634 RIDYARD : THEORETICAL AND PRACTICAL CONSIDERATIONS {Vol. 75 Theoretical and Practical Considerations in the Determination of Aneurine (Vitamin B1) with Special Reference to the Recovery Factor BY H. N. RIDYARD SYNoPsI+-An analysis is made of the factors affecting the “recovery” of aneurine added to extracts in the determination of this substance. The effects of variation of sample and errprs in volume measurements are shown to be considerable. The combination of aneurine during digestion, the possibility of effects a t the oxidation stage, and interference with the emission of fluorescent light in the fluorimeter are considered. Results obtained in practice are given to illustrate the arguments and an investigation of the effects in the fluorimeter cell is included in some detail.The losses that take place in base-exchange methods are mentioned. The remarkable stability of the thiochrome method is showr, both by the expressions developed theoretically and by the practical results. Some of the considerations raised may have application to other estima- tions that involve similar raw materials and methods. THE addition of pure aneurine to extracts prepared for the fluorimetric determination of this substance has frequently been used as a criterion of the efficiency of the analytical pr0cedure.l s 2 y3 s4 s6 For this purpose, two similar extracts of the same sample, to one of which a known amount of aneurine has been added, are submitted to the same analytical procedure. The amount of aneurine found in the extract to which no addition has been made is then sub- tracted from the amount found in the other portion and the difference, “aneurine recovered,” is divided by the amount of aneurine added, to give the “recovery factor.” The added aneurine has commonly been assumed to be subject to the same losses as that already present in the material under examination, and it is this assumption and the accuracy of the method in relation to the physical and chemical factors involved that form the subject of the present investigation.Recovery experiments of this kind are commonly used for checking various analytical procedures and are often of great value, but examination of their theoretical bases appears to have been neglected, although their complete adequacy has been questioned.6 When recovery experiments are applied to a complex system such as we have in the fluorimetric estimation of aneurine they have definite limitations and conceal sources of error, which, if overlooked, can lead to fallacious conclusions.Some of the considerations brought forward in this paper are applicable to a wide variety of recovery experiments, others are specific to the estimation of aneurine. Furthermore, although additions are commonly used as a check on analytical procedure, in the estimation of aneurine the “recovery factor” has sometimes been used to correct the amount found-a procedure much more open to question. It appeared therefore that an analysis of the various factors involved would be of value to workers in this and similar fields, especially as the remarkable stability of the thiochrome method becomes more understandable in the course of the investigation. The fluorimetric determination of aneuririe may be considered in four successive stages- I.Extraction of the aneurine from the raw material either by steeping in acid water or by heating with dilute acid t o boiling-point, adjusting to pH 4.5 and digesting with takadiastase or other source of phosphatase according as aneurine pyrophosphate (co- carboxylase) is absent or present. The extract is sometimes passed over base-exchange materials. 11. Oxidation of the aneurine to thiochrome with alkaline ferricyanide, a process said t o be only about 70 per cent. efficient.’ 111. Extraction of the thiochrome by means of isobutyl alcohol. IV. Fluorimetric examination of this extract.Dec., 19501 I N THE DETERMINATION OF A4NEURINE (VITAMIN B1) I.FACTORS INVOLVED AT THE EXTRACTION STAGE I a . THE MODE OF ADDITION AND THE ERRORS CONCERNED IN THIS- Aneurine can be added to extracts in two ways- 635 I a 1. A definite weight of material can be extracted with acid water or digested in a buffer solution as is necessary, and a similar weight extracted in the same manner with a precisely similar fluid containing a definite amount of aneurine. I u 2. An extract can be prepared, centrifuged, filtered or otherwise clarified, and divided into two or more portions, to one of which is added a solution of aneurine. It will be seen that any factor, chance or otherwise, which affects the two extracts or portions of extracts in a different manner may appear as a considerably increased error in the recovery factor owing to the method of calculation. I a 1.In the first method, the homogeneity of the sample is obviously a factor in the type of result obtained, and many materials remain heterogeneous in spite of careful mixing, or may indeed segregate in the process of mixing. In extracts of samples of imported flour, enriched by synthetic aneurine, a small proportion showed deviations from the mean B, content which were as high as 0.05 pg. per ml. or even more, and this condition was not improved by further mixing of the flour. If only one extract without aneurine and one with were taken, such deviations would frequently act in opposition. If the mean quantity of aneurine