Molecular characterization, phylogenetic comparison and serological relationship of the Imp protein of several ‘Candidatus Phytoplasma aurantifolia’ strains




The immunodominant membrane protein Imp of several phytoplasmas within the ‘Candidatus Phytoplasma aurantifolia’ (16Sr-II) group was investigated. Eighteen isolates from Iran (11), East Asia (5), Africa (1) and Australia (1) clustered into three phylogenetic subgroups (A, B and C) based on the 16S rDNA and imp genes, regardless of geographic origin. The imp gene sequences were variable, with more non-synonymous than synonymous mutations (68 vs 20, respectively), even though many of the non-synonymous ones (75%) produced conservative amino acid replacements. Eight codon sites on the extracellular region of the protein were under positive selection, with most of them (75%) coding for non-conservative amino acid substitutions. Full-length (21 kDa) and truncated (16 kDa) Imp proteins of two economically important Iranian phytoplasmas [lime witches’ broom (LWB) and alfalfa witches’ broom (AlWB-F)] were expressed as His-tagged recombinant proteins in Escherichia coli. An antiserum raised against full-length recombinant LWB Imp reacted in western blots with membrane proteins extracted from LWB-infected periwinkle and lime, indicating that Imp (19 kDa) is expressed in infected plants and is a membrane-associated protein. The same polyclonal antibody also detected native Imp in proteins from periwinkles infected by phytoplasmas closely related to LWB (subgroup C) only, confirming phylogenetic clustering based on 16S rDNA and imp genes. Imp proteins of LWB and AlWB-F isolates were also recognized by an antiserum raised against an enriched preparation of AlWB-F phytoplasma cells, demonstrating the antigenic properties of this protein.


Phytoplasmas are a group of monophyletic plant-pathogenic mollicutes, naturally transmitted by phloem-feeding homopterous insects in a persistent propagative manner. They are intracellular pathogens which colonize both plant and insect cells and are responsible for diseases in over 700 plant species (Hogenhout et al., 2008). Inability to culture them in vitro has hampered their characterization at the functional and molecular levels. However, the complete genomes of four phytoplasmas, including two strains of ‘Candidatus Phytoplasma asteris’ [onion yellows (OY; Oshima et al., 2004) and aster yellows witches’ broom (AY-WB; Bai et al., 2006)], ‘Ca. Phytoplasma australiense’ (Tran-Nguyen et al., 2008) and ‘Ca. Phytoplasma mali’ (Kube et al., 2008) have been sequenced. Genome analyses showed that phytoplasmas have undergone a reductive evolution and lack some crucial metabolic pathways. They depend mainly on hosts for the production of essential metabolites (Oshima et al., 2004; Bai et al., 2006). The genome size of phytoplasmas in comparison with other self-propagating bacterial microorganisms is very small. A considerable number of phytoplasma genes encode membrane and secreted proteins, which probably enable them to communicate with and manipulate their hosts (Hogenhout et al., 2008). Studies on these membrane and secreted proteins can provide a better understanding of host–phytoplasma relationships.

Immunodominant membrane proteins, which are the target of most antibodies raised against phytoplasma cells, are located at the exterior surface of phytoplasma cells, and are the most abundant proteins of the cell membrane (Shen & Lin, 1993; Milne et al., 1995). Using polyclonal and monoclonal antibodies, it has been shown that phytoplasmas have one (Clark et al., 1989) or two (Saeed et al., 1992) immunodominant membrane proteins on their surface with a molecular mass between 18 and 36 kDa. Based on the structure of the protein and the position of the corresponding gene, phytoplasma immunodominant membrane proteins have been classified (Kakizawa et al., 2006a) as: (i) immunodominant membrane protein (Imp) expressed in plants infected by phytoplasmas related to distinct 16Sr groups, including 16Sr-I (Kakizawa et al., 2009), -II (Yu et al., 1998), -III (Neriya et al., 2011), -X (Berg et al., 1999; Morton et al., 2003), and a few others (Kakizawa et al., 2009); (ii) immunodominant protein A (IdPA) identified in western X disease phytoplasma (Blomquist et al., 2001); and (iii) antigenic membrane protein (Amp) identified in different strains of ‘Ca. Phytoplasma asteris’ (Kakizawa et al., 2004; Bai et al., 2006; Galetto et al., 2008), clover phyllody phytoplasma (Barbara et al., 2002), and stolbur (Fabre et al., 2011) phytoplasma. Orthologues of these genes have been sequenced and characterized in several other phytoplasmas.

