Combinations of two amino acids (Ala40 and Phe75 or Ser40 and Tyr75) in the coat protein of apple chlorotic leaf spot virus are crucial for infectivity

Apple chlorotic leaf spot virus (ACLSV) is the type species of the genus Trichovirus. It has filamentous particles approximately 600–700 nm in length that contain a polyadenylated, single-stranded, plus-sense RNA and multiple copies of a single coat protein (CP) of 21 kDa (Yoshikawa & Takahashi, 1988). ACLSV is distributed worldwide and is known to infect fruit-tree species of the family Rosaceae, including apple, peach, pear, plum, cherry and apricot (Lister, 1970; Martelli et al., 1994; Yoshikawa, 2001). Complete nucleotide sequences have been reported for isolates P863 and PBM1 from plum, BAL1 from cherry and P205 from apple (German et al., 1990; Sato et al., 1993; German-Retana et al., 1997). The genome of an apple isolate of ACLSV (P205) consists of 7552 nt and contains three open reading frames (ORFs 1, 2 and 3) (Sato et al., 1993). The 216 kDa protein (KP) encoded by ORF1 is a replication-associated protein (Rep), 50KP encoded by ORF2 is a movement protein (MP) and CP is encoded by ORF3. The similarity of the complete nucleotide sequences of these four ACLSV isolates (P863, P205, PBM1 and BAL1) is 76.2–81.5 %, and those of the amino acid sequences of 216KP, 50KP and CP are 84.3–88.5, 77.2–84.8 and 87–93.3 %, respectively, among these isolates (German-Retana et al., 1997). Thus, the genomes of ACLSV have great molecular variability between isolates.

In Japan, the renewal of apple cultivars has been conducted by top-grafting the scion of other cultivars on apple trees. ACLSV can be transmitted from the scion to the rootstock or in the reverse direction by grafting when either the scion or the rootstock is infected with the virus. ACLSV is one of the causative agents of apple topworking disease and induces lethal decline in apple trees grown on Maruba kaido (Malus prunifolia ringo) rootstocks in Japan (Yanase, 1974). It is thought that the occurrence of apple topworking disease in Japan is due to top-grafting the cultivars imported from North America onto apple trees grown on Maruba kaido rootstocks. At present, it is well known that ACLSV, as well as other apple viruses (apple stem grooving virus and apple stem pitting virus), in single trees consists of a mixture of isolates or sequence variants, and each isolate or variant is thought to be derived from a scion of other trees (Magome et al., 1997, 1999; Yoshikawa et al., 2001). To understand the molecular variability of the viral genome and its effect on biological properties is important to forestall the possible outbreak of viral diseases.

In this study, we compared the amino acid sequences of ACLSV CP among 12 isolates and 109 cDNA clones amplified directly from infected apple tissues. The results showed that the isolates and cDNA clones were separated into two major clusters, in which five amino acids at positions 40, 59, 75, 130 and 184 were highly conserved within each cluster. Covariation between Ala⁴⁰-Val⁵⁹-Phe⁷⁵-Ser¹³⁰-Met¹⁸⁴ and Ser⁴⁰-Leu⁵⁹-Tyr⁷⁵-Thr¹³⁰-Leu¹⁸⁴ was found in the two clusters. We also showed that the combinations of the two amino acids at positions 40 and 75 are crucial for effective virus replication in host plant cells.

Virus isolates and plants.
The ACLSV isolates used in this study were P205 and MO5, originally described by Yanase (1974), A4 and B6 from an apple tree showing fruit russet ring symptoms, and GC10a, GC10c, GC10f, GC10h and GC10j from an apple tree showing green crinkle symptoms. These isolates, except for P205 and MO5, were obtained by a single-lesion transfer five times on Chenopodium quinoa leaves.

Leaf, petal and apple samples from 23 apple trees, summarized in Table 1, were also used for direct PCR amplification of CP genes.

