696 HARRISON, LEES AND WOOD: THE ASSAY [Vol. 76 The Assay of Vitamin B12 Part VI Microbiological Estimation with a Mutant of Escherichia coZi by the Plate Method BY ELEANOR HARRISON, K. A. LEES AND F. WOOD A method for the microbiological cup-plate assay of vitamin B,, with a mutant of Escherichia coli as test organism is presented. The assay is an advance on other methods previously published as Parts I11 and IV of this series because (a) a simple chemically defined medium is used, (b) changes in the EH of the test medium have little effect on the response, (c) the zones of exhibition are reproducible and well defined in character and (d) the degree of specificity and general freedom from interfering effects and inexplicable variations is compatible with a large output of assays on a routine basis.The inoculum consists of a small volume of a culture grown overnight in peptone water, and the assay plates ma-y be incubated at any temperature between 27" and 37°C. Zones of exhibition can be obtained with solutions containing one milli- microgram of vitamin B,, per ml. The dose response line is rectilinear over the range 0.005 to 5.0 pg per ml and hence a (2 + 2) assay design can normally be used. The effects of times of incubation and of standing before incubation have been investigated. A streptomycin-resistant strain of the test organism has been developed for the purpose of assaying directly the vitamin B,, content of Streptomyces griseus fermentation samples. Thymidine does not interfere with the assay. Standard errors of 0.14 to 0.23mm per zone are obtained, 0.18mm being the normal error encountered in assays on Petri dishes or large glass plates.METHODS for the microbiological assay of vitamin B,, in tubes or on plates with strains of These micro-aerophilic bacteria respond uniformly to graded doses of vitamin B!, within a narrow EH range only, so that variations in oxygen tension of test media profoundly influence the growth of the test organism in the presence of sub-maximal quantities of the vitamin. In view of these difficulties, a plate assay with L. Zeichmannii 313 as test organism was developed (unpublished), but it suffered from the disadvantage that certain of the vitamin B,, group of factors proved difficult to assay; in particular vitamin B12c containing the -NO, grouping gave poorly defined zones.Lactobacillus Zeichmannii or L. Zactis Dorner as test organisms have been published. 1 7 2 9 3 9 4,6,6,7Dec., 19511 OF VITAMIN BIZ, PART VI 697 Normally it is not practicable in this country to use wholly synthetic media and the consequent need for some complex constituent, such as casein hydrolysate, adds to the difficulties of preparing reproducible standard media. The receipt of a mutant strain of Escherichia coZi responsive to vitamin B,, and methionine encouraged us to determine its nutritional requirements and the conditions under which it would yield zones of exhibition in an agar medium. E. coZi strains are not as a rule nutrition- ally exigent and are classed as aerobic in their oxygen requirements; this suggested that it should be possible to devise for the assay a simple medium in which the organism would not be sensitive to small changes of oxidation - reduction potential.Preliminary work indicated that inocula grown in peptone water yielded better zones than those grown in synthetic media and that inocula grown under conditions of aerobiosis, i.e., shaken, were more vigorous than those grown in static culture. Initial work was per- formed with a washed inoculum, but later it was found that an unwashed inoculum gave excellent assay plates with no perceptible growth over the background of the plate. Zones of exhibition were obtained on assay plates by using a previously described synthetic medium* enriched with a trace element solution first described by Beadle and T a t ~ m .