Molecular Plant • Volume 2 • Number 5 • Pages 1040–1050 • September 2009RESEARCH ARTICLE
Two Poplar Glycosyltransferase Genes, PdGATL1.1
and PdGATL1.2, Are Functional Orthologs to
PARVUS/AtGATL1 in Arabidopsis
Yingzhen Konga,b,2, Gongke Zhoua,b,2,3, Utku Avcia,b, Xiaogang Gua, Chelsea Jonesa, Yanbin Yinb,c,
Ying Xub,cand Michael G. Hahna,b,1
a Complex Carbohydrate Research Center, The University of Georgia, 315 Riverbend Road, Athens, GA 30602, USA
b BioEnergy Science Center, The University of Georgia, 315 Riverbend Road, Athens, GA 30602, USA
c Computational System Biology Lab, Dept. of Biochemistry and Molecular Biology, and Institute of Bioinformatics, The University of Georgia, Athens, GA
the biosynthesis of xylanin woodyplants, wherethis polysaccharide is a majorcomponent of wood, is poorlyunderstood.
Here, we characterize two Populus genes, PdGATL1.1 and PdGATL1.2, the closest orthologs to the Arabidopsis PARVUS/
GATL1 gene, with respectto their gene expression in poplar,their sub-cellular localization, andtheir ability to complement
the parvus mutation in Arabidopsis. Overexpression of the two poplar genes in the parvus mutant rescued most of the
defects caused by the parvus mutation, including morphological changes, collapsed xylem, and altered cell wall mono-
saccharide composition. Quantitative RT–PCR showed that PdGATL1.1 is expressed most strongly in developing xylem of
poplar. In contrast, PdGATL1.2 is expressed much more uniformly in leaf, shoot tip, cortex, phloem, and xylem, and the
transcript level of PdGATL1.2 is much lowerthan that of PdGATL1.1 inall tissues examined. Sub-cellular localization experi-
ments showed that these two proteins are localized to both ER and Golgi in comparison with marker proteins resident to
these sub-cellular compartments. Our data indicate that PdGATL1.1 and PdGATL1.2 are functional orthologs of PARVUS/
GATL1 and can play a role in xylan synthesis, but may also have role(s) in the synthesis of other wall polymers.
Several genes in Arabidopsis, including PARVUS/AtGATL1, have been implicated in xylan synthesis. However,
Key words: Arabidopsis thaliana; poplar; xylan; glycosyltransferase.
Xylans are polymers with a linear backbone composed entirely
of b-D-Xyl residues connected through (1/4)-linkages that
are partially acetylated and sometimes substituted with glu-
curonic acid and 4-O-methyl glucuronic acid (glucuronoxylan,
GX), arabinose (arabinoxylan), or a combination of acidic and
neutral sugars (glucuronoarabinoxylan). Glucuronoxylans are
mass, composing, for example, 23% of the dry weight of
poplar wood (Simson and Timell, 1978). An understanding
of GX biosynthesis has implications in economically important
industries including biofuel production, where optimization
of the plant cell wall composition to overcome biomass recal-
citrance is a major goal of research (Bevan and Franssen,2006).
GX has typically been viewed as a polysaccharide whose syn-
thesis requires a xylan synthase for backbone formation and
one or more glycosyltransferases for the addition of side
chains. A number of glycosyltransferases have been identified
that appear to be involved in xylan synthesis in Arabidopsis,
including IRX8/AtGAUT12 (Persson et al., 2007; Pen ˜a et al.,
2007), IRX9 (Pen ˜a et al., 2007), IRX7/FRA8 (Pen ˜a et al., 2007;
Brown et al., 2007), PARVUS/AtGATL1 (Brown et al., 2007;
Lee et al., 2007), IRX14 (Brown et al., 2007), IRX10, and
IRX10-L (Brown et al., 2007; Wu et al., 2009). Plants carrying
1To whom correspondence should be addressed. E-mail email@example.com,
fax +01 706 542 4412, tel. +01 706 542 4457
2These authors contributed equally to this work.
3Current address: Qingdao Institute of Bioenergy and Bioprocess Technol-
ogy, No.189 Songling Road, Laoshan District, Qingdao
Republic of China.
