of the cell, the mobile Golgi stacks producing the secretory
product theoretically could cluster at the site of secretion,
eliminating much of the need for long-distance transport and
targeting of vesicles to the apical domain of the plasma mem-
brane. The distribution of Golgi stacks in the cell was compared
at sites of mucilage deposition (apical region of the cell; Figure
9A) and in the portions of the cell distal to the mucilage pockets
(basolateral domains; Figure 9A). Golgi stack distribution was
mapped, but there were no significant differences between the
densities of Golgi stacks in the apical domain versus the baso-
lateral domains at 7 DPA (compared using the Wilcoxon Signed-
Rankstestsforrepeated measures; z¼?1.975, P>0.05)(Figure
9B). Thus, the entire Golgi apparatus of seed coat cells is evenly
distributed around the cell. This was confirmed by live-cell
imaging with ST-GFP as well (Figure 8), where an even distribu-
tion of puncta in seed coat cells at 4 and 7 DPA was observed.
Although the entire population of mobile Golgi stacks does not
cluster near the site of mucilage secretion, there might be a
polarized distribution of mucilage-carrying Golgi in the apical
region. If this were true, it would suggest that there are distinct
populations of Golgi stacks (i.e., specialist Golgi stacks) that
reside near the apex and produce large amounts of pectin. The
distribution of the Golgi labeled with CCRC-M36 was examined
carrying specific products were found in the areas where muci-
lage was incorporated. When the cellular locations of Golgi
stacks containing the CCRC-M36 pectin-epitope or a-XG were
mapped using immunogold TEM, there were virtually no differ-
ences between the percentages of antibody gold-labeled Golgi
stacks in the apical versus basolateral regions of the cells (Figure
9C). This indicates that mucilage-secreting Golgi stacks are
found throughout the cell, not just in the apex, and that the
majority of the Golgi stacks in the cell are involved in the pro-
duction of mucilage at 7 DPA.
Decreased Mucilage (mum4) Mutants’ Golgi Density Still
Doubles during Mucilage Production, but Golgi Stacks
Lack Elaborate TGN
The above data demonstrate that changes in Golgi number
and morphology occur throughout the cell during mucilage
Figure 4. Details of Seed Mucilage (CCRC-M36) or XG (a-XG) Labeling in 7- and 9-DPA Seed Coats.
(A) and (B) Toluidine blue–stained sections of cryofixed Arabidopsis seed coats at 7 and 9 DPA.
(C) and (D) Seed coat cells at 7 and 9 DPA showing a strong CCRC-M36 immunofluorescent label in the mucilage pocket. At 7 DPA (C), punctate
structures in the cytosol also label with CCRC-M36 (arrowheads).
(E) and (F) a-XG labels both the mucilage pocket and the primary cell wall. In addition, a-XG labels the developing secondary cell wall of the columella
(arrows) at 9 DPA (F). Wrinkles in section lead to the appearance of uneven labeling in (E).
Bars ¼ 15 mm.
1628 The Plant Cell
production but do not determine whether the changes are in
response to increases in the bulky pectin product or due to the
intrinsic developmental program that seed coat cells undergo.
Plants with loss-of-function mutations in MUM4 have seed coat
epidermal cells with significantly reduced amounts of mucilage
what happens to Golgi structure and numbers when the surge of
pectin product is drastically reduced. As documented earlier
(Figures 1D, 2, and 10B), wild-type Golgi stacks that are pro-
ducing mucilage have short cisternae with compressed lumens
and swollen margins as well as a convoluted TGN. In compar-
ison, the Golgi cisternae of the stacks observed in mum4 cells at
Figure 5. Single Epitope Immunogold Labeling of Golgi Stacks.
(A) Immunogold labeling of wild-type, 7-DPA seed coat cells with the antimucilage antibody, CCRC-M36, using 10-nm gold particles (squares).
(B) Immunogold labeling of wild-type, 9-DPA seed coat cells with a-XG, using 10-nm gold particles (circles). c, cis face of Golgi stack; m, mucilage;
t, trans face of Golgi stack. Bars ¼ 250 nm.
(C) Percentage of Golgi stacks that immunolabel with 10-nm gold in thin sections of seed coat cells at 7 and 9 DPA. Sections were labeled with either
CCRC-M36 or a-XG. This analysis is based on three separate experiments, which includes a total of 31 cells and 287 Golgi stacks for 7-DPA cells
labeled with CCRC-M36, 28 cells and 252 Golgi stacks for 7-DPA cells labeled with a-XG, 23 cells and 287 Golgi stacks for 9-DPA cells labeled with
CCRC-M36, and 20 cells and 282 Golgi stacks for 9-DPA cells labeled with a-XG. Error bars represent range of percentages obtained in replicates.
Dispersed Plant Golgi during Polarized Secretion1629
7 DPA were longer, with more open lumens and thin, fenestrated
margins (Figure 10D). Strikingly, they had a less complex TGN,
lacking the elaborate clusters of vesicles/swollen networks seen
in 7-DPA wild-type cells. In this regard, the mum4 Golgi stacks
examined at 7 DPA had cisternae that were more similar in
appearance to those of wild-type cells at 4 DPA (cf. Figures 1B
with 10B and 10D).
The similarities and differences in Golgi morphology between
the wild type and mum4 at 7 DPA raised the question of whether
the increased density (number of Golgi stacks/mm2) of Golgi
stacks seen in wild-type cells between 4 and 7 DPA would also
be present in mum4. If the production of mucilage drives Golgi
proliferation, then comparison of 7-DPA wild-type and mum4
Figure 6. Double Immunogold Labeling of Wild-Type, 7-DPA Seed Coat
(A) Labeling density in seed coat mucilage within the mucilage pocket
(see inset for location with respect to the seed coat cell) using CCRC-
M36 (15-nm gold) and a-XG (10-nm gold, circles).
(B) Double immunogold labeling of Golgi stack using CCRC-M36 (15-nm
gold, squares) and a-XG (10-nm gold, circles).
c, cis face of Golgi stack; m, mucilage; t, trans face of Golgi stack. Bar ¼
Figure 7. Quantification of Golgi Stacks throughout Seed Coat Devel-
(A) Average number of Golgi stacks visible in sections of seed coat cells
at 4, 7, and 9 DPA.
