Purification and Properties of Tuberculin-active Protein from Mycobacterium tuberculosis

When Mycobacterium tuberculosis was grown on Sauton medium, intracellular tuberculin-active protein was produced. This product was purified by chromatography on DEAE-cellulose and Sephadex G-200 and was obtained in crystalline form. The crystals were plates, somewhat irregular in shape, about 50 mum in length and 25 mum in width. The yield corresponded to a 1.5% over-all recovery of total protein. Ultracentrifugal analysis showed only one major component with a calculated molecular weight of 9700. Sedimentation velocity analysis gave a sedimentation coefficient at 20 degrees (see article) of 1.73 S. The estimated specific activities of tuberculin-active protein were 6.33 times 10-9 tuberculin units per mg of protein-nitrogen for sensitized guinea pigs and 6.33 times 10-11 tuberculin units per mg of protein-nitrogen for humans. This is the most potent tuberculin-active protein that has yet been obtained.

separated into at least three distinct prot.ein bands by zone electrophoresis.
The material was purified by Seibert (4-7) and called purified protein derivative.
PPDl is widely used and is presently the most popular tuberculin-active protein preparation. I have been able to obtain intracellular tuberculin-active protein from the cells of tubercle bacilli instead of from culture fluid.
It exhibits a higher tuberculin activity than PPD (8), with 10 pg equivalent to 1 TU, as compared to 20 ng of PPD required for a positive Mantoux skin test of a sensitized person. The present paper describes methods for the purification of tuberculin-active protein from human-type tubercle bacilli, Aoyama/B strain, and from the avirulent bovine-type BCG strain and presents some of the properties of these proteins. EXPERIMENTAL  as a standard. Protein-N was calculated on the assumption that the protein measured contained 15.8y0 nitrogen. In some cases determination of protein and nucleic acid was made by the spectrophotometric method of Warburg and Christian (10). Amino Acid Analysis-Amino acids in acid hydrolysates of tuberculin-active protein (200 to 700 rg) were determined as described bj Moore et al. (11) and Spackman et al. (12) with an amino acid analyzer.
Hydrolysis was performed in 6 M of HCl at 110" for 24,48, and 72 hours. Cystine and methionine were determined as cysteic acid and methionine sulfone as described by Hirs (13). Identification of the NH*-terminal residues of tuberculin-active protein by Edman degradation (14,15) was carried out by thin layer chromatography of material absorbed on filter paper as described by . Carbohydrate Analysis-Determinations of glucosamine and galactosamine in hydrolysates used for amino acid analysis were made with an amino acid analyzer.
For determination of carbohydrates, neutral sugars, and amino sugars, residues were converted into silyl derivatives, which were resolved by gas-liquid chromatography (17). The method used was adapted from that of Sweeley and Walker (18).
Ultracentrifugal Analysis-Sedimentation velocity analysis and measurements of the refractive increment were made with a Beckman-Spinco model E ultracentrifuge equipped with schlieren and interference optics as described by Schachman (19) and Svedberg and Pederson (20).
Determination of Molecular Weight-The molecular weight of tuberculin-active protein was determined by the method of Archibald (21) in a synthetic boundary cell in order to create a boundary away from the meniscus. A solution of tuberculin-active protein (3 mg per ml) dialyzed against 0.1 M Tris was put in a synthetic boundary cell, and the solvent was layered on top in the centrifuge. The boundary was left to diffuse until it formed a bell-shaped curve in the schlieren optics, and a photograph was taken. This was done at full speed, and an S form was obtained from a series of pictures (log,r versus 1 was plotted-slope is w?s). The area under the curve was measured.
The experiments were repeated with 0.1 M Tris in the synthetic boundary cell, again using a double sector cell with the same solvent in the second half. Pictures were taken at the same bar angle (60") as used in the synthetic boundary run. The area under the curve between the meniscus and place where the curve reaches the base-line was measured. This was substracted from the area under the curve in the synthetic boundary run, and designated area A in square centimeters. The height (H) in centimeters of the line from the base-line at the meniscus was measured.
The molecular weight, M, was cal-culated from the equation, Hem M(l -0p)p)3 -= r+A RT in which m is the magnification of the optical system in a horizontal direction, r is the distance of the meniscus from the center of rotation in cm, ii is the partial specific volume of the protein, p is the density of the solution, w is the angular speed of the rotor (radians per a), R is the gas constant, and T the absolute temperature.

