Download Tomato LeAGP-1 is a plasma membrane-bound

PHYSIOLOGIA PLANTARUM 120: 319–327. 2004
Printed in Denmark – all rights reserved
Copyright # Physiologia Plantarum 2004
Tomato LeAGP-1 is a plasma membrane-bound,
glycosylphosphatidylinositol-anchored arabinogalactan-protein
Wenxian Suna,c, Zhan Dong Zhaob, Michael C. Harec, Marcia J. Kieliszewskib,c and Allan M. Showaltera,c,*
a
Department of Environmental and Plant Biology,
Department of Chemistry and Biochemistry,
c
Molecular and Cellular Biology Program, Ohio University, Athens, OH 45701-2979, USA
*Corresponding author, e-mail: [email protected]
b
Received 21 March 2003; revised 13 June 2003
Arabinogalactan-proteins (AGPs) are a class of highly glycosylated, hydroxyproline-rich glycoproteins that function in
plant growth and development. Tomato LeAGP-1 represents
a major AGP expressed in cultured cells and plants. Based on
cDNA and amino acid sequence analyses along with carbohydrate
and other biochemical analyses, tomato LeAGP-1 is hypothesized to be a classical AGP localized to the plasma membrane
via a glycosylphosphatidylinositol (GPI) anchor. Here, this
hypothesis was tested and supported with the following experiments. First, tomato (Lycopersicon esculentum, cv. UC82B)
cotyledon protoplasts were isolated following cell wall digestion
with cellulase and pectinase, and LeAGP-1 was immunolocalized
to the plasma membrane with a LeAGP-1 antibody. Second,
LeAGP-1 was shown to be a major AGP component in plasma
membrane vesicles from tomato cv. Bonnie Best suspension-
cultured cells by Western blot analysis with the LeAGP-1
antibody. Third, fluorescence microscopy of plasmolysed, transgenic tobacco (Nicotiana tabacum BY-2) suspension-cultured
cells expressing a green fluorescent protein (GFP)-LeAGP-1
fusion product demonstrated localization to the plasma membrane and Hechtian threads. Fourth, the GFP-LeAGP-1 fusion
protein was present in plasma membrane preparations from
these transgenic tobacco cells by Western blot analysis with a
GFP antibody. Fifth, GFP-LeAGP-1 secreted into the culture
media contained ethanolamine, presumably attached to the
C-terminal amino acid residue, consistent with its processing
and release from the plasma membrane. Thus, these data support the hypothesis that LeAGP-1 is localized to the plasma
membrane via a GPI anchor and suggest possible roles for
LeAGP-1 in cellular signalling and matrix remodelling.
Introduction
Arabinogalactan proteins (AGPs) are a class of
hydroxyproline-rich glycoproteins and form a family of
structurally complex proteoglycans that are widely distributed throughout the plant kingdom from bryophytes
to angiosperms (Fincher et al. 1983, Showalter 1993,
Nothnagel 1997). AGPs occur in all organs and tissues,
predominantly in cell walls, plasma membranes (PMs),
and intercellular spaces (Nothnagel 1997). The use of
monoclonal antibodies to carbohydrate epitopes on AGPs
and b-Yariv phenylglycosides provide evidence that some
AGPs function in signalling and extracellular matrix
interactions during plant growth and development, cell
differentiation, and somatic embryogenesis (Nothnagel
1997, Showalter 2001).
AGPs can be grouped into two broad classes designated ‘classical’ and ‘non-classical’ (Mau et al. 1995, Du
et al. 1996). Classical AGPs consist of three distinct
domains: an N-terminal signal sequence, a Hyp-, Ala-,
Ser-, and Thr-rich domain, and a C-terminal hydrophobic
domain. The C-terminal hydrophobic domain is presumably proteolytically removed and replaced by a glycosylphosphatidylinositol (GPI) lipid anchor, allowing for
attachment to the PM. Indeed, the C-terminal hydrophobic domains predicted from cDNAs encoding two
Abbreviations – AGPs, arabinogalactan proteins; ELISAs, enzyme-linked immunosorbant assays; ESI-MS, electrospray ionization mass
spectrometer; GFP, green fluorescent protein; GPI, glycosylphosphatidylinositol; LC-MS, liquid chromatography- mass spectrometer;
PI-PLC, phosphatidylinositol-specific phospholipase C; PM, plasma membrane.
Physiol. Plant. 120, 2004
319
classical AGPs (NaAGP1 and PcAGP1) were absent in
the mature proteins and replaced with a moiety containing ethanolamine, inositol, glucosamine and mannose
(Youl et al. 1998). Oxley and Bacic (1999) went on to
determine the detailed structure of this GPI anchor in
PM-bound pear AGPs. They found that this structure
was similar to GPI anchors found in animals, protozoa,
and yeast and that it contained a ceramide lipid (Fig. 1).
In addition, rose AGPs were shown to carry a ceramide
class GPI lipid anchor based upon a combination of
techniques including reverse phase HPLC, treatment
with phosphatidylinositol-specific phospholipase C (PIPLC), and chemical structural analysis (Svetek et al. 1999).
Using two-dimensional SDS-PAGE, Western blotting
with AGP antibodies and PI-PLC, Sherrier et al. (1999)
also demonstrated the presence of PM-bound AGPs in
Arabidopsis. Recent analysis of the Arabidopsis genome
for putative AGPs show that there are a large number of
putative AGPs (approximately 47), many of which are
predicted to be GPI anchored (approximately 35) (Schultz
et al. 2002). Similarly, analysis of the Arabidopsis genome
for putative GPI-anchored proteins showed that there are
a large number of putative-GPI anchored proteins
(approximately 210), 40% of which are predicted to be
AGPs or to have AGP modules within them (Borner et al.
