Localization of the Tomato Bushy Stunt Virus ReplicationProtein p33 Reveals a Peroxisome-to-EndoplasmicReticulum Sorting Pathway W

AndrewW. McCartney,a John S. Greenwood,a Marc R. Fabian,b K. AndrewWhite,b and Robert T. Mullena,1

a Department of Molecular and Cellular Biology, University of Guelph, Guelph, Ontario N1G 2W1, Canadab Department of Biology, York University, Toronto, Ontario M3J 1P3, Canada

Tomato bushy stunt virus (TBSV), a positive-strand RNA virus, causes extensive inward vesiculations of the peroxisomal

boundary membrane and formation of peroxisomal multivesicular bodies (pMVBs). Although pMVBs are known to contain

protein components of the viral membrane-bound RNA replication complex, the mechanisms of protein targeting to

peroxisomal membranes and participation in pMVB biogenesis are not well understood. We show that the TBSV 33-kD

replication protein (p33), expressed on its own, targets initially from the cytosol to peroxisomes, causing their progressive

aggregation and eventually the formation of peroxisomal ghosts. These altered peroxisomes are distinct from pMVBs; they

lack internal vesicles and are surrounded by novel cytosolic vesicles that contain p33 and appear to be derived from

evaginations of the peroxisomal boundary membrane. Concomitant with these changes in peroxisomes, p33 and resident

peroxisomal membrane proteins are relocalized to the peroxisomal endoplasmic reticulum (pER) subdomain. This sorting of

p33 is disrupted by the coexpression of a dominant-negative mutant of ADP-ribosylation factor1, implicating coatomer in

vesicle formation at peroxisomes. Mutational analysis of p33 revealed that its intracellular sorting is also mediated by

several targeting signals, including three peroxisomal targeting elements that function cooperatively, plus a pER targeting

signal resembling an Arg-based motif responsible for vesicle-mediated retrieval of escaped ER membrane proteins from the

Golgi. These results provide insight into virus-induced intracellular rearrangements and reveal a peroxisome-to-pER sorting

pathway, raising new mechanistic questions regarding the biogenesis of peroxisomes in plants.

INTRODUCTION

Plus-sense, single-stranded RNA viruses infect most eukaryotes

and are the predominant class of viruses that infect plants.

Tombusviruses belong to a family of positive-strand RNA plant

viruses (Tombusviridae) that possess relatively small genomes

(;4800 nucleotides) and include Cymbidium ringspot virus

(CymRSV), Cucumber necrosis virus (CNV), Carnation Italian

ringspot virus, and Tomato bushy stunt virus (TBSV), the last of

which is probably the best studied in terms of its genome

replication and recombination (reviewed in White and Nagy,

2004). The TBSV genome contains five open reading frames

(ORFs) (Figure 1). ORF1 encodes a 33-kD auxiliary replication

protein (p33), and ORF2 encodes a 92-kD RNA-dependent RNA

polymerase (p92) and is produced by the translational read-

through of the p33 amber stop codon. Both p33 and p92 are

translated directly from the viral genome in infected cells and

interact asmembrane-bound components of theRNA replication

complex (K.B. Scholthof et al., 1995; Rajendran andNagy, 2004).

The remaining three ORFs in the TBSV genome encode a coat

protein of 41 kD (ORF3), a 22-kD protein required for cell-to-cell

movement of the virus (ORF4), and a 19-kD protein that functions

as a suppressor of virus-induced gene silencing (ORF5) (H.B.

Scholthof et al., 1995).

TBSV can infect a variety of plant species, and in all cases the

most conspicuous cytopathological feature of infected cells is

the presence of multivesicular bodies (MVBs) derived from

peroxisomes (reviewed in Martelli et al., 1988). These novel

intracellular structures (referred to herein as peroxisomal multi-

vesicular bodies [pMVBs]) form initially by a progressive inward

vesiculation of the boundary membrane of preexisting perox-

isomes, resulting in the organelle’s interior (matrix) housing up to

several hundred spherical to ovoid vesicles of 80 to 150 nm in

diameter. Eventually, the boundary membrane of individual

pMVBs also produces one or more large, spherical, and vesicle-

containing extrusions that fold back and engulf portions of the

cytosol, yielding doughnut-shaped or sometimes C-shaped

pMVBs that no longer resemble the peroxisomes from which

they are derived.

Although the progressive structural reorganization of perox-

isomes into pMVBs in TBSV-infected cells has been relatively

well documented, the functional role of these complex membra-

nous compartments and the molecular mechanism(s) under-

lying their biogenesis are largely speculative. For instance,

because pMVBs are frequently observed to be in close associ-

ation with segments of the endoplasmic reticulum (ER), this

1 To whom correspondence should be addressed. E-mail rtmullen@uoguelph.ca; fax 519-837-2075.The author responsible for distribution of materials integral to thefindings presented in this article in accordance with the policy describedin the Instructions for Authors (www.plantcell.org) is: Robert T. Mullen(rtmullen@uoguelph.ca).WOnline version contains Web-only data.Article, publication date, and citation information can be found atwww.plantcell.org/cgi/doi/10.1105/tpc.105.036350.

The Plant Cell, Vol. 17, 3513–3531, December 2005, www.plantcell.orgª 2005 American Society of Plant Biologists

endomembrane compartment has been proposed as the mem-

brane source for the numerous vesiculation events that occur at

the peroxisomal boundary membrane during TBSV infection

(Martelli et al., 1988). It has also been proposed that the small

vesicles that are formedwithin pMVBs are the sites of TBSVRNA

replication, because these structures can incorporate tritiated

uridine (Appiano et al., 1983, 1986) and could serve to protect

nascent viral transcripts from host cell RNases. Consistent with

this premise, both p33 and p92 are membrane-associated

proteins; thus, the TBSV replication complex is likely anchored

onto the internal vesicles of pMVBs (K.B. Scholthof et al., 1995).

However, how nascent p33 and p92 are initially targeted to

peroxisomal membranes and participate in pMVB formation

during the TBSV infection process has not been resolved. In fact,

the only significant insights to these events have come from

studies of homologs of p33 in the related CymRSV and CNV. For

instance, modified infectious clones of CymRSV were used to

demonstrate that the N-terminal portion (including the two

predicted membrane-spanning domains) of p33 is essential for

the formation of pMVBs (Burgyan et al., 1996; Rubino and Russo,

1998). The same N-terminal region of CymRSV and CNV p33s

was also shown to contain the targeting information essential for

their sorting to peroxisomal membranes in Saccharomyces

cerevisiae, although a specific targeting sequence in either

protein was not defined (Navarro et al., 2004; Panavas et al.,

2005).

We are interested in understanding all aspects of the TBSV life

cycle, including how this virus exploits peroxisomes as the sites

for viral RNA replication and what this might tell us about plant

peroxisome biogenesis in general. Previously, we and others

showed that newly synthesized peroxisomal membrane proteins

(PMPs) are sorted either directly to peroxisomes from the cytosol

or indirectly to peroxisomes by way of specialized regions of the

ER, known as peroxisomal endoplasmic reticulum (pER), and

pER-derived vesicles (reviewed in Trelease and Lingard, 2005).

Here, we provide evidence that the TBSV replication protein p33

makes use of a novel peroxisome-to-pER vesicle-mediated

sorting pathway. Using a combination of fluorescence and

electron microscopy, as well as site-directed mutagenesis

experiments, we show that the trafficking and molecular target-

ing signals of p33 are far more complex than previously sug-

gested, especially with respect to the involvement of pER.

Overall, these results using p33 as an investigative tool have

important implications for our overall understanding of pMVB

biogenesis and the biogenetic link between peroxisomes and

pER via vesicle transport in plants.

RESULTS

Peroxisomes in TBSV-Transformed BY-2 Cells Are Altered in

Terms of Their Distribution andMorphology, Contain Both

p33 and p92, and Serve as the Sites of Viral RNA Synthesis

To determine the subcellular localization of p33, we initially

examined the protein in plant cells transformedwith TBSV cDNA.

Toward that end, a plasmid containing the full-length cDNAof the

TBSV genome that can launch autonomous TBSV replication

(pHST20; Scholthof, 1999) was introduced by biolistic bombard-

ment into tobacco (Nicotiana tabacum) Bright Yellow-2 (BY-2)

suspension-cultured cells, serving as a model system for study-

ing protein sorting in vivo (Nagata et al., 1992). Expressed p33

along with its translation read-through product, p92, were

localized using fluor-conjugated secondary antibodies bound

to IgGs raised against a synthetic peptide that corresponds to

an amino acid sequence in both proteins (Figure 1, asterisk).

Antibody specificity for both p33 and p92 was confirmed by

immunoblotting of protein extracts from BY-2 protoplasts trans-

formed with TBSV RNA (Figure 2A).

Figure 2B is a grouping of representative micrographs illus-

trating that introduced p33 and p92 colocalized precisely with

the endogenous peroxisomal matrix protein catalase in individ-

ual TBSV-transformed BY-2 cells at 2, 4, 24, and 48 h after

bombardment. These time-course results also revealed dra-

matic and progressive alterations in peroxisome distribution and

morphology. For instance, at 2 and 4 h after bombardment, most

of the peroxisomes in TBSV-transformed cells were larger in size

and fewer in number compared with peroxisomes in neighboring

nontransformed cells. Close inspection revealed that these

globular-like peroxisomes consisted of several individual perox-

isomes (Figure 2B, inset), suggesting that they formed by co-

alescence of preexisting organelles.

Changes in peroxisome morphology in TBSV-transformed

cells were even more pronounced at 24 and 48 h after bom-

bardment; p33 and p92 colocalized with catalase in several large

globular and elongated peroxisomal structures that ranged in

size up to 30 mm in length, were typically restricted to the

perinuclear region of the cell, and had amore diffuse appearance

compared with the distinct aggregated peroxisomal structures

observed at earlier time points (Figure 2B). In contrast with

peroxisomes, the morphology and distribution of other sub-

cellular organelles in TBSV-transformed cells, including mito-

chondria, ER, plastids, and Golgi, were unaltered (Figure 2C).

To determine whether the novel peroxisomal structures in

TBSV-transformed BY-2 cells also contained double-stranded

RNA (dsRNA) intermediates produced during viral replication

(White and Nagy, 2004), cells were dual-labeled with anti-p33/

p92 IgGs and monoclonal antibodies raised against dsRNA.

Figure 2D shows that at 24 h after bombardment, dsRNA in

a TBSV-transformed cell localized exclusively to distinct portions

of the globular p33- and p92-containing peroxisomes, indicating

that viral RNA synthesis occurred at these sites, presumably

Figure 1. Diagram of the TBSV Genome.

The five ORFs coding for p33, p92, p41, p22, and p19 are represented as

boxes, and noncoding regions within the genome are represented as

lines. Read-through of the amber stop codon of p33 (stippled line) results

in the translation of the p92 fusion protein. p22 and p19 are encoded by

overlapping ORFs on the same subgenomic RNA molecule. The two

thick black lines near the 59 end of the genome represent the two putative

membrane-spanning domains in p33 and p92. The asterisk highlights the

relative position of the amino acid sequence used to generate antibodies

that recognize both p33 and p92.

3514 The Plant Cell

Figure 2. Localization of p33 and p92 Replication Proteins in TBSV-Transformed BY-2 Cells.

(A) Immunoblot analysis of TBSV-transformed BY-2 protoplasts. Purified protoplasts were either mock-transformed or transformed with full-length

TBSV RNA transcripts. After a 24-h incubation period, total protein was subjected to SDS-PAGE and immunoblotting with anti-p33/p92 IgGs. The

positions of p33 and p92, as well as the molecular masses (in kD) of marker proteins, are shown at right and left of the blots, respectively.

(B) Temporal dynamics of p33 and p92 localization in BY-2 cells. Cells transformed with TBSV cDNA were fixed in formaldehyde at 2, 4, 24, or 48 h after

biolistic bombardment and then processed for immunofluorescencemicroscopy. The yellow color in the merged images (bottom row) indicates obvious

colocalizations of expressed p33 and p92 (top row) and endogenous peroxisomal catalase (middle row) in the same cells. The inset shows an enlarged

portion of the micrograph identified with an asterisk; arrowheads in the inset denote several individual (punctate) peroxisomes that form part of a larger

globular structure. Bar ¼ 10 mm.

