Báo cáo khoa học: Expression of MsPG3-GFP fusions in Medicago truncatula Ôhairy rootsÕ reveals preferential tip localization of the protein in root hairs pot

Expression of MsPG3-GFP fusions in
Medicago truncatula
Ôhairy rootsÕ
reveals preferential tip localization of the protein in root hairs
Ignacio D. Rodrı
´
guez-Llorente
1
, Javier Pe
´
rez-Hormaeche
1
, Mohammed Dary

1
, Miguel A. Caviedes
1
,
Adam Kondorosi
2,3
, Pascal Ratet
2
and Antonio J. Palomares
1
1
Departamento de Microbiologı

´
a y Parasitologı
´
a, Facultad de Farmacia, Universidad de Sevilla, Spain;
2
Institut des Sciences
Ve
´
ge
´
tales, Centre National de la Recherche Scientifique, Gif sur Yvette, France;
3

Institute of Genetics, Biological Research Center,
Hungarian Academy of Sciences, Szeged, Hungary
Tip growth is a specialized type of polar growth where new
cell wall is deposited in a localized region of the cell, the
growing tip. These cells show a characteristic zonation, with
a high accumulation of secretory vesicles containing cell wall
components at the tip, followed by an organelle-enriched
zone. MsPG3 is a Medicago sativa polygalacturonase gene
isolated in our laboratory, specifically expressed during the
interaction of this plant with its symbiotic partner Sinorhiz-
obium meliloti and which might participate in tip growth
processes during symbiosis. We have used MsPG3-GFP

fusions to study in vivo protein transport processes and
localization during root hair growth. Different MsPG3-GFP
fusions were expressed in Medicago truncatula Ôhairy rootsÕ
following a protocol developed for this study and also tested
by transient expression in onion epidermal cells. Preferential
accumulation of an MsPG3-GFP fusion protein in the tip of
the growing root hair at different developmental stages was
found, confirming the delivery of MsPG3 to the newly
synthesized cell wall. This indicates that this protein may
participate in tip growth processes during symbiosis and, in
addition, that this fusion could be a useful tool to study this
process in plants.

Keywords:GFP;hairyroot;Medicago truncatula; polygal-
acturonase; tip growth.
Plant cells grow either by diffuse growth, over a wide region,
or by tip growth, limited to the apex. Tip growth is a
specialized type of polar growth where new cell wall is
deposited in a localized region of the cell, the growing tip.
These cells show a characteristic zonation, with a high
accumulation of secretory vesicles containing cell wall
components at the tip, followed by an organelle-enriched
zone [1]. Cell wall produced and deposited by tip growth is a
mechanism used in various cellular systems including pollen
tubes, fungal hyphae, developing root hairs and Rhizobium-

induced infection threads. Pollen tubes have been used as a
model system to investigate the tip growth process in plants
[2]. More recently, root hairs in Arabidopsis have become a
model system for tip growth [3]. The isolation and
phenotypic characterization of mutants with defects in
specific aspects of root hair growth has led to the definition
of four stages in Arabidopsis root hair morphogenesis: the
selection of a growing site, bulge formation, tip growth and
polarized extension [3,4]. In the same way, four stages have
been described in Vicia sativa spp. nigra L. (vetch) root hair
development: bulging, growing, growth terminating and full
growth hair [5]. While recent studies with Arabidopsis

mutants have provided new insights into how the tip growth
is governed [1], the mechanisms directing the growth
specifically to the tip are still unknown. Only the importance
of microtubules in this process has been described [6–8].
Polygalacturonases (PGs, EC 3.2.1.15) are cell-wall-
degrading enzymes involved in the degradation of pectins
that are complex polysaccharides found in the middle
lamella and primary cell wall of higher plants. In a previous
study [9], we characterized a Medicago sativa PG gene
(MsPG3) specifically expressed during symbiosis with
Sinorhizobium meliloti. Our results suggested that MsPG3
may participate in several steps of the infection process,

including infection thread formation and reinitiation of the
root hair tip growth induced by the Nod factor during the
early steps of the plant–bacterial interaction [10,11].
The primary aim of this research was to study the cellular
localization of the MsPG3 protein in vivo, using green
fluorescent protein (GFP), in order to understand better its
role during the early steps of symbiosis. GFP is commonly
used for in vivo protein localization, as the mechanism of
fluorophore formation, involving intramolecular autoxida-
tion, does not require exogenous cofactors [12]. A major
advantage of GFP is the maintenance of its fluorescence
when fused to other proteins making it a very useful reporter

protein. To test their functionality, MsPG3-GFP fusions
were transiently expressed in cells from epidermal peels of
onion (Allium cepa) [13]. In addition we have modified the
protocol for the generation of transgenic Ôhairy rootsÕ for
Correspondence to A. J. Palomares, Departamento de Microbiologı
´
a
y Parasitologı
´
a, Facultad de Farmacia, Universidad de Sevilla,
41012 Sevilla, Spain.
Fax: + 34 954556924, Tel.: + 34 954556924;

