Effect of mutations K97A and E128A on RNA binding and self assembly of papaya mosaic coat protein Marie-He´ le` ne Tremblay1, Nathalie Majeau1, Marie-Eve Laliberte´ Gagne´ 1, Katia Lecours2, He´ le` ne Morin1, Jean-Baptiste Duvignaud1, Marile` ne Bolduc1, Nicolas Chouinard1, Christine Pare´F1, Ste´ phane Gagne´ 2 and Denis Leclerc1

1 Centre de Recherche en Infectiologie, Univeriste´ Laval, Que´ bec, Canada 2De´ partement de Biochimie, Universite´ Laval, Que´ bec, Canada O O

Keywords Papaya mosaic potexvirus (PapMV) coat proteinR (CP) was expressed assembly; coat protein; nucleocapsid; (CPDN5) in Escherichia coli and showed to self assemble into nucleocapsid papaya mosaic ; potexvirus like particles (NLPs). Twenty per cent of the purified protein was found as NLPs of 50 nm in length and 80% was foundP as a multimer of 450 kDa (Received 20 July 2005, revised 29 Septem- ber 2005, accepted 25 October 2005) (20 subunits) arranged in a disk. Two mutants in the RNA binding domain of the PapMV CP, K97A and E128A showed interesting properties. The doi:10.1111/j.1742-4658.2005.05033.x proteins of both mutants could be easily purified and CD spectra of these proteins showed secondary and tertiaryD structures similar to the WT pro- tein. The mutant K97A was unable to self assemble and bind RNA. On the contrary, the mutant E128A showed an improved affinity for RNA and self assembled more efficientlyE in NLPs. E128A NLPs were longer (150 nm) than the recombinant CPDN5 and 100% percent of the protein was found as NLPs in bacteria. E128A NLPs were more resistant to diges- tion by trypsin than theT CPDN5 but were more sensitive to denaturation by heat. We discuss the possible role of K97 and E128 in the assembly of PapMV. C

Papaya mosaic potexvirus (PapMV) is a memberE of particles that are very similar to the WT virus [3]. The the potexvirus family. The virion is a flexible rod in vitro assembly of PapMV was shown to be specific 500 nm in length and 15 nm in diameter composed of and triggered by 47 nucleotides of the 5¢ noncoding 1400 subunits of viral coat protein (CP) [1] assembledR region of the virus [6]. This region is free of any dis- around a plus-strand genomic RNA of 6656 nucleo- cernable secondary structure, an important feature for tides (nt) [2]. In vitro reconstitution of PapMV nucleo- initiation of the assembly process [6]. In vitro assembly capsid-like particles (NLPs) was previouslyR studied is specific when performed at a pH of 8.0–8.5 in a buf- using CP prepared from the acetic acid degradation of fer of low ionic strength [4] and elongation proceeds in purified virus [3]. Using this method, PapMV CP can the 5¢fi3¢ direction [7]. One atom of Ca2+ is attached be isolated from the genomicO RNA and used for to each subunit, which is probably important for the in vitro assembly assays [3]. Extracted CP has been structure of the protein [8]. Alignment of PapMV CP found as a variety of aggregates ranging from 14S to with the CP of other reveals that these 25S [4] that include a disk-likeC structure (14 S) made proteins share 35% identity [9]. The N terminus of the of 18–20 subunits. These disks are helical structures proteins is the most divergent and their length is also (two turns of the helix) similar in architecture to the variable. It can reach more than 50 amino acids in the native virus particle [3,5].N The addition of RNA to the case of potato aucuba mosaic virus. Phosphorylation isolated disks triggers the assembly of long rod-shaped of the N terminus of potato virus X CP by host

Abbreviations CP, coat protein; NLP,U nucleocapsid like particles; nt, nucleotide; PapMV, Papaya mosaic potexvirus.

FEBS Journal (2005) ª 2005 The Authors Journal compilation ª 2005 FEBS 1 FEBS 5033 Dispatch: 25.11.05 Journal: FEBS CE: Blackwell Journal Name Manuscript No. B Author Received: No. of pages: 12 PE: Agila RNA binding and assembly of PapMV CP M.-H. Tremblay kinases was shown to render the genomic RNA more Results translatable [10]. It is possible that the incorporation of negative charges by phosphorylation of the N termi- Alignment of the potexvirus coat proteins nus destabilizes the subunits and favours disassembly of the particles, an important step in the initiation of The alignment of the amino acids sequences of 19 infection. potexviruses CPs between amino acid 91–169 of Pap-F Until now, most of the data collected on assembly MV CP revealed a consensus sequence (Fig. 1A). This of the potexvirus family have been obtained from Pap- conserved region shows the charged residues R104, MV in vitro assembly using partially denatured and K133, K137, and R161 that are believedO to be renatured proteins extracted by the acetic acid method involved in interaction with the genomic RNA and [3]. Even though in vitro assembly using this method play an important role in assembly and packaging of has been studied extensively, the nature of the interac- the viral genome [13]. It also containsO acidic residues tions between CP subunits, and between the subunits that could be important for the interaction with the and the genomic RNA remains unknown. genomic RNA, as shown for tobacco mosaic virus In this report, we showed that NLPs produced in CP [14]. Based on these observations,R we hypothes- Escherichia coli were very similar to the WT particles. ized that charged residues in the region 91–169 are Furthermore, extensive mutation analysis in the puta- good candidates for a specific interaction with the tive RNA binding domain of PapMV led to the iden- genomic RNA and excellentP targets for site specific tification of two mutants: one that had lost its mutagenesis. By introduction of mutations in the affinity for RNA, and one that showed an improved putative RNA binding domain, we hope to confirm affinity for RNA. The biochemical properties of the for the first time in the potexvirus family that this mutants were compared with the WT recombinant region is important for binding the genomic RNA proteins. and formation ofD NLPs. E A T C B E R R Fig. 1. (A) Alignment of a consensus sequence derived from 18 known potexvirus coat protein and the papaya mosaic virus coat protein (PapMV CP) in the conserved region 90–169 of PapMV. The amino acid corresponding to position 128 of the PapMV CP is an A in most potexviruses (underlined), an E in three differentO potexviruses including PapMV (bold), and T, V, S or Q occur in this position in the other five potexvirus sequences. The consensus was made using the CP sequences of: BaMV: Bamboo mosaic virus; PlAMV: Plantago asiatica mosaic virus; TVX: Tulip virus X; CsCMV: Cassava common mosaic virus; ClYMV: Clover yellow mosaic virus; HVX: Hosta virus X; LVX: Lily virus X; CymMV: Cymbidium mosaic virus; PAMV: Potato aucuba mosaic virus; NMV: ; PepMV: Pepino mosaic virus; FoMV: Foxtail mosaic virus; ScaVX: ScallionC virus X; WClMV: White clover mosaic virus; SMYEV: Strawberry mild yellow edge virus; PVX: Potato virus X; AltMV: Alternanthera mosaic virus; CVX: Cactus virus X; PapMV: Papaya mosaic virus. (B) Expression and purification of recombinant CPDN5, K97A, R104K105R108 ⁄ A, E128A and E148A. Lanes 1–7: Coomassie staining profile of the recombinant proteins. lane 1: E. coli lysate before induction ofN CPDN5, lane 2: E. coli lysate of CPDN5 16 h postinduction with 1 mM IPTG, lane 3: recombinant CPDN5 purified using a Ni2+ column; lane 4: purified K97A, lane 5: purified R104K105R108 ⁄ A lane 6: purified E128A lane 7: purified E148A, lane 8: PapMV CP from virus purified from plants. Lanes 9–14: Western blotting of purified recombinant proteins revealed with IgG directed against PapMV CP. Lane 9: Purified CPDN5, lane 10: purified K97A, lane 11: R104K105R108 ⁄ A, lane 12; E128A, lane13; E148A and lane 14: purified virus from infected plants.U

