|Home | About | Journals | Submit | Contact Us | Français|
Spiroplasma citri GII3 contains highly related low-copy-number plasmids pSci1 to -6. Despite the strong similarities between their replication regions, these plasmids coexist in the spiroplasma cells, indicating that they are mutually compatible. The pSci1 to -6 plasmids encode the membrane proteins known as S. citri adhesion-related proteins (ScARPs) (pSci1 to -5) and the hydrophilic protein P32 (pSci6), which had been tentatively associated with insect transmission, as they were not detected in non-insect-transmissible strains. With the aim of further investigating the role of plasmid-encoded determinants in insect transmission, we have constructed S. citri mutant strains that differ in their plasmid contents by developing a plasmid curing/replacement strategy based on the incompatibility of plasmids having identical replication regions. Experimental transmission of these S. citri plasmid mutants through injection into the leafhopper vector Circulifer haematoceps revealed that pSci6, more precisely, the pSci6_06 coding sequence, encoding a protein of unknown function, was essential for transmission. In contrast, ScARPs and P32 were dispensable for both acquisition and transmission of the spiroplasmas by the leafhopper vector, even though S. citri mutants lacking pSci1 to -5 (encoding ScARPs) were acquired and transmitted at lower efficiencies than the wild-type strain GII3.
Phytoplasmas and spiroplasmas, two groups of pathogenic mollicutes, are associated with many diseases affecting economically important crops, such as ornamentals, vegetables, fruit trees, and grapevine (6, 10, 26, 36). Whereas most plant-pathogenic bacteria colonize the apoplast of plant tissues, phytoplasmas and spiroplasmas are restricted to the phloem sieve tubes and are transmitted from plant to plant by phloem sap-sucking insects (31, 42), which are therefore responsible for the spread of the diseases. Despite the fact they share common habitats with phytoplasmas, spiroplasmas are distinguishable in that they display a characteristic helical morphology and can be cultured in vitro. Spiroplasma citri is the etiological agent of citrus stubborn disease (33). It also infects many other plants, including the Madagascar periwinkle (Catharanthus roseus), an experimental host plant in which it induces symptoms such as stunting, leaf yellowing, and wilting, eventually leading to plant death. In nature, S. citri is transmitted in a circulative, persistent manner by the leafhoppers Circulifer haematoceps (Hemiptera, Cicadellidae) (in the Mediterranean area and the Near East) (16) and Circulifer tenellus (in the United States) (28).
To complete their transmission cycle, spiroplasmas ingested by leafhopper vectors must cross two physical barriers, the gut epithelium (to move from the lumen to the hemocoel) and then the salivary gland-associated membranes (to reach the salivary duct) (24, 27). Mainly on the basis of electron microscopy observations, a hypothetical model in which spiroplasmas pass through these two barriers by an endocytosis-exocytosis process has been proposed (13). However, the molecular mechanisms that govern the interactions of the spiroplasma with cells of its leafhopper vector are poorly understood. Several proteins have been tentatively associated with insect transmission of S. citri, including Sc76 (the solute binding protein of an ABC transporter) (7), spiralin (the major lipoprotein at the cell surface) (12, 23), and the plasmid-encoded protein P89, or SARP1 (44), which was shown to be required for adhesion to insect cells in vitro (3).
S. citri strain GII3 has seven plasmids, pSciA and pSci1 to -6, five of which (pSci1 to -5) encode eight S. citri adhesion-related proteins (ScARPs) and one of which (pSci6) encodes the cytosolic protein P32 (34). The role of plasmid-encoded determinants in insect transmission of S. citri was first suggested from the observation that, in contrast to S. citri GII3, the non-insect-transmissible strains did not express ScARPs and P32 (22) and did not carry the pSci1 to -6 plasmids encoding these proteins (5). In addition, the finding that pSci6 conferred insect transmissibility to the non-insect-transmissible strain 44 reinforced the idea that plasmid-encoded determinants were essential for insect transmission of S. citri (4).