E-mail:
Abbreviations: BFA, brefeldin A; ER, endoplasmic reticulum;
GFP, green fluorescent protein; MS, Murashige and Skoog media;
PG, polygalacturonase; t-nos, nopaline synthase terminator.
Enzymes: Polygalacturonases (PGs, EC 3.2.1.15).
(Received 10 September 2002, revised 13 November 2002,
accepted 21 November 2002)
Eur. J. Biochem. 270, 261–269 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03384.x
Vicia hirsuta plants described by Quandt et al.[14]toobtain
rapidly transgenic roots of M. truncatula expressing the
GFP fusions in root hairs.
By using this system we showed that MsPG3 is specif-

ically exported to the growing part of the root hair cell
during its development, indicating that the protein may
participate in tip growth processes during symbiosis. In
addition, the protocol described in this work for M. trun-
catula hairy root production may be useful to study in vivo
those proteins that are involved in root hair growth and
elongation, and the fusions could also be used as tools to
examine the secretory system activity in different physiolo-
gical, biochemical and developmental contexts.
Materials and methods
Construction of GFP fusion proteins
Oligonucleotide primers used in this work were:

5038 ()12): 5¢-CTAA
GAATTCACATGGATAGGA
AA-3¢; PG3B (1984): 5¢-GG
GGATCCGCTTCTGCTGC
AGTTGTGC-3¢;EPS-1(72):5¢-CC
CCATGGCTAAT
ATCTTTGATATAAA-3¢;EPS-2(49):5¢-CCACCAG
GATTGGGACCACGCC-3¢;SPS-1()33): 5¢-C
CCCGGG
AGTGAAAAAAGCAAAGTTCAAC-3¢; SPS-2 (105):
5¢-CCC
CCATGGCTCCTCCAAATGATTTTATATC-3¢.

The underlined sequences indicate the EcoRI (5038),
BamHI (PG3B), NcoI (EPS-1 and SPS-2) and SmaI (SPS-1)
restriction enzyme cleavage sites used for cloning. Distances
from the oligonucleotide 5¢ ends to the ATG of the MsPG3
genomic sequence (EMBL data bank, accession no.
Y11118) are given, except for EPS-2, designed from the
gfp sequence [15].
MsPG3-gfp translational fusions were made as described
below.
pgc-gfp-t-nos. The pgc fragment containing the complete
MsPG3 coding sequence including introns was obtained by
PCR amplification from plasmid DNA using oligonucleo-

tides 5038 and PG3B. This fragment was cloned in the
pGEM-T easy vector (Promega) and transferred as an
EcoRI-BamHI fragment to a Bluescript-derived plasmid
(pKSgfp, Stratagene) containing a BamHI, NcoI-gfp-t-nos-
NotI cassette from pmon30049 [15] (t-nos is the nopaline
synthase terminator).
pgte-gfp-t-nos. The pgte fragment containing the two first
exons of MsPG3 and corresponding to an EcoRI-NcoI
fragment of pgc, was cloned as an NcoI fragment in-frame
in the NcoIsiteofthegfp-nos cassette of pKSgfp.
pgwsp-gfp-t-nos. Using oligonucleotides EPS-1 and EPS-2,
the pgwsp fragment lacking the signal peptide was amplified

from fragment pgte-gfp, NcoI restricted and cloned as a
NcoI fragment in-frame with gfp in pKSgfp plasmid.
pgsp-gfp-t-nos. Oligonucleotides SPS-1 and SPS-2 were
used to amplify a 105 bp DNA fragment containing the
sequence of the predicted MsPG3 signal peptide. The NcoI
and SmaI restriction sites generated by PCR were used to
clone the pgsp fragment in-frame with the gfp coding
sequence in pKSgfp plasmid.
All of these MsPG3-gfp-t-nos fragments were transferred
from pKSgfp plasmid to a pK18-derived plasmid [16]
containing the CaMV35S promoter (pK35). In the same
way, a gfp-t-nos fragment from pmon30049 was placed in

this plasmid under the control of CaMV35S promoter.
These plasmids, named pPGC-GFP, pPGTE-GFP,
pPGWSP-GFP, pPGSP-GFP and p35S-cyt-GFP (Fig. 1),
were used for particle bombardment experiments.
For hairy root experiments 35S-pgte-gfp-t-nos and 35S-
pgwsp-gfp-t-nos cassettes were cloned as HindIII-SacI
fragments into pLP100 binary vector [17] to obtain
pPGTE-GFPb and pPGWSP-GFPb plasmids. Plasmid
pLP35GFP [18] was used as a cytoplasmic GFP control.
Finally, these binary vectors were transformed into Agro-
bacterium rhizogenes Arqua1 [14] by electroporation.
A plasmid containing the modified gfp cassette mgfp4-ER

(with a peptide targeting system) [19] has been used as a
control of GFP fluorescence localization in the endoplasmic
reticulum (ER), both in onion and M. truncatula roots. The
construct containing this cassette was called GFP-ER
(Fig. 1).
All the constructs used in this study are listed in Table 1,
indicating their functional domains and their expected
targeting.
Fig. 1. Schematic representation of GFP fusions used in this work.
35S, constitutive promoter from cauliflower mosaic virus; gfp, S65T-
intron green fluorescent protein [15]; t-nos, nopaline synthase termi-
nator; pgc, construction including the MsPG3 complete coding