2 FEBS Journal (2005) ª 2005 The Authors Journal compilation ª 2005 FEBS M.-H. Tremblay RNA binding and assembly of PapMV CP

that the purified protein was indeed PapMV CP Expression of PapMV CP and its mutated forms (Fig. 1B, lane 9). Following the same procedure, we in E. coli have purified several mutated forms of the PapMV CP. We amplified the PapMV CP gene by RT ⁄ PCR and We have generated nine different mutants that har- cloned the resulting PCR fragment in the E. coli bour one, two or three A substitutions (Fig. 1A). Five expression vector pET-3d. The PapMV CP harbours mutants R118-D120-K121 ⁄ A, K133-K137 ⁄ A, D142-F two M residues at positions 1 and 6 of the CP ORF. It D145 ⁄ A, R161A and E166-E167-R168 ⁄ A produced is not clear if both of these initiation codons are used unstable proteins and were undetectable or expressed during replication of the virus. However, it has been at very low levels. It is likely that mutagenesisO in this shown that a large proportion of the CP of the purified conserved region affected the native folding of the CP. virus lacks several amino acids at the N terminus [1]. The mutants K97A, R104K105R108 ⁄ A, E128A and To ensure production of only one ORF in E. coli,we E148A could be expressed to level similarO to CPDN5 removed the N-terminal five amino acids. M6 served as and easily purified using a 6·H tag as shown with the an initiation codon in our expression system. The intro- CPDN5 (Fig. 1B; lane 4, 5, 6 and 7). However, the duction of the initiation codon in the NcoI site intro- removal of imidazole duringR the dialysis made duced an extra A that is found in all the constructs the mutants R104K105R108 ⁄ A and E148A aggregate made in this study. A 6·H tag was added at the C ter- and precipitate. It is likely that the mutations affected minus of the protein to ease the purification process. the folding of the protein.P Therefore, we concentrated The recombinant protein CPDN5 was expressed in our study on the CPDN5, K97A and E128A forms. E. coli BL21 (pLysS) and showed a slightly larger molecular mass than that of WT CP extracted from Self assembly of the purified recombinant purified virus (Fig. 1B, lane 3 and 8). The difference proteins observed between the two proteins is probably caused D by the 6·H tag fusion at the C terminus. The recom- The CPs of several flexuous rod-shaped virus have been binant protein was affinity purified using a Ni2+ col- shown toE self-assemble into NLPs in E. coli in the umn and eluted using 1 m imidazole (Fig. 1B, lane 3). absence of any other viral components [15–17]. The con- We estimated the yield of the purified recombinant struct CPDN5 also followed this rule and self assembled protein at 40–50 mgÆL)1. Western blot assay using an into NLPsT in E. coli as shown by the electron micro- antibody raised against the WT PapMP CP confirmed graph of the purified recombinant protein (Fig. 2B). C E R R O C N

Fig. 2. Electron micrographs of (A) WT PapMV virus. (B) CPDN5 NLPs (C) E128A NLPs (D) K97A purified proteins (E) the immunogold label- ling of recombinant CPDN5 NLPs with antibodies raised against the 6·H TAG. Bars are 200 nm (F) Average length of PapMV virus and the recombinant NLPsU CPDN5 and E128A (n 150). ¼

FEBS Journal (2005) ª 2005 The Authors Journal compilation ª 2005 FEBS 3 RNA binding and assembly of PapMV CP M.-H. Tremblay

