Microbiological Assay for Riboflavin HkRRY .4. KORNBERG', RUTH S. LANGDON, 4 + VERNON ~ H. CHELDELIN Department of Chemistry, Oregon State College, Corvallis, Ore. A microbiological method for the determination of ribofla\in is described. The organism used is Leuconostoc mesenteroides 10,100. Its response to ribofla\in has permitted the detelopment of an assay method which is sensitite to 0.0001 microgram of the kitamin per ml. With the growth medium used, good agreement is obtained among riboflaiin ialues of samples assayed at different le*els as well as good recoteries of added riboflaiin. Determinations may he made turbidimetrically after 14 hours or titrimetrically after 72 hours.
T
HE two methods which hare been used generally for the determination of riboflavin are the Luctobaczllus casei assay,
isms are centrifuged, washed once with sterile physiological saline solution by centrifugation, and resuspended in about 10 ml. of .sterile saline.
developed by Snell and Strong ( I S ) , and the fluorometric method, introduced by Hodson and Sorris (9). Variat,ions of both methods have been proposed. The adsorbent used in the fluorometric assay has been a source of error, as have coilcentration, volume, and clarity of extracts (10). This has been partially corrected by the use of special filters (10) and by using controlled permanganate oxidation of the sample t o he assayed in place of adsorption ( 1 6 ) . For the microbiological method taka-diastase was found to hydrolyze interfering substances present in starch that stiniulate lactic acid bacteria ( 1 7 ) . Filtration of autoclaved samples at pH 4.5 has been suggested to remove protPins ( 2 0 ) . Fat acids were observed to stimulate the growth of L . c a w ( 2 ) , and a procedure for their removal has been published (19). Several changes in the gron-th medium for the assay have been suggested (4,11, 12, 15, 19). The fluorometric and microbiological assay methods have been compared, and it appears that the microbiological assay is favored slightly on the basis of s o m e ~ h a tbett er specificity and reproducibility (1, 5, 7 ) . However, for the a s a y method using L. cusei as the organism, an amount of extract must be used that contains about 0.02 microgram of riboflavin per ml. Since many natural materials have a very low riboflavin content, relatively concentrated extracts must often be prepared, which may contain excessively colored pigments and other interfering substances, such as starches and fats. This paper describes an assay for riboflavin based upon the response of Leuconostoc mesenteroides to concentrations of the vitamin in the order of one fiftieth of that required for an equivalent response of L. casei. The organism's growth may be measwed turbidimetrically after 16 t.o 18 hours, or the acid produced may be tit,rated after 72 hours. The higher sensitivity of this microorganism permits preparation of samples for assay with greater dilution than heretofore; consequently decreased amounts of estraneous matter are present in the assay tubes
Basal Medium. Table I shows the composition of the basal medium. Quantities of various constituents per assay tube are given, as well as quantities per liter and the final per cent concentration. Except for the special preparation of the alkali-treated peptone and the addition of asparagine, it is similar to that used in the Snell and Strong method for riboflavin (18). Each amount given represents the concentrations which the authors have found t o produce thc best growth response in the presence of riboflavin.
Table I.
+
Grams
per Liter 12
% in Final Medium (Diluted) 0.6
20 1.0
1 .o 0.05 0.05
0.2 10 mi. each
o.ni
1 .o
1000 mi.
....
For the preparation of the alkali-treated peptone, to 40 grams of Bacto peptone (Difco) in 250 ml. of water are added a solution of 20 grams of sodium hydroxide in 250 ml. of mater. The re-
sulting mixture is brought to a boil and allowed to stand a t 37" C. for 21 to 48 hours, 28 ml. of glacial acetic acid and 7 grams of anhydrous sodium acetate are added, and the volume is made up to 667 nil., giving a concentration of 60 nig. of original peptone per ml. This solution is stored under toluene. The expedient of bringing the peptone to a boil, followed by incubation of the alkaline solution, was found to destroy virtually all the riboflavin. Prolonged boiling is to be avoided, since heating for even 5 minutes was observed to impair the efficiency of the peptone, possibly by racemizing some of the essential amino acids. Asparagine, not normally present in alkali-treated p e p tone, x a s found to stimulate growth in the presence of riboflavin. .1 more nearly synthetic medium using casein hydrolyzate, glucose, several amino acids, purines and pyrimidines, salts, and amino acids may also be used. I t s formula is similar to that used by Gaines and Stahly (8)for their nutritional studies on a strain of Leuconostoc mesenteroides, identified as P-60.
