- Email: [email protected]
High-level expression of the Listeria monocytogenes listeriolysin O in Escherichia coli and preliminary characterization of the purified protein
Protein Expression and Purification 28 (2003) 78–85 www.elsevier.com/locate/yprep
High-level expression of the Listeria monocytogenes listeriolysin O in Escherichia coli and preliminary characterization of the purified proteinq Camilla Giammarini,a Francesca Andreoni,a Giulia Amagliani,a Annarita Casiere,b Simone Barocci,b and Mauro Magnania,c,* a Centro di Biotecnologie, University of Urbino, Via T. Campanella 1, 61032 Fano, Italy Istituto Zooprofilattico Sperimentale Umbria e Marche, Via G. Salvemini 1, 06126 Perugia, Italy Istituto di Chimica Biologica ‘‘G. Fornaini,’’ University of Urbino, Via Saffi 2, 61029 Urbino, Italy
b c
Received 31 July 2002, and in revised form 6 November 2002
Abstract Listeriolysin O (LLO) is a cholesterol-binding sulfhydryl-activated hemolysin encoded by Listeria monocytogenes hlyA gene. After analyzing the nucleotide coding sequence of this gene from the ATCC 9525 L. monocytogenes strain, we cloned it in a pET vector for expression in Escherichia coli. Thanks to the optimization of the induction protocol, we achieved a high-level LLO synthesis (about 10% of total cell proteins) in hemolytically active form. The expressed hemolysin was then purified to homogeneity, as revealed by SDS–PAGE and Western blot analysis, by a hydroxyapatite adsorption chromatography, followed by an SP Sepharose ion-exchange chromatography. The recombinant protein showed the same properties determined for LLO purified from L. monocytogenes cultures and the characteristics of the sulfhydryl-activated toxins such as inactivation by oxidation and by reaction with cholesterol. By a combination of the pET expression system and the simple purification method, we obtained a significant amount of toxin (4.5 mg/litre cell culture) in a hemolytically active form (1:25 106 HU/mg protein). This procedure can solve the problem of LLO isolation from L. monocytogenes cultures, which is a difficult task, mainly owing to the low levels of toxin released in the culture media. The recombinant hemolysin, purified in sufficient quantities, could be very useful for structural studies and for diagnostic and pharmaceutical applications. Ó 2002 Elsevier Science (USA). All rights reserved.
Listeria monocytogenes is a Gram-positive, foodborne, human and animal pathogen. This bacterium causes an infection, named listeriosis, which especially affects immunocompromised patients, new-borns, and pregnant women and is characterized by a variety of severe syndromes, such as encephalitis, meningoencephalitis, septicemia, and abortion [1]. Owing to its ability to survive in various environmental conditions, such as low pH, high NaCl concentrations, and very low temperatures, L. monocytogenes is present both in raw and processed foods; moreover, being a facultative anq
The nucleotide sequence data reported in this paper appear in DDBJ, EMBL, and GenBank nucleotide sequence databases under Accession No. AF253320. * Corresponding author. Fax: +390721862834. E-mail address: [email protected] (M. Magnani).
aerobe its growth is not significantly affected by vacuum packaging. Following the ingestion of contaminated food, bacteria cross the intestinal barrier, reach the liver and the spleen, and then disseminate to the brain and placenta. L. monocytogenes is a facultative intracellular micro-organism capable of internalizing into macrophages, as well as into cells devoid of phagocytic activity, by inducing its own phagocytosis. After phagocytosis, the pathogen escapes from the bacteriumcontaining vacuole by lysis of the phagosomal membrane and reaches the host cytosol in which it multiplies. The bacteria move inside the cytosol by polymerization of the host cell actin into a polarized comet tail and when they contact the plasma membrane, they induce the formation of bacteria-containing protrusions. Afterwards, these protrusions are internalized into the neighboring cells in which the micro-organisms start a new life cycle. Each
1046-5928/02/$ - see front matter Ó 2002 Elsevier Science (USA). All rights reserved. doi:10.1016/S1046-5928(02)00682-4
C. Giammarini et al. / Protein Expression and Purification 28 (2003) 78–85
step of L. monocytogenes intracellular parasitism depends upon the production of virulence factors among which listeriolysin O (LLO) is one of the most important [1,2]. LLO is an exotoxin, encoded by the hlyA gene, which belongs to the group of cholesterol-binding sulfhydryl-activated toxins, expressed by a large number of Gram-positive bacteria. All the hemolysins in this family have a single unique cysteine, which renders them susceptible to reversible inactivation by oxidation. However, this cysteine can be changed to an alanine without affecting their activity. Their lytic activity is enhanced by reducing agents and is suppressed by exposure to oxygen, cholesterol or anti-streptolysin O antibodies [3]. The molecular mechanism of membrane damage produced by this group of toxins is based on pore formation in membranes containing cholesterol. After binding to cholesterol, 20–80 toxin monomers oligomerize into arcand ring-shaped structures, which cause the passive flux of ions and macromolecules. In contrast to the other sulfhydryl-activated toxins, LLO has optimum activity at pH 5.5 and is almost entirely inactivated at pH 7.0. Owing to this property, the toxin activation occurs only after phagosomal acidification and causes the perforation of L. monocytogenes-containing vacuole and the bacterium diffusion into the cytosol [4]. The deleterious effects of LLO on cytoplasmic membrane are avoided, thanks to its low activity at cellular pH and its rapid degradation in the host-cell cytosol. Indeed, in its NH2 terminus there is a 19-amino acid PEST-like sequence that may target this toxin for degradation [5]. This motif seems to be also essential for bacterial virulence, since its deletion causes a defect in bacterium phagosomal escaping [6]. Another study describes a divergent LLO interaction with cholesterol in solution which inhibits hemolytic activity but does not alter membrane binding [7]. However, the unique LLO molecular mechanism of action is not still completely explained and protein studies are very difficult because the toxin isolation from L. monocytogenes cultures provides very low yields [8,9]. To overcome this problem, we expressed the hemolysin in Escherichia coli, using the pET expression system, obtaining significant amounts of recombinant LLO. The protein purification was then carried out adjusting a method previously described by Traub and Bauer [10] to our conditions. In this paper, we also compare the hemolytic activity and the properties of the recombinant LLO with those of the natural one purified from L. monocytogenes cultures by other authors [8,10].
Materials and methods Construction of the expression vector pET-3a + LLO The L. monocytogenes DNA was extracted by boiling method from the 9525 bacterial strain obtained from
79
American Type Culture Collection (ATCC, Manassas, VA). The region of hlyA gene, coding for the secreted form of LLO, was isolated by polymerase chain reaction (GeneAmp PCR Kit, Roche Molecular Systems, Branchburg, NJ). The amplification was performed using degenerate primers (LLO-3: 50 -CAACAAACTGA ACATATGGACGCATCTGCA-30 ; LLO-2: 50 -ATATT CTTTTACATATGTTAATTCTTC-30 ), which introduced NdeI restriction sites at nucleotides 1559 and 3216 (GenBank Accession No. M24199), respectively. The 1686-bp PCR product was then digested with the NdeI restriction enzyme (Roche Diagnostics, Indianapolis, IN) obtaining a 1657-bp fragment (coding for amino acids 26–529 of LLO). After gel-purification, it was cloned into the NdeI site of a pET-3a vector (Novagen, Madison, WI) to generate the construct referred to as pET-3a + LLO (Fig. 1). Nucleotide sequencing analysis The hlyA coding region of the ATCC 9525 L. monocytogenes strain was isolated by PCR with LLO-8 (50 -CA TCTTTAGAAGCGAATTT-30 ) and LLO-6 (50 -CA AGCTTATTTTTTCGTGTGTG-30 ) primers designed on the hlyA sequence (GenBank Accession No. M24199). Direct sequencing of PCR product was performed by the Sanger method [11] with the ABI PRISM BigDye Terminator Cycle Sequencing Kit (Applied Biosystems, Foster City, CA) following manufacturerÕs protocols, using 90 ng DNA and 3.2 pmol of one of the following sequencing primers: LL5 forward 50 -AACCTATCCAGGTGCTC-30 LL6 reverse 50 -CTGTAAGCCATTTCGTC-30 LLO-5 forward 50 -CGTATGGCCGTCAAGTTTA TTT-30 LLO-6 reverse 50 -CAAGCTTATTTTTTCGTGTG TG-30 The reaction mixtures were purified with Centri-Sep columns (Princeton Separations, Adelphia, NJ) and run on an ABI PRISM 310 Genetic Analyzer (Applied Biosystems). Expression of the recombinant listeriolysin O The recombinant plasmid pET-3a + LLO, obtained as described above, was transformed into E. coli strain BL21(DE3) (Novagen). The induction experiments were initially performed according to manufacturerÕs instructions. Briefly, cells were grown in LB broth plus ampicillin (at a concentration of 50 lg/ml) at 37 °C to early log phase (A600 nm ¼ 0:4). IPTG (Bio-Rad, Hercules, CA) was then added to a final concentration of 0.4 mM to induce the T7 RNA polymerase gene and cultures were incubated further at 37 °C for 4 h.