The NLPs were similar in shape and in diameter to the the CD signal measured with E128A NLPs at 208 nm native virus particles (compare Fig. 2A and B). To ana- superposed with the WT virus (Fig. 3A). lyse and quantify the proportion of the purified protein The CD signal of the isolated disks (high speed that was found as NLPs, we separated NLPs and smal- supernatant of the purified protein) of CPDN5 was ler aggregates by ultracentrifugation at 100 000 g for identical to that of isolated disks from the purified 2 h. Most of the CPDN5 proteins (80%) were found in virus using the acetic acid method (Fig. 3B). It is inter-F the supernatant. NLPs were found in the pellet and esting to notice that the CD signal measured with account for 20% of the total purified recombinant pro- disks in general at 208 nm was less pronounced than tein. However, the purified proteins K97A remained in with NLPs. This result suggests that theO content in the supernatant after ultracentrifugation. On the con- a-helices is increased when the disks assemble in NLPs. trary, the recombinant proteins E128A were found Finally, we compared the folding of the purified pro- totally in the pellet after ultracentrifugation. tein of K97A with high speed supernatantO of CPDN5. The electron microscope revealed that E128A NLPs The two proteins showed identical CD profiles isolated from the high speed pellet are similar to the (Fig. 3C). WT virus (Fig. 2C). We measured the length of 150 We also made spectra betweenR 250 and 350 nm to NLPs for each, CPDN5 and E128A, respectively, and measure the absorption of aromatic residues and try- presented the average length of the NLPs (Fig. 2F). ptophans in the protein. A change in the environment CPDN5 NLPs appeared to be on-tenth the length of those residues affectsP the signal recorded and pro- (50 nm) of the native virus (500 nm). However, E128A vides information on the variation of tertiary structure. NLPs are three times longer than CPDN5 NLPs The NLPs of CPDN5 and E128A appeared to be sim- (Fig. 2F) suggesting that this mutant can more effi- ilar (Fig. 3D); we concluded that the tertiary structures ciently support the initiation and elongation of assem- are the same for the two NLPs. The spectra of CPDN5 bly. Finally, an electron micrograph of the purified disks and K97AD were very also similar, again suggest- K97A protein revealed disorganized aggregates of ing similarity in tertiary structure (Fig. 3E). The slight 15–50 nm in diameter (Fig. 2D). The outline of the differencesE in the intensity of the curves between the aggregates are irregular showing that this protein can samples were probably due to differences in protein not organized itself into NLPs. concentrations. Interestingly, the shape of the curves is differentT between the NLPs and the disks as already reported for the WT virus and the disks extracted by Immunogold labelling of CPDN5 NLPs the acetic acid extraction method [19]. The purification of the NLPs using the Ni2+ column C was efficient suggesting that the 6·H tag is located at Gel filtration of CPDN5 and K97A disks the surface and available for interaction with the affin- ity column. To confirm this hypothesis, we performedE We showed that 80% of the purified recombinant an immunogold labelling experiment on CPDN5 NLPs CPDN5 and all the K97A proteins were found as mul- using anti-6·H tag rabbit antiserum followed by a sec- timers (disks) in the supernatant after ultracentrifuga- ondary donkey antirabbit antibody labelled withR gold tion. To measure the level of multimerization of these particles. As expected, the NLPs were decorated with proteins, we submitted the high speed supernatant of the gold particles, supporting our hypothesis (Fig. 2E). CPDN5 and the purified proteins K97A to a Super- R dexTM 200. For CPDN5, we showed that most of the proteins were eluted as a high molecular weight com- CD spectrophotometry of the recombinant plex of 450 kDa which correspond to a multimer of proteins O approximately 20 subunits if we simply divide this We used CD spectrophotometry to compare the secon- number by the molecular mass of the protein subunit dary structure of the recombinant proteins with that of (23 kDa) (Fig. 4A). The second peak eluting at the WT virus. The secondary structure of CPDN5 was 81.27 mL was collected and loaded on a SuperdexTM 75 estimated to be 49% a-helixC and 15% random coil; the column to improve the resolution. The protein eluted as high a-helical content resembles that of tobacco a 39 kDa molecule (data not shown). A sample from mosaic virus, another rod-shaped virus [18]. The CD this peak separated on SDS ⁄ PAGE and showed a spectra of CPDN5N NLPs and WT virus showed a unique band smaller than CPDN5 that corresponds to a slightly different profile (Fig. 3A). The CD signal degradation product of CPDN5. It is possible that this measured at 208 nm was more pronounced for the WT degraded protein is unable to form a high molecular virus than theU CPDN5 NLPs (Fig. 3A). Interestingly, complex and remains (possibly as a dimer) in solution.

4 FEBS Journal (2005) ª 2005 The Authors Journal compilation ª 2005 FEBS M.-H. Tremblay RNA binding and assembly of PapMV CP

F O O R P D E T C E Fig. 3. CD spectra of CPDN5, E128A, K97A and WT proteins. (A) Far-UV spectra of WT virus (black line), CPA¨ N5 (dotted line) and E128A NLPs (grey line). (B) Far-UV spectra of isolated disks from the WT virus by the acetic acid method (black line) and high-speed supernatant of CPDN5 recombinant proteins (disks) (dotted line). (C) Far-UVR spectra of the CPDN5 disks (dotted line) and K97A purified proteins (grey line). (D) Far UV spectra of the CPDN5 and E128A NLPs between 250 and 350 nm. (E) Far UV spectra of the CPDN5 and K97A disks between 250 and 350 nm. R The elution profile of K97A can be divided into other protein subunits but is unable to assemble into three major peaks (Fig. 4B). The second peak eluted at NLPs in E. coli. 50 mL and overlapped withO the CPDN5 disks (Fig. 4A). The first peak eluted between 41 and 43 mL In vitro assembly of CPDN5 and corresponds to aggregated material that is > 700 kDa in size. The elution pattern is wide and We isolated CPDN5 disks from the high-speed superna- shows a shoulder that suggestsC that this material is not tant of the purified proteins by affinity chromatogra- uniform. Maybe it corresponds to an aggregate of phy. This extract was used for an in vitro assembly K97A disks agglutinated together by nonspecific inter- assay. Disks of 17 nm diameter were isolated as previ- actions. Finally, theN third peak, which was not ana- ously shown by gel filtration (Fig. 5A). We incubated lysed further, probably corresponds to a truncated 50 mL of CPDN5 disks at a concentration of protein as shown for the CPDN5 construct. These 1mgÆmL)1 with 0.05 mg RNA for 30 min at room tem- results confirmU that K97A is able to form disks with perature. We showed by electron microscopy that the

FEBS Journal (2005) ª 2005 The Authors Journal compilation ª 2005 FEBS 5 RNA binding and assembly of PapMV CP M.-H. Tremblay

A RNA in NLPs under those experimental conditions (data not shown).

Measurement of the RNA content by spectroscopy F The K97A and E128A mutations showed completely opposite effects on the PapMV CP. To evaluate if CPDN5 and E128A NLPs contain RNA,O the ratio 280 ⁄ 260 of the different NLPs was measured on the spectrophotometer and compared with the proteins of the purified virus (Table 2). As expected,O the NLPs B showed a smaller 280 ⁄ 260 ratio than disks because of their lower level of RNA. The280 ⁄ 260 ration of the E128A NLPs was comparableR to the purified virus. Interestingly, this ratio was 50% higher with CPDN5 NLPs. The280 ⁄ 260 ratio of the isolated disks of CPDN5 and K97A was comparableP to the disks extrac- ted with acetic acid of the purified virus.

