Characterization of Spiroplasma citri adhesion related protein SARP1, which contains a domain of a novel family designated sarpin

Gene 275 (2001) 57–64 www.elsevier.com/locate/gene

Characterization of Spiroplasma citri adhesion related protein SARP1, which contains a domain of a novel family designated sarpin Michael Berg a, Ulrich Melcher b, Jacqueline Fletcher a,* a

b

Department of Entomology and Plant Pathology, Oklahoma State University, Stillwater, OK 74078, USA Department of Biochemistry and Molecular Biology, Oklahoma State University, Stillwater, OK 74078, USA Received 30 March 2001; received in revised form 13 July 2001; accepted 1 August 2001 Received by A.M. Campbell

Abstract Transmission of the plant pathogen Spiroplasma citri by its leafhopper vector, Circulifer tenellus, involves adherence to and invasion of insect host cells. The S. citri adhesion related protein P89 (SARP1) was purified by immunoprecipitation using anti-SARP1 monoclonal antibodies. The protein’s N-terminal amino acid sequence was determined and used to design a degenerate oligonucleotide. The labeled oligonucleotide hybridized to a 3.5 kb MboI fragment from S. citri DNA, which was then cloned and sequenced. Additionally, a 1.9 kb RsaI fragment of S. citri DNA, partially overlapping the MboI fragment, was isolated and characterized. Sequence analysis of the two clones revealed four open reading frames. ORF1 (675 bp) encodes the C-terminal part of a Soj-like protein. ORFs 1 and 2 were separated from ORFs 3 and 4 by a putative transcription termination site, indicated by a hairpin structure. ORF3 encodes an amphiphilic 798 amino acid long protein with a cleavable signal peptide and a predicted transmembrane helix near the C-terminus. The mature protein of 85.96 kDa has a calculated pI value of 5.5 and has an N-terminal amino acid sequence consistent with that determined from the purified SARP1. At the Nterminus of this protein is a region consisting of six repeats, each 39–42 amino acids, a motif belonging to a previously unrecognized family of repeats found in a variety of bacterial proteins. The taxonomically spotty presence of this ‘sarpin’ domain and the relationship of the repeats to each other suggests a convergent evolution in multiple lineages. q 2001 Elsevier Science B.V. All rights reserved. Keywords: Membrane protein; Mycoplasma; Mollicutes; arp1; Convergent evolution

1. Introduction Spiroplasma citri, a helical, wall-less prokaryote of the class Mollicutes, is the causative agent of lethal stunting diseases in a number of host plants including citrus and several brassicaceous species. Like other phytopathogenic mollicutes (spiroplasmas and phytoplasmas), it is transmitted in nature by phloem-feeding insects in a circulative manner (Purcell, 1982). Spiroplasmas are, in contrast to phytoplasmas, culturable in vitro. Therefore, the relationship between S. citri and its major US vector, the beet leafhopper Circulifer tenellus, has been studied in detail and has become a model for mollicute–vector interactions. Within Abbreviations: bp, base pairs; Ig, immunoglobulin; kb, kilo base pairs; kDa, kilodalton; mAb, monoclonal antibodies; ORF, open reading frame; PAGE, polyacrylamide gel electrophoresis; PVDF, polyvinyl pyrrolidone; SARP1, spiroplasma adhesion related protein; SDS, sodium dodecyl sulfate * Corresponding author. Department of Entomology and Plant Pathology, 127 Noble Research Center, OK 74078, USA. Tel.: 11-405-744-9948; fax: 11-405-744-7373. E-mail address: [email protected] (J. Fletcher).

the insect, spiroplasmas traverse both the intestinal epithelium and salivary glands to complete the transmission cycle. The presence of spiroplasmas within membrane-bound cytoplasmic vesicles of midgut epithelium and salivary gland cells suggests that the mollicutes cross these physical barriers in the insect vector, entering host cells via receptormediated endocytosis and leaving them by exocytosis (Kwon et al., 1999). A model describing the movement of spiroplasmas through these barriers within the leafhopper vector was proposed by Fletcher et al. (1998). However, the molecular and biochemical interactions contributing to the traversal of these barriers are not well understood. It is likely that adherence to the host cell is the initial step in the mollicute–vector interaction. For several mollicutes infecting humans and animals it is well established that adherence to host cell membranes is mediated by mollicute surface proteins called adhesins (Razin and Jacobs, 1992; Baseman et al., 1996; Krause, 1996). Invasion of non-phagocytic host cells by some mollicutes (Andreev et al., 1995; Winner et al., 2000) and their ability to survive intracellularly may be a

