Download Root organization and gene expression patterns

Journal of Experimental Botany, Vol. 47, No. 304, pp. 1613-1628, November 1996
Journal of
Experimental
Botany
REVIEW ARTICLE
Root organization and gene expression patterns
Thomas L. Rost1'3 and John A. Bryant2
1
2
Section of Plant Biology, Division of Biological Sciences, University of California, Davis, CA 95616-8537, USA
Department of Biological Sciences, Washington Singer Laboratory, University of Exeter, Exeter EX4 4QG, UK
Received 18 April 1996; Accepted 1 July 1996
somehow unique to the cylinder, or that the nearby cells
are insensitive to its effect or are structurally isolated.
New tools of microscopy, molecular biology, and
In pea roots, the vascular cylinder can be subdivided
genetics are making it possible for biologists to study
into xylem and phloem sectors (Rost et al., 1988). Xylem
roots with new vigour. Such investigations have
pericycle cells continue to divide in a transverse plane
enabled plant biologists to notice the symmetry,
some distance from the root tip at the three protoxylem
pattern, and simplicity of root structures so that there
points. Since the surrounding cells in the cortex and the
is now an exciting rebirth in the study of roots. The
adjacent phloem sectors are not dividing, this means that
literature of root biology, development and structure
specific signals are perhaps regulating the expression of
is vast. In this short review we will concentrate on
cell cycle genes only in xylem sector cells.
describing our notion of how roots are organized
Evidence of sectors can be observed at the molecular
structurally, and then discuss what is known about
level. When cross-sections of pea roots, about 10 mm
tissue- and zone-specific gene expression in roots.
from the root tip, were probed with labelled mRNA for
histone H2A, the labelled cells occurred only in the three
Key words: Root apex, root anatomy, root development,
xylem sectors, and only in the pericycle cells (Tanimoto
gene expression.
et al., 1993a, b; Tanimoto, 1994). Bruntfield (1943)
observed clonal sectors by plotting chromosome aberrations in Crepis and Vicia roots after treating them with
Root organization
X-rays. The interesting observation here is that the predominant sector type observed included all of the tissues
Levels of organization
of the root. This suggested to him that 3-4 initial cells
Traditionally root structure has been considered in terms
were present in the apex from which all cells in each
of the root apical meristem, the regions of elongation,
sector were derived. This is an interesting idea, but it is
maturation, and secondary growth. Recent analysis connot supported by what is known about apical organization
siders the root body and cap in an ordered way (Rost,
and the activity of root initials at least in mature roots,
with the exception of ferns with apical cells. Another
1994), where the root is composed of cylinders, sectors,
example of observed sector patterns in roots comes from
packets (merophytes) and cell files. In this section these
Taylor and Rogers (1992). They transformed tomato
developmental units are briefly described.
(a) Cylinders and sectors: Aggregates of cellfiles(tissues plants with a transposon linked to a GUS (j3-glucuronidase) expression cassette. Thus, expression of GUS (visualand regions) are organized into cylinders of cells or radial
ized by the familiar blue-staining technique) revealed
sectors (Rost, 1993, 1994) (Fig. 1). For example, the
insertion of the transposon into particular cells or cell
middle-cortex cylinder in pea roots is composed of an
lines. One of the predominant patterns they observed was
aggregate of similar parenchyma cells (Rost et al., 1988).
the occurrence of sectors in which a quarter to a third of
These cells stop dividing and differentiate together as a
the resulting new tissues had GUS activity. The size of
unit. The signals inducing this behaviour must affect all
the
sectors was related to the time when the transposition
cells within the unit and presumably would readily move
event occurred (i.e. earlier insertion led to larger
from cell to cell within it, to the exclusion of other nearby
blue-stained sectors).
cells. This would imply that the hypothetical signal is
Abstract
'To whom correspondence should be addressed: Fax: +1 916 752 5410. E-mail: Urosteucdavi8.edu
O Oxford University Press 1996
1614
Rost and Bryant
phloem sector
cell file
xylem sector
sector
phloem sector
xytem sector
cylinder
vascular cylinder
cortical cylinder
not unlike the merophyte, but may be tissue-specific and
possibly lacks the precise division sequences found in fern
merophytes.
(c) Cell files: Cells within cell files are strongly connected to each other by plasmodesmatal connections
mostly within their transverse walls (Juniper and Barlow,
1969; Gunning, 1978; Zhu et a/., 1996). Since a main
conduit for passage of cytoplasmic signals is through the
plasmodesmata (Gunning and Robards, 1976; Gunning
and Overall, 1983; Lucas and Wolf, 1993), the cell file
becomes a key developmental unit.
Full understanding of root structure and development
must consider all levels of organization. Most likely not
one or even two interacting signals, morphogens, physical
factors, or cues will explain root development. Instead,
since the root is a composite of individual cells, cell
packets, cell files, and tissue sectors and cylinders, a
complicated network of interacting signals must be operating in root development (Rost, 1993, 1994). Before the
signals involved can be analysed it is necessary that the
developmental organization of the root apical meristem
be completely and correctly understood.
The concept of transition points
epidermis cylinder
Fig. 1. Arabidopsis thaliana root apical menstem of a two-week-old
root The three initial tiers and the tissues which, they derive are
clearly shown
(b) Merophytes and packets: Azolla fern root tips have
a large apical cell that divides acropetally to form root
cap initials and sequentially from its three basal surfaces
to make a sequence of merophytes (Gunning, 1982). The
apical cell divides a preprogrammed number of times to
produce these derivative cells in a spiral series. Each of
these in turn divides a predetermined number of times
within the expanding confines of the original derivative
cell wall. This developmental unit, or merophyte, divides,
differentiates and enlarges in a set pattern resulting in
mature cells and tissues. During subsequent development,
cell files differentiate in continuity between merophytes,
which would necessitate communication between them.
A merophyte-system has been reported in other lower
vascular plants (Gifford, 1993), but not in the roots of
higher vascular plants.
Barlow (1987) has described a somewhat analogous
structural unit in corn roots. He found that cells in
cortical cell packets divided according to a somewhat
regular pattern. Cell packets are also found in pea roots
(Webster, 1980). The cell packet is a developmental unit
In the 'classical view' a root tip is considered to be
organized into the following regions: the root cap, the
meristem, the elongation region, and the maturation
region. This view is a reasonable way to designate the
zones where different processes are mostly occurring.
These boundaries do not really exist, however, and each
cell file, or group of files tends to act independently. The
boundaries between the region of the meristem and the
region of elongation, therefore, may be different in cells
of the cortex compared to cells of the epidermis or the
vascular sectors.
The manner in which this happens involves the idea of
transition points. This concept, originally described by
Ivanov (1973), suggests that within each cell file there
are developmental switches, for example, the point where
cell division is turned off at the basal boundary of the
meristem. Another transition point would exist at
the boundary of the elongation region, and another at
the termination of cell maturation. In this way each file
or group of files may act independently of each other.
Exactly what happens at the transition point, is not
known, but perhaps specific genes are expressed which
turn events on or off in a cell file-specific manner.
The position of transition points is not static. That is,
as the rate of growth of a root changes, the location of
the transition point also changes. For example, as root
growth rate speeds up, the position of the maturation of
xylem vessel members becomes farther away from the
root tip, and as the root growth rate slows down the
position of maturation becomes closer to the root tip
Root organization and gene expression patterns
(Rost and Baum, 1988; Reinhardt and Rost, 1995). This
built-in mechanism allows the root to accommodate its
growth rate events spatially (cell division and elongation)
to its cell differentiation events. One observation of this
relationship was made by correlating the height of the
root meristem (relative position where cell division was
terminated in the cortex) and the position of xylem
maturation. It was determined that a close linear relation
existed. In roots with tall meristems maturation occurred
farther back in the root; in roots with short meristems
maturation of the xylem occurred closer to the tip (Rost
and Baum, 1988). In cotton roots a linear relation exists
between root growth rate and the position of protoxylem
maturation (Reinhardt and Rost, 1995). Fast growing
roots show protoxylem maturation farther from the root
tip than do slow growing roots.
Roof apical organization
An important difference between root tips of various
species is the organization of cells in the root apical
meristem (RAM). Popham (1966) categorized roots into
six types and two subtypes, all of these incorporating one
of two common developmental schemes open and closed
apical organization.
In closed organization all longitudinal cell file lineages
can be directly followed to a specific tier of initial cells.
These initials, or histogens (Hanstein, 1868) contribute to
the initiation of different tissues; for example, in the most
common closed type in monocotyledonous plants (e.g.
Poaceae) there are three initial layers, the plerome (vascular cylinder), the periblem/dermatogen (cortex and epidermis) and the calyptrogen (root cap). The most common
closed type in dicotyledonous plants also has three layers
but in a different pattern, the plerome (vascular cylinder),
the periblem (cortex), and the dermatogen/calyptrogen
(epidermis and root cap).
In roots with closed apical organization cell files have
direct lineages to one of the initial tiers and so it might
be assumed that the determination decision which identifies the cell lineage is a result of the initial tier connection.
That this type of mechanism is perhaps true was demonstrated at the molecular level by Dietrich et al. (1989) in
Brassica napus roots. The AX92 gene they characterized
was expressed exclusively in the cortical cells right up to
the periphery of the periblem initial tier.
In root tips with open apical organization, cell lineages
do not connect to distinct initial tiers. Instead, cell files
can be followed to a zone of cells without any obvious
organization (Rost, 1993, 1994). Both monocots and
dicots have variations of this pattern. In roots with open
organization, the identity of specific tissues is often not
clear for some distance back in the root body (Rost
et al., 1988).
