The Plant Cell, Vol.

5,973-982, August 1993 O 1993 American Society of Plant Physiologists

Alanine Scanning Mutagenesis of a Plant Virus Movement

Protein ldentifies Three Functional Domains

Donna Giesman-Cookmeyer and Steven A. Lommel’

Department of Plant Pathology, Box 7616, North Carolina State University, Raleigh, North Carolina 27695-7616

Downloaded from https://academic.oup.com/plcell/article/5/8/973/5984589 by guest on 21 December 2023

Alanine scanning mutagenesis was performed o n the red clover necrotic mosaic virus (RCNMV) movement protein (MP),
and 12 mutants were assayed in vitro for RNA binding characteristics and in vivo for their ability to potentiate RCNMV
cell-to-cell movement. The mutant phenotypes that were identified in vitro and in vivo suggest both that cooperative
RNA binding is not necessary for cell-to-cell movement in vivo and that only a fraction of the wild-type RNA binding
may be required. The MP mutants defined at least three distinct functional regions in the MP: an RNA binding domain,
a cooperative RNA binding domain, and a third domain that is necessary for cell-to-cell movement in vivo. This third
domain may be required for targeting the MP to cell walls and plasmodesmata, interacting with host proteins, folding,
or possibly binding RNA into a functional ribonucleoprotein complex capable of cell-to-cell movement.

INTRODUCTION

Plant viruses must be able to move from cell to cell in a sys- movement, but a smaller deletion of the C-terminal55 amino
temic host to cause disease. Virus infection moves from the acids, which does localize to plasmodesmata, allows cell-to-
initially infected cell to neighboring cells and then throughout cell movement. However, this movement is somewhat impaired
the host. The mechanisms of movement known for animal (Berna et al., 1991; Gafny et al., 1992).
viruses, receptor-mediatedendocytosis and cellular fusion, are Recently, it was also shown that the TMV MP was capable
not available to plant viruses because each plant cell is sur- of binding to single-stranded nucleic acids in vitro (Citovsky
rounded by a cell wall. Available evidence suggests that plant et al., 1990). Because the TMV coat protein is not required
viruses move from cell to cell via the intercellular connections for viral cell-tocell movement (Culver and Dawson, 1989; Saito
known as plasmodesmata (Gibbs, 1976; Atabekov and Dorokov, et al., 1990), this result suggests that it may be the TMV MP
1984). Furthermore,this movement generally requires avirus- that both protects the viral RNA and facilitates its movement
encoded movement protein (MP) (reviewed by Hull, 1989; through the plasmodesmatal intercellular connections in the
Atabekov and Taliansky, 1990; Robards and Lucas, 1990; Deom form of a ribonucleoproteincomplex. Single-stranded DNA-MP
et al., 1992). and RNA-MP complexes have been visualized in vitro using
Comparison of MPs from severa1 different groups of plant electron microscopy. The dimensions of these complexes are
viruses revealed the presence of conserved amino acid mo- such that they would probably not be excluded from move-
tifs that may reflect a common functional ancestry (Melcher, ment through plasmodesmatal channels enlarged by the TMV
1990; Koonin et al., 1991). To date, there have been at least MP (Citovsky et al., 1992). This supports the idea that TMV
two functions associated with MPs. For tobacco mosaic virus may move as a genomic RNA-MP complex.
(TMV), it has been shown that the 30-kD MP localizes to the Red clover necrotic mosaic virus (RCNMV) is similar to TMV
plasmodesmata in cells of both TMV-infected plants and trans- in that it does not require its capsid protein for cell-tocell move-
genic plants expressing TMV MP (3omenius et al., 1987; Atkins ment (Xiong et al., 1993) and that the 35-kD MP is able to bind
et al., 1991) and can be found in the cell wall fraction of in- to single-stranded nucleic acids in vitro (Osman et al., 1992).
fected tissue (Deom et al., 1990). The TMV MP is capable of Like the TMV MP, the RCNMV MP binds to both RNA and
modifying the size exclusion limit of plasmodesmata, making single-stranded DNA in a cooperative manner. A truncated MP,
it possible for larger molecules to pass from cell to cell (Wolf which is missing the C-terminal88 amino acids, has no defect
et al., 1989). A deletion mutant of the TMV MP lacking the in RNA binding but is highly attenuated for viral movement
C-terminal 73 amino acids does not localize to the cell wall and symptom development in cowpea (Osman et al., 1991b).
fraction or to plasmodesmata. This mutant TMV MP does not We were interested in identifying the activities that are as-
modify plasmodesmata and is unable to facilitate cell-to-cell sociated with the 35-kD RCNMV MP and required for cell-tocell
movement and viral infection. To this end, we generated 12*
alanine scanning mutations in the RCNMV MP by site-directed
To whom correspondence should be addressed. mutagenesis and examined their effects on RNA binding in
974 The Plant Cell

vitro and their infectivity in vivo. Alanine scanning mutagene-

sis (Cunningham and Wells, 1989) is based on the rationale
that clusters of charged residues in the primary sequence of
a protein are likely to be found on the surface of the folded 1 2 3 4 5 6 7 8 9 10
structure. By mutagenizing these residues, regions of a pro-
tein involved in macromolecular interactions are targeted and
the number of mutants that must be screened is minimized.
Alanine residues are substituted for the charged residues to
avoid disrupting, as much as possible, the folded structure of
the protein.

Downloaded from https://academic.oup.com/plcell/article/5/8/973/5984589 by guest on 21 December 2023

We developed an in vitro RNA binding assay based on the
work of Osman et al. (1992) and Citovsky et al. (1990) and used

JuL
this assay to quantitatively and qualitatively characterize the
binding of the wild-type and mutant RCNMV MPs to RNA. Also,
using cDNA clones from which RCNMV infectious transcripts
can be derived (Xiong and Lommel, 1991), we were able to
test the effects of the alanine scanning mutations on cell-to-
cell movement and systemic infection in vivo. Our findings in-
dicated the presence of at least three functional domains in
the RCNMV MP, including a region of the protein required for
cooperative binding of the MP to RNA in vitro. We also showed B
that cooperative RNA binding in vitro by the RCNMV MP is
not necessary for its role in cell-to-cell movement and systemic
infection in vivo. Also, MP mutants that bind RNA at 20% of
1 2 3 4 5 6 7 8 9 10
the wild-type level in vitro are still able to facilitate cell-to-cell f\ r\ *\ r \
movement of RCNMV in vivo.