Table 1. Apple-tree samples used for cDNA amplification, cloning and sequence analysis of the CP gene of ACLSV

RNA extraction.
Total RNAs were extracted from apple leaves and petals by using an ISOPLANT kit (Wako) according to the manufacturer's instructions with slight modification, i.e. solution P1 was supplemented with 2 % 2-mercaptoethanol. When apples were used for analysis, pericarps were peeled thinly and homogenized with 5 vols (v/w) CTAB buffer [0.1 M Tris/HCl (pH 9.5), 0.02 M EDTA (pH 7.0), 1.4 M NaCl, 5 % 2-mercaptoethanol, 2 % CTAB]. The homogenates were mixed with an equal volume of chloroform, shaken vigorously for 1 min and centrifuged at 3000 r.p.m. for 30 min. The supernatants were collected, 0.25 vol. (v/v) 10 M LiCl was added and incubated overnight on ice. High-molecular-mass RNAs were precipitated by centrifugation at 3000 r.p.m. for 30 min and then subjected to 2 M LiCl fractionation. The precipitates were dissolved in 100 µl distilled water, treated with DNase I (TaKaRa) and finally dissolved in distilled water. Total RNAs were extracted from C. quinoa and Nicotiana occidentalis leaves as described previously (Yoshikawa et al., 2000). Double-stranded RNAs (dsRNAs) were purified from 2 M LiCl-soluble total RNAs by binding with CC41 (Whatman) as described previously (Isogai et al., 1998).

RT-PCR and sequence analysis.
A cDNA fragment corresponding to the ACLSV CP coding region was amplified by RT-PCR from the total RNAs as follows. Reverse transcription was performed by using an oligo(dT)₂₀ primer and Moloney murine leukemia virus reverse transcriptase (Toyobo). The resulting cDNA was subjected to PCR using Ex Taq polymerase (TaKaRa) and two primers: ACUNI6745(+) (5'-GAGAATTTCAGTTTGCTCGA-3') and ACUNI3'(–) (5'-AGTCTACAGGCTATTTATTATAAGT-3'). The PCR product was ligated to the pGEM-T Easy vector (Promega) or the pT7Blue T-Vector (Novagen), and the plasmids were used to transform Escherichia coli JM109 cells. At least three independent cDNA clones were sequenced by using an ABI PRISM 3100 DNA sequencer (Applied Biosystems) or by a custom sequencing service (Macrogen). Sequence data were collected, assembled and analysed with DNASIS (Hitachi). Multiple alignment of amino acid sequences was obtained with CLUSTAL_W (Thompson et al., 1994). A phylogenetic tree was constructed by the neighbour-joining method (Saitou & Nei, 1987) and the statistical significance of branch order was estimated by performing 100 replications of bootstrap resampling of the original alignment with CLUSTAL_W.