(C) Morphology and distribution of mitochondria, ER, plastids, and Golgi in TBSV-transformed BY-2 cells. Merged images of cells transiently

transformed with p33 and p92 ([a] to [d]; green fluorescence) at 24 h after bombardment and either (immuno)stained (red fluorescence) for endogenous

mitochondrial b-ATPase (a) or endogenous ER with concanavalin A conjugated to Alexa 594 (b) or cotransformed with either plastidial glutathione

reductase fused to the green fluorescent protein (GR-GFP; Chew et al., 2003) (c) or the Golgi nucleotide sugar transporter 1 fused to the yellow

fluorescent protein (GONST1-YFP; Baldwin et al., 2001) (d). Note the similarity in the b-ATPase and concanavalin A staining in TBSV-transformed and

neighboring nontransformed cells in (a) and (b). Note also that the fluorescence attributable to coexpressed GR-GFP (c) and GONST1-YFP (d) in the

merged images was pseudocolored red. Bar in (a) ¼ 10 mm.

(D) Colocalization of p33, p92, and dsRNA in BY-2 cells. The yellow color and arrowheads in the merged image (bottom) indicate obvious

colocalizations of expressed p33 and p92 with dsRNA in a representative TBSV-transformed cell at 24 h after bombardment. Bar ¼ 10 mm.

Trafficking of Tomato Bushy Stunt Virus p33 3515

pMVBs. Neither mock transformations nor omission of anti-p33/

92 and/or anti-dsRNA antibodies yielded immunofluorescence

(data not shown).

p33 Expressed Individually in BY-2 Cells Is Localized Initially

to Aggregated Peroxisomes and Then to a Subdomain of the

ERConcomitant with the Formation of Peroxisomal Ghosts

Similar to the results presented in Figure 2B, p33 expressed on

its own colocalized with endogenous catalase in individual and

aggregated (globular) peroxisomes at the earlier stages of

expression and sorting (i.e., 2 and 4 h) (Figure 3A), suggesting

that p33 targeted directly from its site of synthesis in the cytosol

to peroxisomes. By 24 h after bombardment, however, p33

localized to two distinct subcellular compartments: globular

peroxisomes that had continued to increase in size, and a re-

ticular network that was distributed throughout the entire cell

(Figure 3A). The lack of endogenous catalase in this p33-

containing reticulum suggests that it is not formed by a marked

distension or induced elongation/tubulation of globular perox-

isomes but instead is a separate subcellular compartment(s).

At 48 h after bombardment, p33 remained localized to a diffuse

reticular network as well as to several globular structures,

although the latter were generally smaller than those observed

at 24 h after bombardment and, thus, were more difficult to

discern because of the pervasive nature of the reticular fluores-

cence (Figure 3A). Interestingly, p33-transformed cells at 48 h

after bombardment also were devoid of endogenous catalase

(cf. catalase staining in neighboring nontransformed cells in

Figure 3A). Similar results were observed when these p33-

transformed cells were immunostained for the endogenous per-

oxisomal matrix enzyme isocitrate lyase (data not shown).

The eventual loss of peroxisomal matrix enzymes in cells

expressing p33 was unexpected and suggested that the perox-

isomes in these cells were degraded (e.g., pexophagy) (Farre and

Subramani, 2004). Alternatively, the integrity of the peroxisomal

boundary membranes in these cells may have been compro-

mised such that the small globular structures observed at 48 h

after bombardment were remnants of peroxisomes that pos-

sessed membrane proteins but lacked matrix proteins, similar to

the so-called peroxisomal ghosts described in yeast and mam-

malian cells having mutations in peroxisomal matrix protein

import (reviewed in Purdue and Lazarow, 2001). To test these

possibilities, p33-transfromed cells at 48 h after bombardment

were examined for the presence or absence of two well-

characterized PMPs, ascorbate peroxidase (APX) and a 22-kD

PMP (PMP22) (Lisenbee et al., 2003a; Murphy et al., 2003, and

references therein).

For dual-labeling experiments with p33 and APX, a myc

epitope tag was appended to the C terminus of p33 (p33-myc).

p33-myc resembles wild-type p33 in that it localized to both

a diffuse reticulum and several small globular structures within

the same cell at 48 h after bombardment (Figure 3B, a),

confirming that the addition of the myc epitope to p33 does not

disrupt its normal subcellular sorting. Endogenous APX was also

readily immunodetectable in a p33-myc–transformed cell, and

this resident PMP colocalized with p33-myc in the same small

globular structures (Figure 3B, a and b, closed arrowheads),

indicating that, although lacking endogenous matrix enzymes,

these structures contain APX and likely other PMPs (i.e., perox-

isomal ghosts). Interestingly, endogenous APX in the same

transformed cell also colocalized with p33-myc in the reticular

network (Figure 3B, c and d). By contrast, endogenous APX in

neighboring nontransformed cells localized at its steady state

only to normal individual peroxisomes, as expected (Mullen et al.,

1999) (Figure 3B, b).

Transiently expressed PMP22 was also immunodetectable in

p33-cotransformed cells, and similar to endogenous APX, the

subcellular localization of PMP22 in these cells was substantially

altered. That is, although N-terminal myc-tagged PMP22 (myc-

PMP22) expressed on its own localized exclusively to normal,

individual catalase-containing peroxisomes, as expected (Murphy

et al., 2003) (Figure 3B, e and f), coexpression of myc-PMP22 and

p33 resulted in both proteins localized to the same small globular

peroxisomal ghosts and reticular network (Figure 3B, g and h).

Given the distinct morphology of the reticular compartment to

which p33 (and endogenous APX and coexpressed PMP22 in

p33-transformed cells) localizes and the fact that at least some

plant PMPs, such as APX (Mullen et al., 1999; Nito et al., 2001;

Lisenbee et al., 2003a) and PEX16 (Karnik and Trelease, 2005),

sort indirectly to peroxisomes by way of pER, we speculated that

p33wasalso targeted toER, or a portion thereof. Figure 3C, a and

b, show that the reticular fluorescence patterns attributable to

expressed p33 and concanavalin A–stained ER did not coloc-

alize in the same cell at 48 h after bombardment. A lack of

colocalization was observed also between p33 and the endog-

enous ER resident protein calreticulin (data not shown). By

contrast, p33 colocalized entirely with coexpressed chloram-

phenicol acetyltransferase (CAT) fused to the 36 C-terminal

residues of APX (CAT-APX) (Figure 3C, c and d). This portion of

APX includes the targeting signals responsible for sorting this

PMP to peroxisomes via pER; thus, when CAT-APX is overex-

pressed, the fusion protein localizes throughout its entire sorting

pathway, including pER (Mullen et al., 1999, 2001; Mullen and

Trelease, 2000).

Sorting of p33 from Peroxisomes to pER Is Not Affected

by Cycloheximide or Cold Treatment but Is Disrupted by

a Dominant-Negative Mutant of ADP-Ribosylation Factor1

Although the data presented in Figure 3 are indicative of p33

sorting initially from the cytosol to peroxisomes and then to pER,

it is also possible that newly synthesized p33 sorted directly to

pER at later time points (e.g., 12 h after bombardment) by default,

as a result of the dramatic alterations in peroxisomal integrity

caused by p33 (over)expression and/or saturation of the host cell

machinery responsible for PMP (p33) targeting and integration.

To address this possibility, p33-transformed cells at 4 h after

bombardment were incubated with the protein synthesis in-

hibitor cycloheximide. We selected 4 h for these experiments

because p33 localizes specifically to peroxisomes at this time

(Figure 3A), and observations of p33 in pER in cells that were

subsequently incubated with cycloheximide would be convinc-

ing evidence that the protein sorts from peroxisomes to pER.

Figure 4A shows that expressed p33 at 12 h after bombard-

ment localized to both globular peroxisomes and pER in either

3516 The Plant Cell

Figure 3. Localization of p33 in BY-2 Cells.

(A) Temporal dynamics of p33 localization to peroxisomal structures and a reticular network in BY-2 cells. Transformed cells were fixed in formaldehyde

at 2, 4, 24, or 48 h after bombardment and then processed for immunofluorescence microscopy. The yellow color in the merged images indicates

obvious colocalizations of expressed p33 and endogenous catalase in the same cells. Arrowheads denote the locations of several small p33-containing

globular structures. Bar ¼ 10 mm.

(B) Localization of p33, APX, and PMP22 in BY-2 cells at 48 h after bombardment. Cells were transformed with either p33-myc ([a] and [c]) or myc-

PMP22 (e) or cotransformed with myc-PMP22 (g) and p33 (h). Transformed cells were also immunostained for either endogenous APX ([b] and [d]) or

endogenous catalase (f). The portions of the p33-myc–transformed cells outlined by the stippled box in (a) and (b) are shown at higher magnification in

(c) and (d). Closed arrowheads in (a) and (b), (c) and (d), (e) and (f), and (g) and (h) indicate obvious colocalizations. The open arrowhead in (b) denotes

endogenous APX localized exclusively in individual (punctate) peroxisomes in a nontransformed cell. Bar in (a) ¼ 10 mm.

(C) Localization of p33 to pER in BY-2 cells at 48 h after bombardment. Cells were either transformed with p33 ([a] and [c]) and costained for

endogenous ER with concanavalin A conjugated to Alexa 594 (b) or cotransformed with CAT-APX (d). Arrowheads in (c) and (d) indicate obvious

colocalizations. Bar in (a) ¼ 10 mm.

Trafficking of Tomato Bushy Stunt Virus p33 3517

Figure 4. Effect of Cycloheximide, a Dominant-Negative Mutant of Arf1, and Cold Shock on p33 Localization in BY-2 Cells.

(A) Localization of p33 and CAT-APX in cycloheximide-treated BY-2 cells. Cells were transformed with either p33 or CAT-APX and, at 4 h after

bombardment, either fixed in formaldehyde or incubated for an additional 8 h (12 h total) in the presence or absence of cycloheximide, followed

by formaldehyde fixation. Transformed cells were also immunostained for endogenous catalase. Arrowheads indicate obvious colocalizations. Bar ¼10 mm.

(B) A dominant-negative mutant of Arf1 (Arf1Q71L) blocks sorting of p33 and RFP-HDEL to ER (pER). BY-2 cells were cotransformed with p33 or RFP-

HDEL and wild-type Arf1 or Arf1Q71L. Transformed cells were fixed in formaldehyde at 24 h and stained with concanavalin A (ConA) conjugated to

Alexa 394. Closed arrowheads indicate obvious colocalizations. The differential interference contrast (DIC) image corresponds to the same cell

cotransformed with p33 and Arf1Q71L. Note that RFP-HDEL in cells coexpressing wild-type Arf1 localized entirely to concanavalin A–stained ER,

whereas RFP-HDEL in cells coexpressing Arf1Q71L localized to both ER and Golgi; open arrowheads denote individual Golgi complexes. Bar¼ 10 mm.

(C) Localization of p33 and BS14a in cold-treated BY-2 cells. Cells transformed with either p33 or GFP-BS14a were incubated at either 14 or 258C for 4 h

(p33) or 8 h (GFP-BS14a). Transformed cells were then fixed in formaldehyde and (immuno)stained for endogenous peroxisomal catalase or

endogenous ER with concanavalin A conjugated to Alexa 594. Closed arrowheads indicate obvious colocalizations. The differential interference

contrast image corresponds to the same cell transformed with GFP-BS14a at 258C. Note that GFP-BS14a localizes entirely to the Golgi at 258C, as

expected (Uemura et al., 2004), but localizes only to the ER at 148C. Bar ¼ 10 mm.

3518 The Plant Cell

the absence or presence of cycloheximide, indicating that the

peroxisome is an intermediate sorting site for p33. In a positive

control experiment, CAT-APX expressed in the absence of

cycloheximide for 12 h localized throughout its sorting pathway,

including (aggregated) peroxisomes and pER, as expected

(Mullen et al., 1999) (Figure 4A). In the presence of cyclohexi-

mide, however, CAT-APX accumulated only in peroxisomal

aggregates, consistent with the sorting of this PMP to per-

oxisomes via pER. Additional control experiments with the

Golgi-localized SNARE, BS14a (green fluorescent protein

[GFP]-BS14a), confirmed the effect of cycloheximide on protein

synthesis and protein sorting in transformed BY-2 cells (see

Supplemental Figure 1 online).

