Expression of recombinant proteins in tobacco system

Plant Cell Rep (2004) 23:81–90
DOI 10.1007/s00299-004-0792-0
GENETICS AND GENOMICS
D. Tamilselvi · G. Anand · S. Swarup
A geminivirus AYVV-derived shuttle vector for tobacco BY2 cells
Received: 21 January 2004 / Revised: 1 March 2004 / Accepted: 3 March 2004 / Published online: 7 April 2004
 Springer-Verlag 2004
Abstract We have developed a plant-Escherichia coli
pASV shuttle vector from the essential elements of the
Ageratum yellow vein virus (AYVV). The geminivirus
vector contains the AYVV genome with the coat-protein
deletion, the E. coli vector backbone of pUC19, a unique
cloning site and gene expression cassettes for plant se-
lection and reporter gene activity. The replication of
pASV vectors was compared in Nicotiana benthamiana
and N. tabacum BY2 cells, and the latter were found to be
suitable for long-term maintenance of the vectors in cul-
ture. The vector DNA was detected at regular intervals by
PCR, b-glucuronidase expression analysis and plasmid
rescue during a 4-month culture period. A novel meth-
ylation-based PCR assay was carried out to show de novo
replication for pASV-derived vectors in 2-month-old to-
bacco BY2 cell lines. This is the first report of the ex-
trachromosomal replication of monopartite begomovirus
with stability and foreign gene expression in long-term
cell cultures.
Keywords Ageratum yellow vein virus · Geminivirus ·
Shuttle vector · Replication · Foreign gene expression
Abbreviations ACMV: African cassava mosaic virus ·
AYVV: Ageratum yellow vein virus · CP: Coat protein ·
GUS: b-Glucuronidase · TGMV: Tomato golden mosaic

virus · Tobacco BY2: Tobacco L. cv. Bright Yellow 2
Introduction
Whole plants and plant cells are emerging as viable and
competitive expression systems for large-scale protein
production as a means of obtaining biologically active
and safe biopharmaceutical proteins at affordable prices.
Consequently, there is renewed interest in developing
novel vectors to express foreign proteins in these systems.
While the most widely used method for producing foreign
proteins is via stably transformed plants, plant cell cul-
tures provide an alternative. The main advantages of the
latter lie in the large-scale production of proteins in in-
dustrially sized bioreactors under sterile, defined, and
controllable conditions, all of which makes the plant cell
culture system amenable to standard biomanufacturing
practices. As a result of recent commercial interest, there
have been many advances in the areas of transgenic plants
(Fischer and Emans 2000; Ma et al. 2003), production in
suspension cultures (Fischer et al. 1999a) and the appli-
cation of plant vectors for foreign protein expression
(Fischer et al. 1999b). Vast choices of plant transforma-
tion vectors are currently available due to the early start
researchers have made in this field. Compared to the
stable transformation vectors based on T-DNA, however,
there are only a few reports on extrachromosomal shuttle
vectors for expressing foreign genes.
In plants, single-stranded DNA geminiviruses have
been used as potential sources of extrachromosomal
vector replicons to enable multiplication in the nuclei of
infected cells (reviewed by Davies and Stanley 1989;

Mullineaux et al. 1992; Stanley 1993; Timmermans et al.
1994). Geminiviruses have small single-stranded circular
DNA genomes of 2.5–3.0 kb. They are characterized by
twinned (geminate) icosahedral capsids and can replicate
using plant host machinery via double-stranded DNA
intermediates. The geminiviridae family has three sub-
groups, which can be distinguished based on their genetic
organization, plant host, and insect vector. The genera
Mastrevitesus and Curtovirus have a single genomic
component and infect monocot and dicot species, re-
spectively. Begomovirus members, on the other hand,
Communicated by P.P. Kumar
D. Tamilselvi · G. Anand · S. Swarup (
)
)
Department of Biological Sciences,
National University of Singapore,
Singapore, 117 543
e-mail:
Tel.: +65-6874-7933
Fax: +65-6779-2486
Present address:
G. Anand, Temasek Capital (Private) Limited,
Temasek Tower, Singapore, 068811
exclusively infect dicot species and have a genome
comprising two similar-sized DNA components (DNA A
and DNA B) (reviewed by Gutierrez 2000). DNA A en-
codes a replication-associated protein (Rep), coat-protein
(CP), and proteins that participate in the control of rep-
lication and gene expression. The DNA B component

encodes proteins required for nuclear trafficking and cell-
to-cell movement of the viral DNA. Relatively few be-
gomoviruses have been described to possess a monopar-
tite genome, which resembles DNA A. Those include the
tomato leaf curl virus (TLCV), Ageratum yellow vein
virus (AYVV), and cotton leaf curl virus (CLCuV) (Dry
et al. 1993; Tan et al. 1995; Briddon et al. 2000).
AYVV can replicate and form infectious symptoms in
Ageratum, tomato, French bean and Nicotiana ben-
thamiana (Tan et al. 1995). Transmission of AYVV is by
whiteflies (Tan et al. 1995), and its monopartite genome
has two overlapping virion-sense open reading frames
(ORFs), V1 and V2 that encode the CP and movement
protein, respectively. ORF C1 encodes the replication-
associated protein, ORFs C2 and C3 regulate virion-sense
gene expression and DNA replication, respectively,
ORF C4 is a pathogenicity determinant that may affect
the host cell, and there is an additional ORF C5 of un-
known function. The intergenic region (IR) contains the
initiation site of rolling circle DNA replication.
Irrespective of the monopartite or bipartite nature of
geminiviruses, only their intergenic and the complemen-
tary strand ORFs are necessary for replication (Lazaro-
witz et al. 1989; Kammann et al. 1991; Ugaki et al. 1991).
Hence, a popular strategy used in developing geminiviral
vector backbone is the replacement of the CP gene with
reporter genes. Such vectors have been developed using
monopartite viruses such as wheat dwarf virus (WDV),
maize streak virus (MSV) and bean yellow dwarf virus
(BeYDV) (Ward et al. 1988; Lazarowitz et al. 1989;