One apical region that is common to both types of
1615
roots is the quiescent centre (QC). Clowes (1958) discovered the presence of quiescent cells positioned at the
boundary of the root body and cap. The QC is composed
of cells that are progressing through the cell cycle very
slowly and could include the cells comprising one or more
of the histogen tiers in roots with closed apical organization. The cells peripheral to the QC, the distal and
proximal meristems (Torrey and Feldman, 1977), are
cycling more rapidly and may have an important role in
cell and tissue development especially in roots with open
organization. Also, whether the initial tiers are part of
the QC or not is an important question. In Arabidopsis,
for example, the initial tiers are probably not actually a
component of the QC because mitotic divisions are often
seen there (Baum and Rost, 1996).
Closed root apical organization in Arabidopsis thaliana and
other dicot plants
A new conceptual model for closed apical organization
in Arabidopsis thaliana root tips has been recently suggested (Baum and Rost, 1996). A. thaliana roots have
closed apical organization (Fig. 2), and the dermatogen/
calyptrogen histogen tier is contiguous with the protoderm/epidermis and root cap (Fig. 3). This tier has two
components: (1) the columella initials (~20 cells) generates new root cap columella cells, and (2) around these
cells is a collar of root cap/protoderm initials (RCP).
The initials in both components are not quiescent. A key
feature of this model is that the cells of the RCP collar
are activated in small groups or one at a time, and they
regenerate themselves to maintain the process. The active
RCP initial divides periclinally to make a protoderm
initial (towards the inside) and a peripheral root cap
initial (towards the outside) making a T-division (Fig. 3).
The protoderm initial divides anticlinally to regenerate a
new RCP initial (toward the histogen tier) and the first
protoderm cell (toward the root body). The next active
RCP cell is positioned to the left or right of these cells
depending on the handedness of the spiral. When that
cell division is triggered, it will result in the sequence
being repeated as described, including the regeneration
on a new RCP initial cell. Each regenerated RCP initial
will be triggered to divide again only after the entire
collar of initials has divided in turn and the signalling
event has gone around an entire gyre, returning to the
original cell file (Baum and Rost, 1996). The number of
times a T-division occurs in a given epidermal cell file is
presumably a set number because the primary root is
determinate.
Using phyllotaxis terminology, each cell file could be
considered like an 'orthostichy' where a new T-division
in a particular cell file will be made after the RCP initials
have divided around one gyre. The number of times the
RCP initials divide in one gyre is equal to the number of
cell files in the protoderm.
1616
Rost and Bryant
RCP
protoderm Initial
T- division
columella Initials
peripheral root cap Initial
protoderm cell
\
columella root cap cells
RCP
peripheral root cap cells
Fig. 3. Cell division sequence of the root cap (calyptrogen)/protoderm
(dermatogen) initial tier leading to the continued development of the
protoderm-epidermis, the columella root cap, and the peripheral
root cap.
Fig. 2. Root cells and tissues are structurally organized in a hierarchical
order-cylinders, sectors, cell files and individual cells (from McGraw
Hill 1994 Yearbook of Science and Technology, The McGraw Hill
Companies, 1993). (Epidermis, E; cortex C; vascular cylinder, VC;
initial tiers-root cap and protodenm, RP; ground meristem, G; vascular
tissues, V.) Line scale = 31 ^m.
The histogen tier is also the source of the root cap
columella cells (Fig. 3). Three-dimensionally, the root
cap is actually a series of interconnected cones, where the
columella cells originate directly from the transverse part
of the histogen tier and the peripheral root cap cells
originate from the RCP initials. At any given time only
three to five gyres of root cap cones can be observed
because the old gyres are sloughed off as the root grows.
This kind of spiral also occurs in Azolla and other
ferns and lower vascular plants with apical cells, and, in
flowering plants, something like this has been suggested
only once before. Barlow (1993) reported the suggestion
of a spiral when he plotted the positions of past
T-divisions making the cortex along the axis of tomato
roots.
The second tier of initials (periblem) also has two
components, the central ground meristem initials (CGMI)
are one layer thick and are mostly quiescent (Rost et ah,
1996). This configuration means that Arabidopsis thaliana
is a 'type III' root according to the Popham (1966)
scheme. Around the periphery of the CGMIs is a collar
of ground meristem initials (GMI) (Fig. 4).
In a longitudinal view of a 1-week-old root, the peripheral GMI cell divides periclinal to the root surface
making a T-division (Fig. 4). The inner cell becomes the
endodennis initial (El) and the outer cell the most apical
cortex cell. At this point the cortex is two layers. A fewmicrometers basally, the El will then divide periclinal to
the root surface making another T-division to create the
endodermis to the inside and a middle cortical layer.
After 3 weeks or so the pattern of division changes and
the cortex now has three layers. This is because the
outermost cells of the CGMIs now lose their quiescence
and divide to make two cells. At this point the apical
organization pattern changes to 'type IIIA' (Popham,
1966), only not totally because often the innermost cell
of the CGMI does not divide (Fig. 4). This means that
there are then two GMI layers, the outer one simply
divides anticlinally to make outer cortex cells, and the
inner one is the El and it makes a T-division to form the
endodermis and middle cortex layers.
The three-dimensional pattern of the cortical Tdivisions is not a simple matter since these cells do not
divide at the same level. At the risk of over simplifying,
this division basically occurs in a spiral series where at
Root organization and gene expression patterns
type III stage
outer cortex cell
middle cortex / endodomus initial
. ground meristem initial
central ground menstem initials
typs IMA stags
endodermis
middle cortex
outer cortex
endodermis/middle cortex Initial
outer cortex initial
central ground meristem Initials
Fig. 4. Cell division sequence of the ground meristem (periblem) initial
tier leading to the formation of the ground meristem cortex. This initial
tier has only one layer in the embryo and up to about 1-week-old (type
III). After 1 week the central ground meristem initials divide to form
a two-layered tier (type II1A).
each higher level the position of the dividing cells rotated.
In other words, the T-divisions found around the cortex
initial tier, and the proliferative T-divisions which maintain the added layers of the cortex all seem to occur in
some kind of spiral pattern. Rost et al. (1996) freely
admit, however, that the spiral pattern related to this
initial tier is not as straightforward as that generated by
the protoderm/root cap tier.
The next initial tier, the plerome, forms the procambium and vascular tissues. It consists of approximately
18 rectangular shaped cells. Each of them divides in the
anticlinal plane to produce the next cell in a set number
of files. These files have a clear tissue identity-xylem,
phloem, or pericycle. T-divisions occur in some of these
cells higher in each file correspondent with the increase
in the number of cell files.
Gene expression
The ultimate goal of studying gene expression patterns is
to investigate the relationship between gene expression
and root development. An analysis of this type must be
firmly based on an understanding of root structure and
development including levels of organization, apical patterns, tissue identity and position, differences among roots
of the same plant depending on age and growth
conditions, and developmental event modulation.
Methods for studying gene expression in roots
Ultimately, the expression of a particular gene leads (with
the exception of those genes encoding rRNA, tRNA and
1617
snRNA) to the formation of an active protein. The
process of gene expression, however, may be regulated at
several points between the gene and the protein in its
active form. In order to obtain a full picture of the
expression of a particular gene it is necessary in theory
to assay for as many steps in the process as possible.
Unfortunately, while this may be a laudable goal, the
available techniques are rather more limited, as the
following paragraphs indicate.
(a) Assay of active protein: Many, although not all of
the proteins under consideration are enzymes. Thus, to
study gene expression at its end-point, it is necessary, to
study enzyme activity. However, assay of enzyme activity
does not generally provide the level of sophistication
necessary to locate the enzyme to particular tissues or
cells within organs. Techniques have been developed for
assaying enzymes such as DNA polymerase or topoisomerase in very small amounts of tissue (Bryant et al.,
1981; Chiatante et al., 1990; Olszewska and Kononowicz,
1979), but at best such techniques can only localize
activity to fairly generalized regions of the root. If enzyme
assay is to provide more detailed information on distribution at the cellular level, then it is necessary to develop
histochemical techniques which may be used with sectioned material. Such techniques are available for many
hydrolytic enzymes, and have, for example, enabled the
detection of enzyme activities which act as early markers
for the differentiation of vascular tissue (Gahan, 1981;
Rana and Gahan, 1983; Carmignac et al, 1990).
(b) Immunological methods: Use of antibodies to detect
protein relies on the specificity of the antibodies in
question. This in turn depends on the availability of a
purified protein, or of a chemically synthesized peptide
which is identical to part of the protein, against which to
raise the antibodies. Even if these criteria are met, polyclonal antibodies raised by injection of the protein into a
suitable mammal, may still not have the degree of specificity required. For example, polyclonal antibodies raised
against a particular DNA binding protein from pea
(Al-Rashdi and Bryant, 1994) exhibited strong crossreactivity with several other nuclear proteins (Burton,
1993) presumably because of domains possessed in
common by all the immunologically reacting proteins. In
such instances, it is necessary to raise a panel of monoclonal antibodies and to screen all of these in order to
obtain an antibody of sufficient specificity.
Antibodies may be used in conjunction with assays of
enzyme activity, as described in the section on cell cycle
genes. In such experiments, the use of an immunological
technique alongside a biochemical assay gives information
on the relationship between the level of activity and the
amount of enzyme protein. However, in the context of
this review, antibodies will be most extensively used in
fixed and embedded sectioned plant material. Exposure
of the section to the antibodies results in binding of the
1618
Rost and Bryant
antibodies to the target protein fixed in the section. The
bound antibodies may then be visualized by incubation
with a secondary antibody which is tagged either with an
enzyme whose activity may be detected histochemically
(e.g. peroxidase or with a fluorescent dye). Used carefully,
this technique facilitates the elucidation of the distribution
within tissues and within cells of the protein under
investigation. The technique may also be modified for use
with isolated nuclei in order to study the sub-nuclear
localization of proteins.