sequence; pgte, construction including the MsPG3 first two exons and
the first intron; pgwsp, construction including the pgte part of MsPG3
without the signal peptide; pgsp, construction including only the
MsPG3 signal peptide; mgfp4-ER,modifiedgfp cassette with a peptide
for ER targeting [19]. The restriction sites relevant for the construc-
tions are: B, BamHI; H, HindIII; E, EcoRI; Nc, NcoI; N, NotI; S, SacI;
Sm, SmaI.
262 I. D. Rodrı
´
guez-Llorente et al.(Eur. J. Biochem. 270) Ó FEBS 2003
Transformation of onion cells by particle bombardment
Onion plant material and the gold particle bombardment

protocol are described in Scott et al. [13]. This study used
the same equipment and protocol, apart from the prepar-
ation of cells. Particles were bombarded directly onto onion
pieces that were peeled just before observation, instead of
bombarding onion peels placed on agar plates.
To visualize GFP in the cell wall, onion pieces were
bathed in 20 m
M
sodium phosphate buffer (pH 7.0)
following the 22 h incubation period needed for gfp
expression as described in Scott et al. [13].
Generation of

M. truncatula
Ôhairy rootsÕ
Seeds of M. truncatula (Gaertn.) R-108–1 (3c) were surface
sterilized in 8 gÆL
)1
Bayrochlor mini (Bayrol GMBH) in a
shaking Erlenmeyer flask for 45 min. Thereafter, they were
rinsed six times in sterile water. The seeds were allowed to
swell overnight by incubation in sterile water. Pregermi-
nated seeds were dried briefly by removing all the water and
transferred to agar plates containing 0.5· Murashige and
Skoog (MS) salts and vitamins (Sigma, M 0404), 10 gÆL

)1
sucrose and 9 gÆL
)1
kalys 575 agar (Mayoly Spindler,
France). Ten seedlings were placed on 12 cm square plates
which were then incubated vertically in a 25 °Cgrowth
cabinet in the dark. After 24 h the plantlets were transferred
to a growth chamber with a 16 h photoperiod (120–
130 mEÆm
)2
Æs
)1

), at 25 °C with a relative humidity of 60%.
When the first leaves appeared, around three days later, the
plants were placed again in the dark for two days to
elongate the hypocotyls to facilitate bacterial infection.
When the hypocotyl length reached 3 cm, the plates were
placed again in the growth chamber with the photoperiod
described above.
M. truncatula plantlets having hypocotyls of 3–3.5 cm in
length, 24–48 h after the last transfer into light, were
infected with A. rhizogenes by stab inoculation. One side of
the hypocotyl was stabbed three to five times (approxi-
mately 1/3 distance from the hook of the primary root to the

cotyledons) with an Agrobacterium-containing needle.
Thereafter, A. rhizogenes was taken again from the plate
with the needle and placed carefully on the wounded area of
the hypocotyl. Plates containing infected plantlets were
returned to the growth chamber in the conditions described
previously.
Between 2 and 3 weeks after this procedure, hairy roots
were obtained in 25–30% of the plants. When these roots
were at least 2 cm long, the main root was excised and the
resulting composite plants were transferred to fresh
0.5· MS with 200 mgÆL
)1

cefotaxime (Sigma C7912) to
avoid Agrobacterium proliferation. The composite plants
were grown in the conditions described above, keeping the
roots in the dark. One week later, Ôhairy rootsÕ with a length
of 4 cm were checked for GFP expression by epifluores-
cence microscopy as described below. Positive roots were
then placed in the dark individually on fresh medium with
antibiotic, as every root is a single transformation event. At
this stage, fast lateral root development took place until
enough material for further detailed analysis was obtained.
These transgenic roots can be maintained for months in
plates if transferred on fresh medium every three weeks.

GFP detection
Roots and onion cells were examined using a Polyvar
microscope with two types of filters giving an excitation
spectrum between 450 and 495 nm (B1) or between 475
and 495 nm (B4) and a stop filter at 520 nm (B1) or
between 520 and 560 nm (B4). Images were recorded
using a Leica DC200 camera. Confocal images were
obtained in a Sarastro 2000 Confocal Microscope
(Molecular Dynamics).
Results
Testing the expression capacity of GFP fusions
in onion epidermal cells

Scott et al. [13] have developed a rapid transient expression
system using onion skin cells to express GFP fusion
proteins. The onion epidermis has large, living, transparent
cells, ideal for visualizing GFP. We thus used this system to
test the capacity of various MsPG3-gfp fusions to be
expressed and easily detected before root transformation.
Onion cells expressing the cytoplasmic GFP fusion (cyt-
GFP) showed a cytoplasmic and nuclear GFP localization
(Fig. 2A). This previously reported nuclear localization of
GFP [20] is due to its small molecular weight (27 kDa). As a
second control, we expressed the GFP-ER construct in
onion cells, which targets GFP to the ER (Fig. 1B).