RESULTS

In Vitro RNA Binding of the RCNMV MP

Binding of the RCNMV MP to single-stranded nucleic acids

was first shown by Osman et al. (1992) using both mobility shift Mfcft*
and UV cross-linking assays. We have utilized a similar mo-
bility shift binding assay to study the RNA binding properties
of the RCNMV MP in vitro. A uniformly radiolabeled transcript
generated from the 5'terminal 200 nucleotides of RCNMV RNA
2 was used as the probe in all of the binding studies. The
RCNMV MP was produced in Escherichia coli using the pET
vector system described by Studier and Moffat (1986) and Figure 1. Mobility Shift Assay of Wild-Type RCNMV MP Binding to
Rosenberg et al. (1987). Labeled RNA.
As seen in Figures 1A and 1B, incubation of RCNMV MP (A) Mobility shift assay of wild-type RCNMV MP. Prior to electropho-
purified from E. coli with the labeled RNA transcript prior to resis on a 4°/o polyacrylamide gel, a 33P-labeled transcript (10 ng) of
polyacrylamide gel electrophoresis retarded the mobility of the the 5' terminal 200 nucleotides of RCNMV RNA 2 was incubated with
labeled RNA probe. The mobility shift was abolished when increasing quantities of E. co//-expressed MP at 4°C for 20 min in 40
the reaction products were incubated with proteinase K prior nL of buffer B1. In lanes 1 to 7, 0, 0.1, 0.2, 0.4, 0.8, 1.0, and 2.0 ng of
to electrophoresis (Figure 1A, lanes 8 and 9). The addition of MP were incubated with probe, respectively. In lanes 8 and 9, 20 ng
of proteinase K was added after the binding reaction, and the mixture
BSA neither interfered with the mobility shift caused by the
was incubated for 20 min prior to electrophoresis. Two micrograms
MP (data not shown) nor bound to RNA by itself (Figure 1A, of MP was added in lane 8, and no protein was added in lane 9. In
lane 10). The specificity of RNA binding was examined in a lane 10,10 ng of BSA was incubated with the labeled RNA instead of MP.
competition experiment. A 2.5- to 40-fold molar excess of un- (B) Ten nanograms of 33P-labeled transcript was used in a competi-
labeled RNA was added to the binding reaction at the same tion assay under the same conditions as given in (A). Lane 1 contains
time as the labeled probe (Figure 1B, lanes 2 to 10). Both the only the radiolabeled RNA probe. In lanes 2 to 10, 2 ng of MP was
homologous competitor, the 5'terminal 200 nucleotides of RNA incubated with the labeled RNA. In lanes 3 to 6, 25, 50, 100, and 200
 Plant Virus Cell-to-Cell Movement 975

2, and nonhomologous competitor, yeast ribosomal RNA, ef- were actually caused by a mutant RNA binding profile. An al-
fectively competed with the labeled probe for MP binding, and ternative interpretation is that mutant MPs 122,128,161,204,
the mobility shift was no longer observed (Figure 1B, lanes 242, and 305 (and to some extent 280 and 291) did not alter
3 to 6 and 7 to 10). This suggests that the in vitro binding activ- the RNA binding profile of the MP but were able to migrate
ity of the MP is not sequence specific. The only specificity
exhibited by the RCNMV MP is that it binds only to single-
stranded nucleic acids and not to double-stranded nucleic acids
(Osman et al., 1992).
The binding of the labeled RNA probe in the mobility shift 27-31
MAVHVENLSDLftKTlJDGVAVSLNRYT[jl«Ec3SGVSEAPLIPASMMSK
assay appeared to occur in a highly cooperative fashion. The

Downloaded from https://academic.oup.com/plcell/article/5/8/973/5984589 by guest on 21 December 2023

fact that the probe migrated either as the completely unbound ITDYAKTTAKGNSVALNYTHWLSIAPTIGVAIPGHVTVELINPNVE
or completely bound form in the polyacrylamide gel is consis- 122 128
GPFQVMSGQTLSWSPGAGKPCLMIFSVlSniQLNSllSiEPFRVRITlJTGI
tent with the the idea that once one or a few molecules of MP
144 161
are bound to a labeled RNA molecule, that molecule becomes PTi^SYARCHAYWGFDVGTitnnhfYKSEPARLIELEVGYQRTLLSSIK
a preferred substrate for the rapid binding of additional MP 204
molecules (Citovsky et al., 1990). This would explain the ab- AVEAYVQFTFDTSRM32NPQLCTKSNV11IIPPKAETGSIRGIAPPLS
242 278 280
sence of labeled RNA molecules with intermediate mobilities. VVPNQXSSJjSKVIJCQKGGTGSKTTKLPSLEPSSGSSSGLSMSJiSSijElI
291 301 305