2002).
LeAGP-1 is a major AGP in tomato cell cultures and
plants, and represents one of the best-characterized
AGPs to date. Its cDNA and genomic sequences are
known and predict a classical AGP with four distinct
regions: an N-terminal signal sequence for secretion, a
central hydroxyproline/proline-rich region interrupted
by a short lysine-rich basic region, and a hydrophobic
C-terminal sequence identified as a putative GPI-anchor
addition sequence (Pogson and Davies 1995, Li and
Showalter 1996). The pattern of LeAGP-1 expression in
cultured cells and plants on the mRNA level as well as
the protein level is known, the later analysis being facilitated with a LeAGP-1 specific antibody (Li and Showalter
1996, Gao et al. 1999, Gao and Showalter 2000, Lu et al.
2001). The LeAGP-1 core protein was purified following
deglycosylation and characterized biochemically in terms
of amino acid composition, amino acid sequence, and
SDS-PAGE/Western blot analysis (Gao et al. 1999).
Most recently, native (i.e. glycosylated) LeAGP-1 was
purified and its carbohydrate moiety analysed (Zhao
et al. 2002). Based on this information, LeAGP-1 is
hypothesized to be a classic AGP localized to the PM
via a GPI anchor (Schultz et al. 1998, Gao et al. 1999).
However, most of the biochemical characterization work
to date on LeAGP-1, and indeed on most other AGPs,
involves isolation from cell culture media where AGPs
accumulate in large quantities, and not from PMs, which
represent a more limited, transient, and technically
challenging site of accumulation. Here we present several
experiments that support the hypothesis that LeAGP-1 is
localized to the PM via a glycosylphosphatidylinositol
(GPI) anchor and discuss the functional implications for
such a scenario.
Materials and methods
Isolation of tomato protoplasts and immunolocalization of
LeAGP-1
Cotyledons from 10-day-old tomato (Lycopersicon
esculentum, cv. UC82B) seedlings were collected, cut
into pieces, and incubated with 1% (w/v) cellulase
(Sigma, St. Louis, MO, USA) and 0.5% (w/v) pectinase
(Sigma) on an orbital shaker for at least 4 h. Protoplasts
were tracked with 0.01% (w/v) Fluorescent Brightener 28
(Calcofluor white M2R) (Sigma). Protoplasts were collected and fixed in 2% (w/v) paraformaldehyde and 1%
glutaraldehyde (v/v) in 50 mM citrate-phosphate buffer
(pH 7.4) for 2 h at 4 C, and then mounted onto poly-Llysine covered slides for immunolocalization of LeAGP-1
at the light microscope level using the PAP (i.e. LeAGP-1)
antibody and FITC-conjugated sheep anti-rabbit IgG
antibody as described previously (Gao et al. 1999).
Transgenic tobacco cell lines expressing GFP-LeAGP-1
fusion proteins and fluorescence microscopy
Transgenic tobacco (Nicotiana tabacum BY-2) cell lines
expressing tomato GFP-LeAGP-1 fusion protein and a
variant fusion protein lacking the C-terminal hydrophobic region of LeAGP-1, named GFP-LeAGP-1DC, were
produced as described previously and utilized in this
Fig. 1. GPI anchor structure for plasma membrane-bound AGPs. This chemical structure was determined for a pear PM-bound AGP (Oxley and
Bacic 1999). The phosphoceramide lipid is embedded in the outer leaflet of the PM and is attached to the core oligosaccharide, which is in turn
attached to the C-terminal amino acid residue (i.e. the o residue) of the AGP. The core oligosaccharide may contain a partial b-galactosyl
substitution (*) on the core oligosaccharide. Potential sites of cleavage by phosphatidylinositol-specific phospholipase C (PI-PLC) and D (PI-PLD)
are also indicated. Note that this structure is not drawn to scale. Et, ethanolamine; Man, mannose; Gal, galactose; GlcN, N-acetylglucosamine.
320
Physiol. Plant. 120, 2004
study (Zhao et al. 2002). Specifically, an enhanced version of GFP (i.e. EGFP) was used in these studies (Clontech, Palo Alto, CA, USA). In order to visualize the
fusion proteins in transgenic suspension-cultured cells,
fluorescence microscopy was performed using a Molecular Dynamics Sarastro 2000 confocal laser-scanning
fluorescence microscope equipped with a FITC filter set
consisting of a 488-nm laser wave length filter, a 510-nm
primary beam splitter and a 510-nm barrier filter. Data
were captured by Molecular Dynamics Image Space
software. In some experiments, the tobacco cells were
plasmolysed in 0–4% (w/v) NaCl solutions for 10 min
before microscopic examination.
Isolation of plasma membranes from tomato and
transgenic tobacco suspension-cultured cells
PM vesicles were isolated and purified following the
procedures of Komalavilas et al. (1991) with minor modifications. Tomato cv. Bonnie Best cells and transgenic
tobacco cells expressing GFP-LeAGP-1 or GFPLeAGP-1DC were harvested for PM isolation 10 days
after subculture. Cells were washed thoroughly with distilled water, and homogenized with a model PTA 20S
Polytron (Brinkmann, Westburg, NY, USA) using three
20-s bursts in a homogenizing buffer consisting of 50
mM Tris-HCl, 10 mM KCl, 1 mM EDTA, 0.1 mM
MgCl2 and 8% (w/v) sucrose and 1 mM phenylmethylsulphonylfluoride (pH 7.5). Cell debris was removed by
filtering and centrifugation, and the supernatant was
ultracentrifuged for 45 min at 100 000 g to collect microsomal membranes. Pellets were suspended in a buffer
(0.33 M sucrose, 3 mM KCl, and 5 mM potassium
phosphate [pH 7.8]). PM vesicles were purified by
aqueous two-phase partitioning (Larsson et al. 1987,
Komalavilas et al. 1991).