The sorting of p33 fromperoxisomes to pERwas also analyzed

using a dominant-negative mutant of the ADP-ribosylation fac-

tor1 (Arf1Q71L) that inhibits multiple steps in the secretory

protein transport pathway by disrupting the formation of coat

protein complex I (COPI) coated vesicles (Pepperkok et al., 2000;

Takeuchi et al., 2002). Specifically, we tested whether blocking

COPI vesicle transport with Arf1Q71L changed the steady state

localization of coexpressed p33. Figure 4B shows that at 24 h

after bombardment, p33 coexpressedwith wild-type Arf1 did not

alter the viral protein’s normal localization to (globular) perox-

isomes and pER (cf. with the localization of p33 expressed on its

own at 24 h after bombardment in Figures 3A and 3B). By

contrast, coexpression of p33 with Arf1Q71L altered p33’s

localization at 24 h after bombardment from peroxisomes and

pER to entirely peroxisomes (Figure 4B), indicating that Arf1 is

necessary for sorting p33 from peroxisomes to pER. Similarly,

localization of the ER marker protein RFP-HDEL was altered

when coexpressed with Arf1Q71L (Figure 4B), consistent with

COPI being involved in the retrieval of escaped reticuloplasmins

from the Golgi back to the ER (Denecke, 2003).

To test whether p33 sorts initially to pER and then to perox-

isomes, but does so in such a transient manner that, unlike CAT-

APX, it is not readily detectable in pER, we examined the

localization of p33 at early time points after bombardment (i.e.,

4 h) in transformed cells incubated at 148C. Cold treatment,

among other cellular effects, disrupts protein transport from the

ER by inhibiting ER-derived vesicle formation (Bar-Peled and

Raikhel, 1997; Boevink et al., 1999; Phillipson et al., 2001). Thus,

if p33 sorts to peroxisomes by way of the ER (pER), then cold

treatment of p33-transformed cells should adversely affect this

sorting step and p33 should, at least partially, accumulate in

pER. However, as shown in Figure 4C, the sorting of p33 was

identical when cells were maintained 4 h after bombardment at

either 14 or 258C (i.e., p33 localized exclusively to peroxisomes at

both temperatures). Similar results were observed when p33-

transformed cells were incubated for 4 h after bombardment at

88C (data not shown). By contrast, sorting of GFP-BS14a from

ER to Golgi was disrupted in cells incubated at 148C (Figure 4C).

p33 Expressed in Tobacco Leaves Is Localized to the

Boundary Membranes of Aggregated Peroxisomes and

Novel Cytosolic Vesicles

As noted in the Introduction, TBSV causes progressive rear-

rangements of peroxisomeboundarymembranes that eventually

give rise to the formation of pMVBs (Martelli et al., 1988). Figures

5A to 5C illustrates the results of a series of replicate experiments

that served as a comparative control for this study. Nicotiana

benthamiana plants were rub-inoculated with TBSV RNA, and

systemically infected leaves were analyzed by transmission

electron microscopy. The representative sectional view of the

individual mesophyll cell in Figure 5A shows a cluster of five large

(3 to 5 mm) doughnut-shaped pMVBs, each consisting of

an extended membranous appendage encircling or engulfing a

large portion of the cytosol. Numerous 40- to 170-nm-diameter

vesicles are present within the lumen of each of these pMVBs,

either separate from or, on occasion, connected to the boundary

membrane (Figure 5B, inset), suggesting that they formed by

membrane invagination and vesiculation. Occasionally, pMVBs

were also found in close association with tubular membranous

structures that resembled the ER (Figure 5C), consistent with the

presumption that a biogenetic link exists between these two

compartments (Navarro et al., 2004). All other subcellular organ-

elles were unaltered in TBSV-infected cells. Controls, including

mock-rub-inoculated plants and wild-type untreated plants, did

not yield any noticeable changes in peroxisomal ultrastructure

(see the wild-type peroxisomes shown in Figures 5K and 5L).

Based on the immunofluorescence results shown in Figure 3A,

the large globular peroxisomal structures in p33-transformed

BY-2 cells at 24 h after bombardment are most likely pMVBs;

thus, p33 appears to be sufficient in forming pMVBs in the

absence of other TBSV proteins. However, testing this hypoth-

esis directly at the ultrastructural level using transiently trans-

formed BY-2 cells was difficult because of the low frequency of

cell transformation (0.5%) achieved by biolistic bombardment.

To circumvent this problem, p33 was expressed transiently in

tobacco leaves via Agrobacterium tumefaciens infiltration, a

method that yields transformation frequencies of nearly 100%

(Wroblewski et al., 2005). In addition, two GFP fusion proteins

were used, GFP-SKL (GFP linked to the C-terminal tripeptide

SKL, a prototypic matrix peroxisomal targeting signal) and p33-

GFP (full-length p33 linked to the N terminus of GFP), as vital

markers that provided a convenientmeans of assessing changes

in peroxisome morphology and distribution in infiltrated leaf

tissue before processing for electron microscopy. Three days

after Agrobacterium infiltration, GFP-SKL exhibited a punctate

fluorescence pattern consistent with this protein being localized

to matrices of individual peroxisomes (Figure 5D). By contrast,

p33-GFP exhibited a torus and globular fluorescence pattern

indicative of localization to the boundary membranes of a cluster

of aggregated peroxisomes (Figure 5F, inset), similar to the

localization of wild-type p33 to aggregated peroxisomes in BY-2

cells at 4 h after bombardment (Figure 3A).

Consistent with these fluorescence microscopy results, elec-

tron microscopic analysis of mesophyll cells transformed with

p33-GFP revealed that peroxisomes were highly aggregated,

consisting of clusters of 3 to 8, and up to 15, individual organelles

that varied in size from 200 to 1500 nm in diameter (Figures 5H to

5J). The interiors of these aggregated peroxisomes were as

devoid of vesicles as those observed in pMVBs (cf. Figures 5B

and 5H) and were only slightly more structured and electron

dense than the matrices of wild-type peroxisomes in mesophyll

cells either infiltrated with Agrobacterium containing empty

Trafficking of Tomato Bushy Stunt Virus p33 3519

Figure 5. Electron and Fluorescence Microscopic Analysis of Tobacco Mesophyll Cells Transformed with TBSV or p33-GFP.

(A) to (C) Electron micrographs of N. benthamiana leaves systemically infected with TBSV RNA. Arrowheads in (A) denote individual pMVBs in a TBSV-

infected mesophyll cell. The asterisk in (B) indicates a portion of the pMVB shown at higher magnification in the inset: arrowheads denote two distinct

vesicle-like structures located within the lumen of the pMVB that appear to be connected by a neck to the pMVB boundary membrane. Arrowheads in

(C) highlight a portion of tube-like ER adjacent to two pMVBs.

(D) to (G) Tobacco leaves infiltrated with Agrobacterium containing either GFP-SKL ([D] and [E]) or p33-GFP ([F] and [G]). The inset in (F) shows at

higher magnification the torus fluorescence pattern attributable to p33-GFP in the boundary membranes of several aggregated peroxisomes; compare

with the solid punctate fluorescence attributable to GFP-SKL in the matrix of several individual peroxisomes in (D). Differential interference contrast

images shown in (E) and (G) correspond to the same cells shown in (D) and (F), respectively.

(H) to (L) Tobacco leaves infiltrated with Agrobacterium containing either p33-GFP ([H] to [J]) or empty binary vector (pMAT037) (K) or nontransformed

(L). Asterisks in (H) to (J) denote accumulations of cytosolic vesicles adjacent to aggregated peroxisomes in p33-GFP–transformed mesophyll cells.

Arrowheads in (H) and (J) indicate examples of evaginations in the boundary membranes of aggregated peroxisomes. C, chloroplast; CW, cell wall; Cyt,

cytosol; G, Golgi; M, mitochondrion; P, peroxisome; V, vacuole. Bars in (A) to (C) and (H) to (L) ¼ 0.5 mm; bar in (D) ¼ 10 mm.

binary vector (Figure 5K) or nontransformed (Figure 5L). On the

other hand, the boundary membranes of aggregated peroxi-

somes in p33-GFP–transformed cells were largely distorted, with

contours that gave a rippled or scalloped complexion and

numerous evaginations (Figures 5H and 5J, arrowheads), sug-

gesting that nascent vesicles are externalized into the cytosol.

Consistent with this finding, collections of small (40 to 150 nm),

single membrane-bound vesicles were located in the cytosol

immediately adjacent to the aggregated peroxisomes (Figures

5H to 5J, asterisks). Similar cytosolic vesicles were detected in

mesophyll cells transformed with wild-type p33 (data not shown)

but were not observed in nontransformed cells (Figure 5L), cells

transformedwith vector alone (Figure 5K), or TBSV-infected cells

containing pMVBs (Figures 5A to 5C). Thus, although p33 is

responsible for the formation of the novel cytosolic vesicles, they

are not topologically equivalent to those formed during the

biogenesis of a so-called classical pMVB.

Immunogold labeling of the p33-GFP–transformed samples

represented in Figure 5 was performed to determine the precise

subcellular localization of the viral fusion protein. Virtually all of

the gold particles decorated the periphery (membrane) of the

novel cytosolic vesicles and their neighboring aggregated per-

oxisomes, aswell as thematrix of these peroxisomes (Figure 6A).

All other subcellular structures in these p33-GFP–transformed

cells in which the total surface area was several orders of

magnitude larger than that of the aggregated peroxisomes and

their surrounding vesicles displayed significantly lower frequen-

cies of gold particle labeling (Figures 6B and 6C), equivalent to

that observed when samples were incubated with preimmune

serum (Figure 6D).

The N-Terminal Half of p33 Possesses Both Peroxisome

and pER Targeting Information

To define the targeting information responsible for sorting p33 to

various subcellular compartments (i.e., peroxisomes and pER),

we conducted a series of targeting experiments using chimeras

consisting of different portions of p33 fused to CAT serving as

a passenger protein (Figures 7A and 7B). Although CAT alone

localized only to the cytosol of a transformedBY-2 cell, p33-CAT,

consisting of full-length p33 fused to the N terminus of CAT,

localized exclusively to aggregated (globular) peroxisomes.

These data for p33-CAT are consistent with the peroxisomal

localization of wild-type p33 in BY-2 cells at the same time point

(i.e., 4 h after bombardment; Figure 3A), indicating that the CAT

moiety does not affect p33 sorting. Similar to p33-CAT, p33

1–156-CAT, consisting of residues 1 to 156 of the p33

N terminus, including the N-terminal hydrophilic domain, trans-

membrane domains (TMDs) 1 and 2, and their intervening

hydrophilic loop sequence appended to CAT, localized to

globular peroxisomes. CAT-p33 131–296, however, was cyto-

solic (data not shown), suggesting that, unlike the N-terminal half

of p33, the C-terminal half of the protein lacks sufficient targeting

information.

p33-CAT fusion constructs consisting of smaller portions of

the p33 N-terminal 1 to 156 amino acid domain also targeted to

Figure 6. Localization of p33-GFP in Tobacco Mesophyll Cells.

Electron immunogold microscopy of tobacco leaves infiltrated with Agrobacterium containing p33-GFP. Thin sections represented in Figure 5 were

incubated with anti-p33/p92 IgGs ([A] to [C]) or preimmune serum (D) and protein A conjugated to 5-nm gold particles. The arrowheads in (A) denote

gold particles localized selectively on either cytosolic vesicles or aggregated peroxisomes, whereas the arrowheads in (B) to (D) denote gold particles

located infrequently on these or other subcellular regions (e.g., chloroplasts andmitochondria). The asterisk in (D) denotes the accumulation of cytosolic

vesicles adjacent to aggregated peroxisomes in a p33-GFP–transformed mesophyll cell. C, chloroplast; CW, cell wall; Cyt, cytosol; M, mitochondrion;

P, peroxisomes. Bars ¼ 0.5 mm.

Trafficking of Tomato Bushy Stunt Virus p33 3521

peroxisomes, albeit in an inefficientmanner (i.e., both p33 1–103-

CAT and p33 104–156-CAT localized to the cytosol and only

partially colocalized with endogenous catalase in globular and

individual peroxisomes) (Figure 7). By contrast, p33 1–75-CAT,

which includesmost of the 82–amino acid N-terminal hydrophilic

domain of p33, did not target to peroxisomes but instead

targeted entirely to pER, as indicated by its colocalization with

coexpressed GFP-APX, a fusion protein that, like CAT-APX

(Figure 3B), includes the 36 C-terminal amino acids of APX and

serves as a well-defined marker for pER while en route to

peroxisomes (Lisenbee et al., 2003b). Smaller portions of the

N-terminal 75 amino acid residues of p33 were not sufficient in

sorting CAT to pER or to peroxisomes (data not shown).