Fluorescence in GFP-ER expressing cells could be observed
in the perinuclear region and in the cortical zone of the cell
as a reticulate pattern of fluorescence (Fig. 2B). When the
PGC-GFP construct, corresponding to the fusion of the
complete MsPG3 coding sequence to GFP, was delivered
into onion epidermal cells only a weak expression sur-
rounding the nucleus was found (data not shown), probably
corresponding to a weak gfp expression localized in the ER
surrounding this nucleus. Scott et al. [13] demonstrated that
GFP fusions targeted to the cell wall could not be detected
due to the low pH of this compartment, but could be
visualized by incubating the tissue in a medium buffered at

pH 7.0. As the MsPG3 protein might be exported to the cell
wall compartment, these cells were bathed overnight in a
medium buffered at pH 7.0. Under these conditions,
fluorescence appeared in the border of the cell (Fig. 2C),
suggesting that the fusion protein containing the full
Table 1. Polygalacturonase-gfp andcontrolfusionsusedinthisstudy.
The main characteristics of the constructs and the expected subcellular
localization of the codified proteins are indicated.
GFP fusion Functional domains Expected targeting
cyt-GFP None Cytoplasm and nucleus
PGC-GFP Complete MsPG3 gene Cell wall
PGTE-GFP First two exons and first

intron of MsPG3
Endoplasmic
reticulum?/cell wall?
PGWSP-GFP PGTE without the
signal peptide
Cytoplasm and nucleus
PGSP-GFP MsPG3 signal peptide Endoplasmic reticulum
GFP-ER Peptide targeting system Endoplasmic reticulum
Ó FEBS 2003 Tip localization of MsPG3-GFP fusions (Eur. J. Biochem. 270) 263
MsPG3 coding sequence was exported to the plasma
membrane, or more probably to the cell wall.
Interestingly, onion cells transformed with the PGTE-

GFP construct (containing only the two first exons of
MsPG3) showed a spotted staining pattern with fluorescent
bodies (Fig. 2D). In addition, it was possible to detect
fluorescence in the transvacuolar strands as well as around
the nucleus (Fig. 2D) suggesting ER targeting. To verify
that the spotted structures represented Golgi stacks, their
sensitivity to brefeldin A (BFA) was tested. When onion
peels expressing PGTE-GFP were incubated with BFA the
pattern of fluorescence appeared different (Fig. 2E). Bigger
structures rather than the small spots were detected,
suggesting the localization of GFP within a BFA-sensitive
Fig. 2. Transient expression of the GFP fusion proteins in onion epidermal cells. A-G: epifluorescence microscopy (B1 filter). (A) Cell expressing

cytoplasmic GFP (cyt-GFP). Fluorescence appears inside the nucleus and in transvacuolar strands. (B) Cell expressing the GFP-ER fusion.
Fluorescence is observed in the perinuclear region and in the cortical ER. (C) Cell expressing the PGC-GFP construct and bathed at pH 7.0.
Fluorescence is detected in the cell wall. (D) Cell expressing the PGTE-GFP construct. A spotted pattern of fluorescence is detected, in addition to
fluorescence associated to the transvacuolar strands and to the perinuclear region. (E) Reorganization of PGTE-GFP-labelled structures after BFA
treatment (50 lg/mL). (F) Cell expressing the PGTE-GFP construct and bathed at pH 7.0. Fluorescence is observed in the cell wall. (G) Expression
of PGWSP-GFP construct in onion cells. Fluorescence can be observed in the nucleus and in the cytoplasm (border of the cell and transvacuolar
strands). (H) Expression of PGSP-GFP detected by confocal scanning microscopy. Fluorescence can be observed associated to the cortical ER and
in the ER located around the nucleus and in the transvacuolar strands. TVS, transvascular strands; N, nucleus; C, cortical ER. Bars ¼ 50 lm.
Confocal image has been coloured.
264 I. D. Rodrı
´
guez-Llorente et al.(Eur. J. Biochem. 270) Ó FEBS 2003

compartment probably representing the Golgi apparatus.
When these peels were bathed at pH 7.0 we observed the
same expression pattern as that detected with the complete
MsPG3 (Fig. 2F). These experiments indicate that the
protein fusion containing the first two exons (PGTE-GFP)
is sufficient for exporting the GFP to the cell wall. In
addition, this construct allows the detection of the different
compartments used by the protein during the exportation
process, and was thus later chosen for root transformation
(see below).
The first step in MsPG3 export pathway is probably
peptide penetration in the ER mediated by the presence of a

signal peptide, described in all plant PGs cloned to date. The
activity of the predicted MsPG3 signal peptide was deter-
mined using two new GFP fusions. Using the PCR
technique, the sequence of the predicted signal peptide
was amplified from the pgte fragment to obtain fragment
pgwsp. The PGWSP-GFP construct expressed in onion cells
showed fluorescence in the cytoplasm and in the nucleus
(Fig. 2G), similar to the one observed using the cytoplasmic
GFP (Fig. 2A). Finally, an MsPG3 fragment coding only
for the first 35 amino acids of MsPG3 and including the
22 amino acids of the predicted signal peptide was fused to
gfp (PGSP-GFP) and used to transform onion cells. The

fluorescence pattern observed using the PGSP-GFP con-
struct indicated an ER localization of the GFP (Fig. 2H), as
we obtained the same pattern of expression in cells
transformed with the GFP-ER construct. Using confocal
scanning microscopy, fluorescence was detected in tubules
and lamellar regions of the cortical ER, in the ER present in
transvacuolar strands and in the ER surrounding the
nucleus (Fig. 2H).
MsPG3 export during
M. truncatula
root hair
development