Tuberculosis
Aoyama/B and BCG-The human-type tubercle bacillus, strain Aoyama/B, obtained from the Department of Tuberculosis, National Institute of Health in Tokyo, was inoculated in veal mfusion broth and, when sufficiently grown, transferred to Sauton's synthetic medium (22) in l-liter bottles containing about 200 ml each. After incubation at 37.5" for 6 to 8 weeks, the cultures were shaken and heated in a sterilizer at 120" for 30 min. The bacilli were filtered from the culture fluid through a Buchner funnel, and stored at -15O. The avirulent bovine-type strain BCG, a stock culture of the National Institute of Health in Tokyo, was cultured in the same manner, harvested after 2 and 6 weeks without heating, and was then lyophilized.

Tuberculin
Test-Tests were carried out by Dr. T. Morohashi and his colleagues at the Department of Tuberculosis, National Institute of Health, Tokyo. The PBS buffer (0.1 ml) containing various doses of tuberculin-active protein was administered intracutaneously in the ventral side of guinea pigs that had been immunized with complete Freund's adjuvant containing heat-killed Aoyama/B strain 6 weeks before. Within 24 to 48 hours, the diameters of the erythematous areas were determined with an average of six tests at different ventral sites. Some assays were carried out in humans on the volar surface of the forearm in the same manner as the Mantoux intracutaneous test.

Purification of Tuberculin-Active
Protein from Mycobacterium Tuberculosis Aoyama/B S@e l-Cells (319 g) from Mycobacterium tuberculosis Aoyama/B were prepared at the Japan BCG Laboratory as described above.

2557
Stage S-An acetone-dried powder was prepared from the heated cells (319 g) by stirring them with 23 times (v/w) the volume of acetone at -15" and drying the solution in a hood at room temperature.
The weight of the powder was 50 g which is 15.7% of the wet weight of the starting material.
Stage S-Acetone-dried powder (50 g) from Stage 2 was ground in a mortar with a lcw ionic strength buffer, 0.5 mM Tris, pH 7.0. The powder was centrifuged at 10,000 X, g for 15 min and the supernatant was collected.
The supernatant, consisting of a crude soluble extract (1,311 mg of total protein), was treated with 2% streptomycin sulfate (v/w) and centrifuged at 16,000 x g for 15 min.
Stage &-The supernatant from Stage 3 (1,319 mg of protein, 1,000 ml) was digested with 4 mg each of DNase and RNase at 30" for 60 min. The reaction mixture, saturated with ammonium sulfate, was collected, dissolved in a small amount of 1 mM Tris, and dialyzed against the same buffer overnight at 4". Assay of the resulting solution showed that the specific activity of tuberculin-active protein had increased 4-fold (Table I, Stage 4).
Stage 6: First DEAE-cellulose Column-The resulting solution (283 mg, 10 ml) was added to a DEAE-cellulose column (100 x 1 cm) and eluted at room temperature with 1 mM Tris buffer containing 1 M NaCl. Fractions (2 ml) were collected every minute, and the active fractions were combined (tubes 9 to 15). The protein recovered was 5.67% of the total (Fig. 1, left). The material at this stage contained 2.0% nucleic acid, showing an E280,260 value of 1.11. The specific activity of tuberculin-active protein was increased about 14-fold at this stage.
Stage 6': Second DEAE-cellulose Column-The protein solution from Stage 5 (74.6 mg) was brought to 55% saturation with solid (NH&SO4 at pH 7.0. The mixture was stirred for 30 min at 4" and centrifuged for 15 min at 14,000 x g. The concentration of (NH&S04 in the supernatant was raised to 96.6% saturation, and the mixture was stirred and centrifuged as before. The precipitate was dissolved in a small volume (5 ml) of 10 mM Tris buffer.
The resulting solution was rechromatographed on a Left, first DEAE-cellulose column (100 X 1 cm). Right, second DEAE-cellulose column (106 X 1 cm). The columns were run as described in the text. The samples were 283 mg of total protein (left) and 74.6 mg of total protein (right).
DEAE-cellulose column (100 x 1 cm). Elution of protein from the column was carried out with a sodium chloride gradient obtained by using 500 ml of 1 mM Tris buffer in a mixing chamber with a magnetic stirrer and an equal volume of the same buffer containing 0.1 M sodium chloride in a reservoir. Fractions (2 ml each) were collected every minute (Fig. 1, right). Virtually all of the activity was found in Fractions 4 to 9, with peak activity in Fraction 5. The active protein was precipitated by the addition of (NH&SO4 to saturation. The precipitate was separated by centrifugation and dialyzed against 10 mM Tris buffer for 12 hours. The protein recovery was 2.8%. The material at this stage contained 0.5% nucleic acid and showed an E280,2e0 value of 1.55. The specific activity of tuberculin-active protein in the resulting solution was about 2.0 times that at Stage 5. Stage 7: Sephadez G-200 Column-The solution of protein prepared as described above was added to a Sephadex G-200 column (100 X 1 cm). Elution was carried out with 10 mM Tris buffer and fractions (0.1 ml) were collected every minute at room temperature (Fig. 2). Fractions 28 to 50 were combined. The over-all recovery of protein in the combined fractions was l.S%, and its E~~O/~C,O value (1.8) indicated that it was free of nucleic acid. The resulting solution contained 4.44 x log TU per mg of protein-N and the over-all number of units recovered was 82%. At this stage the active protein was pure enough for crystallization.
Stage 8: First Crystallization-The resulting solution from Stage 7 was dialyzed against 10 mM 'Iris buffer at 4" overnight. Solid (NH&S04 was added to the dialyzed protein solution (5 ml), to 85% saturation, and kept in a cold room (4") for 2 weeks in tightly sealed lo-ml Pyrex bottles The white precipitate was separated from the solution, cold 50% acetone was gently added without mixing, and the upper acetone-rich layer was discarded. The resulting white precipitate was again treated with cold acetone, dissolved in 50% cold acetone, and left at room temperature. Immediately after the above treatment, a silvery sheen began to appear on the surface of the mixture, and the first crystals of the tuberculin-active protein began to form. The crystals were slightly soluble in water and soluble in a salt solution such as 1 mm or higher concentration of Tris buffer, or a solution of 50% saturated (NH&Sod. The crystals were somewhat irregularly shaped plates, about 50 pm in length and 25 pm in width (Fig. 3). They contained 6.33 X log TU per mg of protein-N as assayed in guinea pigs and 6.33 x 10" TU per mg of protein-N 2559 in humans.
The yield corresponded to 1.5% over-all recovery of total protein and 97% over-all recovery of tuberculin act,ivity. Recrystallization was carried out in essentially the same manner as in Stage 8. The specific activity, shape, and size of the crystals did not change significantly on recrystallization. The crystals were in 5OCr, saturated (NH&SC+ pH 7.0, in 10 rnhf Tris buffer. When kept under the conditions at 2-4" and at a concctnration of IO mg per ml, tuberculin-active protein was stable for months. Purification of tuberculin-active protein from Xycobacteriunl tuberculosis Aoyama/R is summarized in Table 1.