Gel shift assays To evaluate if theD ability to make NLPs was directly related to the affinity of CP for RNA, we isolated Fig. 4. Gel filtration analysis of the recombinant proteins CPDN5 disks fromE CPDN5 and the E128A mutants and com- and K97A. (A) CPDN5; 125 lg of the high-speed supernatant of the pared them with K97A. We used the high-speed super- purified protein was loaded on an FPLC SuperdexTM 200 column. natant of CPDN5 for isolation of the 450 kDa Grey line: Molecular weight markers, black line: protein elution pro- multimerT (disks). Since the E128A mutant makes only file. (B) K97A; 500 lg of the high speed supernatant of the purified NLPs in E. coli, we disrupted them using the acetic CPDN5 and 500 lg of purified K97A was loaded on an FPLC Super- dexTM 200 column. Grey line: K97A, black line: CPDN5. acid treatment and isolated E128A disks [20]. Different Camounts of disks were incubated in a volume of 10 lL RNA and the protein were assembled into NLPs of containing 165 fmol of a 32P-labelled RNA probe regular length (150 nm) that correspond to the length of made from a transcript of 80 nucleotides of the 5¢ non- the RNA (5¢ 1800 nt of PapMV) used for the in vitroE coding region of PapMV. The protein ⁄ RNA complex assembly assay (Fig. 5B). This result demonstrates was separated by an EMSA. The disks of CPDN5 clearly that only disks of approximately 20 subunits interacted with the probe in a cooperative manner and serve as building blocks for NLPs in vitro.R We per- induced a shift with 500 ng (22 pmol) of proteins formed the same experiment with the purified K97A (Fig. 6A). This result shows that the CPDN5 disks, recombinant protein. This mutant failed toR assemble the which are free of RNA after isolation, are able to O C N Fig. 5. In vitro assembly with recombinant CPDN5 disks. Purified CPDN5 disks (A) were assembled with 5¢ PapMV RNA tran- U script (B). Bars 200 nm.

6 FEBS Journal (2005) ª 2005 The Authors Journal compilation ª 2005 FEBS M.-H. Tremblay RNA binding and assembly of PapMV CP

A CP N5 (ng) B E128A(ng) 0 50 100 150 200 250 500 0 50 100 150 200 250 500 Protein-RNA Protein-RNA complex complex F O Probe Probe O Fig. 6. EMSA with recombinant CPs. (A) C K97A(ng) EMSA of the purified high speed superna- tant of CPDN5 (disks). (B) EMSA of the 50 100 150 200 250 500 750 1000 1500 0 recombinant E128A disks obtained from the R disruption of the NLPs by using the acetic acid method. (C) EMSA of the purified recombinant K97A proteins. Increasing P amounts of proteins were incubated at RT Protein-RNA for 30 min with 165 fmol of an RNA probe complex labelled with a-P32. The probe was made from an RNA transcript of 80 nts of the 5¢ end of PapMV noncoding region. The free Probe D probe and the RNA protein complex are shown with the arrows. E interact with RNA in vitro only when a molar ratio of the stability of the complex, their resistance to heat was 1000 (disks ⁄ RNA) is reached which, corresponds to a monitoredT by CD spectrophotometry. The resistance weak affinity for RNA. to denaturation of CPDN5 NLPs and the virus were We performed a similar experiment with the isolated remarkable (Fig. 7A). CPDN5 NLPs and virus are both disks of the mutant E128A and the same probe Cresistant to high temperature (>60 C) before showing (Fig. 6B). Interestingly, E128A disks bound RNA any sign of fatigue. The purified PapMV was the most more efficiently than CPDN5. As little as 50 ng of pro- stable structure tested and was able to resist tempera- teins were sufficient to create a protein RNA complex.E tures approaching 100 C. The temperature of inactiva- Purified K97A proteins in the same conditions failed, tion reported for PapMV is 70 C. Even if the virus still even at higher concentration (up to 1500 ng) to induce shows a high level of a-helices at this temperature, it is the formation of a protein RNA complex (Fig.R 6C). A likely that subtle changes to the structure occur and small molecular shift of the probe was noticed on the affect the infectivity of the virus. CPDN5 NLPs were gel but it is likely that this shift was made by the more sensitive than the WT virus probably because the degraded K97A protein that was detectedR as a mono- 6·H tag located at the C terminus affected the stability mer by gel filtration (Fig. 6C). Therefore, this small of the NLPs. The E128A NLPs appeared to be more shift is irrelevant to this study and we conclude that sensitive to heat and already showed signs of fatigue at the K97A multimers did not showO affinity for RNA. 42 C (Fig. 7A). Disks were rapidly denatured at 40 C We repeated the EMSA with RNAs extracted from (Fig. 7B). This result suggests that the packing of the CPDN5 NLPs that were labelled as described previ- disks in the rod structure considerably improves ously. We obtained the same results with this RNA stability. that does not contain theC PapMV packaging signal (data not shown). Trypsin digestion of disks and NLPs N Treatment of PapMV with trypsin results in a clea- CD spectra in denaturating conditions vage, presumably at amino acid 198 at the C terminus. To evaluate the stability of the NLPs and measure if Under those conditions, the remainder of the protein the assembly ofU the disks into a rod structure improved was resistant to the protease [1]. We repeated a similar

FEBS Journal (2005) ª 2005 The Authors Journal compilation ª 2005 FEBS 7 RNA binding and assembly of PapMV CP M.-H. Tremblay