80
C. Giammarini et al. / Protein Expression and Purification 28 (2003) 78–85
Fig. 1. Construction of the plasmid for the expression of the L. monocytogenes toxin listeriolysin O. The steps performed to construct the pET3a + LLO expression vector are illustrated (see Materials and methods for details).
Purification of the recombinant listeriolysin O Cells derived from 1 litre of post-induction culture were harvested by centrifugation at 6000g for 30 min at 4 °C. The resulting pellet was resuspended in 50 ml lysis buffer at pH 6.1 containing 20 mM NaH2 PO4 , 1 mM dithiothreitol (DTT), 5% (v/v) glycerol, 0.5% (v/v) Triton X-100, and an anti-proteolytic cocktail (Sigma–Aldrich, St. Louis, MA). The cell suspension was sonicated five times for 30 s at 200 W and then centrifuged for 45 min at 25,000g at 4 °C. LLO present in the soluble fraction was then purified to homogeneity following a two-step purification procedure based on Traub and BauerÕs protocol [10]. The supernatant obtained from the previous step was applied to 40 ml hydroxyapatite Bio-Gel HTP Gel (Bio-Rad) equilibrated in 20 mM so-
dium phosphate buffer, pH 6.1. The column was washed with the equilibrating buffer and then LLO was desorbed with 0.3 M NaH2 PO4 , 1 mM DTT, and 5% (v/v) glycerol, pH 6.1. The collected fractions were tested for hemolytic activity on blood agar plates and analyzed by SDS–PAGE. Solid ammonium sulfate was slowly added to the pooled fractions, containing LLO from the previous step, to achieve an 85% saturation. The suspension was gently stirred for 30 min and then centrifuged at 10,000g for 30 min. The pellet was dissolved in 10 ml buffer A (50 mM NaH2 PO4 , 1 mM EDTA, 2.7 mM KCl, 1 mM DTT, and 5% (v/v) glycerol, pH 6.4) and then dialyzed twice (1 h each) against 2 litre of the same buffer. The dialyzed protein solution was applied to an SP Sepharose Fast Flow (Amersham Pharmacia Biotech, Uppsala, Sweden) column (2 8 cm) equilibrated
C. Giammarini et al. / Protein Expression and Purification 28 (2003) 78–85
in buffer A. The column was washed with the same buffer and desorption was carried out with a continuous salt gradient (0–1 M NaCl), using buffer B consisting of buffer A plus 1 M NaCl. Finally, the fractions containing LLO were concentrated with Amicon (8200) Stirred Ultrafiltration Cell (Bedford, MA). Hemolytic activity titration The hemolytic activity of a toxin is expressed as hemolytic units (HU) per millilitre of toxin solution or per milligram of total protein. The HU is the smallest amount (highest dilution) of toxin that liberates half the hemoglobin (50% lysis) from a suspension of erythrocytes. The hemolytic activity of purified LLO was determined preparing 1 ml toxin serial dilutions in PBS (pH 5.5) containing 0.1% (w/v) BSA (Sigma–Aldrich) and then adding 0.5 ml human red blood cells suspension. Erythrocyte suspension was used at 2.5% hematocrit, standardized so that 0.5 ml human red blood cells lysed by adding 14.5 ml of 0.1% sodium carbonate solution give an optical absorbance of 0.200 at 541 nm [12]. After 45 min at 37 °C the samples were centrifuged for 5 min at 1700g and the optical absorbance of the hemoglobin in the supernatant was measured at 541 nm. The data were estimated graphically by plotting percent lysis versus toxin dilution on a log probit graph. Determination of protein content Protein concentration was determined using the Bradford method [13] with bovine serum albumin as standard or spectrophotometrically by measuring the absorbance of solutions at 280 nm against appropriate blanks. SDS–PAGE SDS–PAGE of the expressed LLO was performed in 10% polyacrylamide gels according to the method of Laemmli [14] utilizing Bio-Rad equipment. The proteins were visualized by staining with Coomassie brilliant blue. The yield of the recombinant LLO in crude extract was determined approximately by comparison with known concentrations of molecular-weight protein standards, using the Gel Doc 2000 Apparatus (BioRad). The volume of the band corresponding to LLO was compared with the volume of the area from a gel lane containing the total protein, which was also quantified by the Bradford assay.
81
collected 3 weeks later and stored (50 ll aliquots) at )80 °C. Sheep immune serum, diluted 1:100, was used for immunoreactivity against LLO in the Western blot analysis. SDS–PAGE gels were blotted onto Hybond-C Extra nitrocellulose membrane (Amersham Pharmacia Biotech) by the Mini Trans-Blot Electrophoretic Transfer Cell (Bio-Rad). The second antibody was HRP-conjugated rabbit anti-sheep IgG (BIOTREND, K€ oln, Germany), diluted 1:5000, and the substrate was 2,20 -azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) (Sigma–Aldrich). Hemolytic activity at different pH The hemolytic activity of LLO was tested dissolving 10 HU in 1 ml PBS at different pH values: from 5.0 to 8.5. After the addition of RBC suspension and the incubation at 37 °C, the hemolytic activity titration was performed as described above. Hemolytic activity after incubation with cholesterol Cholesterol (Sigma–Aldrich) was dissolved in absolute ethanol at an initial concentration of 1 mg/ml and then serial dilutions were made in ethanol. Ten microliters of various concentrations of cholesterol was added to 1 ml samples containing 30 HU toxin in PBS, pH 6.0, supplemented with 0.1% (w/v) BSA. One milliliter of toxin with 10 ll ethanol was used as control. After 30 min of incubation at 22 °C, 0.5 ml erythrocyte suspension was added to the samples and the hemolytic activity was determined, as previously described. Hemolytic activity after incubation with SH-group reagents As SH-group reagents, we used HgCl2 (Sigma–Aldrich) dissolved in PBS, pH 6.0, at the final concentration of 10 mM and p-chloromercuribenzoate (Sigma–Aldrich) in PBS, pH 6.0, with 0.1 M NaOH at the concentration of 5 mM. Thirty hemolytic units of LLO in 1 ml PBS, pH 6.0, supplemented with 0.1% (w/v) BSA was incubated with 10 ll of various dilutions of the oxidants for 30 min at 22 °C. The reagent concentrations that inhibited the hemolytic activity of 1 HU of toxin were determined, as described above for the cholesterol assay.
Results hlyA gene analysis
Immunoblotting One Sardinian sheep was intravenously infected once with 7 108 CFU/ml of viable L. monocytogenes ATCC 9525 strain. Serum samples of the infected sheep were
The entire coding region of hlyA gene was sequenced (GenBank Accession No. AF253320) and aligned with the corresponding regions of different L. monocytogenes strains available on GenBank (M24199; AL591974;
82
C. Giammarini et al. / Protein Expression and Purification 28 (2003) 78–85
Fig. 2. Alignment of LLO amino acid sequence from L. monocytogenes ATCC 9525 strain with the others available in GenBank. Sequences are indicated with the GenBank accession number and numbers to the right and left indicate the position within the sequence. A 360-amino acid region is not shown because all the residues are conserved. Amino acid substitutions are highlighted in grey. Boxes represent the secretion signal sequence and the PEST-like sequence. The unique cysteine residue is marked in bold and by .