To study MsPG3 localization and export during root hair
development, M. truncatula Ôhairy rootsÕ expressing the cyt-
GFP, GFP-ER, PGWSP-GFP and PGTE-GFP constructs
were obtained. Ten out of 40 of the plants infected with
A. rhizogenes containing T-DNA expressing cyt-GFP pro-
duced between two and four hairy roots. Eight of them were
identified as transgenic roots. The same number of Ôhairy
rootsÕ was generated in approximately 30% (12/40) of the
plants transformed with PGTE-GFP. In this case, 10 roots
showed GFP activity. Similar results were obtained using
the PGWSP-GFP and GFP-ER constructs. Despite the
relatively low efficiency of Ôhairy rootÕ generation of our

protocol, they were easily identified as GFP fluorescent
roots. We used a binary vector carrying a CaMV35S
promoter-gus fusion as a control with similar results. Here,
Ôhairy rootsÕ were produced in 30% of the infected plants,
indicating that the transformation efficiency in these
experiments is independent of the construct used. None of
the control roots showed fluorescence when the roots were
young. The emission of weak yellow autofluorescence was
detected in old roots (data not shown).
Roots expressing GFP were placed individually on plates
and lateral roots emerged very quickly. Young lateral roots
with root hairs were cut and tested for GFP emission under

the epifluorescence microscope. Roots transformed with
cyt-GFP showed cytoplasmic (around the cell, due to the
presence of a large vacuole) and nuclear fluorescence
localization (Fig. 3A). The expression pattern of the cyto-
plasmic GFP was the same in all the cells of the root,
including root hairs (Fig. 3B). This result was also observed
with confocal scanning microscopy (Fig. 4A). Cells from
the root border expressing the cyt-GFP construct showed
fluorescence in the nucleus and the border of the cell. The
same pattern of expression was found in roots transformed
with the PGWSP-GFP construct (fusion without the signal
peptide), both in meristematic root cells (Fig. 3C) and root

hairs (Fig. 3D).
When roots were transformed with the PGTE-GFP
construct, different localization of the fluorescence was
observed. Differentiated cells exhibited stronger GFP fluor-
escence in the periphery of the cell and around the nucleus,
suggesting cell wall localization and possibly perinuclear
ER localization (Fig. 3E), in agreement with the results
obtained with onion cells. Meanwhile, the cells that were
developing to form root hairs at the bulge stage exhibited
fluorescent vesicles accumulating at the place where the root
hair was emerging (Fig. 3E). This specific targeting of the
fluorescence to the growing part of the root hair was also

clearly observed using confocal scanning microscopy, both
in a cell in the bulge stage (Fig. 4B) and in a growing root
hair (Fig. 4C). This pattern of fluorescence probably
represents Golgi stacks or transport vesicles, accumulating
at the growing part of the cell. At a later stage of
development, the growth terminating stage, root hairs
showed GFP fluorescence close to the root hair tip, in the
elongating root hair region, in addition to the perinuclear
localization (Fig. 3F). In roots transformed with the GFP-
ER construct the fluorescence appeared mainly around the
nucleus, where the density of ER is high (Fig. 3G,H). This
fluorescence was also observed in the border of the cells with

weak intensity (Fig. 3G). This pattern of fluorescence had
common localizations with the one described for roots
transformed with the PGTE-GFP construct (everything
related to the ER), but it also showed certain differences.
One difference was clearly observed in the growth termin-
ating stage of the root hairs, as in that case fluorescence was
associated only with the perinuclear region, without the
accumulation in the root hair tip (Fig. 3H).
Discussion
We have previously cloned, sequenced and partially char-
acterized an M. sativa polygalacturonase gene (MsPG3)
specifically expressed during the interaction of this plant

with S. meliloti [9]. Our previous results suggested that
MsPG3 might be involved in the early stages of the
interaction, including infection thread formation and rein-
itiation of the root hair tip growth induced by the Nod
factor during the early steps of the plant–bacterial interac-
tion [10,11]. These two processes are related to tip growth,
defined as a specialized type of growth where organelles are
arranged in zones and where new cell wall is deposited in a
localized region of the cell, the growing tip. This growing
mechanism is used in several important cellular systems
including pollen tubes, fungal hyphae, root hairs and
Rhizobium-induced infection threads.

We have used the reporter protein GFP fused to various
part of the MsPG3 coding sequence [12] to study the
Ó FEBS 2003 Tip localization of MsPG3-GFP fusions (Eur. J. Biochem. 270) 265
Fig. 3. Expression of the GFP fusion proteins in M. truncatula hairy roots detected by epifluorescence microscopy. Nuclear and cytoplasmic
fluorescence localization in root cells (A) and root hair cells (B) transformed with the cyt-GFP. Nuclear and cytoplasmic fluorescence localization in
root cells (C) and root hair cells (D) expressing the PGWSP-GFP fusion. (E,F) GFP expression in roots transformed with PGTE-GFP. (E) Strong
fluorescence can be observed around the nuclear region and in the periphery of the root cell. Fluorescent bodies, with a preferential accumulation in
the region of the emerging root hair, in a cell in the bulge stage (ERH) can be observed. (F) Root hair in growth terminating stage with fluorescence
localized in the root hair tip. Fluorescence around the nucleus is also observed. (G,H) Root cells expressing the GFP-ER fusion. (G) Fluorescence
observed in the cell border and the perinuclear region, containing higher density ER. The intensity of the latter fluorescence is weak. (H) Root hair
in growth terminating stage. Fluorescence is mainly observed around the nucleus, without accumulation at the root hair tip. ERH, emerging root
hair; N, nucleus; T, root hair tip. Bars ¼ 20 lm in panels A, C, E and G and 10 lminpanelsB,D,FandH.