(B) Average surface area of cytosol in thin sections of seed coat cells at
4, 7, and 9 DPA. Cytosol was defined as all areas within the cytoplasm,
excluding the nucleus, vacuole, and starch granules.
(C) Average density (number of Golgi stacks/mm2cytosol) in thin sections
of seed coat cells at 4, 7, and 9 DPA. For this analysis, three separate
experiments were conducted, resulting in a total of 61 cells and 152 Golgi
stacks at 4 DPA, 24 cells and 474 Golgi stacks at 7 DPA, and 17 cells and
333 Golgi at 9 DPA. Raw data were obtained by TEM. Error bars
represent 95% confidence intervals.
1630 The Plant Cell
seed coat cells during intense mucilage production would be
predicted to reveal a lower density of Golgi stacks in the mum4
cells. By contrast, if Golgi proliferation is an intrinsic component
of the development of the seed coat cells, then the mum4 plants
would be expected to display the same increases in Golgi
density as the wild type during mucilage production. High-
resolution electron microscopy was once again employed to
examine differences in Golgi density between wild-type and
mum4 seeds at 4 or 7 DPA (Figure 11). Despite the strong reduc-
tion in the amount of mucilage produced by these cells at 7 DPA,
the density of Golgi stacks in mum4 cells at 7 DPA was not
significantly different from the same developmental stage in the
wild-type seed coat (Figure 11C) (Mann-Whitney U ¼ 601.5, P >
0.05), indicating that even with a large reduction of secretory
product in the mutant, Golgi proliferation occurred between 4
and 7 DPA.
The large increases in polysaccharide secretion in Arabidopsis
seed coat epidermal cells are accompanied by specific changes
in Golgi morphology, especially in the TGN, and a dramatic
in the Golgi stack during mucilage secretion, as well as the
identification of the pectin epitope in the majority of the Golgi
stacks, imply that the individual stacks of the Golgi apparatus
work synchronously to produce mucilage.
Prolific Mucilage Production Leads to Morphological
Changes in the Golgi and TGN
In this study, high levels of mucilage secretion were correlated
with a specific Golgi morphology, which includes flattened
cisternae with swollen, fenestrated margins and a complex
flattened and the margins of the cisternae become increasingly
swollen. This is consistent with what has been documented in
other systems that are active in cell wall matrix production, such
as root hairs (Sherrier and Vandenbosch, 1994) and mucilage-
producing root cap cells (Craig and Staehelin, 1988; Staehelin
et al., 1990; Mollenhauer and Morre, 1991). This correlation is
further strengthened by the fact that a mutant unable to produce
a high level of seed coat mucilage lacks this characteristic Golgi
morphology. A hypothesis has been put forward that attempts to
explain Golgi morphology in the context of biosynthetic function.
Based on freeze-fracture data, it has been postulated that the
bulk of the biosynthetic enzymes are maintained within the
Figure 8. Fluorescence Microscopy Images of Wild-Type Seed Coat
Cells Carrying the Golgi Marker 35S:ST-GFP.
(A) Low-magnification epifluorescence image of seed coat on whole
seed at 7 DPA showing patchy expression of 35S:ST-GFP in the seed
coat. A single cell is outlined in white.
(B) and (C) Single image taken from z-stack of a seed coat cell showing
ST-GFP distribution at higher magnification at 4 and 7 DPA.
(B) Prior to mucilage production, at 4 DPA, there are fewer Golgi stacks
(C) Once mucilage production is underway, at 7 DPA, there are approx-
imately twice as many Golgi stacks in these cells.
Bars ¼ 10 mm.
Dispersed Plant Golgi during Polarized Secretion1631
flattened central portion of the Golgi stack, while the nascent
polysaccharide chains are pushed outwards, forming the swol-
len margins of the cisternae, and concentrating the polysac-
charide product for vesicle transport (Staehelin et al., 1990). If
this model is correct, the decreased amount of mucilage that is
being produced in mum4 would explain why the margins of
cisternae do not appear swollen and the TGN is decreased. It
would be interesting to examine if mum4 Golgi stacks show
similar results to the wild type in freeze fracture, which would
imply that the amount of biosynthetic enzymes embedded within
the cisternae are similar.
During mucilage production, RGI-rich pectin is produced in
large amounts by the Golgi and must be packaged at the trans-
face of the Golgi stack. Tomography and immunogold labeling
demonstrated that the Golgi exit site for pectins at the TGN
consists of interconnected vesicular clusters filled with pectin
product. That this exit site is a network, rather than free vesicles,
suggests that there must be subsequent fission and/or budding
steps to give rise to free secretory vesicles. The clustered
vesicles of the TGN appear to mature into dense secretory
vesicles, which move to the cortical cytoplasm, where they can
be observed among the abundant microtubules lining the muci-
lage secretion domain (McFarlane et al., 2008). This is unlike the
peripheral root cap cells’ mucilage secretion, where large bul-
bous margins of the Golgi stacks become large secretory ves-
icles after a process of clathrin-coated membrane retrieval
(Mollenhauer and Morre, 1991).
In this system, mutant analyses and antimucilage antibody
labeling data demonstrate the importance of the TGN in anter-
ograde transport of cell wall materials. The TGN appeared in
close association with Golgi stacks in this study, which was also
the case in electron tomographic examinations of entire Arabi-
dopsis meristematic cells at all stages of the cell cycle (Segui-
Simarro and Staehelin, 2006). In another polarized secretory
system, RabA4b has been reported to be a marker of a unique
post-TGN compartment that is localized in the secretory zone of
2008). With the emphasis on cargo that was used in our study, a
similar post-Golgi compartment could not be confirmed, but the
existence of a compartment that could mediate vesicle produc-
tion and/or membrane recycling is entirely possible. The TGN
itself has been proposed to be a site where multiple anterograde
(secretion and vacuolar protein sorting) and retrograde (endo-
cytic) pathways converge (reviewed in Lam et al., 2007).