Topfer et al. 1989; Mor et al. 2003) and DNA A genomes
of bipartite virus TGMV (Kanevski et al. 1992). Various
vectors have used both the native CP promoter and
foreign promoters such as the cauliflower mosaic virus
(CaMV) 35S promoter to express foreign genes or re-
porter genes (Ugaki et al. 1991; Mor et al. 2003) for
transient expression. Stable maize and tobacco cell lines
containing the replicating viral episome for MSV or
TGMV, respectively, have also been reported (Kanevski
et al. 1992; Palmer et al. 1999). WDV, a monopartite
mastrevirus, replicon-based shuttle vectors with bacterial
replicons, namely ColE1 and p15A, have been used
(Kammann et al. 1991; Ugaki et al. 1991) for extrachro-
mosomal replication study. However, no shuttle vectors
have been developed and evaluated based on the
monopartite begomoviruses.
We report here the development and evaluation of a
plant-Escherichia coli shuttle vector using the replicon of
the monopartite begomovirus AYVV and a bacterial
replicon derived from pUC19 to ensure a high copy
number in E. coli. Two different methods, namely elec-
troporation and biolistic bombardment, were evaluated
for their efficiency to transform the plant cells. A novel
method based on the unique DNA methylation-specificity
of the plant and E. coli cells was used to study de novo
replication of the shuttle vector in the plant cells. The
successful rescue and maintenance of structural integrity
of the shuttle vector from plant cells into E. coli and the
expression of the reporter gene was demonstrated in 4-
month-old cultures of tobacco BY2 cell suspension cul-

tures.
Materials and methods
Construction of plasmids
Plasmid pASV82, a shuttle vector for Escherichia coli and tobacco
cells, was constructed based on the AYVV geminiviral and E. coli
pUC19 backbone. The genealogy of vector construction is detailed
in Fig. 1. A 2.7-kb AYVV DNA fragment was released from
pHN419 (Tan et al. 1995) following BamHI digestion. This frag-
ment was self-ligated to obtain a circular AYVV DNA template,
which was then used to amplify a 2.3-kb fragment using primers
extending away from the CP gene, thereby creating its deletion.
The CP sense primer (5
0
-AATTCGTACTCATGCCAG-TAATCC-
AGTGTATGC-3
0
) and Mlu antisense primer (5
0
-AATTCATTAC-
CACGCGTGACATCACTAACAC-3
0
) were used with the proof-
reading Vent DNA polymerase. PCR cycling parameters were 95C
for 1 min, 50C for 1 min and 72C for 2 min. The number of PCR
cycles was kept to a maximum of 20 to further minimize proof-
reading errors. EcoRI-compatible ends were generated using
T4 DNA polymerase and the fragment ligated to the unique EcoRI
site of the vector pNKA210.2, a derivative of pIBT210.1 (Haq et al.
1995). pNKA210.2 has a pUC19 (2.68 kb) backbone and a plant
gene expression cassette consisting of a 35S CaMV promoter,

5
0
UTR of the tobacco etch virus (TEV-5
0
UTR), a translational
enhancer, a replaceable stuffer fragment, and a vspB terminator
sequence. This shuttle vector was named pASV. A 1.68-kb neo-
mycin phosphotransferase II (NPTII) expression cassette from
pNGI (Klien et al. 1989) was ligated into the unique HindIII site of
pASV to generate pASVNPT. To facilitate cloning of the foreign
gene in this shuttle vector, we created a unique enzyme site. There
were two HindIII sites in pASVNPT. A single HindIII site was
generated in pASVNPT by destroying one of the two HindIII sites
by partial digestion, followed by blunt-ending and self-ligation to
yield pASV82. Foreign gene expression cassettes can be inserted in
pASV82 (8.2 kb) at its unique HindIII site. The constructs were
selected in E. coli using ampicillin (50 mg/ml) and in plant cells
using kanamycin (50 mg/ml). Plasmid pASVGUS, a derivative of
pASV82 (8.2 kb) containing the GUS expression cassette, was
constructed by inserting the 3-kb HindIII fragment of pRTL2-GUS
(Carrington et al. 1991) into the HindIII site of pASV82.
A replication-defective control plasmid, pASVDIR, was con-
structed by partial digestion of pASVNPT with BamHI, followed
by inverse PCR amplification using primers flanking the IR region
with a mixture of KlenTaq (Fermentas, Hanover, Md.) and Pfu
polymerase (Promega, Madison, Wis.). The primers IRD (5
0
-TA-
CTCTCCTGATACGATTGGGC-3
0

) and C1 (5
0
-AATTCCCAAA-
GTGCCATTCGG-3
0
) were used for PCR with the following pa-
rameters: 25 cycles of 1 min 30 s at 94C, 1 min at 55C, and 8 min
at 72C. The PCR product lacking the IR (pASVDIR) was first
blunted and then self-ligated to circularize it.
All plasmids were constructed and propagated in E. coli strain
DH5a and key junctions of the fragments sequenced after each
construction. Large-scale amplification and purification of plasmids
was performed by the alkaline lysis method followed by CsCl ul-
tracentrifugation (Sambrook et al. 1989) or purified with a Nucle-
obond DNA purification AX500 column (Clontech laboratories,
Palo Alto, Calif.) and compared.
82
Optimization of electroporation conditions
for N. benthamiana mesophyll-derived protoplasts
Protoplast isolation, electroporation medium and culture conditions
for stable transformation were as described by Sala et al. (1989). To
determine the optimum electroporation conditions for introducing
plasmid DNA into protoplasts, we used two electroporation de-
vices, namely T820, a square wave pulse generator (BTX, San
Diego, Calif.) and a Gene Pulser II exponential wave pulse gen-
erator (Bio-Rad, Hercules, Calif.). The optimization methods were
based on Trypan blue uptake and fluorescein diacetate staining to
determine viability according to Saunders et al. (1995) with minor
modifications. A suspension of 110
6