The major problem with this use of antibodies is the
frequent occurrence of non-specific reactions. In addition
to the possibility that antibodies may not be specific for
the proteins against which they are raised (as described
above), it is also possible for either or both the primary
and secondary antibody to bind spuriously to some
component of the tissue giving a signal which bears no
relation to the protein in question. Further, if the secondary antibody is linked to an enzyme, such as peroxidase,
it is quite possible for the tissue's own peroxidase to have
survived fixation and embedding in an active state, again
giving rise to a spurious positive signal (Rost and Bryant,
unpublished data). All these problems serve to indicate
that extensive background experiments and appropriate
controls are necessary in order to interpret properly the
results of immunolocalization experiments.
(c) In situ hybridization for mRNA detection: Of the
techniques for estimating and localizing gene expression,
the hybridization of antisense RNA to fixed, embedded
and sectioned tissue is the most reliable and least subject
to artefacts. All that is needed is an appropriate cloned
sequence such as a cDNA (which need not be full length)
spliced between two different promoters in opposite orientations. The promoters are used to drive the in vitro
transcription of the cloned cDNA to give antisense and
sense RNA molecules. Incorporation of labelled ribonucleotide (preferably labelled with digoxygenin, rather than
with a radionuclide) provides a means of detecting the
probe after the tissue sections have been incubated with
it. Assuming that the particular mRNA in question is
present in the cells, the antisense RNA probe will bind
to it, while the sense RNA probe will not, thus giving an
appropriate control. If no signal is obtained with either
probe, the validity of such a negative result can be
checked by using a probe which will detect an RNA of
universal occurrence, such as rRNA. Used properly, this
technique can provide good data on the relative abundance and cellular distribution of any mRNA for which a
cloned sequence is available.
(d) Use of GUS (^-glucuronidase) as a reporter gene:
While the assay by in situ hybridization of mRNA gives
a clear indication of whether that mRNA accumulates in
particular cells, it does not give information about the
activity of the promoter of the gene from which the
mRNA is transcribed. To gain information of this type it
is necessary to make a promoter GUS fusion, i.e. to make
a hybrid gene construct consisting of the promoter of the
gene in question adjacent to an otherwise promoterless
glucuronidase gene (Jefferson et ai, 1987). The construct
is inserted into a suitable vector and transformed into a
host plant. The activity of the promoter may be assayed
by the now well-known histochemical detection of glucuronidase activity, the presence of the blue stain released
from the chromogenic substrate being indicative of
enzyme activity. A variant of this is to transform plants
with a transposon within which has been inserted a
promoterless GUS gene. In this instance, GUS expression
will only occur if the transposon inserts into a gene in
frame with the promoter of that gene (Springer et al.,
1995).
Although GUS fusions are used very widely to study
gene expression, their use is not without problems, of
which two need to be mentioned here (Chriqui, 1996).
Firstly, the activity of GUS is assayed by diffusing the
chromogenic substrate into living tissue. If this is done
with whole organs, then there is a danger that the
distribution of signal may reflect not the distribution of
the glucuronidase, but the uneven distribution of the
substrate. Secondly, even if the substrate does reach all
the cells, it is still possible for the apparent distribution
of GUS to give spurious information about the activity
of the promoter in question. It is generally assumed that
the product of the action of glucuronidase on the chromogenic substrate gives an insoluble product, i.e. the indigocoloured crystals, which do not diffuse out of the cell.
However, the immediate product of the enzyme's activity
is soluble and diffusible; it is converted to the insoluble
final product by an oxidative process. It is thus possible
for the soluble initial product to diffuse from its site of
production to other cells and thence to be oxidized there.
This would result in a false positive in the latter cells in
respect of assaying promoter activity. However, provided
the investigator is aware of these problems and takes
steps to eliminate them, the use of plant transformed with
GUS genes fused to promoters remains the method of
choice for assaying promoter activity.
(e) Tissue printing: The advent of specific molecular
probes such as antibodies and antisense RNA has led to
a revival of the old technique of tissue printing (Song
et al., 1993; Ye and Varner, 1995). The virtue of this
technique is that it may be used to give information on
the general distribution of a particular protein or mRNA
throughout a whole organ or even a whole plant (particularly with Arabidopsis\). The organ, or plant is very
carefully squashed on to a suitable membrane such as
nitrocellulose or one of the many nylon derivatives. The
cellular contents are allowed to dry on to the membrane,
and after appropriate treatments, e.g. to remove pigments,
the membranes may be probed with antibodies or
antisense RNA as described earlier, giving an acceptably
Roof organization and gene expression patterns
1619
hyoscyamine 6^-hydroxylase. This enzyme was immunolocalized to the pericycle. They also isolated the genomic
clone (H6H) for this enzyme and observed its presence
in the pericycle by in situ hybridization. Last, they made
Root gene expression in tissues, menstems and regions
transgenic tobacco plants (containing the promoter region
It is possible to identify patterns of expression of many
of the H6H gene plus GUS and found it to be expressed
genes which are specific for particular cells and tissues.
in the root tip and in some cases the pericycle of tobacco
In some instances, there is a clear and obvious correlation
roots. This study is very rare in that these workers probed
between the gene and the function of the cell(s) in which
for the protein, the mRNA and also used transgenic
it is expressed. However, for many others, a clear funcplants plus GUS and found the gene and gene product
tional correlation can not be made, particularly for those
expressed in the same location by all three reporter
genes identified by an otherwise anonymous cDNA or by
methods.
an uncharacterized protein revealed only by reaction with
(d) Lateral roots: Hinchee and Rost (1986) have noted
a particular monoclonal antibody. Thus, much of this
that there are three stages to lateral root development—
work remains descriptive.
initiation, organization and emergence. Initiation refers
(a) Root cap: Choi et al. (personal communication with to those events which trigger lateral root formation. There
Professor Renee Sung, University of California, Berkeley)
are two different types of initiation events as pointed out
have screened antibodies made from proteins isolated at
by Charlton (1991). There is the possibility that the
different stages of carrot somatic embryogenesis. One of
pattern of lateral initiation is laid down in the root apex
these antibodies (Mab 64B6) detects a plastid antigen
as part of the general process there of cell production
which is found in the carrot seed radicle cap and root
and determination of tissue pattern.'... 'There is the
cap. Stephenson and Hawes (1994) have investigated
possibility that the pattern of lateral root initiation is
pectin methyl esterase (PME), which is involved in cell
determined later by hormonal means of control and/or
wall solubilization, and is found in root caps which are
by the action of other mobile morphogens.'... 'The
able to separate border cells. Border cells are peripheral
two possibilities may ultimately be perfectly compatroot cap cells which slough off and play an important
ible.' (p. 107)
role in rhizosphere maintenance. They found PME activMost work on identifying genes involved in lateral root
ity in root caps and also observed an increase in transcripinitiation
uses an experimental system where root segtion of PME mRNA prior to enzyme activity. Knox et al.
ments
are
placed in culture in the presence of exogenous
(1991) localized antibodies (JIM13, 14 and 75) of an
auxin.
Auxin-induced
lateral roots are of the second type
arabinogalactan protein to the root cap in carrot roots.
mentioned
above.
Roots
are routinely sampled at different
Lerner and Raikhel (1989) observed the presence of an
times
after
exposure
to
auxin followed by isolation of
mRNA (BLc3) for a lectin by in situ hybridization in the
mRNA
and
identification
of cDNAs expressed in early
root caps of barley roots.
stage lateral root primordia. In tomato, for example,
(b) Quiescent centre: Kerk and Feldman (1995) studied
Taylor and Scheuring (1994) have identified a gene
how the quiescent centre is maintained in corn root tips.
(RSI-1) corresponding to cDNA clone TR132 which is
Their idea is that auxin level is controlled by the activity
expressed in lateral root initiation sites, within 4 h of
of ascorbic acid which is regulated by ascorbic acid
auxin application, and is continuously expressed for 72
oxidase. In their study they demonstrated the histochemh. Keller and Lamb (1989) made transgenic tobacco
ical presence of ascorbic acid and of ascorbic acid oxidase
plants containing the promoter for hydroxyproline-rich
mRNA by in situ hybridization and Northern blots. This
glycoprotein (RRGPnt3) plus the GUS cassette. They
RNA was not exclusively present in the QC, but it was
observed the expression of GUS in the pericycle and
highly concentrated. This is the first identification of the
endodermis in positions corresponding to lateral root
expression of a specific gene in the QC. Sabelli et al.
initiation sites. The expression once started continued to
(1993) studied gene expression patterns in the root cap
be observed in the lateral root apex through emergence.
and quiescent centre. They have observed a QC-specific
Kerk and Sussex (unpublished data) have also identified
cDNA clone from PCR amplified cDNA libraries, but it
a cDNA clone of unknown function (4H11) in radish
has not been identified.
which is expressed within 24 h in lateral root initiation
(c) Pericycle: Knox et al. (1989) and Stacey et al. sites. This particular clone is interesting because its expres(1990) have isolated and screened antibodies made against
sion continues for 48 h and then stops. Miao and Verma
several arabinogalactan proteins which are present in cell
(1993) investigated the membrane channel protein
walls. One of these (JIM4) is present in the pericycle of
(Nodulin 26) in bacterial nodules of legume roots. They
carrot roots. Hashimoto et al. (1992) studied the biosynalso reported Nodulin 26-promoter expression in root
thesis of scopolamine in the roots of Hyoscyamus niger. meristems and lateral root primordia at different stages
They probed these roots with an antibody for the enzyme
rapid indication of the region of the organ or plant in
which expression of the gene in question is occurring
1620
Rost and Bryant
of development in transgenic plants using the GUS
reporter cassette.