The Leuconostoc mesenteroides was kindly furnished by R. J. Williams of the University of Texas. Cultures may be obtained from the Bmerican Type Culture Collection, Georgetown University Medical School, Kashington, D. C., where the organism is listed as No. 10,100. The correct naming of this organism is being investigated by R. P. Tittsler. Although it has been called Leuc. mesenteroides P-60, its fermentation characteristics resemble t.hcse of Leuc. d e s t r a n i c u m more closely (see 3). P-60 cultures described by Dunn et al. (6) and by Pennington (14) d o not require riboflavin.
Assay Procedure. Assays are carried out in 20 X 150 mm. Pyrex test tubes. Red glass test tubes are necessary if the acid production is to be measured after 72 hours' incubation. Samples t o be assayed are introduced into tubes a t four levels, 1 to 4 ml., containing approximately 0.001 to 0.002 microgram of riboflavin per ml. These samples are then diluted t o 5 ml. with distilled water and 5 ml. of the basal medium shown in Table I are added to each. d series of tubes containing, respectively, 0.0, 0.0005, 0.001, 0.002, 0.004, 0.006, 0.008, and 0.01 microgram of riboflavin is used to establish a standard curve with each assay. The tubes are covered with a towel and autoclaved for 10 minutes (if acid production is to be measured after 72 hours instead of
METHOD
Preparation of Inoculum. Stab cultures containing 1% glucose, lyOyeast extract, and 1.5y0agar are used to carry Leucouostoc mesenteroides. They are kept in the refrigerator, and the inoculum for assay is grown from them in 2.5 ml. of basal medium plus 0.05 microgram of riboflavin with sufficient water to make 5 ml. After incubation at' 37" C . for 12 to 24 hours the organ1 Present
Basal Aledium
hfg. per Assay Tube Boiled alkali-treated pepGO tone sodium acetate Glucose 100 5 Asparagine 5 Yeast supplement (rihoflai-in-free) 1 Cystine Salts A and B 0 . 0 5 nil. K a t e r t o make 5 mi. 6.8-7 .0 PH
address, General Electric Co., Richland Wash.
81
82
ANALYTICAL CHEMISTRY
turbidimetric measurement, the tubes are plugged and autoclaved for 15 minutes). The tubes are cooled and inoculated with one drop of the previously grown culture (described above). Turbidimetric readings may be made after 14 t o 16 hours' growth a t 37" C. If the number of tubes is large, they are usually cooled prior to reading, in order to slow growth. Measurements may be made in any reliable turbidimeter. A filter corresponding to the color of the basal medium is used to negate any effects due to slight changes of color in the various tubes. The authors have found that steeper curves result when filters transmitting shorter wave lengths are used (4200 to 4500 A.) because of light scattering by the cells. However, the steeper curve is not particularly advantageous when there is a possibility of introducing errors because of color variation. In Figure lo is shown a typical standard curve for riboflavin, when a 5400 A. filter is used. The turbidimetric readings are given in terms of optical density which is equal to the log per cent of incident light absorbed and diffused (2 log G).
-
Table 11. Riboflavin Analyses and Recoveries hfaterkl
Riboflavin Found
Sample Mg.
Green pea soup. dehydrated
7
2 4
0.0008 0.0016
6
1
+ 0.001 yriboflavin 2 + 0.002 y riboEavin
0.0024 0.0034 0,0013 0.0028
0.5
0.0012
8 Sweet potatoes
1.0 1.5
2.0 0.5 1.O
Turkey feed
++
0.25 0.50 0.75 1.o
0 50
0.75
0.0020 0.0035 0.0044 0.001 y riboflavin 0,0021 0.002 y riboflavin 0.0039 0,0008 0.00185 0.00225 0,0034 0.002 y riboflavin 0,0037 0.003 y riboflavin 0.0051
++
Recovery
Y/Q.
0.40 0.40 0.40 0.43 ,
.
,,
2.4 2.0 2.3 2.2 .,
.,
%
,. ,,
90 100
.. .. 90 95
3.2 3.7 3.0 3.4
.. ..
93 95
Table 111. Comparative Assay Values Using Leuc. mesenteroides and L . casei Material Yeast Navy beans Green pea soup, dehydrated Pea puree Sweet potato Tomato juice Carrots Carrot puree White potatoes White flour, unenriohed
0' 0
,002
.004
RIBOFLAVIN
Figure 1.
