The Cell, Vol. 7, 2101-21 14, December 1995 O 1995 American Society of Plant Physiologists

Tobacco Mosaic Movement Associates with the in Tobacco Cells

8. Gail McLean, John Zupan, and Patricia C. Zambryskil Department of Plant Biology, University of California-Berkeley, Berkeley, California 94720-3102

Tobacco mosaic virus movement protein P30 complexes with genomic viral RNA for transport through plasmodesmata, the plant intercellular connections. Although most research with P30 focuses on its targeting to and gating of plasmodes- mata, the mechanisms of P30 intracellularmovement to plasmodesmata have not been defined. To examine P30 intracellular localization, we used tobacco protoplasts, which lack plasmodesmata, for transfection with plasmids carrying P30 cod- ing sequences under a constitutive promoter and for infection with particles. In both systems, P30 appears as filaments that colocalize primarily with . To a lesser extent, P30 filaments colocalize with filaments, and in vitro experiments suggested that P30 can bind directly to actin and tubulin. This association of P30 with cytoskeletal elements may play a critical role in intracellular transport of the P30-vira1 RNA complex through the cytoplasm to and possibly through plasmodesmata.

INTRODUCTION

To establish a systemic infection, plant must move from that the cytoskeleton acts as a trafficking system for intracel- the infection site to the rest of the plant. For many plant viruses, lular transport, translocating vesicles, organelles, protein, and a virus-encoded product, the movement protein, actively poten- even mRNA to specific cellular locations (Williamson, 1986; tiates viral cell-to-cell spread through plasmodesmata, the Vale, 1987; Dingwall, 1992; Singer, 1992; Wilhelm and Vale, cytoplasmic bridges that function as intercellular connections 1993; Bassell et al., 1994; Hesketh, 1994). An intriguing pos- (Gibbs, 1976; reviewed in Deom et al., 1992; Citovsky and sibility is that directional movement of the P30-RNA complex Zambryski, 1993; McLean et al., 1993; Lucas and Gilbertson, through the cytoplasm occurs on cytoskeletal components. 1994). One of the most extensively studied viral movement pro- In a somewhat analogous situation, many animal viruses teins is the P30 protein of tobacco mosaic virus (TMV), a spread through the host cell by interacting with host cell relatively simple, positive sense, single-stranded RNA virus. cytoskeletal components. Specifically, the network P30 (also termed TMV MP) is proposed to form a complex with appears to play an important role in viral protein trafficking the genomic TMV RNA, to target this protein-nucleic acid com- and distribution in animal cells (Pasick et al., 1994). For ex- plex to plasmodesmata, and to transport it through the now- ample, the viral matrix protein of vesicular stomatitus virus enlarged, open plasmodesmata (Citovsky and Zambryski, associates with tubulin in vitro and in vivo; the in vitro data 1991). Virtually all research on P30-mediated cell-cell move- show that this viral protein can bind to both polymerized and ment focuses on the interaction, that is, targeting and gating, unpolymerized tubulin (Melki et al., 1994). In sensory neurons, with plasmodesmata. Yet, synthesis of P30, replication of TMV the transport of herpes simplex virus occurs in a plus-minus RNA, and presumably, formation of the P3O-vira1 RNA complex direction on microtubules, suggesting that directional trans- all occur in the host cell cytoplasm. Therefore, the P3O-RNA port of herpes simplex virus is mediated by a minus end- complex must move through the cytoplasm to the plasmo- directed motor, such as cytoplasmic dynein (Topp et al., 1994). desmata. Like the animal cytoskeleton,the plant cytoskeleton is com- The mechanism by which P30 moves intracellularly to reach posed of filamentous networks of actin, tubulin (microtubules), the plasmodesmata is not known. Since the protein content and intermediate filaments. Although the plant cytoskeleton and organized nature of the cytoplasm probably restrict diffu- and its components are not as well characterized as those in sion of large molecular complexes, such as protein-nucleic animals, experimental data and evolutionary conservation of acid complexes (Luby-Phelps, 1993, 1994), movement of the the cytoskeletal suggest that both the general mech- P30-RNA complex to plasmodesmata most likely is not by pas- anisms and the functions of the cytoskeleton are conserved sive . Along these lines, numerous studies suggest between animals and (reviewed in Lloyd, 1982; Staiger and Lloyd, 1991; Shibaoka and Nagai, 1994). Thus, both plants To whom correspondence should be addressed and animals may use cytoskeletal filaments and motor proteins 2102 The

Figur . Filamentoue1 s Appearanc f Transientleo y Expressed Wild-Typ Tobaccn i 0 eP3 o Protoplasts P30 was detected by affinity-purified P30 polyclonal antibody and fluorescein-conjugated goat anti-rabbit secondary antibody. Arrows in (B), ) denot (E aggregates 0 d (D)eP3 an , . (A) to (C) P30 in aldehyde-fixed protoplasts. (D) and (E) P30 in a detergent-permeabilized unfixed protoplasts. (F) P30 sb-6 mutant (Citovsky et al., 1993) in an aldehyde-fixed protoplast. (F)o t .) (A urr 0 1 nfo = ) (F Ba n i r

to move macromolecular complexes, such as ribonucleopro- RESULTS tein particles additionn I . , evolutionary studies suggest that viruses act as scavengers, exploiting host cellular genes and processes and adapting them for the viruses' life cycle (Haseloff Expressio Protoplastn i 0 P3 f no s et al., 1984; Zimmern, 1988; Morozov et al., 1989; Citovsky, 1993; Kooni Doljad nan , 1993). Hypothetically virae th , l P30- To examine expressio tobaccn i TMf no 0 VP3 o protoplastse ,th RNA complexes may mimic ribonucleoprotein particles and P30 coding region was cloned into pRTL2, a plant expression use the cytoskeleton as a highway through the cytoplasm to vector containin cauliflowee gth r mosaic virus (CaMVS )35 the plasmodesmata. promoter and the untranslated leader of tobacco etch virus To begin characterizing the process of P30 movement within (Restrep t al.oe , 1990). Consequently alon0 P3 ,produce es i d througd an plane hth t cell cytoplasm examinee ,w intrae dth - in the cell, thus separating P30 expression from other TMV cellular localization of P30 in tobacco cells. P30 must be proteins. Protoplasts were prepared from tobacco cell suspen- confine singla o dt e celsucr fo l h studies preveno .T t transport sion cultures, electroporated with plasmid DNA, and incubated celle th ,f o plasmodesmat t ou 0 oP3 f a absenneee b o dt r o t 18 to 20 hr to allow the expression of the introduced DNA. P30 nonfunctional (Deo t mal.e , 1987; Mesh t al.e i , 1987; Moser was detected by affinity-purified anti-P30 polyclonal antibod- et al., 1988). Consequently, tobacco protoplasts were used. ies, followed by fluorescein-conjugated secondary antibody. Protoplast idean a e l ssystear studyinr mfo g intracellula- lo r Fluorescenc t detecteno s ewa untransfecten d i d protoplasts calizatio proteinsf no protoplase Th . t itsel singla s i f e celld ,an or with preimmune serum. importantly procese ,th removinf so cele gth l wall severs plas- As shown in Figure 1, when expressed alone in protoplasts, modesmata. Thus, P30 is restricted to a single cell, allowing appear0 P3 filamentoua s a s s network reminiscene th f o t cytoplase th n analysii 0 mP3 f beforso e targetin plasmodesgo t - cytoskeleton in both fixed (Figures 1A to 1C) and unfixed mata. Here, we show that P30 forms a filamentous network (Figures 1D and 1E) protoplasts. The appearance of P30 var- that is coincident with the cytoskeleton in tobacco protoplasts. ied from very dense, fine filaments (Figures 1A and 1C) to less Cytoskeletal Associatio f TMno 0 VP3 2103

