A number of immunological tests (enzyme-linked immunosorbent assays) have been developed by several authors for the serodiagnosis of tuberculosis. These tests use antigens derived from reference strains of Mycobacterium tuberculosis H37Rv, M. tuberculosis H37Ra, or even from M. bovis, BCG. In most laboratories, reference strains have been maintained for many years by periodic subculturing in selective media, such as Lowenstein-Jensen’s medium. Changes in structural and antigenic characteristics of the microorganism as a result of repeated subculturing is, therefore, a real possibility. For this reason, we compared patterns of antigen secretion by several recently isolated wild strains of M. tuberculosis with the patterns of antigen secretion displayed by a reference strain of M. tuberculosis H37Rv. This reference strain has been maintained in our laboratory for approximately 10 years. Several colleagues have noted changes, proportional to the time spent in storage, in the protein pattern of mycobacterial extracts prepared by sonication. These changes were observed even when samples were prepared as frozen-stored aliquots that were not subjected to repeated freezing and thawing. One explanation for the observed changes is that they were caused by the activity of proteolytic enzymes extracted from the microorganisms together with other structural proteins. This has prompted the addition of protease inhibitors during the preparation of the sonicates. The work presented here is from studies carried out using the supernatant (filtrate) of 7-week-old cultures of M. tuberculosis. This approach was chosen for two reasons: first, to avoid working with the microorganisms themselves; and second, it was reasoned that the material collected from the 7-week culture filtrates should be maximally autohydrolyzed and, therefore, should remain stable thereafter. Comparisons made between the reference and the wild strains of M. tuberculosis were based on the following two parameters: detection of protein patterns observed in polyacrylamide gels stained with silver stain as described by Merril et al., and recognition of antigens by sera derived from active pulmonary tuberculosis patients and from healthy (non-tuberculous) individuals.

Unless otherwise indicated, chemicals were purchased from Sigma Chemical Co. (St. Louis, MO, USA). Buffer solutions included borate-buffer solution, BBS (0.3092 g of boric acid, and 0.4768 g of sodium tetraborate per liter of 0.15 M NaCl, pH 8.4); phosphate buffer solution, PBS (0.01 M phosphate in 0.15 M NaCl, pH 7.4), and phosphate-citrate buffer, PCB (0.1 M citric acid in 0.2 M disodium phosphate, pH 5.0).

Ten isolates were obtained from sputum or urine specimens from 10 patients with clinically diagnosed tuberculosis. Seven of the 10 isolates were confirmed to be authentic M. tuberculosis on the basis of morphological examination, staining, and biochemical criteria. Of these seven isolates, six were from pulmonary tuberculosis patients and one was from a renal tuberculosis patient. The M. tuberculosis reference strain was obtained approximately 10 years ago from the Trudeau Mycobacterial Collection (M. tuberculosis H37Rv TMC 102) and since then has been maintained in our laboratory by periodically subculturing in Lowenstein-Jensen’s medium.

Both the reference and the wild strains of M. tuberculosis were grown in the protein-free Proskawer-Becker medium, as modified by Youmans (PBY medium). Cultures were maintained without shaking at 37°C for 7 weeks; at this time, confluent superficial growth was always observed. The confluent bacillary mass was separated by decanting and filtration (through filter paper). Microorganisms were then removed from the filtrate by 0.22-?m Millipore filtration before being reduced to a minimum volume by dialysis against sucrose crystals as follows: 500–1000 mL of the filtrates were collected into cellulose dialysis sacs (12,000 Da cut-off). The sacs were then covered with sucrose crystals and in order to facilitate the draining of extracted fluid, the concentration system was placed in a tilted position. Fresh sucrose crystals were periodically added until no more liquid was extracted. This technique resulted in the reduction of the filtrate to about 1/10 of the original volume. The concentrated filtrate was then dialyzed against saline-borate solution (BBS, pH 8.4) to eliminate sucrose, metabolites, and other small sized components of the medium. Finally, the filtrate was recovered and its protein content measured by Lowry’s method and adjusted to 1.0 mg/mL. The sample was then Millipore-sterilized and divided into 1.0-mL aliquots before being frozen for additional use.

Analysis of the protein patterns in the filtrates of each wild and reference M. tuberculosis strain was carried out by polyacrylamide gel electrophoresis (SDS-PAGE) under denaturing and reducing conditions using Laemmli buffer. Twenty to 30 ?L of each filtrate (1.0-mg protein/mL) were fractionated in 12%-polyacrylamide-SDS gels. Ten microliters (1 mg protein/mL) of prestained molecular weight protein markers (SDS-7B, Sigma) were simultaneously run as reference proteins. The gels were run for 4 to 5 h at 20 mA per gel until the tracking dye had migrated 10 cm. Gels were either fixed for silver staining or were transferred to Immobilon P membranes (polyvinylidene difluoride, PVDF, IPVH 000 10, Millipore Corporation, Bedford, MA, USA) for Western blot analysis.

The reactivity of 40 serum samples from healthy (non-tuberculous) individuals and 40 serum samples taken from patients with pulmonary tuberculosis, against secreted protein antigens of M. tuberculosis H37Rv was measured by the ELISA test. Medical personnel at the Hospital General de México in Mexico City obtained the sera after receiving informed consent from the patients. A conventional ELISA was carried out as follows:

(a) wells were coated with 2 ?g of protein from culture filtrate in 100 ?L of borate buffer solution, BBS, pH 8.4;

(b) wells were washed three times with 150 ?L phosphate buffer solution, pH 7.4 containing 0.5% Tween 20 (PBST);

(c) blocking was carried out with 150 ?L of 2% skim milk in BBS, for 30 min at 37°C;

(d) wells were incubated with 100 ?L of sera diluted 1:100 (tuberculous and normal sera) and 1:20 (normal sera) in 2% skim milk in PBS for l h at 37°C;

(e) wells were again washed three times with 150 ?L of PBST;

(f) incubated for 1 h at 37°C with 100 ?L of horseradish peroxidase-labeled goat antibody to human immunoglobulins (1:2000 in 2% skim milk in PBS) (dilutions of serum and peroxidase labeled antibodies were chosen from a previous chessboard titration of the reagents);

(g) again wells were washed three times with 150 ?L of PBST and once with PBS;

(h) 100 ?L of a mixture containing 4.0 mg of o-phenylene diamine (Sigma P1526) and 10 ?L of 30% hydrogen peroxide (Sigma H1009) in 10 mL of phosphate citrate buffer pH 5.0 was added to each well;

(i) 20 ?L of 4N sulfuric acid was added to arrest the reaction; and

(j) absorbency readings were taken at 492 nm in an ELISA reader.

Sera from healthy or tuberculous individuals were selected for their high readings and then used to detect mycobacterial antigens secreted by the several wild and reference strains by Western blot analysis.