,006
,008
- MICROGRAMS
,010
1
,012
Growth Response of Leuc. mesenteroides to Riboflavin
Acid production may be measured after a 3-day period of incubation, using 0.05 h' sodium hydroxide with bromothymol blue indicator. Figure 2 shows the type of standard curve which may be obtained by this means and the range of riboflavin concentrations employed. This organism (a heterofernientative species) is capable of producing about 8 ml. of 0.1 N acid under ideal conditions, and the curve shown here could be extended to include higher riboflavin levels. However, the slope of the curve becomes smaller, and best results are obtained within the recommended range.
Riboflavin, y per Gram Leuconostoc mesenteroides L. casei 45 45 2.6 2.5 1 .oo
0.90 0.95 0.32 1.68 0.16 0.36 0.21
1.01 1.05
0.94 0.35 1.83 0.22 0.52 0.33
flavin content by the present method. It has been shown (W,19) that fat-soluble substances present in many natural materials will stimulate the growth of L. casei. Blthough Leuc. mesenteroides is similarly stimulated, only relatively small amounts of these stimulants are present in the assay tubes, because of the greater dilution, and their effect upon the apparent riboflavin value is correspondingly less. Thus, for example, the authors have found that whereas ether extraction prior to assay lowered the apparent riboflavin content of flour from 0.33 to 0.20 microgram per gram with L. casei, the same treatment produced no lowering of the Leuc. mesenterotdes assay value (0.21 to 0.22 microgram per gram). The present assay method has given consistently satisfactory results for over 6 months, when used with a variety of plant and animal materials.
I
Preparation of Samples for Assay. Extracts are prepared by adding 20 ml. of water and 3 ml. of 1 S sulfuric acid to 1 gram of the dry material to be assayed and autoclaving for 30 minutes. The pH is then adjusted to 4.5 to 5.0 and the extract diluted to contain approximately 0.001 to 0.002 microgram of riboflavin per ml. Materials low in riboflavin content may be diluted t o contain approximately 0.0002 microgram per ml. RESULTS
Table I1 contains t 1.e assay values and riboflavin recoveries from three different materials. Good agreement between the four assay levels has been obtained and recoveries of the added vitamin range from 90 to 10070. Table I11 gives riboflavin values obtained by this method compared with those obtained by the method of Snell and Strong using L. casei (18). In general there is good agreement, although the last two samples appear t o have a considerably lower ribo-
.
~~
0
Figure 2.
,002
,004 ,006 ,008 ,010 RIBOFLAVIN MICROGRAMS
-
,012
Acid Production by Leuc. mesenteroides
V O L U M E 20, NO. 1, J A N U A R Y 1 9 4 8 LITERATURE CITED
(1) Andrews, J. S.,Boyd, H. M., and Terry, D. E., IND.ENG. CHEM., *&SAL. ED.,14,271 (1942). (2) Bauernfeind, J. C., Sotier, A. L., and Boruff, C. S.,Zbid., 14,666 (1942). (3) Bergey, D. H., et a&.,“Manual of Determinative Bacteriology,” 5th ed., Baltimore, Williams & Wilkins Co., 1939. (4) Chattaway, F. W., Happold, F . C., and Sandford, Mary, Biochem. J . , 37,298 (1943). (5) Conner, R. T., and Straub, G. J., IND.EKG.CHEM., ANAL.ED., 13,385 (1941). (6) Dunn, hl. S.,Shankman, S.,Camien, M.N., Franke, W., and Rockland, L. B., J . Biol. Chem., 156,703(1944). (7) Emmett, A. D.,Bird, 0. D., Brown, R. A., Peacock, Gail, and Vandenbelt, J. M.,IND.ENG.CHEM.,ANAL.ED., 13, 219 (1941). (8) Gaines, S.,and Stahly, G. L., J . Bact., 46,441(1943). (9) Hodson, A. F., and Norris, L. C., J . B i d . Chem., 131,621 (1939). (10) Hoffer, A , , Alcock, -4. W.,and Geddes, W.F., Cereal Chem., 21, 515 (1944). (11) Keminerer, A. R., J . Assoc. Oficial A g r . Chem., 27,540 (1944).