Sequence analysis of Amp in phytoplasmas related to ‘Ca. Phytoplasma asteris’ and stolbur phytoplasma revealed that this protein is highly variable because of the accumulation of non-synonymous mutations. It was also suggested that Amp is subject to strong positive selection (Kakizawa et al., 2006b; Fabre et al., 2011). Further studies suggested a role for Amp in transmission specificity of OY and chrysanthemum yellows phytoplasmas (Suzuki et al., 2006; Galetto et al., 2011).

In phytoplasmas related to 16Sr-I and 16Sr-III groups, genes coding for Imp and other immunodominant membrane proteins, Amp or IdpA, are present (Kakizawa et al., 2009; Neriya et al., 2011). The coexistence of imp with other types of immunodominant genes in some phytoplasmas suggests a role as a common ancestor for Imp protein. Genetic diversity analysis has shown that imp genes in different phytoplasmas are more variable than their surrounding genes (Kakizawa et al., 2009). Moreover, the ability of the Imp protein of ‘Ca. Phytoplasma mali’ to bind plant actin has suggested a role for this protein in phytoplasma mobility within the infected cell (Boonrod et al., 2012). However, more sequence information is still needed to better describe imp variability and understand the role of the protein in different groups of phytoplasmas.

In this study, imp gene sequences of several strains of phytoplasmas related to ‘Ca. Phytoplasma aurantifolia’ (16Sr-II; Zreik et al., 1995) according to IRPCM rules (IRPCM, 2004) are provided. Focusing on two economically important phytoplasmas, lime witches’ broom (LWB) and alfalfa witches’ broom (AlWB-F), imp genes were analysed, expressed as recombinant proteins for antibody production and Imp expression studied in natural and laboratory host plants.

Materials and methods

Phytoplasma isolates and DNA extraction

Eighteen phytoplasma isolates, all related to ‘Ca. Phytoplasma aurantifolia’ (16Sr-II group), were used in this work (Table 1). Fifteen isolates from periwinkle were from Shiraz (Table 1, isolates 1–11) and Torino collections (Table 1, isolates 12–15). Total DNAs were extracted from 1·5 g leaf tissues of phytoplasma-infected periwinkle, following the method described by Maixner et al. (1995).

Table 1. List of phytoplasma isolates of the 16Sr-II group used in the present study with their geographical origins, and corresponding 16S rDNA and imp gene accession numbers
Isolate numberPhytoplasma strainAbbreviationOrigin16S rDNA/imp accession numbers
  1. ND, not determined.

 1Lime witches’ broomLWBSouthern Iran U15442/JQ745272
 2Alfalfa witches’ broomAlWB-FCentral Iran DQ233655/JQ745273
 3Alfalfa witches’ broomAlWB-YCentral Iran DQ233656/JQ745274
 4Tomato big budATBBCentral Iran JX083376/JQ745286
 5Carrot witches’ broomCWBCentral Iran JX083372/JQ745280
 6Eggplant big budEBB-ACentral Iran JX083377/JQ745284
 7Eggplant big budEBB-BSouthern Iran JX083373/JQ745276
 8Eggplant big budEBB-FCentral Iran JX083375/JQ745283
 9Parsley witches’ broomPWB-YCentral Iran JX083374/JQ745282
10Periwinkle phyllodyPPh-CSouthern IranND/JQ745275
11Periwinkle phyllodyPPh-JSouthern IranND/JQ745281
12Crotalaria phyllodyCrPhThailand EF193355/JQ745277
13Crotalaria witches’ broomCrWBChina EU650181/JQ745279
14Faba bean phyllodyFBPSudan X83432/JQ745278
15Tomato big budTBB-IAustralia EF193359/JQ745285
16Pear declinePDTWIITaiwan EF193157/GU214177
17Peanut witches’ broomPnWBTaiwan GU113148/GU214176
18Sweet potato witches’ broomSPWBTaiwan DQ452417/U15224