protoplasts in 600 mlof
HeNa/F buffer (10 mM HEPES, pH 7.1, 5 mM CaCl
2
, 150 mM
NaCl, 0.2 M mannitol) was used for electroporation in 4-mm gap
cuvettes obtained from the manufacturers of the two electropora-
tors. The electroporation of protoplasts with the BTX T820 was
optimized with a single pulse of 80 ms and with varied levels of
field strength under low and high voltage modes. The Bio-Rad
Gene Pulser II was operated with the capacitance set at 1,000 mF,
the resistance set at 100–200 W and the time constant at approxi-
mately 18–26 ms with varied field strength. Trypan blue uptake
was observed under bright field microscopy, and the FDA staining
was observed under UV fluorescence in a dark field with the in-
verted fluorescent Leitz Fluovert FU microscope (Leitz, Wetzlar,
Germany) equipped with a UV lamp.
Fig. 1 The Ageratum yellow
vein virus (AYVV) genome and
cloning strategy of the AYYV-
derived plant-Escherichia coli
shuttle vector. In the AYVV
genome, ORF V1 encodes coat-
protein, V2 encodes movement
protein, C1 encodes replication-
associated protein, C2 regulates
virion-sense gene expression,
C3 regulates DNA replication,
C4 encodes pathogenicity de-
terminant, C5 encodes an un-
known function. Exp cas1 Ex-

pression cassette 1, IR inter-
genic region. The recombinant
plasmids described in the Re-
sults are indicated in boxes. The
unique HindIII in pASV82 can
be used to clone expression
cassettes with the gene of in-
terest
83
Transformation of N. benthamiana mesophyll-derived protoplasts
with plasmid DNA and confirmation using PCR
For transformation using our optimized conditions, we generally
electroporated 50 mg of purified closed circular plasmid DNA into
110
6
protoplasts using the BTX electroporator. To confirm the
transformation, we extracted the DNA according to Townsend et al.
(1986) at various intervals during protoplast culture. Linear PCR
for the NPTII gene and NAD5 gene was carried out for 15 cycles
of 1 min at 96C, 1 min at 55C, and 1 min at 72C with primers
NPTfor (5
0
-GAAGGCGATAGAAGGCGA-3
0
) and NPTrev (5
0
-
GGGTGGAGAGGCTATTCGGC-3
0
). To amplify the NAD5 gene,

we used the PCR primers NAD5for (5
0
-TAGCCCGACCGTAGT-
GATGTTAA-3
0
) and NAD5rev (5
0
-ATCACCGAACCTGCACT-
CAGGAA-3
0
) with the following parameters: 15 cycles of 30 s at
96C, 1 min at 55C, 1 min at 72C.
Transformation of tobacco BY2 by particle bombardment
and PCR detection of transformed lines
Cells of N. tabacum L. cv. BY2 were maintained in MS medium
(Murashige and Skoog 1962) supplemented with 0.18 mg/l
K
2
HPO
4
, 100 mg/l myoinositol, 1 mg/l thiamine HCl, 0.5 mg/l
MES, 30 g/l sucrose in the dark at 120 rpm and 25C. A biolistic
particle gun (model PDS-1000/He; Bio-Rad) was used for biolistic
bombardment of the tobacco BY2 cells as described by Kikkert
(1993). Conditions were optimized for various rupture disks (900,
1,100, 1,300 psi), and 1- mm gold particles as microcarriers coated
with pRTL2-GUS vector DNA were used. The BY2 cells were
bombarded and assayed for transient GUS expression 2 days after
bombardment. Under the optimized conditions the biolistic bom-
bardment was carried out for the tobacco BY2 cells using the

shuttle plasmids, and the transformed cell lines were screened on
selection medium containing 50 mg/ml kanamycin. Total DNA was
extracted according to Dellaporta et al. (1983). Transformed to-
bacco cells were confirmed by PCR amplification of the AYVV
region with the primers C1 and CP sense. PCR cycling parameters
were 25 cycles of 1 min at 96C, 1 min at 57C, and 2 min at 72C.
The same DNA extract used for the AYVV primers was also
used for a tobacco chromosomal gene, SAMDC (S-adenosyl me-
thionine decarboxylase). The SAMDC gene was amplified using
the primers samdcfor (5
0
-CGGCTGCTCACATGACTGTTAGTT-
CTGGC-3
0
) and samdcrev (5
0
-AACATGCAAGCACCTTCTCAA-
CCAG-3
0
) with the following PCR cycling parameters: 25 cycles of
1 min at 95C, 30 s at 50C, 30 s at 72C.
Rescue of the pASVNPT and pASVGUS shuttle vectors
from tobacco BY2 cells in E. coli
Total DNA was isolated from tobacco BY2 transformed cells as
described by Dellaporta et al. (1983). Five micrograms of total
plant DNA was transformed into E. coli strain DH5a to rescue the
shuttle vector. Transformed cells were selected on LB plates con-
taining ampicillin (50 mg/ml). Plasmid DNA was isolated for fur-
ther analysis using the alkaline lysis method.
Analysis of replicating DNA by coupled restriction enzyme