dense, thicker filaments (Figures 1B, 1D, and 1E), depending To confirm that the localization of P30 in the transient assay particulaoe nth r protoplast being examined. Protoplasts per- reflected localizatio virus-encodef no d P30, virus-infected pro- meabilized or fixed without DMSO also showed both fine and toplasts were examined. Protoplasts, prepared from tobacco thick filaments, indicating that the presence of DMSO did not suspension cells, were inoculated with TMV in the presence cause bundling of the P30 filaments (data not shown). In some of polyethylene glycol (Maul t al.ee , 1980 incubated )an r dfo protoplasts, small aggregate coul0 seee P3 d b f nso associated 10 halloo rt w expressio P30f no . Previous studies have shown with the P30 filaments. To confirm that the filamentous net- that, in both inoculated tobacco protoplasts and infected cells work results from a specific activity of P30 and not simply from of intact tobacco , P30 accumulates to maximal amounts its overexpression substitutioa , n mutan f P30to , sb-6 alss o,wa durin earle gth y infectiostageV TM f so n (Kibersti al.t se , 1983; transfecte expressed dan d in tobacco protoplasts (Citovskt ye Watanab t al.ee , 1984; Blu t mal.e , 1989; Leht t al.oe , 1990). al., 1993). Although wild-type P30, a tenacious single-stranded In an electron microscopy study, Meshi et al. (1992) found P30 nucleic acid binding protein, bind singled an s A bot- hRN nea nucleue th r TMV-infecten si d protoplasts; however, dur- stranded DNA (Citovsky et al., 1990), sb-6 binds RNA but no ing later stages of virus infection (15 to 24 hr after inoculation), longer binds single-stranded DNA; sb- phose 6b alsn - oca observeP3s 0wa confinet da d cytoplasmareae th r n si ou n I . phorylate cele th l y wall-associatedb d kinase describen di studies, fluorescence microscopy has been used because it Citovsk . (1993)al t ye shows A . Figurn i , sb-e1F 6 doet sno permits gentler and potentially less disruptive fixation of P30- form the filamentous network typical of wild-type P30 in pro- expressing cells. toplasts but rather appears as more random aggregates in the detecteP3s 0wa d with affinity-purified anti-P30 polyclonal cell. antibodies followed by fluorescein-conjugated secondary

B

Figur . Associatioe2 wit0 hP3 f Microtubuleno TMV-lnfecten si d Protoplasts. P30 was detected by affinity-purified P30 polyclonal antibody and fluorescein-conjugated goat anti-rabbit secondary antibody. Microtubules were detected by monoclonal tubulin antibody and rhodamine-conjugated goat anti-mouse secondary antibody. Arrows highlight areas showing colocalized P30 and microtubules. (A) P30 in an aldehyde-fixed protoplast. ) Microtubule(B aldehyde-fixede th n si protoplast show (A)n i . anothen i 0 P3 r) aldehyde-fixe(C d protoplast. (D) Microtubules in the aldehyde-fixed protoplast shown in (C). Bar in (A) = 3.5 urn for (A) and (B); bar in (C) = 3.1 urn for (C) and (0). 2104 Plane Th t Cell

Figur . Colocalizatioe3 wit0 hP3 t Microtubulesno . Wild-typ detectes wa 0 affinity-purifiey deb P3 polyclona0 dP3 l antibod fluorescein-conjugated yan d goat anti-rabbit secondary antibody. Microtu- bules were detected by monoclonal tubulin antibody and rhodamine-conjugated goat anti-mouse secondary antibody. Arrows in (A) to (D) indicate colocalize microtubulesd an 0 dP3 . (A) P30 in an aldehyde-fixed protoplast. (B) Microtubule aldehyde-fixee th n si d protoplast show (A)n i . (C) P30 in another aldehyde-fixed protoplast. (D) Microtubules in the aldehyde-fixed protoplast shown in (C). (E) P30 in an aldehyde-fixed protoplast after incubation at 0°C to depolymerize microtubules. (F) Microtubules in the aldehyde-fixed protoplast shown in (E) after incubation at 0°C. (F)d .an ) (E r fo n ur 0 1 = ) (F n i r (D)o t ba ) ; (A r fo n ur 0 1 = ) (D Ba n i r

antibody. Fluorescence was not detected in uninfected proto- because a genetically altered form of P30 that was transiently plasts or with preimmune serum. As shown in Figures 2A and expresse t higda h levels faile foro dt m filaments (Figure 1F). 2C, P30 appears as filamentous structures and as distinct spots Thus, the P30 transient expression assay provides a simple in protoplasts 10 hr after TMV infection. The presence of P30 system to study cytoplasmic regionalization of newly synthe- filamentspot0 welP3 s sa s a l virus-infecten si d cellhavy sema sized P30. Since TMV-infected protoplasts accumulate other resulted from P30 interacting with other viral proteins. For in- virus-encoded proteins that potentiall interfery yma e wit0 hP3 stance, P30 may accumulate at regions of TMV replication and localization, the transient expression system was used to fur- protein synthesis, in other words, near polyribosomes and the ther characterize P30 expression in protoplasts. foun inclusiodn i n bodies thatypicae tar l of TMV-infected cells observes A . fluorescency db e micros- virus-infecten copyi amoune 0 th , P3 f o t d protoplasts wa s noticeably higher tha P30-transfecten i d protoplasts, where Interactio wit0 hP3 f Microtubuleno Actid an sn P30 is expressed from a CaMV 35S promoter. Filaments Since both virus-infecte P30-transfected dan d protoplasts display filamentous P30 localizatioe ,th transientlf no y expressed In virus-infected protoplasts and in transfected protoplasts ex- P30 appeared to mimic that of P30 during virus infection. This pressing P30 alone, P30 produces a filamentous network. localization most likely represents a biological function of P30 Since the organization of this P30 network resembles the plant Cytoskeletal Associatio0 P3 210V 5TM f no

cytoskeleton examinee ,w localizatioe do th tw wit 0 e hP3 th f no loso t r decreaseso g in d numbe f filamentso r . Indeed, cold major components of the plant cytoskeleton: microtubules and treatment had a dramatic effect on the P30 network, typically actin filaments. For studies comparing the possible colocal- causing a large increase in P30 aggregates, with a concomi- ization of P30 filaments and microtubules, P30 was detected tant decrease in P30 filaments (Figure 3E). Thus, the cold with affinity-purified antibodies and fluorescein-conjugated treatment that disrupte microtubulee dth s (Figur als) - eo3F de secondary antibody; microtubules were detected with a mono- stroye filaments0 P3 d e mosth f . o tAlthoug breakdowe hth n clonal anti-tubulin antibody and rhodamine-conjugated of the microtubule network produced diffuse tubulin staining secondary antibody. Controls showed that rhodamine fluores- and shortened microtubules, the loss of P30 filaments resulted contaminatt cencno d edi fluoresceie eth n signal exampler fo ; , in aggregates or spots of P30. Interestingly, in some cold- protoplast t expressins no t reactin bu 0 ggP3 wit tubulie hth n treated protoplasts, a few P30 filaments remained after intact antibod product no d y di signae fluoresceia e th n i l n channel microtubules coul longeo dn detectee b r d (dat t shown)ano , (compare lower portion f Figure2B)d so an . A s2 suggestin gfilament0 tha associt P3 subse a no te d th sf -di o t As demonstrate numbe a fila 0 , Figure3 n di - P3 d f o r an s2 ate with microtubules. ments were observed to colocalize with microtubules in both Since the other major component of the cytoskeleton, the virus-infecte P30-transfected dan d celld san (FigureD 2 o t A s2 actin filament network t disrupteno s col,wa y db d treatment, Figure 3D)o t confir o A T . s3 interact0 mP3 thae th ts specifi- potentiaa l interactio wit 0 actie hP3 th f nno cytoskeletos nwa cally with microtubules, transfected protoplasts were incubated assayed. Again, P30 was detected with affinity-purified anti- at 0°C to depolymerize the microtubules before fixation; this bodie fluorescein-conjugated san d secondary antibody. Actin cold treatment did not depolymerize the actin filaments (data was detected with rhodamine-phalloidin, which binds actin fila- t shown)no filament0 P3 . s that result from interaction with ments. As noted previously, rhodamine fluorescence did not microtubules should be disrupted by the cold treatment, lead- appea fluoresceie th n i r n channel. Figur showe4 s that some

Figure 4. Colocalization of P30 with Actin Filaments. Wild-type P30 was detected by affinity-purified P30 polyclonal antibody and fluorescein-conjugated goat anti-rabbit secondary antibody. Actin filaments were detected with rhodamine-conjugated phalloidin ) denot(D o .t Arrowe ) colocalize(A actid n si an n 0 filamentsdP3 . (A) P30 in an aldehyde-fixed protoplast. Note the association of P30 with the nucleus in this cell. ) Acti(B n filament aldehyde-fixee th n si d protoplast show (A)n i . anothen i 0 P3 r) aldehyde-fixe(C d protoplast. (D) Actin filaments in the aldehyde-fixed protoplast shown in (C). aldehyde-fixen a n i 0 P3 ) (E d protoplast after pretreatment with cytochalasi disrupo t nD t actin filaments. same ) Actith (F en ni aldehyde-fixed protoplast show) afte(E rn i pretreatmen t with cytochalasi. nD (F)o t .) (C urr 0 n1 fo = ) (D n i r (B)d ba ;an ) (A urr 0 n1 fo = ) (B Ba n i r 2106 The Plant Cell