Purijkation of Tuberculin-active Protein From Avirulent Bovine-type BCG Strain
This procedure was the same as in the purification of tubcrculin-active protein from human-type Aoyama/I% strain except that intact cells from the avirulent bovine-type living 13CG strain were used as starting material in Stage 1. The purification of tuberculin-active protein from this source is summarized in Table II.

Ultracentrifugal
Analysis of Tuberculin-mtive Protein The results of ultracentrifugal analysis of crystalline tuberculin-active protein from Aoyama/B and BCG indicate only one major component.
Sedimentation velocity showed that the coefficient at 20" (s&~) in 0.1 M Tris buffer was 1.73 S for tuberculin-active protein from Aoyamajls strain (Fig. 5). As shown in Fig. 4, such a low S value was very difficult to measure for the estimation of the molecular weight and required the use of synthetic boundary cells in order to create a boundary away from the meniscus.
An analysis of the material at the meniscus gave 9700 for the molecular weight using a partial specific volume of 0.734 (23). Neither value was appreciably dependent on concentration.
Electrophoretic Analysis Electrophoretic analysis was carried out, on a l'erkin-Elmer electrophoresis apparatus using a 2-ml cell and a 8.9 mM Kaz- Aoyama/B. The pictures were taken after 10 min (right) and 30 min (left) of sedimentation at 59,780 rpm. The bar angle was 60", the temperature 17.5", solvent Tris-HCl, pH 7.0, made 0.1 M in 1 mM EDTA.
Concentration of tuberculin-active nrotein was 3.0 me ner ml. The direction of sedimentation is from left to right. I-   HPOh-1 mM EDTA mixture in 0.17 M NaCl. Fig. 6 shows schlieren patterns of a 1% solution of tuberculin-active protein from Aoyama/B in the descending limb of the Tiselus cell. The schlieren pattern remains a sharp, single peak after 120 min.