To define precisely the regions within the N-terminal half

(residues 1 to 156) of p33 responsible for its targeting to

peroxisomes and pER, we examined amino acid sequences

within the protein that resemble a prototypic membrane perox-

isomal targeting signal (mPTS) present in most other PMPs (i.e.,

a stretch of several positively charged residues adjacent to at

least one TMD) (reviewed in Trelease, 2002). We also focused on

sequences of the p33 homologs from CymRSV and CNV that

include the protein’s two TMDs and adjacent hydrophilic do-

mains and that were necessary for targeting to peroxisomes in

yeast cells (Navarro et al., 2004; Panavas et al., 2005). Alignment

of the deduced amino acid sequences for p33 from TBSV,

CymRSV, and CNV (Figure 8A) revealed a high degree of identity,

including two positively charged domains (residues 76 to 80 and

124 to 130; shaded in gray in Figure 8A) that resembled a pro-

totypic mPTS and that were located within the sequences

implicated in the sorting of CymRSV and CNV p33s to perox-

isomes.

To test whether these two positively charged domains in TBSV

p33 conferred necessary peroxisomal targeting information, all

of the basic residues at each region (underlined, -KRRQR- and

-RPSVPKK-) or at both regions were replaced with noncharged

Gly residues, and the efficiency with which the resulting mutant

proteins sorted to peroxisomes at 4 h after bombardment was

compared with that of wild-type p33 (Figures 8B and 8C). Unlike

wild-type p33, which localized only to (globular) peroxisomes at

4 h after bombardment (Figure 3A), p33-K76R77R78R80DG, p33-

R124K129K130DG, and p33-K76R77R78R80R124K129K130DG local-

ized to both globular peroxisomes (Figure 8C, a to f) and pER, as

indicated by their partial colocalizations with endogenous cata-

lase (Figure 8C, a to f), and coexpressed CAT-APX (data shown

for p33-K76R77R78R80R124K129K130DG only; Figure 8C, g and h).

These results indicate that although -K76R77R78R80- and

Figure 7. Localization of p33-CAT Fusion Proteins in BY-2 Cells.

(A) Scheme of various p33-CAT fusion proteins and their corresponding subcellular localization at 4 h after bombardment as peroxisomes (Perox),

cytosol (Cyt), or pER. The numbers in the name of each fusion construct denote the specific amino acid residues from p33 that were fused to the N or C

terminus of CAT. Black boxes in the illustrations denote the two putative TMDs in p33, with the numbers shown above p33-CAT indicating their relative

amino acid positions. Gray boxes denote bacterial CAT.

(B) Representative fluorescence patterns attributable to various constructs illustrated in (A). Each micrograph is labeled at top left with the name of the

transiently expressed (truncated) p33-CAT fusion construct, the endogenous peroxisomal marker protein, catalase, or coexpressed GFP-APX as

a marker protein for pER. The portions of the cells outlined by the stippled boxes in the bottom left two panels are shown at higher magnification in the

bottom right two panels. Data are shown for only selected constructs. Arrowheads indicate obvious colocalizations. Bar ¼ 10 mm.

3522 The Plant Cell

Figure 8. Localization of Modified Versions of p33 in BY-2 Cells.

(A) Amino acid sequence alignment of TBSV, CymRSV, and CNV p33s. Deduced amino acid sequences were obtained from GenBank and aligned using the

ClustalWalgorithm. Identical aminoacid residues ineachprotein are indicatedbyasterisks, and similar residuesare indicatedbycolons. TMDs (underlined; amino

acidresidues83to98and132to154)were identifiedusingTMHMM(version2.0)andvisual inspection.Aminoacidsequences inTBSVp33(andthecorresponding

residues in CymRSV and CNV p33s) proposed to be part of essential targeting signal elements and substituted with Gly residues are shaded gray or boxed.

(B) Subcellular localization of p33 mutants in BY-2 cells at 4 h after bombardment as peroxisomes (Perox), cytosol (Cyt), or pER. The numbers in the

name of each p33 mutant denote the specific amino acid residues that were replaced with Gly residues.

(C) Representative fluorescence patterns attributable to various constructs illustrated in (B). Cells transformed with a p33 mutant ([a], [c], [e], [i], [k],

[m], and [q]) or cotransformed with a p33 mutant ([g], [o], and [s]) and CAT-APX (serving as a pER marker protein) ([h], [p], and [t]) were fixed in

formaldehyde at 4 h after bombardment and then processed for immunofluorescence microscopy. Data are shown for only selected constructs.

(a) to (f) Transiently expressed p33-K76R77R78R80DG (a), p33-R124K129K130DG (c), p33-K76R77R78R80R124K129K130DG (e), and endogenous catalase

([b], [d], and [f]).

Trafficking of Tomato Bushy Stunt Virus p33 3523

-R124K129K130- in p33 are essential for its efficient sorting to

peroxisomes, additional targeting information exists within the

protein.

Because the N-terminal half of p33 is sufficient for sorting CAT

to peroxisomes (p33 1–156-CAT; Figure 7), any peroxisomal

targeting information in p33, in addition to regions 76 to 80 and

124 to 130, is likely located within this portion of the protein.

Indeed, a further inspection of the p33 N terminus revealed two

dibasic motifs (-K5R6- and -K11K12-; boxed in Figure 8A) that

resembled nonprototypic mPTSs identified previously in the

N termini of Arabidopsis thaliana and mammalian PMP22s

(Brosius et al., 2002; Murphy et al., 2003). These PMP22 mPTSs

are considered unique from the more common basic cluster

mPTS because they consist of a smaller stretch of positively

charged residues and are not immediately adjacent to a TMD.

Substitutions at both of these dibasic motifs in p33 with Gly

residues resulted in the mutant protein (p33-K5R6K11K12DG)

being only partially colocalized with endogenous peroxisomal

catalase (Figure 8C, i and j), indicating that this region is essen-

tial for efficiently sorting p33 to peroxisomes. Notably, p33-

K5R6K11K12DG, unlike the other p33 mutants described above,

also partially localized to the cytosol, rather than to pER (i.e., p33-

K5R6K11K12DG did not colocalize with coexpressed CAT-APX in

pER) (see Supplemental Figure 2 online), suggesting that this

region also contains (overlapping) targeting information neces-

sary for sorting p33 to pER.

The partial targeting of p33-K5R6K11K12DG to peroxisomes

was not unexpected, because at least two other peroxisomal

targeting elements within the protein (i.e., -KRRQR- and

-RPSVPKK-) were left intact. Partial localization of p33 to per-

oxisomes was also observed when either of these two regions

was mutated in the context of p33-K5R6K11K12DG (i.e., p33-

K5R6K11K12K76R77R78R80DG and p33-K5R6K11K12R124K129K130

DG both localized to peroxisomes and the cytosol; data not

shown). Only when all three of the regions essential for efficient

peroxisomal targeting of p33 were altered was the resulting

mutant protein (p33-K5R6K11K12K76R77R78R80R124K129K130DG)

no longer targeted to peroxisomes but instead accumulated

entirely in the cytosol (Figure 8B, k and l). Similar results were

observed when the same mutations were introduced into p92

(data not shown), indicating that both protein components of the

TBSV replication complex were targeted to peroxisomes in an

identical manner.

We further investigated the putative pER targeting signal within

the -K5R6K11K12- region of p33 by replacing each pair of basic

residues with Gly residues (underlined, -KRMIWPKK-). Similar to

p33-K5R6K11K12DG, p33-K11K12DG localized to peroxisomes

and pER (Figure 8C, m to p), but p33-K5R6DG localized to

peroxisomes and cytosol (Figure 8C, q to t). Similar substitutions

of -K5R6- with Gly residues in the context of the pER-localized

fusion protein p33 1–75-CAT (Figure 7) also resulted in the

mutant (p33 1–75K5R6DG-CAT) beingmislocalized entirely to the

cytosol (data not shown). Together, these data are consistent

with the -K5R6- region of p33 being essential for sorting to pER.

Disruption of p33 Targeting Signals Negatively Affects the

Replication of TBSV-Defective Interfering RNAs

The identification of distinct targeting sequences within p33

prompted us to investigate whether or not p33’s function in the

replication of viral RNAs was impaired when it was mislocalized

within the cell. Toward this end, in vivo trans-complementation

assays (Oster et al., 1998; Fabian and White, 2004) were used to

quantify viral RNA accumulation after infection under conditions

in which either p33 only was mutated or both p33 and p92 were

mutated.

Figure 9A shows the results of an analysis of viral RNA

replication with a wild-type p92 and different mutant versions

of p33 containing Gly substitutions in either one or combinations

of two or all three of the regions necessary for efficient perox-

isomal and/or pER targeting (Figure 8) (i.e., -K5R6K11K12-,

-K76R77R78R80-, and -R124K129K130-). Viral RNA replication was

assessed by detection and quantification of the accumulation of

a small viral RNA replicon, DI-72, which requires both p33 and

p92 for its replication (White and Morris, 1994). Compared with

cells expressing wild-type p33, those producing mutated forms

of p33 with modifications in one or more of the three targeting

regions showedprogressively reducedDI-72RNA levels (Figures

9A and 9C). Indeed, virtually no DI-72 RNA was detected in cells

expressing the triple mutant p33-K5R6K11K12K76R77R78R80R124

K129K130DG.

In Figure 9B, the replication of DI-72 was investigated under

conditions in which both p33 and p92 contained the targeting

mutations described above. With the exception of mutant p33/

p92-K76R77R78R80DG, the p33/p92 mutants showed similar or

more pronounced decreases in DI-72 RNA accumulation levels

than were observed for the p33 mutants (Figure 9C). Moreover,

DI-72 RNAs accumulated least in all double mutants, the triple

mutant, and the corresponding single p33 mutant that pos-

sessed alterationswithin the region responsible for pER targeting

(i.e., -K5R6K11K12-). These results indicate that robust viral RNA

replication requires proper peroxisome and/or pER targeting,

Figure 8. (continued).

(g) and (h) Transiently coexpressed p33-K76R77R78R80R124K129K130DG (g) and CAT-APX (h).

(i) to (n) Transiently expressed p33-K5R6K11K12DG (i), p33-K5R6K11K12K76R77R78R80R124K129K130DG (k), p33-K11K12DG (m), and endogenous catalase

([j], [l], and [n]).

(o) and (p) Transiently coexpressed p33-K11K12DG (o) and CAT-APX (p).

(q) and (r) Transiently expressed p33-K5R6DG (q) and endogenous catalase (r).

(s) and (t) Transiently coexpressed p33-K5R6DG (s) and CAT-APX (t).

Arrowheads indicate obvious colocalizations. Bar in (a) ¼ 10 mm.

3524 The Plant Cell

and they are consistent with the concept that such targeting

contributes to the proper assembly of the viral RNA replication

complex.

DISCUSSION

For viruses to successfully replicate within host cells, they must

manipulate and coordinate a variety of complex cellular path-

ways. As a result, viral infection often leads to an enhancement of

eventswithin the cell that are otherwise intractable to study; thus,

viruses and individual viral proteins can serve as unique tools to

investigate these cellular processes. Therefore, we used TBSV

as a means to investigate peroxisome biogenesis, a little un-

derstood event, specifically by focusing on the replication pro-

tein p33, which was demonstrated previously to be essential for

pMVB formation (Burgyan et al., 1996; Rubino and Russo, 1998).

Our findings support the notion that p33 successfully appropri-

ates different peroxisomal protein sorting pathways to facilitate

viral replication, including a previously unknown peroxisome-

to-pER sorting pathway.

p33 Causes Vesiculation of the Peroxisomal

Boundary Membrane

TBSV p33 expressed on its own caused the pronounced

accumulation of novel peroxisome-associated cytosolic vesicles

(Figure 5). These vesicles contained p33 (Figure 6), were dis-

persed among populations of aggregated peroxisomes, and

Figure 9. RNA Gel Blot Analysis of Progeny Viral RNAs Isolated from

Cucumber Protoplasts Inoculated with Various Mutant p33 RNA Tran-

scripts.