~ Preliminary details of the assay method have already been published.lO The test media used in these lactobacillus assays are complex and expensive. ASSAY METHOD TEST ORGANISM- A mutant of Escherichia coli, kindly supplied by Dr. Bernard Davis of New York, was used. This organism appeared as Gram negative rods measuring about 2 p by 0.5 p when grown on the synthetic agar slant medium detailed below. A culture of the mutant strain has been deposited with the National Collection of Industrial Bacteria (Chemical Research Laboratory, Teddington, Middlesex), whence sub-cultures can be obtained. For the purpose of determining the vitamin B,, content of StqXomyces gviseus fermenta- tion liquors containing both streptomycin and vitamin B,,, we have developed a streptomycin- resistant strain of the test organism.Our first attempt to induce resistance in E . coZi was not successful. It happened that a vitamin B,, “non-requiring” variant was produced. This proved difficult to separate from the parent culture. Further attempts were made to induce resistance by steps of increasing dosage, the organism being plated out after each transfer. On checking the purity of an isolate by using the colony as a parent inoculum for a few assay plates, increased resistance was recorded, until finally a strain resistant to 20,000 units of streptomycin per ml was obtained. Consequently, growth in the inoculum medium and within the zones on the assay plates is slightly less dense than that of the parent strain, which we prefer to use for the assay of samples that do not contain streptomycin. This resistant variant grew rather less vigorously than the parent mutant. MAINTENANCE OF CULTURE- the composition shown in Table I.Stock cultures of the mutant strain are maintained on a synthetic agar slant medium of TABLE I SYNTHETIC AGAR SLAKT MEDIUM Acid hydrolysed casein (Allen and Hanbury) . . .. 6.0 g Di-potassium hydrogen phosphate . . .. .. . . 0.2 g . . .. . . 5mg Ferrous sulphate (7H,O) . . .. Magnesium sulphate (7H,O) . . . . . . .. .. 0.2 g L-Asparagine . . . . . . . . . . .. . . 0.15 g Distilled water to . . .. .. .. . . 700ml Dissolve the constituents with gentle heating in the order shown in Table I, with a few drops of hydrochloric acid to dissolve the asparagine, and adjust the pH of the solution to 7.2 After cooling add 2.0 g of glycerol and 20.0 g of agar (Davis Gelatine Ltd.).Make up the volume to 1 litre with distilled water and steam to dissolve the 0.1; boil and filter.698 HARRISON, LEES AND WOOD: THE ASSAY [Vol. 76 agar, re-adjust the pH to 7.2 0.1 and add 400 pg of crystalline vitamin B1, (or its equiv- alent). After thorough mixing, dispense the medium in 10-ml amounts in 6-inch by $-inch resistance glass tubes, then plug and autoclave them for 15 minutes at 15 lb pressure. After sterilisation, “slope” the tubes and store them subsequently at 4” C. The master culture is stored at 4” C; from this, sub-master cultures are prepared weekly. Transfers are made daily into peptone water inoculum medium (see below) and incubated a t 37” C for twenty-four hours.INOCULUM MEDIUM- Dissolve 20 g of peptone (Evans Medical Supplies Ltd.) and 5 g of sodium chloride in distilled water and make up to 2 litres. Adjust the pH to 7.2 & 0.1 and dispense the medium into flasks of such dimensions that 100 ml in each occupy a depth of between three and five millimetres.* Many samples of peptone have proved unsuita.ble, but Evans and Difco brands both give satisfactory growth. As SAY MEDI u M- Autoclave the medium for 20 minutes at 15 lb pressure. Prepare the stock solutions shown in Table 11. TABLE I:[ ASSAY MEDIUM Ammonium chloride . .) . . .. .. Ammonium nitrate . . .. .. .. Distilled water to . . .. .. .. Stock solution A- Stock solution B- Potassium dihydrogen phosphate . . . . Di-potassium hydrogen phosphate .. . . Distilled water to . . .. . . . . Stock solution C- Sodium borate . . .. .. . . . . Copper sulphate (5HzO) . . . . . . Ferrous sulphate (7HzO) . . . . .. Manganese chloride (4H20) . . .. .. Zinc sulphate (7H,O) . . .. . . . . Ammonium molybdate .. I . . . Distilled water to . . .. .. . . .. .. .. . . . . . . . . . . . . . . . . . . .. .. .. .. .. .. .. .. .. .. .. .. .. .. 