Previously, we characterized the replication and stability regions of the S. citri GII3 plasmids and showed that each plasmid has its own partition system involving the Soj protein (8). In the present study, we describe plasmid incompatibility in S. citri GII3 and its use in displacing natural plasmids by their mutated/deleted derivatives. Experimental transmission of the produced plasmid mutants proved one pSci6 coding sequence (CDS), pSci6_06, to be essential for transmission of S. citri by its leafhopper vector.
S. citri GII3 was originally isolated from Circulifer haematoceps individuals captured in Morocco (40). Through injection into the leafhopper vector, it can be experimentally transmitted to periwinkle (Catharanthus roseus) plants, in which it multiplies and induces severe symptoms (15). S. citri GII3 contains seven plasmids (pSciA and pSci1 to -6), which have been described elsewhere (34). S. citri 44 was isolated from a stubborn-diseased sweet orange tree in Iran (19). In contrast to GII3, S. citri 44 contains no plasmids (5). Because attempts to infect periwinkle plants through injection to the leafhopper vector repeatedly failed (4, 22), S. citri 44 was described as a non-insect-transmissible strain. S. citri Alcanar was isolated from the columella of a sweet orange, harvested in Spain. This strain contains a plasmid highly similar to pSci6 from S. citri GII3 and was shown to be insect transmissible (S. Duret, S. Richard, and J. Renaudin, unpublished).
Spiroplasmas were grown at 32°C in SP4 medium (43), from which fresh yeast extract was omitted. Electrotransformation of spiroplasmas was carried out as previously described (37), using 1 to 5 μg of purified plasmid or various ligation mixtures. Spiroplasma transformants were first selected by plating them on solid SP4 medium containing 2 μg/ml tetracycline and were further propagated in the presence of 5 μg tetracycline per ml of broth medium.
Plasmids pSci6PT, pSci21NT, pFL4, and pFL5 have been described previously (4, 8). Plasmid pSci6B resulted from self-ligation of the 18,487-bp BstBI restriction fragment of pSci6PT (Fig. (Fig.1).1). The plasmid was generated through transformation of S. citri 44 with a ligation mixture of the BstBI fragments. Spiroplasma transformants carrying pSci6B were identified by PCR amplification with primers BstBIF and BstBIR (see Table S1 in the supplemental material) located on both sides of the BstBI recognition sequence of pSci6 and Southern blot hybridization with probes M, E, and TetM. These probes were generated by PCR amplification with the PMF/PMR, PEF1/PER1, and Tet1/Tet2 primer pairs, respectively (see Table S1 in the supplemental material). Similarly, plasmids pSci6N and pSci6BN were constructed by deleting the 8,104-bp NciI fragments of pSci6PT and pSci6B, respectively (Fig. (Fig.1).1). Plasmid pSci6BA was obtained from pSci6B by further deleting the 5.8-kbp AdhI fragment. Additional deletion of traG through digestion with XbaI yielded pSci6BAX (Fig. (Fig.1).1). To construct pSci6PTΔP32, pSci6PT was first linearized by digestion with PciI, made blunt ended by a fill-in reaction, and then recircularized by ligation before being introduced into S. citri 44 by transformation. In this construct, disruption of the p32 gene by insertion of 4 nucleotides resulted in a truncated translation product of 153 amino acids, compared to 238 for the wild-type protein.
Spiroplasma genomic DNA was prepared from 10-ml cultures by using a Wizard genomic DNA purification kit (Promega), whereas plasmid DNA was purified from 25-ml cultures with a Wizard SV miniprep DNA purification kit (Promega). Southern blot hybridization of spiroplasmal DNA with appropriate [digoxigenin]dUTP-labeled probes has been described elsewhere (1). Probes S235 and P32, specific to pSci1 to -5 and pSci6, respectively, as well as probe TetM, have been described previously (4). Probes P4 and P6, specific to pSci4 and pSci6, respectively, were produced by PCR amplification of the plasmids with the S4F/S4R and PstIF/PstIR primer pairs, respectively (see Table S1 in the supplemental material).