U25452; U25449; U25446; U25443; X15127; and X60035). The analysis revealed an identity of 97% at the nucleotide and of 99% at the amino acid levels. It is noteworthy that the majority of amino acid substitutions were found to be located at the NH2 -terminus, into the PEST-like sequence (Fig. 2). This 19amino acid sequence was submitted to the PEST-FIND program [15], obtaining a score of 7.15. The program produces a score ranging from )50 to +50: a score greater than +5 defines an interesting PEST-like sequence. Optimization of recombinant listeriolysin O expression The aim of our work was to obtain large amounts of LLO toxin in hemolytically active form by expressing this protein in prokaryotic hosts. For this purpose, the L. monocytogenes hlyA gene was inserted into the pET3a vector, generating the expression plasmid pET3a + LLO, as described under Materials and methods.
To avoid secretion of the recombinant protein in the culture medium and to obtain the hemolysin in its active form, this gene was cloned lacking the first 75 nt that code for the amino acid signal sequence for secretion. Thus, the recombinant expressed protein started from amino acid 26 (amino acid sequence deduced from GenBank AF253320) with the addition of a methionine at the NH2 -terminus. In the pET system, the expression of the exogenous gene is carried out by the bacteriophage T7 RNA polymerase, which is inducible adding IPTG to cell growing cultures. First of all, we optimized the induction conditions to obtain the expression of the largest amount of recombinant protein in a soluble and active form. In detail, we examined the influence of length and temperature of induction, the bacterial density, and the IPTG concentration. We performed three induction experiments, the first of which was carried out according to manufacturerÕs instructions: cells were grown with shaking at 37 °C until A600 nm reached 0.4, IPTG was then added to the final concentration
C. Giammarini et al. / Protein Expression and Purification 28 (2003) 78–85
of 0.4 mM, and the incubation at 37 °C was continued for 4 h. Bacterial extract, obtained after sonication of induced cells and centrifugation, as described under Materials and methods, was compared with the uninduced control by 10% SDS–PAGE. Unfortunately, the electrophoretic analysis did not reveal any 56 kDa protein difference (the expected molecular weight for LLO) between the induced and uninduced extract. In the second experiment, we increased the IPTG concentration to 1 mM and after its addition the temperature was lowered to 30 °C and the incubation was continued for 6 h. As in the previous test, no recombinant LLO was found in the soluble extract of induced cells. Then we decided to decrease drastically the induction temperature to avoid the recombinant protein accumulation as inclusion bodies. Therefore, in the third experiment, cells were grown with shaking at 30 °C until A600 nm reached 0.6, IPTG was then added to the final concentration of 1.0 mM and the incubation was continued at 22 °C for 16 h. In this case, a large 56 kDa protein band was pointed out by SDS–PAGE analysis only in the induced cell extract. We also verified the hemolytic activity of this sample, testing it on blood agar plates and detecting clear hemolysis zones. In conclusion, the highest amount of active recombinant LLO in soluble form was obtained in the last experiment, so all the next inductions were performed in the same conditions. Purification of the recombinant listeriolysin O The soluble proteins were obtained from induced E. coli cells containing pET-3a + LLO by a lysis procedure, as described under Materials and methods. The bacterial extract, starting from 1 litre of induced E. coli culture, contained around 250 mg of total proteins, the 10% (mean of five inductions) of which is represented by expressed LLO, as determined by Gel Doc 2000 Apparatus analysis. The hemolysin was then purified to homogeneity, following the purification procedure described by Traub and Bauer [10] based on an adsorption chromatography on hydroxyapatite, followed by an ion-exchange chromatography on SP Sepharose (see Materials and methods). A summary of the purification steps is presented in Table 1. The first step involved adsorption of the sample to hydroxyapatite in batch. The cake was then poured into a column of 5 cm of diameter and the matrix was washed until the A280nm
83
reached baseline. The hemolysin was then desorbed from hydroxyapatite column with 300 ml elution buffer and the 10 ml collected fractions were tested for hemolytic activity on blood agar plates and analyzed by SDS– PAGE. The fractions containing LLO were pooled and proteins were concentrated by ammonium sulfate precipitation. After equilibration in buffer A, the sample was applied to 25 ml SP Sepharose cation-exchange column. The column was washed with 400 ml of the same buffer and elution was obtained with a continuous salt gradient using 80 ml of buffer A and 80 ml of buffer B. The 5-ml collected fractions were tested as previously described and those containing LLO were pooled and concentrated by ultrafiltration to 4 ml. The final average yield was 4.5 mg protein per litre cell culture with a hemolytic activity of 1:25 106 HU/mg protein. In Fig. 3 is shown the analysis of samples from various steps in the purification protocol by SDS–PAGE and the Western blotting technique. Properties of the purified hemolysin The hemolysin obtained at the end of the purification procedure was analyzed by SDS–PAGE and appeared as a single polypeptide chain of 56 kDa, with no relevant traces of minor bands after Coomassie blue staining (Fig. 3A, lane 7). We characterized this recombinant LLO and compared its properties with those of the native one. At first, we verified the effect of pH on hemolytic activity testing the toxin at pH ranging from 5.0 to 8.5. As expected, the hemolysin showed its maximum activity at an acidic pH, that is, at pH 5.5, and the hemolytic activity rapidly decreased when the pH was raised to basic values. At neutral pH, it was about 30% and became almost irrelevant at pH values greater than 8. We also investigated if the purified protein showed the usual properties of the sulfhydryl-activated toxins, such as hemolytic activity inhibition by oxidation and by exposure to cholesterol. The assays were performed, incubating the recombinant LLO with various concentrations of specific reagents, as described under Materials and methods. Cholesterol causes the suppression of the recombinant LLO hemolytic activity to the extent of 20 ng cholesterol for 1 HU. In the same way, the hemolysin was inactivated in consequence of reaction with 1.7 lM HgCl2 or p-chloromercuribenzoate and in both cases the hemolytic activity was fully restored by 2 mM dithiothreitol.
Table 1 Purification of recombinant listeriolysin O Purification step Cell-free extract Hydroxyapatite chromatography SP Sepharose chromatography
Total protein (mg) 250 85 4.5
Total activity (HU) 7
1:39 10 9:44 106 5:62 106
Specific activity (HU/mg) 4
5:56 10 1:11 105 1:25 106
Recovery (%) 100 68 40.4
84
C. Giammarini et al. / Protein Expression and Purification 28 (2003) 78–85
Fig. 3. (A) SDS–PAGE and (B) immunochemical detection of listeriolysin O expressed with the pET system at different purification steps. (A) Lane 1, low-molecular-weight protein standards; lane 2, 10 lg of total post-induction cell proteins of bacterial cells transformed with the pET-3a lacking the hlyA gene; lane 3, 13 lg of total post-induction cell proteins of bacterial cells transformed with the pET-3a + LLO plasmid; lane 4, proteins unbound to hydroxyapatite Bio-Gel HTP Gel column; lane 5, 15 lg proteins concentrated by ammonium sulfate precipitation and dialyzed before loading onto SP Sepharose Fast Flow column; lane 6, 8.5 lg proteins unbound to SP Sepharose Fast Flow column; and lane 7, 3.3 lg of the sample at the end of the purification procedure. (B) An SDS–PAGE similar to those shown in (A) was electroblotted and the blot was probed with a sheep immune serum. Immunoreactive bands were visualized by horseradish-peroxidase-labelled rabbit anti-sheep IgG, followed by ABTS detection.
