March, 19511 WITH Lactobacillzds Zactis DORNER BY THE PLATE METHOD 141 The Assay of Vitamin B12 Part IV* The Microbiological Estimation with Lactobacillus Zeichmannii 313 by the Turbidimetric Procedure BY W. B. EMERY, K. A. LEES AND J. P. R. TOOTILL (Presented at the meeting of the Biological Methods Gvoup on Tuesday, M a y 23rd, 1950) Details are presented of a microbiological tube assay for vitamin B,, with Lactobacillus Zeichmannii 313 as test organism. The procedure employs materials that are readily available in this country. Statistical analyses of a (3 + 3) assay and a standard response curve are given and show that the method is sufficiently sensitive and accurate for routine use. ‘Two papers1 92 from these laboratories describe work on factors affecting the microbiological techniques used for assaying vitamin B12; the cup-plate method is described in detail else- where in this issue.3 In this paper we give the corresponding details for ,the so-called “tube assay’’ with a turbidimetric procedure for measuring the growth of the micro-organism.The conditions used are based on the work described in one of the papers cited above., As was also found by Cuthbertson, Pegler and Lloyd,3 it has been our experience that Lactobacillzcs Zactis Dornel-2 9 5 gives a very irregular growth response to graded doses of vitamin B12. We therefore carried out further work on the use of LactobncilZzcs k i c h m a m i i 313 in the vitamin B,, assay, this organism having been recommended by Skeggs, Huff, Wright and Bosshardt6; other workers have already reported on its use.’ As already briefly reported,* our assay method depends on the inability of L.leichmannii 313 to synthesise vitamin B,, under defined conditions, so that there is a direct relationship between the growth of this organism and the coiicentration of vitamin B,, in the test medium over a certain range. Attempts to devise an assay suitable for routine use, employing the medium of Snell, Icitay and M ~ N u t t , ~ were not successful, the chief difficulty being the production of a dense precipitate during the autoclaving of the medium. This medium was modified by the omission of enzymatic casein digest and reduction of the level of salts B. To prevent precipitation in the tube and consequent high blanks, the medium was preheated in bulk and filtered hot.Hoffman et aZ.7 reported that thioglycollic acid increased the growth response of L. leiclzmannii to vitamin B12. This was thought to be due to protection of the vitamin * For particulars of earlier papers in this series (txro not in The Analyst), see reference list, pp. 145-146.142 EMERY, LEES AND TOOTHILL: ASSAY OF VITAMIN B12 WITH [Vol. 76 during the autoclave cycle. Our own work on the effect of reducing agents and oxygenation,2 as well as the recent work of Koditschek, Hendlin and Woodrufflo on the nutrition of L. Lactis Dorner, indicates that vitamin B1, functions as an anti-oxidant for the test organism, and consequently there is a direct relationship between the E, of the medium and the vitamin B,, requirement. We have been able to show the mathematical relationship between the response of L.Zeichmannii and the depth of medium in the assay tubes.2 TUBE ASSAY TECHNIQUE PREPARATION OF MEDIUM- Solutions A and B are prepared according to the details in Table I and mixed as required, because it has been found that the complete medium does not keep well. TABLE I ASSAY MEDIUM Solution A (double strength)- Sodium acetate, A.R. .. .. . . . . .. Sodium citrate, A.R. .. . . .. .. .. .. .. .. ::it' ;}Snell and Wright" I ' - . . . .. Acid hvdrolvsate of casein (based on total solids) . . Tween-80 (purified for T.B.'test) Sodium oleate . . .. .. .. Glucose . . .. .. * . .. Glass-distilled water to . . .. . . OY Crill 10 (Croda Works, Goole) L-Cystine .. .. .. .. pH adjusted to . . .. .. . * Solution B (forty times single strength)- Adenine, guanine and uracil (of each) DL-Tryptophan .. .. .. .. p-Aminobenzoic acid . . . . .. Asparagine . . .. .. .. Calcium-D-pantothenate . . . . Aneurine hydrochloride . . .. Pyridoxal . . . . .. .. Pyridoxamine . . . . .. .. Nicotinic acid . . .. .. .. Riboflavine . . .. . . .. Biotin . . .. . . .. .. Folic acid .. .. . . . . Glass-distilled water to . . .. .. .. .. .. .. .. .. . . .. . . .. .. . . .. .. .. . . . . .. .. .. .. .. .. .. .. .. .. * . . . . . .. .. .. .. .. . . .. .. .. .. . . .. .. . . .. .. .. .. .. * . .. ,. * . .. .. .. .. .. * . .. .. .. .. .. The pH of solution A is adjusted to 6.8. The solution is then heated to boiling and I t is stored in a Solution B is stored in a refrigerator a t pH 3.0 to 4.0 and may be used for a period of filtered with the aid of kieselguhr and its pH is then re-adjusted to 6.8.refrigerator and may be used over a period of 5 to 6 days. 10 to 14 days. PREPARATION OF ACID HYDROLYSATE OF CASEIN- Five hundred grams of "vitamin-free" casein (Glaxo Laboratories Ltd.) are heated under reffux with 2-5 litres of 6 N hydrochloric acid for 9 hours. The hydrolysate is con- centrated to a syrup in vucuo and diluted to 4.5 litres with glass-distilled water; after the pH has been adjusted to 3.0 with solid sodium hydroxide, the mixture is filtered and treated twice with 50 g of Sutcliffe Speakman No. 5 charcoal, each absorption involving stirring for 30 minutes. The final solution is diluted to 51 litres. This procedure yields a hydrolysate containing 1.0 to 1-25 per cent.w/v of total nitrogen. We normally preserve the hydrolysate by freeze-drying to a dry powder; alternatively it may be stored in a refrigerator under toluene. Each batch of acid-hydrolysed casein is tested over a suitable range of concentration. We have recently found casein hydrolysate supplied by Allen and Hanbury Ltd. to be satisfactory .March, 19511 Lactobacillus Zeichmannii 313 BY TURBIDIMETRIC PROCEDURE 143 CLARIFIED TOMATO JUICE- Commercial canned tomato juice is filtered, with the aid of kieselguhr, through a Whatman No. 1 filter-paper supported on a No. 54 paper. The filtrate is a clear pale straw colour. The filtered tomato juice is stored in the refrigerator and is used over a period of 3 days only. Some brands of commercial tomato juice do not contain the factor required by L.Zeich- mannii. A large quantity of a tested brand is therefore bought and stored in the cold room: we have found “Crest” brand tomato juice to be satisfactory. MAINTENANCE OF TEST ORGANISM AND PREPARATION OF INOCULUM- I.. Zeichmannii is maintained by daily transfer in liquid medium (yeast extract, 1 per cent.; tomato juice, 5 per cent.; separated milk to 100 per cent.). More than three hundred transfers have not resulted in any detectable alteration in the sensitivity of the organism. TABLE I1 ANALYSIS OF DAILY RESPONSE CURVE Dose, mpgpertube 0.23 0.35 0.53 0.79 1.19 1-78 2-67 4.00 Totals Responses Rack I . . Sub Totals Responses Rack I1 .. Sub Totals Responses Rack 111.. Sub Totals Totals . . Orthogonal Coefficients Means ... . 0.15 . . 0.14 . . 0.29 . . 0.19 . . ’ 0.19 . . 0.38 . . 0.17 . . 0.16 . . 0.33 . . 1-00 { z . . 0.167 0.28 0.20 0.48 0.23 0.25 0.48 0.23 0.23 0.46 1-42 - 5 +l 0.237 0.36 0.36 0.72 0.34 0.37 0.71 0.33 0.38 0-7 1 2.14 -3 -3 0.357 0.51 0.68 0.53 0.63 1.04 1.31 0.54 0.64 0.45 0.61 0.99 1.25 0.57 0.65 0.49 0.68 1.06 1.33 3.09 3.89 -1 $1 - 5 -5 0.515 0.648 0.85 0.80 1-65 0.7 1 0.85 1.56 0-94 0.83 1.77 4.98 +3 -3 0.830 1.06 0.9 1 1.97 0.09 0.93 2.02 1.09 1.12 2.21 6.20 +5 +1 1.033 1.21 1.22 2.43 9.89 1-29 1.24 2.53 9.92 1.18 1.24 2-42 10.29 Regression coefficients 7-38 30.10 +0.627,083 + 7 + 77.88 +O-077,262 + 7 + 10.02 + 0*009,94O 1.230 Regression values 0-156 0.251 0.365 0.500 0.655 0.829 1.023 1.237 ANALYSIS OF VARIANCE FOR HOMOGENEITY- Sum of Mean Variance Source D.F.squares square ratio Significance Doses . . . . .. 