In Vitro RNA Binding Profiles of Mutant MPs

Alanine scanning mutations were generated in the RCNMV

MP sequence using site-directed mutagenesis, as shown in B
Figure 2A, and mutant MPs were then purified from £ coli. If)
Mutant RCNMV MPs 27-31,144, 278, and 301 had RNA bind- m CM
ing profiles similar to that of wild-type MP in the mobility shift CM
assay (Figure 2B). Six other mutants (122,128,161,204,242,
and 305) had RNA binding profiles that were unlike that of the dttt
wild type. In general, these mutant MPs retarded the mobility
of the RNA probe much less than the wild-type MR However,
these mutant proteins did bind RNA; the predominant labeled
species migrated more slowly than the unbound probe. Also,
this labeled species represents the fully bound probe, even
though its migration is between that of the unbound RNA probe
and the RNA probe fully bound by the wild-type protein. It is
not an intermediate; the addition of more protein does not pro-
duce any additional species, but rather more of the same
t<
species (data not shown). Mutants 280 and 291 exhibited two
distinct RNA binding activities, one nonmigrating species like
that of the wild-type MP and one quickly migrating species
like the mutant MPs (Figure 2B). At lower MP concentrations
(<100 ng/nL), only the wild-type binding (large mobility shift)
was observed (data not shown). Figure 2. Alanine Scanning Mutations Introduced into the RCNMV
MP and the RNA Mobility Shift Profile of Each Mutant.
Under the conditions of the continuous, native gel system
used for these mobility shift assays, the wild-type MP was un- (A) The primary amino acid sequence of the RCNMV MP in single-
able to migrate into the gel. This assay could not determine letter code. Reverse-shaded residues identify charged amino acids
whether the altered mobilities exhibited by the mutant MPs that were changed to alanine by site-directed mutagenesis. Above each
mutagenized amino acid cluster is the mutant's label. For every mu-
tant except 278, this label identifies the position of the first amino acid
mutagenized in each cluster. For 278, it is the second amino acid.
(B) Autoradiograph of a mobility shift assay conducted with wild-type
(WT) and mutant RCNMV MPs. Four micrograms each of the wild-
Figure 1. (continued). type and mutant MPs were incubated for 20 min at 4°C with 10 ng
ng, respectively, of unlabeled RNA transcript from the 5' 200 nucleo- of the 33P-labeled RNA probe in a total volume of 40 nL of buffer B1.
tides of RCNMV RNA 2 were added to the binding reaction as a The reaction products were electrophoresed in the continuous gel sys-
competitor; in lanes 7 to 10, 50, 100, 200, and 400 ng, respectively, tem described in Methods. The gel was dried and exposed to film to
of ribosomal RNA from yeast were added as a competitor. produce the autoradiograph shown.
976 The Plant Cell

into the gel. The mutations introduced may have altered char-
acteristics (such as net or effective charge) that affect the
electrophoretic properties of the MP but not the RNA binding.
To distinguish between these two possibilities, we conducted 1 2 3 4 5 6 7 8 9 10
mobility shift assays using a discontinuous, native gel system.
The resolution of different proteins and protein complexes was
much greater in a discontinuous gel than in a continuous one
(Chambrach et al., 1976), and the gel system that we employed
had both a different pH and running buffer than the continu-
ous gel system (see Methods).

Downloaded from https://academic.oup.com/plcell/article/5/8/973/5984589 by guest on 21 December 2023

Several results emerged from the analysis of RNA binding
by the wild-type and mutant MPs in the second gel system.
Although the wild-type MP did not migrate into the gel in the
discontinuous gel system, as seen when comparing Figures
2 and 3, the resolution of the unbound probe, bound complexes,
and unincorporated nucleotides was much higher than in the
original continuous system, as shown in Figures 3 and 4. Also,
Mftft«M
the background radioactivity characteristic of the continuous
gel system (seen as a smear in each lane along the length
•apiSHImw
of the gel in Figures 1 and 2) was virtually eliminated. This B
facilitated the quantification of RNA binding by radioanalytic
imaging of the dried gels (see below).
1 2 3 4 5 6 7 8 9 10
CD
JD
O
Q.

IIMMM
Figure 4. Comparison of RNA Binding profiles of Wild-Type and Mu-
tant 305 RCNMV MPs in a Discontinuous Native Gel System.
(A) Binding profile of wild-type MP in a discontinuous gel system. Five
nanograms of 33P-labeled RNA probe was added to increasing
amounts of the wild-type MP isolated from E. coli in a 20-nL volume
of buffer B1 at 4°C for 20 min prior to electrophoresis on a discontinu-
ous gel system (described in Methods). 0, 0.03, 0.07, 0.1, 0.2, 0.4, 0.8,
1, 2, and 4 nL of protein were added in lanes 1 to 10, respectively.
(B) Binding profile of mutant MP 305 in a discontinuous gel system.
Five nanograms of ^P-labeled RNA probe was added to increasing
Figure 3. Autoradiograph of a Mobility Shift Assay Conducted with
amounts of mutant MP 305 under the same conditions given in (A).
Wild-Type and Mutant RCNMV MPs Using a Discontinuous Gel System.
Unincorporated nucleotides are seen at the bottom of the gel for both
Two micrograms each of the wild-type (WT) and mutant MPs (exclud- (A) and (B). The autoradiographs shown here were exposed seven
ing 305, which is presented in Figure 4) were incubated for 20 min to eight times longer than those in Figures 1 to 3.
at 4°C with 5 ng of the 33P-labeled RNA probe in a total volume of 20
nL of buffer 81. The reaction products were electrophoresed on the
discontinuous gel system described in Methods. The gel was dried
and exposed to film to produce the autoradiograph shown. Unincor-
porated nucleotides are visible at the bottom of the gel in all lanes.
 Plant Virus Cell-to-Cell Movement 977

Interestingly, all 12 mutant MPs also failed to migrate into Biological Actlvlty of the Alanine Scanning
the gel when bound to RNA in the discontinuous gel system Mutant MPs
(Figures 3 and 4). However, in some cases, it was possible
to resolve RNA-MP complexes with a higher mobility in addi- RNA 2 transcripts were produced in vitro from cDNA clones
tion to the completely bound, nonmigrating complex (for containing each of the 12alanine scanning mutations and were
example, se8 mutant 305,shown in Figure 46, lanes 6 to 10). coinoculated with the wild-type RNA 1 transcript onto Nico-
These additional complexes appeared to be intermediates in fiana benfhamiana plants. Symptom development was moni-
the binding process and may represent RNA molecules with tored on both the inoculated and noninoculated leaves, and
only one or a few molecules of MP bound. These intermedi- viral infection was assayed by RNA gel blot analysis. Symp-
ates were seen only for mutant MPs 27-31,122,128 (data not toms observed on the inoculated leaves were indicative of the