Figur . eTransien5 t Expressio P30::GFe th f no P Fusion Protei Tobaccn i o Protoplasts. A P30::GFP fusio detectes nwa autofluorescency db f GFPeo . (A) Expression of the P30::GFP fusion in an unfixed protoplast 18 hr after transfection. ) Expressio(B P30::GFe th f no P fusioaldehyde-fixen a n i d protoplas r afteh 8 r1 ttransfection . (C) Expression of GFP alone in an unfixed protoplast 18 hr after transfection. (D) Expression of GFP alone in an aldehyde-fixed protoplast 18 hr after transfection. ) Expressio(E GFP::P3e th f no 0 fusiounfixen a n i d protoplas r afteh 8 r1 ttransfection . (F) Expression of the GFP::P30 fusion in an aldehyde-fixed protoplast 18 hr after transfection. ) Expressio(G P30::GFe th f no P fusio unfixen i d protoplast r afteh 8 rs 4 transfection . (H) Expression of the P30::GFP fusion in an aldehyde-fixed protoplast 48 hr after transfection. Bar in (B) = 10 urn for (A), (B), (E), and (F); bar in (D) = 10 urn for (C), (D), (G), and (H).

of the P30 filaments colocalized with the rhodamine-phalloidin- the actin filaments appeared broke d shortenean n y b d labeled actin cytoskeleton (Figure 4D)o t - A . s4 Thusap 0 P3 , cytochalasin. Cytochalasin D-induced changes in the P30 net- pear associato st e wit t leas hcomponenta o tw t plane th f tso work were less obvious (Figure 4E). Base observationn do s cytoskeleton: microtubules and actin filaments. However, the of large numbers of P30-expressing protoplasts, the number number and extent of P30 filaments that colocalized with actin and sometimes the length of the P30 filaments appeared to filaments typically were lower than those seen filawit0 -hP3 slightle b y decrease cytochalasiy db protoplastw fe A . nD s also ment microtubuled san s (Figure 4D)d an . OccasionallyC s4 , were found in which cytochalasin D treatment disrupted the P30 filaments were observed to colocalize primarily with ac- majority of P30 filaments (data not shown), correlating with filamentn ti show s rathe) s(a 4B Figurenn d i r thaan A ns4 with the previous observation that P30 colocalized primarily with microtubules. Since the protoplasts were not synchronized, actin filament smala n si l numbe f protoplastso r generaln I . , P30 may have associated primarily with actin during a limited, however, when compared with cold-treated protoplasts, most specific stage in the cell cycle. P30 filaments remained intact in cytochalasin D-treated pro- To evaluate further the association of P30 with actin filaments, toplasts. The observation that only a subset of the P30 filaments protoplasts were incubated with cytochalasin D to disrupt the appears affected by cytochalasin D corresponds to the pre- actin filaments before fixation (Cooper, 1987). If P30 binds ac- vious observation that mosfilament0 P3 t s colocalize with filamentsn ti patter localizatioe 0 ,th P3 f no n after cytochalasin microtubules. D treatment should be different from that seen without treat- ment. Both rhodamine-phalloidi actid nan n antibodies were used to check disruption of the actin filaments. Decreases in Potential Dynamic Nature of P30 Filaments bot lengte hth numbed han f actio r n filaments were seen i cytochalasin D-treated protoplasts (Figure 4F). Unlike the ex- Since cytoskeletal elements, such as microtubules, appear to tensive depolymerization of microtubules by cold treatment, play a role in protein trafficking in animal cells, it is plausible Cytoskeletal Association of TMV P30 2107

thaassociatioe th t wit0 hP3 f cytoskeletano l elementy sma cytoskeletal filaments. Most of the P30::GFP at the later time potentiat plasmodesmatae e th transpor o t 0 P3 f to teso T .t this points appeare aggregates da s alon periphere gth proe th -f yo hypothesi examino t d san e protoplasts during extended periods toplasts (Figures 5G and 5H). In some aggregating cells, the of P30 expression for possible P30 movement within the cyto- P30::GFP fusion preferentially accumulated where the cells codin0 plasmP3 e g th ,sequenc terminuN fuses e th e wa do t s wer contacen i t (Figure 5G) aren ,a whican i h secondary plas- of green fluorescent protein (GFP; Chalfi t al.ee , 1994e th n )i modesmata have been reporte foro dt m (Monzer, 1991). Thus, expression vector pRTL2 to form P30::GFR Figure 5 shows the previous positioning of P30 along filaments may represent that, as expected, transient expression of P30::GFP produced an earlier stage in the migration of P30 to the exterior of the a filamentous network in protoplasts (Figures 5A and 5B) similar cell, where, in intact tissue, plasmodesmata may be found. to that detected with P30 antibodies. Again, as seen with im- We next compared the localization pattern of P30 expressed munocytochemically detected P30, the extent of the P30::GFP isteadna y state versu transiensa t expression system thir s.Fo network varied from cel cello t l . P30::GFP appeare filamens da - purpose, protoplasts were isolated from a cell suspension cul- tous arrays, as shown in Figure 5A, but condensed aggregates ture originating from leaves of a P30-transformed plant. In of P30::GFP interspersed alon filamente gth s were presenn i t leaves of intact transgenic plants, P30 is deposited in the cell some of the protoplasts (Figure 5B). In a few cells, large ag- wall, specifically in the secondary plasmodesmata (Deom et gregates could be detected along fine filaments (data not al., 1990; Ding et al., 1992). Enzymatic digestion of the P30 shown). The P30::GFP fusion always displayed more protein transgenic cell suspension culture for several hours did not aggregates than P30 alone. This result suggests that the ad- completely remov cele eth l walls. Consequently transgenie th , c dition of GFP to P30 interferes to some extent with filament "protoplasts" were oblong rather than rounded sees a , n with formation so that P30::GFP aggregates occurred more readily normal protoplasts. Immunocytochemistr thesof y e "pro- than with wild-type P30. Filaments were not present when con- toplasts" revealed that P30 was found predominately around trol pRTL2::GFP alone was expressed in protoplasts (Figures the perimeter near the cell walls, as shown in Figures 6A, 6B, 5D)d 5Can , indicating doe thaP product stGF no filamene eth - detectes cross-wale wa th 0 n dprotoplasi e P3 ; th f 6D o l d tan tous network. The filamentous appearance of P30::GFP was in Figure 6D, suggesting that P30 is deposited in the observed only terminu witfusionN 0 e hP3 th GFPso f t s o . When in these transgenic cell cultures filament0 P3 . s were also pres- P30 was fused to the C terminus of GFP, the intracellular loca- transgenie th en n i t c cells shows a , Figurn differeni a , e6C t typicalltios nwa y cytoplasmic (Figure 5F)d an , resemblinE s5 g focal celplane th l f froeo m that show. 6B Figuren d i an A s6 patteralonee P th GF f ,n o wit h fluorescence localizin cytoo gt - plasmic strands (Figure 5D)d an . ThereforeC s5 likels i t ,i y that C-terminae th l fusion inactivaty sma e P30, possibl alteriny yb g Vitrn I oCytoskeletao t Bindin 0 P3 f go l Proteins conformatio0 P3 maskiny b r n o responsiregio0 e gth P3 f no - formatioe th r filamentse fo th f e no bl . Treatment with both cold and cytochalasin D disrupted the P30::GFe Th P patter protoplastafter n i h 2 r7 transo t 8 s4 - filamentou network0 sP3 , indicating associatetha0 P3 t s with fection differed fro pattere me th Th . n hr see 0 2 n o aftet 8 1 r microtubul actid ean n filament . Thes vivn ei o P30::GFP networks obvious s wera t eno , suggesting that less studies clearly showed coalignment between microtubuled san fusioe ofouns th f nfilamentwa s da associates wa r so d with numerous P30 filaments, but association of actin filaments and B

Figure 6. Localization of P30 Near the Cell Periphery and in Filamentous Arrays in P30 Transgenic Cells.