in the flour extract should be 0.6 pg.per ml. (a normal level) the level without addition could be 0-55 or 0.65 and that with (say) 0-5 pg. per ml. of added aneurine could be 1-15 or 1.05 pg. per ml. The first pair of results would give a recovery of 0-6/0.5 = 120 per cent. and the second pair 0.4/06 = 80 per cent. For samples less homogeneous, the difficulties of obtaining similar fractions and exact weights of material increase, and the chance of error is increased accordingly.8 The question of volumes is of less importance than it is in method I a 2, for the added aneurine is measured as a standard solution and no error in concentration is involved; even a cylinder should measure 50ml. to within &0,25ml. Treated as above this could give recoveries of 101 or 99 per cent. In good work the deviation from the mean value is less than 0.02 pg.per ml. For the great majority of samples it rarely reaches 0.03 pg. per ml. and is of the same order in solutions of any con- centration between 0 and 1.0 pg. per ml. If with a solution containing 0.6 pg. per ml., the greatest error given above operated in opposite ways at the levels with and withod an addition of 0.5 pg. per ml., the results could be 0-57 or 0-63 without addition, and 1.13 or 1-07 with addition, which lead to recoveries of 0*56/0-50 = 112 per cent. and 044/060 = 88 per cent. Both percentage errors will increase rapidly as the quantity of added aneurine decreases. In rare and extreme cases these factors could all operate the same way to give recoveries of 120 x 101 x 112 = 136 per cent.or 80 x 99 x 88 = 70 per cent. These considerations show that to rely upon two extracts only (one with and one without addition) is useless. Nevertheless, because of the accuracy of the volume factor the method has commonly been used in this laboratory when recovery data were desired. The precaution has always been taken of using at least five extracts in all: two without additions, one with a moderate addition and two with larger additions, the results being plotted on squared paper.* If any point diverged seriously from a straight line drawn through the points, the matter was further investigated. I a 2. In the second procedure an extract is prepared, divided into two or more portions and aneurine added to one or more of these. This method is at once subject to the criticism that the extraction stage remains unchecked, although valuable informatiofi can obviously be obtained.The aneurine must be dissolved in the same buffer solution or dilute acid as is used in the extraction (see Table VI below). The manner of volume measurement is important. One mode of addition known to have been used was to add very small volumes of a strong aneurine solution from a graduated pipette and to neglect the increase in volume of the extract. Two different 1-ml. graduated pipettes were found to deliver 0.1 ml. with a standard deviation of 0.002 ml. for six successive deliveries from each, and of 0.008 ml. for six separate deliveries each; the maximum error being 0-011 ml. and the first deliveries always high. A 10 per cent. error in the recovery factor can thus easily be made, and the errors due to determination will still be present to The remaining error to be considered at this point is that of analysis.The mode of addition will be the critical factor.636 RIDYARD : THEORETICAL AND PRACTICAL CONSIDERATIONS [Vol. 75 affect the result. As the volume error decreases with better technique or larger volumes, it is usually preferable to add several ml. at least of aneurine solution, with care to add the same volume of acid water or buffer to the unfortified levels. The question of blank determination (fluorescence of a solution treated with caustic soda but not ferricyanide) must also be considered. This is advisedly determined in extracts with and without additions, and is, subject to the usual error of determination, normally the same in both.But if their independent blanks are subtracted before calculating the recovery factor, the blank errors could be added to those given above; hence in the determina- tion of the recovery factor the aneurine values obtained should be subtracted from one another before subtraction of the blank for determination of the aneunne level in the solution with no addition. The blank may of course be partly oxidised when ferricyanide is added.g*10 I b. COMBINATION OF ADDED ANEURINE DURING DIGESTION- One other factor must be considered at this stage when digestion processes are carried out during extraction of the raw material. Enzymic digestion sets up a state of equilibrium between free and combined forms of aneurine. Aneurine is considered to be phosphorylated by adenosine triphosphatell ; the aneurine pyrophosphate so formed combining with a protein to form carboxylase.In the following treatment no particular mechanism is assumed, but the equilibrium states of one reaction involving aneurine and of three successive reactions, are considered in respect of the effect of adding aneurine to the system. I b 1. ONE EQUILIBRIUM- A + B + C where A and C are free and combined forms of aneurine respectively. If now a is the total concentration of A free and combined, b the concentration of B free and combined and y the amount of C formed, we have at equilibrium .. .. . . . . (I b 1.1) Hence, neglecting the term y2 as very small indeed (since a is of the order g. mol. per litre) . . . .. . ?