200 g 80 g 2 litres 40 g 120 g 2 litres 0-094 g 0.250 g 0.540 g 0.460 g 4.900 g 1 litre 0.020 g Add 36 per cent. WIW hydrochloric acid dropwise until the solution is clear (approximately 0.5 ml is required) From these stock solutions, which can be stored in a refrigerator, the agar plate medium is prepared in the following manner. Mix in the order given- Stock solution A .. .. .. .. . ... Stock solution R .. . . . . .. .. I . Stock solution C .. .. .. .. .. . . 5 per cent. w/v magnesium sulphate (7HzO) . . * . Asparagine . . . . .. . . .. . . .. 0.1 per cent. calcium chloride . . . . . . .. Adjust the pH to . . .. . . .. . . . . Distilled water to . . .. .. .. . . .. 250 ml 250 ml 0.5 ml 10.0 ml 5.0 ml 4.6 litres 7.2 2 0.1 7.5 g Dispense 120-ml volumes into 250-ml conical flasks ;%rid add 2 g of agar (Davis) to each. Plug the flasks and autoclave them at 15 lb pressure for 15 minutes. The sterile medium is stored at 4” C and has a storage life of four to six weeks. USE OF ASSAY MEDIUM- When required for use, the flasks are steamed for thirty to forty minutes to melt the The glucose is sterilised separately agar, and then 10 ml of glucose solution are added to each.* Penicillin flasks (“Glaxo” type) are eminently suitable and can be got from Messrs. Townson and Mercer.Dec., 19511 OF VITAMIN Biz. PART VI 699 as a 13 per cent. w/v solution in an autoclave for 15 minutes at 15 lb pressure. medium is allowed to stand in a water-bath a t a temperature of 46" to 48" C. DETERMINATIOS OF INOCULUM VOLUME- The turbidity of the 24-hour inoculum culture in peptone water is determined with a Spekker photo-electric absorptiometer. An inoculum reference graph has been prepared relating the density of inoculum to the volume required to yield zones of the best character and a dose - response slope giving the greatest difference in response for standard dose increments. It is assumed that the variation from day to day in the ratio of viable to total bacterial count is insignificant.The molten PREPARATION OF ASSAY PLATES- The flasks of molten plate medium are inoculated at a temperature not exceeding 48" C and poured into plates immediately. Into each 9-cm Petri dish 12.5 ml of inoculated medium is poured. High temperatures or delay in pouring the plates a t this stage will materially reduce the initial seeding rate, reduce the density of the growth in the zone of exhibition and thus inflate the mean zone size. With the aid of a guide frame and a stainless steel borer of about 8 mm diameter, six cups are punched into each agar plate, the discs of agar being removed by means of a vacuum attachment. REFERENCE STANDARD- We use sealed vials each containing approximately 50 pg of vitamin B,, freeze-dried on about 40 mg of sodium chloride; the exact vitamin-B,, content is stated on the label in terms of anhydrous vitamin B,,, having been determined by ultra-violet spectrophotometric measurement of E:k at 207 mp.The contents of the vial of standard are diluted appropriately with distilled water and stored in the refrigerator away from direct light. The solutions are not kept for more than four days. The storage life of these prepared plates is about three days. PREPARATION OF SAMPLE DILUTIOW- In attempting to evaluate the potency of pure crystalline vitamin B12a, B,,, and BIZd in terms of vitamin B,,, we experienced difficulty through lack of consistency in the assay from day to day and through lack of parallelism between the curves for the crystalline factors on certain days.Consequently accurate evaluation of the potency of the crystalline vitamin B,, factors was not possible, although the results indicated an order of activity very similar to that of vitamin B,, referred to an anhydrous weight basis or the colour of the solution determined on the Spekker photo-electric absorptiometer. Having been informed by Cuthbertsonll that lack