ª The Author 2009. Published by the Molecular Plant Shanghai Editorial
Office in association with Oxford University Press on behalf of CSPP and
IPPE, SIBS, CAS.
doi: 10.1093/mp/ssp068, Advance Access publication 24 August 2009
Received 10 June 2009; accepted 18 July 2009
and decreased xylan and xylose content. However, these gly-
cosyltransferases appear to be involved in different aspects
of the biosynthetic process. For example, IRX9, IRX14, IRX10,
and IRX10-L (Pen ˜a et al., 2007; Brown et al., 2007; Wu et al.,
2009; Brown et al., 2009) appear to be involved in elongation
of the xylan backbone, whereas IRX7, IRX8/AtGAUT12, and
PARVUS/AtGATL1 (Persson et al., 2007; Pen ˜a et al., 2007; Lee
et al., 2007) appear to be involved in the synthesis of a galac-
turonic acid-containing tetramer that is located at the reduc-
ing end of xylan. This reducing-terminal tetrasaccharide
appears to play an important role during xylan synthesis, al-
though it is not known whether this oligosaccharide acts as
a primer or a terminator (York and O’Neill, 2008). The exact
role(s) of each of these proteins in the synthesis of xylan remain
unclear, in part, due to the absence of functional in vitro assays
of enzyme activity. Two of these proteins, IRX8/AtGAUT12 and
PARVUS/AtGATL1, are related by sequence to a functionally
characterized galacturonosyltransferase (AtGAUT1) (Sterling
et al., 2006), suggesting that they might be involved in the syn-
thesis of the xylan-terminal tetrasaccharide via the addition of
an a-linked GalA residue to the growing tetrasaccharide (Pers-
son et al., 2007; Pen ˜a et al., 2007). Another possibility is that
these two proteins are involved in the synthesis of a structure
example, a specific pectic polysaccharide (Mohnen, 2008).
Despite these advances in Arabidopsis, little is known about
the genes involved in wood formation in trees, which contain
xylan as a major hemicellulosic component (Ebringerova ´ et al.,
2005). Populus trichocarpa has been fully sequenced and a to-
tal of 45 555 gene models have been predicted (Tuskan et al.,
2006). The completion of the P. trichocarpa genome sequence
provides an opportunity to advance our knowledge of wood
formation. However, the scarcity of loss-of-function mutants
complicates the studyof
Arabidopsis has been suggested as a model system for the
study of secondary growth because this herbaceous species,
under specific growing conditions, can be induced to develop
features that exhibit many of the characteristics common to
secondary growth in tree species (Chaffey et al., 2002; Ko
and Han, 2004).
Recent studies have shown that PoGT43B and PoGT47C, the
poplar orthologs of IRX9 and IRX7/FRA8, respectively, are able
to rescue the xylan defects of irx9 and irx7/fra8 mutants in Ara-
bidopsis (Zhou et al., 2006, 2007). These findings indicate that
PoGT43B and PoGT47C are likely to be involved in xylan syn-
thesis during wood formation. These results also established
the feasibility of using Arabidopsis as a model plant in which
to study the functions of poplar glycosyltransferases that par-
ticipate in wood formation.
In this study, we report molecular and genetic characteriza-
tion of two poplar genes, PdGATL1.1 and PdGATL1.2, that are
orthologous to the Arabidopsis PARVUS/AtGATL1 gene. These
two poplar genes are highly expressed in developing wood
(Aspeborg et al., 2005), and are specifically up-regulated in sec-
andSundberg,2008)anddown-regulatedduring tension wood
formation (Andersson-Gunnera ˚s et al., 2006), which indicates
thatthey may playrolesinwoodformation.However, the exact
function(s) of the proteins encoded by these two genes are still
unclear. So, inorder togainfurtherinsight intothefunctionsof
these two poplar genes, we examined their expression patterns
ing proteins, and their ability to complement the parvus/gatl1
mutant in Arabidopsis.
Gene Structure and Expression Profiles of Poplar GATL1.1
Two poplar genes named GATL1.1 and GATL1.2 were identi-
fied from thePopulus trichocarpa
(www.jgi.doe.gov/poplar) on the basis of their sequence sim-
ilarity to the Arabidopsis PARVUS/AtGATL1 gene (Figure 1A).