(A) Analysis of trans-expression activities of p33 mutants. Cucumber

protoplasts were cotransfected (via electroporation) with the appropriate

transcripts, and total nucleic acids were isolated after a 24-h incubation

and subjected to gel electrophoresis, RNA gel blotting, and autoradiog-

raphy. RNA transcripts used in the inoculations included the full-length

TBSV genome; DI83DII encoding wild-type p33; DI83DIIN, which is

identical to DI83DII except that a NcoI site was added to the 59 end of the

p33 ORF to allow for convenient cloning of the p33 mutants; DI83DIIN-

p33 mutants with Gly substitutions in either one or a combination of two

or all three of the regions necessary for targeting (the numbers in the

name of each mutant denote the specific amino acid residues that were

replaced with Gly residues); HS175, encoding wild-type p92; and DI-72.

Note that together, DI83DIIN (or DI83DII) and HS175 encode a functional

TBSV replicase capable of amplifying DI-72 to detectable levels. The

relative positions of TBSV genomic RNA (TBSV genome) and the two

TBSV subgenomic RNAs (sg1 and sg2 RNA) are shown at left, and the

position of DI-72 RNA is indicated at right.

(B) Analysis of trans-expression activities of p33/p92 mutants in the

context of the full-length TBSV genome. Cucumber protoplasts were

cotransfected with DI-72 and either RTD23, RTD23N, or RTD23N-p33/

p92, and DI-72 RNAs were analyzed by RNA gel blotting/autoradiogra-

phy as described for (A). RTD23 and RTD23N maintain the coding

sequence for p33 and the read-through sequence of p92, which allows

for p33 targeting sequence mutations to be integrated into both ORFs

simultaneously. Similar to DI83DIIN in (A), RTD23N contains a NcoI site

at the 59 end of the p33/p92 ORF for convenient cloning of p33/p92

mutants. The relative positions of RTD23 and DI-72 RNAs are shown at

left.

(C) Relative accumulation levels of DI-72 produced by the TBSV

replicase in the context of wild-type or mutant versions of either p33

(DI83DIIN) in (A) (black bars) or p33 and p92 (RTD23N) in (B) (white bars).

DI-72 accumulation levels in cells transfected with mutants are shown

relative to the accumulation levels obtained from wild-type p33 (DI83-

DIIN) or p33 and p92 (RTD23N) and were obtained using a phosphor

imager.

Trafficking of Tomato Bushy Stunt Virus p33 3525

most likely arose from the peroxisomal boundary membrane, as

also suggested for the novel vesicles observed in plant cells

transformed with CymRSV p33 (Bleve-Zacheo et al., 1997).

Because peroxisomes normally undergo outward vesiculation

or budding, albeit at a low frequency (Jedd and Chua, 2002), it

appears that p33 usurps the host cell molecular machinery

needed to accomplish peroxisome budding. However, the

identity of these host cell components and how they contribute

to peroxisome vesiculation remain to be determined.

The outward vesiculation of the peroxisomal boundary mem-

brane causedbyp33expressedon its ownprovides a reasonable

explanation for the relocalization of this viral protein, and of host

cell PMPs, from peroxisomes to pER (Figure 3B). The redistri-

bution of APX andPMP22 (alongwith p33) from their steady state

localization in peroxisomes (Mullen et al., 1999; Murphy et al.,

2003) to pER is likely attributable to their incorporation into the

peroxisome-derived vesicles that, after release into the cytosol,

traffic to and fuse with pER. The outward vesiculation of

peroxisomal membranes may also explain the formation of

peroxisomal membrane remnants or ghosts in p33-transformed

cells (Figure 3B); vesiculation may compromise the organelle’s

integrity such that the endogenous matrix constituents (e.g.,

catalase and isocitrate lyase) are either degraded and/or diluted

in the cytosol to below immunodetection. Indeed, transient

expression of a peroxisomal matrix marker protein (RFP-SKL)

in p33-cotransformed BY-2 cells resulted in the marker protein

being completely mislocalized to the cytosol at later time points

(e.g., 48 h; data not shown), suggesting that peroxisomal matrix

constituents, unlike p33 and host cell PMPs, are not incorpo-

rated into peroxisome-derived vesicles destined for pER.

p33 Uses a Peroxisome-to-pER Sorting Pathway in Plants

All plant PMPs examined seem to be synthesized in the cytosol

on free polyribosomes and sorted either directly to peroxisomes

or indirectly to peroxisomes via pER and pER-derived vesicles

(reviewed in Trelease and Lingard, 2005). One of the most

remarkable findings of this study is that p33 sorted initially from

the cytosol to peroxisomes and then to pER, presumably via

peroxisome-derived vesicles. Although one could argue that the

peroxisome-to-pER sorting of p33 was the result of protein

overexpression, results from several sources suggest otherwise.

For instance, in contrast with APX,which at early time points after

biolistic bombardment of BY-2 cells is readily detectable in pER

while en route to peroxisomes (Mullen et al., 1999) (Figure 4), p33

was not detectable in pER at similar times, even when cells were

cold-treated to block ER (pER) protein transport (Figures 3A and

4C). p33 also did not insert into purified microsomal membranes

in vitro (data not shown), whereas APX does so in an efficient

manner (Mullen et al., 1999), indicating that pER is not an

authentic sorting site for newly synthesized p33 in the cytosol.

Furthermore, experiments with cycloheximide (Figure 4A) and

a dominant-negative mutant of Arf1 (Arf1Q71L) (Figure 4B)

yielded the most convincing evidence that p33 localized in

peroxisomes is eventually sorted to pER.

Although our conclusion for peroxisome-to-pER sorting of

TBSV p33 is different from that drawn by Navarro et al. (2004) for

the sorting of CymRSV p33, these two proteins probably use

similar intracellular sorting pathways. Navarro and coworkers

(2004) reported that CymRSV p33 localized exclusively to

peroxisomes when expressed on its own in wild-type S. cerevi-

siae but was localized to ER in a yeast mutant lacking per-

oxisomes as a result of a disruption in the gene encoding the

PMP receptor/chaperone Pex19p. The authors suggested that

CymRSV p33 accumulated in ER in Pex19 mutant cells either

because ER served as a preperoxisome sorting site or because,

in the absence of peroxisomes, sorting to ER occurred simply by

default, as is commonly observed for yeast PMPs (Hettema et al.,

2000). We favor the latter explanation and also propose that

differences observed in sorting p33 in plant and yeast cells are

attributable to differences in the biogenetic relationship between

ER and peroxisomes in these two organisms (reviewed in

Trelease and Lingard, 2005). Thus, CymRSV p33 might not

target to ER from peroxisomes in yeast simply because yeast do

not possess this sorting pathway or because the plant viral

protein cannot use the yeast variant of this pathway.

Properties of the p33 Peroxisomal Membrane

Targeting Signal

A comprehensive analysis of the targeting information in p33

revealed that at least three distinct regions within the N-terminal

half of the protein (i.e., -K11K12-, -K76R77R78R80-, and -R124K129

K130-) constitute the mPTS and were necessary for its initial

sorting from sites of syntheses in the cytosol to peroxisomes

(Figures 7 and 8). One of these regions, -K76R77R78R80-, along

with a C-terminal region (-R213R214R215R216-), were previously

proposed to be the targeting elements responsible for the sorting

of CymRSV p33 to peroxisomes in yeast (Navarro et al., 2004).

However, neither of these regions was demonstrated experi-

mentally to be necessary and sufficient for the peroxisomal

targeting of full-length CymRSV p33, although substitution of the

-R213R214R215R216- with Ala residues prevented peroxisomal

targeting of a minimal CymRSV p33-GFP fusion protein (Navarro

et al., 2004). In our study, gain-of-function and loss-of-function

experiments demonstrated that the C-terminal half of TBSV p33

possessed no peroxisomal targeting information (Figure 7).

Although the relative position (and number) of targeting elements

within TBSV and CymRSV p33 appear to be inconsistent, these

types of mutagenic studies must be interpreted carefully. Aber-

rant protein folding and/or function, rather than disruptions in key

components of peroxisomal targeting signals, can result in

subcellular mislocation. This caveat is especially relevant with

respect to the putative C-terminal targeting element within

CymRSV p33 (Navarro et al., 2004), as this and adjacent regions

of the protein are likely involved in RNA binding (Panaviene et al.,

2003) and p33–p33/p92 protein–protein interactions (Rajendran

and Nagy, 2004).

The mPTS of p33, like most mPTSs, is relatively long, consist-

ing of the N-terminal 156 amino acid residues, including the three

proposed targeting regions and both putative TMDs. All of the

p33-CAT fusion proteins that contained shorter fragments of the

p33 N terminus were either not sorted to peroxisomes or in-

efficiently sorted to peroxisomes (i.e., partiallymislocalized to the

cytosol). Furthermore, disruptions to any of the individual target-

ing regions within the mPTS resulted in p33 being partially

3526 The Plant Cell

mislocalized to the cytosol or pER (Figures 8B and 8C) as well as

inactive in terms of its replicase function (Figure 9). Thus, the

three targeting regions and two TMDswithin themPTS appear to

cooperate to ensure the fidelity of the complexmultistep process

involved in the correct targeting, membrane integration/assem-

bly, and functionality of p33, as proposed previously for the role

of multiple targeting elements in the biogenesis of other PMPs

(Jones et al., 2001; Wang et al., 2001; Murphy et al., 2003).

p33 Possesses a pER Targeting Signal That Resembles an

Arg-Based Motif Responsible for the COPI-Dependent

Retrieval of Resident ER Membrane Proteins

In addition to a mPTS, p33 possesses a pER targeting signal

consisting, in part, of a pair of N-terminal positively charged

residues (-K5R6-). Extensivemutational analysis revealed at least

two important functional properties of the p33 pER targeting

signal. First, the signal overlaps with the N-terminal peroxisome

targeting element whose core component includes the -K11K12-

region, as sorting of p33-K5R6DG to peroxisomes was inefficient

(Figures 8B and 8C). Second, the pER targeting signal functions

from either the cytosol or peroxisomes, but it does so in

a context-dependent manner. For instance, in the case of p33

1–75-CAT, the pER targeting signal was sufficient for sorting

to pER directly from the cytosol, whereas in the context of

full-length p33, the protein was first sorted to peroxisomes (via

the mPTS) before the signal could redirect it to the pER. This

also explains why mutations to any one or more of the three

peroxisomal targeting elements in p33 resulted in the modified

protein(s) being partially localized to pER after only 4 h of

expression (Figure 8), yet wild-type p33 localized exclusively to

peroxisomes at the same time point; these mutations diminish

the overall efficiency of the mPTS and thereby allow the pER

targeting signal to function outside its normal context and

mediate sorting directly from the cytosol to pER. Such a contex-

tual dependence of the pER targeting signal makes it difficult to

elucidate how it functions in an integrative manner with the

protein’smPTS, as noted previously for the overlapping pER (ER)

targeting signal and mPTS in APX (Mullen and Trelease, 2000)

and Pex15p (Elgersma et al., 1997).

To date, several targeting signals responsible for the localiza-

tion of membrane proteins to the ER have been identified,

including cytosolically exposed C-terminal dilysines and aro-

matic amino acid–enriched motifs and N-terminal or internally

positioned Arg residues (Pelham, 2000). In each case, these

targeting signals function in the retrieval of membrane proteins

from the Golgi back to the ER by serving as recognition sites for

specific cytosolic receptor factors, includingCOPI (Aniento et al.,

2003). Because a vesicle transport pathway is most likely re-

sponsible for sorting p33 from peroxisomes to pER, the p33 pER

targeting signal might function in a manner similar to those

involved in COPI-directed transport. There are several striking

similarities between the properties of the -K5R6- region in p33

and the Arg-based ER retrieval motif, a position-independent

signal with a flexible consensus sequence (e.g., -R-R/K- and

-R/K-X-R-) that functions in a COPI-dependent manner in evolu-

tionarily diverse organisms (Schutze et al., 1994; Zerangue et al.,

2001; Yuan et al., 2003). Consistent with these observations, we

showed that replacement of -K5- with -R- in p33, yielding a

prototypic diarginine ER retrieval motif, preserved its sorting to

pER (see Supplemental Figure 3A online). Furthermore, substi-

tution of the N terminus of Arabidopsis a-glucosidase I, a type II

membrane protein that possessed a diarginine ER retrieval motif

(Gillmor et al., 2002), with the N terminus of p33, including the

-K5R6- region, preserved the ER localization of an a-glucosidase

I–GFP fusion protein (see Supplemental Figures 3B and 3C

online). COPI also appears to be involved in the sorting of p33

from peroxisomes to pER, because Arf1Q71L caused coex-

pressed p33 to accumulate exclusively in (aggregated) perox-

isomes (Figure 4B). These results suggest that the Arf1 mutation

prevented the outward vesiculation of the peroxisomal boundary

membrane, as expected based on the established role of this

protein in Golgi-derived vesicle budding (Aniento et al., 2003).