digestion-random amplification PCR (CREDRA-PCR)
and Southern blot analysis
CREDRA-PCR has been used previously to identify DNA meth-
ylation in plants (Cai et al. 1996; Prakash and Kumar 1997). We
modified CREDRA-PCR to study plasmid replication in plant cells
as follows. Total DNA (5 mg) from DH5a, DNA from the tobacco
transformed cell line, and rescued DNA from DH5a were restricted
with m
6
A methylation-sensitive (DpnI) and methylation-resistant
(BclI) enzymes for 5 h in a 20-ml reaction volume. The digested
DNA was re-purified and amplified using AYVV primers (C1 and
CP sense) under the conditions described earlier. Five-microliter
aliquots of the PCR products were fractionated by agarose gel
electrophoresis and analyzed by Southern blot hybridization ac-
cording to manufacturer’s protocol (Boehringer Mannheim, Ger-
many). The 1.7-kb AYVV gene probe was generated by PCR
amplification and was labeled with digoxygenin-labeled dUTP by
random priming.
Histological GUS assays
Putative transformed cell lines and control BY2 cells were tested
for histochemical localization of GUS (Gallagher 1992). The cells
were incubated overnight at 37C in an assay buffer consisting of
100 mM NaPO
4
(pH 7.0), 1 mM X-Gluc (5-bromo-4-chloro-3-in-
dolyl glucuronide cyclohexylammonium salt), examined for fluo-
rescence, and photographed under a bright field using an inverted
Lietz Fluovert FU microscope (Leitz, Wetzlar, Germany).
Results and discussion

Construction of pASV plant-E. coli shuttle vectors
We chose a monopartite AYVV clone from pHN419 to
construct a pASV series of plant-E. coli shuttle vectors
pASVNPT, pASV82, pASVGUS and pASVDIR (Fig. 1).
The vector pASV has all of the essential elements of
AYVV except the CP (V1) to allow AYVV viral replicon
for extrachromosomal replication in plant cells. The
pUC19 backbone enables the plasmids to replicate in E.
coli under ampicillin selection for easy manipulation and
recovery of the clones from plants in E. coli. The NPTII
expression cassette [consisting of the CaMV 35S pro-
moter and nopaline synthase (NOS) gene terminator] in
pASVNPT allows kanamycin selection in plant cells. The
pASVDIR plasmid, with the deletion of the intergenic
region in pASVNPT, serves as a replication-defective
control. An intermediate cloning vector, pASV82, was
derived from pASVNPT and contains a unique HindIII
site for subsequent cloning of the expression cassette of
genes of interest. The GUS expression cassette was
cloned at this HindIII site to yield pASVGUS for its ex-
pression in plant cells. The expression cassette has a dual
CaMV 35S promoter, 5
0
UTR translation enhancer of TEV
and a CaMV 35S terminator. An alternate expression
cassette1 with the CaMV 35S promoter, 5
0
UTR of TEV, a
stuffer fragment, and vspB terminator is also available in
this vector. This stuffer fragment can be replaced with the

gene of interest between NcoI and KpnI sites. However,
the unique HindIII site is recommended for the insertion
of foreign gene expression cassette since NcoI and KpnI
are not unique sites in the vector.
Optimization of electroporation condition
Our objective was to study the replication of pASV-de-
rived vectors in N. benthamiana and N. tabacum proto-
plasts. We first standardized the electroporation condi-
tions for introducing the vectors to the isolated mesophyll
protoplasts of N. benthamiana using two electroporators.
A comparison of the effect of the two different pulse
84
types revealed a general trend. Freshly isolated proto-
plasts had 50–60% viability, and in both types of pulse
generators the viability dropped below 10% with higher
pulse strength (Fig. 2). The number of Trypan blue-
stained protoplasts increased and viability rapidly de-
creased with increasing field strength with both electro-
porators. To maintain a protoplast viability of 50%, we
determined that the optimum field strength should be
0.3 kV/cm for the exponential wave pulse generator and
0.5–0.6 kV/cm for the square wave pulse generator. The
protoplasts showed decreased viability under a low volt-
age mode with BTX T820 (data not shown). In subse-
quent experiments we used the BTX electroporator with a
pulse field strength of 0.55 kV/cm and a single pulse of
80 ms.
Transformation of mesophyll-derived protoplasts
of N. benthamiana with pASVNPT
Using the optimized electroporation conditions with high-

quality (circular form) input DNA (derived from Nucle-
obond column-purified kit), we electroporated mesophyll
protoplasts with pASVNPT and followed the replication
of vector DNA over a 6-day period. Microscopic obser-
vation showed no active protoplast division during this
period. The NPTII PCR products of pASVNPT were de-
tectable throughout the period tested, but their levels
declined with time in the cells (Fig. 3b). As expected,
there was no increase in the level of PCR products for the
chromosomal gene NAD5 (Fig. 3c). The steady-state
levels of the pASVNPT vector decreased slightly during
the 6-day period; this could be due to either decreased
replication or to the same level of replication ability but
an increased turnover of vector DNA. In practical terms,
either situation would lead to a lower copy number of
vector molecules. These results show that the pASVNPT
vector did not replicate efficiently in the N. benthamiana
cells. Previous reports of transient replication experiments
with protoplasts derived from mesophyll cells have shown
that host cell division is a prerequisite for the replication
of some types of begomovirus but not for all. For ex-
ample, replication of the ACMV requires host cell divi-
sion (Townsend et al.1986), while TGMV does not
(Brough et al. 1992). In other geminiviruses also, such
dependency on host cell division varies. In the Mastre-
virus group, WDV replication is dependent on cell divi-
Fig. 2 Electroporation efficiency of mesophyll protoplasts. Try-
pan-blue uptake and viability as determined by fluorescein diace-
tate staining of mesophyll protoplasts electroporated using: a BTX
model T820 (square wave pulse generator), b Bio-Rad Gene