Isolation of AGPs from PM vesicles
AGPs were extracted from PM vesicles by incubating
with 1% (w/w) Triton X-100 overnight at 4 C. The
detergent-treated PM vesicles were centrifuged for 1 h
at 100 000 g at 4 C. Supernatants were incubated with
(b-D-galactosyl)3 Yariv reagent to precipitate and purify
AGPs as previously described (Gao et al. 1999). Alternatively, absolute ethanol was added to the supernatant
to a final concentration of 80% and incubated overnight
at 4 C. The precipitating AGPs were collected by centrifugation, washed twice with ethanol, and then air-dried.
Anhydrous hydrogen fluoride (HF) deglycosylation of
AGPs and Reverse phase (RP)-HPLC of deglycosylated
AGPs
Anhydrous HF deglycosylation was carried out as previously described by Kieliszewski et al. (1994). Briefly,
anhydrous HF was added to completely dried tomato
PM-AGPs (PM-AGPs) at a concentration of 20 mg ml1
AGP and incubated for 1 h at 4 C, with occasional
Physiol. Plant. 120, 2004
agitation. The reaction was quenched by the addition
of ice-cold distilled water to a final HF concentration
of 10%, dialysed against distilled water, and lyophilized.
Deglycosylated AGPs (0.5 mg) were dissolved in a small
amount of dH2O and spun briefly at 12 000 g. The supernatant was subjected to reverse-phase HPLC by injection
into a Hamilton PRP-1 column, using gradient elution
with 0.1% trifluoroacetic acid (TFA) and 80% acetonitrile in 0.1% TFA (Gao et al. 1999, Zhao et al. 2002).
Western blot and enzyme-linked immunosorbent assays
(ELISAs)
Tomato PM-AGPs were examined with a LeAGP-1 specific antibody (PAP antibody), while PM-AGPs from
transgenic tobacco cell lines were examined using an
anti-GFP antibody (Clontech, Palo Alto, CA, USA) in
order to detect GFP-LeAGP-1 fusion proteins. Quantification of proteins was accomplished with a Bio-Rad
DC protein assay kit II (Bio-Rad, Hercules, CA, USA).
Western blotting was conducted as described previously
(Gao et al. 1999). ELISAs were performed according to a
standard protocol using alkaline phosphatase conjugated
goat anti-rabbit IgG (Sigma) as the secondary antibody
(Harlow and Lane 1988).
Ethanolamine analysis of GFP-LeAGP-1 and
GFP-LeAGP-1DC
GFP-LeAGP-1 and GFP-LeAGP-1DC were purified
from transgenic tobacco cell culture media and treated
with chymotrypsin to remove GFP as previously
described (Zhao et al. 2002). The resulting LeAGP-1
and LeAGP-1DC glycoproteins were deglycosylated
with HF as described above and then hydrolysed in 6 N
HCl under vacuum at 110 C for 18 h. Samples were dried
under vacuum, washed with ethanol, water, triethylamine (2:2:1), and redried under vacuum. Samples, as
well as an ethanolamine standard, were derivatized with
AQC and analysed by reverse phase HPLC (Waters
AccQ-Tag C18 column, 150 4.6 mm i.d) using the gradient recommended by Waters for analysing collagen
hydrolysates (Crimmins and Cherian 1997, Van Wandelen
and Cohen 1997). Fluorescence was monitored by
flow-through detection using a Hewlett-Packard 1100
series fluorometer (excitation ¼ 250 nm; emission ¼
395 nm). The reversed phase peak corresponding to
AQC-derivatized ethanolamine (34 min elution time)
and a 34 min peak from the AQC-derivatized LeAGP-1
hydrolysate were collected for further analysis by electrospray ionization mass spectrometry. Although the control (LeAGP-1DC) lacked a fluorescent peak at 34 min
(indicating the control lacked ethanolamine), the eluant
from the column at 34 min was nevertheless collected for
further analysis. The 34 min Waters AccQ-Tag reversed
phase fractions described above were injected into a
Vydac reverse phase C18 column (1 50 mm i.d) at a
flow rate of 0.75 ml min1. Buffer A was 0.02% (v/v)
TFA in water. Buffer B was 95% (v/v) aqueous
321
acetonitrile. The gradient went from 1 to 15% B in 15 min,
then 15–50% B in 10 min. The Vydac column effluent was
fed directly into an electrospray ionization mass spectrometer (ESI-MS). Liquid chromatography-MS (LC-MS)
spectra were acquired using electrospray ionization on a
Bruker Esquire ion trap instrument operating in the
positive ion mode. The electrospray capillary was set at
4 kV. The sample flow was nebulized by nitrogen gas at
40 psi and dried using a countercurrent flow of nitrogen
at 7 l min1 and 300 C.
Upon request, all novel materials described in this
publication will be made available in a timely manner
for non-commercial research purposes.
Results
Immunolocalization of LeAGP-1 on tomato plasma
membranes
Protoplasts were isolated from tomato cotyledons using
cellulase and pectinase. The PAP antibody (a LeAGP-1specific antibody) and preimmune serum (as a control)
were used to immunolocalize LeAGP-1 to the PM of
these protoplasts (Fig. 2A). Pre-immune controls showed
no background immunofluorescence (Fig. 2B).