Site-directed mutagenesis and construction of deletion mutants.
Substitutions of single and multiple sites at positions 40 (Ala to Ser), 59 (Val to Leu), 75 (Phe to Tyr), 130 (Ser to Thr) and 184 (Met to Leu) of ACLSV CP were conducted as follows. To introduce substitutions of the four amino acids at positions 59, 75, 130 and 184, the 3' terminus, containing the partial CP coding region corresponding to nucleotide positions 6888–7558, was amplified from pCLSF (an infectious cDNA clone based on isolate P205; Satoh et al., 1999) by PCR using a forward primer, CLS6888EcoRI(+) [5'-GCGAATTCACAGACACTGGAGGC-3', containing an EcoRI site (underlined)], and a reverse primer, CLS3'Xba(–) [5'-TTCTAGATTTTTGTAGTAAAATATTTAAAAGT-3', containing an XbaI site (underlined)]. The PCR product was double-digested with EcoRI and XbaI and ligated to pUC19 restricted with the same enzymes. The resulting plasmid, p19CLSCP, was subjected to in vitro mutagenesis using an LA PCR in vitro mutagenesis kit (TaKaRa) and mutational primers containing one or two nucleotide changes (bold italic) of each codon [ACCPmut59(+) 5'-TTCCTGGACTTGCTGGTG-3', ACCPmut75(+) 5'-TGGGGTCATACAATCTGAAG-3', ACCPmut130(+) 5'-AACCTCTTTACTACTATGCC-3', ACCPm184(+) 5'-AGGCGAAACTGTCGTCTGT-3', and ACCPmm184(+) 5'-AGGCGAAACTCTCGTCTGT-3']. An amino acid substitution at position 40 was introduced with PCR using two primers: CLS6574EcoRI(+) [5'-GCGAATTCGATCTGCTAGCCTGAG-3', containing an EcoRI site (underlined)] and ACCPmut40Nru(–) [5'-TTCGCGAAGATGGACTCCA-3', containing a nucleotide change (italic) and an NruI site (underlined)]. The PCR product was double-digested with EcoRI and NruI and ligated to p19CLSCP restricted with the same enzymes. The resulting plasmids, containing single or multiple amino acid substitutions, were double-digested with NruI or NheI and XbaI, and ligated to pCLSF (Satoh et al., 1999) restricted with the same enzymes. The resulting cDNA clones were designated pCPm40, pCPm75, pCPm130, pCPm184, pCPmm184, pCPm40m75, pCPm59m75, pCPm59m130, pCPm75m130, pCPm40m59m75, pCPm40m75m130, pCPm59m75m130, pCPm59m130m184, pCPm40m59m75m130, pCPm59m75m130m184 and pCPm40m59m75m130m184 (Fig. 3a). These plasmids were purified from large-scale cultures by using a plasmid midi kit (Qiagen) and used for mechanical inoculation onto C. quinoa at a concentration of 0.5 µg µl^–1.

(32K): Fig. 3. Mutational analysis of the five conserved amino acids of ACLSV CP. (a) Summary of the CP amino acid substitution mutants based on pCLSF and their infectivity on C. quinoa. Bold lines in the enlarged CP boxes indicate the position of amino acid changes from Ala40 to Ser40, Val59 to Leu59, Phe75 to Tyr75, Ser130 to Thr130 and Met184 to Leu184. + and – indicate the presence and absence of infectivity of the mutants, respectively. P, 35S promoter of cauliflower mosaic virus; T, nopaline synthase terminator; Rep, 216 kDa replication-associated protein; MP, 50 kDa movement protein; CP, coat protein. (b) Immunoblot analysis of ACLSV CP extracted from C. quinoa leaves infected with pCLSF and the CP amino acid substitution mutants at 7 days post-inoculation, using polyclonal antibodies against ACLSV particles. The arrow and arrowhead indicate positions of CP with different mobility.

To create a P216 deletion mutant (pBICLΔRep), the region corresponding to nucleotide positions 4304–5429 of the ACLSV genome (Sato et al., 1993) was removed from pBICLSF (Yaegashi et al., 2007) by cutting with KpnI. The P50 deletion mutant (pBICLΔMP) and CP mutants with one or two amino acid substitutions (pBICLCPstop, pBICLCPm40, pBICLCPm75 and pBICLCPm40m75) were constructed by replacing the BamHI–XbaI fragment of pBICLSF with the same region of pΔStuNhe (Yoshikawa et al., 2000), pCPstop, pCPm40, pCPm75 and pCPm40m75, respectively. The resulting Ti plasmids were introduced into Agrobacterium tumefaciens strain C58C1 by a freeze–thaw method. Cultures of A. tumefaciens carrying Ti plasmid constructs were infiltrated at an OD₆₀₀ of 2.0 on N. occidentalis leaves as described previously (Yaegashi et al., 2007). The infiltrated plants were kept in a growth chamber at 24 °C.

Immunoblot analysis.
Total protein samples from C. quinoa leaves inoculated with the plasmids described above were electrophoresed in a 12.5 % polyacrylamide/SDS gel and transferred electrophoretically to a PVDF membrane (Millipore). The membrane was incubated with an antiserum against ACLSV particles or the 50 kDa MP, followed by an anti-rabbit IgG (H&L) alkaline phosphatase-linked antibody (Cell Signaling Technology) and immersed in development solution containing a Fast Red TR salt (Sigma) and a naphthol AS-MX phosphate (Sigma).