pulser II (exponential pulse generator). Arrows indicate the opti-
mum conditions used in subsequent experiments
Fig. 3a–c PCR detection of
pASVNPT DNA in Nicotiana
benthamiana protoplasts. a
pASVNPT construct. IR Inter-
genic region. b Amplification
products of the vector-borne
NPTII gene. L Low-mass DNA
ladder. c Amplification prod-
ucts of the chromosomal NAD5
gene. L 1-kb ladder (Fermen-
tas). Lanes 1–6 PCR for days 1–
6 in duplicate, C negative con-
trol
85
sion of the host Triticum monococcum (Matzeit et al.
1991), while in another study with maize it was shown to
proceed in the absence of cell division (Timmermans et
al. 1992). Hence, the dependency of geminiviral replica-
tion on host cell division may be affected by both viral
and host factors.
Because there was a lack of replication of pASVNPT
in the N. benthamiana system in our studies, we conclude
that AYVV replication is dependent on host cell division
and that a slowly or non-dividing host system such as N.
benthamiana would not be suitable for pASV-derived
vectors. Consequently, we selected the related N. tabacum
BY2 cells, which are known for sustaining rapid division
in long-term cultures, for further testing.

Transformation of pASV-derived vectors
in N. tabacum BY2 cells by particle bombardment
In order to efficiently transform tobacco BY2 cells we
optimized the particle bombardment conditions using
pRTL2-GUS. In transient GUS expression studies with
pRTL2-GUS, the highest number of uniformly blue-col-
ored cells (4,423€225 cells) with 1,300 psi were observed
2 days post-bombardment. This is in contrast to the low
number of stained cells with 900 and 1,100 psi (271€11,
1,019€45 cells), respectively, and no stained cells ob-
served in the control mock bombarded cells. Subsequent
experiments were conducted under these optimized con-
ditions.
With the objective to study the replication of pASV-
derived vectors in tobacco BY2 cells, the pASVNPT,
pASVGUS and a replication-defective pASVDIR DNA
were bombarded into BY2 tobacco cells for clonal se-
lection of transformed lines on antibiotic-containing me-
dium. Two hundred putative transformed independent
lines from each vector were further cultured on selection
plates for an extended period of 4 months. The calluses
transformed with pASVNPT and pASVGUS appeared as
tiny white clumps between 20 days and 30 days post-
bombardment, thereby allowing clonal selection of
transformed cell lines. Both non-transformed tobacco
calluses and cells transformed with pASVDIR showed no
callus proliferation on selection plates (Fig. 4e,f), while
the growth of control non-transformed calluses on non-
selection plates was normal. The absence of colonies with
the negative control pASVDIR-transformed cells proved

that the intergenic region is necessary for viral replication
and that the selection process of transformed cells was
efficient. This observation is in accordance with deletion
analyses of other geminiviruses such as WDV (Ugaki et
al. 1991) and MSV (Shen and Hohn 1994), as this region
contains the initiation site for rolling circle DNA repli-
cation. The transformed calluses maintained for 4 months
kept their ability to stably replicate the pASV vectors. In
other analyses cell suspension cultures of TGMV and
MSV transformed tobacco and maize lines were main-
tained for 6 months and 1 year, respectively (Kanevski et
al. 1992; Palmer et al. 1999). Because calluses require
less frequent transfers than cell suspensions, the methods
described here would allow easier maintenance in pro-
longed periods.
Replication studies of pASV-derived shuttle vectors
As the maintenance of long-term cultures provides only
indirect evidence for the stable replication of vectors,
replication was studied using four methods to obtain in-
dependent validation: (1) detecting the presence of the
vector DNA by PCR in long-term cultures of transformed
calluses, (2) rescuing pASVGUS from plant cells into E.
coli, (3) CREDRA-PCR assays based on methylation
differences of DNA replicated in plant and E. coli cells,
and (4) assaying for foreign reporter genes in transformed
tobacco BY2 cells.
Fig. 4a–f Selection of trans-
genic tobacco BY2 calluses
30 days following bombard-
ment with pASV vectors. a

pASVNPT-transformed calluses
in selection medium. b pASV-
GUS-transformed calluses in
selection medium. c, d Control
tobacco calluses in the absence
of kanamycin. e Control tobac-
co calluses in selection medium.
f pASVNPTDIR-transformed
replication-defective control on
selection medium
86
PCR detection of pASVNPT and pASVGUS vectors
in transformed calluses
Randomly picked healthy calluses at the 6-week stage
growing on selection plates were screened for the pres-
ence of the vector. A 1.7-kb AYVV DNA fragment from
the pASVNPT and pASVGUS constructs was amplified
from the DNA of the transformed calluses. A 0.3-kb
SAMDC DNA fragment from the chromosomal gene
was amplified as a control (Fig. 5b). Twenty transformed
calluses that showed the presence of pASVGUS DNA
were transferred to MS medium with kanamycin for
vector selection in suspension culture.
Shuttling ability and rescue of pASVGUS
from plant cells to E. coli
Suspensions of transformed cell lines, as described above,
were used for studying shuttle replication of the con-
structs in plant and E. coli cells. The suspension cell
cultures were maintained by subculturing in selection
medium once a week. The morphology and growth rates