The corresponding genes, PdGATL1.1 and PdGATL1.2, were
then cloned and sequenced from Populus deltoides xylem-
derived cDNA. Except for the P. trichocarpa sequences used
for generation of the phylogenetic tree, all the other sequen-
from Populus deltoides. PdGATL1.1 encodes a protein of 360
amino acids and PdGATL1.2 encodes a protein of 353 amino
acids. Pair-wise comparisons of the amino acid sequences
showed that these two proteins are highly similar, having
93% sequence identity with each other. Further, these two
poplar proteins have 82 and 81% identity, respectively, at
the amino acid level with PARVUS/AtGATL1 (Figure 1B).
The expression profiles of the two genes were examined by
quantitative real-time PCR using primers that were specific to
each gene. Cortex, phloem, xylem, shoot tip, leaf, and root tis-
sues were harvested from young Populus deltoides trees
grown in a greenhouse. Ubquitin was used as an internal con-
trol. As indicated in Figure 2, PdGATL1.1 and PdGATL1.2 are
pression levels are very different. PdGATL1.1 is highly expressed
in xylem compared to other tissues. PdGATL1.2 expression is
highest in xylem, but overall transcript levels are more uniform
in all tissues examined, except root and PdGATL1.2 transcript
lem, the expression level of PdGATL1.1 is 26 times that of
PdGATL1.1 and PdGATL1.2 Are Targeted to the Secretory
Predictions about the sub-cellular localization of the PdGATL1
proteins were made by subjecting the PdGATL1 amino acid
sequences to analyses using publicly available bioinformatics
packages, including SOSUI, TMHMM 2.0, and PSORT (see
Methods). PdGATL1.2 was predicted to have no transmem-
brane domain by all programs used, and is predicted by PSORT
to have a cleavable N-terminal signal peptide that directs the
Kong et al.
dPoplar Orthologs to PARVUS/AtGATL1 | 1041
protein into the secretory pathway. For PdGATL1.1, no consen-
sus was found among the prediction programs used. PSORT
predicts the presence of a cleavable N-terminal signal peptide
could be an integral membrane protein with one transmem-
brane domain. Similarly, the PARVUS/AtGATL1 protein is pre-
dicted to have a signal peptide sequence at the N-terminus by
Figure 1. Phylogenetic Tree and Multiple Sequence Alignments Were Constructed from the Deduced Amino Acid Sequences of the 10-
Member Arabidopsis thaliana GATL Family and the Orthologous Proteins from Populus trichocarpa and P. deltoides.
(A) The protein sequences from A. thaliana and P. trichocarpa were aligned using MAFFT v6.603 (Katoh et al., 2005) and the resulting
alignment was used to perform maximum likelihood phylogeny reconstruction using PhyML v2.4.4 (Guindon and Gascuel, 2003). P. tricho-
carpa GATL protein sequences are identified by their NCBI RefSeq accessions (www.ncbi.nlm.nih.gov/RefSeq/).
(B) Amino acid sequence alignment of PtGATL1.1, PtGATL1.2, PdGATL1.1, PdGATL1.2, and PARVUS/AtGATL1. Gaps (marked with dashes)
were introduced to maximize the sequence alignment. Identical and similar amino acid residues are shaded with black and gray, respec-
1042 | Kong et al.
dPoplar Orthologs to PARVUS/AtGATL1
PSORT, and no transmembrane domain is predicted for this
protein by TMHMM 2.0. However, SOSUI predict that PAR-
VUS/AtGATL1 has a 23-amino acid transmembrane domain
that includes the N-terminal amino acids 1 to 23. This would
leaveno cytosolicN-terminaldomain,whichwould beunusual
for a type II membrane. So it seems most likely that both PAR-
domain, but rather just an N-terminal signal peptide directing
the protein into the endomembrane secretory system. Consid-
ering the high amino acid similarity of PdGATL1.1 with these
two proteins, it is highly likely that PdGATL1.1 also has no
The PARVUS/GATL1 protein was previously suggested to be
localized in endoplasmic reticulum (ER) based on the localiza-
tion of a heterologously expressed enhanced yellow fluores-
cence protein (EYFP)-tagged PARVUS/GATL1 recombinant
protein in carrot protoplasts (Lee et al., 2007). To investigate
whether PdGATL1.1 and PdGATL1.2 have the same sub-cellular
localization as PARVUS/GATL1, we fused PdGATL1.1 and
PdGATL1.2 with EYFP at the C-terminus and transformed
the recombinant protein into Nicotiana benthamiana leaves.