Biological Control of Golgi Proliferation
The upregulation of pectin synthesis for the production of mu-
cilage in the seed coat epidermis occurs prior to 7 DPA and is
correlated with a dramatic increase in the number of Golgi
stacks.Inmum4,both the sizeand timingof the increase in Golgi
density remain unchanged, even though the production of mu-
cilage is dramatically reduced. Although polysaccharide secre-
tion required for the formation of the columella still occurs (at 9
The fact that the timing of the proliferation of Golgi stacks in the
mum4 mutant is identical to the wild type, despite a dramatic
reduction in the secretory demands on the mum4 cell, implies
Figure 9. Distribution of Golgi Stacks in Wild-Type, 7-DPA Seed Coat
(A) Schematic representation of how the apical and basolateral regions
of seed coat cells are defined.
(B) Comparison of the Golgi concentration in apical and basolateral
regions of 7-DPA seed coat cells. This analysis was based on three
separate experiments, which resulted in a total of 24 cells and 474 Golgi
stacks being examined. Error bars represent 95% confidence intervals.
Differences were not considered significant (P > 0.05).
(C) Comparison of regions of 7-DPA cells that label with immunogold
(anti-mucilage (CCRC-M36) or a-XG), indicating an even distribution
of epitopes in all regions of the cell. For this analysis, a total of 31 cells
and 287 Golgi stacks using CCRC-M36, and 28 cells and 252 Golgi
stacks using a-XG were examined, in three separate experiments.
Error bars represent the range of percentages obtained in different
1632The Plant Cell
that Golgi proliferation is not a direct consequence of poly-
saccharide production. Instead, Golgi proliferation may be in-
dependently activated by developmental signals during cell
differentiation. The nature of such developmental signals has
yet to be elucidated. The mechanism of Golgi doubling in these
cells is also not known, but, in plants, Golgi fission is considered
to be the mostly likely mechanism (Bosabalidis, 1985; Craig and
Staehelin, 1988; Hirose and Komamine, 1989; Langhans et al.,
2007). However, de novo production of Golgi stacks (Langhans
et al., 2007) could be involved as well.
Golgi Stacks Respond Synchronously and Collectively
Represent a Single Golgi Apparatus
Current views of cell wall architecture, based on antibody local-
laid down during development (Knox, 1997). In the past, this has
raised the question of what proportion of the hundreds of
streaming stacks in the plant Golgi apparatus contribute to a
particular epitope’s production and secretion. The detection of
monoclonal antibody, allowed us to test this question. Both the
uniform morphological changes of the Golgi stacks and the high
proportion of stacks that colocalize with CCRC-M36 label sup-
port the view that production of mucilage occurs synchronously
in the vast majority of stacks in the Golgi apparatus.
Dispersed Production of Matrix Polysaccharides by Golgi
Stacks for Polarized Secretion
During the development of the Arabidopsis seed coat, the
targeted secretion of matrix polysaccharides is not accompa-
nied by a concentration of Golgi stacks near the site of cargo
deposition. This indicates that the clustering of Golgi stacks is
not a mechanism for the proper targeting of cell wall carbohy-
drates to their destination in the developing mucilage pocket.
has been observed. In root hairs (Sherrier and Vandenbosch,
1994) and trichomes (Lu et al., 2005) and in cell plate formation
during cytokinesis (Nebenfu ¨hr et al., 2000; Segui-Simarro and
Staehelin, 2006), targeted secretion of cell wall products is
accompanied by a clustering of Golgi in the region of the cell
where deposition is high. In the case of cell plate formation, the
overall density of Golgi stacks in meristematic root cells remains
relatively constant despite local increases in the number of Golgi
Figure 10. Comparison of Wild-Type and mum4 Seed Coat Cell Morphology and Golgi Ultrastructure at 7 DPA.
(A) A wild-type, 7-DPA seed coat cell. Note the large mucilage pocket and central cytoplasmic column, with starch granules in amyloplasts above the
(B) Close-up of a 7-DPA Golgi stack.
(C) A mum4 seed coat cell at 7 DPA. Cells are flatter than wild-type cells, with very small mucilage pockets and virtually no central cytoplasmic column.
Vacuoles are approximately the same size in both wild-type and mum4 cells.
(D) Close-up of mum4 Golgi stack.
a, amyloplast; c, cis face of Golgi stack; m, mucilage; n, nucleus; t, trans face of Golgi stack; V, vacuole. Bars ¼ 5 mm in (A) and (C) and 250 nm in
(B) and (D).
Dispersed Plant Golgi during Polarized Secretion1633
stacks near the cell plate (Segui-Simarro and Staehelin, 2006).
The nonpolarized distribution of Golgi stacks seen here is more
similar to the arrangement in expanding interphase plant cells
(Boevink et al., 1998; Segui-Simarro and Staehelin, 2006), es-
pecially in mucilage-secreting root cap cells (Staehelin et al.,
1990). However, even in such cases of diffuse growth, there is
evidence of targeted secretion of macromolecules to specific
of the root (Roudier et al., 2005).
If the Golgi stacks are dispersed, yet specifically deposit
pectins and other matrix components at the mucilage pocket,
then post-Golgi mechanisms forvesicle targeting mustbeacting
to ensure that cell wall products are targeted to the appropriate
domain of the plasma membrane. Plant homologs of the molec-
ular machinery involved in this process in other eukaryotes have
been identified in silico (Ju ¨rgens and Geldner, 2002), but their
exact roles in plants have not been determined in many cases.
The major plasma membrane targeting complex in yeast and
animals is the exocyst (Guo et al., 2000), an eight-protein
complex that most likely forms a rod-like structure to facilitate
docking of the incoming vesicles at the target membrane
(Munson and Novick, 2006). All eight subunits have been iden-
tified in Arabidopsis, and there is some evidence that three of the
subunits, RTH1/SEC3 (Wen and Schnable, 1994; Wen et al.,
2005), SEC8 (Cole et al., 2005), and EXO70 (Elias et al., 2003;
Synek et al., 2006), play a role in targeting vesicles to the plasma
membrane. In mutant analyses of these three subunits, targeted
secretion appears to be disrupted, resulting in decreased ability
of certain cells, such as root hairs or pollen tubes, to elongate.
However, the phenotype of impaired secretion of seed coat
mucilage in these mutants, as well as in those of other exocyst
mutants, has not been determined.