F O O R P

Fig. 7. CD spectra heat denaturation curves of the purified recom- binant proteins. (A) Temperature-induced denaturations of CPDN5, E128A NLPs and WT virus. (B) Temperature-induced denaturations D of CPDN5 and WT disks. Spectra are presented in units of mean residue ellipticity ([h]). Unfolding was monitored by recording [Q] at 222 nm as a function of temperature. All CD spectra shown were )1 E generated with proteins at a concentration of 1 mg mL in 10 mM Fig. 8. Trypsin digests of WT, recombinant NLPs and disks. –, Neg- NaP buffer pH 7.2. ative controls without trypsin; +, trypsin was added to the sample. Each digestT contains 10 mg protein and 0.2 mg trypsin. Two micro- grams of protein was separated by SDS ⁄ PAGE and stained with assay on the purified virion and recombinant NLPs Coomassis blue. MW markers are shown on the left of the gel. and disks. Under our experimental conditions, PapMV C did not seem to be affected by trypsin (Fig. 8). We confirmed by electron microscopy that treated virus using this method generates partially denatured pro- was identical in appearance to untreated virus (dataE teins that are incapable of assembly and difficult to not shown). However, the isolated disks from CPDN5 remove from the extract. The E. coli expression system were very sensitive to trypsin and several bands of of the WT protein described in this study, allowed the lower molecular weight corresponding to degradedR identification of the building block of the PapMV fragments were generated (Fig. 8, lane 3). Similarly, rods; a 450-kDa disk composed of approximately 20 the CPDN5 NLPs were very sensitive to trypsin subunits. We also showed that the N-terminal five (Fig. 8, lane 5), suggesting that severalR positively amino acids of PapMV CP and the introduction of an charged residues are exposed and available at the sur- A after the initiation codon do not affect the capacity face of the NLPs. This is a major difference between of the protein to assemble into NLPs. The average size the CPDN5 rods and the WT virus.O E128A (Fig. 8, of the CPDN5 NLPs is rather small in bacteria prob- lane 7), was very resistant to trypsin like the WT virus. ably because the protein is specific for unstructured RNA similar to the PapMV packaging signal [2] that are not common in the bacteria. Furthermore, we Discussion C showed that the fusion of a peptide (6·H) to the C ter- Previous studies on potexvirus assembly were per- minus is tolerated and exposed to the surface of NLPs. formed with proteins extracted from purified virus by This finding is consistent with a tritium planigraphy acetic acid treatmentN [21]. This method generated var- analysis of the potato virus X CP that proposes the C ious aggregated forms, from which it was difficult to terminus to be exposed at the surface of the virus [22]. identify the building block of the virus particles [5]. Most of the mutations that we introduced in the pre- Furthermore,U the disassembly of the PapMV virus dicted RNA-binding domain located between amino

8 FEBS Journal (2005) ª 2005 The Authors Journal compilation ª 2005 FEBS M.-H. Tremblay RNA binding and assembly of PapMV CP acids 104 and 198 [13] (R117D119R120 ⁄ A, The K97A mutation generated a protein that har- K134K137 ⁄ A, D142D145 ⁄ A, R151A and E166E167R- bour the same secondary structure than the WT pro- 168 ⁄ A) (R104K105R108 ⁄ A and E148A) were very tein but that completely abolished the interaction with unstable or precipitated upon dialysis. This region of the RNA and maintained the protein in the disk structure. protein is conserved in the potexvirus family and is very This residue is a conserved amino acid in the potex- sensitive to mutations. However, the mutation E128A virus family (Fig. 1) that is probably critical for RNAF improved the affinity of the protein for RNA as shown binding. This result suggests that RNA is not import- by the EMSA (Fig. 6). E128A disks could make a strong ant for the multimerization of the protein into the disk gel shift with only one-tenth of the amount of protein structure. O that was necessary with CPDN5 disks. It is probable During PapMV infection, we propose that newly that E128A proteins in E. coli rapidly interact with synthesized CP will multimerize in a disk that accumu- RNAs and initiate the formation of NLPs with an lates until the assembly initiates. DisksO are probably improved efficiency as compared with the recombinant the only form of PapMV CP found in the infected protein CPDN5. Consequently, the improved affinity of plant cells. Because they have a weak affinity for viral E128A disks led to 100% efficiency of assembly in the RNA, they accumulate until sufficientR amount of viral bacteria and formation of longer NLPs. The E128A RNA is available for packaging. Assembly will be trig- NLPs are also more resistant to digestion by trypsin gered only when disks and RNA are sufficiently abun- than CPDN5 which suggests that they are more compact dant. Since the binding ofP the disks are cooperative, it at their surface. is likely that upon binding with RNA, the structure of The CD signal of E128A at 208 nm was more pro- the protein changes and become favourable for recruit- nounced than CPDN5 NLPs suggesting that the ment of other disks that allows the elongation to pro- a-helix content of E128A is higher than that of ceed. The assembly of PapMV, like most , is a CPDN5. The mutation of E128A somehow, compensa- late event in theD infection that is possibly dependant ted the detrimental influence of the 6·His on the sec- on the abundance of the disks and the genomic RNA. ondary structure of the protein and restored the a- It is clearE that nucleation of the disks on the RNA helix content to the WT level. E128A NLPs showed is essential to trigger the assembly process. The muta- also the same 280 ⁄ 260 ratio than the WT virus which tion K97A appears to inhibit the nucleation of the suggests that the RNA content is identical between the disksT on the RNA. Consequently, K97A can not two types of particles. CPDN5 NLPs showed a lower assemble in NLPs. E128A is a ‘super assembler’ in ratio probably because of their shorter length, since E.coli because this mutation improved the affinity of short RNAs are known to absorb less at 260 than lon- Cthe protein for RNA. Because E128A binds more ger RNAs. Finally, we showed that E128A NLPs are efficiently the PapMV genome, it could be interesting more sensitive to denaturation by heat than the WT to use this mutant in a transgenic approach to protect virus or CPDN5 NLPs. It is possible that residue E128E papaya plants against PapMV infection. is involved in an electrostatic interaction that stabilizes the disks in the rod structure. Several K residues found Experimental procedures in this region can potentially be involved in anR interac- tion with E128. It is possible that the repulsion Cloning of the PapMV CP gene between the negatively charged RNA backbone and E128 contributes to the low affinity ofR the disks for The PapMV CP gene was amplified by RT ⁄ PCR from RNA. This could be a regulation mechanism that isolated viral RNA using the forward and reverse oligonu- ensures that assembly of the genomic RNA occurs cleotide primers 5¢-AGTCCCATGGATCCAACGTCCAAT only when the CP is sufficiently abundant in the infec- CTTCTG-3¢ and 3¢-GAAGGTGGGGGGCTTGTGGTAG O CCTAGG ted cell. We also showed that the spectrum measured TGGTAGTGGTAATCATT CGTA-5¢, respect- between 250 and 350 nm was similar for CPDN5 and ively. The PCR product was digested with NcoI and E128A NLPs and CPDN5 and K97A disks. The muta- BamHI and inserted into the vector pET-3d to generate the tions apparently did not affectC the tertiary structure of clone CPDN5, in which five amino acids at the N terminus the proteins. Interestingly, the shape of the signal was were deleted from the WT sequence. CPDN5 also harbours clearly different for the disks and the NLPs as already the insertion of an alanine at position 2 of the recombinant protein. All the mutants presented in this manuscript were reported [19]. Therefore,N the environment of the aro- derived from this construct. matic and tryptophan residues is different between The E128A mutation was introduced by PCR on the these two structures. U CPDN5 clone using the oligonucleotides forward (E128A)