7 6.12 1,62 Racks . . . . . . 2 0.006,20 0*003,100 1.251 none Interaction . . . . 14 0*029,87 0*002,134 0.861 none Error . , . . . . 24 0*059,50 0.002,479 Total . . . . . . 47 6*217,19 .2NALYSIS OF VARIANCE FOR REGRESSION- Sum of Mean Variance D.F. squares square ratio Significance Quadratic . . . . 1 0*099,60 0*099,60 40.178 very Linear . . .. . . 1 6*017,16 6.017,16 2427.656) high Deviation .. . . 5 @*004,86 0*000,97 0.391 none Doses . . . . . . 7 6*121,62 Standard error = 0.0498 The inoculum is prepared by two daily transfers in basal assay medium enriched with purified liver extract to give a final vitamin B,, concentration of 0.015 pg-per ml. Volumes of 0.5 ml are subcultured into 10-ml volumes of enriched basal medium in 6 inch x 2 inch tubes and the cultures are incubated at 37” C for 18 to 20 hours.The cells from the second144 [Vol. 76 transfer in inoculum medium are centrifuged, washed once in sterile physiological saline and re-suspended in sterile saline to Burroughs ’Wellcome opacity 4. This suspension is diluted 1 to 8 in a mixture of equal volumes of sterile autoclaved physiological saline and clarified tomato juice, the latter having been previously sterilised by filtration through a sintered-glass filter; 0-5 ml of this inoculum is used for each assay tube. EMERY, LEES AND TOOTHILL: ASS.4Y OF VITAMIN B12 WITH PREPARATION AND DISTRIBUTION OF TUBE MEDIUM- Solutions A and B are mixed in the following proportions and 84ml are introduced For 500 assay tubes the following are mixe(1- into each of several 6 inch x 2 inch standard tubes by means of an Ayling filler.Solution A . . .. .. .. .. . . . . 2500ml Solution B . . .. .. .. . . .. .. 125ml Glass-distilled water to a . .. .. . . . . 4250ml By the additions of 1 ml of suitably diluted test sample and 0.5 ml of inoculum, the volume is increased to 10 ml and the medium reduced to “single strength.” The pH is re-adjusted to 6.8 if necessary. 0~0002 u 00 0 2 04 0.6 0-8 1.0 1.2 1.4 Response Fig. 1. A typical I-esponse curve The assay tubes are covered by aluminium caps and autoclaved a t 151b pressure for 10 minutes, the pressure in the autoclave being released quickly and reduced to atmospheric pressure in 2 minutes. We have found that slight differences in the autoclave cycle pro- foundly influence the slope of the growth resportse curve.Increasing the autoclave cycle to 30 minutes decreased the accuracy of the test by reducing the range of growth available for the daily response curve. We have not fount1 that increased autoclaving will produce a medium in which L. Zeichmannii no longer requires vitamin B,,, as reported by Shawl2 for L. Zactis Dorner. The daily reference curve is obtained by six-tube replicates, each of eight doses of pure crystalline vitamin B12, logarithmically spaced in the range 0.00016 and 0.0027 pg per tube. A few blank tubes are set up as controls each day to indicate what growth, if any, occurs in the absence of vitamin BI2. The analysis of a daily response curve is given in Table 11; it shows that the standard deviation of a reading is approximately 0.05. Test samples are diluted according to the estimated potency and sterilised by filtration through sintered-glass filters.One millilitre of each dilution is added to the assay tubes and followed by 06ml of inoculum. The tubes are shaken to ensure thorough mixing of the inoculum and incubated in an incubator room at 37” C for 17 hours. Small laboratory type incubators have not been found suitable. After incubation, the growth is stopped by the addition of 2 drops of 40 per cent. formaldehyde solution to each tube or by steaming; the turbidities are read on a Spekker photo-electric absorptiometer in the normal way. Dr. W. F. J. Cuthbertson (personal