RNA analysis.
For Northern blot analysis of ACLSV RNA, heat-denatured RNAs from infiltrated leaves were separated on a 1 % agarose gel containing 6 % formaldehyde and transferred to a Hybond-N+ membrane (Amersham Biosciences). After UV cross-linking, the membrane was hybridized with a digoxigenin (DIG)-labelled RNA probe complementary to the plus-strand RNA (nt 152–1133) or minus-strand RNA (nt 6888–7552). The hybridized membrane was immunodetected with an anti-DIG Fab fragment coupled to alkaline phosphatase (Roche) and visualized with a chemiluminescent substrate, CSPD (Amersham Biosciences), on X-ray films.

Covariation of the five amino acids at positions 40, 59, 75, 130 and 184 in ACLSV CP
The CP amino acid sequences of nine Japanese isolates [P205 (GenBank accession no. D14996[GenBank] ), MO5 (AB326225[GenBank] ), A4 (AB326223[GenBank] ), B6 (AB326224[GenBank] ), GC10a (AB326226[GenBank] ), GC10c (AB326227[GenBank] ), GC10f (AB326228[GenBank] ), GC10h (AB326229[GenBank] ) and GC10j (AB326230[GenBank] )] and three European isolates [P863 (M58152[GenBank] ), PBM1 (AJ243438[GenBank] ) and BAL1 (X99752[GenBank] )] were compared. Multiple alignment of the amino acid sequences of CP revealed that covariation of the five amino acids at positions 40, 59, 75, 130 and 184 occurred, i.e. Ala⁴⁰, Val⁵⁹, Phe⁷⁵, Ser¹³⁰ and Met¹⁸⁴ for isolates P205, A4, GC10c, GC10h and GC10j, and Ser⁴⁰, Leu⁵⁹, Tyr⁷⁵, Thr¹³⁰ and Leu¹⁸⁴ for isolates B6, MO5, GC10a, GC10f, P863, PBM1 and BAL1, except that the amino acid at position 184 of BAL1 was Ile instead of Leu (Fig. 1). We designated the isolates containing the former combination of the five amino acids as P205 type and the isolates containing the latter combination as B6 type.

(27K): Fig. 1. Alignment of CP amino acid sequences of ACLSV isolates, including nine Japanese isolates from apple (P205, A4, B6, MO5, GC10a, GC10c, GC10f, GC10h and GC10j) and three European isolates from plum (P863 and PBM1) and sweet cherry (BAL1). Asterisks indicate the five amino acids that show covariation between isolates at amino acid positions 40, 59, 75, 130 and 184. Amino acids that differ from those of isolate P205 are shown below the P205 sequence.

In order to investigate whether the covariation of the five amino acids is found in other ACLSV isolates or variants, we analysed the CP amino acid sequences of 109 cDNA clones amplified directly by RT-PCR from 23 apple-tree samples (Table 1). The results showed that most cDNA clones have the combination of the five amino acids of either the P205 type or the B6 type at positions 40, 59, 75, 130 and 184. In a phylogenetic tree constructed from the CP amino acid sequences, ACLSV isolates and cDNA clones were divided into two large clusters that each corresponded fully to the P205 type or the B6 type based on these five amino acids (Fig. 2).

Table 1). Two clusters (types P205 and B6) in the tree are circled. To simplify the phylogenetic tree, detailed bootstrap values and the names of cDNA clones are not shown. Bar, 0.1 amino acid replacements per site.