Transmission of S. citri to periwinkle plants through injection to the leafhopper vector C. haematoceps was adapted from a previously described method (15). In brief, the leafhoppers were microinjected with 0.2 μl of a spiroplasma culture (approximately 105 spiroplasma cells), and the injected insects were caged on stock (Matthiola incana) plants. After a 2-week latency period, the infected insects were transferred to young periwinkle plants (with 5 insects per plant and 5 to 15 plants per spiroplasma strain, depending on the experiment) for a 10-day inoculation access period. After the leafhoppers were removed, plants were kept at 30°C in the greenhouse and checked for symptom production over a period of 8 weeks. Under these conditions, S. citri wild-type strain GII3 repeatedly induced symptoms in all plants within 2 weeks after insect removal. Spiroplasma isolation from infected insects and plants was carried out as described previously (12, 14). Acquisition of spiroplasmas through feeding of the leafhoppers on spiroplasma-infected periwinkles has been described elsewhere (4). Whereas the acquisition assay determines the ability of the spiroplasmas to cross the gut epithelium and reach the hemolymph, in which they multiply, the transmission assay through injection of spiroplasma cultures directly into the insect hemolymph shortcuts the gut epithelium barrier and therefore reflects the ability of the spiroplasmas to invade the salivary glands and reach the saliva duct.
Statistical significance was determined using the Fisher exact probability test, in which significance is indicated by a P value of <0.05. The data were obtained from the combination of the two sets of plants (in transmission assays) or insects (in acquisition assays).
Partial recovery of transmissibility by the non-insect-transmissible strain S. citri 44 when transformed by pSci6 suggested that this plasmid encodes determinants of insect transmission (4). Furthermore, the pSci6-encoded protein P32 had been tentatively associated with insect transmission (22). In an attempt to identify these determinants, S. citri 44 was transformed with a derivative of pSci6, in which the candidate gene p32 was disrupted. Experimental transmission of the pSci6ΔP32 transformant through injection to the leafhopper vector showed that it was transmitted less efficiently than S. citri GII3 but similarly to S. citri 44/pSci6PT carrying the wild-type plasmid (see Table S2 in the supplemental material). These results indicated that the pSci6-encoded P32 protein was not essential for transmission of S. citri 44/pSci6PT. It is known that, in addition to distinct plasmid contents, S. citri strains have distinct chromosome sizes (9) and hence different chromosomal coding capacities. Therefore, determining the role of plasmids from S. citri GII3 in the S. citri 44 chromosomal background could be misleading. To avoid the probable bias due to distinct genetic environments between strains GII3 and 44, the role of plasmids in insect transmission was further investigated by constructing plasmid mutants of S. citri GII3 through plasmid incompatibility and comparing their insect transmissibility to that of the wild-type strain.
Plasmid curing of S. citri GII3 by the conventional methods, which includes growth at a sublethal temperature and/or in the presence of curing agents, resulted in the isolation of spiroplasma subclones having lost one, two, and three (pSci1, -3, and -5) out of six plasmids (5). However, further propagation of one such clone (carrying the plasmids pSci2, -4, and -6) in the presence of novobiocin did not result in any additional plasmid loss. No plasmid-free spiroplasma was obtained (see Fig. S1 in the supplemental material). Therefore, we explored the use of incompatibility (30) for plasmid curing/replacement in S. citri. Following transformation of S. citri GII3 with pSci21NT, pSci4NT, and pSci6PT, the tetracycline-selectable derivatives of pSci2, pSci4, and pSci6, respectively, tetracycline-resistant transformants were subcultured for up to 3 passages (approximately 10 generations) and examined for the presence of natural plasmids pSci1 to -6 by PCR and Southern blot hybridization (Fig. (Fig.2;2; see also Table S3 in the supplemental material).
When transformed by pSci21NT, a pSci2 derivative containing the replication and stability regions, most of the transformants carried pSci1 (9 out of 10 transformants) and pSci3 to -6 (10/10 transformants) but lacked pSci2 (9/10 transformants) (Fig. (Fig.2a).2a). Similarly, all 16 pSci4NT transformants specifically lacked pSci4 (see Table S3 in the supplemental material). In the case of pSci6PT also, nearly all transformants carried pSci1 to -5 but all of them lacked pSci6 (Fig. (Fig.2c).2c). Hybridization profiles of DNAs from pSci21NT and pSci6PT transformants having lost pSci2 and pSci6 are presented in Fig. 2b and d, respectively. These data indicate that, when introduced into S. citri GII3 by transformation, the marked pSci derivatives displace the corresponding wild-type plasmids, specifically. In the case of pSci6B transformants, the complete loss of wild-type pSci6 was further confirmed by the failure to amplify the 461-bp fragment of pSci6 with the PstIF/PstIR primer pair. In contrast, the 2.8-kbp fragment of pSci6B comprising the tetM gene was amplified (Fig. (Fig.2e,2e, lanes 1 to 6).