LeAGP-1 present in tomato plasma membrane vesicles
To further verify the presence of LeAGP-1 at the PM,
biochemical separation strategies were used for purification of LeAGP-1 from tomato PM vesicles. Tomato
suspension-cultured cells were collected and homogenized in a homogenizing buffer. Cell debris was separated
from microsomal membranes through filtering and
centrifugation. PMs were purified from other microsomal membranes by aqueous two-phase partitioning.
AGPs were fully extracted from PM with 1% Triton X100 overnight and precipitated with (b-D-galactosyl)3
Yariv reagent. PM-AGPs were deglycosylated with
anhydrous hydrogen fluoride, and the resulting protein
backbones were fractionated by RP-HPLC (Fig. 3A).
A
A
B
Fig. 2. Immunolocalization of LeAGP-1 to the plasma membrane
of a tomato cotyledon protoplast. (A) Tomato protoplast treated
with the LeAGP-1 antibody/FITC-labelled secondary antibody. (B)
Tomato protoplast treated with preimmune serum/FITC-labelled
secondary antibody as a control. Bars ¼ 10 mm.
Two major peak fractions (PM-1 and PM-3) and three
minor peak fractions (PM-2, PM-4, and PM-5) were
collected and subjected to further analysis. Analysis of
the five peak fractions using the LeAGP-1 antibody by
ELISAs (data not shown) and by Western blotting
(Fig. 3B) demonstrated that LeAGP-1 is only present in
the PM-3 fraction as a deglycosylated protein with an
apparent molecular weight of 48 kDa.
Fluorescence microscopy of transgenic tobacco cells
expressing GFP-LeAGP-1 and GFP-LeAGP-1DC fusion
proteins
Tomato LeAGP-1 gene constructions with and without
the C-terminal hydrophobic region were expressed as
fusion proteins with GFP in transgenic tobacco suspensioncultured cells as previously described (Zhao et al. 2002).
Preliminary results indicated that GFP-LeAGP-1 was
present on the cell surface of these transgenic tobacco
cells and in PMs and Hechtian strands of plasmolysed
cells. This was substantiated and examined in more detail
using the two cell lines, respectively, expressing GFPLeAGP-1 and GFP-LeAGP-1DC fusion proteins and
treating them with various concentrations of NaCl
B
Fig. 3. Biochemical separation and
identification of LeAGP-1 from
plasma membrane vesicles isolated
from tomato suspension-cultured
cells. (A) Reverse-phase HPLC
separation of (b-D-galactosyl)3
Yariv reagent-precipitable, HFdeglycosylated AGPs from tomato
plasma membrane vesicles. (B)
Western blot analysis of HPLC
peak PM-3 using the LeAGP-1
antibody/alkaline phosphataselinked secondary antibody.
Molecular masses (kDa) of the
protein size standards are shown
along with the calculated
molecular mass of the reacting
protein (i.e. deglycosylated
LeAGP-1).
322
Physiol. Plant. 120, 2004
ranging from 0 to 4% (w/v) (Fig. 4). Cells expressing
GFP-LeAGP-1 and not treated with NaCl (0% NaCl)
demonstrated green fluorescence at the cell surface. With
increasing percentages of salt treatment, plasmolysis was
observed and green fluorescence was increasingly
observed at the PM and in Hechtian strands, which are
connections between the PM and cell wall (Fig. 4B-F).
Cells expressing GFP-LeAGP-1DC also demonstrated
cell surface fluorescence; however, this fluorescence was
uniformly distributed in the cell wall–PM interface and
did not clearly discern the PM or Hechtian threads
(Fig. 4G). Moreover, cells expressing GFP-LeAGP-1DC
consistently secreted 6–10 times more fusion protein (i.e.
4.9–6.3 mg l1) into the culture media than cells expressing GFP-LeAGP-1 (0.46–1.1 mg l1).
GFP-LeAGP-1 present in plasma membrane vesicles of
transgenic tobacco cells expressing GFP-LeAGP-1, but
not GFP-LeAGP-1DC
Since the C-terminal hydrophobic domain of LeAGP-1
contains a putative GPI anchor addition site, and
transgenic GFP-LeAGP-1DC is apparently not tightly
associated with the PM, it was hypothesized that GFPLeAGP-1 should be anchored to the PM, while GFPLeAGP-1DC should not. In order to test this hypothesis,
A
GFP-LeAGP-1:
GFP-LeAGP-1∆C:
Signal peptide domain
GFP domain
Lys-rich domain
C-terminal hydrophobic domain
B
C
PM/CW
D
F
Physiol. Plant. 120, 2004
E
PM
H
G
“Classical” AGP domain
Fig. 4. Domain organization and
expression of GFP-LeAGP-1
fusion constructs in transgenic
tobacco cell suspension cultures.
(A) Organization of the GFPLeAGP-1 and GFP-LeAGP-1DC
fusion proteins expressed in
transgenic tobacco cells. GFPLeAGP-1 consists of the
N-terminal LeAGP-1 signal
peptide, an inserted GFP domain,
a central classical AGP domain
sandwiching a highly basic Lysrich domain, and a C-terminal
hydrophobic domain; GFPLeAGP-1DC has the same
organization except the
C-terminal hydrophobic domain,
presumably essential for GPIanchor addition, is absent. (B-G)
Detection of GFP fluorescence in
tobacco cells expressing GFPLeAGP-1 following treatment
with culture media supplemented
with 0% NaCl (B), 1% NaCl (C),
2% NaCl (D), 3% NaCl (E), and
4% NaCl (F) for 10 min.
Detection of GFP fluorescence in
tobacco cells expressing GFPLeAGP-1DC following treatment
with culture media supplemented
with 4% NaCl (G). Hechtian
strands are clearly visible in
plasmolysed tobacco cells
expressing GFP-LeAGP-1,
particularly following treatment
with 2–4% NaCl. PM, plasma
membrane; CW, cell wall; H,
Hechtian threads. Bars ¼ 10 mm.