P30 was detected by affinity-purified P30 polyclonal antibody and fluorescein-conjugated goat anti-rabbit secondary antibody. (A) to (C) Different focal planes of P30 in the same aldehyde-fixed P30-transgenic protoplast. (D) P30 in a different aldehyde-fixed P30-transgenic protoplast. (D)o t ) . (A ur0 r 1 nfo Ba = r Plane Th 210t Cel8 l

P30 filaments often was less pronounced. Usually, when -96 compared with that seen with microtubules aree , th colocal f ao - ization of a single P30 filament with an actin filament was quite •68 colocalize0 limitedP3 d an , d with actin filaments over short distances only. To test further the interaction of P30 with actin, in vitro binding assays were performed. Sinc accumu0 eP3 - actin •*• •45 lates only transientl TMV-infecteyn i d plantsmors i 0 e P3 ,readil y produce purified an dn i d from Escherichia co//fo vitrn i r- oex P30-» periments. However, £. co//-produced P30 is often found in -29 inclusion bodies and subsequently exhibits low solubility. The high-speed centrifugation use n typicai d l actin filament sedimentation assays cause alon0 sP3 sedimento et . Thus, two alternate experimental approaches were used. B 1234567 As shown in Figure 7A, the association of P30 and actin was investigated by sedimentation with in vitro-translated P30. In- 96- soluble 35S-P30 represented onl minoya r componene th f to 68- translation reaction removes s (Figurwa d lan, y dan b e) 7A e3 centrifugation. Then the soluble 35S-P30 was incubated with unlabeled polymerized actin. Figur lan, , showe7A 2 s that 35 —. —. 4- P30 S-P30 cosedimented with actin filaments, suggesting that it binds directl actinyo t negativ e .Th e control vitrn i oe trans,th - 29- lation product of T7 gene 10, did not cosediment with actin (dat t shown) aseconno e th n I . d method show Figurn i , e7B 12345678 P30-actin interactions were tested using biotinylated actin and 96- streptavidin-coupled magnetic beads. Biotinylated monomeric 68- — ••-BSA actin or polymerized actin (consisting of a mixture of biotinylated «*» *-tubulin and unbiotinylated actin) was incubated with unlabeled P30. 45 - Streptavidin-coupled magnetic beads were use collecdo t e tth 4-P30 biotinylated actin, followe incubatioy db samplS SD en i buffer 29- to cleave the biotin cross-linker and release actin and any act in- associated proteins. After removing the beads magnetically, the supernatant containing acti associated nan d proteins swa Figure 7. In Vitro Assays of P30 Binding to Actin and Tubulin. electrophoresed. Centrifugation was not used at any point in (A) Actin sedimentation assay wit vitro-translatehn i RNA0 e dP3 .Th this assay, thus avoiding potential cosedimentatio wit0 hP3 f no of a 10.5% SDS-polyacrylamide protein gel assaying actin due to low P30 solubility. This method also permitted in- translation products that cosediment with rabbit muscle actin is shown. teractions of P30 with both monomeric and polymerized actin detectepositions e (a Th 0 Coomassi y P3 actif d sb o d nan e blue stain- to be tested easily. The results presented in Figure 7B show ingindicatee ar ) t leftproteid a an , n molecular weight markere sar recoveres wa tha0 P3 t d from both monomeric (Figur lan, e7B e indicated at right in kilodaltons. Lane 1 contains the pellet from co- filamentoud 3)an s (Figur lan, actie) 7B e6 n samples. Control sedimentatio actif ntranslatioo d nan ns productwa A RN s0 wheP3 o nn addetranslatioe th o dt n system; lan containe2 e pellesth t from experiments with BSA indicated that BSA does not bind ei- cosedimentation of actin and soluble translation products when P30 ther monomeri filamentour co s actin (Figur , laned e7B an s4 RNaddetranslatios e th Awa o dt n system land containan e;3 e sth converse 7).Th e experiment using biotinylate unlad an -0 dP3 pellet from centrifugation of P30 translation products. This insoluble beled actin gav same eth e result (dat shown)t ano . These data materia removes i l d before incubation with polymerized actin. sugges binn ca dt tha0 directlP3 t boto yt h monomerid can Analysi) (B associatio0 actiP3 f so d nan n using magnetic beadse .Th filamentous actin. 10.5% SDS-polyacrylamide protein gel was stained with Coomassie blue to detect actin and proteins that associated with rabbit muscle actin positione Th . denotee f actinsar o A t rightd,a BS P30d d an , an , the protein molecular weight markers are at left in kilodaltons. Lane sie blu deteco et t porcine brain tubuli associated nan d proteinse Th . contain1 alone0 sP3 ; lan , monomerie2 c biotinylated actin wito hn positions of tubulin, P30, and BSA are denoted at right, and the pro- added P30; lan monomeri, e3 c biotinylated actin after incubation with tein molecular weight markers are at left in kilodaltons. Lane 1 contains P30; lan , monomerie4 c biotinylated actin after incubation with BSA; P30 alone; lan biotinylate, e2 d tubulin; lan unbiotinylate, e3 d tubulin; lane 5, polymerized biotinylated actin with no added P30; lane 6, poly- lan polymerize, e4 d biotinylated tubulin after incubation with P30; lane merized biotinylated actin after incubation with P30 d lan, ;an e7 5. unpolymerized biotinylated tubulin after incubation with P30; lane polymerized biotinylated actin after incubation with BSA. 6. polymerized biotinylated tubulin after incubation with BSA; lan, e7 (C) Analysis of tubulin and P30 association using magnetic beads. polymerized biotinylated tubulin with no added P30; and lane 8, BSA The 10.5% SDS-polyacrylamide protein gel was stained with Coomas- alone. Cytoskeletal Association of TMV P30 2109

From the in vivo data, the coalignment of P30 with microtu- transported through the cytoplasm. P30 along cytoskeletal bules is more widespread and obvious than that of P30 with filaments may be a precursor to localization of P30 to the plas- actin filaments; consequently, the in vitro interaction of P30 modesmata and may reflect the association of P30 with a motor with tubulin was not tested as extensively as that of P30 with protein that shuttles P30 filaments through the cell. That ex- actin. Nevertheless, in vitro experiments suggest that P30 tended expression of P30::GFP resulted in lower amounts of directly binds tubulin as well as actin. Essentially as described cytoplasmic filamentous P30 and increased amounts of P30 for the experiment testing the association of P30 with actin, aggregates along the outer edges of the cell suggests that the shown in Figure 78, P3O-tubulin interactions were assayed P30 filamentous localization is transient and may precede P30 using biotinylated tubulin and streptavidin-coupled magnetic movement to the plasma membrane and ultimately to the plas- beads. The results presented in Figure 7C show that P30 was modesmata. An additional observation supporting this idea recovered from both polymerized (Figure 7C, lane 4) and un- is that P30 in protoplasts isolated from transgenic plants is polymerized (Figure 7C, lane 5) tubulin samples. Control found predominately along the cell wall and plasma membrane experiments with BSA indicated that BSA does not bind tubu- and that some P30 is found in a filamentous pattern in the lin (Figure 7C, lane 6). These data suggest that P30 can bind cytoplasm of these protoplasts. directly to both polymerized and unpolymerized tubulin. Based on the observed colocalization of P30 with cytoskeletal components and the studies of membrane, organelle, and RNA transport in animals, a model for intracellular transport of P30 DlSCUSSlON can be proposed. In brief, the P30-vira1 RNA complex could use a “linked” system of microtubules and actin filaments for active transport. This system would be similar to that proposed The experiments we have presented show that the TMV move- for squid giant axon organelles in which microtubules are ment protein P30 forms a filamentous network in the plant thought to provide tracks for long-distance movement, whereas cytoplasm that colocalizes with the filamentous arrays of the actin filaments are thought to direct short-distance movement plant cytoskeleton. Since a P30 mutant, sb-6, did not form this to local sites (Atkinson et al., 1992; Kuznetsov et al., 1992; network, the filamentous appearance of wild-type P30 most Langford, 1995). Similarly, both microtubule motor and ac- likely results from a function of P30 rather than its overexpres- tin- systems appear to actively transport various sion. The P30 network observed in the present study coincides mRNAs, as RNA-protein complexes, in animal cells (reviewed with random cortical arrays of cytoskeletal filaments in plant in Wilhelm and Vale, 1993). Since viruses tend to exploit cel- protoplasts. As suggested from the pattern of P30 expression, lular mechanisms, the interaction of P30 with both microtubules P30 in both transfected and virus-infected protoplasts colocal- and actin filaments may mimic transport of the RNA-protein izes with cytoskeletal filaments, primarily with tubulin and to complexes and organelles by using microtubule motors and a lesser extent with actin. Microtubule-depolymerizing treat- actin-myosin systems for long- and short-distance transport, ment disrupted the P30 filamentous pattern, indicating that respectively (Langford, 1995). P30 interacts specifically with microtubules. Disruption of ac- As shown in Figure 8, in the model for P30 intracellular move- tin filaments by cytochalasin D treatment rarely affected the ment, P30 binds to viral RNA to form a P30-vira1 RNA complex P30 filamentous pattern, suggesting that the P30 interaction (step 1). We propose from previous evidence (Citovsky et al., with actin is not as extensive as that with microtubules. 1992a) that the P30-RNA complex is quite elongated and thin. Whether the in vivo interaction of P30 with the plant cyto- This structure would reduce diffusion within the cytoplasm skeleton occurs directly through P30 or through a cellular while favoring organized or directed movement along cytoskele- protein bound to actin and tubulin is not known. For example, tal elements. According to our model, the cytoskeleton provides P30 may interact with the cytoskeleton by binding both a a track for the long unfolded P30-RNA complexes and facili- microtubule-associated protein and an actin binding protein. tates linear, directed transport. P30 could be associated with Alternatively, P30 may associate indirectly with microtubules the cytoskeleton either before, during, or after RNA complex via a microtubule-associated protein and directly with actin or formation. The P30-RNA complex would first use microtubules vice versa. However, in vitro experiments suggest that P30 can for long-distance, possibly bidirectional, movement through bind directly to both actin and tubulin. Along these lines, the cytoplasm (step 2). Then it would associate with actin another plant viral protein, the 65-kD heat shock protein filaments for short-distance unidirectional movement to and 70-related protein of beet yellows closterovirus, has been found possibly through plasmodesmata because plasmodesmata ap- to bind microtubules in vitro (Karasev et al., 1992). Although pear to contain actin (step 3; White et al., 1994). In addition P30 appears to bind both actin and tubulin in vitro, the in vivo to the structural actin in plasmodesmata, actin filaments also cellular environment may favor the interaction of P30 with a can be seen extending from both sides of the plasmodes- cytoskeletal-associatedcomponent, such as a motor protein, mata into the adjoining cells (White et al., 1994). Thus, P30 rather than the direct interaction between P30 and tubulin or is proposed to interact with plasmodesmata-associated actin actin. filaments by using them to target and move through plasmodes- The colocalization of P30 filaments with cytoskeletal fila- mata into the cytoplasm of adjacent cells (step 4). Since TMV ments lends support to the hypothesis that P30 is actively P30 itself moves between cells (Waigmann and Zambryski, 2110 The Plant Cell