Confocal microscopy was used to determine the sub-cellular
localization of recombinant PdGATL1.1 and PdGATL1.2. As
shown in Figure 3B, 3E, 3H, and 3K, EYFP-tagged PdGATL1.1
and PdGATL1.2 showed both punctuate and network-like
localization patterns in tobacco leaf epidermal cells. Co-
localization experiments (Figure 3C, 3F, 3I, and 3L) revealed
that these patterns overlap with those of both Gmct-ECFP,
an enhanced cyan fluorescent protein (ECFP)-tagged Golgi
marker (Saint-Jore-Dupas et al., 2006; Nelson et al., 2007),
and ECFP-WAK2-HDEL, an ER marker (Nelson et al., 2007)
(Figure 3A, 3D, 3G, and 3J).
PdGATL1.1 and PdGATL1.2 Rescue Arabidopsis parvus/
atgatl1 Mutant Phenotypes
To determine whether PdGATL1.1 and PdGATL1.2 have the
same function as PARVUS/AtGATL1, we attempted to comple-
ment the Arabidopsis parvus/gatl1mutant with the full-length
poplar genes. The open reading frames of the two P. deltoides
genes were cloned from xylem-derived cDNA fused with the
cauliflower mosaic virus (CaMV) 35S promoter, and introduced
individually into a heterozygous parvus/gatl1 mutant line of
Arabidopsis (homozygous parvus/gatl1 plants have very low
fertility (Lao et al., 2003; Shao et al., 2004; Lee et al., 2007),
necessitating the use of the heterozygous line). Transgenic
lines were tested for the presence of the PdGATL1.1 and
Figure 2. Expression Analysis of the PdGATL1.1 and PdGATL1.2
Genes by Quantitative Real-Time PCR.
Relative expression levels in all samples were normalized using
ubiquitin as a constitutively expressed internal control and the
PdGATL1.2 expression levels in root are set to 1. Data are the aver-
ages 6 SE of three biological replicates.
Figure 3. Sub-Cellular Localization of EYFP Tagged PdGATL1.1 and
EYFP-tagged PdGATL1.1 and PdGATL1.2 were transiently expressed
in leaf epidermal cells of Nicotiana benthamiana plants, and their
microscope. Scale bars represent 20 lm.
(A–C) Tobacco leaf epidermal cells expressing both PdGATL1.1–
EYFP and ECFP–WAK2–HDEL constructs. (A) localization of ECFP–
WAK2–HDEL ER marker protein (green). (B) localization of
PdGATL1.1–EYFP protein (red) in the same cell as in (A). (C) merged
image of (A) and (B), showing co-localization of PdGATL1.1–EYFP
and Gmct–ECFP constructs. (D) Localization of Gmct–ECFP Golgi
(red) in the same cell as in (D). (F) Merged image of (D) and (E),
showing co-localization of PdGATL1.1–EYFP and Gmct–ECFP.
(G–I) Tobacco leaf epidermal cells expressing both PdGATL1.2–EYFP
and ECFP-WAK2-HDEL constructs. (G) Localization of ECFP–WAK2–
HDEL ER marker protein (green). (H) Localization of PdGATL1.2–
EYFP protein (red) in the same cell as in (G). (I) Merged image of
(G) and (H), showing co-localization of PdGATL1.2–EYFP and
(J–L) Tobacco leaf epidermal cells expressing both PdGATL1.2–EYFP
and Gmct–ECFP constructs. (J) Localization of Gmct–ECFP Golgi
marker protein (green). (K) Localization of PdGATL1.2–EYFP pro-
tein (red) in the same cell as in (J). (L) Merged image of (J) and
(H), showing co-localization of PdGATL1.2–EYFP and Gmct–ECFP.
