Isolation and Partial Characterization of Glycolipid Fractions from

0 downloads 0 Views 1MB Size Report
Florida 333101,1 and Department of Veterinary Pathology4 and Department of Veterinary Physiology,. Pharmacology and Toxicology,3 School of Veterinary ...
Vol. 61, No. 1

INFECTION AND IMMUNI, Jan. 1993, p. 1-7

0019-9567/93/010001-07$02.00/0 Copyright © 1993, American Society for Microbiology

Isolation and Partial Characterization of Glycolipid Fractions from Mycobacterium avium Serovar 2 (Mycobacterium paratuberculosis 18) That Inhibit Activated Macrophages MURRAY E. HINES II,1* JESSE M. JAYNES,2 STEVEN A. BARKER,3 JOSEPH C. NEWTON,4 FREDERICK M. ENRIGHT,' AND THERON G. SNIDER III4

Division of Comparative Pathology, School of Medicine, University of Miami Box 016960 (R-46), Miami, Florida 333101,1 and Department of Veterinary Pathology4 and Department of Veterinary Physiology, Pharmacology and Toxicology,3 School of Veterinary Medicine, Department of Biochemistry,2 and Department of Veterinary Science,5 Louisiana State University, Baton Rouge, Louisiana 70803 Received 18 June 1992/Accepted 5 October 1992

Glycolipid fractions from Mycobacterium avium serovar 2 (Mycobacterium paratuberculosis 18) inhibited the killing of Candida albicans by activated bovine peripheral-blood-derived macrophages. Fractions were derived by using the matrix solid-phase dispersion technique, which is a new method of simultaneous lysis and partial fractionation of components of bacterial cells. Further purification of active fractions was performed by concanavalin A affinity chromatography, centrifugal filtration, and differing solvent solubility. Three different fractions were isolated and partially characterized. Two of these fractions have characteristics typical of glycolipids, and the third fraction has characteristics compatible with a peptidoglycolipid. This peptidoglycolipid fraction has been purified and named MIF-A3. M. paratuberculosis 18 (M. avium serovar 2) that markedly inhibit the killing ability of bovine monocyte-M4. We used a new method of lysis and fractionation termed matrix solidphase dispersion (MSPD) (1, 21, 27, 28). (This work was conducted by M. E. Hines II in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Interdepartmental Program in Veterinary Medical Sciences, Department of Veterinary Pathology, Louisiana State University.)

Mycobacterium paratuberculosis is the cause of bovine paratuberculosis (Johne's disease), a chronic, debilitating, granulomatous, enteric disease of cattle (7). M. paratuberculosis may be a human pathogen, since it has been isolated from non-AIDS patients with Crohn's disease (8-10, 30, 31). Similarities in morphology and a close antigenic relationship between M. paratuberculosis and Mycobacterium avium surface glycopeptidolipids have been noted (5). Comparisons of various mycobacterial species by DNA probe and restriction fragment length polymorphism analysis also have shown that M. paratuberculosis 18 is identical to M. avium serovar 21 (26, 32). Like other mycobacteria, M. paratuberculosis is phagocytized by and replicates within macrophages (M4O). The organism is protected intracellularly against humoral and cellular defense mechanisms (34). Cellular immunity is thought to be of major importance in resistance to infection with M. paratuberculosis. Bovine monocyte-derived M4 have been shown to phagocytize M. paratuberculosis more readily than freshly adherent monocytes do; optimal phagocytosis requires serum concentrations of 5 to 20% (49). Ingested organisms have been shown to multiply by as much as 250% over a 7-day incubation period in both bovine monocytes and monocyte-derived M4 (49). Compounds in the cell walls of some species of mycobacteria appear to inhibit M4 killing. However, very little is known about the production of a factor(s) by M. paratuberculosis that inhibits killing by bovine M4. M4-inhibitory factors have been investigated with Mycobacterium tuberculosis (3, 6, 38, 46, 47) and M. leprae (36, 38-40). Elucidation of the intracellular protective mechanisms of M. paratuberculosis may provide information that would contribute to the understanding, prevention, and treatment of related animal and human mycobacterial diseases. In this study, we provide evidence for three glycolipid fractions derived from