0378-1119/01/$ - see front matter q 2001 Elsevier Science B.V. All rights reserved. PII: S 0378-111 9(01)00655-2

58

M. Berg et al. / Gene 275 (2001) 57–64

factor in the ability of the microbes to cross the mucosal barrier, thereby gaining access to the bloodstream. Spiroplasmas are reported to cytadhere to insect host and non-host cells, both within the intact insect and in tissue culture (Steiner et al., 1982; Liu et al., 1983). Yu et al. (2000) found that loss and restoration of the ability of S. citri to adhere to a monolayer of cultured C. tenellus cells is clearly associated with degradation and restoration of a specific spiroplasma membrane protein, P89. Their results suggest that P89 (designated SARP1, for spiroplasma adhesion related protein) is directly involved in spiroplasma– insect cell interaction. In this work we describe the isolation of SARP1 and the detection, cloning, and analysis of the encoding gene. 2. Materials and methods 2.1. Spiroplasmas S. citri strain BR3 was originally isolated from diseased horseradish plants (Fletcher et al., 1981). The line BR3-T, derived from the triply cloned original isolate (BR3-3X) by repeated transmission from turnip to turnip via its insect vector C. tenellus, is still transmissible (Wayadande and Fletcher, 1995). S. citri was cultured in LD8 broth medium (Chen and Davis, 1979) at 288C. 2.2. Escherichia coli and plasmids E. coli strain XL1-blue MRF 0 and plasmid pBluescript SK (Stratagene, La Jolla, CA) were used for the cloning of DNA fragments. Bacteria were grown in Luria–Bertani (LB) media and plasmid DNA was extracted by alkaline lysis following standard procedures (Sambrook et al., 1989). 2.3. Immunoprecipitation Cells from a 200 ml culture of S. citri (BR3-T) were collected by centrifugation (20,000 £ g, 15 min), rinsed twice with PBS and resuspended in 30 ml PBS containing 1% Triton X-114. The suspension was incubated on ice for 1 h and subjected to centrifugation (20,000 £ g, 15 min). The supernatant was incubated for 1 h at 378C, enabling micelle aggregation and phase separation (Bordier, 1981). After centrifugation (5000 £ g, 5 min) at room temperature the aqueous upper phase was discarded. The Triton X-phase was mixed with 1 vol. cold acetone and incubated overnight at 2208C for protein precipitation. After centrifugation (20,000 £ g, 15 min), the pellet was desiccated and resuspended in 0.5 ml PBS containing 1% Triton X-114. Nondissolved material was removed by centrifugation (12,000 £ g, 10 min). About 50 mg mAb (1-2E8) IgG (mAb were developed against P89 recovered after SDSPAGE with the assistance of the Hybridoma Center for Agricultural and Biological Sciences of Oklahoma State University) (Berg and Fletcher, unpublished data) was added to the