Interestingly, transformation of S. citri GII3 with pSci2 deletion derivatives pFL4 and pFL5, both of which possess the replication region but lack the stability region (8), also led to specific loss of wild-type pSci2, indicating that incompatibility is driven by the replication region. While most of the transformants contained pSci1 and pSci3 to -6, only 1 pFL4 transformant out of the 38 tested still carried pSci2 (Fig. (Fig.2a;2a; see also Table S3 in the supplemental material). Nevertheless, it is noteworthy that among spiroplasmal transformants that lost pSci2, a few lacked one, two, or more additional plasmids, as illustrated in Fig. Fig.3.3. For example, pFL4 transformants lacking two (pSci2 and pSci5 in Fig. Fig.3a,3a, lane 2) and up to four (pSci2, -3, -5, and -6 in lane 1 and pSci1 to -4 in lane 4) plasmids were isolated. Likewise, some pFL5 transformants lacked pSci1 and pSci2 (Fig. (Fig.3a,3a, lane 10), pSci2 and pSci3 (lane 8), pSci2 and pSci5 (lane 7), and pSci1, -2, and -5 (lane 11). For one of the pFL5 transformants, only pSci6 was detected (lane 9), indicating that pSci1 to -5 were lost. Figure Figure3b3b shows that, in addition to pSci2, which was by far the more frequently lost (97.4%), the other five plasmids were occasionally lost. Plasmids pSci1 and pSci5 were more frequently missing (in 31.6 and 26.3% of the transformants, respectively), whereas loss of pSci6 was rarely observed (in 5.3% of the transformants). In summary, these results showed that transformation of S. citri GII3 by a given selectable pSci plasmid resulted in the specific displacement of the related native plasmid and, in a few cases, the fortuitous loss of one or several additional plasmids. As a result, transformation of S. citri GII3 by the pSci2 derivatives yielded a large set of spiroplasmal mutants differing in their plasmid contents. We have previously established that, due to defective partitioning, pSci2 derivatives lacking soj are lost during propagation of the spiroplasmal transformants in the absence of selection (8). Indeed, further propagation of pFL4 or pFL5 transformants in nonselective medium resulted in the loss of these tetM-marked plasmids, yielding spiroplasmas with native plasmids, exclusively. S. citri strains G/1.4, with pSci1 and pSci4, and G/6, with only pSci6 (Fig. (Fig.3a,3a, lanes 1 and 9), were obtained this way.
In previous studies, we have reported the apparent correlation between the presence of plasmids pSci1 to -6 and the ability of S. citri strains to be experimentally transmitted through injection to its leafhopper vector (5). In particular, pSci6 was shown to confer insect transmissibility to the non-insect-transmissible strain 44 (4). To further assess the role of pSci6 in transmission, we compared the abilities of the two S. citri GII3 mutants G/145 (lacking pSci6) and G/6 (having pSci6 only) to be transmitted through injection to the leafhopper vector C. haematoceps. In the experiment summarized in Table Table1,1, successful transmission was revealed by symptom production in the periwinkle plants on which the infected leafhoppers fed. The results showed that G/145 was not transmissible, as none of the 30 plants developed symptoms, and that no spiroplasma was detected in these plants. In contrast, symptomatic plants were obtained (11 out of 17 plants) in the case of G/6, suggesting that pSci6 or at least pSci6-encoded determinants were required for transmission of S. citri GII3 by the leafhopper C. haematoceps.