266 I. D. Rodrı
´
guez-Llorente et al.(Eur. J. Biochem. 270) Ó FEBS 2003
localization of our protein. GFP coding sequences have
been fused at either the 5¢ or 3¢ end of the coding region of a
DNA sequence of interest and the resulting N-terminal or
C-terminal fusions used for in vivo studies on vesicular
trafficking, protein localization and cellular compartmenta-
tion in plants [20]. Potential problems using GFP fusions are
conformational changes in the attached protein, which
could activate localization signals that are normally seques-
tered in the absence of GFP, and improper folding or

instability of the encoded chimeric GFP, so that little or no
fluorescence is detectable. Negative results of this type are
rarely described in the literature [21]. Using this reporter
system, simple and rapid transient assays have been
developed to test the GFP fusion proteins before stable
transformation. For example, the onion epidermal cell
bombardment protocol described by Scott et al. [13] is very
suitable for testing GFP fusion proteins before stable
transformants are attempted. Onion epidermis, which has
large, living and transparent cells in a single layer, appears to
be particularly useful material for visualizing GFP in
transient assays.

In the work presented here we have developed an
alternative to the onion cell system for studying protein
localization in relation to cell growth, absent in the onion
cells. This protocol is based on the production of transgenic
hairy root on wounded hypocotyls of young seedlings of
M. truncatula, a diploid autogamous legume that is cur-
rently being developed as a model plant for the study of root
endosymbiotic associations [22]. Due to their fast and
hormone-independent growth, hairy root cultures represent
a material of choice to study roots and they have been
obtained from more than 100 different species [23]. This
transformation system is faster and cheaper than complete

plant transformation and has the advantages of stable
transgenic material over transient assays, in which damage
often occurs during DNA incorporation and for which
there is variability in the amount of DNA delivered.
All the plant PGs cloned to date have a N-terminal
hydrophobic signal sequence that targets the protein to the
lumen of the ER. The presence of a 22 amino acid
hydrophobic N-terminal section in MsPG3, displaying the
properties of a signal peptide [24], strongly suggests post-
translational cleavage of the protein and secretion of the
mature protein. In this work we have used two different
controls: a cytoplasmic GFP (cyt-GFP) expressed from the

CaMV35S promoter fusion, that showed the previously
reported [20] cytoplasmic and nuclear localization of the
fluorescence, and a modified GFP with a target sequence
that keeps the protein in the lumen of the ER [19]. This
second control helped us to explain part of our results. The
localization in our work of the PGSP-GFP fusion (MsPG3
signal peptide fused to GFP) in the ER, and of the PGWSP-
GFP fusion in the cytoplasm and in the nucleus suggest, as
expected, that the predicted MsPG3 signal peptide is
enough to target GFP to the ER but is not enough for
cell wall localization.
When the pgc-gfp fusion, containing the complete

MsPG3, was expressed in onion cells, we observed only a
weak GFP fluorescence around the nucleus, probably
corresponding to the ER. The GFP fluorescence is pH
dependent, with fluorescence intensity decreasing at low pH
[25]. Thus the lack of visualization of GFP in the cell wall
can be attributed to the low pH of this compartment. Scott
et al. [13] showed indeed that GFP fluorescence appeared in
the cell wall 4 h after buffering the cells at pH 7.0. As the
development of the GFP fluorophore takes approximately
4 h [26], it was thus suggested that newly synthesized GFP
was exported to the cell wall. Because PGs are supposed to
be localized within the cell wall, we bathed onion peels

transformed with this fusion overnight at pH 7.0, and
fluorescence appeared associated with the border of the cell.
The pattern of fluorescence did not change when cells
expressing the cytoplasmic GFP underwent the same
treatment (data not shown). Thus, our results suggest that
GFP is targeted to the cell wall when fused to the complete
MsPG3 peptide.
Interestingly, when the pgte-gfp fusion corresponding to
the two first exons of the gene was expressed in onion cells,
ER localization was also observed, but in addition we
detected fluorescence associated with transport vesicles
Fig. 4. Confocal scanning microscopy images of roots transformed with GFP. (A) Root transformed with the cyt-GFP construct. No preferential

accumulation of fluorescence in the root hair tip was observed. (B,C) Roots transformed with the PGTE-GFP construct. Preferential localization of
the fluorescence was observed in the region of the emerging root hairs (ERH), both in a hair in the bulge stage (B) and in a growing root hair in the
phase of organelle zonation (C). Images 1–4 are sequential confocal images of the same cells. ERH, emerging root hair; N, nucleus. Bars ¼ 10 lm.
Images have been coloured.
Ó FEBS 2003 Tip localization of MsPG3-GFP fusions (Eur. J. Biochem. 270) 267
probably representing Golgi stacks, as suggested by the
BFA treatment. BFA is a drug that induces reorganization
of the Golgi apparatus and blocks protein exportation [27].
Finally, cell wall localization of GFP fluorescence was
revealed when cells expressing this fusion were bathed at
pH 7.0 as found for the full length fusion. Thus, this
construct was useful because it allowed us to detect the

various steps of the exportation process followed by
the MsPG3 protein. These results indicate either that the
presence of the entire MsPG3 peptide allowed a more
complete or faster export of the protein to its final location,
or that this shorter fusion may not contain all the
information (such as glycosylation sites) necessary for the
efficient targeting of the protein. Another possibility is that
the conformation of the hybrid protein does not allow its
proper exportation, resulting in its partial retention in the
different compartments (ER, Golgi), or finally that the
shorter fusion is more fluorescent that the longer one and
allows a better detection. In conclusion, the fusion including

