Mycoplasma agalactiae, the causative agent of contagious agalactia in small ruminants, produces a protein, named P80, that is detectable in all wild-type isolates examined to date and that appears expressed during the early phase of infection. We describe here the identification, cloning and expression of the gene encoding P80 (ma-mp81). The deduced amino acid sequence is consistent with a hydrophobic and basic protein that possesses a lipoprotein signal peptide. Sequence analysis of gene ma-mp81 suggests that P80 is a membrane lipoprotein that shows significant homology with other putative lipoproteins of M. pneumoniae. An internal 1-kb fragment of ma-mp81 was expressed in Escherichia coli as a 6×His-tagged protein. The purified recombinant protein greatly reacted with polyclonal anti-P80 sera raised in lamb.
Mycoplasma agalactiae is the causative agent of contagious agalactia (CA) in small ruminants . CA is particularly widespread around the Mediterranean basin and it causes significant financial losses due to reduced milk production and increased lamb mortality. In Sardinia (Italy), the disease was unknown until 1980 when it was introduced by sheep from the mainland. Then, a very rapid spread of CA occurred, possibly due to the combination of several factors: the large number of animals (3.5 million milking sheep, corresponding to one third of the total Italian stock), high animal density (149/km2 against 23/km2 on the mainland), breeding conditions (primitive herding practices and transhumance) and, finally, inadequate diagnosis and prophylaxis . The first outbreaks of CA in Sardinia were mainly characterised by acute and subacute clinical signs, such as agalactia, mastitis, keratoconjunctivitis and arthritis. More recently, however, less severe clinical forms were observed (i.e., without keratitis and arthritis), as is typical of an endemic situation. Control and eradication of CA can be achieved through an improvement of management practices, better diagnostic tests and by developing more effective vaccines. Serological [3,4] and genetic [5–7] assays have been developed to detect M. agalactiae in clinical samples.
In order to develop a protective vaccine, it is crucial to understand the pathogenesis and immunology of M. agalactiae infection. This involves the definition of the microorganism's antigenic structure during the natural infectious process. Recently, we have identified several integral membrane proteins recognised by sera collected from naturally infected sheep. These proteins are integral membrane proteins with molecular masses spanning between 18 kDa and 80 kDa . These observations have prompted us to evaluate different bacterial inactivating agents and to produce a protective saponin-M. agalactiae vaccine . The efficient protection obtained with inactivated bacteria suggests that the administration of purified immunogenic proteins may also stimulate a specific and protective immunity.
Recently, several M. agalactiae genes have been characterised: the P48 gene and vsp-related genes [10–12]. P48 has homology to Mycoplasma fermentans MALP-404 lipoprotein and MALP-2 lipopeptide, which has a macrophage-stimulatory activity , whereas variable surface proteins (Vsp) represent a set of immunodominant lipoproteins undergoing high-frequency phase and size variation. We have focused on an integral membrane protein of 80 kDa (P80), since this protein was detected in all M. agalactiae wild-type strains isolated from naturally infected sheep examined to date. Furthermore, P80 was recognised by animal sera collected during the early phase of infection .
In this study, the P80 was purified and its amino-terminal sequence served to identify and clone the ma-mp81 gene. Finally, we have expressed and purified an internal fragment of P80 as a 6×His-tagged fusion protein (MA-MP81-B).
Materials and methods
Bacterial strains and growth conditions
Type strain PG2 (Bga) and six strains of M. agalactiae isolated from milk samples of different flocks with CA (Bus1, Ras1, 233 M, 129 M, 137 M and 141 M) were used. These strains were grown at 37°C in modified Hayflick medium containing 10% equine serum . Bacteria were harvested from exponential-phase broth cultures by centrifugation at 20 000×g for 30 min and washed twice with phosphate-buffered saline (0.1 M phosphate, 0.33 M NaCl, pH 7.4). The final pellets were used for both DNA extraction and protein analysis.
Purification and concentration of P80
P80 was sliced from the SDS–PAGE gel after electrophoresis of total proteins of M. agalactiae. The protein was then electroeluted using the Model 422 electro-eluter (Bio-Rad) with running buffer (rb, 25 mM Tris, 192 mM glycine, 0.1% SDS, pH 8.3) for 3 h at 10 mA per glass tube and with rb without SDS for another hour. Protein was concentrated with Microcon 30 (Amicon, Bedford, MA, USA) following the manufacturer's instructions.
Amino-terminal sequence of P80
Following electroelution, P80 was subjected to electrophoresis and transferred to a PVDF membrane in CAPS buffer (10 mM cyclohexylaminopropane sulfonic acid, 10% (v/v) methanol, pH 11). P80 was stained with Coomassie blue R250 and sliced from the membrane. N-terminal sequencing was performed by Ceinge (Biotecnologie Avanzate, Naples, Italy).