Downloaded from https://academic.oup.com/plcell/article/5/8/973/5984589 by guest on 21 December 2023

shown), and 305 (Figure 46). The loss of the “all-or-none” bind- ability of an MP to facilitate cell-to-cell movement. On the other
ing characteristic seen with the wild-type MP is consistent with hand, symptoms on the noninoculated leaves demonstrated
a loss of cooperative binding, as originally described by the ability of the virus to move long distance or systemically
Citovsky et al. (1990).It is also possible that one or more of (Xiong et al., 1993).Plants inoculated with viral transcripts of
these mutants still bind RNA cooperatively but notas efficiently the MP mutants 27-31,204,242,280, 301,and 305 developed
as the wild-type MP. The other eight mutants formed only a symptoms on the inoculated and noninoculated leaves of N.
single RNA-MP complex that did not migrate into the gel, as benfhamiana between 4 and 6 days postinoculation. Gel blot
in the case of the wild-type MP (data not shown). In contrast, analysis of total RNA extracted from inoculated leaves showed
in the continuous gel system, severa1 mutants exhibited that RNA 2 was only detected in tissue isolated from plants
RNA-MP complexes that migrated into the gel, migrating more exhibiting symptoms. MP mutants that produced no symptoms
quickly than the wild-type complex (compare Figures 2 and 3). did not accumulate a detectable amount of viral RNA in the
Because the wild-type and mutant MPs shift the mobility inoculated leaves, as shown in Figure 6. Each mutant was in-
of the labeled RNA probe to the same extent in the discon- oculated on at least two plants at a time and the inoculations
tinuous gel system, it was possible to directly compare the were repeated at least three times.
levels of RNA binding exhibited by the various mutant MPs.
We observed that mutations 122 and 128 have the most sig-
nificant effect on the levels of RNA binding (Figure 3)and that, 1 O0
in general, mutations in the 5’ half of the MP gene reduce the
RNA binding capacity of the MPs more than those in the 3‘
half of the MP gene. It is possible, therefore, that amino acids
122 to 129 in the MP represent an RNA binding site and that
the surrounding regions are also important for RNA binding.
We quantified the RNA binding of the wild-type and mutant
MPs in the mobility shift assay by measuring the radioactivity
U
C
a
60
-
_O_

-
_c_

__t_
wtprotein
mutant 27-31
mutant242
mutant278
mutant 305
O
associated with the free and bound species in a dried poly- n
acrylamide gel by radioanalytic imaging, as shown in Figure 8 40
5. The RNA binding profile of the wild-type MP was character-
istic of cooperative binding and is consistent with results
previouslydescribed for the TMV and RCNMV MPs (Citovsky 20
et al., 1990;Osman et al., 1992).
From our analysis of the RNA binding profiles of the mutant
MPs, three phenotypic classes could be distinguished. One O
O 1 O0 200
class, composed of mutants 144,161,204,242,280, and 291,
retained the cooperative binding properties of the wild-type [protein] ln nghL
MP but had lower levels of RNA binding (represented by mu-
Figure 5. Quantification of Mobility Shift Assays of the Wild-Type
tant 242 in Figure 5).A second class, including mutants 27-31,
RCNMV MP and Three Phenotypic Classes of RCNMV MP Alanine
122,128,and 305,not only affected the levels of RNA binding Scanning Mutants.
by the MP but also disrupted the cooperative binding of MP
to RNA (represented by mutants 27-31 and 305 in Figure 5). Five nanograms of the 33P-labeledtranscript described above was
added to increasing amounts of the E. coli-produced MPs (as described
The loss of cooperative RNA binding was demonstrated by
in Methods) in a total volume of 20 pL. The RNA binding profiles of
the linear nature of the RNA binding curves (Figure 5) and
the four mutants, 27-31,242,278,and 305,are compared to the wild-
was confirmed by the presence of bound species with inter- type (WT) Mi? RNA binding data were obtained from quantification
mediate mobilities in the polyacrylamide gel (Figure 48,lanes of reaction mixtures electrophoresed on a discontinuous gel system,
6 to 10).Mutants 278 and 301 defined a third class of mutants as described in Methods. The dried gel was quantified on an Ambis
that was indistinguishable from the wild-type MP, affecting nei- scanner, and the resulting data were plotted, as described in Methods,
ther the leve1 nor the cooperative FINA binding in vitro. to produce the graph shown.
978 The Plant Cell

Viral RNA was isolated from plants inoculated with each of with virus titer or viral spread. Nevertheless, these data do
the six infectious MP mutants, and the RNA was sequenced. suggest that the RCNMV MP has a significant role in symp-
Five of the mutants retained the mutations that were originally tom development and that mutations in the MP may affect the
introduced (data not shown). However, mutant 280 consistently rate and/or severity of viral infection.
reverted to the wild type in three separate experiments. All six
infectious mutants elicited symptoms on N. benthamiana that
were consistent with infection by RCNMV, such as necrosis, Cell-to-Cell Movement Activity Does Not Correlate
stunting, ringspots, and mosaic patterns, but the severity of with Cooperative RNA Binding in Vitro
these symptoms varied widely. Mutants 204 and 242 caused
much more severe necrosis and stunting than the wild-type The mobility shift assay allowed us to measure the ability of

Downloaded from https://academic.oup.com/plcell/article/5/8/973/5984589 by guest on 21 December 2023

MP, whereas mutants 27-31 and 301 did not cause necrosis, each MP mutant to cooperatively bind RNA and to determine
stunting, or mosaic patterns but rather exhibited only small the relative levels of RNA binding. In addition, we were able
circular lesions (ringspots) on the inoculated and systemically to test each mutant for its ability to facilitate cell-to-cell move-
infected leaves. Plants inoculated with mutant 305 exhibited ment of the virus infection in the plant. Figure 7 summarizes
predominantly mosaic symptoms and stunting with little necro- the three activities assayed for each RCNMV MP alanine scan-
sis; those inoculated with mutant 280 were, not surprisingly, ning mutant. From the comparison of these activities, we
indistinguishable from the wild type. Although there appears concluded that the cell-to-cell movement function of the MP
to be a correlation between the severity of symptoms elicited was not dependent upon cooperative binding or on a high level
by the mutant MPs and the quantity of viral RNA observed in of RNA binding. For example, mutants 27-31 and 305, which
our RNA gel blot analysis, it is important to note that the gel facilitated cell-to-cell movement and caused systemic infec-
blot analysis was conducted using total RNA from the inocu- tion in N. benthamiana, bound to RNA poorly and did not bind
lated leaves, whereas the majority of symptom differences to RNA in a cooperative manner. In this study, no mutant MP
observed were on the systemically infected leaves. Also, the was identified that had absolutely no RNA binding activity. How-
timing and severity of symptoms do not necessarily correlate ever, mutants 27-31 and 305 displayed 20% of the level of RNA
binding activity exhibited by the wild-type MP (Figures 3 to 5,
and 7). In contrast to mutants 27-31 and 305, mutant 278 bound
cooperatively to RNA as efficiently as wild-type MP (Figures
3 to 5, and 7) but was not capable of potentiating the cell-to-
cell movement of RCNMV. Consequently, we concluded that
T-^-cMcooT-1-m
S CNI
CM
00
cvi CM
CM -* <o 5 <*
r>- -co o>
- - -o- mutation 278 disrupted a region of the MP that was essential
CM i- T- i- T- CM CMCM CM CM CO
for biological activity but not for cooperative RNA binding.