Mutational analysis of the five amino acids in the CP conserved among isolates
To investigate whether the five amino acids in the CP are important for viral infectivity and pathogenesis, we constructed 16 viral cDNA clones with one or multiple amino acid changes from the P205 type to the B6 type (pCPm40, pCPm59, pCPm75, pCPm130, pCPm184, pCPm40m75, pCPm59m75, pCPm59m130, pCPm75m130, pCPm40m59m75, pCPm40m75m130, pCPm59m75m130, pCPm59m130m184, pCPm40m59m75m130, pCPm59m75m130m184 and pCPm40m59m75m130m184; Fig. 3a). These clones were inoculated mechanically onto C. quinoa plants to test for their infectivity. C. quinoa plants inoculated with pCLSF, pCPm59, pCPm130, pCPm184, pCPm40m75, pCPm59m130, pCPm40m59m75, pCPm40m75m130, pCPm59m130m184, pCPm40m59m75m130 or pCPm40m59m75m130m184 all showed chlorotic and necrotic spots on inoculated leaves at 4–5 days post-inoculation. These plants all developed systemic symptoms in upper leaves similar to those caused by P205 infection. In contrast, neither symptoms nor virus accumulation were found on inoculated or upper uninoculated leaves of C. quinoa plants inoculated with pCPm40, pCPm75, pCPm59m75, pCPm75m130, pCPm59m75m130 or pCPm59m75m130m184. Comparison of the mutated positions of the five amino acids between infectious and non-infectious cDNA clones suggested strongly that the combinations of two amino acids (Ala⁴⁰ and Phe⁷⁵ or Ser⁴⁰ and Tyr⁷⁵) are necessary for infectivity on C. quinoa (Fig. 3a). There was no difference of symptoms on infected C. quinoa and N. occidentalis between pCPm59, pCPm130, pCPm184, pCPm40m75, pCPm59m130, pCPm40m59m75, pCPm40m75m130, pCPm59m130m184, pCPm40m59m75m130 and pCPm40m59m75m130m184 (data not shown).

Immunoblot analysis using an antiserum against ACLSV particles showed that viral CP was detected from symptomatic C. quinoa leaves inoculated with the ten clones listed above, and CP species with two different electrophoretic mobilities could be distinguished (Fig. 3b). The difference in electrophoretic mobility of CP depends on the amino acid at position 59, i.e. the CPs containing Leu⁵⁹ migrate faster than those with Val⁵⁹. These results were also confirmed by the analysis of natural isolates (slower-migrating type: P205, A4, GC10a, GC10c, GC10h and GC10j; faster-migrating type: B6 and GC10f).

We next assessed the stability of the mutations introduced into the CP by RT-PCR and sequence analysis of progeny viral RNA extracted from upper leaves of infected C. quinoa plants. Sequence analysis of three randomly selected cDNA clones revealed that the mutation(s) were maintained in CP from leaves infected with pCPm59, pCPm130, pCPm40m75, pCPm59m130, pCPm40m59m75, pCPm40m75m130, pCPm59m130m184, pCPm40m59m75m130 or pCPm40m59m75m130m184. In leaves infected with pCPm184, however, the amino acid at position 184 had reverted to the original amino acid (Met). To prevent this reversion, pCPmm184, carrying a mutation by two nucleotide substitutions of the first and third positions of a codon, was constructed. C. quinoa plants inoculated with pCPmm184 developed symptoms on inoculated and upper leaves similar to those caused by pCLSF infection, and the amino acid mutation at position 184 was maintained in the progeny virus.

Combinations of the two amino acids at positions 40 and 75 of CP are important for effective viral replication
To investigate whether amino acid substitution at position 40 or 75 has an effect on RNA accumulation or on the cell-to-cell movement of the virus, we utilized an Agrobacterium-mediated efficient viral inoculation system that is able to deliver viral cDNA to a large number of cells of leaves (English et al., 1997). This inoculation system would be useful for measuring the RNA accumulation level, independently of the ability of virus cell-to-cell movement (Voinnet et al., 2000; Chiba et al., 2006). A series of mutants was constructed based on pBICLSF (Fig. 4a) (Yaegashi et al., 2007). In addition to pBICLCPm40, pBICLCPm75 and pBICLCPm40m75, Rep, MP and CP-defective mutants, denoted pBICLΔRep, pBICLΔMP and pBICLCPstop, respectively, were used as comparable controls.