Discussion Listeria monocytogenes releases in the culture medium low levels of LLO, therefore as previously described by other authors, the isolation of this toxin requires high culture volumes, is time-consuming, and provides a very low yield [8,9,10]. Our aim was to obtain a significant amount of LLO in pure form for use in future studies. To this purpose, we expressed the protein in E. coli cloning the L. monocytogenes hlyA gene in a pET vector. The analysis of hlyA gene and of the deduced amino acid sequence from L. monocytogenes ATCC 9525 strain showed some differences with the corresponding sequences of other strains. Amino acid substitutions are located mostly in the PEST-like region and the analysis of our PEST-like sequence with the PEST-FIND program [15] revealed a score (7.15) higher than those previously reported [5,6], indicating a very likely PEST sequence (see Results). Thanks to the optimization of the induction procedure, we got a high-level expression of the recombinant LLO, which was purified to homogeneity adapting the protocol of Traub and Bauer [10] to our particular conditions. First of all in the E. coli-induced cells the toxin was expressed in the cytoplasmic compartment while in the case of L. monocytogenes cultures the protein was secreted in the medium. Therefore, to purify the recombinant LLO it was necessary to lyse the induced bacteria by sonication and to recover the proteins in soluble form by centrifugation. In the first chromatographic step, some problems occurred due to the compression of the matrix. These difficulties were overcome, applying the sample to the matrix in batch and then pouring the mixture in a wide column. The column was washed with a high volume of buffer to desorb all the unbound proteins and after LLO elution
the fractions with hemolytic activity were pooled. The proteins were then concentrated by ammonium sulfate precipitation and reequilibrated in a suitable buffer. The ion-exchange chromatography was basically performed as previously described [10], and finally, the recombinant LLO was concentrated approximately to 4 ml. The 56-kDa protein appeared to be immunologically pure in terms of immunoblot reactivity with sheep immune serum. Some characteristics of the recombinant LLO were compared to those of the natural ones. No difference was detected in the hemolytic activity; in fact, in both cases 1 HU corresponded to about 1 ng of protein [8,10]. As expected, recombinant LLO activity was maximum at pH 5.5 and rapidly decreased with the increase of the pH, but in contrast with that previously assessed [8], in our findings the purified toxin was not completely inactive at neutral pH. It also displayed the classical properties of the sulfhydryl-activated toxins, since its hemolytic activity was inhibited after binding to cholesterol and reagents that modify the SH-group inactivated the hemolysin, the toxicity of which was restored by reducing agent. Starting from 1 litre of cell culture, we obtained a high amount of purified recombinant LLO (on an average of 4.5 mg) in a hemolytically active form (around 5:62 106 HU total) that will be very useful for further structural and functional characterization (e.g., protein crystallization studies). The current protein concentration is 1.1 mg/ml. Although this is not high enough for crystallographic studies, it can be increased further by ultrafiltration starting from a large biomass (i.e. 5 litres). Moreover, the recombinant purified LLO should be used for several diagnostic and pharmaceutical applications. The diagnosis of listeriosis is currently performed by bacterial isolation and by serum aggluti-
C. Giammarini et al. / Protein Expression and Purification 28 (2003) 78–85
nation test. The first method takes a few days while the second is non-specific because of antigenic cross-reactivity between L. monocytogenes and other Gram-positive bacteria. LLO has been recently chosen as antigen for the development of new serodiagnostic tests because this toxin is a major virulence factor and is not found in other Listeria species [16,17]. Untill now a large-scale production of immunologic tests based on this toxin has not been possible because of the difficult isolation procedures, so the use of the recombinant protein could solve this problem. From another point of view, LLO could permit an efficient delivery of macromolecules across the cell membrane barrier into the cytosol. This hemolysin has been used successfully encapsulated inside liposomes along with other molecules, allowing the evasion of the cellular degradative pathway. Thus, adopting the strategy used by L. monocytogenes to enter the cytoplasm, the efficiency to deliver antigens or therapeutic molecules into cells could be considerably increased [18,19].
Acknowledgments This work was supported by ‘‘Istituto Zooprofilattico Sperimentale dellÕUmbria e delle Marche’’ (Ricerca Corrente 2001).
References [1] J. Rocourt, P. Cossart, Listeria monocytogenes, in: M.P. Doyle, L.R. Beuchat, T.J. Montville (Eds.), Food Microbiology: Fundamentals and Frontiers, ASM Press, Washington, DC, 1997, pp. 337–352. [2] D.A. Portnoy, S. Jones, The cell biology of Listeria monocytogenes infection (escape from a vacuole), Ann. N. Y. Acad. Sci. 730 (1994) 15–25. [3] M. Palmer, The family of thiol-activated, cholesterol-binding cytolysins, Toxicon 39 (2001) 1681–1689. [4] K.E. Beauregard, K.-D. Lee, R.J. Collier, J.A. Swanson, pHdependent perforation of macrophage phagosomes by listeriolysin O from Listeria monocytogenes, J. Exp. Med. 186 (1997) 1159– 1163.
85
[5] A.L. Decatur, D.A. Portnoy, A PEST-like sequence in listeriolysin O essential for Listeria monocytogenes pathogenicity, Science 290 (2000) 992–995. [6] M.-A. Lety, C. Frehel, I. Dubail, J.-L. Beretti, S. Kayal, P. Berche, A. Charbit, Identification of a PEST-like motif in listeriolysin O required for phagosomal escape and for virulence in Listeria monocytogenes, Mol. Microbiol. 39 (2001) 1124–1139. [7] T. Jacobs, A. Darji, N. Frahm, M. Rohde, J. Wehland, T. Chakraborty, S. Weiss, Listeriolysin O: cholesterol inhibits cytolysis but not binding to cellular membranes, Mol. Microbiol. 28 (1998) 1081–1089. [8] C. Geoffroy, J.-L. Gaillard, J.E. Alouf, P. Berche, Purification, characterization, and toxicity of the sulfhydryl-activated hemolysin listeriolysin O from Listeria monocytogenes, Infect. Immun. 55 (1987) 1641–1646. [9] A.L. Baetz, I.V. Wesley, Detection of anti-listeriolysin O in dairy cattle experimentally infected with Listeria monocytogenes, J. Vet. Diagn. Invest. 7 (1995) 82–86. [10] V.H. Traub, D. Bauer, Simplified purification of Listeria monocytogenes listeriolysin O and preliminary application in the enzyme-linked immunosorbent assay (ELISA), Zbl. Bakt. 283 (1995) 29–42. [11] F. Sanger, S. Nicklen, A.R. Coulson, DNA sequencing with chain-terminating inhibitors, Proc. Natl. Acad. Sci. USA 74 (1977) 5463–5467. [12] J.E. Alouf, M. Viette, R. Corvazier, M. Raynaud, Preparation and properties of antistreptolysin O horse sera, Ann Inst. Pasteur. 108 (1965) 476–500. [13] M.M. Bradford, A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein–dye binding, Anal. Biochem. 72 (1976) 248–254. [14] U.K. Laemmli, Cleavage of structural proteins during the assembly of the head of bacteriophage T4, Nature 227 (1970) 680–685. [15] M. Rechsteiner, S.W. Rogers, PEST sequences and regulation by proteolysis, Trends Biochem. Sci. 21 (1996) 267–271. [16] P. Berche, K.A. Reich, M. Bonnichon, J.-L. Beretti, C. Geoffroy, J. Raveneau, P. Cossart, J.-L. Gaillard, P. Geslin, H. Kreis, M. Veron, Detection of anti-listeriolysin O for serodiagnosis of human listeriosis, Lancet 335 (1990) 624–627. [17] A. Bourry, B. Poutrel, Bovine mastitis caused by Listeria monocytogenes: kinetics of antibody responses in serum and milk after experimental infection, J. Dairy Sci. 79 (1996) 2189–2195. [18] K.-D. Lee, Y.-K. Oh, D.A. Portnoy, J.A. Swanson, Delivery of macromolecules into cytosol using liposomes containing hemolysin from Listeria monocytogenes, J. Biol. Chem. 271 (1996) 7249– 7252. [19] C.J. Provoda, K.-D. Lee, Bacterial pore-forming hemolysins and their use in the cytosolic delivery of macromolecules, Adv. Drug. Deliv. Rev. 41 (2000) 209–221.