XG in the Mucilage and the Columella
Our data demonstrate that both the mucilage and the columella
is usually associated with the network of primary cell wall
cellulose microfibrils and that cellulose has been found to be a
component of mucilage in the Arabidopsis seed coat (see
Supplemental Figure 1 online; Willats et al., 2001a; Macquet
et al., 2007), the presence of XG in the mucilage is not surprising.
On the other hand, the columella is considered to be a
secondary cell wall by definition, due to the fact that it is laid
down after the completion of cell expansion. XG is the hemicel-
lulose most commonly associated with type I primary cell walls
Figure 11. Quantification of Changes in Golgi Stacks in Wild-Type and
mum4 Seed Coat Cells at 4 and 7 DPA.
(A) Comparison of the average number of Golgi stacks visible in thin
sections of wild-type and mum4 seed coat cells.
(B) Average surface area of cytosol in thin sections wild-type and mum4
seed coat cells.
(C) Average density (number of Golgi stacks/mm2cytosol) in thin sections
of 7-DPA wild-type and mum4 seed coat cells. For this analysis, three
separate experiments were conducted. For the wild type, a total of 61
cells and 152 Golgi stacks at 4 DPA and 24 cells with 474 Golgi stacks at
7 DPA were examined. In the case of mum4, 46 cells and 115 Golgi
stacks were examined at 4 DPA and 57 cells with 800 Golgi stacks were
examined at 7 DPA. In all cases, data were acquired using TEM. Error
bars represent 95% confidence intervals.
1634 The Plant Cell
(Carpita and Gibeaut, 1993), whereas secondary cell walls in
other cell typesof Arabidopsis,suchas xylemand interfascicular
fibers, show reactivity with antixylan antibodies (such as LM10
and LM11) (McCartney et al., 2005; Persson et al., 2007). In
developing hybrid aspen (Populus tremula 3 P. tremuloides)
secondary xylem, fucosylated XGs were localized with CCRC-
M1 to the compound middle lamella region of the secondary cell
wall of fibers and are believed to link the primary and secondary
cell walls during secondary cell wall deposition (Bourquin et al.,
2002). Therefore, the presence of XG throughout the secondary
cell wall layer in the seed coat is intriguing.
Tracking Polysaccharide Production with
We have shown that the CCRC-M36 antibody reacts only with
mucilage in the developing seed, providing an invaluable tool for
the analysis of seed coat mucilage secretion. Although there are
other antibodies that label seed coat mucilage, including a-XG
(reported here) and PGA/RGI (Western et al., 2004), CCRC-M36
has the advantage that it is specific to mucilage, rather than
primary cell wall, in the Arabidopsis seed coat. This fact has
allowed us to selectively track cargo that is being targeted to a
specific cell wall domain. Thus far, we have been able to show
that Golgi stacks carrying cargo destined for polar secretion do
not cluster near the target site but are instead evenly distributed
throughout the cell. Future experiments made possible with this
cargo destined for the apoplast are packaged and sorted by the
Plant Materials and Growth Conditions
Wild-type Arabidopsis thaliana was Columbia-2 ecotype (Lehle Seeds).
The mutant line mum4-1 (here referred to as mum4) was isolated in
Westernet al.(2001).Seedsweregrownonprepared soilmix(Sunshine 5
Professional Growing Mix; Sungro Horticulture Canada) that had been
fertilized with AT media (Haughn and Somerville, 1986) in growth cham-
bers at 218C under continuous light (90 to 120 mE m?2s?1PAR). Flowers
were initially staged as by Western et al. (2000). In addition, stages were
defined by the morphology of seed coat cells upon examination at the
light microscopy level (see below).
High-Pressure Freezing and Freeze Subsititution
Seeds 4, 7, and 9 DPA (Western et al., 2004) were excised from siliques,
stabbed with an insect pin (in the case of 7- and 9-DPA seeds), and
prepared for TEM by high-pressure freezing, freeze substitution, and
loaded into copper hats (Ted Pella) filled with 1-hexadecene and high-
pressure frozen using a Bal-Tec HPM 010 high-pressure freezer (Balzers
Instruments). The hats were immediately transferred to frozen cryovials
containing freeze substitution medium consisting of either 2% (w/v)
osmium tetroxide in acetone with 8% (v/v) dimethoxypropane for mor-
phological assays or 0.25% (v/v) glutaraldehyde and 2% (w/v) uranyl
acetate in acetone with 8% (v/v) dimethoxypropane for immunolabeling
assays. Freeze substitution was performed for 4 to 6 d at ?808C by
incubation in an acetone/dry ice slush, followed by transfer of the
cryovials into a metal block precooled to ?208C, which was warmed
over 20 h to allow reaction of fixatives. The samples were removed from
the sample holders, rinsed in anhydrous acetone several times, and
slowly infiltrated and embedded in either Spurr’s resin (Spurr, 1969) for
morphological examination or LR White (London Resin Company) for
Thick (0.5 mm) sections, stained with toluidine blue, were examined by
light microscopy for intact seed coat cells at each stage of development
and to ensure that embedded seeds had seed coats at the appropriate
developmental stage, based on morphological criteria described previ-
ously (Western et al., 2000). Samples containing appropriately staged
seeds were thin sectioned using a Leica Ultracut UCT (Leica Micro-
systems) and mounted on 150 mesh Formvar-coated grids. Sections
were poststained with 2% (w/v) aqueous uranyl acetate in 70% (v/v)
examined using a Hitachi H-7600 PC-TEM microscope.
To calculate the density of Golgi in thin sections (number of Golgi
stacks/mm2cytosol), the number of Golgi stacks visible in a given TEM
section of a cell was counted, and images of the counted cells were
collected. The cytosol was defined as all areas within the cytoplasm,
except the nucleus, amyloplasts, and vacuole. The surface area of the
cytosol was calculated using ImageJ (National Institutes of Health; http://
rsb.info.nih.gov/ij/). The number of cells measured was chosen so that
there were approximately equivalent amounts of cytosolic surface area
wild-type cells and 152 Golgi stacks at 4 DPA, 24 cells and 474 Golgi
stacks at 7 DPA, and 17 cells and 333 Golgi at 9 DPA. In the case of
with 800 Golgi stacks at 7 DPA. Extrapolation of stack density in
sectioned material to stack density in whole cells was done according
to Elias and Hyde (1983).