FEBS Journal (2005) ª 2005 The Authors Journal compilation ª 2005 FEBS 9 RNA binding and assembly of PapMV CP M.-H. Tremblay

5¢-GCTCCTGCCAATTGGGCGGCTTCAGGATACAAG- constructs, and maintained in 2·YT medium containing ) 3¢ and reverse (E128A) 3¢-CGAGGACGGTTAACCCGCC ampicillin (50 lgÆmL 1). Bacterial cells were grown at 37 C GAAGTCCTATGTTC-5¢. Two small PCR products were to an optical density of 0.6 at 600 nm and protein expres- first generated using the forward (E128A) primer with the sion was induced with 1 mm isopropyl-b-d-thiogalactopyr- CPDN5 reverse primer and the forward (CPDN5) primer anoside (IPTG). Induction was continued for 16 h at with the reverse (E128A) primer. These two PCR products 25 C. Bacteria were harvested by centrifugation for 15F min were ligated together by PCR using the forward and the at 6000 r.p.m. The pellet was resuspended in ice-cold lysis

reverse primer at each end of the CPDN5 construct and buffer (50 mm Na H2PO4 pH 8.0, 300 mm NaCl, 10 mm ) cloned in pET-3d using the NcoI ⁄ BamH1 restriction sites as imidazol, 20 lm phenylmethanesulfonyl fluiride,O 1 mgÆmL 1 described above to generate a clone expressing the E128A lysosyme) and bacteria were lysed by sonication. The lysate recombinant protein. The clone K97A, R104K105R108 ⁄ A, was centrifuged twice for 30 min at 13 000 r.p.m. to elimin- R118D120K121 ⁄ A, K133K137 ⁄ A, D142D145 ⁄ A, E148A, ate cellular debris. The supernatant was incubated with R161A and E166E168 ⁄ A were generated using the same 1 mL Ni–NTA (Qiagen, Turnberry Lane,O Valencia, CA) approach as E128A. The primers used for the PCRs are under gentle agitation for 4 h at 4 C. Lysates were loaded shown at Table 1. onto a column and the beads were washed with 3 · 15 mL m m washing buffer (50 m Na H2POR4 pH 8.0, 300 m NaCl) containing increasing concentrations of imidazole (10 mm, Expression and purification of recombinant 20 mm and 50 mm). The beads were then washed with proteins from E. coli 15 mL working buffer (10P mm Tris ⁄ HCl pH 8 or 10 mm The E. coli expression strain BL21(DE3) RIL (Stratagene, soduim phosphate buffer pH 7.2). Proteins were eluted in Torrey Pines Road, La Jolla, CA) was transformed with working buffer containing 1 m imidazol. The purity of the the plasmid pET-3d containing either of the different proteins was determinedD by SDS ⁄ PAGE and confirmed by Table 1. Primers used to introduce mutations in the PapMV CP by PCR for amplification of the mutated forms of the CP. Name Sequence E K97A Forward 5¢-GCACAATTGGCTAGTATTGTCT GCAGCTTCCGGCACTTCCCTT-3¢ Reverse 3¢-CGTGTTAACCGATCATAACAGCGTCGAAGGCCGTGAAGGGAA-5¢ R104-K105-R108 ⁄ A Forward 5¢-GCTTCCGGCACTTCCCTTGCAGCATTCTGCGCGTACTTCGCGCCAATA-3¢ Reverse 3¢-CGAAGGCCGTGAAGGGAAC CGTCGTAAGACGCGCATGAAGCGCGGTTAT-5¢ R118-D120-K121 ⁄ A Forward 5¢-ATAATCTGGAATCTGGCGACGGCCGCAATGGCTCCTGCCAATTGG-3¢ Reverse 3¢-TATTAGACCTTAGACE CGCTGCCGGCGTTACCGAGGACGGTTAACC-5¢ E128A Forward 5¢-GCTCCTGCCAATTGGGCGGCTTCAGGATACAAG-3¢ ReverseR 3¢-CGAGGACGGTTAACCCGCCGAAGTCCTATGTTC-5¢ K133-K137 ⁄ A Forward 5¢-GCCTCAGGATACGCACCAAGCGCCGCCTTTGCCGCGTTC-3¢ Reverse 3¢-CGGAGTCCTATGCGTGGTTCGCGGCGGAAACGGCGCAAG-5¢ D142-D145 ⁄ A R Forward 5¢-TTTGCCGCGTTCGCCTTCTTCGCCGGGGTGGAGAAT-3¢ Reverse 3¢-AAACGGCGCAAGCGGAAGAAGCGGCCCCACCTCTTA-3¢ E148a ForwardO 5¢-TTCTTCGACGGGGTGGCGAATCCGGCGGCCATG-3¢ Reverse 3¢-AAGAAGCTGCCCCACCGCTTAGGCCGCCGGTAC-5¢ R161A ForwardC 5¢-CAACCCCCTTCGGGACTAATCGCGTCGCCGACCCAGGAAGAGCGG-3¢ Reverse 3¢-GTTGGGGGAAGCCCTGATTAGCGCAGCGGCTGGGTCCTTCTCGCC-5¢ E166-E167-R168 ⁄ A Forward 5¢-CTAATCAGGTCGCCGACCCAGGCAGCGGCGATTGCCAATGCTACCAACAAA-3¢ ReverseN 3¢-GATTAGTCCAGCGGCTGGGTCCGTCGCCGCTAACGGTTACGATGGTTGTTT-5¢ Amorces communes 2 N-terminutesus Forward 5¢-ATCGCCATGGCATCCACACCCAACATAGCCTTCCCCGCCATCACC-3¢ C-terminutesusU Reverse 3¢-GAAGGTGGGGGGCTTGTGGTAGTGGTAGTGGTAATCATTCCTAGGCGTA-5¢

10 FEBS Journal (2005) ª 2005 The Authors Journal compilation ª 2005 FEBS M.-H. Tremblay RNA binding and assembly of PapMV CP

3 Table 2. OD ratio 280 ⁄ 260 of the proteins. OD measurements were taken three times on the spectrophotometer on different protein prepa- rations. The results were consistent between measurements. The recombinant CPDN5 rods and disks were separated by ultracentrifugation; disks: supernatant, NLPs: pellet. The measurement was taken directly on the purified recombinant proteins E128A and K97A.