323
total AGPs were isolated from PM vesicles of transgenic
tobacco cells expressing GFP-LeAGP-1 and GFPLeAGP-1DC. Western blotting analysis using a GFP
antibody indicated that GFP-LeAGP-1 was indeed present in these PM preparations, while GFP-LeAGP-1DC
was not (Fig. 5A). Moreover, some of the 28 kDa GFP
was apparently cleaved from the higher molecular weight
fusion protein. Such cleavage can largely be attributed to
heating the protein samples prior to gel loading, since
GFP-LeAGP-1 samples loaded on a gel without heating
had considerably less of the 28 kDa GFP protein released
from the fusion protein compared to that of the heated
sample (Fig. 5B).
Ethanolamine analysis of the GFP-LeAGP-1 fusion
protein
In order to provide evidence for a GPI anchor on
LeAGP-1, GFP-LeAGP-1 was purified from transgenic
tobacco cell culture media, as was GFP-LeAGP-1DC, and
analysed for the presence of ethanolamine, a known
substituent attached to the C-terminal residue of GPIanchored proteins even after their secretion. Indeed,
reversed phase fractionation of 6-aminoquinolyl-Nhydroxysuccinimidylcarbamate (AQC)-derivatized LeAGP-1
hydrolysate yielded a small peak that eluted at 34 min,
coincident with AQC-derivatized ethanolamine standard,
whereas the control, LeAGP-1DC, did not (Fig. 6).
Further fractionation of the 34 min peak on a Vydac
column yielded a peak eluting at 7 min. The ethanolamine
A
Discussion
The association of AGPs with the PMs of plant cells is
well documented in the scientific literature; however,
only in the past 5 years have researchers elucidated the
nature of this association. Certain AGPs are now known
to be attached to the PM via a GPI lipid anchor (Fig. 1).
Since GPI anchors require a C-terminal consensus
sequence for their addition, data mining of gene
sequences can be performed to reveal a large number of
putative GPI-anchored AGP sequences. Using two complementary approaches, two research groups performed
such analyses. Specifically, Borner et al. (2002) used a
search algorithm to first identify 210 putative GPI
anchored proteins in the Arabidopsis genome and then
found that 40% of these proteins are predicted to be
AGPs or include AGP modules. In contrast, Schultz
B
Fig. 5. Identification of GFP-LeAGP-1 from plasma membrane
vesicles isolated from transgenic tobacco suspension-cultured cells.
(A) SDS-PAGE/Western blot analysis of plasma membrane (PM)
proteins from transgenic tobacco cell cultures expressing GFPLeAGP-1 or GFP-LeAGP-1DC fusion proteins using a GFP
antibody/FITC-labelled secondary antibody. Lanes 1 and 3: 10 mg
and 5 mg total PM protein from transgenic tobacco cells expressing
GFP-LeAGP-1; lanes 2 and 4: 10 mg and 5 mg total PM protein from
transgenic tobacco cells expressing GFP-LeAGP-1DC. B. SDSPAGE/Western blot analysis of GFP-LeAGP-1 treated with (1) or
without (–) heat (100 C for 5 min) prior to gel loading. The GFP
antibody/alkaline phosphatase-labelled secondary antibody was
used for detection.
324
standard also eluted from the Vydac column at 7 min. The
ethanolamine standard, which eluted at 7 min on the
Vydac column, had a molecular mass of 231 Da, as did
the 7 min Vydac peak from LeAGP-1.
Fig. 6. Identification of AQC-derived ethanolamine (EA-AQC)
prepared from (A) an ethanolamine standard, or from (B) the
acid hydrolysate of HF-deglycosylated LeAGP-1. Samples were
fractionated by reverse phase column chromatography and the
peaks eluting at 34 min were collected for atomic mass
determination by LC-MS. The additional peaks occurring in
panels A and B correspond to the remaining AQC-derived amino
acids in the LeAGP-1 hydrolysate (B) and by-products of the AQC
derivatization reaction (A and B). The fluorescence (LU,
luminescence units) was measured using 250 nm and 395 nm as the
excitation and emission wavelengths, respectively.
Physiol. Plant. 120, 2004
et al. (2002) used a search algorithm to first identify 47
putative AGP genes from the Arabidopsis genome and
subsequently predicted that 35 of these genes encode
GPI-anchored products. These complementary data
mining studies, coupled with other biochemical analyses
of PM bound AGPs, indicate that AGPs are encoded by
a large gene family with a substantial number of these
AGPs predicted to be GPI-anchored to the outer leaflet
of the PM where they likely constitute a major surface
decoration in the form of arabinogalactan glycomodules.
While several AGPs are predicted to be GPI anchored
to the PM, only a few AGPs are actually known to
contain this anchor based on biochemical analysis
(Youl et al. 1998, Oxley and Bacic 1999, Sherrier et al.
1999, Svetek et al. 1999). LeAGP-1, a major AGP present in tomato cultured cells and plants, is hypothesized
to be GPI-anchored to the PM based on the presence of
a C-terminal consensus sequence for GPI addition, as
well as immunolocalization at the cell surface, and here
immunochemical and biochemical evidence is presented
to support this hypothesis. This work was undertaken,
not simply to support this hypothesis, but to provide
additional information on LeAGP-1 to supplement the
wealth of structural information on this novel AGP with
its characteristic Lys-rich subdomain and thereby facilitate its functional identification.