Live-Cell Imaging of Golgi Stacks
Seeds from transgenic plants bearing the trans-Golgi marker ST-GFP,
driven by the 35S promoter (35S:ST-GFP) (Boevink et al., 1998) were
staged, removed from siliques, and mounted in SlowFade Antifade
reagent (Invitrogen). Samples were visualized by obtaining z-stacks of
entire cells using a Zeiss LSM510 Meta confocal microscope (Carl Zeiss
Canada) or by epifluorescence of whole cells using a Leica DM6000B
microscope with fluorescent source (Leica Microsystems) at 363 mag-
nification immediately after the seeds were removed from siliques.
ImageJ software was used to identify and count Golgi-sized punctate
structures in individual seed coat cells. Thirty-three 4-DPA cells were
counted, with a total of 776 Golgi stacks, whereas 42 cells were counted
at 7 DPA, yielding 2070 Golgi stacks.
Previously published antibodies that were used include JIM5 and JIM7
against methyl-esterified homogalacturonans (Knox et al., 1990), a-RGI/
PGA against rhamnogalacturonan/polygalacturonic acids, a-XG against
XG (Moore et al., 1986), PAM1 against unesterified homogalacturonans
(Willats et al., 1999), LM5 against anti-(1-4)-b-D-galactan (Jones et al.,
1997), LM7 against partially methylesterified homogalacturonan (Willats
et al., 2001b), LM10 against xylans (McCartney et al., 2005), CBD-OG
(McLean et al., 2000), CCRC-M1 against fucosylated XG, and CCRC-M2
and CCRC-M7 against RGI (Puhlmann et al., 1994). Another set of anti-
bodies used (CCRC-M30, CCRC-M31, CCRC-M32, CCRC-M34, CCRC-
M36, and CCRC-M38) was generated against extracted Arabidopsis
Dispersed Plant Golgi during Polarized Secretion1635
seed mucilage (T. Bootten, Z. Popper, C. Deng, R. Jia, W.S. York, M.A.
O’Neill, and M.G. Hahn, unpublished data), and these antibodies are
available from CarboSource (http://cell.ccrc.uga.edu/;carbosource/
CSS_home.html). A further set of antibodies (CCRC-M48, CCRC-M54,
CCRC-M57, and CCRC-M58) was generated against tamarind seed XG
(Z. Popper, T. Bootten, S. Tuomivaara, A.G. Swennes, R. Jia, W.S. York,
and M.G. Hahn, unpublished data), and these antibodies are also
available from CarboSource.
For immunoflourescence, 0.3- to 0.5-mm LR White sections were
mounted on 10-well, Teflon-coated microscope slides. Nonspecific pro-
tein binding was blocked by incubating slides in Coplin jars filled with of
5% (w/v) nonfat dry milk (NFDM) in Tris-buffered saline/0.1% (v/v) Tween
20 (TBST) for 20 min. Samples were then incubated at room temperature
for 1 h with primary antibodies at 1:20 (v/v) dilutions in 1% (w/v) NFDM in
TBST and for 1 h in secondary antibodies at 1:100 (v/v) dilution. Control
experiments were performed to test specific binding of primary anti-
bodies in CCRC-M36 and a-XG label experiments by preincubating the
antibodies with their specific purified antigen to block antibody binding
sites. Antigen for CCRC-M36 was mucilage that had been extracted from
wild-type seeds by vortexing whole seeds, and the antigen for anti-XG
was 2 mg/mL of tamarind XG (Megazyme International Ireland). In both
cases, preincubation of primary antibody with excess antigen completely
blocked binding to the sections (see Supplemental Figure 2 online).
Rinses were performed before and after incubations in Coplin jars with
TBST. Samples were mounted in 90% (v/v) glycerol in water and exam-
ined via epifluorescence using a Leica DMR light microscope (Leica
Microsystems). For whole seed immunolabeling (see Supplemental Fig-
ure 1 online), we followed the same protocol, except that incubations
were done in small batches (10 to 50) of whole seeds, in 1.5-mL
microcentrifuge tubes with gentle agitation on an orbital shaker.
LR White samples were cut into 70-nm sections and mounted on 200
mesh, fine bar, nickel grids (Ted Pella). Nonspecific protein binding was
blocked with 5% (w/v) NFDM in TBST for 20 min. Excess solution was
blotted off, and grids were transferred to a drop of primary antibody,
diluted 1:5 (CCRC-M36) or 1:20 (a-XG) with 1% (w/v) NFDM in TBST, and
incubated for 1 h at room temperature. After washing grids in three
subsequent washes of TBST for 15 s each, grids were transferred to
secondary antibodies and diluted 1:100 with 1% (v/v) NFDM in TBST for
1 h. Secondary antibodies were goat anti-mouse IgG þ IgM (for CCRC-
M36) or goat anti-rabbit IgG (for a-XG), conjugated to 10-nm colloidal
gold (Ted Pella). Grids were washed again, followed by three washes of
distilled water for 15 s each, and then poststained with 2% (w/v) uranyl
acetatefor8minandReynold’slead citratefor2min.Forthedouble label
mouse IgG þ IgM was used to locate CCRC-M36. Primary or second-
ary antibodies were mixed together and treated as described above.
287 Golgi stacks for 7-DPA cells labeled with CCRC-M36, 28 cells and
stacks for 9-DPA cells labeled with CCRC-M36, and 20 cells and 282
Golgi stacks for 9-DPA cells labeled with a-XG. For the double label, 57
Golgi stacks were examined in five different cells.
Data analysis was conducted with SPSS 13 software. Since the data
failed the Kolgomorov-Smirnov test for normalcy, nonparametric tests
wereused.Forcomparison ofthreeindependent samples(i.e.,4,7,and9
DPA), significance was analyzed using the Kruskal-Wallis test (H) to
hoc test, taking into account Bonferroni’s correction for significance
levels. When there were two independent samples, a Mann-Whitney test
(U) was performed. For repeated measures on the same sample (e.g.,
apical versus basolateral calculations), a Wilcoxon signed ranks test
(reported as a z-score) was used. In all cases, results were deemed
significant if P < 0.05.