of parallelism and inconsistencies in the assay could be eliminated by treating the assay samples with potassium cyanide solu- tion, we now use the following dilution technique. The assay samples are accurately diluted with distilled water to about 20 pg per ml in graduated flasks and to 1 ml of this solution in a graduated flask one drop of 1 per cent. potassium cyanide solution is added, the solution being allowed to stand for fifteen minutes and then further diluted with distilled water as required.Three drops of test or standard solution are now added to the appropriate cup from a pipette in which the dropper has a platinum tube, of 0.0365-inch external and 0.0295-inch internal diameter and about 10 mm long, fused into the tip at an angle of about 130" to take up a position normal to the Petri dish during delivery. The diameter of the zones of exhibi- tion can be measured after 18 hours' incubation, but this time is not critical and little change in zone size or appearance is detectable on further incubation. For the purpose of determining the amount of vitamin B,, in crude sources of the vitamin such as feeding materials, liberation of the combined vitamin can be achieved by steaming the preparation in solution for 30 minutes a t a pH value between 5.0 and 6.0.The liberation is facilitated by the presence of potassium cyanide, as recommended by Wijmenga, Veer and Lens.,, In view of the recent reports13 of factors active in this assay and also for EzgZena gracilis, but not active for L. Zeichmanii 313, it is advisable to include in any preliminary investigations of crude sources of vitamin B,, a qualitative control by paper chromatography to demonstrate the absence of factors other than vitamin B12. It is advisable to use for this700 HARRISON, LEES AND WOOD: THE ASSAY [Vol. 76 chromatography a developing solvent containing a trace of potassium cyanide to ensure that the vitamin B,, is not converted to slower moving factors on the strip, as reported by Woodruff and Foster.14 ASSAY DESIGN- The dose - response curve is rectilinear for doses between 5.0 and 0.005 pg of vitamin B,, per ml applied to the plate.Any of the conventional microbiological plate assay designs can be used and results computed from the mean slopes of standard and sample dilutions by reference to suitably We normally use a (2 + 2) assay with a dilution ratio of 1 to 10. 12 x 12 LATIN SQUARE DESIGN Totals . . H E 13.9 18.7 F I 18.7 19.2 L K 13-7 13.7 B C 19.8 14.0 19.0 13.5 13.8 19.3 E H 19.3 13-7 A F 19-3 19.3 D G 13.4 13.6 I D 18-4 13.7 G A 13.8 18.8 K B 12.9 18.0 76.0 75-5 J L C J G B D 13.2 18.8 14.0 E D G 19.2 13.8 13.8 13.8 19.8 20.0 D I F 13.8 19.7 19.7 K C R 13.7 13.5 10.5 B A L 19.2 19.2 14-0 A K B 19.5 13.9 19.0 19.3 13.4 13.4 H E K 13-6 18.8 13.5 L F J 13.4 19.0 19.0 I L C 18.8 13-3 13-4 F H A 18.0 12.9 18.3 C J I J G H 75.5 76.1 77.6 I L F 19.6 13.7 18.7 H C L 13.7 13.7 13.7 G A D 14.0 19.5 13.6 L G A 13.8 13.7 19.0 A B H 19-5 19.4 13.7 D H K 13.9 13.9 13.9 13.5 19.0 19.0 E K C 19.3 13.6 13.7 19.1 18.8 19.1 K E G 13-4 18-9 14.0 B F E 18.7 18.9 19.1 18.4 13.2 18.4 C I J F J B J B I 76.9 76.3 75.9 C J A K 13.8 19.0 18.7 13.4 A K J B 18.7 13.5 18.7 18.7 B F H E 19.3 19.3 13.7 19.3 K H E J 13.6 13.7 19.0 19.0 F G D I 19.0 13.8 13.7 19.2 E I F G 19.3 19.3 19.5 13.7 D L G F 13.0 13-4 13.5 18.9 I B L D 19.2 19.2 13.7 13-4 L C I A 13.7 13.0 18.9 18-9 H A B C 13.3 18.9 18.8 13.3 J D K H 18.7 13.2 13.0 13.0 G E C L 12.9 18.0 13.0 12.9 74-5 74.3 74.2 73.7 Totals 75.5 75.4 79.7 78.8 77.5 79.0 75.7 76.8 74.4 74.1 72.7 66.9 906.5 1.which prepared tables or nomograms. gives the maximum likelihood solution of any assay. To use the nomogram, abstract from a (2 + 2) assay the values of (RT, - RS,) and (RS, - RT,), where RS,, RS,, RT, and RT, refer to the mean zone sizes of the responses of low and high standard and low and high doses of sample dilution respectively. By joining these two values on the ordinates of the nomo- gram with a perspex ruler, the