imp gene amplification, cloning, sequencing and prediction of protein structure

Primer pair IMF1\IMR1 (IMF1: 5′-AATTGAAGGCGATATTGAATCT-3′; IMR1: 5′-ATTTGGTTTGTAGGGGTTCA-3′) was designed on regions flanking the imp sequence of sweet potato witches’ broom (SPWB) phytoplasma and used to amplify this gene from phytoplasma isolates 1–15 listed in Table 1. PCR amplification was carried out with 50 ng total nucleic acid extracted from the plants in 50 μL reaction volumes following a standard protocol. The imp-containing amplicons were ligated into the pTZ57R/T cloning vector (Fermentas). The recombinant vectors were used to transform Escherichia coli DH5α. Three clones for each phytoplasma isolate were sequenced (SEQLAB, Göttingen, Germany) and the sequence verified by blast searches within the NCBI database. Furthermore, imp sequences of SPWB (Yu et al., 1998), Taiwan pear decline (PDTWII, Liu et al., 2011) and peanut witches’ broom phytoplasmas (PnWB; Lin & Lin, 1998) (Table 1, isolates 16–18) were obtained from GenBank and used in this study. The complete ORFs coding for Imp were retrieved and amino acid sequences were deduced in silico. Prediction for transmembrane domains was performed using topcons ( and hmmtop ( with advanced settings and submitting all homologous Imp proteins clustering within the imp-C subgroup. Signal peptides were analysed by SignalP (v. 3) (

Sequence analyses and phylogenetic tree construction

Multiple alignments of imp and 16S rDNA gene sequences were obtained using clustalW v. 1.7 ( The 16S rDNA gene partial sequences (1·2 kb) were trimmed to R16F2n and R16R2 (Gundersen & Lee, 1996) for phylogenetic analyses, with and without ‘Ca. Phytoplasma mali’ (AJ542542) as an out-group. The mean distance values for imp genes were computed using the Tajima–Nei model implemented in mega v. 4.0 (

Phylogenetic trees for both genes were inferred by the neighbour-joining (NJ) method implemented in mega v. 4.0 through 100 bootstrap replicates. The random effect likelihood (REL) approach implemented in the hyphy program ( was used to identify codons under positive selection. For this purpose the translated amino acid sequences of the imp genes were aligned using clustalW in the dambe program ( The codon-based alignment and the NJ tree were used to find codon sites under positive selection across imp. The REL approach allows both synonymous and non-synonymous substitution rates (dS and dN, respectively) to vary among sites. The program was set to consider a codon site under positive selection if the posterior probability of dN–dS >0 for the site was more than 0·95. A Bayes factor cut-off of 100, which corresponds to a low P value (P∼1/Bayes factor) was used in the REL analysis.

Grantham’s matrix (Grantham, 1974) was used as a tool to identify sites with radical amino acid substitution in the amino acid alignment of different Imp proteins. Based on this matrix, variable sites including amino acids with a Grantham’s distance of more than 100 (corresponding to dissimilar amino acid substitution) were recognized. Grantham’s distance is a method of pairwise amino acid comparison based on physicochemical properties and is believed to be an appropriate method for distinguishing between conservative and radical amino acid replacements.