Electron Microscopy Tomography
Tomographic analysis of 7-DPA seed coat cells was done according to
Donohoe et al. (2006), with some minor changes. Briefly, samples were
cryofixed and embedded as described above. Spurr’s sections (200 nm
thick) were picked up on 0.75% (w/v) Formvar-coated copper/rhodium
slot grids and stained with 2% (w/v) uranyl acetate in 70% (v/v) methanol
for 20 to 25 min, followed by Reynold’s lead citrate for 6 to 8 min. After
staining, 15-nm unconjugated colloidal gold particles wereadded to both
sides of the grid to be used as fiducial markers to align the series of tilted
Tomograms were acquired on an FEI Technai TF30 intermediate
voltage electron microscope, operating at 300 kV. Tilt series were
acquired at 23,0003, from ?608 to þ608 at 18 intervals about two
camera,giving acalculated pixel sizeof1nm.Dual-axistomograms were
generated using the Etomo software interface, a part of the IMOD
software package (Kremer et al., 1996). Tomograms were displayed
andanalyzedwith3dmod, thegraphics componentoftheIMODsoftware
package. Golgi stacks, vesicles, and other structures were modeled
manually according to Donohoe et al. (2006).
The following materials are available in the online version of this article.
Supplemental Figure 1. Antibody Screen of Mucilage from Hydrated
Supplemental Figure 2. Fluorescent Labeling Antibody Controls.
We thank Andrew Staehelin (University of Colorado, Boulder, CO) for
constructive criticism of the manuscript and Zach Gergely (University of
Colorado) and the staff of the Boulder Lab for three-dimensional
electron microscopy of cells for invaluable help with tomography. The
assistance of the staff of the University of British Columbia BioImaging
Facility (Vancouver, Canada) is gratefully acknowledged. We also thank
David Bird (University of Winnipeg, Winnipeg, Canada) for insightful
discussion on statistics. CBD-OG was a generous gift from Douglas
Kilburn (Michael Smith Laboratories, University of British Columbia,
Vancouver, Canada). This work was supported by the National Science
and Engineering Research Council of Canada Discovery Grants to
G.W.H., T.L.W., and A.L.S. Generation and characterization of the
monoclonal antibodies against plant cell wall polysaccharides was
supported by a grant (DCB-0421683) from the U.S. National Science
Foundation Plant Genome Program (to M.G.H.).
1636 The Plant Cell
Received February 15, 2008; revised April 5, 2008; accepted May 16,
2008; published June 3, 2008.
Beeckman, T., De Rycke, R., Viane, R., and Inze, D. (2000). Histo-
logical study of seed coat development in Arabidopsis thaliana. J.
Plant Res. 113: 139–148.
Boevink, P., Oparka, K., Santa Cruz, S., Martin, B., Betteridge, A.,
and Hawes, C. (1998). Stacks on tracks: The plant Golgi apparatus
traffics on an actin/ER network. Plant J. 15: 441–447.
Bosabalidis, A.M. (1985). A possible new mode of dictyosome dupli-
cation in plant cells. Tissue Cell 17: 279–286.
Bourquin, V., Nishikubo, N., Abe, H., Brumer, H., Denman, S.,
Eklund, M., Christiernin, M., Teeri, T.T., Sundberg, B., and
Mellerowicz, E.J. (2002). Xyloglucan endotransglycosylases have a
function during the formation of secondary cell walls of vascular
tissues. Plant Cell 14: 3073–3088.
Carpita, N.C., and Gibeaut, D.M. (1993). Structural models of primary
cell walls in flowering plants: Consistency of molecular structure with
the physical properties of the walls during growth. Plant J. 3: 1–30.
Carpita, N.C., and McCann, M.C. (2000). The Cell Wall. In Biochemistry
and Molecular Biology of Plants, B.B. Buchanan, W. Gruissem, and
R.L. Jones, eds (Rockville, MD: American Society of Plant Physiolo-
gists), pp. 52–109.
Cole, R.A., Synek, L., Zarsky, V., and Fowler, J.E. (2005). SEC8, a
subunit of the putative Arabidopsis exocyst complex, facilitates pollen
germination and competitive pollen tube growth. Plant Physiol. 138:
Cosgrove, D.J. (2005). Growth of the plant cell wall. Nat. Rev. Mol. Cell
Biol. 6: 850–861.
Craig, S., and Staehelin, L.A. (1988). High pressure freezing of intact
plant tissues evaluation and characterization of novel features of the
endoplasmic reticulum and associated membrane systems. Eur. J.
Cell Biol. 46: 80–93.
Doblin, M.S., Vergara, C.E., Read, S., Newbigin, E., and Bacic, A.
(2003). The plant cell wall: Making the bricks. In The Plant Cell Wall,
J.K.C. Rose, ed (Oxford, UK: Blackwell Publishing), pp. 183–222.
Donohoe, B.S., Mogelsvang, S., and Staehelin, L.A. (2006). Electron
tomography of ER, Golgi and related membrane systems. Methods
Elias, H., and Hyde, D.M. (1983). A Guide to Practical Stereology.
(Basel, Switzerland: S. Karger AG).
Elias, M., Drdova, E., Ziak, D., Bavlnka, B., Hala, M., Cvrckova, F.,
Soukupova, H., and Zarsky, V. (2003). The exocyst complex in
plants. Cell Biol. Int. 27: 199–201.
Griffing, L.R. (1991). Comparisons of Golgi structure and dynamics in
plant and animal cells. J. Electron Microsc. Tech. 17: 179–199.
Guo, W., Sacher, M., Barrowman, J., Ferro-Novick, S., and Novick,
P. (2000). Protein complexes in transport vesicle targeting. Trends
Cell Biol. 10: 251–255.
Haughn, G.W., and Somerville, C. (1986). Sulfonylurea-resistant mu-
tants of Arabidopsis thaliana. Mol. Gen. Genet. 204: 430–434.
Hirose, S., and Komamine, A. (1989). Changes in ultrastructure of
Golgi apparatus during the cell cycle in a synchronous culture of
Catharanthus roseus. New Phytol. 111: 599–605.
Jarvis, M.C., and McCann, M.C. (2000). Macromolecular biophysics of
the plant cell wall: Concepts and methodology. Plant Physiol. Bio-
chem. 38: 1–13.