DNA was extracted from type strain PG2 according to the method described by Ausubel et al. . Hybridisation using 5′-fluorescein 45-mer oligonucleotides was performed as previously described . DNA fragments were extracted from agarose gel with CONCERT gel extraction systems (Life Technologies). DNA cloning was performed with cloning vectors pZero (Invitrogen) or pQE30 (Qiagen, Chatsworth, CA, USA), and transforming into Escherichia coli TOP10F′ cells or DH5α (pREP4), respectively. Cloning into pQE30 makes it possible to produce recombinant proteins linked to a polyhistidine (6×His) stretch that binds strongly to nickel-chelated columns (Pierce, Rockford, IL, USA). pZero derivatives were selected in low-salt–LB agar containing isopropyl-1-thio-β-d-galactoside (IPTG) and zeocin at 50 μg ml−1 final concentration. pQE30 derivatives were selected in LB agar plates supplemented with 100 μg ml−1 ampicillin and 50 μg ml−1 kanamycin. Plasmid DNA was extracted with Qiagen Plasmid Maxi Kit (Qiagen).
Production of polyclonal antibodies against P80
A 45-day-old sarda lamb was purchased from a farm with no clinical history of CA. The lamb was treated once with levamisole chlorohydrate/nyclosamide polyvalent anthelminth (Atenas: Fatro) and once with ivermectin against ecto/endoparasites (Ivomec: Msd AGVet). Absence of previous contacts with M. agalactiae was assessed by immunoblotting analysis of animal sera. The lamb was immunised s.c. with 200 μl of concentrated P80 protein emulsified 1:1 with Freund's complete adjuvant. Three weeks later, two more s.c. injections of P80 with an equal volume of Freund's incomplete adjuvant were given at approximately 8-day intervals. After the last i.m. injection of 200 μl of P80 protein, the animal was blooded and serum aliquoted and stored at −20°C. Test sera were collected from naturally infected sheep belonging to different areas of Sardinia.
Purification of M. agalactiae P80 and identification of gene ma-mp81
In order to purify P80, whole cell lysates of six M. agalactiae wild-type strains were pooled and subjected to SDS–PAGE. The band corresponding to P80  was electroeluted from the Coomassie-stained gels, and concentrated (Fig. 1A). The purified protein was utilised to deduce the N-terminal sequence. Based on the amino acid sequence obtained (AKCVDKDYEELGKDT) a degenerate 45-mer oligonucleotide was synthesised and labelled with fluorescein (5′-(F)-45P). Total PG2 M. agalactiae DNA was digested with HindIII, EcoRV and SpeI endonucleases and subjected to Southern hybridisation with 5′-(F)-45P. After hybridisation, only HindIII-digested DNA presented a single positive band at approximately 5500 bp of size (data not shown). The 5′-(F)-45P-positive HindIII fragment was extracted from the agarose gel and ligated into HindIII-digested pZero vector, obtaining pGS98. The nucleotide sequence of the 5.5-kb HindIII genomic fragment was determined. Sequence analysis showed an open reading frame (ORF) of 2166 bp in size encoding a protein of 721 amino acids (Fig. 2) with a calculated molecular mass of 81 007.3 Da. The gene, named ma-mp81 (M. agalactiae membrane protein 81), predicted a very basic protein (pI 9.12) with a high concentration of lysine (13.5%) and with an instability index computed to be 24.37. This classifies the protein as stable (http://www.ExPASy.ch/tools/ptotpartef.html) [17–19]. The start codon is preceded by a typical ribosome binding sequence (AGGA), and potential −10 (TATAAT) and −35 (TATTCA) promoter regions. Finally, a DNA sequence consistent with the P80 amino-terminus was located 69 nucleotides from the ATG start codon. The first 23 amino acids constitute a typical prokaryotic lipoprotein signal peptide. The cleavage region, deduced from a mature protein, contains the sequence AVS↓AKC instead of VSA↓KC as predicted with http://www.cbs.dtu.dk/services/SignalP or the method of Nielsen et al.  for lipoproteins of other known prokaryotes. The single Cys residue located at the N-terminus of the mature protein is a likely site for lipid modification and membrane anchorage [21,22].
A search through the nucleic acid and protein databases using the BLAST search tool  revealed that the predicted amino acid sequence of P80 is related to lipoprotein A65-ORF794 (GenBank accession number S73877) , to lipoprotein GT9-ORF798 (S62791)  and to lipoprotein PO2-ORF793 (S73622)  of M. pneumoniae. The predicted sequence of P80 is about 25% identical to that of the hypothetic lipoproteins of M. pneumoniae. The nucleotide sequence of the ma-mp81 gene of M. agalactiae reported here has been deposited in GenBank under accession number AF299294.