(I b 1.2) . . . . ab ’’ K’ + a + b * ’ From (I b 1.2) we see that for y to be just detectable (Le., = 10-8g. mol. per litre), if a = 10-6 g. mol. per litre then -# + lo-,. The free aneurine measured by the thiochrome b K method will be a - y = a ). If now we add aneurine of a concentration a’ b to the solution, the free aneurine measured will be (a + a’) recovery factor determined will be the difference between these two observed values divided by a’, i.e., (I b 1.3) (a + a’)b ab Recovery factor = 1 - _____ a’(K’ + a + a’ + b) a’(K’ + a + b) Since a and a’ are of the order the denominators of the two fractions in (I b 1.3) are virtually the same unless both K‘ and b are less than lo4, and the expression becomes b 1 - - and is independent of aneurine concentration.K’ + b 1 b 2. THREE EQUILIBRI-4- D + E + B A + B + C + F C + H - + M Let d , e, a, f, h, be the initial total concentrations of D, E, A, F, H, and at equilibrium let x of D and E be transferred, y of A, and x of H; K,, K,, K,, being the (reciprocal) equilibrium constants of the three reactions. Solving for x and z in the equilibrium equations of theDec., 19501 IN THE DETERMINATION OF ANEURINE (VITAMIN B1) 637 first and third reactions (rejecting terms involving second powers of x , y or x) and substituting in the second equation, we find that- . . . . + 12) = K,KJ(K, + d + e) + (K, 3. h)(ad + ae + de) - ride (I b 2.1) The first and third reactions have a buffering effect, and from considerations similar to those advanced above in the case of one equilibrium it seems unlikely that the recovery factor will be noticeably affected by aneurine concentrations, although the possibility must be borne in mind.Although these theoretical aspects have been considered for several years, it is only recently that practical results have been obtained which indicate that added aneurine may be combined to a marked extent at the digestion stage in the case of two or three materials. As these observations were made during investigations with a new adaptation of the sand absorption m e t h ~ d , ~ which will be described in a forthcoming paper, and very little work has yet been done on the phenomenon, further comment is not possible here. I c. BASE-EXCHANGE PURIFICATION AND RECOVERY- The use of zeolitic materials to purify extracts needs special mention.These usually lessen the optical interferences (see Table VII, Eluates), but may lead to losses by irreversible adsorption of aneurine, displacement by other basic material (organic or inorganic) in the extracts used, or inadequate adsorptive capacity. Irreversible adsorption, always found with De~also,~ might be proportional t o the weight of zeolite used, the concentration of aneurine, the time of contact, or to all three. Hence it cannot be assumed that recovery is thesame at the levels of addition and no addition. One example has been met with in which a 96 to 100 per cent. recovery was obtained with a lower analysis value than that obtained by another method giving a similar recovery. This might have been due to a chance combination of factors mentioned above, or to a genuine separation of aneurine from another substance fluorescing after oxidation. Safeguards lie in recovery experiments at more than one level of extract concentration and total aneurine - zeolite ratio and by comparison with other methods (see also I11 and IV, wheats, below).11. ERRORS AT THE OXIDATION STAGE Little special information has been obtained concerning any factor affecting recovery at the oxidation stage, and, with the possible exception of a few materials which have given serious and complicated difficulties in other ways, there is no experimental reason to suspect differences in the proportion of aneurine oxidised to thiochrome in solutions with and without addition, or indeed that the recovery at this stage is less than with pure aneurine solutions. The oxidation stage with pure aneurine solutions probably includes not only the main reaction, aneurine -+ thiochrome, but also a side reaction.If complex extracts introduce a substance that can combine with the aneurine during the oxidation process, the three reactions might be represented thus- Kl A + B K2 A + C K3 A + D + E aneurine -+ thiochrome side reaction Let the initial concentrations of A and D be a and d respectively, and let the amounts of B, C and E formed at time t be p , q and Y respectively, and let 9 + q = y. Then 3 + dq - = - dY = (K, + K2)(a - y - r) . . .. . . (I1 1) dt dt dt = K,(a - y - r)(d - Y ) . . . . ar dt -____ K3 (d - r) andr = d (1 - e ''1 7 Kz ") (compare R. Wegescheider12). Hence - = _____ dr dY Kl + K2638 RIDYARD THEORETICAL AND PRACTICAL CONSIDERATIONS [Vol.75 This value of r is expanded in series, terms involving y to powers greater than 1 being Hence, on integration, discarded (since y is of the order 10-6), and substituted in (I1 1). transferring to the exponential form and expanding, we find that- K3 y*- . . .] 1 . . (I1 3) The stability of this oxidation reaction, in spite of all kinds of foreign substances in the various extracts used, is well expressed by this equation, since p and q will not vary appreciably from their values with pure solutions unless dK,/(K, + K,) is comparable with 1, i.e., either d must be of the order a x lo4, or, if d is of the same order as a, K, must be of the order (K, + K2)1O4. 