the signal peptide and the two first exons of the MsPG3
protein is sufficient and necessary to detect the protein along
the exportation pathway and to localize it to the cell wall.
We took advantage of this construct to study the localiza-
tion of the MsPG3 protein in developing root hairs.
In M. truncatula root cells that were developing to form a
root hair, the PGTE-GFP fusion was detected in the ER
apparatus as well as inside the cell at the site of the emerging
root hair at bulge stage and at the apical part of a growing
hair. The apical region in a growing hair corresponds to the
transport vesicles-rich region [5]. Similarly, this fusion was
specifically detected at the tip of the mature root hairs, rich

in secretory vesicles containing cell-wall components. The
result observed in roots transformed with the GFP-ER
construct helped us to understand which part of the pattern
of fluorescence detected in PGTE-GFP expressing roots is
related to the localization of the fusion protein in the ER
and which one represents a more specific targeting. In root
hairs expressing the GFP-ER fusion, the tip localization of
the fluorescence described in PGTE-GFP expressing root
hairs could not be observed, suggesting further exportation
of MsPG3 protein in the developing root hair. The
fluorescent pattern in roots transformed with the PGTE-
GFP construct might result from a specific localization of

the MsPG3 protein to the tip of these cells, but we can not
exclude the possibility that it also represents the localization
of all proteins that are excreted following the secretion
pathway in the root hairs and thus is a reflection of the cell
biology of a developing root hair. In all cases, it indicates
that the MsPG3 protein can be exported to the root hair tip
and thus can, by its enzymatic activity, participate to the tip
growth processes during symbiosis.
In addition to giving us information about the localiza-
tion of the MsPG3 protein, this fusion protein turned out to
be a useful tool to visualize protein trafficking and
localization in developing root hairs. Thus, the experimental

system described in this work may be used to study in vivo
and at the cellular level different aspects of root hair tip
growth. Because these hairy roots are suitable for hormone
or drug treatments, the system could be used to study the
secretory system activity in different physiological, bio-
chemical and developmental contexts. The transformation
protocol described allows the generation of composite
plants, consisting of transgenic roots on M. truncatula
untransformed shoots, which can be nodulated successfully
by their symbiotic partner. Recently, Boisson-Dernier et al.
[28] described a protocol for hairy root production in
M. truncatula that is probably faster than our one, making

this kind of study even easier. Thus our fusions can also be
used to study localization of proteins involved in infection
thread formation during symbiosis, another tip growth
based process. Finally, the protein fusions used in this work
could also be used as a tool to examine the secretory system
in other contexts, such as pollen tubes, wound sites or
abcission zones.
Acknowledgements
We are grateful to Fundacio
´
nRamo
´

n Areces, Junta de Andalucı
´
a
(CVI-181) and Ministerio de Educacio
´
n y Cultura (DGESIC PB95-
1268 and DGESIC PB98-1158) for supporting this work. IRLL was a
Fundacio
´
nRamo
´
n Areces and FPI Ministerio de Educacio

´
nyCultura
fellowship recipient. JPH was an MIT Ministerio de Educacio
´
ny
Cultura fellowship recipient. MD was a Ministerio de Asuntos
Exteriores fellowship recipient. We will like to thank Dr Beatrice
Satiat-Jeunemaitre and I. Couchy for technical advices and discussions.
References
1. Mathur,J.&Hu
¨
lskamp, M. (2001) Cell growth: How to grow and

where to grow. Curr. Biol. 11, 402–404.
2. Yang, Z. (1998) Signaling tip growth in plants. Curr. Opin. Plant
Biol. 1, 525–530.
3. Schiefelbein, J.W. (2000) Constructing a plant cell: The genetic
control of root hair development. Plant Physiol. 124, 1525–1531.
4. Parker, J.S., Cavell, A.C., Dolan, L., Roberts, K. & Grierson, C.S.
(2000) Genetic interactions during root hair morphogenesis in
Arabidopsis. Plant Cell. 12, 1961–1974.
5. Miller, D.D., Leferink-ten Klooster, H.B. & Emons, A.M. (2000)
Lipochito-oligosaccharide nodulation factors stimulate cyto-
plasmic polarity with longitudinal endoplasmic reticulum and
vesicles at the tip in vetch root hairs. Mol. Plant-Microb. Interac.

13, 1385–1390.
6. Wang, H., Lockwood, S.K., Hoeltzel, M.F. & Schiefelbein, J.W.
(1997) The ROOT HAIR DEFECTIVE 3 gene encodes an evo-
lutionary conserved protein with GTP-binding motif and is
required for regulated cell enlargement in Arabidopsis. Genes Dev.
11, 799–811.
7. Bibikova, T.N., Blancaflor, E. & Gilroy, S. (1999) Microtubules
regulate tip growth and orientation in root hairs of Arabidopsis
thaliana. Plant J. 17, 657–665.
8. Molendijk, A.J., Bischoff, F., Rajendrakumar, C.S., Friml, J.,
Braun,M.,Gilroy,S.&Palme,K.(2001)Arabidopsis thaliana
Rop GTPases are localized to tips of root hairs and control polar