The implication that Arf1 and COPI are involved in peroxisome

vesiculation and peroxisome-to-pER sorting of p33 is not without

precedent. Rat Pex11p contains a C-terminal dilysine ER retrieval

motif that can bind COPI, and although overexpression of Pex11p

causes a proliferation of peroxisomes, cells with defects in the

coatomer vesicle coat possess peroxisomes with pronounced

morphological changes (e.g., tubulation and elongation), consis-

tent with impairment in peroxisome vesiculation (Passreiter et al.,

1998). These data led the authors to propose a dynamicmodel for

the role of Pex11p and COPI in peroxisomal vesiculation in which

the recruitment of COPI to the peroxisomemembrane ismediated

by Pex11p and probably other peroxisomal components (Anton

et al., 2000). Nascent peroxisome-derived vesicles either form into

new mature peroxisomes or function in the return of ER resident

proteins and/or peroxisomal substrates that previously escaped

from the ER via ER-derived vesicles.

Modifications and extensions of the Anton et al. (2000) model

are necessary to incorporate new and previous data on the

possible roles of p33 during pMVB formation. Nascent p33 in

TBSV-infected cells is targeted directly from the cytosol to

peroxisome membranes (via a mPTS), where it is assembled

along with p92 (Rajendran and Nagy, 2004) and viral RNA

(Panaviene et al., 2003) to form a functional replication complex.

This interaction of p33 and p92, and possibly other viral and/or

host cell factors, leads to the inward vesiculation of the perox-

isomal boundary membrane during the initial stages of pMVB

biogenesis. Because the amount of p33 in infected cells is

;20-fold higher than that of p92 (K.B. Scholthof et al., 1995),

excess p33 could also function in a manner similar to that pro-

posed for Pex11p, mediating COPI-dependent budding of the

peroxisomal boundary membrane. The resulting peroxisome-

derived vesicles containing p33 (and other peroxisomal sub-

strates such as early peroxins that function at steady state in the

ER [e.g., Arabidopsis Pex10p; Flynn et al., 2005]) would then be

transported to pER, where, by some unknown means, the viral

protein would modulate the trafficking (via pER-derived vesicles)

of additional membrane constituents required for the increase in

surface area during pMVB biogenesis. Although p33 did not

accumulate in pER in TBSV-infected cells under steady state

conditions (Figure 2B), disruption of the pER targeting signal in

p33 (and p92) in the context of the full-length TBSV genome

resulted in the virus being inactive in terms of its replicase

function (Figures 9B and 9C). Thus, the sorting of p33 to pER

Trafficking of Tomato Bushy Stunt Virus p33 3527

appears to be essential for viral replication, a premise consistent

with electron micrographs of TBSV-infected plant cells showing

pMVBs frequently connected to ER strands that were previously

suggested to be involved in pMVB biogenesis (Martelli et al.,

1988) (Figure 5C). It is possible that, similar to mammalian

Pex11p (Passreiter et al., 1998), only a small amount of p33 is

sorted from peroxisomes to pER and/or that p33 is readily

incorporated into nascent pER-derived vesicles and, thus,

localizedmostly at steady state in pMVBs in TBSV-infected cells.

Although this is a hypothetical model, the mechanism is

appealing because it incorporates pertinent research results

and interpretations of the current working models for the forma-

tion of pMVBs and peroxisome biogenetic pathways in diverse

organisms, particularly in plants. Properly conceived and exe-

cuted experiments are now needed to help resolve the role(s) of

p33 in the context of the wild-type virus and to provide direct

evidence for the intracellular trafficking of this viral protein and

how it functions with, or is modulated by, other proteins in TBSV-

infected cells to control distinct membrane vesiculation events

at peroxisomes/pMVBs and pER, in addition to its role as an

essential component of the viral RNA replication complex. These

types of studies will clearly have an impact on our understand-

ing of the complex means by which p33 along with other TBSV

proteins modify their cellular environment to facilitate viral

replication. Meanwhile, this model not only serves as a basis on

which to analyze unexplored aspects of the cellular processes

underlying TBSV replication, it also reinforces the notion of how

viruses can serve as a useful tool for studying new avenues of

peroxisome biogenesis, including the biogenetic link between

ER and peroxisomes.

METHODS

Recombinant DNA Procedures and Reagents

Standard recombinant DNA procedures were performed as described by

Sambrook et al. (1989). Restriction enzymes and other DNA-modifying

enzymes were purchased from New England Biolabs, and custom syn-

thetic oligonucleotides were obtained from either Invitrogen Life Technol-

ogies, Sigma-Aldrich, or the University of Guelph Laboratory Services.

PCR site-directed mutagenesis of plasmid DNA was performed using

a GeneAmp PCR system 2400 programmable thermal controller (Perkin-

Elmer Canada) and the QuikChange site-directed mutagenesis kit ac-

cording to the manufacturer’s recommendations (Stratagene). Isolation

of DNA fragments and plasmids was performed using Qiagen reagents.

All DNA constructs were confirmed by fluorescent dye-terminator cycle

sequencing at the University of Guelph Molecular Supercentre using an

Applied Biosystems Prism 377 automated sequencer.

Construction of Plasmids

A complete description of all plasmids used in this study and a list of the

sequences of oligonucleotide primers used in plasmid constructions are

available in the online version of this article (see supplemental protocols).

Tobacco BY-2 Cell Cultures, Microprojectile Bombardment, and

Immunofluorescence Microscopy

Tobacco (Nicotiana tabacum) BY-2 suspension-cultured cells were

maintained and prepared for biolistic bombardment as described pre-

viously (Banjoko and Trelease, 1995). Transient transformations were

conducted using 10 mg of plasmid DNA (or 5 mg of each plasmid DNA for

cotransformations) and a PDS1000 biolistic particle delivery system (Bio-

Rad Laboratories). After bombardment, BY-2 cells were incubated at

either 25 or 148C (cold shock) in the dark for 2 to 48 h to allow for

expression of the introduced gene(s) and intracellular protein sorting.

Bombarded cells were then fixed in 4% (w/v) formaldehyde, incubated

with 0.1% (w/v) pectolyase Y-23 (ICN Canada), and permeabilized with

0.3% (v/v) Triton X-100 (Sigma-Aldrich) (Mullen et al., 2001). Cyclohex-

imide (Sigma-Aldrich) was dissolved in BY-2 transformation buffer

(Banjoko and Trelease, 1995) to a final concentration of 100 mM as de-

scribed previously (Imanishi et al., 1998). Briefly, cells at 4 h after bom-

bardment were scraped from filter papers, resuspended in buffer plus

or minus cycloheximide, spread on filter papers, and incubated for an

additional 8 h before fixation.

Fixed and permeabilized BY-2 cells were processed for immunofluo-

rescence microscopy as described by Trelease et al. (1996). Primary and

dye-conjugated secondary antibodies and sources were as follows:

mouse anti-myc (clone 9E10) and rabbit anti-myc IgGs (Berkeley Anti-

body Company); mouse anti-a-tubulin IgGs (clone DM 1A) and rabbit

anti-CAT IgGs (Sigma-Aldrich); mouse anti-CAT hybridoma medium

(provided by S. Subramani, University of California, San Diego); rabbit

anti-cottonseed catalase IgGs (provided by Richard Trelease, Arizona

State University; Kunce et al., 1988); rabbit anti-cottonseed APX IgGs

(provided by R. Trelease; Corpas et al., 1994); mouse monoclonal anti-

tobacco catalase hybridoma medium (Princeton University Monoclonal

Antibody Facility; Chen et al., 1993); mouse anti-maize b-ATPase (pro-

vided by Tom Elthon, University of Nebraska; Luethy et al., 1993); mouse

anti-dsRNA hybridomamedium (clone K2) (provided byWolfgang Nellen,

Kassel University; Schonborn et al., 1991; Bonin et al., 2000); goat anti-

mouse and goat anti-rabbit Alexa 488 IgGs (Molecular Probes); and

goat anti-mouse Cy3 IgGs and goat anti-rabbit rhodamine red-X IgGs

(Jackson ImmunoResearch Laboratories). Rabbit anti-p33 IgGs raised

against a synthetic peptide corresponding to the p33 amino acid sequence

VEPARELKGKDGEDLLTGSR (residues 184 to 203) and purified using

a p33 peptide-Sepharose–linked column were generated by Bethyl Lab-

oratories. For staining ER in BY-2 cells that were also labeled with

antibodies, 5mL of a 1mg/mL stock solution of concanavalin A conjugated

to Alexa 594 or Alexa 394 (Molecular Probes) was added to the cells during

the final 20 min of a 60 min incubation with the secondary antibodies.

Fluorescent images of labeled BY-2 cells were acquired using an

Axioskop 2 MOT epifluorescence microscope (Carl Zeiss) equipped with

a Zeiss 63X Plan Apochromet oil-immersion objective and a Retiga 1300

charged-couple device camera (Qimaging). Images were deconvolved

and/or adjusted for brightness and contrast (Northern Eclipse 7.0; Empix

Imaging) and then composed into figures using Adobe Photoshop 8.0. All

fluorescent images of cells shown in individual figures are representative

of >50 transformed cells from at least two independent transformation

experiments.

Rub-Inoculation and Agrobacterium Infiltration of Tobacco Leaves

Nicotiana benthamiana plants stably transformed with pMAT037/GFP-

SKL were grown in chambers at 218C with a 12-h-light/12-h-dark cycle,

and 5 d before rub-inoculation plants were transferred to a laboratory

bench top with lower light conditions to facilitate the infection process

(Scholthof, 1999). Rub-inoculations (including mock inoculations) were

performed with six- to eight-leaf-stage plants and 5 mg of TBSV RNA

transcripts diluted in 30 mL of RNA inoculation buffer (Scholthof, 1999).

Approximately 5 to8dafter inoculations, tissues fromsystemically infected

upper leaves (Burgyan et al., 1996) were processed for electron micros-

copy.

For transformations of tobacco cv Xanthi using Agrobacterium tume-

faciens infiltration, the binary expression plasmids pMAT037/p33-GFP,

3528 The Plant Cell

pMAT037/p33, pMAT037/GFP-SKL, and pMAT037 were introduced in-

dependently into Agrobacterium (strain LBA4404) using the freeze-thaw

method of Hofgen and Willmitzer (1988). Transformed Agrobacterium

were grown overnight (Hofgen and Willmitzer, 1988), resuspended in

10 mM MgCl2 and 40 mM acetosyringone (Sigma-Aldrich), and then

infiltrated into the upper epidermis of six- to eight-leaf-stage tobacco

plants using a 10-cc syringe as described previously (Grimsley, 1995).

Infiltrated tobacco plants were incubated for 2 or 3 d at the same growth

conditions as those used before the transformation procedure (i.e., 218C

with a 12-h-light/12-h-dark cycle). Approximately 1-cm2 peels of the

lower epidermis of Agrobacterium-infiltrated leaf tissue or 1-cm2 pieces

of whole leaf tissue infiltrated with Agrobacterium were examined using

an epifluorescence microscope or processed for electron microscopy,

respectively.

Electron Microscopy and Immunogold Labeling

TBSVRNA-infiltratedN.benthamiana leaves andAgrobacterium-infiltrated

tobacco leaves were processed for transmission electron microscopy by

fixing tissue pieces in 4% (v/v) glutaraldehyde (Fisher Scientific) and 1%

(v/v) acrolein (Sigma-Aldrich) in 0.05 M KPO4 buffer, pH 7.2, for 2 h in

a vacuum chamber and then at 48C overnight. Samples were then

postfixed overnight at 48C in 1% (w/v) osmium tetroxide (Fisher Scientific),

dehydrated in a graded series of ethanol (50 to 100% [v/v]), and embedded

in Spurr’s medium-grade epoxy (Spurr, 1969). Thin sections were cut with

a glass or diamond knife, mounted on copper grids, and poststained

with 2% (w/v) uranyl acetate and Reynold’s lead citrate (Venable and

Coggeshall, 1965) for 10 and 3min, respectively. All images were acquired

at 80 kV with a Phillips CM10 transmission electron microscope.