To delineate the pSci6 region comprising the genetic determinants required for insect transmission, we took advantage of plasmid incompatibility to replace the wild-type pSci6 plasmid with its selectable deletion derivatives pSci6B, pSci6N, and pSci6BN (Fig. (Fig.1).1). These plasmids were introduced into both S. citri GII3 and the S. citri mutant G/6 by transformation. Spiroplasmal transformants were selected on tetracycline plates and further grown in the presence of the antibiotic for 5 to 10 passages before their plasmid contents were characterized by PCR and Southern blot hybridization (data not shown). The whole set of PCR amplifications used to confirm the plasmid content of the various spiroplasmal transformants is presented in Table S4 in the supplemental material. For example, pSci6BN was characterized by successful amplification with the BstBF/BstBR primer pair and failed amplification with the P32F/P32R and TraGF/TraGR primer pairs.
Experimental transmission assays through injection to the leafhopper vector revealed that S. citri GII3/6B carrying pSci6B was efficiently transmitted by the leafhopper, as a large majority of plants (14/15 and 6/8 plants) showed symptoms (Table (Table2).2). In contrast, no symptomatic plants (0/15 and 0/8 plants) were obtained in the cases of GII3/6N, indicating that these spiroplasmas were not transmitted to plants, in spite of their ability to multiply in the leafhopper. Similar data were obtained with S. citri G/6 transformants. Nine plants out of 12 were infected with G/6B, but none of the 15 plants tested with G/6N or G/6BN either showed symptoms or contained spiroplasmas.
Similar data were obtained with S. citri 44 (see Table S5 in the supplemental material). Whereas S. citri transformants 44/6 and 44/6B were transmitted (though at low efficiency), those carrying pSci6N (44/6N) were not. In strain S. citri Alcanar also, replacement of pSci6Alc by pSci6BN resulted in loss of insect transmissibility (see Table S5 in the supplemental material).
These results suggested that genetic determinants required for insect transmission were located within the 10.4-kbp NciI fragment of pSci6 (Fig. (Fig.1).1). This region comprises 12 CDSs, including two truncated copies of CDS E (proved to be the replication protein) (8), two copies of the CDS cluster K traG M, CDS N, two hypothetical CDSs, and a truncated copy of CDS F. In pSci6BA, the 10.4-kbp NciI fragment was further reduced to a 4.6-kbp region, in which the traG and M CDSs were the only CDSs specifically encoded by pSci6 (Fig. (Fig.1).1). Interestingly, strain G/6BA harboring this plasmid was consistently transmitted. Depending on the experiment, 3 to 6 of the 10 tested plants were infected (Table (Table2).2). In contrast, transmission assays of strain G/6BAX carrying pSci6BAX, in which traG sequences had been deleted (Fig. (Fig.1),1), repeatedly failed. None of the 20 periwinkle plants tested showed symptoms, indicating that pSci6 traG sequences were required for transmission of S. citri through injection to its leafhopper vector C. haematoceps.
In this experiment, healthy leafhoppers were fed on periwinkle plants infected with various S. citri strains. After a 2-week feeding period, the insects were transferred to stock plants for a 2-week latency period before the number of infected leafhoppers was determined (Table (Table3).3). The results showed that S. citri strains GII3, GII3/6B, and G/6 were efficiently acquired (with 83 to 86, 60, and 45 to 46% insects infected, respectively), suggesting that pSci1- to -5-encoded proteins, in particular the adhesion-related ScARPs, are not essential for the acquisition of the spiroplasma by the leafhopper vector C. haematoceps. Moreover, the finding that strains 44/6 and Alcanar, both of which carry pSci6, were acquired at very low efficiencies (with fewer than 2% of insects infected, compared to 45% for S. citri G/6) suggests that chromosomal (rather than plasmid) determinants might be required for the spiroplasmas to efficiently cross the gut epithelium barrier.
Although bacterial plasmids are not considered to be essential, they usually encode genetic determinants that sustain survival in a given environment and promote adaptation to the host (38, 41). The widespread occurrence of pSci plasmids among S. citri strains (8) is consistent with the fact that these plasmids play a key role in the biological cycle of the spiroplasma. By constructing mutant strains through plasmid curing/replacement, we showed that pSci6 sequences are associated with insect transmission. S. citri GII3 contains low-copy-number plasmids pSci1 to -6, having a Soj/ParA-dependent, active partition system (8, 34). As a result, the plasmids are stably maintained for many generations. In spite of their high homology, all six plasmids, pSci1 to -6, coexist within the same cell. Similar coexistence between highly related replicons has also been described to occur in Pseudomonas syringae strains, some of which carry up to six pT23A-like plasmids (18). In this case, the suppression of incompatibility was thought to result from the variability of the repA sequences and the poor conservation of putative maintenance determinants.