Expression of ma-mp81 gene in E. coli
Analysis of the ma-mp81 sequence marked the presence of four universal stop codons (TGA) that encode tryptophan in Mycoplasma. These codons, two near the N-terminus and two near the C-terminus (see Fig. 2), disable the expression of a full-length protein in E. coli. In order to overexpress and purify a large region of MA-MP81 in E. coli, pairs of primers were designed to amplify the ma-mp81 gene internal sequences with no TGA codons. A set of primers (forward, 5′-GGG GTA CCC CTG CAA TAA CTT CAT-3′ [ma-mp81-A/F]; reverse, 5′-CCC AAG CTT GGG TCA TTT ACC TGG-3′ [ma-mp81/R]) were synthesised to amplify and to clone a 1491-bp fragment. A shorter fragment of 1014 bp was also amplified by replacing forward primer ma-mp81-A/F with forward primer 5′-GGG GTA CCC CGA TGG AAA ATG G-3′ (ma-mp81-B/F) (Fig. 3). The 1491-bp and 1014-bp ORFs encoded respectively proteins of 497 and 338 amino acids with estimated molecular masses of approximately 55 kDa and 37 kDa. The amplified products were cloned in the pQE-30 vector and introduced into competent DH5αE. coli cells containing pREP-4 repressor plasmid (lacIq). MA-MP81-A and MA-MP81-B clones, containing the 1491-bp and the 1014-bp fragment respectively, were grown and expression of the respective 6×His fusion proteins were then induced with 2 mM IPTG. Immunological analysis with the anti-P80 lamb serum and with the sera collected from naturally infected sheep showed P80 fragments of the expected sizes in both MA-MP81-A and MA-MP81-B bacterial cells lysates (Fig. 4). The production of recombinant protein from clone MA-MP81-B was very high and increased during the time course analysed (30, 60, 90 min), whereas MA-MP81-A clone had low-level expression and the recombinant product further decreased from 60 min to 90 min after induction.
Characterisation of antigens expressed during the course of an M. agalactiae infection would be helpful in elucidating the pathogenesis of the disease. Identification of immunogenic antigens would also be useful to develop tools for the rapid diagnosis of CA and more effective vaccines.
Previous studies from our laboratory, based on electrophoretic and serological analyses of several wild-type strains from different areas of Sardinia, showed that a surface membrane protein with an apparent molecular mass of 80 kDa was expressed by all strains examined, and induced antibody production in the early phase of infection . In this paper, we described the amino-terminal sequence of P80, the cloning and characterisation of the ma-mp81 gene, and the production of recombinant proteins linked to a polyhistidine domain.
Sequence analysis of the ma-mp1 gene product provided evidence that P80 is indeed a lipoprotein. A posttranslational cleavage site for signal peptidase II, deduced from the amino-terminal sequence, was located between residues Ser-23 and Ala-25 (i.e., AVS↓AKC) instead of Ala-25 and Lys-26 (i.e., VSA↓KC) used by other Gram-positive and Gram-negative bacteria [20,22]. This is in agreement with the consensus sequence found in other M. agalactiae lipoproteins: P48 and Vpma [10,11]. Both lipoproteins are subjected to a cleavage that occurs two amino acids before the first Cys residue: P48 (VA↓ASC) and Vpma (A↓AKC). Therefore, all three mature lipoproteins started with an Ala residue. These data confirm the hypothesis that the peptidase II of M. agalactiae possesses a different proteolytic processing compared to that found in other prokaryotic signal peptidases II.
Although recombinant DNA technology has made possible the cloning and expression of mycoplasma DNA in E. coli, expression of a whole protein is generally unfeasible. In fact, most Mycoplasma utilise the universal stop codon TGA to encode tryptophan [26,27]. Since we obtained the entire sequence of the ma-mp81 gene, gene expression in E. coli could be realised by three different approaches: (i) gene expression in an E. coli opal suppressor strain, (ii) site-directed mutagenesis to convert TGA into TGG or (iii) cloning regions of the ma-mp81 gene that lack TGA codon(s). The first strategy was tested with E. coli strain YN2980 (kindly provided by C. Minion, Iowa State University, Ames, IA, USA), but P80 expression was not efficient under this condition (data not shown). Expression of a large fragment of P80 (clone MA-MP81-A) turned out to be somewhat toxic for the cells. This led us to assume that the expression of the entire ma-mp81 gene would have caused the arrest of the growth of E. coli. Therefore, site-directed mutagenesis of the four TGA codons was not performed. Instead, the MA-MP81-B clone, expressing a smaller P80 fragment (37 kDa), yielded high amounts of recombinant protein that strongly reacted with the anti-P80 lamb serum.
Further studies are required to better understand the biological function of the M. agalactiae P80 lipoprotein, whether it possesses a role in pathogenesis, and its potential utilisation to develop vaccines against the agent of CA.
We are grateful to Francesca Masia for technical assistance. This research was funded by Grant 12104/01 from the Sardinia Region.