Tobacco leaf material was processed for transmission electron mi-

croscopy immunogold labeling as described above except that postfix-

ation in osmium tetroxide was omitted and thin sections were mounted

onto nickel grids. Sections were pretreated with saturated sodium meta-

periodate followed by incubation in a blocking buffer consisting of 33

PBS, 0.05% (w/v) Gly, 0.05% (v/v) Tween 20 (Fisher Scientific), and 0.5%

(w/v) BSA. After incubation in blocking buffer, samples were incubated for

30 min in primary antibodies (rabbit anti-p33 IgGs or preimmune sera)

diluted in blocking buffer, washed a second time in blocking buffer, and

then incubated for 30 min with secondary antibodies (goat anti-rabbit

IgGs conjugated to 5-nm gold particles; Sigma-Aldrich) also diluted in

blocking buffer. Sections were then washed in 13 PBS, washed with

water, air-dried, stainedwith uranyl acetate (Greenwood et al., 2005), and

viewed using a Phillips CM10 transmission electron microscope.

Isolation and Inoculation of BY-2 and Cucumber Protoplasts

Protoplasts prepared from either 3-d-old BY-2 cells (Komoda et al., 2004)

or 6- to 8-d-old cucumber (Cucumis sativus cv Straight 8) cotyledons

(Jones et al., 1990) were quantified by bright-field microscopy using

a hemocytometer. Purified protoplasts were electroporated with 5 mg of

viral RNA transcripts as described previously (White andMorris, 1994). All

TBSV or TBSV-derived transcripts described in Figure 9 were generated

in vitro via transcription of SmaI-linearized template DNAs and using the

Ampliscribe T7 RNA polymerase transcription kit (Epicentre Technolo-

gies) (Oster et al., 1998). Isolated transcripts were quantified spectro-

photometrically and their integrity verified by agarose gel electrophoresis.

After electroporation, BY-2 protoplasts (;1.0 3 106) were pelleted by

centrifugation (500g for 5 min), resuspended in 1 mL of BY-2 culture

medium (Banjoko and Trelease, 1995), and incubated overnight at 268C in

the dark. The next day, protoplasts were centrifuged as described above

and resuspended by gentle pipetting in an equal volume of 23 SDS-

PAGE loading buffer. Aliquots of total protein were separated on 12%

SDS-polyacrylamide gels and electroblotted onto Hybond nitrocellulose

(Amersham Biosciences). Membranes were incubated with anti-p33/p92

antiserum (1:2000) and goat anti-rabbit antiserum conjugated to horse-

radish peroxidase (1:10,000), and immunoreactive proteins were visual-

ized using a Western Lighting kit (Perkin-Elmer) along with a Molecular

Dynamics Storm PhosphorImager (Amersham Biosciences).

Cucumber protoplasts (;33 105) after electroporation were incubated

in a growth chamber under fluorescent lighting at 228C for 24 h and

processed for RNA gel blot analysis as described below.

RNA Gel Blot Analysis of Viral RNAs

RNA was extracted from total nucleic acids harvested from cucumber

protoplasts 24 h after inoculation as described previously (White and

Morris, 1994). Aliquots of isolated RNA were separated in nondenaturing

1.4% (w/v) agarose gels, and viral RNAswere detected by electrophoretic

transfer to nylon (Hybond-N; Amersham Biosciences) followed by in-

cubation with a 32P-end-labeled P9 oligonucleotide probe complemen-

tary to the 39 terminal 23 nucleotides of the TBSV genome (White and

Morris, 1994). Labeled RNAswere visualized and quantified using a phos-

phor imager (Amersham Biosciences).

Accession Numbers

GenBank/EMBL accession numbers for TBSV proteins, including p33

and p92, as well as other proteins described in this study are as follows:

TBSV p33 (NP_062898), TBSV p92 (NP_062897), CymRSV p33

(NP_613261), CNV p33 (AAA42903), Arabidopsis PMP22 (NP_192356),

cottonseed APX (T09845), Arabidopsis Golgi GDP mannose transporter

(NP_178987), pea glutathione reductase (CAA62482),ArabidopsisBS14a

(NP_191376),Arabidopsis a-glucosidase I (NP_176916), andArabidopsis

Arf1 (AAA32729).

Supplemental Data

The following materials are available in the online version of this article.

Supplemental Figure 1. Localization of GFP-BS14a inCycloheximide-

Treated Tobacco BY-2 Cells.

Supplemental Figure 2. p33-K5R6K11K12DG Colocalizes with CAT-

APX in Aggregated Peroxisomes, but Not in pER, in Tobacco BY-2

Cells.

Supplemental Figure 3.p33Contains an Arg-Based ERRetrieval Motif.

ACKNOWLEDGMENTS

We thank our colleagues for their generous gifts of antibodies and

plasmids used in this study. We also thank Robb Flynn and Brian Ellis for

providing N. benthamiana and N. tabacum seeds, respectively, Rosa Di

Leo for constructing pRTL2/X-myc, Chris Trobacher for constructing

pRTL2/RFP-HDEL, and Preetinder Dhanoa for constructing monomeric

GFP cassette vectors and pRTL2/GFP-BS14a and for guidance in

experiments involving BS14a. We are also grateful to Kelly Wakeling for

constructing several p33-CAT plasmids, Kimberley Gibson for assis-

tance with sectioning material used for electron microscopy, Ian Smith

for his guidance with Adobe Photoshop, and Baozhong Meng, Annette

Nassuth, John Dyer, Peter Kim, Derek Bewley, and members of our

laboratories for their helpful discussions during the course of this work

and for their comments during the preparation of the manuscript. This

work was supported by grants from the Natural Sciences and Engi-

neering Research Council of Canada (NSERC) to J.S.G., K.A.W., and

R.T.M. K.A.W. holds a Tier II Canada Research Chair, A.W.M. was sup-

ported by an Ontario Graduate Scholarship, M.R.F. was supported by

an NSERC Postgraduate Scholarship, and K.A.W. and R.T.M. are recip-

ients of Ontario Premier’s Research in Excellence Awards.

Trafficking of Tomato Bushy Stunt Virus p33 3529

Received July 19, 2005; revised September 26, 2005; accepted October

12, 2005; published November 11, 2005.

REFERENCES

Aniento, F., Helms, J.B., and Memon, A.R. (2003). How to make

a vesicle: Coat protein-membrane interactions. In The Golgi Appara-

tus and the Plant Secretory Pathway, D.G. Robinson, ed (Oxford, UK:

Blackwell Publishing), pp. 36–62.

Anton, M., Passreiter, M., Lay, D., Thai, T.-P., Gorgas, K., and Just,

W.W. (2000). ARF- and coatomer-mediated peroxisomal vesiculation.

Cell Biochem. Biophys. 32, 27–36.

Appiano, A., Bassi, M., and D’Agostino, G. (1983). Cytochemical and

autoradiographic observations on tomato bushy stunt virus-induced

multivesicular bodies. Ultramicroscopy 12, 162.

Appiano, A., D’Agostino, G., Bassi, M., Barbieri, N., Viale, G., and

Dell’Orto, P. (1986). Origin and function of tomato busy stunt virus-

induced inclusion bodies. II. Quantitative electron microscope auto-

radiography and immunogold cytochemistry. J. Ultrastruct. Mol.

Struct. Res. 97, 31–38.

Baldwin, T.C., Handford, M.G., Yuseff, M.I., Orellana, A., and

Dupree, P. (2001). Identification and characterization of GONST1,

a Golgi-localized GDP-mannose transporter in Arabidopsis. Plant Cell

13, 2283–2295.

Banjoko, A., and Trelease, R.N. (1995). Development and application

of an in vivo plant peroxisome import system. Plant Physiol. 107,

1201–1208.

Bar-Peled, M., and Raikhel, N.V. (1997). Characterization of AtSEC12

and AtSAR1. Proteins likely involved in endoplasmic reticulum and

Golgi transport. Plant Physiol. 114, 315–324.

Bleve-Zacheo, T., Rubino, L., Melillo, M.T., and Russo, M. (1997). The

33 K protein encoded by cymbidium ringspot virus localizes to

modified peroxisomes of infected cells and of uninfected transgenic

plants. J. Plant Pathol. 79, 197–202.

Boevink, P., Martin, B., Oparka, K., Santa Cruz, S., and Hawes, C.

(1999). Transport of virally expressed green fluorescent protein

through the secretory pathway in tobacco leaves is inhibited by

cold shock and brefeldin A. Planta 208, 392–400.

Bonin, M., Oberstrab, J., Luckas, N., Ewert, K., Oesterschulze, E.,

Kassing, R., and Nellen, W. (2000). Determination of preferential

binding sites for anti-dsRNA antibodies on double-stranded RNA by

scanning force microscopy. RNA 6, 563–580.

Brosius, U., Dehmel, T., and Gartner, J. (2002). Two different targeting

signals direct human peroxisomal membrane protein 22 to perox-

isomes. J. Biol. Chem. 277, 774–784.

Burgyan, J., Rubino, L., and Russo, M. (1996). The 59-terminal region

of a tombusvirus genome determines the origin of multivesicular

bodies. J. Gen. Virol. 77, 1967–1974.

Chen, Z., Silva, H., and Klessig, D.F. (1993). Active oxygen species in

the induction of plant systemic acquired resistance by salicylic acid.

Science 262, 1883–1886.

Chew, O., Rudhe, C., Glaser, E., and Whelan, J. (2003). Character-

ization of the targeting signal of dual-targeted pea glutathione re-

ductase. Plant Mol. Biol. 53, 341–356.

Corpas, F.J., Bunkelmann, J., and Trelease, R.N. (1994). Identification

and immunochemical characterization of a family of peroxisome

membrane proteins (PMPs) in oilseed glyoxysomes. Eur. J. Cell Biol.

65, 280–290.

Denecke, J. (2003). Retrograde transport from the Golgi. In The Golgi

Apparatus and the Plant Secretory Pathway, D.G. Robinson, ed

(Oxford, UK: Blackwell Publishing), pp. 90–99.

Elgersma, Y., Kwast, L., van den Berg, M., Snyder, W.B., Distel, B.,

Subramani, S., and Tabak, H.F. (1997). Overexpression of Pex15p,

a phosphorylated peroxisomal integral membrane protein required for

peroxisome assembly in S. cerevisiae, causes proliferation of the

endoplasmic reticulum membrane. EMBO J. 16, 7326–7341.

Fabian, M.R., and White, K.A. (2004). 59-39 RNA-RNA interaction

facilitates cap- and poly(A) tail-independent translation of tomato

bushy stunt virus mRNA: A potential common mechanism for Tom-

busviridae. J. Biol. Chem. 279, 28862–28872.

Farre, J.C., and Subramani, S. (2004). Peroxisome turnover by micro-

pexophagy: An autophagy-related process. Trends Cell Biol. 14,

515–523.

Flynn, C.R., Heinze, M., Schumann, U., Gietl, C., and Trelease, R.N.

(2005). Compartmentalization of the plant peroxin, AtPex10p, within

subdomain(s) of ER. Plant Sci. 168, 635–652.

Gillmor, C.S., Poindexter, P., Lorieau, J., Palcic, M.M., and Somerville,

C. (2002). Alpha-glucosidase I is required for cellulose biosynthesis

and morphogenesis in Arabidopsis. J. Cell Biol. 156, 1003–1013.

Greenwood, J.S., Helm, M., and Gietl, C. (2005). Ricinosomes and

endosperm transfer cell structure in programmed cell death of the

nucellus during Ricinus seed development. Proc. Natl. Acad. Sci. USA

102, 2238–2243.

Grimsley, N. (1995). Agroinfection. Methods Mol. Biol. 44, 325–342.

Hettema, E.H., Girzalsky, W., van Den Berg, M., Erdmann, R., and

Distel, B. (2000). Saccharomyces cerevisiae pex3p and pex19p are

required for proper localization and stability of peroxisomal membrane

proteins. EMBO J. 19, 223–233.

Hofgen, R., and Willmitzer, L. (1988). Storage of competent cells for

Agrobacterium transformation. Nucleic Acids Res. 16, 9877.

Imanishi, S., Hashizume, K., Nakakita, M., Kojima, H., Matsubayashi,

Y., Hashimoto, T., Sakagami, Y., Yamada, Y., and Nakamura, K.