(e) Epidermis: Clone BLc3 is a barley lectin isolated
from embryos. Using in situ hybridizations for its mRNA,
Lerner and Raikhel (1989) observed its presence on the
outer surfaces of the epidermis in embryo radicles. Knox
et al. (1991) isolated various arabinogalactan proteins
from cells of carrot. The antibodies for one of these,
JIM13, was observed in immature xylem tissue. Stacey
et al. (1990) localized another of these antibodies, JIM4,
in epidermis and other tissues. Pichon et al. (1992)
inoculated transgenic alfalfa (Medicago truncatula) roots
containing GUS with Rhizobium meliloti. Within 3—6 h
the nodulin gene MtENOD12 was expressed, based on
GUS activity, in epidermal cells near the root tip.
(f) Cortex: Dietrich et al. (1989) isolated and characterized a cortex specific clone, AX92. Using in situ hybridizations this clone was observed to be present in the cortex
of Brassica napus roots, but not in the initial cell tier.
Dietrich et al. (1992) made transgenic rape seed plants
and studied the distribution of the AX92 promoter plus
GUS. They determined that only a certain portion of the
promoter was needed to regulate the expression of this
gene and that it was expressed in the embryo radicle
cortex 20 d after fertilization.
John et al. (1992) also identified cortex-specific
mRNAs, but in corn roots. One of them ZRP3 accumulated in the outer 3-4 layers of the cortex. Held et al.
(1993) observed another clone, ZRP4. This one was
found in the endodermis at 4 cm from the root tip of 9-dold corn roots, and in both the endodermis and the
exodermis at 10 cm from the root tip. This clone was
identified to be Omethyl transferase.
(g) Vascular tissue—xylem, phloem and procambium:
Gahan (1988) used traditional histochemical techniques
to study the distribution of the enzyme carboxyesterase
in the procambium in radicles of pea (Pisum sativum) and
a number of other species. This enzyme was notably
absent from the quiescent centre. Mueller (1991) observed
the presence of the enzyme esterase in the procambium
of Trifolium pratense roots also using traditional histochemical techniques.
Yamamoto et al. (1991) characterized a root-specific
gene in tobacco called TobRB7. The mRNA for this gene
was observed in the vascular cylinder and root tip of
tobacco roots. TobRB7 promoter + GUS transgenic
tobacco plants were produced by the same group. GUS
staining was observed in the root tip, vascular cylinder
and in lateral root primordia. Demura and Fukuda (1994)
have identified three genes with vascular tissue specific
expression— TED2, 3 and 4. The mRNA for TED2 was
observed by in situ hybridization to be located in the
procambium of Zinnia roots. In situ's with TED3 and 4
also localized to the procambium, but with different
patterns.
Hydroxyproline-rich glycoproteins (HRPG) have been
localized to different sites in plant tissues. Stiefel et al.
(1990) showed HRPG gene expression in early stages of
vascular tissues in corn roots and in other places.
(h) Elongation zone: Ludevid et al. (1992) identified a
tonoplast intrinsic protein {TIP) in A. thaliana. The
mRNA for TIP was observed in the elongating cells of
roots, and the same labelling pattern was observed in
roots of transgenic plants using GUS as a reporter gene
for the promoter. Tanimoto (1994) reported an mRNA
clone in pea roots (RP25) which was also localized
exclusively to the elongating cells. Sequences of this clone
showed it to be homologous to a membrane intrinsic
protein (MIP) which is quite similar to the TIP membrane
protein mentioned above.
(i) Root apical meristem: Many gene expression patterns
have been localized to the root apical meristem (RAM)
using antibodies for specific proteins, mRNA clones and
GUS transgenic plants. Most of the genes observed in
the RAM have been cell cycle related, and only a few
genes not involved in cell cycles have been observed. One
of these TobRBl (Yamamoto et al., 1991) was observed
in the RAM using in situ hybridizations and also in
transgenic plants by GUS expression. It is curious to note
that GUS expression in the RAM seems to occur in many
studies. It is difficult to assess if this means that all kinds
of genes are switched on in the RAM or if this is a
peculiarity of meristems or the GUS assay.
/3-glucosidase cleaves cytokinin-O-glucoside, the mobile
form of cytokinin. Brzobohat et al. (1993) localized this
enzyme by in situ hybridization of the clone Zm-p60. The
enzyme was found only in the RAM, exclusively in
regions of cycling cells. Antibodies to the enzyme also
localized to the same cells.
The cell cycle in roots
Genes concerned with the cell cycle may be divided into
three general groups. First, there are genes which are
involved in the specification of meristem identity and the
regulation of meristem morphology and activity in relation to root growth and development. Because these genes
are almost exclusively known only from studies of developmental mutants, they are dealt with later, in the section
on root mutants. Second, there are genes which are
concerned with regulation of the cell cycle, such as those
encoding the cyclin-dependent protein kinases. Third,
there are the genes which encode the enzymes and other
proteins which mediate the specific events of the cell cycle,
for example, the enzymes of DNA replication.
(a) Cell cycle control genes: Of the proteins involved in
overall regulation of the plant cell cycle, the 34 kDa,
cyclin-dependent protein kinase (p34), encoded by the
plant homologue(s) of the fission yeast cdcl gene, is the
best known (Ferreira et al., 1994c). In fission yeast, this
Roof organization and gene expression patterns
gene is absolutely required for entry into the cell division
cycle and for progress through the cycle (Nurse, 1990).
This regulatory role can not be carried out by p34 on its
own. Its protein kinase activity is dependent on association with cyclins and is regulated by the phosphorylation
status of p34 itself (thus there are interactions with other
protein kinases and with protein phosphatases). In animal
cells, the transcription of the p34 gene(s) is under strict
control in relation to cell cycle activity; there is no
expression of the gene(s) in cells which are not dividing
but there is a high level of expression in dividing cells
(Ferreira et al., 1994c). In plants, there are at least two
classes of genes which are homologous to cdcl and it is
now clear that one of these is expressed in roots and the
other in aerial parts of the plant (Miao et al., 1993). The
root-specific gene is expressed strongly in dividing cells,
as indicated by expression of GUS-fusion constructs, by
in situ hybridization of anti-sense RNA probes and more
recently by immuno-localization of proteins containing
the cdcl PSTAIR motif (Mews et al., 1996). Thus, apical
meristems of primary and lateral roots, lateral root primordia and vascular cambium all show strong signals
(Hemerly et al., 1993; Ferreira et al., 1994c; Mews et al.,
1996). However, expression of p34cdc2 also occurs in cells
which are not dividing, including pericycle and parenchyma. These cells may express the gene at a lower level
than is seen in meristems, and may also exhibit
up-regulation of expression in response to auxin, or, as
is seen in legumes, during the induction of nodulation by
Rhizobium (Miao et al., 1993). Nonetheless, the fact
remains that expression of cdcl is not confined to dividing
cells. Rather, it is associated with competence to divide;
the presence of p34 does not on its own indicate division
activity. It thus appears in plants that strict cell cycle
control may be vested in other components of the p34
regulatory system. Relevant to this is the observation that
in Arabidopsis roots, the expression of the cyclin lAt gene
(a mitotic cyclin) is totally confined to dividing cells
(Ferriera et al., 1994ft). Further, during lateral root
formation, induction of the expression of the cyclin lAt
gene is a very early event. Root cells also express genes
encoding for at least three more mitotic cyclins, cyclin
2aAt, 2bAt and 3bAt (Ferriera et al., 1994a). Of these,
expression of cyclin 2bAt is root-specific although all of
them are expressed in roots. Like the expression of cyclin
lAt, the expression of these genes is tightly correlated
with cell division activity. The specific roles for several
different mitotic cyclins are not known, but these data
certainly support the notion that overall regulation of
progress through the cell cycle does not rely on regulating
the transcription of the p34 cdc2 gene, but of the genes
which encode proteins with which p34 interacts.
(b) Genes encoding proteins which mediate the events of
the cell cycle: Most of the work on proteins which mediate
the events of the cell cycle has concentrated on the
1621
enzymes and other proteins involved in DNA replication.
Of this work, the majority has been concerned not with
detailed analysis of gene expression per se, but with
studies of enzyme activity in relation to the occurrence
of DNA replication. A favourite system for experimentation is the plant embryo during the transition from
quiescence to active growth which occurs in seed germination. Although most researchers of this topic have dealt
with whole embryos rather than with just the root (radicle)
it is the root meristem which resumes activity first and
so, if the experiments have concentrated on the first round
of DNA replication and cell division, they have effectively
concentrated on the root meristem. The picture which
emerges from these studies is perhaps predictable: roots
in quiescent seeds have low but measurable activities
of the enzymes of DNA replication, as illustrated by
DNA polymerase (Robinson and Bryant, 1975), DNA
ligase (Daniel and Bryant, 1988) and topoisomerase I
(Chiatante and Bryant, 1994) and these activities increase
markedly as the cells in the meristem prepare for resumption of cell cycle activity. On this basis, it could be argued
that the activity of such enzymes is correlated with cell
proliferation. Such a conclusion would be supported by
the finding that in the roots of established pea seedlings,
DNA polymerase-a activity is present at high levels only
in the zone containing the root apical meristem (Bryant
et al., 1981). However, these studies do not give any
information about how these changes in activity are
regulated. This is a significant omission in view of the
fact that several enzymes of DNA replication are subject
to post-translational regulation, e.g. by phosphorylation
(Bryant and Dunham, 1988) and of the finding that in
roots of geminating peas, topoisomerase I may exhibit
marked changes in activity which are not based on
changes in the amount of enzyme protein (Chiatante and
Bryant, 1994). Further, these studies do not give a specific
indication of the detailed distribution of enzyme protein
in relation to the cells which are actually undergoing
DNA replication or cell division. An in situ method is
available for localizing DNA polymerase activity, but
this does not distinguish between different forms of
DNA polymerase. Thus, in roots of Tulipa kaufmanniana
(Olszewska and Kononowicz, 1979) DNA polymerase
activity is present in all dividing cells and not just in
those going through S-phase. DNA polymerase is also
present in differentiating cells which have ceased any form
of cell cycle activity (including DNA endoreduplication).