were similar in all cell suspension cultures up to 2–
3 months in comparison to the control BY2 cells without
kanamycin. The control BY2 cells on kanamycin did not
multiply in suspension cells. Subculturing was continued
for a period of 4 months.
We hypothesized that presence of the 1.7-kb AYVV
PCR fragment would not directly confirm the shuttling
ability of the vector. Hence, we attempted to rescue the
vector from plant cells into E. coli at monthly intervals.
This also allowed us to study any major rearrangements in
the vector as a consequence of replication in both the
plant and E. coli cells. The rescue of pASVGUS from
plant cells into E. coli was carried out on 1-, 2- and 3-
month-old suspension cultures maintained on selection
medium. Between 10 and 20 E. coli colonies were ob-
tained with 5 mg of total DNA derived from cells of the
tobacco suspension culture, while DNA prepared from the
untransformed control BY2 cells did not yield any E. coli
transformants, as was expected. This further confirmed
that pASVGUS DNA was propagated along with the
chromosomal DNA in transformed plant cells. The
structural integrity of the AYVV vector was studied by
PCR and the sizing of the constructs was determined by
restriction profiling. Plasmid DNA was prepared from the
rescued colonies and screened for the presence of the
AYVV and GUS regions, respectively (Fig. 6a,b). All of
the rescued clones that were screened showed the 1.7-kb
AYVV fragment and the 0.7-kb GUS gene fragment upon
PCR-based amplification. Uncut plasmids were compared
to detect any size differences, but no major differences in

the sizes of the plasmids were found. Additional restric-
tion analyses of the rescued clones were performed to
further investigate the possibilities of any vector DNA
rearrangements: no significant restriction fragment length
polymorphisms were seen between the rescued and con-
trol vector DNA (Fig. 6c). Similar results were obtained
from 1-, 2-, and 3-month-old cultures. Taken together,
these results confirm that the pASV backbone vector can
replicate in both plant and E. coli without undergoing any
detectable size alterations or rearrangements. Previous
studies on the rescue of geminiviral shuttle vectors were
carried out on 6-day- and 7-day-old cultures with the
graminaceous host containing mastrevirus, WDV (Ugaki
et al. 1991; Kammann et al. 1991). We report here long-
term maintenance of the structural integrity of bego-
movirus-based pASV vectors.
Although, these results show no major rearrangements,
minor ones would not have been detected using the
methods described here. Also, the rearranged vectors
could not have been rescued if the rearrangement had
affected their ability to replicate in E. coli. The versatility
of the vectors to replicate in other host plants needs to be
tested with respect to broader applications.
Fig. 5a, b PCR detection of pASVGUS in tobacco BY2 trans-
formed calluses. a pASVGUS construct. b Lanes: L 1-kb ladder
(Fermentas), 1 negative control lacking total DNA from calluses, 2
PCR products of AYVV gene from pASVNPT, 3 pASVGUS-
transformed callus DNA, 4, 5 PCR products of SAMDC gene—
internal positive control for chromosomal DNA from calluses
87

Replication studies based on DNA methylation differences
We reasoned that direct evidence for vector replication in
plant cells would require either distinguishing the
replicative intermediates of the vector or showing the
presence of DNA replication features specific to plants.
We chose the latter line of proof and based our experi-
ments on the m
5
C methylation that occurs during the
replication of DNA in plants but which is absent in E.
coli. The converse argument that m
6
A methylation is
found in E. coli but is absent in eukaryotic DNA was also
used in these studies. Other researchers have proposed
that the extrachromosomal replicons may not be acces-
sible for the host methylases, although the barrier pre-
venting access of the methylases was not defined (Brough
et al. 1992; Doerfler 1993). Based on these considera-
tions, we studied the differences in the methylation status
of the input DNA obtained from E. coli and that of the
DNA obtained from newly transformed plant cell lines
using a combination of methylation-sensitive restriction
enzymes followed by PCR.
To study the replication of extrachromosomal DNA
of the pASVGUS shuttle vector we used a modified
methylation-based PCR method termed CREDRA-PCR
(Fig. 7a). This method was carried out on total ge-
nomic DNA isolated on the third day of subculture from
Fig. 6a–c Rescue of pASVGUS shuttle vectors from transformed

BY2 plant cells in E. coli. PCR detection of pASVGUS plasmid
DNA with: a AYVV primers, b GUS primers. Lanes: L 1-kb ladder
(Fermentas), C negative control with non-transformed tobacco cell,
2 PCR with control pASVGUS plasmid DNA, 3–8 PCR with res-
cued pASVGUS clones (RC 3–8) from E. coli clones. c Restriction
profiles of pASVGUS rescued clones RC 3 (3*) and RC 4 (4*)
from panels a and b. Lanes: 1, 4, 7 Non-digested DNA controls, 2,
5, 8 BamHI-digested, 3, 6, 9 HindIII-digested. The experiment was
repeated with 1-, 2-, and 3-month-old suspension cultures. Results
from the 2-month-old suspension cultures are shown in this figure
Fig. 7a–c Verification of
pASVGUS shuttle vector repli-
cation in tobacco cells using
CREDRA-PCR. a Experimental
setup for studying extrachro-
mosomal replication. b CRE-
DRA-PCR products using
AYVV primers for pASVGUS
detection. Lanes: 1–3 Control
input pASVGUS m
6
A methyl-
ated DNA, 4–6 total genomic
DNA from pASVGUS-trans-
formed tobacco cells, 7–9 res-
cued plasmid DNA in DH5a (U
un-digested, BclI BclI-digested,
DpnI DpnI-digested). c South-
ern blot hybridization of the
CREDRA-PCR products from

the gel shown in panel b with
DIG-labeled 1.7-kb AYVV
fragment
88
2-month-old cell suspension cultures. To discriminate
between the input and the newly replicated DNA, both the
DNA from suspension tobacco cells and the DNA rescued
from E. coli were digested with either dam-methylation-
dependent restriction enzyme DpnIorBclI followed by
PCR amplification and Southern analyses. DpnI is de-
pendent on the methylation of adenine in the recognition
sequence GA
6
/TC and, consequently, it cleaved the both
the input and rescued DNA that was methylated in E.
coli. BclI did not cleave the m
6
A methylated sequence
(Fig. 7b,c). In contrast, the replicative DNA in tobacco
cells was resistant to digestion with DpnI since they
lacked the m
6
A site, consequently giving a PCR product
with the AYVV primers. BclI did cleave the de novo
synthesized DNA that lacked m
6
A methylation in the
tobacco cells. Hence, vector DNA was not amplified and
there was no corresponding signal in either the agarose
gel or its Southern blot probed with vector DNA (Fig. 7c)