pGEM7Zf+ (Promega foro t ) m pGEMP30. Before cloning, Ndes wa l fille witn d i Klenoe hth w fragmen Escherichiaf to polymerase A coliDN . pGEMPSON was constructed using oligonucleotide-directed mutagen- esis to insert an Ncol at the start of the P30 coding sequence. pGEMPSON was digested completely with BamHI and digested par- tially with Nco releaso t l entire ope0 eth eP3 n reading framn a s ea 800-bp Ncol-BamHI fragment; this fragment was cloned into the respec- tive site f pRTLso 2 (Restrep t al.oe , 1990 foro t ) m pRTLPSO. pRTLP30::GFP contains the entire P30 coding sequence fused in- frame with the coding sequence of green fluorescent protein, GFP 3'::5 0 construc3P o (5%T P3 ' GF ' t pRTLP30::GFP BspHa , I sits ewa introduce oligonucleotide-directey db d mutagenesi terminae th t sa - tio pGEMP30Nn i n codo0 P3 f no . This change terminatio0 P3 e dth n codon to a methionine, forming pGEMP30NB. pGEMPSONB was di- •P30 - microtubules gested with BspH releaso t I 670-bea p fragment that containee dth P30 coding sequence from 43 to the newly added methio- - viral RNA nine at the end of the coding sequence. To fuse P30 to GFP, the 670-bp acti- n filaments BspHI fragmen clones wa t d intcompatible oth e Ncol sit f pKSIIeo - . PetersG GFPN/ d an , Universite gifCa ( f Drs o tRo . .J f Californiayo , Figure 8. Model for the Movement of the P30-RNA Complex through Berkeley, CA), giving pKS-P30::GFP. pKSII-GFPN/C contain muta-sa the Cytoplasm to and through Plasmodesmata. genized version of the GFP gene in pBluescript II KS+ (Stratagene); an Ncol site was engineered at the 5' of GFP, and the cauliflower mo- Step 1 shows that P30 forms a complex, either in the cytoplasm or saic virus (CaMV polyadenylatioS )35 n signal from pRTL addes 2wa d on the cytoskeleton, with viral RNA. Step 2 shows that the P30-RNA at the 3' of the GFP. The fusion between P30 and GFP was cloned complex moves long distances throug cytoplase hth microtubulesmn o , from pKS-P30::GFP as a 900-bp Clal-Pstl fragment into the Clal-Pstl possibly by interacting with a microtubule motor. Step 3 shows that sites in pRTLPSO, producing the plasmid pRTLP30::GFP. The sequence P30-RNe th A complex moves short distance plasmodesme th o st a of pRTLPSO: :GF confirmes Pwa dideoxy db y sequencing (Sequenase; on plasmodesma-associated actin filaments, possibly via a myosin mo- U.S. Biochemical) expression controP a s GF A . r lfo 850-be ,th p Ncol- tor. Step 4 shows that the P30-RNA complex then may move through Pstl fragment from pKSII.GFPN containin codinP GF ge sequencgth e plasmodesm adjaceno at samte cellth en s o plasmodesma-associate d plus the CaMV 35S polyadenylation signal was cloned into the respec- actin filaments. tive sites of pRTL2, producing pRTLGFP. The plasmids pRTLPSO, pRTLP30::GFP, and pRTLGFP were used for transient expressio f P30no , P30::GFP GFPd an , , respectivelyn i , shuttlee b y 1995)ma d 0 througP3 , h plasmodesmate th y ab protoplasts. actin filaments extending betweeno t cells 0 abilite P3 .Th f yo increase the plasmodesmal size exclusion limit (Wolf et al., 1989, 1991; Waigmann et al., 1994) may also be related to its interaction with actin, because actin was found in the neck Plant Cell Cultur Electroporatiod ean n region of plasmodesmata, where the size exclusion limit is though regulatee b o t t d (Whit t al.ee , 1994). Suspension cultures of Nicotians tabacum (line XD) were grown in a As suggeste modele th intracellulay e db ,th r associatiof no Murashige and Skoog-based medium, TXD (88 mM sucrose, 1 x wit0 hP3 both microtubule actid san n filament facilitaty sma e Murashige and Skoog salts [GIBCO], 1.3 mg/L niacin, 0.25 mg/L thia- cell-to-cell movemen P30-RNe th f o t A comple providiny xb g mine, 0.25 mg/L pyridoxine, 0.25 mg/L calcium pantothenate mg/4 , L p-chlorophenoxyacetic acid mg/,5 L kinetin mg/0 20 ,L inositol mg/0 13 ,L active transport functions. Further investigation0 P3 f o s asparagine, pH 5.8), and shaken at 200 rpm at 25°C. Protoplasts were interaction with the plant cytoskeleton will focus on whether prepared according to Howard et al. (1992). transport occurs by microtubule motors, actin-myosin, or both, For electroporation m 1 f ,electroporatioL o n Hepes M buffem 0 ,r(1 whethed an r cytoskeletal elements plagatine rolyth a en gi mannitolCaCINaCIM M M 7.2inoculates H m m 2 p 0. 0 2,5 , )wa 15 , d of plasmodesmata by P30. Studies of P30 intracellular trans- mixe addedx 10s iagwit 5 th 65 7 -wa ) hd protoplasts 2. o t an , 0 (5 A DN . modea s a t l systeac porn tca analyzm o t functioe eth planf no t min0 1 .o t Protoplast5 r fo turplace s e eic wa n sdo were electroporated cytoskeletal elements in intracellular and possibly intercellu 50-msea - r fo V c0 eitheelectroporatiopulsn 25 a t a n eri n apparatus lar transpor proteinf to othed san r macromolecular complexes. constructed to the specifications described in Fromm et al. (1985) or microfarad0 50 , V 0 aBio-Raa 15 t sn i d electroporator. After electropo- ration, protoplasts were incubated for 10 min on ice and then for 10 METHODS min at room temperature. Protoplasts were diluted to 2.5 x 105 pro- toplasts per mL in TXD medium containing 0.4 M mannitol (TXD-M) and incubated t 25°Covernigha r h . 0 Normallyr ~12 fo t8o t , pRTL2 Plasmid Constructs constructs electroporated into protoplasts have transformation efficien- assayes a % proteiy d90 b o t n0 cie5 expression f so efficience Th . y The P30 open reading frame was cloned from pETP30 (Citovsky et of transformation for pRTL2::P30, however, was typically 20 to 40%, al., 1990) as an Ndel-BamHI fragment into Smal-BamHI sites of suggesting that expressio deleterious i 0 celle P3 th f n.o o t s Cytoskeletal Association of TMV P30 21 11