In our study, we showed that transformation of S. citri GII3 by selectable pSci2, pSci4, or pSci6 derivatives resulted in the nearly specific loss of their native counterparts. The finding that the deletion derivative pFL5, like pSci21NT, was most incompatible to pSci2 is consistent with the fact that the replication region plays a key role in plasmid incompatibility (30). However, in a few pFL4/pFL5 transformants, plasmids other than pSci2 were also missing. In these transformants, pSci1, -3, and -5 were the most frequently lost, similarly to the situation in which S. citri GII3 was propagated at the 37°C sublethal temperature (5), suggesting that the loss of these additional plasmids would be independent of incompatibility. However, the interference of pFL4/pFL5 with the replication of plasmids other than pSci2 cannot be fully excluded. Indeed, constructing pFL4/pFL5 through deletions of pSci21NT might have resulted in loss of regulatory sequences and subsequent deregulation of replication determinants. This could explain the reduced specificity of the incompatibility between pFL4/pFL5 and pSci2. Yet, the inconsistency of the plasmid contents from the various pFL4/pFL5 transformants is not understood.
We took advantage of plasmid incompatibility to construct S. citri GII3 mutants differing in their plasmid contents, with the aim of identifying plasmid-encoded determinants involved in insect transmission. We have previously shown that the presence or absence of the soj gene determines the segregational stability of pSci plasmids and that the soj-free plasmids are rapidly lost in the absence of selection pressure (8). Therefore, depending on the stability or instability of the transforming plasmid used (i.e., the presence or absence of the soj gene), replacement or loss of the native counterpart could be achieved through further propagation of the spiroplasmal transformant in the absence of selection pressure. By using this plasmid curing/replacement strategy based on incompatibility, we obtained S. citri strains having one single plasmid (pSci6) or various deleted derivatives. Experimental transmission assays proved that pSci6, more precisely, the 0.9-kbp traG fragment, encodes determinants that are critical for insect transmission of S. citri GII3 by its natural vector C. haematoceps. These results are consistent with our previous studies showing that pSci6 from S. citri GII3 conferred insect transmissibility to the non-insect-transmissible strain 44 (4) and with the fact that strain Alcanar, having a pSci6-related plasmid, also was transmitted. In turn, the finding that strains GII3/6B and 44/6B lacking the p32 gene were readily transmitted indicates that the hydrophilic protein P32 is not a marker of insect transmissibility as previously suggested (22). Indeed, the findings that P32 was dispensable for both transmission through injection and acquisition through insect feeding clearly show that P32 was not required for the spiroplasmas to cross the gut epithelium and salivary gland membrane barriers. The eight ScARPs from S. citri GII3 are encoded by pSci1 to -5 and share strong homology with the surface protein P89, or SARP1, of S. citri BR3 (3). This protein was thought to be involved in the adherence of spiroplasmas to leafhopper (C. tenellus) cells in vitro, as the reduced adherence of protease-treated spiroplasmas was associated with a decrease in the amount of protein detected (44). Likewise, we have shown recently that the non-insect-transmissible strain 44, which lacks pSci1 to -6 and hence does not possess ScARPs, is impaired in its ability to invade the C. haematoceps cell line Ciha-1 (11). Taken together, these data suggested the involvement of ScARPs in the interactions of the spiroplasma with the insect midgut epithelial or salivary gland cells in vivo during the transmission process. In the present work, however, successful acquisition and transmission of ScARP-less S. citri GII3 mutants (such as strain G/6) by C. haematoceps clearly proved the pSci1- to -5-encoded ScARPs to be nonessential in these processes. Nevertheless, the finding that S. citri G/6 was acquired and transmitted at lower efficiencies than the wild-type strain GII3 (P < 0.001; Fisher's exact probability test) suggests that pSci1- to -5-encoded proteins (possibly the ScARPs), in spite of being nonessential, might function as aggregative partners, facilitating the interactions between spiroplasmas and leafhopper cells to complete the infection cycle.