Heterologous expression of Imp and antibody production

Primer pairs IMPF/IMPR (IMPF: 5′-ATGAATCACAAAGAAATTTTTTAC-3′; IMPR: 5′-TTATGATAATTTTAAATCTG-3′) and IMPT/IMPR (IMPT: 5′-GGAAACCTTTTAACATAACA-3′) were used to amplify the entire and a truncated part of the imp sequence devoid of transmembrane domain at the N-terminus, respectively. The entire imp sequence was amplified from the LWB phytoplasma, whereas the truncated part was amplified from both the LWB and AlWB-F phytoplasmas. The purified PCR products were directly ligated in the expression vector pQEUA (QIAGEN) in fusion with 6 × His tag at the N-terminus. Escherichia coli M15 competent cells were transformed using the recombinant pQEUA plasmids. The presence and orientation of the gene were verified by PCR and sequencing. Clones harbouring each construct were induced by IPTG (1 mm) to express fusion protein following the manufacturer’s instruction. Induced cells were then grown for 4 h, harvested and lysed in Tris-buffered saline (TBS) with or without 8 m urea for purification under denaturing or native conditions, respectively. The expression was confirmed using SDS-PAGE analysis. For antibody production, the full-length recombinant LWB Imp was purified from a large-scale culture (500 mL) using a Ni-NTA agarose column under native conditions. Full-length recombinant Imp was administered to a rabbit by four injections (100 μg each) at 1-week intervals, and antiserum was harvested 2 weeks after the fourth injection. The rabbit serum was collected and precipitated by ammonium sulphate. The precipitate was dissolved in PBS and dialysed against fresh PBS overnight at 4°C. IgG was purified by DEAE-cellulose chromatography (Harlow & Lane, 1988). The concentration of IgG was adjusted to 1 mg mL−1. This antibody against recombinant LWB Imp, together with a polyclonal antibody previously prepared against purified AlWB-F phytoplasma cells (Salehi et al., 2010), were used in western blot analysis.

Protein extraction and western blot

Total proteins from healthy or infected periwinkle and healthy or naturally LWB-phytoplasma-infected lime plants were extracted by crushing 0·5 g midribs in Laemmli sample buffer (Laemmli, 1970). Membrane proteins were also isolated by the phase fractionation method with Triton X-114, as described by Bordier (1981). Total proteins in Laemmli sample buffer, membrane and soluble proteins extracted with Triton X-114, and recombinant proteins expressed in E. coli were separated through 12% SDS-PAGE (Laemmli, 1970). For western blot analysis, proteins separated in SDS-PAGE were then blotted onto a polyvinyl difluoride (PVDF, Bio-Rad) membrane. After 1 h blocking with 3% bovine serum albumin (BSA) in TBS containing 0·1% Tween (TBS-T), PVDF membranes were incubated with either the anti-LWB Imp (1:1000) or anti-AlWB-F (1:1000) antibodies for 2 h at room temperature. After washings, membranes were incubated at room temperature for 2 h with horseradish peroxidase (HRP)-conjugated goat anti-rabbit antibody (Sigma-Aldrich) diluted (1:10 000) in 3% BSA in TBS-T, developed with Super Signal West Pico reagents (Pierce) and visualized using a VersaDoc model 4000 imaging system (Bio-Rad).

Lectin assay

The ability of recombinant LWB Imp to bind glycoproteins was tested. A lectin assay was conducted based on Killiny et al. (2005). The glycoproteins ovalbumin and ArtI (Galetto et al., 2008), a phytoplasma membrane protein recombinantly expressed and known for its lectin properties (L. Galetto, unpublished data), were included as positive controls. Proteins were separated in a 12% SDS-PAGE and blotted onto PVDF membranes. After 2 min blocking with 2% Tween in TBS, PVDF membranes were incubated for 2 h in TBS-T with glycoprotein ovalbumin (10 μg mL−1), washed, treated with HRP-conjugated concanavalin A (Sigma-Aldrich) diluted at 1 μg mL−1 in TBS-T and developed as described for western blot.