Virus lnfection of Protoplasts in PBS). Following incubations with secondary antibodies, protoplasts again were washed with a large volume of PBS. To prevent fading of Protoplasts were prepared from N. tabacum (line XD) suspension cell the fluorescent signal, protoplasts were resuspended in Citifluor culture according to Howard et al. (1992) and inoculated with TMV using PBS-Citifluor glycerol (Ted Pella, Inc., Redding, CA; combined 1:2). polyethylene glycol as described by Maule et al. (1980). lnoculated For P30-actin studies, protoplasts were treated as described above protoplasts were cultured in TXD medium under continua1 light at 25OC except that P30 affinity-purified polyclonal rabbit antibody was used and assayed immunocytochemically for P30 expression after 10 hr. in primary antibody incubations and fluorescein-conjugated goat anti-rabbit IgG (Calbiochem; diluted 1:30 in PBS) and 0.33 to0.66 pM rhodamine phalloidin (Molecular Probes, Eugene, OR) were used in secondary antibody incubations. Transgenic Cell Suspension Cultures During some experiments, protoplasts were treated with either cold or cytochalasin D before fixation. For cold treatment, 5 x 105 pro- A transgenic tobacco line expressing P30, produced as described in toplasts were incubated at O°C for 0.5 to l hr and then aldehyde fixed. Citovsky et al. (1992b), was grown under sterile conditions. Leaves For cytochalasin D treatment, 5 x 105 protoplasts were incubated with were removed from the plants, cut from the midrib to the margin with 50 to 100 pM cytochalasin D for 0.5 to 1 hr at room temperature. Pro- toplasts were washed briefly with TXD-M and then aldehyde fixed. a scalpel every 3 to 4 mm, and incubated in 1% cellulase, 0.1% macero- For P30::GFP fusions, protoplasts were either observed directly with- zyme in K3+ (10 mM CaCI2, 0.45 M sucrose, 1 x Murashige and out fixation or briefly aldehyde fixed. For fixation, protoplasts were Skoog salts [GIBCO], 1.3 mglL niacin, 10.25 mglL thiamine, 1.25 mg/L incubated for 10 to 15 min in 3 to 4% formaldehyde (freshly prepared pyridoxine, 0.25 mglL calcium pantothenate, 0.1 mglL 2,4-dichlorophen- oxyacetic acid, 0.2 mg/L Bbenzylaminopurine,1 mg/L naphthaleneacetic from paraformaldehyde), 10 mM EGTA, 5 mM MgS04, and 10% acid, 250 mglL xylose, 100 mglL myoinositol, 1 mglL nicotinic acid, DMSO in 100 mM Pipes, pH 6.9. Protoplasts were rinsed for 10 min pH 5.8) overnight at 25% with gentle shaking. The undigested in 100 mM Pipes, pH 6.9, and then briefly rinsed with PBS. tissue and veins were removed and the remaining solution centrifuged Observations were performed on a Zeiss Axiophot epifluorescence at 100 to 1509 for 5 min to collect protoplasts. The protoplasts were microscope (Carl Zeiss, Inc., Thornwood, NY). lmages were captured washed twice with K3+ and then cultured in K3+ at 25OC without shak- with a cooled CCD (model TEAlCCD-1400TK; Princeton Instruments, ing. The high osmolarity concentration was gradually reduced by Inc., Trenton, NJ) controlled by IPLab (Signal Analytics, Vienna, VA). diluting K3+ with TXD every 7 days. After 7weeks, cultures were gently Digital images were processed and figures assembled using Adobe shaken to facilitate growth and then at 10 weeks were maintained at Photoshop (Adobe Systems, Mountain View, CA). 25OC and shaken at 200 rpm. After 12 to 15 weeks, the cell suspen- sion culture was used for protoplasts as described above.

P30:Actin Cosedimentation Assay lmmunocytochemistry and GFP Fusion Protein Expression As a template for in vitro transcription, pGEMP3O was linearized with The P30 protein was purified from E. coli as described by Citovsky BamHI. Messenger P30 RNA was transcribed from the template using et al. (1990) and used for production of polyclonal antibodies in rab- T7 DNA polymerase (New England Biolabs, Beverly, MA) and Fmethyl- bits. Specific P30 antibodies were purified from the rabbit antisera by guanosine cap (Pharmacia Biotechnology) as described in Promega affinity for the P30 protein immobilized on polyvinylidene difluoride Protocols and Application Guide. The P30 RNA was translated in a membrane (Immobilon P; Millipore, Bedford, MA; Smith and Fisher, wheat germ translation system (Promega); a typical50-pL translation 1984). reaction contained 25 pL wheat germ extract, 10 pCi 3%-methionine Before immunocytochemistry with P30 antibodies, protoplasts were (Amersham), and -1 to 2 pg RNA from the transcription reaction. The either aldehyde fixed or detergent permeabilized (Kengen and de Graaf, reaction was incubated at 25OC for 1 hr. Labeled translation products 1991; Kengen and Derksen, 1991). For fixation, protoplasts were in- were resolved by 12.5% SDS-PAGE and detected by fluorography. cubated for 0.5 to 1 hr in 3 to 4% formaldehyde (freshly prepared from Before cosedimentation assays, 1 pL of the translation reaction was paraformaldehyde), 10 mM EGlA, 5 mM MgS04, 1 mM phenylmethyl- diluted with 99 pL of F buffer (50 mM Hepes, pH 7.5, 0.2 mM CaCI2, sulfonyl fluoride, 2 pg/mL aprotinin, 2 pg/mL leupeptin, 1 pglmL 0.1 M KCI, 5 mM MgCI2, and 1 mM ATP) and then centrifuged for 10 pepstatin, and 10% DMSO in 100 mM Pipes, pH 6.9, and then rinsed min at 80,OOOg to sediment insoluble proteins. lmmediately following for 10 min in 100 mM Pipes, pH 6.9. For detergent permeabilization, centrifugation, aliquots of the translation reaction were used for actin protoplasts were incubated at room temperature for 30 min in 100 mM cosedimentation assays. Rabbit or chick muscle actin was polymer- Pipes, pH 6.9, 10% DMSO, 10 mM EGTA, 5 mM MgS04, 0.05% ized by adjusting buffer conditions from G buffer (50 mM Hepes, pH Nonidet P-40 (vlv), and 0.4 M mannitol. 7.5, 0.2 mM CaCI2, 0.2 mM ATP) to that of F buffer. Cosedimentation The following procedure was used with both aldehyde-fixed and assays then were performed by mixing 25 pL of in vitro translation detergent-permeabilized protoplasts. For P30-microtubulestudies, pro- products with 5 to 15 pg of polymerized actin. The mixture was in- toplasts were incubated for l hr at room temperature with P30 affinity- cubated at room temperature or on ice for 1 hr and then centrifuged purified polyclonal rabbit antibody and tubulin mouse monoclonal anti- for 10 min at 80,OOOg either directly or layered over a cushion of 10% body. Protoplasts were washed with a large volume of PBS (10 mM sucrose in F buffer. After removal of the supernatant, the pellet was phosphate, pH 7.4, 150 mM NaCI) to remove primary antibodies and washed with F buffer and solubilized with sample buffer. The pellet then incubated for 30 min at room temperature with fluorescein- and supernatant fractions were electrophoresed by 10.5% SDS-PAGE. conjugated goat anti-rabbit IgG (Calbiochem; diluted 1:30 in PBS) and Actin was detected by Coomassie Brilliant Blue R 250 staining, and rhodamine-conjugatedgoat anti-mouse IgG (Calbiochem; diluted 1:30 cosedimenting translation products were detected by fluorography. 21 12 The Plant Cell