*

MATERIALS AND METHODS General methods. Cultures of M. paratuberculosis 18 (American Type Culture Collection, Rockville, Md.) were grown as previously described (21). Standard compounds and solvents of the highest purity available (Sigma Chemical and Fisher Chemical) and triple-distilled type I high-performance liquid chromatography (HPLC)-grade water were used. The presence and quantity of lipopolysaccharide (LPS) in stock solutions and fractions were determined by using a Limulus amebocyte lysate assay (E-toxate; Sigma). The presence and quantity of protein were determined by the Markwell modification of the Lowry et al. technique (29). Lysis and fractionation of M. paratuberculosis samples (0.5 g of pelleted wet weight) were performed by the MSPD technique as previously described (21). Sample treatments. Blanks of all sample treatments consisted of similarly treated type I HPLC-grade water unless otherwise noted. Protein in the acetonitrile and H20 fractions was degraded in the following manner. Portions of protease K (20 mg of dry weight, 0.5 U of activity) brought out of solution by covalent binding to 4.0% agarose beads (Sigma) were washed 5 times with 0.5 ml of H20 in a centrifugal filtration unit (spin X; Costar, Cambridge, Mass.) containing a 0.45-pLm-pore-size cellulose acetate membrane at 5,200 x g for 5 min. The protease K-agarose was resuspended within the upper chamber with 0.5 ml of each sample and transferred to a fresh tube containing the remain-

Corresponding author. 1

2

HINES ET AL.

der of each sample. Samples were incubated for 24 h at 37°C with occasional shaking and then filtered in a microcentrifuge at 12,000 x g with 0.45-,um-pore-size centrifugal filtration units to remove the protease K-agarose. Lipid was removed from the acetonitrile and H20 fractions by adding 3 volumes of chloroform (3.0 ml), mixing thoroughly, and then centrifuging at 650 x g for 4 min. The supernatant was removed and transferred to a clean tube. Chloroform extraction was repeated three times, and then the samples were lyophilized to remove traces of chloroform. Carbohydrate groups were degraded with sodium metaperiodate as previously described (47). Samples of the acetonitrile and H20 fractions were reacted with antilipoarabinomannan (LAM) monoclonal antibodies obtained from M. tuberculosis H37Ra (kindly provided by P. J. Brennan, Colorado State University). Samples (1 ml) of 1:200 dilutions of the anti-LAM antibodies were added to the lyophilized samples of the acetonitrile and H20. Samples were incubated at 4°C for 48 h and then dialyzed in dialysis tubing (3,500-molecular-weight cutoff) against four changes of buffer (4 liters each) for 48 h at 4°C. Dialysis was done to remove any traces of the sodium azide present in the antibody. Samples were lyophilized, resuspended in 1.0 ml of H20, and stored frozen at -20°C before testing. The dried acetonitrile extract from 1.5 g of M. paratuberculosis was resuspended in 1.0 ml of H20 and centrifuged with a Centricon 30 microconcentrator (Amicon, Danvers, Mass.) in a 450 fixed-angle bench-top centrifuge at 1,500 x g for 3 h. The retentate was resuspended in 1.0 ml of H20. The H20-lyophilized extract from 1.5 g of M. paratuberculosis was resuspended in 3.0 ml of H20 and subjected to concanavalin A (ConA) affinity chromatography to selectively bind and then elute compounds containing mannan and glucan groups (i.e., LAM, lipomannan, arabinogalactan, etc.) as previously described by Daniel et al. (12-15) with the following modifications. ConA covalently bound to 4.0% agarose beads type V-A (ConA-agarose; Sigma) in 50% aqueous suspension (10.0 ml) was poured into a plastic disposable 15-ml Econo-Column chromatography column (Bio-Rad, Rockville Center, N.Y.) and allowed to settle. The column was washed with 10 column volumes (50 ml) of phosphate-buffered saline (pH 7.5) (PBS). The H20 extract (3.0 ml) was allowed to flow through the column by gravity flow. The column was then washed with 2 column volumes (10 ml) of PBS. The column was eluted in step fashion with 2 column volumes (10 ml) of 1% D-glucose in PBS followed by 2 column volumes (10 ml) of 10% D-glucose in PBS. The 1% D-glucose, 10% D-glucose, and column flowthrough fractions were dialyzed in a cellulose membrane (3,500molecular-weight cutoff) against four changes of 4 liters of HPLC-grade H20. The blanks consisted of the ConA-agarose column flowthrough fraction and buffer. The dialyzed samples were then lyophilized to dryness. Purification of MIF-A3. The acetonitrile extract from 1.5 g of M. paratuberculosis was resuspended in 1.0 ml of H20, and the aqueous portion was transferred to a tube containing 9.0 ml of acetone. After 15 min at room temperature, the sample was centrifuged at 4,400 x g for 30 min at 4°C. The supernatant was transferred to another tube and dried under nitrogen gas. The dried sample was resuspended in 1.0 ml of H20 and examined for purity by sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) and thinlayer chromatography (TLC). This purified fraction was named M4-inhibitory factor A3 (MIF-A3). M+-Candida albicans killing assay. Bovine peripheral blood mononuclear cells were obtained from peripheral