growth. EMBO J. 20, 2779–2788.
9. Mun
˜
oz, J.A., Coronado, C., Pe
´
rez-Hormaeche, J., Kondorosi, A.,
Ratet, P. & Palomares, A.J. (1998) MsPG3, a Medicago sativa
polygalacturonase gene expressed during the alfalfa–Rhizobium
meliloti interaction. Proc. Natl Acad. Sci. USA 95, 9687–9692.
10. de Ruijter, N.C.A., Rook, M.B., Bisseling, T. & Emons, A.M.C.
(1998) Lipochito-oligosaccharides re-initiate root hair tip growth
in Vicia sativa with high calcium and spectrin-like antigen at the

tip. Plant J. 13, 341–350.
11. Catoira, R., Galera, C., de Billy, F., Penmetsa, R.V., Journet,
E.P.,Maillet,F.,Rosenberg,C.,Cook,D.,Gough,C.&
De
´
narie
´
,J.(2000)FourgenesofMedicago truncatula controlling
components of a Nod Factor transduction pathway. Plant Cell.
12, 1647–1665.
268 I. D. Rodrı
´

guez-Llorente et al.(Eur. J. Biochem. 270) Ó FEBS 2003
12. Cody, C.W., Prasher, D.C., Westler, W.M., Prendergast, F.G. &
Ward, W.W. (1993) Chemical structure of the hexapeptide chro-
mophore of the Aequorea green-fluorescent protein. Biochem. 32,
1212–1218.
13.Scott,A.,Wyatt,S.,Tsou,P.L.,Robertson,D.&Stromgren
Allen, N. (1999) Model system for plant cell biology: GFP imaging
in living onion epidermal cells. Biotechnical 26, 1125–1132.
14. Quandt, H.J., Puhler, A. & Broer, I. (1993) Transgenic root
nodules of Vicia hirsuta: a fast and efficient system for the study of
gene expression in indeterminate-type nodules. Mol. Plant-
Microb. Interac. 6, 699–706.

15. Pang, S.Z., DeBoer, D.L., Wan, Y., Ye, G., Layton, J.G., Neher,
M.K., Armstrong, C.L., Fry, J.E., Hinchee, M.A.W. & From,
M.E. (1996) An improved green fluorescent protein gene as a vital
marker in plants. Plant Physiol. 112, 893–900.
16. Pridmore, R.D. (1987) New and versatile cloning vectors with
kanamycin-resistance marker. Gene 56, 309–312.
17. Szabados, L., Charrier, B., Kondorosi, A., de Bruijn, F.J. & Ratet,
P. (1995) New plant promoter and enhancer testing vectors. Mol.
Breed. 1, 419–423.
18. Kamate
´
,K.,Rodrı

´
guez-Llorente, I.D., Scholte, M., Durand, P.,
Ratet, P., Kondorosi, E., Kondorosi, A. & Trinh, T.H. (2000)
Transformation of floral organs with GFP in Medicago truncatula.
Plant Cell Report 19, 647–653.
19. Haseloff, J., Siemering, K., Prasher, D. & Hodge, S. (1997)
Removal of a cryptic intron and subcellular localization of green
fluorescent protein are required to mark transgenic Arabidopsis
plants brightly. Proc. Natl Acad. Sci. USA 94, 2122–2127.
20. Chiu, W.L., Niwa, Y., Zeng, W., Hirano, T., Kobayashi, H. &
Sheen, J. (1996) Engineered GFP as a vital reporter in plants. Curr.
Biol. 6, 325–330.

21. Hanson,M.R.&Ko
¨
hler, R.H. (2001) GFP imaging: methodology
and application to investigate cellular compartmentation in plants.
J. Exp. Bot. 52, 529–539.
22. Cook, D. (1999) Medicago truncatula: a model in the making!
Curr. Opin. Plant Biol. 2, 301–304.
23. Shanks, J.V. & Morgan, J. (1999) Plant Ôhairy rootÕ culture. Curr.
Opin. Biotechn 10, 151–155.
24. Verner, K. & Schatz, G. (1988) Protein translocation across
membranes. Science. 241, 1307–1313.
25. Di Sansebastiano, G.P., Paris, N., Marc-Martin, S. & Neuhaus,

J.M. (1998) Specific accumulation of GFP in a non-acidic
vacuolar compartment via a C-terminal propeptide-mediated
sorting pathway. Plant J. 15, 449–457.
26. Heim, R., Prasher, D.C. & Tsien, R.Y. (1994) Wavelength
mutations and posttranslational autoxidation of green fluorescent
protein. Proc.NatlAcadSci.USA91, 12501–12504.
27. Satiat-Jeunemaitre, B., Cole, L., Bourett, T., Howard, C. &
Hawes, C. (1996) Brefeldin A effects in plant and fungal cells:
Something new about vesicle trafficking? J. Microsc. 181,
162–177.
28. Boisson-Dernier, A., Chabaud, M., Garcia, F., Be
´

card, G.,
Rosenberg, C. & Barker, G.D. (2001) Agrobacterium rhizogenes-
transformed roots of M. truncatula for the study of nitrogen-fixing
and endomycorrhizal symbiotic associations. Mol. Plant-Microb.
Interac. 14, 695–700.
Ó FEBS 2003 Tip localization of MsPG3-GFP fusions (Eur. J. Biochem. 270) 269