that showed a higher sensitivity of detection. Based on
these observations, we conclude that de novo replication
of the pASVGUS vector occurred in plant cells in 2-
month-old suspension cells. The mastrevirus replicating
shuttle vector in protoplasts of Triticum monococcum
(Kammann et al. 1991), maize endosperm-derived pro-
toplasts for WDV (Timmermans et al. 1992), and tobacco
NT1 cells for BeYDV (Mor et al. 2003) have also been
analyzed by methylation sensitivity. In these reports,
methylation-based studies of replication were carried out
during the first week following electroporation. In order
to study de novo replication in long-term plant cell cul-
tures, we adopted a methylation-based PCR assay—the
CREDRA-PCR method—which is more sensitive, re-
quires little starting material, and is simpler to perform
than the previous methods.
Expression of foreign reporter gene
Once the replication ability of pASVGUS was estab-
lished, we further tested its application as an extrachro-
mosomal expression vector using the GUS reporter gene.
Transgenic cell lines and control BY2 cells were tested
for histochemical localization of GUS over an extended
period of time spanning 4 months. GUS assays were
performed daily for 7 days beginning with the first day of
subculturing on 2-week- and 1-, 2-, 3,- and 4-month-old
suspension cultures. In all cases GUS expression was
detected in 0.001–0.006% of the cells screened. GUS
expression was maximal on third day, and expression was
not found when the cells had reached the seventh day, a
time when the tobacco BY2 cells are likely to be in the

stationary stage (Nagata et al.1992). The reason for
maximal expression on third day is not yet clear. The
highest percentage of GUS-expressing cells occurred in
the first month, with the percentage declining gradually
later in the fourth month. However, we speculate that this
could be due to changes in the copy number of the vector
during the different growth phases of the culture. This
variation was also noticed with TGMV (Kanevski et al.
1992) and BeYDV in tobacco cells (Mor et al. 2003). The
difference in copy number could arise from interference
of geminiviral replication with plant cell-cycle machinery
or other host cell pathways (Gutierrez 2000).
Conclusion
We have shown here the utility of pASV-derived shuttle
vectors for long-term stable maintenance of constructs
and expression of foreign genes in cultured plant cells.
Using methylation-based PCR assays we have also shown
de novo replication of pASV vectors in long-term cul-
tures. On the basis of the system we have described here
pASV vectors will be suitable for use when researchers
are looking to express foreign proteins in a closed sterile
system rather than in whole transgenic plants. They will
be useful in long-term cell cultures by improving the
replication and expression levels through extensively
characterizing the host factors in synchronized cell cul-
tures and promoters highly active in early cell division of
BY2 cells.
Acknowledgements The authors would like to thank Dr. Wong Sek
Man for providing pHN419 containing the full-length AYVV
coding sequence, Prof. Charles Arntzen and Dr. Hugh S. Mason for

their gift of the pIBT210.1 plasmid, and Dr. Jaideep Mathur for
providing the tobacco BY2 suspension cells. TS and GA have been
supported by NUS research scholarships.
References
Briddon RW, Mansoor S, Bedford ID, Pinner MS, Markham PG
(2000) Clones of cotton leaf curl geminivirus induce symptoms
atypical of cotton leaf curl disease. Virus Genes 20:19–26
Brough CL, Gardiner WE, Inamdar NM, Zhang XY, Ehrlich M,
Bisaro DM (1992) DNA methylation inhibits propagation of
tomato golden mosaic virus DNA in transfected protoplasts.
Plant Mol Biol 18:703–712
Cai Q, Guy CL, Moore GA (1996) Detection of cytosine methyl-
ation and mapping of a gene influencing cytosine methylation
in the genome of citrus. Genome 39:235–242
Carrington JC, Freed DD, Leinicke AJ (1991) Bipartite signal se-
quence mediates nuclear translocation of the plant potyviral
NIa protein. Plant Cell 3:953–962
Davies JW, Stanley J (1989) Geminivirus genes and vectors.
Trends Genet 5:77–81
Dellaporta SL, Wood J, Hicks JB (1983) Maize DNA miniprep. In:
Malmberg R, Messing J, Sussex I (eds) Molecular biology of
plants: a laboratory course manual. Cold Spring Harbor Labo-
ratory Press, Cold Spring Harbor
Doerfler W (1993) Pattern of de novo methylation and promoter
inhibition: studies on the adenovirus and the human genome.
In: Jost JP, Saluz HP (eds) DNA methylation: molecular biol-
ogy and biological significance. Birkhauser Vedlag, Basel,
pp 262–299
Dry IB, Rigden JE, Krake LR, Mullineaux PM, Rezaian MA (1993)
Nucleotide sequence and genome organization of tomato leaf