Five lines of experimental evidence are presented here
in support of this hypothesis. First, LeAGP-1 can be
immunolocalized to the PM in tomato cotyledon protoplasts (Fig. 2). Years earlier, Larkin (1978), working with
several plant species, demonstrated that AGPs are at the
surface of protoplasts by treating protoplasts with Yariv
reagent and causing their agglutination. Similarly, antibodies developed against carbohydrate epitopes on AGPs
also show PM localization, but such epitopes are shared
by multiple AGPs (reviewed in Showalter 2001). In contrast to Yariv reagent and AGP antibodies directed
against carbohydrate epitopes, which react with multiple
AGPs, the LeAGP-1 antibody used here is highly selective and allows for specific identification of LeAGP-1
(Gao et al. 1999).
Second, LeAGP-1 is present in PM vesicles produced
from cultured tomato cells (Fig. 3). This evidence was
ascertained by examining deglycosylated AGPs (i.e. the
AGP core proteins) obtained from these vesicles, again
using the LeAGP-1 antibody. Notably, other AGPs are
apparently present in these vesicles, and furthermore the
HPLC profile of the eluting core proteins is similar to
that previously observed for the culture media (Gao et al.
1999). Thus, as many as five different AGP core proteins
may be present with LeAGP-1 (peak PM-3) and the
AGP represented in PM-1 being the most abundant.
The LeAGP-1 core protein has an apparent molecular
weight of 48 kDa, again identical to that previously
observed for the LeAGP-1 core protein isolated from
culture media (Gao et al. 1999). It should be noted,
however, that the LeAGP-1 gene predicted a core protein
(sans signal and C-terminal hydrophobic peptides) with a
molecular weight of approximately 16 kDa. This indicated
Physiol. Plant. 120, 2004
either a crosslinked dimer/trimer or anomalous gel
migration of this Hyp-rich protein. Analysis of deglycoslyated LeAGP-1 by MALDI-TOF mass spectrometry
yielded a mass of approximately 16 kDa (unpublished
results) consistent with the later alternative.
Third, GFP-LeAGP-1 fusion proteins were expressed
in transgenic tobacco cells and demonstrated GFP fluorescence at the PM and its extensions (i.e. Hechtian
strands) following plasmolysis. While GFP fluorescence
is seen at the cell surface (i.e. at the CW/PM interface) in
turgid cells, it was only after plasmolysis that localization
to the PM and Hecht’s threads was clearly observed
(Fig. 4). The presence of LeAGP-1 in Hechtian strands
may indicate roles for this AGP in these poorly characterized entities, which serve as attachment sites between
the PM and CW. In contrast, the truncated version of
LeAGP-1, lacking the C-terminal hydrophobic domain,
did not clearly display PM or Hechtian thread fluorescence, but was still present in the CW/PM interface in
plasmolysed cells and indeed accumulated to much
higher levels in the culture media than the full length
LeAGP-1 fusion product. These data are consistent
with the prediction that the C-terminal domain contains
a consensus signal for GPI addition, which serves to
retain LeAGP-1 at the PM.
Fourth, the GFP-LeAGP-1 fusion protein is present in
PM vesicles produced from the transgenic tobacco
cultured cells (Fig. 5). However, the truncated version
of LeAGP-1 lacking the C-terminal hydrophobic domain
did not appear in these vesicle preparations, consistent
with the predicted role of the C-terminus in signalling
addition of a GPI membrane anchor and with the corresponding GFP localization data that failed to demonstrate clear PM fluorescence (Fig. 4G). Specifically, two
bands are observed in the Western blot analysis of PM
vesicle preparations using the GFP antibodies. One high
molecular weight band (approximately 98–120 kDa)
represents the glycosylated fusion protein. This product
was previously isolated and characterized from the culture media and demonstrated to be glycosylated with
large amounts of arabinose and galactose and lesser
amounts of glucuronic acid and rhamnose (Zhao et al.
2002). The broadness of this high molecular weight band
reflects the heterogeneous nature of core protein glycosylation. The smaller product was 28 kDa and represents
GFP which is cleaved from the fusion protein in block
upon heat treatment, perhaps through peptide bond
cleavage via an N!O acyl shift. (Xu et al. 1999).
Related and complementary to the above information,
Takos et al. (2000) fused a Clostridium thermocellum
endoglucanase E reporter gene (celE0 ) to the C-terminal
hydrophobic domain of LeAGP-1. They expressed these
constructs in tobacco protoplasts and demonstrated that
the C-terminal hydrophobic domain of LeAGP-1 directed
the addition of a GPI anchor to the endoglucanase
reporter protein based on sensitivity to PI-PLC digestion.
Fifth, ethanolamine was identified in the GFPLeAGP-1 fusion protein secreted into the transgenic
tobacco cell culture media. Ethanolamine is added to
325
the so-called C-terminal o residue of GPI-anchored
proteins during anchor addition and remains with the
protein even upon release from the PM (Udenfriend and
Kodukula 1995, Schultz et al. 1998). The identification
of ethanolamine indicates LeAGP-1 indeed contains a
GPI anchor. The truncated version of LeAGP-1 lacking
the C-terminal hydrophobic domain, however, did not
contain ethanolamine, consistent with its predicted inability
to carry out the C-terminal proteolytic processing required
for anchor addition.
It is now clear that LeAGP-1 is localized to the PM via
a GPI anchor and is subsequently processed for release
to the cell wall, extracellular space, and culture media.
Indeed, previous biochemical and immunolocalization
studies on LeAGP-1 in tomato as well as GFP-LeAGP-1
in transgenic tobacco have documented LeAGP-1 in
these extracellular sites of accumulation (Gao et al. 1999,
Gao and Showalter 2000, Zhao et al. 2002). In particular, it is worth noting that LeAGP-1 as well as other
tomato AGPs appear to be present in both the PM and
in the culture media based upon their similar HPLC core
protein profiles as mentioned above. Such processing
most likely involves the action of PI-PL C or D (Fig. 1).