Magnetic Assay for P30:Actin and P3O:Tubulin lnteractions supernatant was electrophoresed by 10.5% SDS-PAGE and analyzed by Coomassie blue staining. Rabbit muscle actin (3 to 4 mg) was labeled with biotin-HPDP as directed by the manufacturer (Pierce, Rockford, IL). Uncoupled biotin- HPDP was removed over a desalting (Bio-Gel P; Bio-Rad), ACKNOWLEDGMENTS and biotinylated actin was stored in G buffer. Linkage of biotin to actin was confirmed by dot blot. Presumptive biotinylated actin was absorbed to lmmobilon P in a dot blot filtering apparatus.The filter was removed, We thank Dan Oppenheimer and Dr. Ted Wong for rabbit and chick blocked for 1 hr at room temperature with 5% BSA-TBS (50 mM Tris, actin, the laboratory of Dr. Zac Cande for the tubulin antibody, Drs. 200 mM NaCI, pH 7.4), and incubated for 30 min at room temperature Judith Roe and Gary Peters for pKSII.GFPN/C, and Dr. Steve Ruzin with streptavidin-alkaline phosphataseconjugate diluted 1:5000 in 1% and Hans E.E. Holtan for their invaluable help with imaging and mi- BSA-TBS. After extensive washing, the blot was developed using nitro croscopy. We thank Dr. Vitaly Citovsky and Dr. Andy Jackson for critical blue tetrazolium and 5-bromo-4-chloro-3-indolylphosphate. As deter- reading of this manuscript and for encouragement. We gratefully ac- mined by sedimentation assays, biotinylation of actin did not appear knowledge members of the Zambryski laboratory for many helpful to affect polymerization significantly. discussions. This work was supported by National lnstitutes of Health For actin binding assays, biotinylatedactin was polymerized by add- Grant No. GM45244 awarded to P.C.Z. ing 1 pg native actin per 2 pg biotinylated actin and adjusting buffer conditions to that of F buffer. Either monomeric biotinylated or poly- merized biotinylated actin (1.5 pg) was incubated at room temperature for 20 to 30 min with 0.2 to 0.4 pg P30 purified from E. coli (Citovsky Received July 31, 1995; accepted October 4, 1995. et al., 1990) or, as a negative control, with 0.4 pg BSA. Streptavidin paramagnetic beads (5 to 7 pg; Promega; protein/antibody qualified) were incubated with the actin-P30 mixture at room temperature for REFERENCES 15 min. The beads were magnetically attracted to the side of the microcentrifuge tube, allowing the supernatant to be removed. The beads were washed two to three times by 5- to 10-min incubations Atkinson, S.J., Doberstein, S.K., and Pollard, T.D. (1992). Moving at room temperature with gentle shaking. The washes were removed off the beaten tracks. Curr. Biol. 2, 326-328. by again magnetically attracting the beads to the side of the micro- Bassell, G.J., Taneja, K.L., Kislauskis, E.H., Sundell, C.L., Posers, centrifuge tube. Washes were performed in either G buffer for the C.M., Ross, A., and Singer, R.H. (1994). Actin filaments and the monomeric actin-P30 mixture or F buffer for the polymerized actin-P30 spatial positioning of mRNAs. In Actin: Biophysics, Biochemistry mixture. The biotinylated actin and any actin-associated protein were and Cell Biology, J.E. Estes and P.J. Higgins, eds (New York: Ple- released from the biotin by incubating the beads in SDS sample buffer num Press), pp. 183-189. 5 to 10 min at 55OC to cleave the biotin cross-linker. The beads again Blum, H., Gross, H.J., and Beier, H. (1989). The expression of the were magnetically attached to the side of the microcentrifuge tube, TMV-specific 30-kDa protein in tobacco protoplasts is strongly and allowing removal of the supernatant containing the actin and associated selectively enhanced by actinomycin. Virology 169, 51-61. protein in sample buffer. The supernatant was electrophoresed by 10S0/o Chaltie, M., Tu, Y., Euskirchen, G., Ward, W.W., and Prasher, D.C. SDS-PAGE and analyzed by Coomassie blue staining. (1994). Green fluorescent protein as a marker for gene expression. For tubulin binding assays, biotin-porcine brain tubulin (Cytoskele- Science 263, 802-805. ton, Denver, CO) was polymerized at a concentration of 1 to 3 pg/pL in G-PEM-1 (80 mM Pipes, pH 6.8, 1 mM MgS04, 1 mM EGTA, 1 mM Citmky, V. (1993). Probing plasmodesmal transport with plant viruses. GTP) plus 10 to 15% glycerol (v/v) for 30 min at 35OC. Either unpolymer- Plant Physiol. 102, 1071-1076. ized or polymerized biotinylated tubulin (3 pg) was incubated at room Citovsky, V., and Zambryski, R (1991). How do nucleic acids temperature for 20 to 30 min with 0.2 to 0.4 pg P30 purified from E. move through intercellular connections? BioEssays 13, 373-379. coli (Citovsky et al., 1990) or, as a negative control, with 0.4 pg BSA. Citovsky, V., and Zambryski, P. (1993). Transport of nucleic acids Streptavidin paramagnetic beads (5 to 7 pg; Promega; proteinlanti- through membrane channels: Snaking through small holes. Annu. body qualified) were incubated with the tubulin-P30 mixture at room Rev. Microbiol. 47, 167-197. temperature for 15 min. The beads were magnetically attracted to the Citovsky, V., Knorr, D., Schuster, G., and Zambryski, P. (1990). The side of the microcentrifuge tube, allowing the supernatant to be re- P30 movement protein of tobacco mosaic virus is a single strand moved. The beads were washed two to three times by 5- to 10-min nucleic acid binding protein. Cell 60, 637-647. incubations at room temperature with gentle shaking. The washes were Citwky, V., Wong, M.L., Shaw, A.L., Prasad, B.V.V., and Zambryski, removed by again magnetically attracting the beads to the side of the P. (1992a). Visualization and characterization of tobacco mosaic vi- microcentrifuge tube. Washes were done in either G-PEM-0.1 (80 mM rus movement protein binding to single-stranded nucleic acids. Plant Pipes, pH 6.8, 1 mM MgS04, 1 mM EGTA, 0.1 mM GTP) for the un- Cell 4, 397-411. polymerized tubulin-P30 mixture or G-PEM-1 plus glycerol for the polymerized tubulin-P30 mixture. Unlike the linkage between actin Citovsky, V., Zupan, J., Warnick, D., and Zambryski, P. (1992b). Nu- and biotin, the linkage between tubulin and biotin was not cleavable. clear localization of Agrobacterium VirE2 protein in plant cells. Thus, the biotinylated tubulin and any tubulin-associated protein were Science 256, 1803-1805. released from the beads by boiling the beads in SDS sample buffer Citovsky, V., McLean, B.G.,Zupan, J.R., and Zambryski, P. (1993). for 10 min. The beads again were magnetically attached to the side Phosphorylation of tobacco mosaic virus cell-to-cell movement pro- of the microcentrifuge tube, allowing removal of the supernatant Con- tein by a developmentally regulated plant cell wall-associated protein taining the tubulin and associated protein in sample buffer. The kinase. Genes Dev. 7, 904-910. Cytoskeletal Association of TMV P30 21 13