Because the non-replicative DNA polymerase-/? is high in
activity in the zones of the root where activity of the
replicative DNA polymerase-a is low (Bryant et al., 1981),
it is difficult to interpret the histochemical observations
for Tulipa in relation to the cellular distribution of cell
cycle enzymes.
In a limited number of instances changes in nuclear
protein content during germination have been studied
1622
Rost and Bryant
immunologically. For each protein studied, the data show
some unexpected features, again indicating that in the
context of gene expression, enzyme activities alone may
not be a good indication of what is happening. The first
example is PCNA (Citterio et cil., 1992; Suzuka et al.,
1989). This has no enzyme activity of its own, but acts
as an auxiliary protein for DNA polymerase-8, one of
the two (or possibly three) DNA polymerases involved in
DNA replication. Nuclei isolated from root meristems of
6 h geminated pea seeds react strongly with PCNA antibodies, although they are several hours away from resuming DNA replication. The reaction of nuclei with the
antibodies then declines to a very low level before increasing again shortly before the resumption of DNA replication. With antibodies raised against topoisomerase I and
topoisomerase II, there is little or no change in the nuclear
immunogenic response during germination (Levi et al.,
1994), despite the marked changes in activity, at least of
topoisomerase I (mentioned above). However, for topoisomerase II, there are very clear changes in the subnuclear
distribution of the signal, possibly associated with the
association of the enzyme with the resolution of daughter
chromosomes during cell division. The sub-nuclear location of the topoisomerase I by contrast, does not change
during germination.
Finally, there have been only a few investigations of
the detailed distribution patterns, within tissues, of proteins or of expression of the genes encoding the proteins.
As indicated above, several antibody probes are available
to detect replicative enzymes and related proteins, but to
date, the only one which has been used to study distribution within tissues is anti PCNA. The results obtained
with these antibodies suggest that PCNA is present in all
meristematic cells (Daidoji et al., 1992; Rost and Bryant,
unpublished data), despite earlier unconfirmed reports
that it was confined to cells going through the S-phase.
When studies of gene expression per se are considered,
data are again very scarce, despite the existence of a
number of cloned DNAs and/or upstream control elements. So far, the only data (apart from those on histones
which are dealt with elsewhere in this review) concern,
firstly, PCNA and, secondly, a potentially very exciting
gene family, the MCM genes. For PCNA, the data are
consistent with those obtained with antibodies and show
that the expression of GUS driven by the PCNA promoter
from rice occurs throughout the meristems in transgenic
tobacco (Kosugi et al., 1991).
In yeast, the MCM genes are essential for normal cell
cycle activity; one or more of the MCM proteins is
implicated in the activation of DNA replication origins
(reviewed in Toyn et al., 1995). In animal cells, one or
more MCM protein is an absolute requirement for entry
into S-phase (Todorov et al., 1994), almost certainly
because it forms part of a small complex of proteins
which interacts ith origins of DNA replication, licensing
them for one round of DNA replication (Chong et al.,
1995; Madine et al., 1995). Studies on plant MCM
proteins are in their very early stages. However, it is
becoming apparent that MCM proteins are detectable in
nuclei of monocots and dicots (Ivanova et al., 1994;
Bryant, Moore and Brice, unpublished data) and that in
Zea mays roots, an mRNA encoding an MCM protein
may be isolated from meristematic cells, but not from
non-dividing cells (Sabelli et al., 1996). More substantial
evidence comes from Arabidopsis, in which transposon
mutagenesis (with a transposon incorporating a GUS
gene in order to study expression patterns, as described
earlier) has revealed an essential cell cycle gene (named
PROLIFERA) with a very marked homology to the yeast
MCM genes (Springer et al., 1995). Further, the expression of PROLIFERA is confined to dividing cells.
(c) Histones: Histones H2A, H2B, H3, and H4 interact
with DNA to form nucleosomes, and histone HI is
responsible for aspects of chromatin organization (K5hler
et al., 1992). Most histones are replicated during DNA
synthesis (replication-dependent), but some histone transcription and translation is replication independent and
involved in such events as DNA repair. Both types of
histones have been observed in roots (Tanimoto et al.,
1993ft).
Koning et al. (1991) characterized the histone H2A
gene from a tomato cDNA library. In situ hybridization
showed that this mRNA was expressed in the tomato
shoot and root apical meristem in an apparently random
distribution in the cycling cells. The homologue for this
gene was identified in pea roots, and it showed the same
distribution. Tanimoto et al. (1993a, ft) further characterized the localization of histone H2A in pea roots by
co-localizing H2A mRNA with a digoxigenin (DIG) nonradioactive probe coupled with 3H-thymidine-labelled
DNA. In this way they showed that about 10% of the
cells were only 3H-labelled, 10% were DIG-labelled and
80% were double-labelled. They interpreted this to mean
that H2A transcription began in late Gx and continued
through most of S-phase. Tanimoto et al. (1993ft) also
showed that the H2A mRNA was found in all types of
cycling cells, including the xylem pericycle several millimetres beyond the boundary of the root apical meristem.
This distribution corresponded to the location of cycling
cells previously reported by Rost et al. (1988). Kohler
et al. (1992) found histone H2A, H2B, H3, and H4
homologues in a barley cDNA library and localized them
by in situ hybridizations showing expression of all of
these clones in barley root tip meristems. They did not
show photographs of in situ's for root tips, but those for
the shoot tip indicated the apparently random labelling
as one would expect if only the cells in S-phase-were
expressing histone mRNA.
In other histone studies, Lepetit et al. (1992) used the
histone H3 and H4 promoter from Arabidopsis coupled
Roof organization and gene expression patterns
to the GUS cassette to transform tobacco. In the roots
of the resulting plants, GUS was expressed in the root
tips and lateral root primordia, and indicated that most,
but not all of the histone GUS activity was replicationdependent. Terada et al. (1992) used a wheat histone H3
promoter plus GUS to transform rice cells. The transformed plantlets expressed GUS in the root tips and
lateral root primordia corresponding to the-localization
of cycling cells.
Mutants of root development
Mutations affecting initiation of roots in the embryo
The patterns of development and organization of roots
described earlier in the review depend on the correct
functioning in space and time of cell division within the
meristem and of the enlargement, differentiation and
functioning of those displaced cells. The control of these
processes is first seen in the early stages of embryogenesis
and it is during these early stages that mutational analysis
reveals the activity of genes which regulate the initiation
and location of the root meristem. An attempt to analyse
the specific function of the genes which have been discovered so far reveals a complex situation, in which particular functions are apparently regulated by several genes,
acting in order at particular times. Many of these mutants
affect the initiation of cell lineages. For example, the
mutant phenotype may be a failure to initiate by division
a particular cell group from which particular organs or
tissues are derived. However, the situation for some of
these cell lineage genes is actually more complex than
this: position effects within the embryo may in some
instances exert more influence than the cell lineage genes
(Scheres et al., 1994; Vandenberg et al., 1995; Scheres,
1996). So, when a particular progenitor cell (or group of
cells) is missing, its function may be at least partly taken
over by other cells. In fact, in monocots positional
information seems in normal development to be much
more important than cell lineage, with roots and shoots
emerging from what appears to be an unstructured mass
of embryonic cells (Barlow and Parker, 1996).
Root meristem progenitors are denned early in embryogenesis, well before the heart stage in dicot embryos.
Indeed, on the basis of strict lineage models, the meristem
can be traced back to the basal cell arising from the
earliest asymmetric division in the zygote. Mutants in
which the basal cell is deleted may lack the root and
hypocotyl altogether, or more frequently as in the
STUMP phenotype of Arabidopsis (Berleth et al., 1996)
give rise to very truncated roots with apparently normal
meristems which later in embryogenesis cease division (cf
MONOPTEROS, described later). Several Arabidopsis
mutants have been described which affect cell lineages
originating in the hypophyseal cell. This cell arises by
1623
division of the basal cell (Fig. 5) and is clearly involved
in denning the apical/basal plane of the embryo. Indeed,
ablation of the hypophyseal cell has a dramatic effect on
the subsequent development of the meristem such that it
has been suggested that the hypophyseal cell or its immediate derivatives (progenitors of the quiescent centre)
functions in the organization of an active root meristem
with the appropriate signals for cell division and cell
specification. Divisions of the hypophyseal cell give rise
to the quiescent centre, while divisions of the embryo
proper give rise basally to a series of initials. Several
mutations affecting these cell groups have been denned:
HOBBIT, BOMBADIL, ORC, and GREMLIN affect
the hypophyseal cell (Scheres, 1996). Each of these regulates a function which is involved in programming the
hypophyseal cell to carry out its role of specifying meristem structure and functions.