High-level expression of the Listeria monocytogenes listeriolysin O in Escherichia coli and preliminary characterization of the purified proteinq Camilla Giammarini,a Francesca Andreoni,a Giulia Amagliani,a Annarita Casiere,b Simone Barocci,b and Mauro Magnania,c,* a Centro di Biotecnologie, University of Urbino, Via T. Campanella 1, 61032 Fano, Italy Istituto Zooprofilattico Sperimentale Umbria e Marche, Via G. Salvemini 1, 06126 Perugia, Italy Istituto di Chimica Biologica ‘‘G. Fornaini,’’ University of Urbino, Via Saffi 2, 61029 Urbino, Italy
b c
Received 31 July 2002, and in revised form 6 November 2002
Abstract Listeriolysin O (LLO) is a cholesterol-binding sulfhydryl-activated hemolysin encoded by Listeria monocytogenes hlyA gene. After analyzing the nucleotide coding sequence of this gene from the ATCC 9525 L. monocytogenes strain, we cloned it in a pET vector for expression in Escherichia coli. Thanks to the optimization of the induction protocol, we achieved a high-level LLO synthesis (about 10% of total cell proteins) in hemolytically active form. The expressed hemolysin was then purified to homogeneity, as revealed by SDS–PAGE and Western blot analysis, by a hydroxyapatite adsorption chromatography, followed by an SP Sepharose ion-exchange chromatography. The recombinant protein showed the same properties determined for LLO purified from L. monocytogenes cultures and the characteristics of the sulfhydryl-activated toxins such as inactivation by oxidation and by reaction with cholesterol. By a combination of the pET expression system and the simple purification method, we obtained a significant amount of toxin (4.5 mg/litre cell culture) in a hemolytically active form (1:25 106 HU/mg protein). This procedure can solve the problem of LLO isolation from L. monocytogenes cultures, which is a difficult task, mainly owing to the low levels of toxin released in the culture media. The recombinant hemolysin, purified in sufficient quantities, could be very useful for structural studies and for diagnostic and pharmaceutical applications. Ó 2002 Elsevier Science (USA). All rights reserved.
Listeria monocytogenes is a Gram-positive, foodborne, human and animal pathogen. This bacterium causes an infection, named listeriosis, which especially affects immunocompromised patients, new-borns, and pregnant women and is characterized by a variety of severe syndromes, such as encephalitis, meningoencephalitis, septicemia, and abortion [1]. Owing to its ability to survive in various environmental conditions, such as low pH, high NaCl concentrations, and very low temperatures, L. monocytogenes is present both in raw and processed foods; moreover, being a facultative anq
The nucleotide sequence data reported in this paper appear in DDBJ, EMBL, and GenBank nucleotide sequence databases under Accession No. AF253320. * Corresponding author. Fax: +390721862834. E-mail address: [email protected] (M. Magnani).
aerobe its growth is not significantly affected by vacuum packaging. Following the ingestion of contaminated food, bacteria cross the intestinal barrier, reach the liver and the spleen, and then disseminate to the brain and placenta. L. monocytogenes is a facultative intracellular micro-organism capable of internalizing into macrophages, as well as into cells devoid of phagocytic activity, by inducing its own phagocytosis. After phagocytosis, the pathogen escapes from the bacteriumcontaining vacuole by lysis of the phagosomal membrane and reaches the host cytosol in which it multiplies. The bacteria move inside the cytosol by polymerization of the host cell actin into a polarized comet tail and when they contact the plasma membrane, they induce the formation of bacteria-containing protrusions. Afterwards, these protrusions are internalized into the neighboring cells in which the micro-organisms start a new life cycle. Each
1046-5928/02/$ - see front matter Ó 2002 Elsevier Science (USA). All rights reserved. doi:10.1016/S1046-5928(02)00682-4
C. Giammarini et al. / Protein Expression and Purification 28 (2003) 78–85
step of L. monocytogenes intracellular parasitism depends upon the production of virulence factors among which listeriolysin O (LLO) is one of the most important [1,2]. LLO is an exotoxin, encoded by the hlyA gene, which belongs to the group of cholesterol-binding sulfhydryl-activated toxins, expressed by a large number of Gram-positive bacteria. All the hemolysins in this family have a single unique cysteine, which renders them susceptible to reversible inactivation by oxidation. However, this cysteine can be changed to an alanine without affecting their activity. Their lytic activity is enhanced by reducing agents and is suppressed by exposure to oxygen, cholesterol or anti-streptolysin O antibodies [3]. The molecular mechanism of membrane damage produced by this group of toxins is based on pore formation in membranes containing cholesterol. After binding to cholesterol, 20–80 toxin monomers oligomerize into arcand ring-shaped structures, which cause the passive flux of ions and macromolecules. In contrast to the other sulfhydryl-activated toxins, LLO has optimum activity at pH 5.5 and is almost entirely inactivated at pH 7.0. Owing to this property, the toxin activation occurs only after phagosomal acidification and causes the perforation of L. monocytogenes-containing vacuole and the bacterium diffusion into the cytosol [4]. The deleterious effects of LLO on cytoplasmic membrane are avoided, thanks to its low activity at cellular pH and its rapid degradation in the host-cell cytosol. Indeed, in its NH2 terminus there is a 19-amino acid PEST-like sequence that may target this toxin for degradation [5]. This motif seems to be also essential for bacterial virulence, since its deletion causes a defect in bacterium phagosomal escaping [6]. Another study describes a divergent LLO interaction with cholesterol in solution which inhibits hemolytic activity but does not alter membrane binding [7]. However, the unique LLO molecular mechanism of action is not still completely explained and protein studies are very difficult because the toxin isolation from L. monocytogenes cultures provides very low yields [8,9]. To overcome this problem, we expressed the hemolysin in Escherichia coli, using the pET expression system, obtaining significant amounts of recombinant LLO. The protein purification was then carried out adjusting a method previously described by Traub and Bauer [10] to our conditions. In this paper, we also compare the hemolytic activity and the properties of the recombinant LLO with those of the natural one purified from L. monocytogenes cultures by other authors [8,10].
Materials and methods Construction of the expression vector pET-3a + LLO The L. monocytogenes DNA was extracted by boiling method from the 9525 bacterial strain obtained from
79
American Type Culture Collection (ATCC, Manassas, VA). The region of hlyA gene, coding for the secreted form of LLO, was isolated by polymerase chain reaction (GeneAmp PCR Kit, Roche Molecular Systems, Branchburg, NJ). The amplification was performed using degenerate primers (LLO-3: 50 -CAACAAACTGA ACATATGGACGCATCTGCA-30 ; LLO-2: 50 -ATATT CTTTTACATATGTTAATTCTTC-30 ), which introduced NdeI restriction sites at nucleotides 1559 and 3216 (GenBank Accession No. M24199), respectively. The 1686-bp PCR product was then digested with the NdeI restriction enzyme (Roche Diagnostics, Indianapolis, IN) obtaining a 1657-bp fragment (coding for amino acids 26–529 of LLO). After gel-purification, it was cloned into the NdeI site of a pET-3a vector (Novagen, Madison, WI) to generate the construct referred to as pET-3a + LLO (Fig. 1). Nucleotide sequencing analysis The hlyA coding region of the ATCC 9525 L. monocytogenes strain was isolated by PCR with LLO-8 (50 -CA TCTTTAGAAGCGAATTT-30 ) and LLO-6 (50 -CA AGCTTATTTTTTCGTGTGTG-30 ) primers designed on the hlyA sequence (GenBank Accession No. M24199). Direct sequencing of PCR product was performed by the Sanger method [11] with the ABI PRISM BigDye Terminator Cycle Sequencing Kit (Applied Biosystems, Foster City, CA) following manufacturerÕs protocols, using 90 ng DNA and 3.2 pmol of one of the following sequencing primers: LL5 forward 50 -AACCTATCCAGGTGCTC-30 LL6 reverse 50 -CTGTAAGCCATTTCGTC-30 LLO-5 forward 50 -CGTATGGCCGTCAAGTTTA TTT-30 LLO-6 reverse 50 -CAAGCTTATTTTTTCGTGTG TG-30 The reaction mixtures were purified with Centri-Sep columns (Princeton Separations, Adelphia, NJ) and run on an ABI PRISM 310 Genetic Analyzer (Applied Biosystems). Expression of the recombinant listeriolysin O The recombinant plasmid pET-3a + LLO, obtained as described above, was transformed into E. coli strain BL21(DE3) (Novagen). The induction experiments were initially performed according to manufacturerÕs instructions. Briefly, cells were grown in LB broth plus ampicillin (at a concentration of 50 lg/ml) at 37 °C to early log phase (A600 nm ¼ 0:4). IPTG (Bio-Rad, Hercules, CA) was then added to a final concentration of 0.4 mM to induce the T7 RNA polymerase gene and cultures were incubated further at 37 °C for 4 h.