INFECT. IMMUN.

blood by Ficoll-Hypaque density gradient centrifugation with Histopaque 1083 (Sigma). Cells were washed three times with PBS (10 ml) and then resuspended in RPMI 1640 buffered with sodium bicarbonate containing 100 ng of Escherichia coli LPS per ml, 100 ng of phorbol myristate acetate per ml, 100 ,ug of ampicillin (Sigma) per ml, and 100 ,g of streptomycin (Sigma) per ml to a final concentration of approximately 50,000 M4 per 300 RI. The percentage of M4~ was determined by counting and nonspecific esterase staining. The M4 solution (300 ,u) was added to the wells of a 96-well plastic tissue culture plate and allowed to incubate for 16 h at 37°C with 5% CO2. The stimulants LPS and phorbol myristate acetate were added to directly activate the monocyte-M4 as well as induce the release of gamma interferon by contaminating lymphocytes, resulting in hyperstimulated monocyte-M4. Wells were then washed three times with 300 pI of fresh medium (without LPS and phorbol myristate acetate) to remove the lymphocytes and traces of LPS and phorbol myristate acetate. The plates were allowed to incubate as described above for 8 h. Aliquots of 20 jig (25-,ul volume; final concentration of 100 pug/ml) of each fraction and approximately 10,000 C. albicans cells diluted in 200 RI of RPMI 1640 buffered with sodium bicarbonate containing 100 ,ug of ampicillin and 100 ,ug of streptomycin were simultaneously added to each well and allowed to incubate for 16 h at 37°C with 5% CO2- Candida yeast cells were cultured overnight (24 h) in modified Winge medium (0.2% glucose and 0.3% yeast extract) containing 100 ,ug of ampicillin and 100 ,ug of streptomycin. All samples were tested with a minimum of three replicates. The viability of MX was also assessed by exclusion of trypan blue 2 h after the addition of each fraction; viability was consistently greater than 95%. Plates were centrifuged at 450 x g for 10 min at 4°C, and the supernatants were discarded. A 0.02% SDS aqueous solution (300 RI) was added to each well to lyse the M+, and then the pellets were resuspended. Dilutions of 102 and 103 were prepared in 1.5-ml microcentrifuge tubes with sterile H20 as the diluent. Aliquots (50 ,ul) of each dilution were plated on culture plates (60 by 15 mm) containing Sabouraud dextrose medium (Difco). After 16 h of incubation at 37°C, the CFU of C. albicans were enumerated by a method modified from that of Decker et al. (16). Statistical analysis (Tukey's Studentized range test) of the obtained surviving CFU was done to analyze the activities of the suspected active fractions versus those of controls. Cytotoxicity of fractions. The cytotoxicities of the active fractions were determined with a rapid colorimetric assay utilizing the tetrazolium salt MTT as previously described (35) with the following modifications. The human M+-like cell line U937 (2.5 x 104 and 5.0 x 104 cells) that had previously been incubated for 24 h at 37°C and 5% CO2 with various concentrations of the active fractions (MIF-A3, H20-ConA-agarose-1% D-glucose, H20-ConA-agarose10% D-glucose) was used. Plates were read on an enzymelinked immunosorbent assay microplate reader with a test wavelength of 570 nm and a reference wavelength of 650 nm. PBS served as the negative control. Candida growth assay. The effects of the fractions on the growth of C. albicans were tested by using a rapid colorimetric assay utilizing the tetrazolium salt MTT as previously described (35) with the following modifications. Plates (96 wells) containing 1.0 x 104 Candida cells per well were incubated for 16 h at 37°C and 5% CO2 with various concentrations of the active fractions (MIF-A3, H2O-ConA-

agarose-1% D-glucose, H20-ConA-agarose-10% D-glucose). Plates were read on an enzyme-linked immunosorbent