(60K): Fig. 4. Analysis of genomic RNA accumulation of ACLSV mutants in N. occidentalis leaves. (a) Schematic representation of pBICLSF-based mutants used for agroinfiltration. Bold lines in the CP box of pBICLCPm40, pBICLCPm75 and pBICLCPm40m75 indicate the amino acid substitutions at positions 40 (Ala to Ser) and 75 (Phe to Tyr). Abbreviations are as described in the legend to Fig. 3(a). (b) Northern blot analysis of viral genomic RNA extracted from leaves infiltrated with buffer only (Mock) or agrobacteria carrying each mutant shown in Fig. 4(a) at 3 days post-infiltration. The arrow and arrowhead indicate positions of the genomic RNA derived from pBICLSF and pBICLΔMP, respectively. Ethidium bromide staining of rRNA is shown as a loading control. (c) Immunoblot analysis of viral proteins (MP and CP) extracted from leaves infiltrated with buffer only (Mock) or agrobacteria carrying pBICLSF, pBICLCPm40, pBICLCPm75 or pBICLCPm40m75, using polyclonal antibodies against MP (Anti-MP) or ACLSV particles (Anti-CP) at 3 days post-infiltration. The arrows indicate positions of MP and CP. (d) Northern blot analysis of viral dsRNAs extracted from leaves infiltrated with agrobacteria carrying pBICLSF, pBICLCPm40, pBICLCPm75 or pBICLCPm40m75 at 3 days post-infiltration. The arrows indicate the positions of genomic RNA (g), subgenomic RNA1 (sg1) and subgenomic RNA2 (sg2).

N. occidentalis leaves were infiltrated with agrobacteria carrying each clone, and then the accumulation levels of viral RNA in infiltrated regions were analysed by Northern blotting at 3 days post-infiltration. As shown in Fig. 4(b), genomic RNA could be detected in leaves infiltrated with agrobacteria carrying pBICLSF and pBICLΔMP, but not with pBICLΔRep (Fig. 4b), indicating that the system can be used for the estimation of the RNA accumulation level in cells, as reported for other viruses (Voinnet et al., 2000; Chiba et al., 2006; Gopinath et al., 2005). When leaves were infiltrated with agrobacteria carrying pBICLCPstop, pBICLCPm40 or pBICLCPm75, the accumulation of genomic RNA was suppressed strongly compared with those infiltrated with agrobacteria carrying pBICLSF or pBICLΔMP (Fig. 4b). In contrast, the accumulation level of genomic RNA in leaves infiltrated with agrobacteria carrying pBICLCPm40m75 was almost equivalent to that of pBICLSF-infiltrated leaves.

To resolve whether an amino acid change at position 40 or 75 in CP has affected viral replication, immunoblot analysis of leaves infiltrated with agrobacteria carrying pBICLSF, pBICLCPm40, pBICLCPm75 or pBICLCPm40m75 at 3 days post-infiltration was carried out by using polyclonal antibodies against MP or virus particles. As shown in Fig. 4(c), accumulation of both MP and CP was below detectable levels in samples infiltrated with agrobacteria carrying pBICLCPm40 or pBICLCPm75. In contrast, both proteins were detected in leaves infiltrated with agrobacteria carrying pBICLSF or pBICLCPm40m75. Analysis of viral dsRNAs by Northern blotting using a minus-strand-specific probe revealed that accumulation of viral dsRNAs was reduced considerably in leaves infiltrated with agrobacteria carrying pBICLCPm40 or pBICLCPm75 compared with that in leaves infiltrated with pBICLSF or pBICLCPm40m75 (Fig. 4d). These results suggest strongly that the combinations of the two amino acids at positions 40 and 75 of CP are important for effective replication in host plant cells.