supernatant. After incubation for 2 h on ice, the suspension was mixed with 100 ml protein A-Sepharosee 6MB (Amersham Pharmacia Biotech, Piscataway, NJ). The mixture was incubated for 16–20 h on ice with gentle agitation, and the supernatant was removed after allowing the Sepharose to settle out. The Sepharose protein-complex was washed five times with PBS. After removing the PBS, 50 ml SDS sample buffer (Sambrook et al., 1989) was added and the mixture was incubated for 1 h at 568C and then for 5 min at 958C. The Sepharose was removed by centrifugation (10,000 £ g, 5 min) and the supernatant was collected. 2.4. Microsequencing of SARP1 Supernatants from several immunoprecipitation preparations were combined and concentrated by filtration (Ultrafree w-15 centrifuge filter device, Sigma, St. Louis, MO). The concentrate was subjected to SDS-PAGE (Laemmli, 1970). To avoid N-blocking of the desired protein, the time for gel polymerization was extended (16–20 h) and 0.1 mM Na-thioglycolate was added to the running buffers. Following electrophoresis, the proteins were electroblotted for 1 h at 100 V to PVDF membrane (BioRad, Hercules, CA) using blotting buffer (25 mM Tris, 192 mM glycine, 20% methanol (pH 8.3)). Resolved bands were visualized by staining with Coomassie G-250, and a section of the PVDF membrane containing P89 was excised, transferred into the reaction chamber of an automated gas phase sequencer (491 protein sequencer, Applied Biosystems, Foster City, CA), and subjected to Edman sequence analysis according to the manufacturer’s protocols. 2.5. DNA isolation and manipulation S. citri cells from a 50 ml culture were collected by centrifugation (15 min at 20,000 £ g) and resuspended in DNA extraction buffer (1.4 M NaCl, 68.6 mM hexadecyltrimethylammonium bromide (CTAB), 20 mM EDTA, 100 mM Tris–HCl (pH 8.0)). After incubation at 608C for 30 min, the suspension was extracted with phenol/chloroform and chloroform, successively. DNA was precipitated with isopropanol, washed with 70% (v/v) ethanol and dissolved in TE (Tris–EDTA) buffer. All routine DNA manipulation techniques including restriction endonuclease analysis, ligation and end-labeling of oligonucleotides with [g- 32P]dATP were performed as described by Sambrook et al. (1989). 2.6. Southern blot analysis Southern blot analysis was performed as described by Sambrook et al. (1989) with the following modifications. Prehybridization and hybridization were performed at 388C in Church buffer (0.5 M Na-phosphate (pH 7.2), 1 mM EDTA, 0.24 M SDS) containing 150 mg/ml salmon sperm DNA. Subsequent washing was done four times for 20 min each at 388C with 40 mM sodium phosphate buffer (pH 7.2) containing 1% (w/v) SDS and 1 mM EDTA.

M. Berg et al. / Gene 275 (2001) 57–64

2.7. Cloning and sequencing of the SARP1 gene Fragments of the appropriate size from digested S. citri DNA were recovered after agarose gel electrophoresis using a QIAquick gel extraction kit (Qiagen Inc., Valencia, CA). The DNA fragments were ligated into plasmid vector pBluescript following standard procedures (Sambrook et al., 1989) and E. coli cells were transformed by electroporation following the manufacturer’s instructions (BioRad). Recombinant clones were selected and grown in broth medium. Plasmid DNA was extracted by alkaline lysis according to Sambrook et al. (1989). Cloned DNA was screened by dot blot hybridization. The extracted plasmid DNA was heated at 1008C for 10 min in 0.4 M NaOH and 10 mM EDTA and 2 ml of each sample was spotted manually to a positively charged nylon membrane. The membrane was subjected to hybridization following the procedure described above. Cloned DNA fragments reacting positively in dot blots were fully sequenced using dye-terminated thermal cycle sequencing and an Applied Biosystems/PerkinElmer 373 sequencer (Perkin-Elmer Inc., Wellesley, MA). The GenBank Accession number for the sequence reported in this work is AJ297706. 2.8. Sequence analysis Prediction of putative ORFs comprised by the DNA sequence was done using ‘ORF Finder’ (http:// www.ncbi.nlm.nih.gov/gorf/gorf.html), and protein sequence similarity searches were performed using ‘PSI-

59

BLAST’ and ‘BLASTp’ (http://www.ncbi.nlm.nih.gov/ BLAST/). Definitional, algorithmic and statistical refinements of ‘PSI-BLAST’ and ‘BLASTp’ are described by Altschul et al. (1997). The program ‘SignalP V1.1’ (Nielsen et al., 1997) was used to identify the signal sequence and the ‘Compute pI/Mw tool’ (ExPASy) was used for the calculation of the molecular mass and the isoelectric point (pI). The transmembrane region was predicted with ‘TMpred’ (Hofmann and Stoffel, 1993). CLUSTAL (http://www.ebi.ac.uk/clustalw/) was used for multiple sequence alignments. 3. Results 3.1. Isolation and microsequencing of SARP1 SARP1 was isolated and purified by immunoprecipitation using mAb 1-2E8 and SDS-PAGE (Fig. 1A). Sequence analysis of the N-terminal region of SARP1 revealed 17 amino acids (AVPLTNTLTSN[S/C]NP[D/A]Y[L/D]) (at three positions where the results were ambiguous, the two most predominant amino acids are given in brackets). Considering the codon usage of S. citri, a degenerate oligonucleotide (5 0 -GCW GTW CCW TTA ACW AAT ACW TTA ACW TCA AAT TCW AAT CC-3 0 ) was designed for the detection of the encoding gene. 3.2. Detection and cloning of the SARP1 gene (arp1) Using the oligonucleotide as a probe in Southern hybridization, DNA fragments of distinct size were detected in S.