communication) has now found it possible to replace the Tween 80 by Estax 36 (Watford Chemical Company) and the tomato juice filtrate by a A typical response curve is shown in Fig.1.March, 19511 LactobaciZZus Zeichmannai 313 BY TURBIDIMETRIC PROCEDURE 145 mixture of 250 pg of fumaric acid, 250 pg of sodium ethyloxalacetate* and 100 pg of DL-alanine per millilitre of single strength medium. These constituents may be incorporated in the “forty times strength” vitamin solution. Dr. Cuthbertson has also reported that certain batches of inoculum medium, i.e., basal medium enriched with vitamin BIZ, that failed to produce satisfactory growth after storage for a few days were greatly improved by the addition ANALYSIS OF (3 + 3) ASSAY Responses of standard (-.-h-.--..-- -7 0~0008 0.0012 0.0018 0.93 1.12 1.38 0.83 1.15 1.29 0.54 0.98 1-30 0.56 0.74 1.28 1 -06 1.12 1.28 0.86 1-05 1.32 (pg per tube) I Responses of unknown A > 4/9u 2/3u U 0.83 1-15 1.33 0.87 1-08 1.33 0-96 0.95 1.31 0.92 1.13 1.34 0.80 1.09 1.20 1.03 1.05 1.32 ‘Totals .. . . . . . . 5-10 6.16 7-85 5.41 6-49 7-83 38.84 Sample difference . . . . -1 -1 -1 +1 -!-1 i- 1 + 0.62 Linear regression . . . . --1 0 +1 -1 0 -t 1 +5*17 Departure from parallelism . . + 1 0 -1 -1 0 $1 - 0.33 Combined curvature . . . . $1 -2 +1 f l -2 i- 1 + 0.89 Opposed curvature . . . . -1 4- 2 -1 +1 - 2 i-1 -0.37 AXALYSIS OF VARIANCE- Source D.F. Sample difference . . . . 1 Linear regression . . . . 1 Departure from parallelism . . 1 Opposed curvature . . .. 1 Combined curvature . . . . 1 Error . . .. . . .. 30 Sum of squares 0.01068 1.11370 0.00454 0~01100 0~00190 0.33 173 Mean Variance square ratio Siplicance 0-01068 0.97 not Significant 1.11370 100.70 P < 0.001 0.00464 0.41 not significant 0.01 100 0.99 not significant 0.00 190 0.17 not signscant 0.01 106 Total .. . . .. . . 35 1.43755 Potency ratio: 1.07 Fiducial limits: (P = 0.95) 0.93 to 1.22 (93 to 122 per cent.) of small amounts of thioglycollic acid or ascorbic acid ; this suggested that an accumulation of peroxides had occurred. Excessive amounts of these reducing agents, however, led to the production of unsatisfactory inocula and consequently boiling and cooling of the inoculum tubes immediately before inoculation is recommended. We wish to acknowledge valuable technical assistance by Miss F. M. Ord and Mr.D. G. Goodinson. REFERENCES 1. 2. 3. 4. 5. 6. 7. Cuthbertson, W. F. J., and Lloyd, J. T., J. gew. Microbid., 1951, 5, 416. Lees, K. A., and Tootill, J. P. R., in pr~pnration. Cuthbertson, W. F. J., Pegler, H. F,, and Lloyd, J. T., Anrtlyst, 1951, 76, 133. Shorb, M. S., J. Biol. Chem., 1947, 169, 455. - , Science, 1948, 107, 396. Skeggs, H. R., Huff, J. W., Wright, 1,. D., and Bosshardt, D. K., J. Biol. Chem., 1948, 176, 1469. Hoffmann, C. E., Stokstad, E. L. R., Hutchings, H. L., Dornbush, A. C., and Jukes, T. H., Ibid., 1949, 181, 635. * Commercial brands have not been found suitable and laboratory preparation according to Wislicenus, -4 nnalen, 1888, 246, 31 5, is recommended.146 COATES, HARRISON AND KON: THE CHICK ASSAY OF 8. 9. 10. 11. 12. Lees, K. A., and Emery, W. B., Biochem. J., Pvoc. Biochem. Soc., 1949, 45, ii. Snell, E. E., Kitay, E., and McNutt, W. S., J . Bid. Chem., 1948, 175, 473. Koditschek, L. K., Hendlin, D., and Woodruff, H. B., Ibid., 1949, 179, 1093. Snell, E. E., and Wright, L. D., Ibid., 1941, 139, 675. Shaw, G., Nature, 1949, 164, 186. [Vol. 76 NOTES-References 1, 2 and 3 are to Parts I, I1 and 111 of this series. Since this paper was written the following two papers on the microbiological assay of vitamin B,, Thompson, H. T., Dietrich, L. S., and Elvehjem, C. A., J . Bid. Chem., 1950, 184, 175. Skeggs, H. R., Hepple, H. M., Valentik, I<. A., Huff, J. W., and Wright, L. D., Ibid., 1950, 184, have appeared- 211. GLAXO LABORATORIES LTD. SEFTON PARK STOKE POGES, BUCKS.