Oxley and Bacic (1999) suggested that PI-PL D is likely
to operate, with or without prior PI-PL C action, since a
secreted pear AGP (PcAGP1) lacks the phosphoceramide moiety. While plants contain several phospholipases, only one report documents the existence of a
plant PL with the ability to cleave GPI anchors, namely
PI-PL C from peanut (Butikofer and Brodbeck 1993).
Several functional scenarios are envisioned for LeAGP-1,
which are in part based on the transgenic expression of
LeAGP-1 in tobacco. This transgenic work likely reflects
the natural situation in tomato, since tobacco and tomato
are close evolutionary relatives, both species express
LeAGP-1 (the orthologous gene in tobacco is called
NaAGP4), and transgenic LeAGP-1 is processed and
modified in tobacco similar to that of endogenous
LeAGP-1 in tomato (Gilson et al. 2001, Zhao et al.
2002). Thus, given that LeAGP-1 is GPI anchored and
released from the PM, the following functional scenarios
are envisioned for this AGP. First, LeAGP-1 may serve
as a marker of cellular identity. In this context, the
arabinogalactan polysaccharides that decorate noncontiguous Hyp, and the oligoarabinosides attached to
contiguous Hyp residues of LeAGP-1 would constitute
part of a glycocalyx containing information-rich, molecular surface markers (Zhao et al. 2002). The ability to
shed or turnover such markers is inherently associated
with GPI-anchored proteins, and could be regulated in
response to cellular, developmental, or environmental
cues where remodelling of the cell surface is required
(Nosjean et al. 1997). Second, LeAGP-1 may serve as a
cell membrane receptor or mediator of cellular signalling. Although LeAGP-1 lacks a cytoplasmic domain, it
may associate with other membrane proteins, such as
transmembrane proteins, which do contain cytoplasmic
domains capable of initiating an intracellular signalling
cascade. For example, LeAGP-1 may function analo326
gously to certain GPI-anchored heparan sulfate proteoglycans which serve to bind ligands for delivery to
a membrane receptor tied into a signalling network
(Schlessinger et al. 1995). It is also possible that the
GPI moiety of LeAGP-1 may serve as a signal itself
following release of LeAGP-1 from the PM, since
phosphatidyl inositol, inositol phosphoglycan and
ceramides, breakdown products of the GPI anchor,
are all known intracellular signalling molecules (Jones
and Varela-Nieto 1998). Third, LeAGP-1 may function
as a linker between the PM and cell wall. While
LeAGP-1 is clearly found in sites of PM-CW adhesion
(i.e. Hecht’s threads) (Fig. 4F), it is also distributed
throughout the PM. Thus, its distribution is not
restricted to Hecht’s threads, leaving unanswered the
questions of whether a punctuate distribution is
required for suspected PM-CW adhesion molecules
(Gens et al. 2000) and whether LeAGP-1 is simply
‘pulled’ into the Hecht’s strands by virtue of its apparent uniform distribution on the PM. Finally, these
and any other functional scenarios should also be considered in light of the possibility that LeAGP-1, as well
as other GPI-anchored proteins, may exist in lipid rafts.
GPI-anchored proteins in mammalian cells are essentially universally targeted to lipid rafts, sphingolipidand cholesterol-rich PM microdomains, where they
have proposed functions in polarized sorting, signal
transduction, and regulation of cell-surface hydrolytic
activity (Simon and Toomre 2000, Brown 2002). Testing
of the above and other functional scenarios for LeAGP-1
and its Arabidopsis homologues is underway in the
hope of elucidating the role of this well characterized,
abundant, GPI-anchored PM AGP in plant growth
and development.
Acknowledgements – This work was supported by grants from the
National Science Foundation (IBN-9727757 and IBN-0110413) to
A.M.S and M.J.K. The authors thank Dr Li Tan and Dr Jianfeng
Xu for technical assistance with HPLC and HF deglycosylation,
Mr Jeff Thuma for confocal scanning microscopy, and Dr Li-Wen
Wang and Mr Ming Chen for helpful advice.