Fig. 1. (A) SDS-PAGE silver stained; isolation of P89 (SARP1) using mAbs (1-2E8) and protein A-Sepharose. Lane 1, marker (95 and 55 kDa); lane 2, Triton X-114 fraction of S. citri (BR3-T) proteins; lane 3, proteins after immunoprecipitation (P89 and heavy chain of mAb IgG). (B) Southern blot hybridization of S. citri DNA digested with restriction enzymes EcoRI, HindIII, MboI, and XbaI. A labeled oligonucleotide derived from the amino acid sequence of the Nterminus of P89 was used as a probe.

60

M. Berg et al. / Gene 275 (2001) 57–64

Fig. 2. (A) Schematic representation of putative ORFs comprising the two cloned S. citri DNA fragments pP89A and pP89B. arp1 represents the gene encoding the spiroplasma adhesion related protein (SARP1). (B) Domain organization of S. citri adhesion related protein (SARP1). SP, signal peptide; TM, transmembrane segment; rep1–6, motif repeats.

citri DNA digested separately with restriction enzymes EcoRI, HindIII, MboI, and XbaI (Fig. 1B). While the detected EcoRI, HindIII, and XbaI fragments were larger than 10 kb, the two MboI fragments detected were about 3.5 and 4.5 kb, respectively. The 3.5 kb fragment was cloned in E. coli using a plasmid vector. Sequence analysis revealed that the cloned MboI fragment contained only the 5 0 -end of the ORF encoding SARP1. To obtain the remainder of arp1, S. citri DNA, digested with restriction enzyme RsaI, was probed by Southern hybridization with a labeled oligonucleotide (5 0 -TTT AAC ATC AAC CGA ACC C-3 0 ) derived from the sequencing results of the MboI fragment. Southern hybridization revealed reactions with three distinct fragments, which were 1.9, 2.5, and 3.8 kb in size. The 1.9 kb RsaI fragment was cloned and sequenced and found to contain 1949 bp. The 5 0 -most 1235 bp overlapped the 3 0 -end of the cloned MboI fragment.

region between ORF2 and ORF3, a termination site of transcription was indicated by a hairpin structure. ORF3 (2397 bp), which encodes SARP1, was preceded by a Shine– Dalgarno sequence, 5 0 -AGAAAGGA-3 0 , 9 bp from the initiation codon ATG, and stopped at the translation termination codon TAA. ORF4 started 2 bp from the stop of arp1. The ORF4 putative ribosomal binding site (5 0 GGAGGTGA-3 0 ) is located within ORF3 and 9 bp from the ATG initiation codon. Database search BLASTp did not reveal any significant matches between the peptide derived from ORF4 and known proteins. A signal sequence with a cleavage site between positions 24 and 25 was predicted by the program SignalP 1.1 (Nielsen et al., 1997). Analysis with program TMpred revealed a putative transmembrane helix consisting of 21 amino acids starting at position 25 of the mature protein and suggesting an Nterminus located outside of the cell (Hofmann and Stoffel, 1993).