Cooper, J.A. (1987). Effects of cytochalasin and phalloidin on actin. Lehto, K., Bubrick, P., and Dawson, W.O. (1990). Time come of TMV J. Cell Biol. 105, 1473-1478. 30K protein accumulation in intact leaves. Virology 174, 290-293. Deom, C.M., Shaw, M.J., and Beachy, R.N. (1987). The30-kilodalton Lloyd, C.W. (1982). The Cytoskeleton in Plant Growth and Develop- gene product of tobacco mosaic virus potentiates virus movement. ment. (New York: Academic Press). Science 327, 389-394. Luby-Phelps, K. (1993). Effect of cytoarchitecture on the transport and Deom, C.M., Schubert, K.R., Wolf, S., Holt, C.A., Lucas, W.J., and localization of protein synthetic machinery. J. Cell. Biochem. 52, Beachy, R.N. (1990). Molecular characterization and biologicalfunc- 140-147. tion of the movement protein of tobacco mosaic virus in transgenic Luby-Phelps, K. (1994). Physical properties of cytoplasm. Curr. Opin. plants. Proc. Natl. Acad. Sci. USA 87, 3284-3288. Cell Biol. 6, 3-9. Deom, C.M., Lapidot, M., and Beachy, R.N. (1992). Plant virus move- Lucas, W.J., and Gilbertson, R.L. (1994). Plasmodesmata in rela- ment proteins. Cell 69, 221-224. tion to viral movement within leaf tissues. Annu. Rev. Phytopathol. Ding, B., Haudsenshleld, J.S., Hull, R.J., Wolf, S.,Beachy, R.N., 32, 387-411. and Lucas, W.J. (1992). Secondary plasmodesmata are specific Maule, A.J., Boulton, M.I., Edmunds, C., and , K.R. (1980). sites of localization of the tobacco mosaic virus movement protein Polyethylene glycol-mediated infection of cucumber protoplasts by in transgenic tobacco plants. Plant Cell 4, 915-928. cucumber mosaic virus and virus RNA. J. Gen. Virol. 47, 199-203. Dingwall, C. (1992). Soluble factors and solid phases. Curr. Biol. 2, McLean, B.G., Waigmann, E., Citovsky, V., and Zambryski, P. (1993). 503-505. Cell-to-cellmovement of plant viruses. Trends Microbiol. 1, 105-109. Fromm, M., Taylor, L.P., and Walbot, V. (1985). Expression of genes Melki, R., Gaudin, Y., and Blondel, D. (1994). lnteraction between transferred into monocot and dicot plant cells by electroporation. tubulin and the viral matrix protein of vesicular stomatitis virus: Pos- Proc. Natl. Acad. Sci. USA 82, 5822-5828. sible implications in the viral cytopathic effect. Virology 202,339-347. Gibbs, A.J. (1976). Viruses and plasmodesmata. In lntercellular Com- Meshi, T., Watanabe, Y., Saito, T., Sugimoto, A,, Maeda, T., and munication in Plants: Studies on Plasmodesmata. B.E.S. Gunning Okada, Y. (1987). Function of the 30 kD protein of tobacco mosaic and A.W. Robards, eds (Berlin: Springer-Verlag), pp. 149-164. virus: lnvolvement in cell-to-cell movement and dispensability for Haseloff, J., Goelet, P., Zimmern, D., Ahlquist, P., Dasgupta, R., replication. EMBO J. 6, 2557-2563. and Kaesberg, P. (1984). Striking similarities in amino acid sequence Meshi, T., Hosokawa, D., Kawagishi, M., Watanabe, Y., and Okada, among nonstructural proteins encoded by RNA viruses that have Y. (1992). Reinvestigationof intracellular localization of the 30K protein dissimilar genomic organization. Proc. Natl. Acad. Sci. USA 81, in tobacco protoplasts infected with tobacco mosaic virus RNA. Virol- 4358-4362. Ogy 187, 809-813. Hesketh, J. (1994). Translation and the cytoskeleton: A mechanism Monzer, J. (1991). Ultrastructure of secondary plasmodesmataforma- for targeted protein synthesis. MOI. Biol. Rep. 19, 233-243. tion in regenerating Solanum nigrum-protoplast cultures. Protoplasma Howard, E.A., Zupan, J.R., Citovsky, V., and Zambryski, P. (1992). 165, 86-95. The VirD2 protein of A. tumefaciens contains a C-terminal bipartite Morozov, Dolja, V.V., and Atabekov, J.G. (1989). Probable reas- nuclear localization signal: lmplications for nuclear uptake of DNA S., sortment of genomic elements among elongated RNA-containing in plant cells. Cell 68, 109-118. plant viruses. J. MOI. Evol. 29, 52-62. Karasev, A.V., Kashlna, A.S., Gelfand, V.I., and Dolja, V.V. (1992). HSP70-related 65 kDa protein of beet yellows closterovirus is a Moser, O., Gagey, M.J., Godefroy-Colburn, T., Stussi-Garaud, C., microtubule-binding protein. FEBS Lett. 304, 12-14. Ellwart-Tscurtz, M., Nitschko, H., and Mundry, K.-W. (1988). The fate of the transport protein of tobacco mosaic virus in systemic and Kengen, H.M.P., and de Graaf, B.H. (1991). Microtubules and actin hypersensitive tobacco hosts. J. Gen. Virol. 69, 1367-1373. filaments co-localize extensively in non-fixed cells of tobacco. Pro- toplasma 163, 62-65. Pasick, J.M., Kalicharran, K., and Dales, S. (1994). Distribution and Kengen, H.M.P., and Derksen, J. (1991). Organization of microtubules trafficking of JHM coronavirus structural proteins and virions in pri- and in protoplasts from suspension cells of Nicotiana mary neurons and the OBL-21 neuronal cell line. J. Virol. 68, plumbaginifolia: A quantitative analysis. Acta Bot. Neerl. 40,29-40. 2915-2928. Klberstis, P.A., Pessi, A., Atherton, E., Joacson, R., Hunter, T., Restrepo, M.A., Freed, D.D., and Carrington, J.C. (1990). Nuclear and Zimmern, D. (1983). Analysis of in vitro and in vivo products transport of plant potyviral proteins. Plant Cell 2, 987-998. of the TMV 30 kDa open reading frame using antisera raised against Shibaoka, H., and Nagai, R. (1994). The plant cytoskeleton.Curr. Opin. a synthetic peptide. FEBS Lett. 164, 355-360. Cell Biol. 6, 10-15. Koonin, E.V., and Dolia, V.V. (1993). Evolution and taxonomyof positive- Singer, R. (1992). The cytoskeleton and mRNA localization.Curr. Opin. strand RNA viruses: lmplications of comparative analysis of amino Cell Biol. 4, 15-19. acid sequences. Crit. Rev. Biochem. MOI. Biol. 28, 375-430. Erra- tum. Crit. Rev. Biochem. MOI. Biol. 28, 546. Smith, D.E., and Fisher, P.A. (1984). Identification, developmental regulation, and response to heat shock of two antigenically related Kuznetsov, S.A., Langford, G.M., and Weiss, D.G. (1992). Actin- fOrmS of a major nuclear envelope protein in Dfosophila : dependem organelle movement in squid axoplasm. Nature 356, Application of an improved method for affinity purification of anti- 722-725. bodies using polypeptides immobilized on nitrocellulose blots. J. Langford, G.M. (1995). Actin- and microtubule-dependem organelle Cell Biol. 99, 20-28. motors: lnterrelationships between the two motility systems. Curr. Staiger, C.J., and Lloyd, C.W. (1991). The plant cytoskeleton. Curr. Opin. Cell Biol. 7, 82-88. Opin. Cell Biol. 3, 33-42. 2114 The Plant Cell

Topp, K.S., Meade, L.B., and LaVail, J.H. (1994). Microtubule polar- White, R.G., Badelt, K., Overall, R.L., and Vesk, M. (1994). Actin ity in the peripheral processes of trigeminal ganglion cells: Relevance associated with plasmodesmata. Protoplasma 180, 169-184. for the retrograde transport of herpes simplex virus. J. Neurosci. 14, 318-325. Wilhelm, J.E., and Vale, R.D. (1993). RNA on the move: The mRNA localization pathway. J. Cell Biol. 123, 269-274. Vale, R.D. (1987). lntracellular transport using microtubule-based mo- tors. Annu. Rev. Cell Biol. 3, 347-378. Williamson, R. (1986). Organelle movements along actin filaments Waigmann, E., and Zambryski, P. (1995). Tobacco Mosaic Virus move- and microtubules. Plant Physiol. 82, 631-634. ment protein-mediated protein transport between trichome cells. Plant Cell 7, 2069-2079. Wolf, S., Deom, C.M., Beachy, R.N., and Lucas, W.J. (1989). Move- ment protein of tobacco mosaic virus modifies plasmodesmatalsize Waigmann, E., Lucas, W.J., Citovsky, V., and Zambryski, P. (1994). exclusion limit. Science 246, 377-379. Direct functional assay for tobacco mosaic virus cell-to-cell move- ment protein and identification of a domain involved in increasing Wolf, S., Deom, C.M., Beachy, R., and Lucas, W.J. (1991). Plasmodes- plasmodesmal permeability. Proc. Natl. Acad. Sci. USA 91, matal function is probed using transgenic tobacco plants that express 1433-1437. a virus movement protein. Plant Cell 3, 593-604. Watanabe, Y., Emori, Y., Ooshika, I., Meshi, T., Ohno, T., and Okada, Y. (1984). Synthesis of TMV-specific and proteins at the early Zimmern, D. (1988). Evolution of RNA viruses. In RNA Genetics, J. stage of infection in tobacco protoplasts: Transient expression of the Holland, E. Domingo, and P. Ahlquist, eds (Boca Raton, FL: CRC 30K protein and its mRNA. Virology 133, 18-24. Press), pp. 211-240.