Sequence analyses and phylogenetic tree construction

Representative NJ phylogenetic trees based on 16S rDNA and imp gene alignments are shown in Fig. 1. In the imp-based phylogenetic tree, phytoplasma strains were grouped into three subgroups (A, B and C). The highest and the lowest mean sequence distances were between subgroups A and C (0·21) and subgroups A and B (0·08), respectively. Isolates in subgroup C were most variable, with a mean distance of 0·021. In the 16S rDNA-based phylogenetic tree, phytoplasma strains were grouped into three subgroups corresponding to those identified on the basis of imp analysis, with the exception of the carrot witches’ broom (CWB) isolate, which clustered within the 16S rDNA-A and imp-C subgroups. Most phytoplasma isolates from East Asia grouped within the 16S rDNA-B and imp-B subgroups, with the exception of crotalaria phyllody (CrPh), which grouped within the C subgroup. In contrast, Iranian isolates grouped within the A and C subgroups, suggesting that the geographical origin of phytoplasmas has little effect on their phylogenetic divergence based on these two genes.

Figure 1.

 Phylogenetic trees of 16Sr RNA (a) and imp (b) gene sequences of phytoplasma isolates. Trees were constructed using the neighbour-joining method. Numbers on branches show the bootstrap values. Phytoplasma abbreviations are listed in Table 1.

The length of the imp gene varied from 507 to 519 nucleotides (coding for a protein of about 19 kDa, theoretical pI 8·86–9·8) among the 18 phytoplasma strains used in this study (accession numbers in Table 1). Pairwise sequence identity scores showed that nucleotide-level similarity of imp sequences ranged from 79·4 [LWB vs eggplant big bud (EBB)-A] to 100% [tomato big bud (TBB)-I vs tomato big bud (ATBB)]. Figure 2 shows the alignment of Imp amino acid sequences of all isolates used in this study, with the exception of ATBB and TBB-I (identical to EBB-F) and pear decline [PDTWII; identical to peanut witches’ broom (PnWB)]. Imp showed a short stretch of conserved hydrophilic amino acids and a membrane-spanning helix (positions 16–37). The main part of the protein, following the transmembrane domain up to the C-terminus, was predicted to be exposed outside of the cell and hosted most of the amino acid differences among isolates (Fig. 2). Signal peptide analyses for all the Imp proteins under study, based on the neural networks algorithm, detected a signal sequence with a cleavage score (C-score) lower, and other scores (S-score, Y-score, D-score and S-mean) higher than the default. These results were not confirmed following the hidden Markov model analysis.

Figure 2.

 Alignment of Imp amino acid sequences of ‘Candidatus Phytoplasma aurantifolia’ isolates. All isolates used in this study (listed in Table 1) are aligned, with the exception of ATBB, TBB-I and PDTWII. Black and grey boxes show identical and similar amino acids at each site, respectively. The transmembrane domain was predicted by the hmmtop program and is shown by a dark line. Sites under positive selection, calculated by the REL approach, are indicated by asterisks above. Sites containing amino acid substitutions with a Grantham’s distance of over 100 are indicated by arrowheads below.

The codon-based comparison revealed 68 non-synonymous mutations and 20 synonymous changes among a total of 173 aligned codons. Despite the abundance of non-synonymous mutations, only 17 amino acid substitutions showed a Grantham’s distance >100 (Fig. 2). The REL method, used to identify positively selected codon sites across the imp gene, revealed eight sites under positive selection (Fig. 2). The dN–dS values calculated for these sites ranged between 1·156 in site 140 and 1·213 in site 156, with Bayes factors of 273·44 and 3265·8, respectively. The highest Bayes factor obtained among these eight sites was 3350·8 at position 145, with a dN–dS value of 1·212. All of the sites inferred to be under positive selection were located in the extracytoplasmic region of the protein. Most of the codon sites undergoing positive selection (six out of eight) coded for highly dissimilar amino acids, with a Grantham’s distance of over 100 (Fig. 2), suggesting that part of these radical amino acid substitutions in the Imp can be explained by positive selection.

Expression of LWB fusion Imp and western blot analysis

The whole imp ORF (519 bp) of the LWB phytoplasma was cloned in E. coli in fusion with a 6 × His tag. This fusion protein (21 kDa), expressed with a very low yield, was purified under native conditions and used to raise an antiserum in rabbit. Truncated imp ORFs (lacking the first 40 amino acids at the N-terminus, 16 kDa) of the LWB and AlWB-F phytoplasmas were expressed with higher yields (5 mg L−1). Recombinant LWB truncated Imp showed no lectin activity (data not shown).