3.3. DNA sequence analysis 3.4. SARP1 gene analysis Analyses of the cloned DNA fragments revealed the presence of two complete and two partial ORFs (Fig. 2). The partial ORF1 (675 bp) encodes the C-terminal part of a protein homologous to Soj, which is involved in chromosome partitioning in Bacillus subtilis (Sullivan and Maddock, 2000). The hypothetical protein encoded by ORF2 (519 bp) was not significantly similar to any other sequence in the databases (GenBank CDS translations, PDB, SwissProt, PIR, PRF, Trembl, and TremblNew). A putative ribosomal binding site (5 0 -AAGGAG-3 0 ) 14 bp from the initiating ATG of ORF2 overlapped with the termination codon TAA of ORF1. Within the 415 bp intergenic

ORF3, representing arp1, encodes 798 amino acids. The predicted mature protein, omitting a 23 amino acid long Nterminal signal peptide that was suggested by the SignalP V1.1 analysis program (Nielsen et al., 1997), has an Nterminal amino acid sequence (AVPLTNTLTSNSNNDYL) consistent with that determined from the purified P89. Of the 17 amino acids determined by peptide sequencing only one did not agree with the sequence deduced from arp1. In each of the three cases where the sequencing results were ambiguous, one of the two most likely amino acids agreed with the sequence encoded by ORF3. The deduced mature

Fig. 3. Multiple sequence alignment of repeat units of the sarpin domain. Prediction of the accessibility of residues (3-state from PHD; see Rost and Sander, 1994) is shown in the top line (b, buried; B, highly likely to be buried; e, exposed). The next line shows the PHD predicted secondary structure. Reliability values associated with the prediction are in the following line. Polypeptides are identified as follows: Sci, Spiroplasma citri CAC10363; Sku, S. kunkelii SkARP1; Myx, Myxococcus xanthus BAB40338; Syn, Synechocystis sp. S75099; Pae, Pseudomonas aeruginosa AAG03857; Pgi, Porphyromonas gingivalis AAD51078; VchT, transcription regulator Vibrio cholerae AAF95226; VchG, GGDEF-containing protein V. cholerae AAF94511; Tth, Thermus thermophilus CAA71198; Xfa, Xylella fastidiosa AAF84139; Zmo, Zymomonas mobilis AAD19713. Numbers following these designations reflect the repeat order in the primary amino acid sequence.

M. Berg et al. / Gene 275 (2001) 57–64

61

62

M. Berg et al. / Gene 275 (2001) 57–64

protein is predicted to be an amphiphilic membrane protein with a pI value of 5.5 and a molecular weight of 85.96 kDa. Anchoring of the polypeptide within the plasma membrane is predicted to be accomplished by a transmembrane helix (residues 713–733) near the C-terminus with the hydrophilic N-terminus located outside of the cell. The amphiphilicity of SARP1 is confirmed by the protein’s behavior in Triton X-114 phase partitioning experiments. A protein dotplot revealed six tandem repeats consisting of 39–42 amino acids each (Figs. 2 and 3), starting ten amino acids beyond the N-terminus of the mature protein. Domains similar to the SARP1 tandem repeats were found to occur in 11 proteins of nine taxonomically diverse bacterial genera. A PSI-BLAST search initiated with residues 24–277 of S. citri SARP1 retrieved seven polypeptides (E ¼ 3 £ 10235 to 5 £ 10 291 on iteration three). One (sLL0241) was the only member of the ProDom PD134053 family. Searches of other domain and motif databases yielded no matches to SARP1. In BLAST searches using each of seven sequences as queries, each was recognized by at least three retrieved sequences. These searches revealed additional sequences, AAG03857 and CAA71198. An eleventh member of this domain family, SkARP1, was found by a TBLASTN search of the developing genome database of Spiroplasma kunkelii (http://www.genome.ou.edu/spiro.html), a plant pathogenic spiroplasma closely related to S. citri. SkARP1 is similar throughout its length to SARP1. The degree of divergence between the two is not detectably different from that for other pairs of genes from the two species for which such comparisons can be made. The 11 sequences contained five (AAD19713), six (CAA71198, CAC10363, AAD51078) or seven (AAG03857, S75099, SkARP1) repeats or about double those numbers (13 in AAF94511 and BAB40338, 14 in AAF95226 and 17 in AAF84139), approximately 40 residues each. Preliminary alignment of all repeats revealed that repeats within a polypeptide were more similar than they were to those in other polypeptides. Therefore, they were aligned with one another before that group of repeats was aligned with those from other polypeptides (Fig. 3). The repeats were predicted (Rost and Sander, 1994) to form bsheets (Fig. 3) with buried side chains. Repeated sheet structures occur in propellers of PQQ domains (Ghosh et al., 1995) and indeed, PQQ domains were retrieved in the PSI-BLAST search (E . 0:002). This domain, which we designate ‘sarpin’, occurs in a variety of contexts. While some sarpin domains are attached to additional sequences, several are not, strengthening the view that sarpins are structural domains. SARP1, SkARP1 and AAD51078 likely are surface-exposed proteins. AAF94511 has a C-terminal GGDEF domain (Pei and Grishin, 2001), while AAF95226 may be a transcriptional regulator. Functions for the others are unknown. Sarpin domains were detected only in eubacteria. Five diverse groups were represented: the CFB (Cytophaga-Flexibacter-Bacteroides) and Thermus/Deinococcus groups,