In contrast to pSci1 to -5, we found that pSci6, more precisely, CDS pSci6_06, annotated traG (34), was required for insect transmission of S. citri GII3. This CDS encodes a 287-amino-acid protein sharing limited homology with the C-terminal part of ATPases of the TraG/VirD4 family involved in type IV secretion systems (32, 35). In particular, a domain of the P-loop nucleoside triphosphatase (NTPase) superfamily including the Walker B motif was predicted. However, unlike most TraG proteins, including those two of 600 and 721 amino acids encoded by the S. citri GII3 chromosome, the pSci6_06-encoded protein lacks a signal peptide, the N-terminal transmembrane segments, and the Walker A motif (i.e., one of two functional domains of the TraG proteins). Therefore, it is doubtful that the loss of insect transmissibility associated with deletion of the pSci6_06 protein results from the inactivation of a type IV secretion system. In addition, despite the presence of genes encoding proteins with strong similarities to components of type IV secretory pathways (2, 21, 34), evidence for functional type IV secretion systems in spiroplasmas is still lacking.
In insect-transmitted microbes, a large majority of proteins involved in the interactions with insect cells are membrane proteins. In Plasmodium falciparum, for instance, all of the proteins involved in invasion of the salivary glands are membrane proteins (17). In the plant-pathogenic mollicute “Candidatus Phytoplasma asteris,” insect vector specificity is determined by the interaction between the cell surface membrane protein Amp and the insect microfilament complex (39). In S. citri also, membrane-associated proteins were shown to interact with insect microfilaments, as suggested by colocalization of spiroplasmas with the actin filaments of the leafhopper salivary glands (25). In these studies, protein overlay assays revealed that two membrane-associated proteins, a lipoprotein and the glycolytic enzyme phosphoglycerate kinase (PGK), interacted with actin in vitro. By use of the Ciha-1 leafhopper cell culture system (11), it was shown that PGK, by specifically binding to actin, was involved in internalization of spiroplasmas into insect cells (25). In contrast to these membrane-associated proteins, the pSci6_06 protein was predicted to be cytosolic, making a direct interaction with leafhopper cells unlikely. It is noteworthy that transmission of the Lyme disease bacterium Borrelia burgdorferi by the tick Ixodes scapularis involves not only chromosome-encoded surface proteins such as OspA/B (29) but also cytosolic, plasmid-encoded proteins such as IMP dehydrogenase and GMP synthase (20). By providing a growth advantage to spirochetes in the ticks, these proteins are critical for B. burgdorferi mouse infectivity. In the case of S. citri, in silico analyses of the pSci6_06 protein did not predict any function, despite the presence of a putative nucleotide binding domain. Interestingly, whereas in vitro spiroplasma cultures reached similar titers regardless of their plasmid contents, the mutant strains rarely reached the titers of the wild-type strain GII3 in the leafhopper, suggesting that plasmid determinants could provide a growth advantage to spiroplasmas in the insect vector. However, given the high variability of spiroplasma titers in the leafhoppers (as determined under our experimental conditions), no clear-cut difference was noticed between S. citri strains having and those lacking the pSci6_06 CDS. Whether and how this pSci6_06-encoded protein might contribute to the spiroplasma fitness in the leafhopper vector is unknown. However, from the findings that S. citri G6/BAX (devoid of this protein) multiplied in the insect hemolymph but failed to cross the salivary gland barrier, one may hypothesize that this protein is involved in binding to the salivary gland membrane or in the survival of the spiroplasmas within the endocytosis vesicles within the proposed pathway through the salivary glands to the salivary duct (13).
This work was funded by INRA and the Université Victor Segalen Bordeaux 2. Support for M. Breton was provided by the Ministère de l'Enseignement Supérieur et de la Recherche.
We are grateful to L. Béven for performing statistical analyses. We also thank K. Guionneaud and D. Lacaze for growing plants and insects.
Published ahead of print on 19 March 2010.
†Supplemental material for this article may be found at http://aem.asm.org/.