Comparison of the amino acid sequences of ACLSV CP between 12 isolates and 109 cDNA clones amplified directly from 23 apple trees in Japan revealed covariation of the five amino acids at positions 40, 59, 75, 130 and 184 of CP into two types (P205 and B6), which correspond fully to two major clusters in a phylogenetic tree constructed from amino acid sequences of CP (Fig. 2). The results suggest that covariation of these five amino acids defines the two lineages phylogenetically. Alternatively, covariation of the five amino acids found in ACLSV CP may be associated with host adaptation. Recently, it has been reported that serial passages of two carmoviruses, Hibiscus chlorotic ringspot virus and Pelargonium flower break virus, in C. quinoa plants resulted in covariation of eight and five site-specific amino acids of CP of the progeny viruses, respectively (Liang et al., 2002; Rico et al., 2006). Additionally, it has been suggested that these changes are not likely to be random events, but may be due to host-associated positive selection or accelerated genetic drift.

Mutational analysis of the five amino acids of CP using an infectious cDNA clone indicated that the combinations of two amino acids (Ala⁴⁰ and Phe⁷⁵ or Ser⁴⁰ and Tyr⁷⁵) are crucial for infectivity on C. quinoa and N. occidentalis plants (Fig. 3). Furthermore, subsequent experiments using an agroinfiltration assay suggested that viral replication was inhibited strongly in N. occidentalis leaves infiltrated with mutants (pBICLCPm40 and pBICLCPm75) carrying an amino acid substitution at position 40 or 75, but not mutants carrying two substitutions at both positions (pBICLCPm40m75) (Fig. 4). Thus, the combinations of the two amino acids (Ala⁴⁰ and Phe⁷⁵ or Ser⁴⁰ and Tyr⁷⁵) are important for an efficient viral replication in cells. This idea is also supported by the results of sequence analysis, showing that the combinations of these two amino acids are fully conserved among all ACLSV isolates and cDNA clones amplified directly from infected apple trees. In contrast, the three amino acids at positions 59, 130 and 184 were not related directly to infectivity or symptom severity on herbaceous host plants. These experimental results suggest that the variation of the three amino acids at positions 59, 130 and 184 is probably due to stochastic divergence between lineages, whereas the variation at positions 40 and 75 represents an example of functional co-evolution.

Because the amino acid at position 40 of CP is within the region overlapping ORF2, the mutation at this position resulted in a change of the amino acid of MP at position 392. Therefore, we cannot strictly rule out the possibility that the combination of the two amino acids of MP (position 392) and CP (position 75) is important for viral infectivity. However, it has been reported that the C-terminal region of MP (aa 287–457) is not essential for the known functions of MP (Satoh et al., 2000; Isogai & Yoshikawa, 2005). In addition, when pCPm40 was inoculated onto N. occidentalis expressing MP (Yoshikawa et al., 1999) by particle bombardment, infection of pCPm40 could not be complemented (data not shown). Collectively, it is thought that the combination of the two amino acids in the CP, rather than those between MP and CP, may be necessary for infectivity and effective replication.

It is well known that the virus capsid protein has many functions, including replication, symptom modulation, cell-to-cell movement, systemic spread and suppression of RNA silencing, in addition to virion formation (Callaway et al., 2001; Thomas et al., 2003; Lu et al., 2004). Our data presented in this paper also suggest that ACLSV CP is involved in efficient RNA replication. It is reported that alfalfa mosaic virus and ilarvirus RNAs are not infectious unless a few molecules of CP or a CP subgenomic transcript are added to the inocula (Bol, 1999, 2003; Jaspars, 1999). In these viruses, the binding of CP to the 3' termini of viral RNAs is required for viral replication (Neeleman et al., 2004; Guogas et al., 2005; Krab et al., 2005). The role of ACLSV CP in the virus replication cycle is under investigation.

This work was in part supported by a Grant-in-Aid for the 21st Century Centers of Excellence Program from the Ministry of Education, Culture, Sports, Science and Technology of Japan.