potency of the sample referred to the highest dose is obtained. This value is then multiplied by the dilution that has been made on the sample. In view of the slope of the dose - response curve (1.2 to 1.7 mm for 100 per cent. increase in dose) we consider it necessary to employ replication on ten plates if the zone diameters are being measured only to the nearest 0.5 mm.We prefer to read the zone diameters to the nearest 0.1 mm with replication on five plates, For this measurement we use slide calipers modified by the addition of needle tips by which zones are measured on the surface of the agar, as shown in Fig. 2. This method of reading avoids errors due to parallax. For the accurate evaluation of special samples or the critical examination of interfering agents on the assay we prefer to use larger assay plates to obtain more precise comparisons. These large assay plates are made from plate glass on the lines previously reported by Brownlee, Loraine and Stephens16 for penicillin and streptomycin assays.The use of these large plates for microbiological assay can only be justified if a statistical design is used to obviate effects similar to those reported by Brownlee et al. We prefer to use the nomogram shown in Fig.RT,- RS , RS,-RT, Fig. I . Nomogram for (2 + 2) assay, 10 to 1 dilutionsRT,- RS , RS,-RT, Fig. I . Nomogram for (2 + 2) assay, 10 to 1 dilutionsDec., 19511 OF VITAMIN BIZ. PART VI 701 The precision of the assay is demonstrated in Tables I11 to V. A 12 x 12 Latin square assay was performed to investigate the effect of the pH and of the cyanide content of sample dilutions on the zone size. Table I11 shows the plan of the assay; the letters indicate the experimental treatment being assayed. The letters were randomised against the treatments as shown in Table IV.The solutions were introduced into the cups in horizontal row order beginning H, E, G, . . . and finishing E, C , L. The use of the design has improved the TABLE IV TREATMENT TABLE All dilutions performed in 0-05 M phosphate buffer Treatment and dose, p g per ml r > A Vitamin B,, KCN pH 5-0 0.1 0.0 0.01 0.0 pH 5.0 0.1 0.3 0-0 1 0.03 pH 5.0 0.1 15.0 0.01 1.5 0.0 1 0.0 pH 8-0 0.1 0.3 0.01 0.03 pH 8.0 0.1 15.0 0.01 1.5 pH 8.0 0.1 0.0 Randomised plating code E C J L B D A G I H Total of twelve responses 228.1 162.5 228.0 162.8 228.0 162.7 228.3 163.4 229.7 162.5 F 228-1 K 162.1 TABLE V ANALYSIS OF VARIANCE Source D.F. Mean square .. 11 0.9960 Rows .. .. .. Columns . . .. .. .. 11 0.1 146 .. .. .. 1 .. .. 1 Level of CN .... .. 1 Effect of CN . . .. .. 1 .. .. 1 Zone size Effect of pH . . .. .. 1 Slope Effect of pH . . 1 Effe'c't of 6N 1 Zone size pH interaction with Slope pH interaction with 1 1 1 Error .. 110 Level of CN 1086-7656 0-0059 0.0460 0-0153 0.0126 0.0201 0.0201 0.00 17 0,0376 0.0767 0.0176 0.0548 Significance highly significant significant highly significant not significant not significant not significant not significant not significant not significant not significant not significant not significant not significant not significant accuracy of the experiment by eliminating significant extraneous variation in the rows and the columns. Hence it can be concluded that the only effect that is significant is the slope of the dose - response line, which demonstrates that neither pH nor cyanide within the ranges examined have any interfering effects in the E .coli assay. The analysis recorded in Table V showed that- The variation in responses within the rows and columns was highly significant. This effect has been eliminated in the analysis. There was no observable difference in zone size or slope of the dose - response line between pH 6.0 and pH 8.0. Potassium cyanide, at either level, had no observable effect on zone size or slope. There were no interactions between cyanide and pH treatments. The standard error of the response zones was 0.23 mm.702 HARRISON, LEES AND WOOD: THE ASSAY [Vol. 76 It must be noted that the tolerance with respect to pH