Mutations are also known which affect the apical basal
axis of the early embryo leading to aberrations in the
basal region of the developing 'heart' stage structure, in
the region which will give rise to initials. The earliest
acting of these is GNOM (Mayer et al., 1993) where the
mutant phenotype is again a failure to form a meristem.
MONOPTEROS mutants specifically lack the basal tier
in the heart stage embryo (Berleth and Jurgens, 1993).
This lesion can be traced back to the octant stage embryo;
cells of the embryo proper and the uppermost cell of the
suspensor do not exhibit normal patterns of division and
thereby fail to set up the normal basal body structures.
Cotyledons are normal, (although they may be wrongly
positioned), but the root is equivalent to just one file of
cells. Any roots which do form are extremely rudimentary
and also exhibit lesions in vascular development.
HD
20>jm
Fig. 5. Early stage embryo in a typical crucifer. The hypohyseal cell,
seen initially as the proximal cell of the suspensor, clearly contributes
tiers of cells to the embryo itself, as seen at the globe stage of
development. (Suspensor, S; hypophyseal cell, H; hypophyseal derivatives, HD.) (Redrawn from Cuming, 1993).
1624 Rost and Bryant
However, MONOPTEROS-mutants can be induced to
develop roots from callus cultures, showing that the
lesion is not a lesion in root formation per se, but in
the development of the normal hypocotyl-root axis in
embryogenesis. The ability to unhook function from
development, as in root formation in callus cultures of
MONOPTEROS mutants is not surprising: for example,
cell division in the absence of an organized meristem is a
'normal' feature of callus and suspension cultures.
Mutants affecting continuing meristem activity in
Arabidopsis
Mutants RMLl and RML2 in Arabidopsis are so-called
because they are said to be 'root-meristemless' (Cheng
et al., 1995). However, this is a misnomer. In the mutant
phenotype, an apparently normal root is formed in the
embryo, but in these roots, cell division is not resumed
after germination. It is concluded that the RML gene(s)
is involved in the reactivation of the cell cycle in the root
apical meristem. Intriguingly, this phenotype is also seen
in certain gibberellic acid-deficient mutants of tomato.
However, mutant seedlings can initiate lateral and adventitious roots which may grow normally, or, with some
mutant alleles, stop growing in a similar way to the main
root, which contains only c. 1500 cells.
A related phenotype is exhibited by the STP mutants
of Arabidopsis, in which there is no increase in the number
of cells in the meristem after it has been laid down in
embryogenesis (Baskin et al., 1995). The same mutation
also causes a failure to elongate in the differentiating cells
although elongation of dividing cells is not affected. The
roots of mutant plants are not responsive to cytokinins,
whereas tissue cultures respond normally, i.e. the STP
locus affects only the responses of roots to cytokinins.
and one mutant which specifically lacks the endodermal
layer. On the other hand, mutations have been described
in both tomato (Nakielski and Barlow, 1995) and tobacco
(Traas et al., 1995) in which there is an extra layer of
cortical cells, the progenitors of which are sandwiched
between the pole of the stele and the cap initials. This
extra layer arises because of the occurrence of an extra
round of periclinal divisions. Further, in tobacco, all the
cell layers also go through an extra round of anticlinal
division. In tomato, the described phenotype is seen in
Gibl mutants, which are deficient in production of gibberellin. However, the hyp2 mutant of tobacco is not gibberellin-deficient, although the application of gibberellic acid
to mutant plants can partially alleviate the effects of this
mutation. The authors conclude that the hyp2 gene and
gibberellins independently control these aspects of root
architecture and further state pithily that 'gibberellins
have a more direct effect on the decision of a cell to
divide in a particular plane than is normally proposed in
the literature'.
Concluding remarks
The root is an apparently simple organ and yet within it
there is an array of tissues and cells organized with a
beautiful precision. This in turn reflects a temporally
ordered series of developmental events. This precision,
which is seen for example in the occurrence of completely
different developmental pathways in adjacent files of cells,
implies that cells are exposed to and can respond to
detailed positional information. This positional information affects gene activity, but also is a result of the activity
of other genes. Thus, the development of the root may
be ascribed to a series of organ, tissue and cell identity
genes coming into play as an ordered progression. The
precision of this is further reflected in the fact that
Mutants which affect the radial structure of the root
particular genes are involved in the determination of the
specific number of cell layers or files which make up a
The mutations described so far either affect roots via the
particular tissue. The products of the genes involved in
apical/basal organization of the embryo leading to lesions
this developmental cascade are largely unknown as are
in meristem initiation or affect the division function of
the mechanisms by which one genetic/developmental
the meristem. However, there are mutations which affect
event
sets off the next one.
the radial organization of the meristem, thus causing
It is also clear that although root development appears
variation in the number of cell layers and the identity
to be genetically determined, the genetic programming
and function of the cells in those layers. Some of these
can be disturbed. There is thus a paradox that tight
mutations are of genes which affect the radial anatomy
programming is coupled with developmental plasticity.
of the root meristem itself during embryogenesis. In
Arabidopsis, the mutants known as wooden leg and gollum Thus the normal pattern of development can be disturbed
by both extrinsic and intrinsic factors. The former is
affect the organization of the vascular tissue, while short
exemplified by the effects of the application of plant
root, scarecrow and pinocchio affect cortex and endohormones to developing roots and by the effects of
dermis (Scheres et al., 1995). However, for most of the
environmental factors (including nutrition) on root
mutants of this type which have been described, the
growth and architecture; the latter is seen in the ability,
precise timing of the lesion is not denned, although the
albeit limited, of cell layers in the early embryo to take
effect of the lesion is clearly described. Thus, Coomber
up the function of layers which are missing because of
and her colleagues (Holding et al., 1994) described two
mutants of Arabidopsis with excessive radial expansion genetic mutation. The responses of cells to such factors
Root organization and gene expression patterns
gives a clue that at least some of the genes which regulate
root development may be involved in the production
and/or perception of signalling molecules. It is therefore
interesting that the functions of certain developmental
genes may be replaced, experimentally, by exogenous
application of plant hormones. Further, the ability of
cells in the root to communicate in all directions with
adjacent cells, coupled with the occurrence of specific
mechanisms for both transport and metabolism of signalling molecules (mechanisms which are, of course, themselves the result of gene action) allows for the formation
of complex interacting gradients of these molecules.
This brings us back to consider the linkage between
the different levels at which genes act. Taking the meristem
as an example, it is clear that there are genes which
regulate the identity, position and size of the root meristem. But how does this specification of meristem identity
lead to the expression of the genes which regulate the
activities of meristematic cells, e.g. DNA replication and
cell division? Signalling molecules may again be part of
the answer with the evidence that in different situations,
cytokinins and/or auxins may be directly involved in
regulation of cell cycle genes.
Study of genes and development in plants is still
providing us with more questions than answers. Use of
the root as an experimental system is going to play a key
role in answering those questions. We therefore hope that
this review stimulates further interest in roots.
References
Al-Rashdi J, Bryant JA. 1994. Purification of a DNA binding
protein from a multi-protein complex associated with DNA
polymerase-a in pea (Pisum sativum). Journal of Experimental
Botany 45, 1867-71.
Barlow PW. 1987. Cellular packets, cell division and morphogenesis in the primary root meristem of Zea mays. New
Phytologist 105, 27-56.
Barlow PW. 1993. The cell division cycle in relation to root
organogenesis. In: Ormrod JC, Francis D, eds. Molecular
and cell biology of the plant cell cycle. The Netherlands:
KJuwer Academic Publishers, 179-99.
Barlow PW, Parker JS. 1996. Microtubular cytoskeleton and
root morphogenesis. In: Barlow PW, ed. Plant roots: from
cells to systems. Proceedings of 14th Long Ashton
International Symposium (in press).
Baskin TI, Cork A, Williamson RE, Gorst JR. 1995. STUNTEDPLANT 1, a gene required for expansion in rapidly elongating
but not in dividing cells, and mediating root-growth responses
to applied cytokinin. Plant Physiology 107, 233-43.
Baum SF, Rost TL. 1996. Apical organization in Arabidopsis
roots: the root cap and protoderm. Protoplasma 192, 178-88.
Berleih T, Jurgens G. 1993. The role of the MONOPTEROS
gene in organizing the basal body region of the Arabidopsis
embryo. Development 118, 575-87.
Berleth T, Hardtke SC, Przemeck GKH, Mattsson J, Muller J.
1996. Mutational analysis of root initiation in the Arabidopsis
embryo. In: Barlow PW, ed. Plant roots: from cells to systems.
1625
Proceedings of 14th Long Ashton International Symposium
(in press).
BrumrkM RT. 1943. Cell-lineage studies in root meristems by
means of chromosome rearrangements induced by X-rays.
American Journal of Botany 30, 101-9.
Bryant JA, Durham VL. 1988. Replication of nuclear DNA.
Oxford Surveys of Plant Molecular and Cell Biology 5, 23-56.
Bryant JA, Jenns SM, Francis D. 1981. DNA polymerase
activity and DNA synthesis in roots of pea (Pisum sativum
L.). Phytochemistry 20, 13-15.
Brzobohaty B, Moore I, Kristoffersen P, Bako L, Campos N,
Shell J, Palme K. 1993. Release of active cytokinin by a pglucosidase localized to the maize root meristem. Science
262, 1051-4.
Barton SK. 1993. DNA binding-proteins associated with DNA
polymerase-a in pea. PhD thesis, University of Exeter, UK.
Cannignac DF, Gahan PB, Loukopoulou Z. 1990.