80
C. Giammarini et al. / Protein Expression and Purification 28 (2003) 78–85
Fig. 1. Construction of the plasmid for the expression of the L. monocytogenes toxin listeriolysin O. The steps performed to construct the pET3a + LLO expression vector are illustrated (see Materials and methods for details).
Purification of the recombinant listeriolysin O Cells derived from 1 litre of post-induction culture were harvested by centrifugation at 6000g for 30 min at 4 °C. The resulting pellet was resuspended in 50 ml lysis buffer at pH 6.1 containing 20 mM NaH2 PO4 , 1 mM dithiothreitol (DTT), 5% (v/v) glycerol, 0.5% (v/v) Triton X-100, and an anti-proteolytic cocktail (Sigma–Aldrich, St. Louis, MA). The cell suspension was sonicated five times for 30 s at 200 W and then centrifuged for 45 min at 25,000g at 4 °C. LLO present in the soluble fraction was then purified to homogeneity following a two-step purification procedure based on Traub and BauerÕs protocol [10]. The supernatant obtained from the previous step was applied to 40 ml hydroxyapatite Bio-Gel HTP Gel (Bio-Rad) equilibrated in 20 mM so-
dium phosphate buffer, pH 6.1. The column was washed with the equilibrating buffer and then LLO was desorbed with 0.3 M NaH2 PO4 , 1 mM DTT, and 5% (v/v) glycerol, pH 6.1. The collected fractions were tested for hemolytic activity on blood agar plates and analyzed by SDS–PAGE. Solid ammonium sulfate was slowly added to the pooled fractions, containing LLO from the previous step, to achieve an 85% saturation. The suspension was gently stirred for 30 min and then centrifuged at 10,000g for 30 min. The pellet was dissolved in 10 ml buffer A (50 mM NaH2 PO4 , 1 mM EDTA, 2.7 mM KCl, 1 mM DTT, and 5% (v/v) glycerol, pH 6.4) and then dialyzed twice (1 h each) against 2 litre of the same buffer. The dialyzed protein solution was applied to an SP Sepharose Fast Flow (Amersham Pharmacia Biotech, Uppsala, Sweden) column (2 8 cm) equilibrated
C. Giammarini et al. / Protein Expression and Purification 28 (2003) 78–85
in buffer A. The column was washed with the same buffer and desorption was carried out with a continuous salt gradient (0–1 M NaCl), using buffer B consisting of buffer A plus 1 M NaCl. Finally, the fractions containing LLO were concentrated with Amicon (8200) Stirred Ultrafiltration Cell (Bedford, MA). Hemolytic activity titration The hemolytic activity of a toxin is expressed as hemolytic units (HU) per millilitre of toxin solution or per milligram of total protein. The HU is the smallest amount (highest dilution) of toxin that liberates half the hemoglobin (50% lysis) from a suspension of erythrocytes. The hemolytic activity of purified LLO was determined preparing 1 ml toxin serial dilutions in PBS (pH 5.5) containing 0.1% (w/v) BSA (Sigma–Aldrich) and then adding 0.5 ml human red blood cells suspension. Erythrocyte suspension was used at 2.5% hematocrit, standardized so that 0.5 ml human red blood cells lysed by adding 14.5 ml of 0.1% sodium carbonate solution give an optical absorbance of 0.200 at 541 nm [12]. After 45 min at 37 °C the samples were centrifuged for 5 min at 1700g and the optical absorbance of the hemoglobin in the supernatant was measured at 541 nm. The data were estimated graphically by plotting percent lysis versus toxin dilution on a log probit graph. Determination of protein content Protein concentration was determined using the Bradford method [13] with bovine serum albumin as standard or spectrophotometrically by measuring the absorbance of solutions at 280 nm against appropriate blanks. SDS–PAGE SDS–PAGE of the expressed LLO was performed in 10% polyacrylamide gels according to the method of Laemmli [14] utilizing Bio-Rad equipment. The proteins were visualized by staining with Coomassie brilliant blue. The yield of the recombinant LLO in crude extract was determined approximately by comparison with known concentrations of molecular-weight protein standards, using the Gel Doc 2000 Apparatus (BioRad). The volume of the band corresponding to LLO was compared with the volume of the area from a gel lane containing the total protein, which was also quantified by the Bradford assay.
81
collected 3 weeks later and stored (50 ll aliquots) at )80 °C. Sheep immune serum, diluted 1:100, was used for immunoreactivity against LLO in the Western blot analysis. SDS–PAGE gels were blotted onto Hybond-C Extra nitrocellulose membrane (Amersham Pharmacia Biotech) by the Mini Trans-Blot Electrophoretic Transfer Cell (Bio-Rad). The second antibody was HRP-conjugated rabbit anti-sheep IgG (BIOTREND, K€ oln, Germany), diluted 1:5000, and the substrate was 2,20 -azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) (Sigma–Aldrich). Hemolytic activity at different pH The hemolytic activity of LLO was tested dissolving 10 HU in 1 ml PBS at different pH values: from 5.0 to 8.5. After the addition of RBC suspension and the incubation at 37 °C, the hemolytic activity titration was performed as described above. Hemolytic activity after incubation with cholesterol Cholesterol (Sigma–Aldrich) was dissolved in absolute ethanol at an initial concentration of 1 mg/ml and then serial dilutions were made in ethanol. Ten microliters of various concentrations of cholesterol was added to 1 ml samples containing 30 HU toxin in PBS, pH 6.0, supplemented with 0.1% (w/v) BSA. One milliliter of toxin with 10 ll ethanol was used as control. After 30 min of incubation at 22 °C, 0.5 ml erythrocyte suspension was added to the samples and the hemolytic activity was determined, as previously described. Hemolytic activity after incubation with SH-group reagents As SH-group reagents, we used HgCl2 (Sigma–Aldrich) dissolved in PBS, pH 6.0, at the final concentration of 10 mM and p-chloromercuribenzoate (Sigma–Aldrich) in PBS, pH 6.0, with 0.1 M NaOH at the concentration of 5 mM. Thirty hemolytic units of LLO in 1 ml PBS, pH 6.0, supplemented with 0.1% (w/v) BSA was incubated with 10 ll of various dilutions of the oxidants for 30 min at 22 °C. The reagent concentrations that inhibited the hemolytic activity of 1 HU of toxin were determined, as described above for the cholesterol assay.
Results hlyA gene analysis
Immunoblotting One Sardinian sheep was intravenously infected once with 7 108 CFU/ml of viable L. monocytogenes ATCC 9525 strain. Serum samples of the infected sheep were
The entire coding region of hlyA gene was sequenced (GenBank Accession No. AF253320) and aligned with the corresponding regions of different L. monocytogenes strains available on GenBank (M24199; AL591974;
82
C. Giammarini et al. / Protein Expression and Purification 28 (2003) 78–85
Fig. 2. Alignment of LLO amino acid sequence from L. monocytogenes ATCC 9525 strain with the others available in GenBank. Sequences are indicated with the GenBank accession number and numbers to the right and left indicate the position within the sequence. A 360-amino acid region is not shown because all the residues are conserved. Amino acid substitutions are highlighted in grey. Boxes represent the secretion signal sequence and the PEST-like sequence. The unique cysteine residue is marked in bold and by .