has been demonstrated in the presence of 0-05 M phosphate buffer and no information was obtained as to the pH of the solutions after diffusion into the agar medium.REPRODUCIBILITY OF ASSAYS- The results for several samples, each assayed on two days, are shown in Table VI. Six cups were punched on each plate and the equivalent of two (2 + 2) assays were performed on five plates, i.e., two samples are referred to one set of dilutions of standard. It follows, therefore, that a total of one hundred and thirty plates has been used to assay the twenty-six samples. TABLE 'VI REPRODUCIBILITY OF ASSAYS Sample 1 2 3 4 5 6 7 8 9 10 11 12 13 Day 1 5.6 2-8 20.0 15.5 19.5 17.5 19.5 0.22 2.2 2.3 2-2 1-9 7.8 Day 2 5.6 3.1 2 1.0 17.0 21.0 17.6 18.5 0.18 2.0 2.0 2.2 1.7 7.0 Reported value 5.5 3.0 20.5 16.5 20.0 17.5 19.0 0.20 2.1 2.1 2.2 1.8 7.5 Sample 14 15 16 17 18 19 20 21 22 23 24 25 26 Day 1 6.0 5.2 27.0 27-0 27.0 23.6 21.6 22.6 23.5 4.6 0.25 19.5 1.60 Day 2 5.0 4-7 25.0 26-0 26-5 21.0 18.0 19.5 26.0 4-8 0-2 1 20.8 1.42 Reported value 6.5 5.0 26.0 26.5 27.0 22.0 19.5 21-0 24.0 4.7 0.23 20.0 1-46 THE EFFECT OF PRE-COOLING ON THE DIFFUSION OF VITAMIN BIZ THROUGH THE AGAR- It was thought that, if diffusion of vitamin B,, could be allowed to take place before growth of the test organism, not only would the zone sizes be larger, but that the differences between zone sizes given by graded doses of vitamin B,, might also be increased.In this manner the gradient of the response line would be increased and the assay made more precise. Accordingly an experiment was performed in which the growth of the organism was delayed by refrigeration for various times up to three hours, while graded doses of vitamin B,, were diffusing through the agar.Following this refrigeration the plates were incubated in the normal manner and the resultant zones measured to the nearest 0.1 mm. Five equally spaced times of pre-refrigeration, 0, 45, 90, 135 and 180 minutes, were combined in all possible ways with five logarithmically spaced doses of vitamin B,,, 0.2, 0.1, 0.05, 0.025 and 0.0125 pg per ml. The resulting twenty-five treatments were replicated six times by using in all thirty five-hole plates. In order that the results should be as precise as possible the twenty-five treatments were arranged in an incomplete block design. Table VII shows the analysis of variance of this experiment ; the estimated and calculated regression values are shown in Table VIII.TABLE 'VII ANALYSIS OF VARIANCE Source D.F. Mean square Linear effect of time . . .. .. 1 119.347 Linear effect of dose . . .. .. 1 541.417 Interaction of time and dose.. .. 1 1.980 Deviation . . .. .. .. 21 0.082 Error .. .. .. . . .. 96 0.060 Significance significant highly significant not significant not significant Intra-plate error : + 0-229 Inter-plate error : 5 0.42 1 Error of weighted results : + 0-244 iO.100 Error of estimates (mean of 6) :Dec., 19511 OF VITAMIN Biz. PART VI 703 TABLE VIII ESTIMATED AND CALCULATED REGRESSION VALUES Dose Time 0 45 90 135 0-2 Estimate 25.1 1 24-27 23-33 22-80 Regression 25.09 24.34 23-60 22.85 0.1 Estimate 23.69 22.95 22.28 21-65 Regression 23.63 22.94 22.25 21-57 0.05 Estimate 22.35 21.64 21.00 20.22 Regression 22.17 21.54 20.91 20.28 0.025 Estimate 20.75 19.99 19-50 19.11 Regression 20.71 19.57 19.57 18.99 0.0125 Estimate 19.27 18.70 18.16 17.54 Regression 19.26 18.74 18-22 17.71 Standard error of estimate : & 0.1 mm.This experiment showed that- 180 22-18 22.10 21.02 20-88 19.52 19.65 18.48 18-42 17.32 17.19 (i) For any given dose, the zone size increased linearly with increasing diffusion time in the refrigerator at a mean rate of 0.84 mm per hour. (G) For any given pre-refrigeration time the zone size is linearly related to the logarith- mic dose and therefore suitable for assay. The mean slope over all pre-refrigeration times is 1.34 mm per doubling dilution. (iii) The slope of the linear response of (ii) increases with increasing diffusion time in the refrigerator