Amino Acid Composition of Tuberculin-active Protein
The amino acid compositions of the tuberculin-active protein from two strains of Mycobacterium tuberculosis are given in Table  III.
No amino sugars nor carbohydrate was detected in the protein of either strain.
The amino acid composition of the Aoyama/B strain protein is based on a molecular weight of 9700 (1). The molecular weight of BCG strain is not known, so the amino acid composition of this protein is based on the assumption of 1 mol of histidine per mol of protein.
The amino acid compositions of the tuberculin-active protein from the two strains are very similar except that the protein of the BCG strain has less tyrosine, methionine, and half-cystine, and more arginine than the Aoyama/B protein.
No free sulfhydryl groups were detected by the Ellman procedure (24). Oxidation of cystine and methionine at -10" for 246 hours with performic acid yielded 92'% each of cysteic acid and methionine sulfone. The KHzterminal analysis by the Edman procedure (14, 15) revealed a single NHz-terminal residue (arginine), confirms that tuberculinactive protein has been isolated in a homogenous form.

Tuberculin Activity of Tuberculin-active Protein
Guinea Pigs-Various doses of tuberculin-active protein from Aoyama/B strain or HCG dissolved in 0.1 ml PBS buffer, were injcctcd into guinea pigs intracut'aneously.
As little as 1.0 ng was sufficient to produce a tuberculin-positive skin test in guinea pigs immunized by heat-killed Aoyama/B strain (Fig. 7). This amount of tuberculin-active protein is equivalent to 1 TU, v;hich is equal to 20 ng of PPD (Fig. 7). No skin reaction was observed in guinea pigs without sensit.ization by Aoyama/B strain or BCG.
There is no difference in tuberculin activity between tuberculin-active protein from BCG and that from Aoyama/B.  Humans-An injection of 0.1 ml of PBS buffer containing 10 pg of tuberculin-active protein from Aoyama/B was sufficient to produce a positive skin test in a man immunized with tubercle bacilli Aoyama/B strain or BCG (Fig. 8). The above amount of tuberculin-active protein is equivalent to 1 TU, which indicates that its activity is about 200 times higher than that of PPD.
The sensitivity of humans to tuberculin-active protein is about 100 times higher than that of guinea pigs.

Intracellular
tuberculin-active protein might be closely associated with the cytoplasmic membrane.
Pollock (25) discovered two immunologically distinct types of /3-lactamase in cultures of Bacillus cereus. Over 90% of the enzyme is of one type (a-type) and is found almost exclusively in the culture supernatant fluid. The other immunological type of p-lactamase, termed y-type, is invariably bound to the cell. There is evidence (26) that the T-type enzyme, in contrast to the small fraction of cr-type enzyme absorbed in the cells, is firmly attached to the cytoplasmic membrane.
According to one report (27) part of the y-type appears to be bound to ribosomes.
The enzyme is largely obtained in solution when the cells are extensively disintegrated (24), and it can be released from isolated membrane preparation by sonic disintegration or by treatment with organic solvents (25). Tuberculin-active protein and y-type enzyme resemble each other in their localization and method of preparation.
Al- though the extent of purification of tuberculin-active protein from the avirulent bovine-type BCG strain is much higher than from human-type Aoyama/B strain, this might be due to the heat used to kill the cells of the latter at the first stage of purification.
The two preparations showed equal antigenic activity against guinea pigs that had been immunized with killed Aoyama/B strain. As a source of tuberculin-active protein, human-type Aoyama/B strain and bovine-type BCG are essentially the same.
An assessment of the significance of the apparent similarities and differences between tuberculin-active protein obtained by this method and PPD by Seibert (28,29) derives from the sources of the starting materials.
Whether intracellular tuberculin-active protein is the precursor of extracellular PPD by passage through the cell walls or cytoplasmic membranes remains to be ascertained.
As shown in Fig. 7, a parallel relationship is obtainable between activities in both tuberculin-active protein and PPD when they are administered subcutaneously in guinea pigs sensitized with killed Aoyama/B strain or living BCG. It is postulated that tuberculin-active protein has greater activity in a sensitized human subject than laboratory animals such as guinea pigs. When tuberculin-active protein is injected in humans, only 10 pg, the equivalent to 20 ng of PPD (1 TU), is required to get a positive Mantoux skin test. About 100 times as much tuberculin-active protein, 1000 pg, is required to get a positive skin test with sensitized guinea pigs. PPD shows a wider range of tuberculin-activity, and 20 ng may be administered either to humans or guinea pigs with the same aflinity. Another apparent difference between tuberculin-active protein and PPD is the relationship between the protein concentration