DISCUSSION

RNA-2 Deletion mutagenesis has been used previously to identify the

various functional domains in plant virus MPs. Using this ap-
proach, Citovsky et al. (1990, 1992) were able to identify two
putative RNA binding domains in the TMV MP. Similarly,
Osman et al. (1991b) and Xiong et al. (1993) demonstrated that
the C terminus of the RCNMV MP is not necessary for cell-to-
cell movement in a limited number of systemic hosts. In addi-
tion, Berna et al. (1991) and Erny et al. (1992) were able to
Figure 6. Summary of Data from Plants Inoculated with the RCNMV identify sequences that are apparently necessary for cell wall
MP Alanine Scanning Mutants and RNA Gel Blot Analysis of Total RNA
localization of the TMV and alfalfa mosaic virus MPs, respec-
from Inoculated Leaves.
tively. We have used alanine scanning mutagenesis to
Autoradiograph of a gel blot with total RNA isolated from the inocu- investigate functional domains of the RCNMV MR We believe
lated leaves of plants infected with each alanine scanning mutant. this method is both more informative and more subtle than
Approximately 20 ng of total RNA isolated from plants infected with deletion analysis because alanine substitutions are not likely
the indicated mutant was added to each well of a 1.0% agarose gel.
to significantly perturb tertiary structure (Cunningham and
RNA was transferred onto a nylon membrane and hybridized with a
Wells, 1989). Therefore, this type of mutagenesis may facili-
labeled RNA 2 cDNA clone. The blot was exposed to film for 5 hr at
room temperature. A (+) below the gel lane indicates the observation tate the isolation of mutants with interesting phenotypes.
of symptoms and a (-) a lack of symptoms on both the inoculated and Using this strategy, we have demonstrated the presence of
systemic leaves of N. benthamiana 6 to 10 days after inoculation with at least three functional domains in the RCNMV MP by charac-
each of the 12 alanine scanning mutants. terizing MP mutants both in vitro and in vivo. We confirmed
 Plant Virus Cell-to-Cell Movement 979

su--w 0.0

1.2

Mutant 27-31 122 128144 161 204 242 278 280 291 301 305

Downloaded from https://academic.oup.com/plcell/article/5/8/973/5984589 by guest on 21 December 2023

Movement t 1 1 " + + = + - + +
RNA binding 20 10 11 32 28 32 56 109 42 59 100 25
(% of w
Cooperativity I = I t + + + + + . I . + =
Figure 7. Cell-to-Cell Movement, RNA Binding, and Cooperative FINA Binding Activities for Each RCNMV MP Alanine Scanning Mutant.
A (+) indicates that the activity was detected and a (-) indicates that the activity was not detected. Numerical results were derived from data
obtained in quantificationof mobility shift assays. The surface probabilityplot (Eminiet al., 1985) illustratesthe likelihoodthat mutagenized residues
are localized to the surface of the folded protein. WT, wild type.

the results of Osman et al. (1992), who showed that the RCNMV N-terminal half of the MP appear to have a greater effect on
MP has RNA binding activity in vitro and that this binding is RNA binding than those in the C-terminal region. Mutations
cooperative. One difference between our results and those of 122 and 128, which reduce the level of RNA binding to 10%
Osman et al. (1992) is that the RCNMV isolate that we have of that of the wild-type MP, are in a region of the protein that
characterized (Australian) does not enter the polyacrylamide is either directly or indirectly involved in RNA binding. How-
gel under the electrophoretic conditions common to the two ever, this analysis does not distinguish phenotypically between
studies (a continuous, native gel system), whereas the MP of mutants that may prevent folding of the protein into an active
the Czechoslovakian or TpM-34 isolate described by Osman conformation and mutants that fold correctly but have an al-
et al. (1992) does. Although in both cases the MP clearly binds tered active site. Further, changes in RNA binding activity
to the RNA probe, one interpretationof our observations is that cannot be correlated directly with the biological activity of the
the MP may be aggregating or forming partially insoluble com- mutant MPs. For example, mutant 278 binds RNA as well as
plexes with the RNA probe prior to electrophoresis rather than the wild-type MP in vitro but is not capable of facilitating cell-
binding RNA in a cooperative manner. However, we do not be- to-cell movement in the systemic host plant N. benthamiana.
lieve that this is the case. Some of the mutant MPs do migrate Mutants 27-31 and 305 bind RNA at 20% of the level of the
into the gel (Figure 2B), for example, and among these, some wild-type protein but facilitate rapid systemic infections of N.
bind RNA cooperatively (204 and 242) and some do not (128 benthamiana.
and 305). In a discontinuous, native gel system with different Assuming that the RNA binding activity characterized in vitro
buffers and pH, the same mutant MPs do not migrate into the reflects RNA binding in vivo, it is possible that relatively little
gel (Figures 4A and 48). Consequently, we conclude that the RNA binding is actually required for cell-to-cell movement of
E. coli-produced MPs are not aggregating or becoming insolu- RCNMV ((20% of wild type). We speculate that the high lev-
ble, but rather that the electrophoretic properties (pH, pl, net, els of cooperative RNA binding exhibited by the wild-type MP
or effective charge) of the proteins determine whether the MPs may be required only under certain conditions. For example,
migrate into the gel. In addition, the TpM-34 isolate of RCNMV it may be important in systemic hosts other than N. benthami-
has only 80% amino acid sequence identity with the Australian ana or under a particular set of environmental conditions.
isolate, and near the C terminus, between amino acids 237 Preliminary evidence in our laboratory from a host range anal-
and 288, the proteins are only 44% identical (Osman et al., ysis of these alanine scanning mutants suggests that the MP
1991a). Therefore, it is not surprising that these two RCNMV may be involved in long-distance (systemic) movement of the
MPs have different electrophoretic mobilities under the same virus out of the inoculated leaves. Indeed, some of the mu-
gel conditions. tants with altered levels of RNA binding, such as 27-31 and
Using a mobility shift assay, we have shown that RCNMV 204, appear restricted to the inoculation leaves of N. edward-
MP mutations alter its RNA binding profile. Mutations in the soniiand cowpea (D.Giesman-Cookmeyer and S. A. Lommel,
980 The Plant Cell