References
Borner GHH, Sherrier DJ, Stevens TJ, Arkin IT, Dupree P (2002)
Prediction of glycosylphosphatidylinositol (GPI) -anchored
proteins in Arabidopsis: a genomic analysis. Plant Physiol 129:
486–499
Brown D (2002) Structure and function of membrane rafts. Int J
Med Microbiol 291: 433–477
Butikofer P, Brodbeck U (1993) Partial purification and characterization of a (glycosyl) inositolphospholipid-specific phospholipase C from peanut. J Biol Chem 268: 17794–17802
Crimmins DL, Cherian R (1997) Increasing the sensitivity of
6-aminoquinolyl-N-hydroxysuccinimidyl carbamate amino acid
analysis: a simple solution. Anal Biochem 244: 407–410
Du H, Simpson RJ, Clarke AE, Bacic A (1996) Molecular characterization of a stigma-specific gene encoding an arabinogalactanprotein (AGP) from Nicotiana alata. Plant J 9: 313–323
Fincher GB, Stone BA, Clarke AE (1983) Arabinogalactanproteins: structure, biosynthesis, and function. Ann Rev Plant
Physiol 34: 47–70
Gao M, Kieliszewski MJ, Lamport DTA, Showalter AM (1999)
Isolation, characterization and immunolocalization of a novel,
Physiol. Plant. 120, 2004
modular tomato arabinogalactan-protein corresponding to the
LeAGP-1 gene. Plant J 18: 43–55
Gao M, Showalter AM (2000) Immunolocalization of LeAGP-1, a
modular arabinogalactan-protein, reveals its developmentally
regulated expression in tomato. Planta 210: 865–874
Gens JS, Fujiti M, Pickard BG (2000) Arabinogalactan protein
and wall-associated kinase in a plasmalemmal reticulum with
specialized vertices. Protoplasma 212: 115–134
Gilson P, Gaspar YM, Oxley D, Youl JJ, Bacic A (2001) NaAGP4
is an arabinogalactan protein whose expression is suppressed by
wounding and fungal infection in Nicotiana alata. Protoplasma
215: 128–139
Harlow E, Lane D (1988) Antibodies: a Laboratory Manual. Cold
Spring Harbor Laboratory, New York, NY, pp. 564–565
Jones DR, Varela-Nieto I (1998) The role of glycosyl-phosphatidylinositol in signal transduction. Int J Biochem Cell Biol 30:
313–326
Kieliszewski MJ, Showalter AM, Leykam JF (1994) Potato lectin: a
modular protein sharing sequence similarity with the extensin
family, the hevein lectin family, and snake venom disintegrins
(platelet aggregation inhibitors). Plant J 5: 849–861
Komalavilas P, Zhu J-K, Nothnagel EA (1991) Arabinogalactanproteins from the suspension culture medium and plasma
membrane of rose cells. J Biol Chem 266: 15956–15965
Larkin PJ (1978) Plant protoplast agglutination by artificial carbohydrate antigens. J Cell Sci 30: 283–292
Larsson C, Widell S, Kjellbom P (1987) Preparation of high-purity
plasma membranes. Meth Enzymol 148: 558–568
Li S-X, Showalter AM (1996) Cloning and developmental/stressregulated expression of a gene encoding a tomato arabinogalactan protein. Plant Mol Biol 32: 641–652
Lu H, Chen M, Showalter AM (2001) Developmental expression and pertubation of arabinogalactan-proteins during
germination and seedling growth in tomato. Physiol Plant 112:
442–450
Mau SL, Chen CG, Pu ZY, Moritz RL, Simpson RJ, Bacic A,
Clarke AE (1995) Molecular cloning of cDNAs encoding the
protein backbones of arabinogalactan-proteins from the filtrate
of suspension-cultured cells of Pyrus communis and Nicotiana
alata. Plant J 8: 269–281
Nosjean O, Briolay A, Roux B (1997) Mammalian GPI proteins:
sorting, membrane residence and functions. Biochimica Biophysica Acta 1331: 153–186
Nothnagel EA (1997) Proteoglycans and related components in
plant cells. Int Rev Cytol 174: 195–291
Oxley D, Bacic A (1999) Structure of the glycosylphosphatidylinositol
anchor of an arabinogalactan protein from Pyrus communis
suspension-cultured cells. Proc Natl Acad Sci USA 96:
14246–14251
Pogson B, Davies C (1995) Characterization of a cDNA encoding
the protein moiety of a putative arabinogalactan protein from
Lycopersicon esculentum. Plant Mol Biol 28: 347–352
Schlessinger J, Lax I, Lemmon M (1995) Regulation of growth
factor activation by proteoglycans: what is the role of the low
affinity receptors? Cell 83: 357–360
Schultz C, Gilson P, Oxley D, Youl J, Bacic A (1998) GPI-anchors
on arabinogalactan-proteins: implications for signalling in
plants. Trends Plant Sci 3: 426–431
Schultz CJ, Rumsewicz MP, Johnson KL, Jones BJ, Gaspar YM,
Bacic A (2002) Using genomic resources to guide research directions. The arabinogalactan protein gene family as a test case.
Plant Physiol 129: 1448–1463
Sherrier DJ, Prime TA, Dupree P (1999) Glycosylphosphatidylinositol-anchored cell-surface proteins from Arabidopsis. Electrophoresis 20: 2027–2035
Showalter AM (1993) Structure and function of plant cell wall
proteins. Plant Cell 5: 9–23
Showalter AM (2001) Arabinogalactan proteins: structure, expression and function. Cell Mol Life Sci 58: 1399–1417
Simons K, Toomre D (2000) Lipid rafts and signal transduction.
Nature Rev: Mol Cell Biol 1: 31–41
Svetek J, Yadav MP, Nothnagel EA (1999) Presence of a glycosylphosphatidylinositol lipid anchor on rose arabinogalactan proteins.
J Biol Chem 274: 14724–14733
Takos AM, Dry IB, Soole KL (2000) Glycosyl-phosphatidylinositolanchor addition signals are processed in Nicotiana tabacum. Plant
J 21: 43–52
Udenfriend S, Kodukula K (1995) Prediction of o site in nascent
precursor of glycosylphosphatidylinositol protein. Meth Enzymol
250: 571–582
Van Wandelen C, Cohen SA (1997) Amino acid analysis of
unusual and complex samples based on 6-aminoquinolyl-Nhydroxysuccinimidyl carbamate derivatization. J Chromatogr
A763: 11–22
Xu Q, Buckley D, Guan C, Guo H-C (1999) Structural insights into
the mechanism of intramolecular proteolysis. Cell 98: 651–661
Youl JJ, Bacic A, Oxley D (1998) Arabinogalactan-proteins from
Nicotiana alata and Pyrus communis contain glycosylphosphatidylinositol membrane anchors. Proc Natl Acad Sci USA 95:
7921–7926
Zhao ZD, Tan L, Showalter AM, Lamport DTA, Kieliszewski MJ
(2002) Tomato LeAGP-1 arabinogalactan-protein purified from
transgenic tobacco corroborates the Hyp contiguity hypothesis.
Plant J 31: 431–444
Edited by D. Van Der Straeten
Physiol. Plant. 120, 2004
327