curl geminivirus. J Gen Virol 74:147–151
Fischer R, Emans N (2000) Molecular farming of pharmaceutical
proteins. Transgenic Res 9:279–299
Fischer R, Emans N, Schuster F, Hellwig S, Drossard J (1999a)
Towards molecular farming in the future: using plant cell-
suspension cultures as bioreactors. Biotechnol Appl Biochem
30:109–112
89
Fischer R, Vaquero CM, Sack M, Drossard J, Emans N, Com-
mandeur U (1999b) Towards molecular farming in the future:
transient protein expression in plants. Biotechnol Appl Bio-
chem 30:113–116
Gallagher SR (1992) GUS Protocols: using the GUS gene as a
reporter of gene expression. Academic, San Diego, Calif.
Gutierrez C (2000) DNA replication and cell cycle in plants:
learning from geminiviruses. EMBO J 19:792–799
Haq TA, Mason HS, Clements JD, Arntzen CJ (1995) Oral im-
munization with a recombinant bacterial antigen produced in
transgenic plants. Science 268:714–716
Kammann M, Schalk HJ, Mazeit V, Schaefer S, Schell J, Gronen-
born B (1991) DNA replication of wheat dwarf virus, a gemi-
nivirus, requires two cis-acting signals. Virology 184:786–790
Kanevski IF, Thakur S, Cosowsky L, Sunter G, Brough C, Bisaro
D, Maliga P (1992) Tobacco lines with high copy number of
replicating recombinant geminivirus vectors after biolistic
DNA delivery. Plant J 2:457–463
Kikkert JR (1993) The biolistic PDS-1000/He device. Plant Cell
Tissue Organ Cult 33:221–226
Klien TM, Kornstein L, Stanford JC, Fromm ME (1989) Genetic
transformation of maize cells by particle bombardment. Plant

Physiol 91:440–444
Lazarowitz SG, Pinder AJ, Damsteegt VD, Rogers SG (1989)
Maize streak virus genes essential for systemic spread and
symptom development. EMBO J 8:1023–1032
Ma JKC, Drake PMW, Christou P (2003) The production of re-
combinant pharmaceutical proteins in plants. Nat Rev Genet
4:794–805
Matzeit V, Schaefer S, Kammann M, Schalk HJ, Schell J, Gro-
nenborn B (1991) Wheat dwarf virus vectors replicate and
express foreign genes in cells of monocotyledonous plants.
Plant Cell 3:247–258
Mor TS, Moon YS, Palmer KE, Mason HS (2003) Geminivirus
vectors for high level expression of foreign proteins in plant
cells. Biotechnol Bioeng 81:430–437
Mullineaux PM, Davies JW, Woolston CJ (1992) Geminiviruses
as gene vectors. In: Michael T, Wilson A, Davies JW (eds)
Genetic engineering with plant viruses. CRC, Boca Raton,
pp 187–215
Murashige T, Skoog F (1962) A revised medium for rapid growth
and bioassays with tobacco tissue cultures. Physiol Plant
15:473–497
Nagata T, Nemoto Y, Hasezawa S (1992) Tobacco BY-2 cell line
as the ‘HeLa’ cells in the cell biology of higher plants. Int Rev
Cytol 132:1–30
Palmer KE, Thomson A, Rybicki EP (1999) Generation of maize
cell lines containing autonomously replicating maize streak
virus-based gene vectors. Arch Virol 144:1345–1360
Prakash AP, Kumar PP (1997) Inhibition of shoot induction by
5-azacytidine and 5-aza-2
0

deoxycitidine in Petunia involves
DNA hypomethylation. Plant Cell Rep 16:719–724
Sala F, Marchesi ML, Castiglione S, Paszkowski J, Saul M,
Potrykus I, Negrutiu I (1989) Direct gene transfer in protoplasts
of Nicotiana plumbaginifolia. In: Bajaj YPS (ed) Biotechnol-
ogy in agriculture and forestry, vol 9. Plant protoplasts and
genetic engineering II. Springer, Berlin Heidelberg New York,
pp 217–227
Sambrook J, Fritsch EF, Maniatis T (1989) Molecular cloning: a
laboratory manual. Cold Spring Harbor Laboratory Press, Cold
Spring Harbor
Saunders JA, Lin CH, Hou BH, Cheng JP, Tsengwa N, Lin JJ,
Smith CR, McIntosh MS, Wert SV (1995) Rapid optimization
of electroporation conditions for plant cells, protoplasts and
pollen. Mol Biotechnol 3:181–189
Shen WH, Hohn B (1994) Amplification of the b-glucuronidase
gene in maize plants by vector based on maize streak virus.
Plant J 5:227–236
Stanley J (1993) Geminiviruses: plant viral vectors. Curr Biol 3:91–
96
Tan HN Priscilla, Wong SM, Wu M, Bedford ID, Saunders K,
Stanley J (1995) Genome organization of ageratum yellow vein
virus, a monopartite whitefly-transmitted geminivirus isolated
from a common weed. J Gen Virol 76:2915–2922
Timmermans MCP, PremDas O, Messing J (1992) Trans replica-
tion and high copy numbers of wheat dwarf virus vectors in
maize cells. Nucleic Acids Res 20:4047–4045
Timmermans MCP, PremDas O, Messing J (1994) Geminiviruses
and their uses as extrachromosomal replicons. Annu Rev Plant
Physiol Plant Mol Biol 45:79–112

Topfer R, Gronenborn B, Schell J, Steinbiss HH (1989) Uptake and
transient expression of chimeric genes in seed-derived em-
bryos. Plant Cell 1:133–139
Townsend R, Watts J, Stanley J (1986) Synthesis of viral DNA
forms in Nicotiana plumbaginifolia protoplasts inoculated with
cassava latent virus (CLV); evidence for the independent rep-
lication of one component of the CLV genome. Nucleic Acids
Res 14:1253–1265
Ugaki M, Ueda T, Timmermans MCP, Vieira J, Elliston KO,
Messing J (1991) Replication of a geminivirus derived shuttle
vector in maize endosperm cells. Nucleic Acids Res 19:371–
377
Ward A, Etessami P, Stanley J (1988) Expression of a bacterial
gene in plants mediated by infectious geminivirus DNA.
EMBO J 7:1583–1587
90