MYCOBACTERIAL GLYCOLIPIDS INHIBIT ACTIVATED M4

VOL. 61, 1992

Candida-Macrophage Killing Assay CFU, 103 0

50

100

150

200

Blank

Hexane

Methylene C12 Acetonitrile Methanol H20

FIG. 1. Inhibition of killing of C. albicans by M+ with M. paratuberculosis fractions obtained by MSPD as shown by surviving CFU (n = 6). Multiple assays produced similar results.

assay microplate reader with a test wavelength of 570 nm and a reference wavelength of 650 nm. Sterile water and 1% dextrose in PBS served as negative and positive controls, respectively. SDS-PAGE and TLC. Molecular weight standards for SDS-PAGE were commercially obtained (low-molecularweight electrophoresis calibration kit; Pharmacia LKB Biotechnology, Piscataway, N.J.). Electrophoresis was performed with a 4.5% acrylamide stacking gel and a 12.0% acrylamide separating gel at pH 6.6 and 8.8, respectively (11, 25, 41). Aliquots of each sample (100 to 200 ,ul) were lyophilized or precipitated in acetone and then solubilized by heating for 3 min at 100°C in 30 pul of sample buffer (10% glycerol, 2% SDS, 5% 2-mercaptoethanol, 0.125 Tris-HCl [pH 6.8], 0.01% bromophenol blue). The gels were stained with silver by the techniques of Wray et al. (48) for proteins and of Tsai and Frasch (44) for carbohydrates. Molecular masses were determined by comparison with protein standards of known molecular mass: phosphorylase b, 94 kDa; albumin, 67 kDa; ovalbumin, 43 kDa; carbonic anhydrase, 30 kDa; trypsin inhibitor, 20.1 kDa; and ot-lactalbumin, 14.4 kDa. TLC of the active fractions was performed as previously described (23) with the following modifications. Samples of the MIF-A3 fraction (30 ,ug), the H20-ConA-agarose-1% D-glucose fraction (30 ,ug), and the H20-ConA-agarose-10% D-glucose fraction (30 jig) were spotted on a silica gel-coated glass TLC plate (catalog no. JT7009-04; J.T. Baker Co., Phillipsburg, N.J.) and allowed to dry for 30 min. The mobile phase used was 50% chloroform-40% methanol-10% H20.

RESULTS

The purpose of this study was to isolate and characterize potential virulence factors from M. paratuberculosis 18 that may allow the organism to evade destruction and to replicate within M4). Several crude fractions were obtained from M. paratuberculosis by the MSPD technique. A marked increase in Candida survival was observed when hyperstimulated bovine M, had been exposed to the acetonitrile and H20 fractions (Fig. 1). Activity was occasionally but not

3

consistently found in the methylene chloride fractions; therefore, only the acetonitrile and H20 fractions were used in subsequent tests. No difference in the level of phagocytosis (determined by direct microscopic evaluation of new methylene blue-stained wells) resulting from any of the treatments was observed. Levels of endotoxin in these fractions, samples of water, media, stock solutions, and MSPD-derived fractions (except methanol extract) were less than 0.125 endotoxin unit per ml. The methanol extract apparently contained an inhibitor to the assay, since the positive LPS control containing this fraction did not gel. No LPS was detected in the methanol extract with an assay for 2-keto-3-deoxyoctonic acid, an integral component of LPS (24). The results of various enzymatic, chemical, and antibody treatments of the active fractions are summarized in Table 1. A significant reduction in M4-inhibitory activity resulted from treatment of the acetonitrile fraction with protease K, anti-LAM antibody, and sodium metaperiodate from extraction with chloroform, whereas similar treatment of the H20 fraction showed considerable reduction in activity after sodium metaperiodate treatment and chloroform extraction. After separation of the acetonitrile extract by centrifugal filtration into >30-kDa and