proteobacteria, firmicutes, and cyanobacteria. Four other mollicutes, Mycoplasma genitalium, Mycoplasma pneumoniae, Mycoplasma pulmonis and Ureaplasma urealyticum, whose complete genome sequences are available, do not encode sequences resembling the sarpin domain. Analysis of the results of a BLASTp search of available databases (GenBank CDS translations, PDB, SwissProt, PIR, PRF) revealed an additional periodicity in the remaining protein sequence. The highest detected similarity was to the STARP antigen of Plasmodium falciparum. This antigen consists of multiple tandem repeats of an approximately ten residue motif characterized by a thr·asp dipeptide followed by multiple asn residues. Examination of the BLAST alignment of the STARP sequence with the SARP1 sequence revealed that a significant contribution to the good expected value came from the coincidence of SARP1 asp residues with the asp of the STARP thr·asp dipeptide. A secondary contributor was the high content (10%) of asn in the section of SARP1 C-terminal to the six tandem repeats coupled with the high asn content of the STARP antigen. This section (excluding the putative transmembrane helix, but including the C-terminal tail) could be divided into 41 segments of an average length of 12 (mean, median and mode: range 9–17) of which 73% began with an asp residue. Ala, thr, and asn accounted for most of the exceptions.

4. Discussion The interactions between plant pathogenic mollicutes (phytoplasmas and spiroplasmas) and their insect vectors are remarkably specific but not well characterized at the molecular level. Therefore, the objectives of this research were to isolate the S. citri adhesion related protein P89 (SARP1) (Yu et al., 2000) and to identify, clone and characterize the encoding gene. The cloned DNA fragments comprise not only the ORF encoding SARP1 but also the adjacent regions. ORF1 encodes the C-terminus of a Soj-like protein. In B. subtilis, Soj protein is involved in chromosome partitioning and negatively regulates expression of several sporulation genes by binding to the promoter regions and inhibiting transcription (Sullivan and Maddock, 2000). The function of the predicted peptide encoded by ORF2 could not be clearly identified by a database search. The relatively large spacing between ORF2 and arp1, and a hairpin structure following ORF2, indicate that the SARP1 gene belongs to a separate transcription unit. A functional relationship between ORF1/ORF2 and the P89 gene is therefore not likely. In many other organisms genes immediately adjacent to and downstream of soj (parA) encode proteins that are likewise involved in replication and/or cell division. We did not find an indication that the protein encoded by ORF2 has any relationship to those proteins. However, frequent rearrangements in the S. citri chromosome, partly due to integrated