Jones, L., Seymour, G.B., and Knox, J.P. (1997). Localization of pectic
galactan in tomato cell walls using a monoclonal antibody specific to
(1[->]4)-[beta]-D-galactan. Plant Physiol. 113: 1405–1412.
Ju ¨rgens, G. (2005). Cytokinesis in higher plants. Annu. Rev. Plant Biol.
Ju ¨rgens, G., and Geldner, N. (2002). Protein secretion in plants: From
the trans-Golgi network to the outer space. Traffic 3: 605–613.
Knox, J.P. (1997). The use of antibodies to study the architecture and
developmental regulation of plant cell walls. Int. Rev. Cytol. 171: 79–120.
Knox, J.P., Linstead, P.J., King, J., Cooper, C., and Roberts, K.
(1990). Pectin esterification is spatially regulated both within cell walls
and between developing tissues of root apices. Planta 181: 512–521.
Kremer, J.R., Mastronarde, D.N., and McIntosh, J.R. (1996). Com-
puter visualization of three-dimensional image data using IMOD.
J. Struct. Biol. 116: 71–76.
Lam, S.K., Tse, Y.C., Robinson, D.G., and Jiang, L. (2007). Tracking
down the elusive early endosome. Trends Plant Sci. 12: 497–505.
Langhans, M., Hawes, C., Hillmer, S., Hummel, E., and Robinson,
D.G. (2007). Golgi regeneration after brefeldin A treatment in BY-2
cells entails stack enlargement and cisternal growth followed by
division. Plant Physiol. 145: 527–538.
Lu, L., Lee, Y.R., Pan, R., Maloof, J.N., and Liu, B. (2005). An internal
motor kinesin is associated with the Golgi apparatus and plays a role
in trichome morphogenesis in Arabidopsis. Mol. Biol. Cell 16: 811–823.
Lynch, M.A., and Staehelin, L.A. (1992). Domain-specific and cell type-
specific localization of two types of cell wall matrix polysaccharides in
the clover root tip. J. Cell Biol. 118: 467–479.
Macquet, A., Ralet, M.-C., Kronenberger, J., Marion-Poll, A., and
North, H.M. (2007). In situ, chemical and macromolecular study of the
composition of Arabidopsis thaliana seed coat mucilage. Plant Cell
Physiol. 48: 984–999.
ment methods that preserve resolution. J. Struct. Biol. 120: 343–352.
McCartney, L., Marcus, S.E., and Knox, J.P. (2005). Monoclonal
antibodies to plant cell wall xylans and arabinoxylans. J. Histochem.
Cytochem. 53: 543–546.
McFarlane, H.E., Young, R.E., Wasteneys, G.O., and Samuels, A.L.
(2008). Cortical microtubules mark the mucilage secretion domain
of the plasma membrane in Arabidopsis seed coat cells. Planta 227:
McLean, B.W., Bray, M.R., Boraston, A.B., Gilkes, N.R., Haynes,
C.A., and Kilburn, D.G. (2000). Analysis of binding of the family 2a
carbohydrate-binding module from Cellulomonas fimi xylanase 10A to
cellulose: Specificity and identification of functionally important amino
acid residues. Protein Eng. 13: 801–809.
Mollenhauer, H.H., and Morre, D.J. (1991). Perspectives on Golgi
apparatus form and function. J. Electron Microsc. Tech. 17: 2–14.
Moore, P.J., Darvill, A.G., Albersheim, P., and Staehelin, L.A. (1986).
Immunogold localization of xyloglucan and rhamnogalacturonan I in
the cell walls of suspension-cultured sycamore cells. Plant Physiol.
Moore, P.J., and Staehelin, L.A. (1988). Immunogold localization of the
cell-wall-matrix polysaccharides rhamnogalacturonan I and xyloglu-
can during cell expansion and cytokinesis in Trifolium pratense L.:
Implication for secretory pathways. Planta 174: 433–445.
Moore, P.J., Swords, K.M.M., Lynch, M.A., and Staehelin, L.A.
(1991). Spatial organization of the assembly pathways of glycopro-
teins and complex polysaccharides in the Golgi apparatus of plants. J.
Cell Biol. 112: 589–602.
Munson, M., and Novick, P. (2006). The exocyst defrocked, a frame-
work of rods revealed. Nat. Struct. Mol. Biol. 13: 577–581.
Nebenfu ¨hr, A., Frohlick, J.A., and Staehelin, L.A. (2000). Redistribu-
tion of Golgi stacks and other organelles during mitosis and cytoki-
nesis in plant cells. Plant Physiol. 124: 135–151.
Nebenfu ¨hr, A., Gallagher, L.A., Dunahay, T.G., Frohlick, J.A.,
Mazurkiewicz, A.M., Meehl, J.B., and Staehelin, L.A. (1999).
Dispersed Plant Golgi during Polarized Secretion1637
Stop-and-go movements of plant Golgi stacks are mediated by the Download full-text
acto-myosin system. Plant Physiol. 121: 1127–1142.
Nebenfu ¨hr, A., and Staehelin, L.A. (2001). Mobile factories: Golgi
dynamics in plant cells. Trends Plant Sci. 6: 160–167.
Northcote, D.H., and Pickett-Heaps, J.D. (1966). A function of the
Golgi apparatus in polysaccharide synthesis and transport in the root-
cap cells of wheat. Biochem. J. 98: 159–167.
Odell, J.T., Nagy, F., and Chua, N.H. (1985). Identification of DNA
sequences required for activity of the cauliflower mosaic virus 35S
promoter. Nature 313: 810–812.
Persson, S., Caffall, K.H., Freshour, G., Hilley, M.T., Bauer, S.,
Poindexter, P., Hahn, M.G., Mohnen, D., and Somerville, C.
(2007). The Arabidopsis irregular xylem8 mutant is deficient in
glucuronoxylan and homogalacturonan, which are essential for sec-
ondary cell wall integrity. Plant Cell 19: 237–255.
Pesacreta, T.C., and Lucas, W.J. (1985). Presence of a partially-
coated reticulum in angiosperms. Protoplasma 125: 173–184.
Preuss, M.L., Schmitz, A.J., Thole, J.M., Bonner, H.K., Otegui, M.S.,
and Nielsen, E. (2006). A role for the RabA4b effector protein
PI-4Kbeta1 in polarized expansion of root hair cells in Arabidopsis
thaliana. J. Cell Biol. 172: 991–998.