Carboxylesterase activities during the establishment of the
procambium and procortex in embryos of Pisum sativum L.
Cell Biochemistry and Function 8, 233-5.
Charlton WA. 1991. Lateral root initiation. In: Waisel Y, Eshel
A, Kafkafi U, eds. Plant roots: the hidden half. New York:
Marcel Dekker, Inc. 103-28.
Cheng JC, Seeky KA, Sung ZR. 1995. RML 1 and RML 2,
Arabidopsis genes required for cell proliferation at the root
tip. Plant Physiology 107, 365-76.
Chiatante D, Bryant JA. 1994. Activity of DNA topoisomerase
I and quiescence of embryo cells in seeds of Pisum sativum
L. Journal of Experimental Botany 46, 959-65.
Chiatante D, Fitchett PN, Bryant JA. 1990 Topoisomerase I
and II activities in nuclei from pea root meristem cells.
Journal of Experimental Botany 42, 813-20.
Ctaong JPJ, Mahbubani HM, Khoo CY, Blow JJ. 1995.
Purification of an MCM containing complex: a component
of the DNA-replication licensing system. Nature 375, 418-21.
Chriqui D. 1996. Rol genes and root initiation and development.
In: Barlow PW, ed. Plant roots: from cells to systems.
Proceedings of 14th Long Ashton International Symposium
(in press).
Citterio S, Sgorbati S, Levi M, Colombo BN, Sparvoli E. 1992.
PCNA and total nuclear protein content as markers of cell
proliferation in pea tissue. Journal of Cell Science 102, 71-8.
Clowes FAL. 1958. Development of quiescent centres in root
meristems. New Phytologist 57, 85-8.
Cuming AC. 1993. Seed development. In: Bryant JA, ed.
Methods in plant biochemistry, Vol. 10. Molecular biology.
London: Academic Press, 217-49.
Daidoji H, Takasaki Y, Nakane PK. 1992. Proliferating cell
nuclear antigen (PCNA/cyclin) in plant proliferating cells:
immunohistochemical and quantitative analysis using autoantibody and murine monoclonal antibodies to PCNA. Cellular
Biochemistry and Function 10, 123-32.
Daniel PP, Bryant JA. 1988. DNA ligase in pea {Pisum sativum)
seedlings: changes in activity during germination, and effects
of deoxyribonucleosides. Journal of Experimental Botany
39, 481-6.
Demura T, Fukuda H. 1994. Novel vascular cell-specific genes
whose expression is regulated temporally and spatially during
vascular system development. The Plant Cell 6, 967-81.
Dietrich RA, Maslyar DJ, Heupel RC, Harada JJ. 1989. Spatial
patterns of gene expression in Brassica napus seedlings:
identification of a cortex-specific gene and localization of
mRNAs encoding isocitrate lyase and a polypeptide
homologous to proteinases. The Plant Cell 1, 73-80.
Dietrich RA, Radke SE, Harada JJ. 1992. DNA sequence
downstream of the protein coding region are required to
1626 Rost and Bryant
activate a gene expressed in the root cortex of embryos and
seedlings. The Plant Cell 4, 1371-82.
Ferreira P, Hemerly A, Engler JD, Bergounioux C, Burssens S,
van Montagu M, Engler G, Inze D. 1994a. Discrete classes of
Arabidopsis cyclins are expressed during different intervals
of the cell cycle. Proceedings of the National Academy of
Sciences, USA 91, 11313-17.
Ferreira PCG, Hemerly AS, Engler JD, van Montague M,
Engler G, Inze D. 19946. Developmental expression of the
Arabidopsis cyclin gene eye lat. The Plant Cell 6, 1763-74.
Ferreira P, Hemerly A, van Montagu M, Inze D. 1994c. Control
of cell proliferation during plant development. Plant
Molecular Biology 26, 1289-1303.
Gahan PB. 1981. An early cytochemical marker of commitment
to stelar differentiation in meristems from dicotyledonous
plants. Annals of Botany 48, 769-75.
Gahan PB. 1988. Xylem and phloem differentiation in perspective. In: Roberts LW, Gahan PB, Aloni R, eds. Vascular
differentiation and plant growth regulators. Berlin: SpringerVerlag, 1-21.
Gifford Jr EM. 1993. The root apical meristem of Equisetum
diffusum: structure and development. American Journal of
Botany 80, 468-73.
Gunning BES. 1978. Age-related and origin-related control of
the numbers of plasmodesmata in cell walls of developing
Azolla roots. Planta 143, 181-90.
Gunning BES. 1982. The root of the water fern Azolla: cellular
basis of development and multiple roles for cortical microtubules. In: Subtelmy S, Green PB, eds. Developmental order: its
origin and regulation. Berlin: Springer-Verlag, 379—421.
Gunning BES, Overall RL. 1983. Plasmodesmata and cell-tocell transport in plants. BioScience 33, 260-5.
Gunning BES, Robards AW (eds). 1976. Intercellular communication in plant: studies on plasmodesmata. Berlin: SpringerVerlag.
Hanstein J. 1868. Die scheitelzellgruppe im vegetationspunkt
der phanerogamen. In: Festschrift zum 50 jarigen Jubilaum
Universitat Bonn, 109-43.
Hashimoto T, Kanegae T, Kajiya H, Yamada Y. 1992. Expression
of an alkaloid biosynthesis gene at the pericycle. Abstracts of
the Root Research Conference, City University of New York.
Held BM, Wang H, John I, Wurtele ES, Colbert JT. 1993. An
raRNA putatively coding for an O-methyltransferase accumulates preferentially in maize roots and is located predominantly in the region of the endodermis. Plant Physiology
102, 1001-8.
Hemerly AS, Ferreira P, Engler JD, van Montagu M, Engler G,
Inze D. 1993. Cdc 2a expression in Arabidopsis is linked with
competence for cell division. The Plant Cell 5, 1711-23.
Hinchee MAW, Rost TL. 1986. The control of lateral root
development in cultured pea seedlings. I. The role of seedling
organs and plant growth regulators. Botanical Gazette
147, 137^*7.
Holding DR, McKenzie RJ, Coomber SA. 1994. Genetic analysis
and structural analysis of five Arabidopsis mutants with
abnormal root morphology generated by the seed transformation method. Annals of Botany 74, 193-204.
Ivanova MI, Todorov IT, Atanossova L, Dewitte W, van Onckelen
HA. 1994. Co-localization of cytokinins with proteins related
to cell proliferation in developing somatic embryos of Dactylis
glomerata L. Journal of Experimental Botany 45, 1009-17.
Ivanov VB. 1973. Growth and reproduction of cells in roots.
In: Obrucheva NV, ed. Physiology of roots. 1-40.
Jefferson RA, Kavanagfa TA, Bevan MW. 1987. Gus fusions: 0glucuronidase as a sensitive and versatile gene fusion marker
in higher plants. EMBO Journal 13, 3901-7.
John I, Wang H, Held BM, Wurtele ES, Colbert JT. 1992. An
mRNA that specifically accumulates in maize roots delineates
a novel subset of developing cortical cells. Plant Molecular
Biology 20, 821-31.
Juniper BE, Barlow PW. 1969. The distribution of plasmodesmata in the root tip of maize. Planta 89, 352-60.
Keller B, Lamb CJ. 1989. Specific expression of novel cell wall
hydroxyproline rich glycoprotein gene in lateral root initiation. Genes and Development 3, 1639-46.
Kerk NM, Feldman LJ. 1995. A biochemical model for the
initiation and maintenance of the quiescent center: implications for organization of root meristems. Development
121, 2825-33.
Knox JP, Day S, Roberts K. 1989. A set of cell surface
glycoproteins forms an early marker of cell position, but not
cell type, in the root apical meristem of Daucus carota L.
Development 106, 47-56.
Knox JP, Linstead PJ, Peart J, Cooper C, Roberts K. 1991.
Developmentally regulated epitopes of cell surface arabinogalactan proteins and their relation to root-tissue pattern
formation. The Plant Journal 1, 317-26.
Kohler S, Coraggio I, Becker D, Salamini F. 1992. Pattern of
expression of meristem-specific cDNA clones of barley
(Hordeum vulgare L.). Planta 186, 227-35.
Koning AJ, Tanimoto EY, Kiehne K, Rost T, Comai L. 1991.
Cell specific expression of plant histone H2A genes. The
Plant Cell 3, 657-65.
Kosugi S, Suzuki I, Ohashi Y, Murakami T, Arai Y. 1991.
Upstream sequences of rice proliferating cell nuclear antigen
(PCNA) gene mediate expression of PCNA-GUS chimeric
gene in meristems of transgenic tobacco plants. Nucleic Acids
Research 19, 1571-6.
Lepetit M, Ehling M, Chaubet N, Gigot C. 1992. A plant
histone gene promoter can direct both replication-dependent
and -independent gene expression in transgenic plants.
Molecular and General Genetics 231, 276-85.
Lerner DR, Raikbel NV. 1989. Cloning and characterization of
root specific barley lectin. Plant Physiology 91, 124-9.
Levi M, Sparvoli E, Corbetta N. 1994. An antibody against a
sequence of human topoisomerase II gives different immunofluorescence patterns in quiescent and proliferating nuclei of
Pisum sativum L. Journal of Experimental Botany 45, 1157-62.
Ludevid D, Hofte H, Himelblau E, Chrispeeb MJ. 1992. The
expression pattern of the tonoplast intrinsic protein gamma
TIP in Arabidopsis thaliana is correlated with cell enlargement.
Plant Physiology 100, 1633-9.