at a mean rate of 0.0766 mm per doubling dilution per hour. Thus the slope without any pre-diffusion, corresponding to existing methods of assay, is 1-23 mm per doubling dilution, while after 3 hours pre-diffusion this is improved to 1.46 mm per doubling dilution.This corresponds to a reduction of a 10 per cent. assay error to about 7 per cent. (iv) In view of the low experimental error of 20.1 mm per estimate, the above effects are known to be linear with considerable accuracy. A further experiment on similar lines to that described above, making use of refrigeration times greater than seven hours, was performed. In this the slope of the dose - response line increased at a mean rate of 0.09 mm per doubling dilution per hour, resulting in a final slope of 2.0 mm for a doubling dilution after the full refrigeration time.The zones of exhibition resulting from the vitamin B,, solutions that have been permitted to diffuse for seven hours before incubation are indistinguishable in character from zones resulting from normal treatment. For routine assay we do not find it convenient to employ a pre-diffusion period, but when critical comparisons or determination of effects are required we use such a pre-diffusion treat- ment (up to seven hours if possible) to improve the precision of the determinations. The effect of incubating the assay plates before adding the samples to the cups was examined in a similar manner. The results, together with those from the above refrigeration experiment, are shown in Fig.3, where it can be seen that the slope of the dose - response curve decreases at a mean rate of 0.11 mm per hour per doubling dilution. DISCUSSION OF RESULTS The optimal conditions for the growth of the inoculum, together with other unpublished results from the use of the E. coli mutant in tube assays, indicated that a highly aerated medium favours rapid dense growth of this strain. The thin agar layer in the assay plates therefore results in sharply defined dense zones of exhibition. The character of these zones may be seen in Fig. 2. It is important that the growth on the assay plates should be preceded by growth of an inoculum in shallow layers of medium in order to obtain zones of the best character superimposed on a clear transparent background.As previously reported,1° zones of exhibition can be obtained by using ammonium dihydrogen phosphate, potassium chloride, magnesium sulphate and dextrose as the only constituents of the medium, but the growth within the zones is not so dense as with the recommended complete medium, and we prefer this for routine assay.704 HARRISON, LEES AND WOOD: THE ASSAY [Vol. 76 We have found that incubation of the assay plates may be performed either at 27" C or 37" C, a small increase in the slope of the dose - response line sometimes being observed at 27" C. With the Lactobacillus Zaclis Dorner plate assay,, vitamin BlZc often gives assay vdues three to four times as high as those obtained by tube methods or physico-chemical procedures. We have also found the same effect with L.Zeichmannii 313 plate assays, in which the zones are characterised by a certain fuzziness, but are entirely different in appearance from the zones given by desoxyribosides. With the E . col% mutant plate assay, however, activity approximately equivalent on a weight basis to that of the vitamin B,, reference standard is obtained for vitamin B1*, and there is no unusual appearance of the zones. 