M. Berg et al. / Gene 275 (2001) 57–64

sequences of spiroplasma viruses and leading to a remarkable genome instability, were shown earlier (Ye et al., 1996; Melcher et al., 1999). The gene encoding SARP1 (arp1) was completely cloned and characterized. A transmembrane segment near the Cterminus, which presumably functions as a membrane anchor, suggested an amphiphilic nature for SARP1. This was consistent with its solubility in Triton X-114 buffer, commonly used for the enrichment of integral membrane proteins. Previous results showing the accessibility of P89 to proteolytic digestion without cell disruption (Yu et al., 2000) and the presence of an N-terminal signal sequence also support the conclusion that SARP1 is a membranebound surface protein. That the predicted signal sequence functions as a signal peptide was confirmed by the finding that the amino acid sequence, revealed from microsequencing of the SARP1 N-terminus, agreed with the predicted Nterminus of the protein after removal of the first 23 amino acids. Signal peptides are commonly involved in protein targeting and translocation across membranes (Driessen, 1994) in eukaryotes as well as in prokaryotes. The slight discrepancy between the results from N-terminal sequencing and the amino acid residues predicted from the DNA sequence is probably due to the low concentration of protein deployed and the difficulties in distinguishing the amino acids from background signals. Similarities of the SARP1 repeats with repeat motifs of other bacteria were obvious, although the functions of most of them remain unclear. The taxonomically spotty presence of sarpins could be explained by horizontal transfer of sarpin genes, selective retention of sarpins in only few species, or repeated de novo evolution of the domain. The latter is supported by relationships among repeats. SARP1 and SkARP1 repeats, both from spiroplasmas, were significantly less diverged from one another than they were from the repeats in other proteins or than the latter repeats were from each other. This observation suggests that these sarpin domains arose in a common ancestor of the two spiroplasmas by iterated duplication of an ancestral 40 amino acid unit. Since the sarpin domain was not found in other mollicutes, the SARP1/SkARP1 domains may have evolved recently. That they differ in the number of repeats suggests that such duplication occurred readily in evolution. The view of convergent evolution in multiple lineages is also supported by repeats within a polypeptide being more similar to one another than they are to repeats in most other polypeptides. The low similarities among polypeptides may explain why this family was recognized only after the discovery of SARP1 with its less diverged repeat sequences. Interactions between adhesins and host cell receptors are well studied in human and animal pathogenic mollicutes of the genus Mycoplasma. Host cell membrane receptors responsible for mycoplasma attachment, identified so far, are mostly sialoglycoconjugates and sulfated glycolipids (Razin, 1985; Razin and Jacobs, 1992; Zhang et al.,

63

1994). Detailed information on the molecular properties of characterized mycoplasma adhesin proteins and their encoding genes is summarized in several recent reviews (Baseman et al., 1996; Krause, 1996). Although mycoplasma cytadherence and surface colonization have been studied extensively for years, it was only after the recent discovery of certain human mycoplasma species within eukaryotic cells (Lo et al., 1989; Winner et al., 2000) that the intracellular presence and host cell invasion of mycoplasmas began to be investigated. Despite the notably different host systems of vertebrates versus insects, the molecular mechanisms leading to adherence and internalization of mollicutes might be similar. Therefore, the spiroplasma–insect relationship may provide a model for how other mollicutes invade host cells and cross cellular barriers.

Acknowledgements We thank our colleagues from the Recombinant DNA/ Protein Resource Facility and the Hybridoma Center for Agricultural and Biological Sciences of Oklahoma State University for their excellent service in DNA synthesis, sequencing and antibody production. We also thank Dr Astri C. Wayadande, Dr John R. Sauer, and Dr Margaret K. Essenberg for reviewing the manuscript. This research was supported by grants from the United States Department of Agriculture and the Oklahoma Agricultural Experiment Station.

References Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller, W., Lipman, D.J., 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25 (17), 3389–3402. Andreev, J., Borovsky, Z., Rosenshine, I., Rottem, S., 1995. Invasion of HeLa cells by Mycoplasma penetrans and the induction of tyrosine phosphorylation of a 145-kDa host cell protein. FEMS Microbiol. Lett. 132, 189–194. Baseman, J.B., Reddy, S.P., Dallo, S.F., 1996. Interplay between mycoplasma surface proteins, airway cells, and the protean manifestations of mycoplasma-mediated human infections. Am. J. Respir. Crit. Care Med. 154, 137–144. Bordier, C., 1981. Phase separation of integral membrane proteins in Triton X-114 solution. J. Biol. Chem. 256, 1604–1607. Chen, T.A., Davis, R.E., 1979. Cultivation of spiroplasmas. In: Whitcomb, R.F., Tully, J.G. (Eds.), The Mycoplasmas. Academic Press, New York, pp. 65–79. Driessen, A.J.M., 1994. How proteins cross the bacterial cytoplasmic membrane. J. Membrane Biol. 142, 145–159. Fletcher, J., Schultz, G.A., Davis, R.E., Eastman, C.E., Goodman, R.E., 1981. Brittle root disease of horseradish: evidence for an etiological role of Spiroplasma citri. Phytopathology 71, 1073–1080. Fletcher, J., Wayadande, A., Melcher, U., Ye, F., 1998. The phytopathogenic mollicute-insect vector interface: a closer look. Phytopathology 88, 1351–1358. Ghosh, M., Anthony, C., Harlos, K., Goodwin, M.G., Blake, C., 1995. The refined structure of the quinoprotein methanol dehydrogenase from ˚ . Structure 3, 177–187. Methylobacterium extorquens at 1.94 A