The polyclonal antibody raised against the full-length LWB phytoplasma Imp recognized the homologous recombinant full-length LWB protein (data not shown) and also the truncated fusion proteins from AlWB-F and LWB phytoplasmas in western blot analysis (Fig. 3a). This antibody also detected a 19-kDa band, probably the native Imp, in membrane or total proteins of periwinkle plants infected by LWB, AlWB-F, faba bean phyllody (FBP), CrPh, CWB and periwinkle phyllody (PPh)-C phytoplasmas (all belonging to subgroup C; Fig. 4a), as well as from field-collected LWB-infected limes (Fig. 4b). The same antibody did not detect any antigen in soluble protein extracts from periwinkles infected by group-C isolates (AlWB-F, FBP, CrPh, CWB; Fig. 4a), thus confirming the membrane localization of Imp. The antibody did not react with proteins extracted from periwinkles infected by subgroup A (TBB-I) and B [crotalaria witches’ broom (CrWB)] or those extracted from healthy periwinkle (Fig. 4a) and lime plants (Fig. 4b).

Figure 3.

 Western blot of transformed Escherichia coli proteins expressing truncated Imp of lime witches’ broom (LWB) and alfalfa witches’ broom (AlWB-F) phytoplasmas. Total proteins from E. coli transformed with truncated Imp of LWB (1) and AlWB-F (2) phytoplasmas were probed with antibodies against LWB full-length recombinant Imp (a) or AlWB-F phytoplasma (b).

Figure 4.

 Western blots of membrane (M), soluble (S) and total (T) proteins from healthy and phytoplasma-infected periwinkle (a, c) and lime (b) plants probed with antisera against LWB full-length recombinant Imp (a, b) or AlWB-F phytoplasma cells (c). Phytoplasma abbreviations are listed in Table 1.

To assess whether Imp is an antigenic membrane protein (immunodominant), a polyclonal antibody previously produced against purified AlWB-F phytoplasma cells (Salehi et al., 2010) was used. This polyclonal antibody recognized the recombinant Imp of both AlWB-F and LWB phytoplasmas (Fig. 3b), as well as the 19-kDa native Imp in membrane or total proteins from periwinkle infected by AlWB-F and LWB phytoplasmas (Fig. 4c). The anti-AlWB-F antibody recognized an additional band of 36 kDa in total and membrane proteins extracted from periwinkle infected by the AlWB-F phytoplasma (Fig. 4c). Therefore, two proteins of 19 kDa (Imp) and 36 kDa are the main immunodominant proteins of the AlWB-F phytoplasma, and their membrane location is confirmed by the presence of an antigenic signal in the membrane fraction proteins of the infected plants.


The imp genes from several phytoplasmas belonging to ‘Candidatus Phytoplasma aurantifolia’ (16Sr-II group) were cloned and sequenced, and the full-length sequence of LWB, as well as partial sequences of LWB and AlWB-F Imp proteins, were expressed as fusion antigens. The results of this study showed that the identity between imp gene sequences in isolates related to the 16Sr-II phytoplasma group was lower than that recorded for the corresponding 16S rDNA gene sequences, as already suggested for phytoplasmas within the 16Sr-X ribosomal group (Morton et al., 2003). The imp-based phylogenetic tree confirmed the grouping obtained by 16S rDNA sequence analysis, with the exception of the Iranian CWB isolate. Western blot analysis with the antibody raised against the LWB phytoplasma Imp confirmed grouping of the CWB isolate within subgroup C of the imp-based tree. SPWB, CrWB, PnWB and PDTWII, all from East Asia, clustered in the same subgroup in both phylogenetic trees. Grouping of the other isolates was not correlated with their geographical origin, with isolates from Iran clustering in two subgroups in both trees.