Lucas WJ, Wolf S. 1993. Plasmodesmata: the intercellular
organelles of green plants. Trends in Cell Biology 3, 308-15.
Madine MA, Kboo C-Y, Mills AD, Laskey RA. 1995. MCM3
complex required for cell-cycle regulation of DNA replication
in vertebrate cells. Nature 375, 421-4.
Mayer U, Buttner G, Jurgens G. 1993. Apical-basal patternformation in the Arabidopsis embryo: studies on the role of
theGNOM gene. Development 117, 149-62.
Mews M, Baluska F, Volkmann D. 1996. Tissue and
development-specific distribution of PSTAIR-proteins in cells
of control and wounded maize root apices. Journal of
Experimental Botany 47, 819-29.
Miao GH, Hong ZL, Verma DPS. 1993. Two functional
soybean genes encoding p34 (cdc 2) protein-kinases are
regulated by different plant developmental pathways.
Proceedings of the National Academy of Sciences, USA
90, 943-47.
Miao G, Verma DPS. 1993. Soybean nodulin-26 gene encoding
a channel protein is expressed only in the infected cells of
Root organization and gene expression patterns
nodules and is regulated differently in roots of homologous
and heterologous plants. The Plant Cell 5, 781 -94.
Mueller RJ. 1991. Identification of procambium in the primary
root of Trifolium praiense (Fabaceae) American Journal of
Botany 78, 53-62.
Nakiebki J, Barlow PW. 1995. Principal directions of growth
and the generation of cell patterns in wild-type and GIB-1
mutant roots of tomato (Lycopersicon esculentum, Mill)
grown in vitro. Planta 196, 30-9.
Nurse P. 1990. Universal control mechanism regulating onset
of M phase. Nature 344, 503-8.
Olszewska MJ, Kononowicz AK. 1979. Studies of DNA
polymerases and RNA polymerases detected in situ in growing
and differentiating cells of root cortex. Histochemistry
59,311-23.
Picbon M, Journet E, Dedieu A, deBilly F, Trucbet G, Barker
DG. 1992. Rhizobium mehloti elicits transient expression of
the early nodulin gene ENOD12 in the differentiating root
epidermis of transgenic alfalfa. The Plant Cell 4, 1199-1211.
Popham R. 1966. Laboratory manual for plant anatomy. St
Louis, MO: CV Mosby 120.
Rana MA, Gahan PB. 1983. A quantitative cytochemical study
of determination for xylem-element formation in response to
wounding in roots of Pisum sativum L. Planta 157, 307-16.
Robinson NE, Bryant JA. 1975 Development of chromatinbound and soluble DNA polymerase activities during germination of Pisum sativum L. Planta 127, 69-75.
Reinhardt DH, Rost TL. 1995. On the correlation of primary
root growth and tracheary element size and distance from
the tip in cotton seedlings grown under salinity. Environmental
and Experimental Botany 35, 575-88.
Rost TL. 1993. The organization of root apical meristems in
flowering plants and patterns of gene expression. In: Parker
SB, ed. McGraw Hill Yearbook of Science and Technology
1994. New York: McGraw Hill Inc. 324-5.
Rost TL. 1994. Root tip organization and the spatial relationships of differentiation events. In: Iqbal M, ed. Growth
patterns in vascular plants. Portland OR: Dioscorides Press,
Chapter 2.
Rost TL, Baum S. 1988. On the correlation of primary root
length, meristem size and protoxylem tracheary element
position in pea seedlings. American Journal of Botany
75, 414-24.
Rost TL, Baum SF, Nichol S. 1996. Root apical organization
in Arabidopsis thaliana and a comment on root cap structure.
In: Barlow PW, ed. Plant roots; from cells to systems.
Proceedings of the 14th Long Ashton Symposium, 1995
(in press).
Rost TL, Jones TJ, Falk RH. 1988. The distribution and
relationship of cell division and maturation events in Pisum
sativum (Fabaceae) seedling roots. American Journal of
Botany 75, 1571-83.
Sabelli PA, Burgess SR, Carbajosa JV, Parker JS, Halford NG,
Shewry PR, Barlow PW. 1993. Molecular characterization of
cell populations in the maize root apex. In: Ormrod JC,
Francis D, eds. Molecular and cell biology of the plant cell
cycle. The Netherlands: KJuwer Academic Publishers, 97-109.
Scheres B. 1996. Genetic analysis of root development in
Arabidopsis. In: Barlow PW, ed. Plant roots: from cells to
systems. Proceedings of 14th Long Ashton International
Symposium (in press).
Scheres B, DiLaurenzio L, Willemsen V, Hauser MT, Janmaat
K, Weisbeek P, Benfey DN. 1995. Mutations affecting the
radial organization of the Arabidopsis root display specific
effects throughout the embryonic axis. Development 121,
53-62.
1627
Scberes B, Wolkenfelt H, Willemsen V, Terlouw M, Lawson E,
Dean C, Weisbeek P. 1994. Embryonic origin of the
Arabidopsis primary root and root-meristem initials.
Development 120, 2475-87.
Song YR, Ye Z-H, Varner JE. 1993. Tissue print hybridization
on membrane for localization of messenger RNA in plant
tissue. Methods in Enzvmology 218, 671-81.
Springer PS, McCombie WR, Sundaresan V, Martienssen RA.
1995. Gene trap tagging of PROLIFERA, an essential MCM
2-3-5-like gene in Arabidopsis. Science 268, 877-80.
Stacey NJ, Roberts K, Knox JP. 1990. Patterns of expression
of the JIM4 arabinogalactan protein epitope in cell cultures
and during somatic embyrogenesis in Daucus carota L. Planta
180, 285-92.
Stephenson MB, Hawes MC. 1994. Correlation of pectin
methylesterase activity in root caps of pea with root border
cell separation. Plant Physiology 106, 739-45.
Stiefel V, Rulz-Avila L, Raz R, Valles MP, Gomez J, Pages M,
Martinez Izquierdo JA, Ludevid MD, Langdale JA, Nelson T,
Puigdomenech P. 1990. Expression of a maize cell wall
hydroxyproline-rich glycoprotein gene in early leaf and root
vascular differentiation. The Plant Cell 2, 785-93.
Suzuka I, Daidoji H, Matsuoka M, Kadowaki K, Takasaki Y,
Nakane PK, Moriuchi T. 1989. Gene for proliferating-cell
nuclear antigen (DNA polymerase auxiliary protein) is
present in both mammalian and higher plant genomes.
Proceedings of the National Academy of Science. USA
86, 3189-93.
Tanimoto EY. 1994. Identification and localization of differentiation-specific mRNAs in the primary roots of pea, Pisum
sativum var. Alaska. PhD Dissertation, University of
California, Davis, CA.
Tanimoto EY, Rost TL, Comai L. 1993a. DNA replication
dependent histone H2A mRNA expression in pea root tips.
Plant Physiology 103, 1291-7.
Tanimoto EY, Rost TL, Comai L. 19936. Histone H2A
expression during S-phase: Histological co-localization of
H2A mRNA and DNA synthesis in pea root tips. In: Ormrod
JC, Francis D, eds. Molecular and cell biology of the plant
cell cycle. The Netherlands: KJuwer Academic Publishers,
85-95.
Taylor BH, Rodgers LC. 1992. Visualization of tranposable
element-induced clonal sectors in tomato roots. Abstracts of
the Root Research Conference, New York University, New
York
Taylor BH, Scbeuring CF. 1994. A molecular marker for lateral
root initiation: The RSI-1 gene of tomato Lycopersicon
esculentum Mill) is activated in early lateral root primordia.
Molecular and General Genetics 243, 148-57.
Terada R, Nakayama T, Iwabuchi M, Shimato K. 1992. A wheat
histone H3 promoter confers cell division-dependent and
-independent expression of the gus a gene in transgenic rice
plants. The Plant Journal 3, 241-52.
Todorov I, Pepperkok R, Philipova R, Kearsey S, Ansorge W,
Werner D. 1994. A human nuclear protein with sequence
homology to a family of early S-phase proteins is required
for entry into S-phase and for cell division. Journal of Cell
Science 107, 253-65.
Torrey JG, Feldman LJ. 1977. The organization and function
of the root apex. American Scientist 65, 334-44.
Toyn JH, Toone WM, Morgan BA, Johnston LH. 1995.
Activation of DNA replication in yeast. Trends in Biochemical
Sciences 20, 70-3.
Traas J, Laufi P, Jullien M, Cabocbe M. 1995. A mutation
affecting etiolation and cell elongation in Nicotiana
plumbaginifolia causes abnormal division plane alignment
1628
Rost and Bryant
and pattern formation in the root meristem. The Plant
Journal 7, 785-96.
Webster PL. 1980. Analysis of heterogeneity of relative division
rates in root apical meristems. Botanical Gazette 141, 353-9.
Vandenberg C, Willemsen V, Hage W, Weisbeek P, Scheres B.
1995. Cell fate in Arabidopsis root meristem determined by
directional signalling. Nature 378, 62-5.
Yamamoto YT, Taylor CG, Acedo GN, Cheng C, Cookling MA.
1991. Characterization of cw-acting sequences regulating
root-specific gene expression in tobacco. The Plant Cell
3, 371-82.
Ye Z-H, Varner JE. 1995. Differential expression of two
o-methyltransferases in lignin biosynthesis in Zinnia elegans.
Plant Physiology 108, 459-67.
Zhu T, Baum SF, Lucas WJ, Rost TL. 1996. Dynamic
intercellular communications in Arabidopsis root apical
meristem. Abstracts of the Third International Workshop
on basic and applied research in plasmodesmata biology,
Zichron-Yakov, Israel.