unpublished observations). Other mutants, such as 242 and of RCNMV that still facilitates cell-to-cell movement (Xiong et
305, appear to facilitate systemic movement more rapidly than al., 1993).A deletion mutant that is lacking amino acids 279
the wild-type MP. It is possible, therefore, that the RNA bind- to 317of the RCNMV MP is only functional in N. benthamiana.
ing properties of the RCNMV MP are important for its role in This particular host has been found in many systems to be
long-distance movement. unusually permissive for viral infection. We believe that mu-
Mutant MPs 27-31,122,128,and 305 appear to disrupt a tant 278 must affect a function of the MP that is distinct from
functional domain responsible for cooperative protein-protein RNA binding and cooperative binding. This function could be
interactions during RNA binding. That is, binding of one MP folding, targeting to the plasmodesmata, modifying the plas-
molecule to RNAenhances the binding of additional MP mol- modesmata, interacting with host proteins, or trafficking
ecules. This domain is genetically distinct from that required through the plasmodesmata.

Downloaded from https://academic.oup.com/plcell/article/5/8/973/5984589 by guest on 21 December 2023

for RNA binding, as demonstrated by the isolation of mutant
MPs that interact cooperatively but have a reduced leve1 of METHODS
RNA binding (Figure 7).However, no mutants were isolated
that affected cooperative RNA binding without also affecting
the absolute levels of RNA binding. Site-Directed Mutagenesis
Finally, a third functional domain of the RCNMV MP has been
defined by the isolation of mutant 278.This mutant binds RNA Alanine scanning mutants of the red clover necrotic mosaic virus
as cooperatively and to the same extent as the wild-type MP. (RCNMV)movement protein (MP) gene were generatedby site-directed
However, it is not capable of facilitating RCNMV movement mutagenesis using oligodeoxynucleotide primers of 30 to 35 nucleo-
tides in length (Kunkel et al., 1987). A total of 12 clusters of charged
from cell to cell. In addition, three other mutants, 144,161, and
amino acid residues identified in the RCNMV MP were changed to
291,have higher levels of RNA binding and bind more cooper- alanines. Clusters of charged amino acids were identified using the
atively to RNA than infectious mutants 27-31 or 305 (Figure complete nucleotide sequence described by Lommel et al. (1988) and
7)but, again, do not facilitate cell-to-cell movement. These mu- the surface probability algorithm described by Emini et al. (1985). Each
tants may be movement defective because they disrupt a mutant was created directly in the vector pRC21G5'Nco,from which
putative third domain of the MP required for cell-to-cell move- infectious RNA 2 transcripts could be derived. An Ncol restrictionsite
ment. The TMV movement protein has been shown to localize was engineered at the initiation codon of the RCNMV MP gene in
to plasmodesmata and to alter the size exclusion limit of plas- pRC21G (Xiong and Lommel, 1991) by site-directed mutagenesis
modesmata during infection (Tomenius et al., 1987;Wolf et al., (Kunkel et al., 1987) to make the vector pRC21G"Nco. This vector was
1989;Atkins et al., 1991).It is possible that the third functional used for all of the site-directedmutagenesis and yields an infectious
RNA transcript indistinguishable from the original infectious clone.
domain of the RCNMV MP may be involved in one or both of
Reactionswere performedas described in the Mutagene Kit (BioRad).
these activities, perhaps by targeting the MP to the plasmodes- CJ236 was the dut ung Escherichia coli host used, and DH5aF' was
mata or by interacting with host factors to modify the the wild-type host. Mutant DNAwas sequenced using the Sequenase
plasmodesmata. Alternatively, this third domain may simply protocol and reagents (U.S. Biochemical Corp.).
be required for folding or for binding RNA into a functional vi-
ral ribonucleoprotein complex that is capable of cell-to-cell
movement. Citovsky et al. (1992)suggest that the TMV MP MP Expression in E. coli
must unwind the viral RNA and relieve its secondary struc-
ture for the RNA to move through the plasmodesmata. Mutants Once the 12 MP mutants were generated,the entire RNA 2 gene from
such as 278 may bind RNA but be unable to eliminate the sec- the wild-type and mutagenizedvectors was then excised by Ncol re-
ondary structure and/or unwind the RNA. striction and cloned into the Ncol site of the pET3d expression vector
(Studier and Moffat, 1986; Rosenberg et al., 1987). The resulting plas-
Interestingly, mutations 278 and 291 lie in the C-terminal re-
mids were transformed into the BLPl(DE3)pLys.Sstrain of E. coE. For
gion of 88 amino acids, which was not absolutely required for the wild type and each of the 12 mutant proteins, a 100" culture
cell-to-cell movement in cowpea (Osman et al., 1991b). Yet of the transformed bacteria was grown to an ODm of 0.6 and then
these mutations are movement defective. Although these data induced with the addition of 400 pL of 100 mM isopropyl p-o-
may seem contradictory, it is possible that deletion of the C thiogalactopyranoside. After 2 hr of incubation at P C , the cells were
terminus may not affect the function of the rest of the protein harvested and resuspended in 2.5 mL of buffer 81 (10 mM Tris-HCI,
as drastically as the presence of an altered C terminus. For pH 8.0,1 mM EDTA, 10% glycerol, and 200 mM NaCI). Cells were lysed
example, the amino acid changes in the MP mutants 278 and in a French press at 20,000 psi, and cell debris was pelleted and washed
291 alter the net charge of the protein and may affect folding twice with high-salt buffer B2 (61 + 1 M NaCI) and twice with 4 M
and/or other macromolecular interactions. As mentioned pre- urea buffer 83 (61 + 1 M NaCl + 4 M urea). Pelletswere resuspended
in 1.5 mL of 8 M urea buffer 84 (61 + 1 M NaCl + 8 M urea); the
viously, the RCNMV TpM-34 MP studied by Osman et al. (1992)
debris was then pelleted by centrifugation,and the supernatant was
is only 80% homologous to the Australian isolate that we have dialyzed at room temperature against buffers 83, 82, and 81 for 30
studied, and the greatest divergence between them is at the min each. The insoluble debris resulting from the dialysis was pelleted
C terminus. We have not isolated adeletion mutant larger than by centrifugation, and the concentrations of the soluble fractions
39 amino acids from the C terminus in the Australian isolate were determined by the modified Bradford assay (Bio-Rad). The
 Plant Virus Cell-to-Cell Movement 981