U25452; U25449; U25446; U25443; X15127; and X60035). The analysis revealed an identity of 97% at the nucleotide and of 99% at the amino acid levels. It is noteworthy that the majority of amino acid substitutions were found to be located at the NH2 -terminus, into the PEST-like sequence (Fig. 2). This 19amino acid sequence was submitted to the PEST-FIND program [15], obtaining a score of 7.15. The program produces a score ranging from )50 to +50: a score greater than +5 defines an interesting PEST-like sequence. Optimization of recombinant listeriolysin O expression The aim of our work was to obtain large amounts of LLO toxin in hemolytically active form by expressing this protein in prokaryotic hosts. For this purpose, the L. monocytogenes hlyA gene was inserted into the pET3a vector, generating the expression plasmid pET3a + LLO, as described under Materials and methods.
To avoid secretion of the recombinant protein in the culture medium and to obtain the hemolysin in its active form, this gene was cloned lacking the first 75 nt that code for the amino acid signal sequence for secretion. Thus, the recombinant expressed protein started from amino acid 26 (amino acid sequence deduced from GenBank AF253320) with the addition of a methionine at the NH2 -terminus. In the pET system, the expression of the exogenous gene is carried out by the bacteriophage T7 RNA polymerase, which is inducible adding IPTG to cell growing cultures. First of all, we optimized the induction conditions to obtain the expression of the largest amount of recombinant protein in a soluble and active form. In detail, we examined the influence of length and temperature of induction, the bacterial density, and the IPTG concentration. We performed three induction experiments, the first of which was carried out according to manufacturerÕs instructions: cells were grown with shaking at 37 °C until A600 nm reached 0.4, IPTG was then added to the final concentration
C. Giammarini et al. / Protein Expression and Purification 28 (2003) 78–85
of 0.4 mM, and the incubation at 37 °C was continued for 4 h. Bacterial extract, obtained after sonication of induced cells and centrifugation, as described under Materials and methods, was compared with the uninduced control by 10% SDS–PAGE. Unfortunately, the electrophoretic analysis did not reveal any 56 kDa protein difference (the expected molecular weight for LLO) between the induced and uninduced extract. In the second experiment, we increased the IPTG concentration to 1 mM and after its addition the temperature was lowered to 30 °C and the incubation was continued for 6 h. As in the previous test, no recombinant LLO was found in the soluble extract of induced cells. Then we decided to decrease drastically the induction temperature to avoid the recombinant protein accumulation as inclusion bodies. Therefore, in the third experiment, cells were grown with shaking at 30 °C until A600 nm reached 0.6, IPTG was then added to the final concentration of 1.0 mM and the incubation was continued at 22 °C for 16 h. In this case, a large 56 kDa protein band was pointed out by SDS–PAGE analysis only in the induced cell extract. We also verified the hemolytic activity of this sample, testing it on blood agar plates and detecting clear hemolysis zones. In conclusion, the highest amount of active recombinant LLO in soluble form was obtained in the last experiment, so all the next inductions were performed in the same conditions. Purification of the recombinant listeriolysin O The soluble proteins were obtained from induced E. coli cells containing pET-3a + LLO by a lysis procedure, as described under Materials and methods. The bacterial extract, starting from 1 litre of induced E. coli culture, contained around 250 mg of total proteins, the 10% (mean of five inductions) of which is represented by expressed LLO, as determined by Gel Doc 2000 Apparatus analysis. The hemolysin was then purified to homogeneity, following the purification procedure described by Traub and Bauer [10] based on an adsorption chromatography on hydroxyapatite, followed by an ion-exchange chromatography on SP Sepharose (see Materials and methods). A summary of the purification steps is presented in Table 1. The first step involved adsorption of the sample to hydroxyapatite in batch. The cake was then poured into a column of 5 cm of diameter and the matrix was washed until the A280nm
83
reached baseline. The hemolysin was then desorbed from hydroxyapatite column with 300 ml elution buffer and the 10 ml collected fractions were tested for hemolytic activity on blood agar plates and analyzed by SDS– PAGE. The fractions containing LLO were pooled and proteins were concentrated by ammonium sulfate precipitation. After equilibration in buffer A, the sample was applied to 25 ml SP Sepharose cation-exchange column. The column was washed with 400 ml of the same buffer and elution was obtained with a continuous salt gradient using 80 ml of buffer A and 80 ml of buffer B. The 5-ml collected fractions were tested as previously described and those containing LLO were pooled and concentrated by ultrafiltration to 4 ml. The final average yield was 4.5 mg protein per litre cell culture with a hemolytic activity of 1:25 106 HU/mg protein. In Fig. 3 is shown the analysis of samples from various steps in the purification protocol by SDS–PAGE and the Western blotting technique. Properties of the purified hemolysin The hemolysin obtained at the end of the purification procedure was analyzed by SDS–PAGE and appeared as a single polypeptide chain of 56 kDa, with no relevant traces of minor bands after Coomassie blue staining (Fig. 3A, lane 7). We characterized this recombinant LLO and compared its properties with those of the native one. At first, we verified the effect of pH on hemolytic activity testing the toxin at pH ranging from 5.0 to 8.5. As expected, the hemolysin showed its maximum activity at an acidic pH, that is, at pH 5.5, and the hemolytic activity rapidly decreased when the pH was raised to basic values. At neutral pH, it was about 30% and became almost irrelevant at pH values greater than 8. We also investigated if the purified protein showed the usual properties of the sulfhydryl-activated toxins, such as hemolytic activity inhibition by oxidation and by exposure to cholesterol. The assays were performed, incubating the recombinant LLO with various concentrations of specific reagents, as described under Materials and methods. Cholesterol causes the suppression of the recombinant LLO hemolytic activity to the extent of 20 ng cholesterol for 1 HU. In the same way, the hemolysin was inactivated in consequence of reaction with 1.7 lM HgCl2 or p-chloromercuribenzoate and in both cases the hemolytic activity was fully restored by 2 mM dithiothreitol.
Table 1 Purification of recombinant listeriolysin O Purification step Cell-free extract Hydroxyapatite chromatography SP Sepharose chromatography
Total protein (mg) 250 85 4.5
Total activity (HU) 7
1:39 10 9:44 106 5:62 106
Specific activity (HU/mg) 4
5:56 10 1:11 105 1:25 106
Recovery (%) 100 68 40.4
84
C. Giammarini et al. / Protein Expression and Purification 28 (2003) 78–85
Fig. 3. (A) SDS–PAGE and (B) immunochemical detection of listeriolysin O expressed with the pET system at different purification steps. (A) Lane 1, low-molecular-weight protein standards; lane 2, 10 lg of total post-induction cell proteins of bacterial cells transformed with the pET-3a lacking the hlyA gene; lane 3, 13 lg of total post-induction cell proteins of bacterial cells transformed with the pET-3a + LLO plasmid; lane 4, proteins unbound to hydroxyapatite Bio-Gel HTP Gel column; lane 5, 15 lg proteins concentrated by ammonium sulfate precipitation and dialyzed before loading onto SP Sepharose Fast Flow column; lane 6, 8.5 lg proteins unbound to SP Sepharose Fast Flow column; and lane 7, 3.3 lg of the sample at the end of the purification procedure. (B) An SDS–PAGE similar to those shown in (A) was electroblotted and the blot was probed with a sheep immune serum. Immunoreactive bands were visualized by horseradish-peroxidase-labelled rabbit anti-sheep IgG, followed by ABTS detection.