Kong et al.
dPoplar Orthologs to PARVUS/AtGATL1 | 1043
PdGATL1.2 transgenes in a homozygous parvus/gatl1 back-
ground. Using poplar gene-specific primers for transgene am-
plification, poplarmRNA expression
Arabidopsis lines was confirmed by RT–PCR (Figure 4B and
4C). The absence of PARVUS/AtGATL1 gene expression in the
transgenic lines was also confirmed using primers specific to
the wild-type (w.t.) Arabidopsis gene (Figure 4B and 4C).
Homozygous parvus/gatl1 mutants show dark-green leaves,
reduced plant stature, reduced size of all organs, including
leaves, floral organs, and fruits, and reduced fertility (Lao
et al., 2003; Shao et al., 2004; Lee et al., 2007). Expression of
either PdGATL1.1 or PdGATL1.2 in the parvus/gatl1 mutant res-
cued the morphological defects in the mutant. Indeed, the
morphology of the complemented plants is indistinguishable
from that of w.t. Arabidopsis plants (Figure 4A).
The parvus/gatl1 mutant also exhibits collapsed xylem ves-
sels and thinner secondary cell walls, which are largely due to
defects in xylan synthesis (Brown et al., 2007; Lee et al., 2007).
In parvus/gatl1 mutants, the xylose content is decreased by
about 47% compared to w.t. (Lee et al., 2007). To demonstrate
whether the morphological complementation by PdGATL1.1
and PdGATL1.2 could be correlated with a rescue of xylan syn-
thesis, xylem morphology and xylan immunolocalization were
compositions of the transgenic plants were determined. As
shown in Figure 5, the shapes of xylem vessels in either the
PdGATL1.1 or PdGATL1.2 complemented plants were essen-
tially indistinguishable from those of xylem vessels in w.t.
pression of either PdGATL1.1 or PdGATL1.2 in a parvus/gatl1
background restored the level of xylose to 86 and 80% of
w.t. levels, respectively (Figure 6).
Immunolocalization of xylan using the xylan-directed
monoclonal antibodies LM10 and LM11 was done to further
investigate whether the increased xylose content and the
complemented phenotype in the transgenic plants correlate
with the rescue of xylan synthesis in secondary cell walls.
but not to arabinoxylan and glucuronoarabinoxylan, whereas
LM11 interacts with both 4-O-methylglucuronoxylan and
Figure 4. Restoration of Plant Size in Arabidopsis parvus Plants by
Overexpression of the Poplar PdGATL1.1 and PdGATL1.2 Genes,
The results shown are representative of six independent transgenic
Arabidopsis lines. All transgenic plants were confirmed to have
a homozygous parvus background by PCR detection of the
T-DNA insertion and absence of an endogenous PARVUS gene com-
pared with the w.t. (upper two panels in (B) and (C)).
(A) The parvus mutant (right) has a short inflorescence stem and
a small rosette size, and overexpression of PdGATL1.1 and
PdGATL1.2, respectively, in parvusplants(middle)restoredthestem
height and rosette size to those of the w.t. (left).
(B) PCR detection of the PdGATL1.1 transgene and its transcript in
the transgenic parvus plants. The expression of the ACTIN gene was
used as an internal control.
(C) PCR detection of the PdGATL1.2 transgene and its transcript in
the transgenic parvus plants. The expression of the ACTIN gene was
used as an internal control. WT, wild-type.
Figure 5. Restoration of Secondary Wall Thickness of Vessels in the
Transgenic Arabidopsis parvus Plants Overexpressing the Poplar
PdGATL1.1 and PdGATL1.2 Genes, Respectively.
Stems and hypocotyls of 8-week-old plants were sectioned (250 nm
thick) and stained with toluidine blue for examination of vessels.
Arrows indicate collapsed vessels. ve, vessel. Images for each tissue
are taken at the same magnification. Bars = 50 lm.
(A–D) Transverse sections taken from the base of stems of w.t.,
PdGATL1.1-complemented parvus, PdGATL1.2-complemented par-
vus, and parvus, respectively.
parvus, and parvus, respectively.
1044 | Kong et al.
dPoplar Orthologs to PARVUS/AtGATL1