83 (12) Kent-Jones, D. W., and Meiklejohn, M., Analyst, 69, 330 (1944). (13) Ibid., 69,372(1944). (14) Pennington, D., Science, 103,397 (1946). (15) Roberts, E.C.,and Snell, E. E., J . Bid. Chem., 163,499(1946). (16) Rubin, S.H., De Ritter, E., Schuman, R. L., and Bauernfeind, J. C., ISD.ENG.CHEW,ANAL.ED.,17, 136 (1945). (17) Scott, M.L.,Randall, F. M., and Hessel, F. H., J. Biol. Chem., 141,325 (1941). (18) Snell, E. E., andstrong, F. M., ISD. ESG.CHEM., ANAL.ED.,11, 346 (1939). (19) Strong, F. M., and Carpenter, L. E., Zbid., 14,902(1942). (20) Wegner, M. E., Kemmerer, A. R., and Fraps, G. S., J . Bid. Chem., 146,547 (1942). RECEIVED June 27, 1947. A preliminary report was presented a t the fifth meeting of the Oregon Academy of Science, Portland, February 1947. Published with the approval of the Monographs Publications Committee, Oregon S t a t e College, Research Paper 114, School of Science, Department of Chemistry. This work was supported b y t h e Nutrition Foundation, Inc., New York, and b y the General Research Council, Oregon S t a t e System of Higher Education, Corvallis, Ore.
Volumetric Determination of Small Amounts of Soluble Sulfates C. L. OGG, C. 0. WILLITS, AND F. J . COOPER, Eastern Regional Research Laboratory, PhiZadeZphia 18, Pa.
A new technique is described for the volumetric determination of sulfur, in w-hich standard barium chloride is used with the indicators dipotassium rhodizonate or tetrahydroxyquinone. The technique permits continuous following of the colors developed, so that the true end point is easily identified. This makes the titration readily reproducible and results in an appreciable gain in accuracy. The barium chloride solution is standardized, and samples are titrated by the same procedure, eliminating a titration correction factor.
T
H E volumetric method for determination of sulfur as sulfate with barium chloride, in which the internal indicator tetrahydroxyquinone disodium salt (THQ) or the dipotassium or disodium rhodizonate is used, has been described by a number of investigators (f-fg), each of whom has attempted to present a method whereby a sharp or reproducible end point may be obtained. Although satisfactory in t’he hands of skilled technicians, these methods have not been generally accepted because ‘ of difficulty in identifying the true end point. I n the titration of sulfate ions with barium chloride, the indicator changes the color of the solution from yellow to orange red. As the end point is approached, but before it is reached, there is a localized formation of the red barium salt, of the indicator, which is dispersed through the solution by stirring. This produces a pseudo end point, which don-ly disappears as the barium associated with the indicator reacts with remaining sulfate ions. Because of the low concentration of sulfate ions, the reaction becomes very slow as the end point is approached. Unlike most reactions, the true end point is not indicated by a sudden change in color but instead is that, point where no fading of the barium-indicator color (orange red) occurs. The success of the titration depends upon the ability of the analyst to distinguish that point in the titration a t which fading no longer occur^. Throughout the titration, there is a slight but gradual change in the stable color from yellow to orange red. As the equivalence point is passed, there is a shift in the color from orange to red, the same as that which occurs with the premature or pseudo end points but distinguishable from them because i t is nonfading. If the end point is passed and the titration is continued, the shift in color toward the red is more rapid than that before the end point is reached, and will continue until sufficient barium has been added to combine with the indicator present.
The end point should be easily recognizable if it can be compared with a standard which has the same orange-red hue as the solution which has been titrated just past the equivalence point. However, the fading of the color must be followed continuously to avoid selection of a pseudo end point. An apparatus and titration technique have been developed which not only identify the end-point color but permit the continuous comparison of the color of the solution with a standard color filter. The apparatus (Figure l ) , consists of a rectangular titration vessel (23 x 45 X 50 mm. high optical absorption cell), a standard 25 x 45 mm. glass color filter, two 5-ml. burets graduated to 0.01 ml., and a titration stand. The titration vessel and the light filter are mounted side by side on an opal glass plate. I1lumination is from below, preferably by fluoresrent light. All thr opal glass is masked except that covered by the titration vessel and the color filter. Best results are obtained when no overhead artificial illumination is used. PROCEDURE
Sulfate solutions high in carbonates are acidified with nitric acid and boiled. All solutions, eit,her acidic or basic,, are neutralized to pH 6.5 to 7.5, the neutral solution is transferred to the. titration vessel, and the volume is adjusted to approximately 15 ml. About 0.08 gram of the commercial tetrahydroxyquinone indicator is added and dissolved, and the solution is diluted with an equal volume of 9370 et,hanol. The solution is titrated with standard barium chloride until the permanent color matches the color filter. At the end point, stirring must be continued long enough (1 to 2 minutes) to make sure that no fading of the orange-red color will occur. h rubber-tipped stirrer is used to prevent scratching the bottom of the absorption cell. The end point is reached when one additional drop of barium chloride causes t,he solution to appear a deeper red t,han the color filter. The 0.02 A- barium chloride is standardized against a solution containing 1.8140 grams of reagent grade potassium