concentration of the wild-type protein was estimated using BSA as a 2 transcripts were coinoculated with the wild-type RNA 1 transcript.
standard. The relative concentrations of the 12 MPs were then deter- Each complete set of mutantswas inoculated and assayed a minimum
mined using the wild-type protein as a standard. Severa1 batches of of three times. Plants were maintained under normal glasshouse
the MPs were isolated from E. coli, including at least four for the wild- conditions.
type protein and twofrom some of the poor RNA binding MPs. In each The accumulation of viral genomic FINA with mutations in the MP
case, the isolated proteins did not vary in function or solubility, as de- gene in the inoculated leaves of N. benthamiana was determined by
termined by the assays described in the text. RNA gel blot analysis. Nucleic acids were phenol extracted from inoc-
ulated leaves 7 days postinoculation. The presence of viral RNA in
the 2 M LiCl insoluble fraction was determined as described by
Mobility Shift Assay Carrington and Morris (1984). Gel blots were hybridized with the full-
length RNA 2 cDNA and nick translated in the presence of u - ~ ~ P -

Downloaded from https://academic.oup.com/plcell/article/5/8/973/5984589 by guest on 21 December 2023

The infectious transcript vector pRC2lG”Nco was restrictedwith Ndel dCTP (Maniatis et al., 1982).
and the first 200 nucleotides of the RCNMV MP gene were uniformly
labeled with a-33P-UTPin a transcription reaction (Xiong and Lommel,
1991). The resulting RNA transcript was purified over a G50 Sepha-
dex column, extracted with phenol and chloroform, and precipitated ACKNOWLEDGMENTS
with ethanol. Typically, 5 ng of transcript and 0.01 to 4 pg of protein
were used in a binding reaction. Reactionswere conducted for 20 min
on ice in buffer 61, and the reaction products were electrophoresed We are grateful to Tim Kendall and Shelley Woloshuk for technical
at 100 V in a continuous, nondenaturing (native) 4% (19% T 1% C) assistance and to Tim Petty, Jim Moyer, Eric Miller, Olke Uhlenbeck,
polyacrylamide gel in 1 x Tris-buffered EDTA at 4OC. A discontinu- and Jack Keene for stimulating discussions and critical reviews of the
ous, native gel system was also used in some experiments. For this manuscript. This research was supported in part by grants from the
system, a stacking gel of 3.125% polyacrylamide (6.25% T 20% C), North Carolina Biotechnology Center and the U.S. Department of
pH 6.9, and a separating gel of 4% (19% T 1% C) polyacrylamide, pH Agriculture (No. 91-37303-6431).
8.48, were used. The gel was electrophoresed at 100 V with an upper
tank buffer composed of 37.6 mM Tris-HCI and 40 mM glycine, pH 8.89
and a lower tank buffer composed of 63 mM Tris-HCI, pH 7.47. 60th Received May 3, 1993; accepted June 25, 1993.
types of gel were then dried and autoradiographed from 2 to 24 hr
at room temperature. RNA binding was quantified by radioanalytic im-
aging of the dried gel on an Ambis scanner (San Diego, CA). The REFERENCES
percent of RNA bound was determined by dividing the radioactivity
measured in the retarded species by the total input radioactivity (total
counts present in each lane). The percent bound is in every case less Atabekov, J.G., and Dorokov, Y.L. (1984). Plant virus-specific trans-
than 100% for two reasons. First, some of the input radioactivity is
port function and resistance of plants to viruses. Adv. Virus Res.
in the form of unincorporated nucleotides, despite column purifica-
29, 313-364.
tion of the radiolabeled RNA probe. The unincorporated nucleotides
are easily distinguished from the probe in the discontinuous gel sys- Atabekov, J.G., and Taliansky, M.E. (1990). Expression of a plant virus-
tem and can be quantified. Second, the RNA probe is often degraded coded transport function by different viral genomes. Adv. Virus Res.
38, 201-248.
when incubated with MP mutants that do not bind RNA well, espe-
cially at higher concentrations. This can be quantified by measuring Atkins, D., Hull, R., Wells, B., Roberts, K., Moore, P., and Beachy,
the amount of radioactivity that migrates with unincorporated nucleo- R.N. (1991). The tobacco mosaicvirus 30K mwement protein in trans-
tides in the presence and absence of MI? The degree of protection genic tobacco plants is localized to plasmodesmata.J. Gen. Virol.
from degradation is, in fact, another measure of the ability of the mu- 72, 209-211.
tant MP to bind RNA. In the case of mutant 278, for example, which Berna, A., Gafny, R., Wolf, W.J., Holt, C.A., and Beachy, R.N. (1991).
binds RNA slightly better than the wild-type MP, we found that the The TMV movement protein: Role of the C-terminal73 amino acids
amount of RNA degraded by this mutant MP was lower than for the in subcellular localization and fractionation. Virology 182,682-689.
wild-type MI? Conversely, most of the other mutants exhibited higher Carrington, J.C., and Morris, T.J. (1984). Complementary DNA clon-
levels of RNA degradation. ing and analysis of carnation mottle virus FINA. Virology 139,22-31.
Chambrach, A., Jovin, T.M., Svedsen, P.J., and Rodbard, D. (1976).
Plant lnoculations and Phenotype Analysis Analytical and preparative polyacrylamidegel electrophoresis. An
of the MP Mutants objectively defined fractionation route, apparatus, and procedures.
In Methodsof Protein Separation, Vol. 2(New York: Plenum Press),
The mutant and wild-type RCNMV RNA 2 cDNA clones were linear- pp. 27-144.
ized with Smal prior to in vitro transcriptionwith bacteriophageT7 RNA Citovsky, V., Knorr, D., Schuster, G., and Zambryski, P. (1990). The
polymerase.Capped transcripts from a 50-pL reaction ( 4 0 pg RNA) P30 movement proteinof tobacco mosaic virus is a single-stranded
were resuspended in GKP buffer (50 mM glycine, 30 mM K2HP04, nucleic acid binding protein. Cell 60, 637-647.
pH 9.2, 1% bentonite, 1% celite) and inoculated into a total of four leaves Citovsky, V., Wong, M.L., Shaw, A.L., Prasad, B.V.V., and Zambryski,
on two N. benthamiana plants, as described previously (Xiong and P. (1992). Visualization and characterization of tobacco mosaic vi-
Lommel, 1991). A pair of the next to the lowest set of leaves were inoc- rus movement protein binding to single-stranded nucleic acids. Plant
ulated when the plants were at the six-to-eight leaf stage. Mutant RNA Cell 4, 397-411.
982 The Plant Cell