(1998). Differential induction by methyl jasmonate of genes encoding

ornithine decarboxylase and other enzymes involved in nicotine

biosynthesis in tobacco cell cultures. Plant Mol. Biol. 38, 1101–1111.

Jedd, G., and Chua, N.H. (2002). Visualization of peroxisomes in living

plant cells reveals acto-myosin-dependent cytoplasmic streaming

and peroxisome budding. Plant Cell Physiol. 43, 384–392.

Jones, J.M., Morrell, J.C., and Gould, S.J. (2001). Multiple distinct

targeting signals in integral peroxisomal membrane proteins. J. Cell

Biol. 153, 1141–1149.

Jones, R.W., Jackson, A.O., and Morris, T.J. (1990). Defective-

interfering RNAs and elevated temperatures inhibit replication of tomato

bushy stunt virus in inoculated protoplasts. Virology 176, 539–545.

Karnik, S.K., and Trelease, R.N. (2005). Arabidopsis thaliana peroxin

16 (AtPex16p) coexists at steady state in peroxisomes and endo-

plasmic reticulum. Plant Physiol. 318, 1967–1981.

Komoda, K., Naito, S., and Ishikawa, M. (2004). Replication of plant

RNA virus genomes in a cell-free extract of evacuolated plant

protoplasts. Proc. Natl. Acad. Sci. USA 101, 1863–1867.

Kunce, C.M., Trelease, R.N., and Turley, R.B. (1988). Purification and

biosynthesis of cottonseed (Gossypium hirsutum L.) catalase. Bio-

chem. J. 251, 147–155.

Lisenbee, C.S., Heinze, M., and Trelease, R.N. (2003a). Peroxisomal

ascorbate peroxidase resides within a subdomain of rough endoplasmic

reticulum in wild-type Arabidopsis cells. Plant Physiol. 132, 870–882.

Lisenbee, C.S., Karnik, S.K., and Trelease, R.N. (2003b). Overexpres-

sion and mislocalization of a tail-anchored GFP redefines the identity

of peroxisomal ER. Traffic 4, 491–501.

Luethy, M.H., Horak, A., and Elthon, T.E. (1993). Monoclonal anti-

bodies to the a- and b-subunits of the plant mitochondrial F1-ATPase.

Plant Physiol. 101, 931–937.

Martelli, G.P., Gallitelli, D., and Russo, M. (1988). Tombusviruses. In

The Plant Viruses. Vol. 3. Polyhedral Virions with Monopartite RNA

Genomes, R. Koenig, ed (New York: Plenum Press), pp. 13–72.

3530 The Plant Cell

Mullen, R.T., Lisenbee, C.S., Flynn, C.R., and Trelease, R.N. (2001).

Stable and transient expression of chimeric peroxisomal membrane

proteins induces an independent ‘‘zippering’’ of peroxisomes and an

endoplasmic reticulum subdomain. Planta 213, 849–863.

Mullen, R.T., Lisenbee, C.S., Miernyk, J.A., and Trelease, R.N. (1999).

Peroxisomal membrane ascorbate peroxidase is sorted to a membra-

nous network that resembles a subdomain of the endoplasmic

reticulum. Plant Cell 11, 2167–2185.

Mullen, R.T., and Trelease, R.N. (2000). The sorting signals for

peroxisomal membrane-bound ascorbate peroxidase are within its

C-terminal tail. J. Biol. Chem. 275, 16337–16344.

Murphy, M.A., Phillipson, B.A., Baker, A., and Mullen, R.T. (2003).

Characterization of the targeting signal of the Arabidopsis 22-kD

integral peroxisomal membrane protein. Plant Physiol. 133, 813–828.

Nagata, T., Nemoto, Y., and Hasezawa, S. (1992). Tobacco BY-2 cell

line as the ‘‘Hela’’ cell in the cell biology of higher plants. Int. Rev.

Cytol. 132, 1–30.

Navarro, B., Rubino, L., and Russo, M. (2004). Expression of the

Cymbidium ringspot virus 33-kilodalton protein in Saccharomyces

cerevisiae and molecular dissection of the peroxisomal targeting

signal. J. Virol. 78, 4744–4752.

Nito, K., Yamaguchi, K., Kondo, M., Hayashi, M., and Nishimura, M.

(2001). Pumpkin peroxisomal ascorbate peroxidase is localized on

peroxisomal membranes and unknown membranous structures. Plant

Cell Physiol. 42, 20–27.

Oster, S.K., Wu, B., and White, K.A. (1998). Uncoupled expression of

p33 and p92 permits amplification of tomato bushy stunt virus RNAs.

J. Virol. 72, 5845–5851.

Panavas, T., Hawkins, C.M., Panaviene, Z., and Nagy, P.D. (2005).

The role of the p33:p33/p92 interaction domain in RNA replication and

intracellular localization of p33 and p92 proteins of cucumber necrosis

tombusvirus. Virology 338, 81–95.

Panaviene, Z., Baker, J.M., and Nagy, P.D. (2003). The overlapping

RNA-binding domains of p33 and p92 replicase proteins are essential

for tombusvirus replication. Virology 308, 191–205.

Passreiter, M., Anton, M., Lay, D., Frank, R., Harter, C., Wieland,

F.T., Gorgas, K., and Just, W.W. (1998). Peroxisome biogenesis:

Involvement of ARF and coatomer. J. Cell Biol. 141, 373–383.

Pelham, H.R.B. (2000). Using sorting signals to retain proteins in

endoplasmic reticulum. Methods Enzymol. 327, 279–283.

Pepperkok, R., Whitney, J.A., Gomez, M., and Kreis, T.E. (2000).

COPI vesicles accumulating in the presence of a GTP restricted Arf1

mutant are depleted of anterograde and retrograde cargo. J. Cell Sci.

113, 135–144.

Phillipson, B.A., Pimpl, P., daSilva, L.L., Crofts, A.J., Taylor, J.P.,

Movafeghi, A., Robinson, D.G., and Denecke, J. (2001). Secretory

bulk flow of soluble proteins is efficient and COPII dependent. Plant

Cell 13, 2005–2020.

Purdue, P.E., and Lazarow, P.B. (2001). Peroxisome biogenesis. Annu.

Rev. Cell Dev. Biol. 17, 701–752.

Rajendran, K.S., and Nagy, P.D. (2004). Interaction between the

replicase proteins of tomato bushy stunt virus in vitro and in vivo.

Virology 326, 250–261.

Rubino, L., and Russo, M. (1998). Membrane targeting sequences in

tombusvirus infections. Virology 252, 431–437.

Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989). Molecular

Cloning: A Laboratory Manual, 2nd ed. (Cold Spring Harbor, NY:

Cold Spring Harbor Laboratory Press).

Scholthof, H.B. (1999). Rapid delivery of foreign genes into plants by

direct rub-inoculation with intact plasmid DNA of a tomato bushy stunt

virus gene vector. J. Virol. 73, 7823–7829.

Scholthof, H.B., Scholthof, K.B., Kikkert, M., and Jackson, A.O.

(1995). Tomato bushy stunt virus spread is regulated by two nested

genes that function in cell-to-cell movement and host-dependent

systemic invasion. Virology 213, 425–438.

Scholthof, K.B., Scholthof, H.B., and Jackson, A.O. (1995). The

tomato bushy stunt virus replicase proteins are coordinately ex-

pressed and membrane associated. Virology 208, 365–369.

Schonborn, J., Oberstrass, J., Breyel, E., Tittgen, J., Schumacher,

J., and Lukacs, N. (1991). Monoclonal antibodies to double-stranded

RNA as probes of RNA structure in crude nucleic acid extracts.

Nucleic Acids Res. 19, 2993–3000.

Schutze, M.P., Peterson, P.A., and Jackson, M.R. (1994). An

N-terminal double-arginine motif maintains type II membrane proteins

in the endoplasmic reticulum. EMBO J. 13, 1696–1705.

Spurr, A.R. (1969). A low-viscosity epoxy resin embedding medium for

electron microscopy. J. Ultrastruct. Res. 26, 31–43.

Takeuchi, M., Ueda, T., Yahara, N., and Nakano, A. (2002). Arf1

GTPase plays roles in the protein traffic between the endoplasmic

reticulum and the Golgi apparatus in tobacco and Arabidopsis

cultured cells. Plant J. 31, 499–515.

Trelease, R.N. (2002). Peroxisomal biogenesis and acquisition of

membrane proteins. In Plant Peroxisomes: Biochemistry, Cell Biology

and Biotechnological Applications, A. Baker and I. Graham, eds (Dor-

drecht, The Netherlands: Kluwer Academic Publishers), pp. 305–337.

Trelease, R.N., Lee, M.S., Banjoko, A., and Bunkelmann, J. (1996).

C-terminal polypeptides are necessary and sufficient for in vivo target-

ing of transiently-expressed proteins to peroxisomes in suspension-

cultured plant cells. Protoplasma 195, 156–167.

Trelease, R.N., and Lingard, M.J. (2005). Participation of the plant ER

in peroxisomal biogenesis. In The Plant Endoplasmic Reticulum (Plant

Cell Monographs), D.G. Robinson, ed (Heidelberg, Germany:

Springer-Verlag), in press.

Uemura, T., Ueda, T., Ohniwa, R., Nakano, A., Takeyasu, K., and Sat,

M.H. (2004). Systematic analysis of SNARE molecules in Arabidopsis:

Dissection of the post-Golgi network in plant cells. Cell Struct. Funct.

29, 49–65.

Venable, J.H., and Coggeshall, R. (1965). Simplified lead citrate stain

for use in electron microscopy. J. Cell Biol. 25, 407–408.

Wang, X., Unruh, M.J., and Goodman, J.M. (2001). Discrete targeting

signals direct Pmp47 to oleate-induced peroxisomes in Saccharomy-

ces cerevisiae. J. Biol. Chem. 276, 10897–10905.

White, K.A., and Morris, T.J. (1994). Nonhomologous RNA recombi-

nation in tombusviruses: Generation and evolution of defective in-

terfering RNAs by stepwise deletions. Virology 68, 14–24.

White, K.A., and Nagy, P.D. (2004). Advances in the molecular biology

of tombusviruses: Gene expression, genome replication and recom-

bination. Prog. Nucleic Acid Res. Mol. Biol. 78, 187–226.

Wroblewski, T., Tomczak, A., and Michelmore, R.W. (2005). Optimi-

zation of Agrobacterium-mediated transient assays of gene expression

in lettuce, tomato and Arabidopsis. Plant Biotechnol. J. 3, 259–273.

Yuan, H., Michelsen, K., and Schwappach, B. (2003). 14-3-3 dimers

probe the assembly status of multimeric membrane proteins. Curr.

Biol. 13, 638–646.

Zerangue, N., Malan, M.J., Fried, S.R., Dazin, P.F., Jan, Y.N., Jan,

L.Y., and Schwappach, B. (2001). Analysis of endoplasmic reticulum

trafficking signals by combinatorial screening in mammalian cells.

Proc. Natl. Acad. Sci. USA 98, 2431–2436.

Trafficking of Tomato Bushy Stunt Virus p33 3531

DOI 10.1105/tpc.105.036350; originally published online November 11, 2005; 2005;17;3513-3531Plant Cell

Andrew W. McCartney, John S. Greenwood, Marc R. Fabian, K. Andrew White and Robert T. MullenPeroxisome-to-Endoplasmic Reticulum Sorting Pathway

Localization of the Tomato Bushy Stunt Virus Replication Protein p33 Reveals a

 This information is current as of March 3, 2021

 

Supplemental Data /content/suppl/2005/11/04/tpc.105.036350.DC1.html

References /content/17/12/3513.full.html#ref-list-1

This article cites 66 articles, 28 of which can be accessed free at:

Permissions https://www.copyright.com/ccc/openurl.do?sid=pd_hw1532298X&issn=1532298X&WT.mc_id=pd_hw1532298X

eTOCs http://www.plantcell.org/cgi/alerts/ctmain

Sign up for eTOCs at:

CiteTrack Alerts http://www.plantcell.org/cgi/alerts/ctmain

Sign up for CiteTrack Alerts at:

Subscription Information http://www.aspb.org/publications/subscriptions.cfm

is available at:Plant Physiology and The Plant CellSubscription Information for

ADVANCING THE SCIENCE OF PLANT BIOLOGY © American Society of Plant Biologists