Virus and NLPs Extracted disks Purified PapMV CPDN5 NLPs E128A NLPs Purified PapMV CPDN5F K97A Ratio 280 ⁄ 260 0.75 1.10 0.75 1.5 1.64O 1.55 Western immunoblot analysis using rabbit polyclonal anti- RNA transcripts and EMSA bodies generated against purified PapMV virus. The RNA probe was generated by transcriptionO in vitro using a RiboMAX Large Scale RNA Production System- Electron microscopy and immunogold labelling T7 kit (Promega P1300, Woods Hollow Road, Madison, NLPs or viruses were diluted in 10 mm Tris ⁄ HCl pH 8 and WI) and a clone of 80 nt of the 5R¢ end of PapMV in front were absorbed for 3 min on a carbon-coated formvar grid. of the T7 promoter. The clone was linearized with EcoRI Grids were blocked with 8 lL BSA (10 lgÆlL)1) for 30 s before in vitro transcription. The RNA transcript was puri- and washed with NaCl ⁄ Pi. Grids were incubated for 30 min fied on a G-50 Quick SpinP Column for DNA ⁄ RNA purifi- at room temperature with a rabbit anti-6·H tag antibody cation (Roche no. 1273 965; Indianapolis, IN). The same

(Amersham, Pittsburgh, PA) diluted 1 : 10 in NaCl ⁄ Pi. method was used to generate a transcript of the 5¢ 1800 Grids were then washed three times with NaCl ⁄ Pi and incu- nucleotides of PapMV for the in vitro assembly assay. The bated at room temperature for 30 min with donkey antirab- RNA probe was dephosphorylated using shrimp alkaline bit antibodies conjugated with 6-nm gold particles (Jackson phosphatase (Fermentas,D Connelley Drive, Suite A, Han- Immuno-Research, West Baltimore Pike, West Grove, PA) over, MD, USA, EF0511) and labelled with gamma 32 and diluted 1 : 20 in NaCl ⁄ Pi. Grids were then washed with P-ATP usingE T4 polynucleotide kinase (NEB, Ipswich, deionized water and stained as described above. MA). The probe was then purified using the G-50 Quick Spin Columns as before. Labelled RNA was incubated with recombinant proteins at room temperature for 60 min. We SDS/PAGE and electroblotting T used 165 fmol of RNA for each reaction and various Proteins were mixed with one-third the final volume of amounts of purified recombinant proteins in the in vitro loading buffer containing 5% SDS, 30% glycerol, 0.01% Cassembly buffer, which contained 7.5 U RNase inhibitor Bromophenol blue. SDS ⁄ PAGE was performed as des- (Amersham Biosciences, 27-0816-01). The final volume of cribed elsewhere [11]. the reaction was 10 lL; 2 lL loading dye were added to E the sample before loading onto a 5% native polyacrylamide gel. Electrophoresis was performed in 0.5· Tris ⁄ bor- CD spectroscopy ate ⁄ EDTA buffer for 90 min at 10 mA. The gel was dried CD spectra were recorded on an Olis RSM 1000 (Olis, and subjected to autoradiography for 16 h on Kodak Bio- R Max MS film (Amersham Biosciences, V8326886) and Conway DriveSuites A and B, Bogart, GA) rapid scanner monochromator at 20 C. For far UV CD (260–190 nm), developed. thermostated quartz cells of 0.1 cm path length were used. R 2 Mean residue ellipticity values ([h]MRW in deg · cm · Purification of PapMV and isolation of disks dmol) were calculated using the equation: [h]MRW ¼ [h]*MRW ⁄ (10 · c · l), where [h] is the ellipticity in PapMV was purified from infected papaya leaves that degrees, MRW is the average molecularO weight of the resi- showed mosaic symptoms. Infected leaves (100 mg) were dues in the protein (108 was used in this study), c is the ground in 100 mL 50 mm Tris ⁄ HCl pH 8,0 containing protein concentration in gÆmL)1 and l is the path length 10 mm EDTA in a commercial blender. The ground leaves in cm [12]. C were filtered through cheese-cloth and 1% of Triton X-100 Near UV CD spectra were recorded (250–350 nm) at was added to the filtrate and gently stirred for 10 min. room temperature in a Jasco Model J-710 instrument (Jas- Drop by drop, chloroform was added until one-quarter the co, Commerce Dr Easton,N MD). Recordings were made volume of the filtrate. The solution was stirred for an addi- using quartz cuvettes (pathlength 0.1 cm). Spectra were tional 30 min at 4 C and centrifuged 20 min at 10 000 g to averaged from 10 scans of 0.2 nm steps at a rate of remove the precipitate. The supernatant was subjected to a 100 nmÆmin)1. Rod and disks samples were, respectively, at high speed (100 000 g) for 120 min in a Beckman 50.2Ti a concentrationU of 1.5 mgÆmL)1 and 5.5 mgÆmL)1. rotor. The viral pellet was resuspended and subjected to