Preuss, M.L., Serna, J., Falbel, T.G., Bednarek, S.Y., and Nielsen, E.
(2004). The Arabidopsis Rab GTPase RabA4b localizes to the tips of
growing root hair cells. Plant Cell 16: 1589–1603.
Puhlmann, J., Bucheli, E., Swain, M.J., Dunning, N., Albersheim, P.,
Darvill, A.G., and Hahn, M.G. (1994). Generation of monoclonal
antibodies against plant cell-wall polysaccharides. Plant Physiol. 104:
Rensing, K.H., Samuels, A.L., and Savidge, R.A. (2002). Ultrastructure
of vascular cambial cell cytokinesis in pine seedlings preserved by
cryofixation and substitution. Protoplasma 220: 39–49.
Roudier, F., Fernandez, A.G., Fujita, M., Himmelspach, R., Borner,
G.H.H., Schindelman, G., Song, S., Baskin, T.I., Dupree, P.,
Wasteneys, G.O., and Benfey, P.N. (2005). COBRA, an Arabidopsis
extracellular glycosylphosphatidylinositol-anchored protein, specifi-
cally controls highly anisotropic expansion through its involvement in
cellulose microfibril orientation. Plant Cell 17: 1749–1763.
Segui-Simarro, J.M., and Staehelin, L.A. (2006). Cell cycle-dependent
changes in Golgi stacks, vacuoles, clathrin-coated vesicles and
multivesicular bodies in meristematic cells of Arabidopsis thaliana: A
quantitative and spatial analysis. Planta 223: 223–236.
Sherrier, D.J., and Vandenbosch, K.A. (1994). Secretion of cell wall
polysaccharides in Vicia root hairs. Plant J. 5: 185–195.
Somerville, C., Bauer, S., Brininstool, G., Facette, M., Hamann, T.,
Milne, J., Osborne, E., Paredez, A., Persson, S., Raab, T., Vorwerk,
S., and Youngs, H. (2004). Toward a systems approach to under-
standing plant cell walls. Science 306: 2206–2211.
Spurr, A.R. (1969). A low-viscosity epoxy resin embedding medium for
electron microscopy. J. Ultrastruct. Res. 26: 31–43.
Staehelin, L.A., Giddings, T.H., Kiss, J.Z., and Sack, F.D. (1990).
Macromolecular differentiation of Golgi stacks in root tips of Arabi-
dopsis and Nicotiana seedlings as visualized in high pressure frozen
and freeze-substituted samples. Protoplasma 157: 75–91.
Synek, L., Schlager, N., Elias, M., Quentin, M., Hauser, M.T., and
Zarsky, V. (2006). AtEXO70A1, a member of a family of putative
exocyst subunits specifically expanded in land plants, is important for
polar growth and plant development. Plant J. 48: 54–72.
Thole, J.M., Vermeer, J.E., Zhang, Y., Gadella, T.W., Jr., and Nielsen,
E. (2008). ROOT HAIR DEFECTIVE4 encodes a phosphatidylinositol-
4-phosphate phosphatase required for proper root hair development
in Arabidopsis thaliana. Plant Cell 20: 381–395.
Usadel, B., Kuschinsky, A.M., Rosso, M.G., Eckermann, N., and
Pauly, M. (2004). RHM2 is involved in mucilage pectin synthesis and
is required for the development of the seed coat in Arabidopsis. Plant
Physiol. 134: 286–295.
Wen, T.J., Hochholdinger, F., Sauer, M., Bruce, W., and Schnable,
P.S. (2005). The roothairless1 gene of maize encodes a homolog
of sec3, which is involved in polar exocytosis. Plant Physiol. 138:
Wen, T.-J., and Schnable, P.S. (1994). Analyses of mutants of three
genes that influence root hair development in Zea mays (Gramineae)
suggest that root hairs are dispensable. Am. J. Bot. 81: 833–842.
Western, T.L., Burn, J., Tan, W.L., Skinner, D.J., Martin-McCaffrey,
L., Moffatt, B.A., and Haughn, G.W. (2001). Isolation and charac-
terization of mutants defective in seed coat mucilage secretory cell
development in Arabidopsis. Plant Physiol. 127: 998–1011.
Western, T.L., Skinner, D.J., and Haughn, G.W. (2000). Differentiation
of mucilage secretory cells of the Arabidopsis seed coat. Plant
Physiol. 122: 345–356.
Western, T.L., Young, D.S., Dean, G.H., Tan, W.L., Samuels,
A.L., and Haughn, G.W. (2004). MUCILAGE-MODIFIED4 encodes a
putative pectin biosynthetic enzyme developmentally regulated by
APETALA2, TRANSPARENT TESTA GLABRA1, and GLABRA2 in
the Arabidopsis seed coat. Plant Physiol. 134: 296–306.
Willats, W.G., Gilmartin, P.M., Mikkelsen, J.D., and Knox, J.P. (1999).
Cell wall antibodies without immunization: Generation and use of de-
esterified homogalacturonan block-specific antibodies from a naive
phage display library. Plant J. 18: 57–65.
Willats, W.G., McCartney, L., and Knox, J.P. (2001a). In-situ analysis
of pectic polysaccharides in seed mucilage and at the root surface of
Arabidopsis thaliana. Planta 213: 37–44.
Willats, W.G., Orfila, C., Limberg, G., Buchholt, H.C., van Alebeek,
G.J., Voragen, A.G., Marcus, S.E., Christensen, T.M., Mikkelsen,
J.D., Murray, B.S., and Knox, J.P. (2001b). Modulation of the degree
and pattern of methyl-esterification of pectic homogalacturonan in
plant cell walls. Implications for pectin methyl esterase action, matrix
properties, and cell adhesion. J. Biol. Chem. 276: 19404–19413.
Windsor, J.B., Symonds, V.V., Mendenhall, J., and Lloyd, A.M.
(2000). Arabidopsis seed coat development: Morphological differen-
tiation of the outer integument. Plant J. 22: 483–493.
Zhang, G.F., and Staehelin, L.A. (1992). Functional compartementation
of the Golgi apparatus of plant cells. Plant Physiol. 99: 1070–1083.
1638 The Plant Cell