30.0 - 25.0 - E & L s 5 2 20.0 - 4 c 15.0 - Normal assay treatment it.. plating out of dilutions I I I I 1 7 6 5 4 3 2 I 0 I 2 3 12.0 Time, hours Fig. 3. The effects of refrigeration and incubation of the assay plates on the assay of vitamin B,, by the Escherichia coli mutant The standard error of the (12 x 12) Latin square assay quoted above is 0.23 mm per zone, rather higher than we normally experience. With (11 x 5) or (13 x 4) cyclic incomplete block designs it is usual to find experimental errors of 0.14 to 0.20 mm per zone.This high error of 0.23 mm we attribute to the unusually large design employed, for no evaluation of the effect of time of plating out on the slope of the dose - response line could be made and any consequent error is inflating the assay error. This effect of time of plating out on the slope of the dose - response line is not very great when smaller designs are being used. The time taken to plate out the (12 x 12) design was forty-four minutes. Methionine, as noted by Davis,16 gives zones of exhibition with the E. coZi mutant. A solution of 1 mg per ml gave a zone of approximately 35 mm, and dilutions of 100 and 10 pg per ml also stimulated growth, but the zones were indistinct and not measurable.A full account of the interfering effects of various substances in this assay method has been published by Cuthbertson, Pegler, Quadling and Herbert.,' Consideration of Fig. 3 shows the effect of deviation from normal assay treatment in plating out and incubation of the assay plates. The graphs serve to emphasize that the assay plates must be retained in the refrigerator until required for plating out and must during this operation be as cold as possible. For accurate assay comparisons an obvious gain in precision will be obtained if pre- refrigeration of the assay plate between plating out and incubation be performed. This mayDec., 19511 OF VITAMIN BIZ.PART VI 705 conveniently be done when the large assay plates are used, since they are easily handled as individual units. The analysis of the (12 x 12) Latin square experiment in Table V demonstrates the wide tolerance of the assay method to differences in pH and cyanide content of the test samples. Neither alteration in the pH values of the dilutions of the vitamin B,, reference standard from 5.0 to 8.0 nor an alteration in the cyanide content from three times to one hundred and fifty times the vitamin B,, concentration significantly affected the response. We are much indebted to Dr. Bernard Davis of the Public Health Laboratory, Cornell University Medical College, New Uork City, not only for having originally called Dr. E. Lester Smith’s attention to the mutant, but also for supplying the culture that enabled us to carry out the work reported here. We wish to acknowledge the help given by our colleagues Miss A. Gibson and Mr. J. P. R. Tootill for developing the streptomycin resistant strain and for the statistical design and analysis, respectively. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. REFERENCES Caswell, M. C., Koditschek, L. K., and Hendlin, D., J . Bid. Chem., 1949, 180, 125. Cuthbertson, W. F. J., Pegler, H. F., and Lloyd, J. T., Analyst, 1951, 76, 133. Emery, W. B., Lees, K. A., and Tootill, J . P. R., Ibid., 1951, 76, 141. Foster, J. C., Lally, J. A., and Woodruff, H. B., Science, 1949, 110, 507. Larkin, F. E., and Stuckey, R. E., Analyst, 1951, 76, 150. Skeggs, H. R., Hepple, €3. M., Valentik, K. -\., Huff, J. W., and Wright, L. D., J . Biol. Chem., Thompson, H. T., Dietrich, L. S., and Elvehjem, C. A., Ibid., 1950, 184, 175. Tatum, E. L., and Lederberg, J., J . Bact., 1947, 53, 672. Beadle, G. W., and Tatum, E. L., Apner. J . Rot., 1945, 32, 678. Bessell, C. J., Harrison, E., and Lees, K. A., Chem. & Ind., 1950, 561. Cuthbertson, W. F. J. (Personal communication). Wijmenga, H. G., Veer, W. L. C., and Lens, J., Biochim. Biophys. Acta, 1950, 6, 229. Coates, M. E., Ford, J. E., Harrison, G. F., Kon, S. I<., Porter, J. 117. G., Cuthbertson, W. F. J., and Pegler, H. F., Biochem. J . , 1951, 49, lxvii. Woodruff, H. B., and Foster, J. C., J . Biol. Chem., 1950, 183, 569. Brownlee, K. A., Loraine, P. K., and Stephens, J., J . Gen. Microbiol., 1949, 3, 347. Davis, B. D. (Personal communication). Cuthbertson, W. F. J., Pegler, H. F., Quadling, C., and Herbert, V., Analyst, 1951, 76, 540. 1950, 184, 211. NOTE-References 2, 3 and 17 are to Parts 111, V and IV of this series, respectively. GLAXO LABORATORIES LIMITED SEFTON PARK STOKE POGES, BUCKS. June, 1951