FEBS Journal (2005) ª 2005 The Authors Journal compilation ª 2005 FEBS 11 RNA binding and assembly of PapMV CP M.-H. Tremblay another high speed centrifugation through a sucrose cush- 9 Dolja VV, Boyko VP, Agranovsky AA & Koonin EV ion (30% sucrose) at 100 000 g for 3.5 h. The final viral (1991) Phylogeny of capsid proteins of rod-shaped and pellet was resuspended in 10 mL 50 mm Tris pH 8.0. If the filamentous RNA plant viruses: two families with dis- colouration persisted, an additional clarification with chlo- tinct patterns of sequence and probably structure con- roform was performed and the purified virus was collected servation. Virology 184, 79–86. by ultracentrifugation. Isolation of the disks by acetic acid 10 Atabekov JG, Rodionova NP, Karpova OV, KozlovskyF degradation was performed as described before [3]. SV, Novikov VK & Arkhipenko MV (2001) Transla- tional activation of encapsidated potato virus X RNA by coat protein phosphorylation. Virology 286, 466–474. Trypsin digest O 11 Scha¨gger H & von Jagow G (1987) Tricine-sodium We incubated 10 lg of proteins at 37 C in a volume of dodecyl sulfate-polyacrylamide gel electrophoresis for 50 lL for 120 min in a 100 mM Tris ⁄ HCl pH 8.5 with the separation of proteins in the range from 1 to 100 0.2 lg trypsin (Roche, 1418475). The reaction was stopped kDa. Anal Biochem 166, 368–379. O by addition 10 lL loading dye containing 5% SDS, 5 mm 12 Johnson WC (1996) Circular dichroism instrumentation. dithiothreitol and 40% glycerol. The sample was boiled for Circular Dichroism and the Conformational Analysis of 5 min prior loading on a SDS ⁄ PAGE. The proteins were Biomolecules (Fasman GD, ed.),R pp. 635–652. Kluwer visualized by Coomassie blue staining. 1 Academic ⁄ Plenum Publishers. 13 Abouhaidar MG & Lai R (1989) Nucleotide sequence of the 3¢-terminal regionP of clover yellow mosaic virus Acknowledgements RNA. J Gen Virol 70, 1871–1875. We thank the Natural Sciences and Engineering 14 Stubbs G (1999) Tobacco mosaic virus particle structure Research Council of Canada for funding our research and the initiation of disassembly. Philos Trans R Soc program on papaya mosaic virus, Dr Helen Rothnie Lond B Biol Sci 354, 551–557. 15 Anindya R, JosephD J, Gowri TD & Savithri HS (2004) for critical reading of our manuscript and Franc¸ois Complete genomic sequence of Pepper vein banding Otis for his help with the CD spectrum (CREFSIP). virus (PVBV): a distinct member of the genus Potyvirus. Arch VirolE149, 625–632. References 16 Jacob T & Usha R (2002) Expression of Cardamom mosaic virus coat protein in Escherichia coli and its 1 Zhang H, Todderud E & Stubbs G (1993) Crystalliza- T assembly into filamentous aggregates. Virus Res 86, tion and preliminary X-ray analysis of papaya mosaic 133–141. virus coat protein. J Mol Biol 234, 885–887. 17 Voloudakis AE, Malpica CA, Aleman-Verdaguer ME, 2 Sit TL, Abouhaidar MG & Holy S (1989) Nucleotide C Stark DM, Fauquet CM & Beachy RN (2004) Struc- sequence of papaya mosaic virus RNA. J Gen Virol 70, tural characterization of Tobacco etch virus coat protein 2325–2331. mutants. Arch Virol 149, 699–712. 3 Erickson JW, Bancroft JB & Horne RW (1976) The E 18 Blanch EW, Hecht L, Syme CD, Volpetti V, Lomonoss- assembly of papaya mosaic virus protein. Virology 72, off GP, Nielsen K & Barron LD (2002) Molecular 514–517. structures of viruses from Raman optical activity. J Gen 4 Erickson JW, Abouhaidar M & Bancroft JB (1978) The R Virol 83, 2593–2600. specificity of papaya mosaic virus assembly. Virology 19 Erickson JW, Bancroft JB & Stillman MJ (1981) Circu- 90, 60–66. lar dichroism studies of papaya mosaic virus coat pro- 5 Erickson JW, Hallett FR & Bancroft JB (1983) Sub- R tein and its polymers. J Mol Biol 147, 337–349. assembly Aggregates of papaya mosaic-virus protein. 20 Abouhaidar M & Bancroft JB (1978) The initiation of Virology 129, 207–211. papaya mosaic virus assembly. Virology 90, 54–59. 6 Sit TL, Leclerc D & AbouHaidar MG (1994) The mini- 21 Fraenkel-Conrat H (1957) Degradation of tobacco mal 5¢ sequence for in vitro initiationO of papaya mosaic mosaic virus with acetic acid. Virology 4, 1–4. potexvirus assembly. Virology 199, 238–242. 22 Baratova LA, Grebenshchikov NI, Shishkov AV, 7 Abouhaidar MG & Bancroft JB (1980) The polarity of Kashirin IA, Radavsky JL, Jarvekulg L & Saarma M assembly of papaya mosaic-virusC and tobacco mosaic- (1992) The topography of the surface of potato virus X: virus RNAs with PMV-protein under conditions of non- tritium planigraphy and immunological analysis. J Gen specificity. Virology 107, 202–207. Virol. 73, 229–235. 8 Durham AC & BancroftN JB (1979) Cation binding by papaya mosaic virus and its protein. Virology 93, 246– 252. U

12 FEBS Journal (2005) ª 2005 The Authors Journal compilation ª 2005 FEBS Author Query Form

Journal: FEBS

Article: 5033

Dear Author, During the copy-editing of your paper, the following queries arose. Please respond to these by marking up your proofs with the necessary changes/additions. Please write your answers on the query sheet if there is insufficient space on the page proofs. Please write clearly and follow the conventions shown on the attached corrections sheet. If returning the proof by fax do not write too close to the paper’s edge. Please remember that illegible mark-ups may delay publication.

Many thanks for your assistance.

Query reference Query Remarks Q1 Place of publication? Q2 Au: Is this name correct? Q3 Table 2 has not been found in the text. MARKED PROOF ÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐ Please correct and return this set ÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐ Please use the proof correction marks shown below for all alterations and corrections. If you wish to return your proof by fax you should ensure that all amendments are written clearly in dark ink and are made well within the page margins.

Instruction to printer Textual mark Marginal mark Leave unchanged under matter to remain Stet Insert in text the matter New matter followed by indicated in the margin Delete through matter to be deleted Delete and close up through matter to be deleted Substitute character or through letter or through New letter or new word substitute part of one or word more word(s) Change to italics under matter to be changed Change to capitals under matter to be changed Change to small capitals under matter to be changed Change to bold type under matter to be changed Change to bold italic under matter to be changed Change to lower case Encircle matter to be changed Change italic to upright type (As above) Insert `superior' character through character or where under character required e.g. Insert `inferior' character (As above) over character e.g. Insert full stop (As above) Insert comma (As above) Insert single quotation marks (As above) and/or Insert double quotation (As above) and/or marks Insert hyphen (As above) Start new paragraph No new paragraph Transpose Close up linking letters Insert space between letters between letters affected Insert space between words between words affected Reduce space between letters between letters affected Reduce space between words between words affected