64

M. Berg et al. / Gene 275 (2001) 57–64

Hofmann, K., Stoffel, W., 1993. TMbase – a database of membrane spanning protein segments. Biol. Chem. Hoppe-Seyler 347, 166. Krause, D.C., 1996. Mycoplasma pneumoniae cytadherence: unravelling the tie that binds. Mol. Microbiol. 20, 247–253. Kwon, M.-O., Wayadande, A.C., Fletcher, J., 1999. Spiroplasma citri movement into the intestines and salivary glands of its leafhopper vector, Circulifer tenellus. Phytopathology 89, 1144–1151. Laemmli, U.K., 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 136, 180–184. Liu, H.-Y., Gumpf, D.J., Oldfield, G.N., Calavan, E.C., 1983. The relationship of Spiroplasma citri and Circulifer tenellus. Phytopathology 73, 585–590. Lo, S.C., Dawson, M.S., Wong, D.M., Newton, P.B.D., Sonoda, M.A., Engler, W.F., Wang, R.Y., Shih, J.W., Alter, H.J., Wear, D.J., 1989. Identification of Mycoplasma incognitus infection in patients with AIDS: an immunohistochemical, in situ hybridization and ultrastructural study. Am. J. Trop. Med. Hyg. 41, 601–616. Melcher, U., Sha, Y., Ye, F., Fletcher, J., 1999. Mechanisms of spiroplasma genome variation associated with SpV1-like viral DNA inferred from sequence comparisons. Microb. Comp. Genomics 4, 29–46. Nielsen, H., Engelbrecht, J., Brunak, S., von Heijne, G., 1997. Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites. Protein Eng. 10, 1–6. Pei, J., Grishin, N.V., 2001. GGDEF domain is homologous to adenylyl cyclase. Proteins 42, 210–216. Purcell, A.H., 1982. Insect vector relationship with procaryotic plant pathogens. Annu. Rev. Phytopathol. 20, 397–417. Razin, S., 1985. Mycoplasma adherence. In: Razin, S., Barile, M.F. (Eds.), The Mycoplasmas. Academic Press, Orlando, FL, pp. 161–202.

Razin, S., Jacobs, E., 1992. Mycoplasma adhesion. J. Gen. Microbiol. 138, 407–422. Rost, B., Sander, C., 1994. Combining evolutionary information and neural networks to predict protein secondary structure. Proteins 19, 55–72. Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning: A Laboratory Manual, 2nd Edition. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Steiner, T., McGarrity, G.J., Phillips, D.M., 1982. Cultivation and partial characterization of spiroplasmas in cell cultures. Infect. Immun. 35, 296–304. Sullivan, S.M., Maddock, J.R., 2000. Bacterial sporulation: pole-to-pole protein oscillation. Curr. Biol. 10, 159–161. Wayadande, A.C., Fletcher, J., 1995. Transmission of Spiroplasma citri lines and their ability to cross gut and salivary gland barriers within the leafhopper vector Circulifer tenellus. Phytopathology 85, 1256– 1259. Winner, F., Rosengarten, R., Citti, C., 2000. In vitro cell invasion of Mycoplasma gallisepticum. Infect. Immun. 68, 4238–4244. Ye, F., Melcher, U., Rascoe, J.E., Fletcher, J., 1996. Extensive chromosome aberrations in Spiroplasma citri strain BR3. Biochem. Genet. 34, 269– 286. Yu, J., Wayadande, A.C., Fletcher, J., 2000. Spiroplasma citri surface protein P89 implicated in adhesion to cells of the vector Circulifer tenellus. Phytopathology 90, 716–722. Zhang, Q., Young, T.F., Ross, R.F., 1994. Glycolipid receptors for attachment of Mycoplasma hyopneumoniae to porcine respiratory ciliated cells. Infect. Immun. 62, 4367–4373.