These findings suggest that forces other than geographical separation influence imp divergence, as already suggested for different 16S ribosomal phytoplasma groups (Kakizawa et al., 2009). Genetic diversity analysis of the imp gene within 16Sr-II phytoplasmas showed a higher accumulation of non-synonymous (77%) to silent mutations, although only about 25% of them resulted in a radical amino acid substitution. Diversifying positive selection pressure acted on several codon sites of imp, although it could explain only about 35% of the total radical amino acid substitutions. All the codon sites under positive selection and all the radical amino acid mutations of 16Sr-II phytoplasma Imp proteins were located in the extracellular domain of the protein, while the N-terminal part, up to the transmembrane region, was much more conserved, as already described for Imp proteins of other 16S ribosomal groups (Morton et al., 2003; Kakizawa et al., 2009). Selection pressure acting on the extracellular portion of another immunodominant phytoplasma membrane protein, Amp, was demonstrated in ‘Ca. Phytoplasma asteris’ strains (Kakizawa et al., 2006b). Moreover, the Amp of two ‘Ca. Phytoplasma asteris’ strains specifically bind to only a few insect vector proteins, suggesting a role for this polypeptide in determining vector specificity (Suzuki et al., 2006; Galetto et al., 2011).

The molecular weight of the detected Imp protein in infected plants was in line with the predicted mass from amino acid sequence analysis, therefore confirming the uncleavable status of the N-terminal signal sequence in isolates related to the 16Sr-II group, in agreement with Imp proteins of phytoplasmas belonging to other groups (Yu et al., 1998; Berg et al., 1999; Kakizawa et al., 2009). The detection of Imp only in the Triton X-114 fraction of infected plants confirmed the membrane localization of this polypeptide, as already reported for apple proliferation phytoplasma Imp (Berg et al., 1999), and as predicted by hmmtop and topcons for 16Sr-II phytoplasma Imp proteins.

The polyclonal antibody against AlWB-F phytoplasma cells reacted with recombinant Imp from AlWB-F and LWB phytoplasmas. The same antibody detected a 19-kDa antigen from proteins of LWB and AlWB-F phytoplasma-infected periwinkles, therefore suggesting that Imp is the antigenic membrane protein of these phytoplasmas. Imp is also the unique antigenic membrane protein of SPWB phytoplasma (Yu et al., 1998) and some members of the 16Sr-X group (Clark et al., 1989; Berg et al., 1999). In the case of AlWB-F, a specific 36-kDa signal in the membrane-protein fraction probed with AlWB-F antibody was also detected, although with a weaker signal than the 19-kDa Imp. Saeed et al. (1992) also detected two antigens of 18 and 36 kDa in proteins from FBP-infected plants probed with an antibody against purified phytoplasma cells. In this case, the 36-kDa protein signal was putatively ascribed to dimerization of the 18-kDa antigen. According to the present results, this can be excluded because of the absence of any signal in the correct MW range (36–38 kDa) when membrane-protein extracts from AlWB-F infected plants were probed with the anti-Imp antibody. Therefore, two proteins of 36 and 19 kDa are the major antigens on the membrane surface of the AlWB-F phytoplasma. On the other hand, the antibody raised against AlWB-F detected only the 19-kDa protein (Imp) in extracts from LWB-infected plants, suggesting that the 36-kDa protein is less conserved than Imp across subgroup-C phytoplasmas. However, under the experimental conditions of this study, the presence of this protein could not be confirmed in isolates other than AlWB-F.

Recombinant LWB Imp showed no lectin activity, so it cannot be involved in interactions with the glyco-proteome of the insect vector, in contrast with what is reported for spiralin, the major antigenic protein of Spiroplasma citri (Killiny et al., 2005).

The high variability of the Imp protein within 16Sr-II phytoplasmas, resulting from positive selection pressure, highlights the importance of understanding the role of this membrane protein in interactions with insect and plant hosts.


This research was supported by funds from the Iranian council of centre of excellence and comprehensive programme of Lime Witches’ Broom Management, and by the Italian Regional Grant (Piemonte Region) ‘Studi sui fattori che favoriscono le epidemie di Flavescenza dorata in Piemonte e loro superamento’.