Culver, J.N., and Dawson, W.O. (1989). Tobacco mosaic virus coat Maniatis, T., Fritsch, E.F., and Sambrook, J. (1982). Molecular Clon-
protein: An elicitor for the hypersensitive reaction but not required ing: A Laboratory Manual. (Cold Spring Harbor, NY Cold Spring
for the development of mosaic symptoms in Nicotianasy/vestris.Virol- Harbor Laboratory).
Ogy 173, 755-758.
Melcher, U. (1990). Similarities between putative transport proteins
Cunnlngham, B.C., and Wells, J.A. (1989). High-resolution epitope of plant viruses. J. Gen. Virol. 7l, 1009-1018.
mapping of hGH-receptor interactions by alanine-scanning mula-
genesis. Science 244, 1081-1085. Osman, T.A.M., Miller, S.J., Marrlott, A.C., and Buck, K.W. (1991a).
Nucleotidesequence of RNA-2 of a Czechoslovakian isolate of red
Deom,C.M.,Schubert, K.R., Wolf, S., Holt, C.A., Lucas, W.J., and
clover necrotic mosaic virus. J. Gen. Virol. 72, 213-216.
Beachy, R.N. (1990). Molecular characterization and biological func-
tion of lhe movement protein of tobacco mosaic virus in transgenic Osman, T.A.M., Ingles, P.J., Miller, S.J., and Buck, K.W. (1991b).
plants. Proc. Natl. Acad. Sci. USA 87, 3284-3288. A spontaneous red clover necrotic mosaic virus mutant with a trun-

Downloaded from https://academic.oup.com/plcell/article/5/8/973/5984589 by guest on 21 December 2023

Deom, C.M., Lapidot, M., and Beachy, R.N. (1992). Plantvirus move- cated movement protein. J. Gen. Virol. 72, 1793-1800.
ment proteins. Cell 69, 221-224. Osman, T.A.M., Hayes, R.J., and Buck, K.W. (1992). Cooperative bind-
Eminl, E.A., Hughes, J.V., Perlow, D.C., and Boger, J. (1985). Induc- ing of the red clover necrotic mosaicvirusmovement proteintosingle
tion of hepatitis A virus-neutralizing antibody by a virus-specific stranded nucleic acids. J. Gen. Virol. 73, 223-227.
synthetic peptide. J. Virology 55, 836-839.
Robards, A.W., and Lucas, W.J. (1990). Plasmodesmata.Annu. Rev.
Erny, C., Schoumacher, F., Jung, C., Gagey, M.J., Godefroy- Plant Physiol. Plant MOI. Biol. 41, 369-419.
Colbum, T., Stussl-Garaud, C, and Berna, A. (1992). An N-proximal
Rosenberg, A.H., Lade, B.N., Chiu, O., Lin, S., Dunn, J.J., and
sequence of the alfaifa mosaic virus movement protein is neces-
Studier, F.W. (1987). Vectors for selective expression of cloned DNAs
saryforassociation with cell walls in transgenic plants. J. Gen. Virol.
by T7 RNA polymerase. Gene 56, 125-135.
73, 2115-2119.
Gafny, R., Lapldot, M., Berna, A., Holt, C.A., Deom, C.M., and Salto, T., Yamanaka, K., and Okada, Y. (1990). Long-distance move-
Beachy, R.N. (1992). Effects of terminal deletion mutationson func- ment and vira1 assembly of tobacco mosaic virus mutants. Virology
tion of the movement protein of tobacco mosaic virus. Virology 187, 176, 329-336.
499-507. Studler, F.W., and Moffat, B.A. (1986). Use of bacteriophage T7 RNA
Gibbs, A. (1976). Viruses and plasmodesmata. In lntercellular Com- polymerase to direct selective high-leve1expression of cloned genes.
munication in Plants. Studies on Plasmodesmata, B.E.S. Gunning J. Moi. Biol. 189, 113-130.
and A.W. Robards, eds (Berlin: Springer-Verlag), pp. 149-164. Tomenius, K., Clapham, D., and Meshi, T. (1987). Localization by
Hull, R. (1989). The movement of viruses in plants. Annu. Rev. immunogold cytochemistry of the virus-coded 30K protein in plas-
Phytopathol. 27, 213-240. modesmata of leaves infected with tobacco mosaic virus. Virology
Koonln, E.V., Musheglan, A.R., Ryabov, E.V., and Dolja, V.V. (1991). 160, 363-371.
Diverse groups of plant RNA and DNA viruses share related move- Wolf, S., Deom, C.M., Beachy, R.N., and Lucas, W.J. (1989). Move-
ment proteins that may possess chaperone-like activity. J. Gen. Virol.
ment protein of tobacco mosaic virus modifies plasmodesmatal size
72, 2895-2903. exclusion limit. Science 246, 377-379.
Kunkel, T.A., Roberts, J.D., and Zakour, K.A. (1987). Rapid and effi-
cient site-specific mutagenesis without phenotypic selection. In Xiong, Z.,and Lommel, S.A. (1991). Red clover necrotic mosaic vi-
Methods in Enzymology, Vol. 155, R. Wu, ed (San Diego: Academic rus infectious transcripts synthesized in vitro. Virology 182,388-392.
Press), pp. 367-382. Xlong, Z.,Kim, K.H., Giesman-Cookmeyer, D., and Lommel, S.A.
Lommel, S.A., Weston-Fina, M., Xlong, Z.,and Lomonossoff, G.P. (1993). The roles of the red clover necrotic mosaic virus capsid and
(1988). The nucleotide sequence and gene organization of red clo- cell-to-cellmovement proteins in systemic infection. Virology, 192,
ver necrotic mosaic virus RNA-2. Nucl. Acids Res. 16, 8587-8602. 27-32.