Discussion Listeria monocytogenes releases in the culture medium low levels of LLO, therefore as previously described by other authors, the isolation of this toxin requires high culture volumes, is time-consuming, and provides a very low yield [8,9,10]. Our aim was to obtain a significant amount of LLO in pure form for use in future studies. To this purpose, we expressed the protein in E. coli cloning the L. monocytogenes hlyA gene in a pET vector. The analysis of hlyA gene and of the deduced amino acid sequence from L. monocytogenes ATCC 9525 strain showed some differences with the corresponding sequences of other strains. Amino acid substitutions are located mostly in the PEST-like region and the analysis of our PEST-like sequence with the PEST-FIND program [15] revealed a score (7.15) higher than those previously reported [5,6], indicating a very likely PEST sequence (see Results). Thanks to the optimization of the induction procedure, we got a high-level expression of the recombinant LLO, which was purified to homogeneity adapting the protocol of Traub and Bauer [10] to our particular conditions. First of all in the E. coli-induced cells the toxin was expressed in the cytoplasmic compartment while in the case of L. monocytogenes cultures the protein was secreted in the medium. Therefore, to purify the recombinant LLO it was necessary to lyse the induced bacteria by sonication and to recover the proteins in soluble form by centrifugation. In the first chromatographic step, some problems occurred due to the compression of the matrix. These difficulties were overcome, applying the sample to the matrix in batch and then pouring the mixture in a wide column. The column was washed with a high volume of buffer to desorb all the unbound proteins and after LLO elution
the fractions with hemolytic activity were pooled. The proteins were then concentrated by ammonium sulfate precipitation and reequilibrated in a suitable buffer. The ion-exchange chromatography was basically performed as previously described [10], and finally, the recombinant LLO was concentrated approximately to 4 ml. The 56-kDa protein appeared to be immunologically pure in terms of immunoblot reactivity with sheep immune serum. Some characteristics of the recombinant LLO were compared to those of the natural ones. No difference was detected in the hemolytic activity; in fact, in both cases 1 HU corresponded to about 1 ng of protein [8,10]. As expected, recombinant LLO activity was maximum at pH 5.5 and rapidly decreased with the increase of the pH, but in contrast with that previously assessed [8], in our findings the purified toxin was not completely inactive at neutral pH. It also displayed the classical properties of the sulfhydryl-activated toxins, since its hemolytic activity was inhibited after binding to cholesterol and reagents that modify the SH-group inactivated the hemolysin, the toxicity of which was restored by reducing agent. Starting from 1 litre of cell culture, we obtained a high amount of purified recombinant LLO (on an average of 4.5 mg) in a hemolytically active form (around 5:62 106 HU total) that will be very useful for further structural and functional characterization (e.g., protein crystallization studies). The current protein concentration is 1.1 mg/ml. Although this is not high enough for crystallographic studies, it can be increased further by ultrafiltration starting from a large biomass (i.e. 5 litres). Moreover, the recombinant purified LLO should be used for several diagnostic and pharmaceutical applications. The diagnosis of listeriosis is currently performed by bacterial isolation and by serum aggluti-
C. Giammarini et al. / Protein Expression and Purification 28 (2003) 78–85
nation test. The first method takes a few days while the second is non-specific because of antigenic cross-reactivity between L. monocytogenes and other Gram-positive bacteria. LLO has been recently chosen as antigen for the development of new serodiagnostic tests because this toxin is a major virulence factor and is not found in other Listeria species [16,17]. Untill now a large-scale production of immunologic tests based on this toxin has not been possible because of the difficult isolation procedures, so the use of the recombinant protein could solve this problem. From another point of view, LLO could permit an efficient delivery of macromolecules across the cell membrane barrier into the cytosol. This hemolysin has been used successfully encapsulated inside liposomes along with other molecules, allowing the evasion of the cellular degradative pathway. Thus, adopting the strategy used by L. monocytogenes to enter the cytoplasm, the efficiency to deliver antigens or therapeutic molecules into cells could be considerably increased [18,19].
Acknowledgments This work was supported by ‘‘Istituto Zooprofilattico Sperimentale dellÕUmbria e delle Marche’’ (Ricerca Corrente 2001).
References [1] J. Rocourt, P. Cossart, Listeria monocytogenes, in: M.P. Doyle, L.R. Beuchat, T.J. Montville (Eds.), Food Microbiology: Fundamentals and Frontiers, ASM Press, Washington, DC, 1997, pp. 337–352. [2] D.A. Portnoy, S. Jones, The cell biology of Listeria monocytogenes infection (escape from a vacuole), Ann. N. Y. Acad. Sci. 730 (1994) 15–25. [3] M. Palmer, The family of thiol-activated, cholesterol-binding cytolysins, Toxicon 39 (2001) 1681–1689. [4] K.E. Beauregard, K.-D. Lee, R.J. Collier, J.A. Swanson, pHdependent perforation of macrophage phagosomes by listeriolysin O from Listeria monocytogenes, J. Exp. Med. 186 (1997) 1159– 1163.
85
[5] A.L. Decatur, D.A. Portnoy, A PEST-like sequence in listeriolysin O essential for Listeria monocytogenes pathogenicity, Science 290 (2000) 992–995. [6] M.-A. Lety, C. Frehel, I. Dubail, J.-L. Beretti, S. Kayal, P. Berche, A. Charbit, Identification of a PEST-like motif in listeriolysin O required for phagosomal escape and for virulence in Listeria monocytogenes, Mol. Microbiol. 39 (2001) 1124–1139. [7] T. Jacobs, A. Darji, N. Frahm, M. Rohde, J. Wehland, T. Chakraborty, S. Weiss, Listeriolysin O: cholesterol inhibits cytolysis but not binding to cellular membranes, Mol. Microbiol. 28 (1998) 1081–1089. [8] C. Geoffroy, J.-L. Gaillard, J.E. Alouf, P. Berche, Purification, characterization, and toxicity of the sulfhydryl-activated hemolysin listeriolysin O from Listeria monocytogenes, Infect. Immun. 55 (1987) 1641–1646. [9] A.L. Baetz, I.V. Wesley, Detection of anti-listeriolysin O in dairy cattle experimentally infected with Listeria monocytogenes, J. Vet. Diagn. Invest. 7 (1995) 82–86. [10] V.H. Traub, D. Bauer, Simplified purification of Listeria monocytogenes listeriolysin O and preliminary application in the enzyme-linked immunosorbent assay (ELISA), Zbl. Bakt. 283 (1995) 29–42. [11] F. Sanger, S. Nicklen, A.R. Coulson, DNA sequencing with chain-terminating inhibitors, Proc. Natl. Acad. Sci. USA 74 (1977) 5463–5467. [12] J.E. Alouf, M. Viette, R. Corvazier, M. Raynaud, Preparation and properties of antistreptolysin O horse sera, Ann Inst. Pasteur. 108 (1965) 476–500. [13] M.M. Bradford, A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein–dye binding, Anal. Biochem. 72 (1976) 248–254. [14] U.K. Laemmli, Cleavage of structural proteins during the assembly of the head of bacteriophage T4, Nature 227 (1970) 680–685. [15] M. Rechsteiner, S.W. Rogers, PEST sequences and regulation by proteolysis, Trends Biochem. Sci. 21 (1996) 267–271. [16] P. Berche, K.A. Reich, M. Bonnichon, J.-L. Beretti, C. Geoffroy, J. Raveneau, P. Cossart, J.-L. Gaillard, P. Geslin, H. Kreis, M. Veron, Detection of anti-listeriolysin O for serodiagnosis of human listeriosis, Lancet 335 (1990) 624–627. [17] A. Bourry, B. Poutrel, Bovine mastitis caused by Listeria monocytogenes: kinetics of antibody responses in serum and milk after experimental infection, J. Dairy Sci. 79 (1996) 2189–2195. [18] K.-D. Lee, Y.-K. Oh, D.A. Portnoy, J.A. Swanson, Delivery of macromolecules into cytosol using liposomes containing hemolysin from Listeria monocytogenes, J. Biol. Chem. 271 (1996) 7249– 7252. [19] C.J. Provoda, K.-D. Lee, Bacterial pore-forming hemolysins and their use in the cytosolic delivery of macromolecules, Adv. Drug. Deliv. Rev. 41 (2000) 209–221.