111. THE Z’SOBUTANOI, EXTRACTION STAGE In isobutanol extraction, interferences are less likely, with the important exception of the effect of high concentrations of dissolved materials in the aqueous solution changing the proportion of aneurine passing to the isobutanol layer.This is most likely to occur when using base-exchange methods with eluates containing high concentrations of a salt, for changes in the concentration of the salt may lead to serious error^.^^^^ IV. FACTORS INVOLVED IN THE FLUORIMETRIC MEASUREMENTS Although it has been upon this stage that attention has been concentrated in published references to recovery, the somewhat complicated process has not been adequately treated. Fluorescence is considered to be due to a molecule of the fluorescing substance absorbing a quantum of light of short wavelength and correspondingly high energy content, and sub- sequently emitting a quantum of longer wavelength and lower energy content ; the difference in energy due to the change in wavelength being dissipated as heat.Three factors can interfere with this simple relationship between absorbed and emitted light- IVa. Short wavelength light may be absorbed by some substance other than that fluorescing. IV b. The molecule of fluorescent substance may lose its acquired energy by collision with other molecules, either of the same substance, or other substances. The latter process is tenned “quenching,” but in the literature concerned with aneurine estimation the term appears to be used loosely to cover all forms of loss of output of fluorescent light.IVc. The fluorescent light may be absorbed before it leaves the solution. IV a. ABSORPTION OF SHORT WAVELENGTH LIGHT IN THE CELL BY SUBSTANCES OTHER THAN In view of the complex nature of the extracts used for aneurine estimation, it is obvious that this effect is likely to be present in the fl~orirneter.~ Now the addition of aneurine to an extract will ultimately involve a change in the total light absorbed in the cell, of which the light absorbed by the thiochrome is only a part. As the light absorbed varies logarith- mically as compared with the concentrations of the absorbing substances and since, further, in dilute solution the fluorescence is proportional to the light absorbed by the thiochrome and is measured at some distance up the cell, it can be seen that the results of changing the concentration of the aneurine are not simple. A detailed examination of the light absorption in the cell, and a determination of the extinction coefficients concerned, was made in order to investigate the matter.In the Spekker fluorimeter a narrow band of light passes vertically through a rectangular cuvette, the fluorescent light emitted being collected by a circular selenium cell from a circular patch of indefinite extent; an area in the centre of the cuvette having the greatest effect (Fig. 1). Let a beam of light incident on the base of the cuvette have a width w and an intensity I,; and the circle from which light falls on the photo-cell a radius Y and centre situated a t a height h above the base of the cuvette. THIOCHROME-Dec., 19501 IN THE DETERMINATIOX OF ANEURINE (VITAMIN B1) 639 From the Beer - Lambert law, the light absorbed in this circular region by the thiochrome in a solution containing other absorbing substances will be- (IV a 1) where K,C, are the absorption coefficient and concentration of thiochrome in the cuvette, and K,C, the absorption coefficient and concentration of another absorbing substance.Other factors, K,C,, etc., could be added if necessary without changing the argument. Front view of cuvecte Side VICW of cuvettc Fig. 1. Optical effects in fluorimeter cuvette This expression, rearranged, integrated in series, and simplified (terms beyond the second . . ( I V a 2) The light absorbed by a corresponding thiochrome solution free from other absorbing agents will be given by (IV a 2) with the omission of KbCb wherever it occurs .. . . ( I V a 3 ) The relative amount of light absorbed by the thiochrome in the two will be given by (IV a 2) divided by (IV a 3), i.e., will be e } .. . . . . .. . . ( I V a 4 ) In a particular solution, the actual aneurine concentration, multiplied by this factor, gives the apparent aneurine concentration that would be obtained in the analysis if light absorption by substances other than thiochrome were the only disturbing condition. The factor (IV a 4) would then be the “recovery factor”; it is denoted by “R” in the remainder of this discussion. being discarded as extremely small), reduces to K,C, m%Ioe - h(KaCa 4- K*cb){ 1 + Q[r(K,C, + K,Cb)]e) . . V Fig. 2. Arrangement of apparatus for measurement of light absorption640 RIDYARD : THEORETICAL AND PRACTICAL CONSIDERATIONS [Vol.75 Experimental deteymination of absorption coeficients-The determination of the absorption coefficients was undertaken with the simple apparatus shown in Fig. 2. The right-hand side block of the Spekker fluorimeter was removed, and the light from the mercury discharge lamp A, after passing a filter B (Chance’s 0x1, UV), the lens C normally mounted on the fluorimeter, and an 8mm. cardboard diaphragm D, was further concentrated by another lens E, on to a series of cardboard diaphragms F, G, H, with holes 16mm. in diameter, extended over a distance of about 1 metre. By this means a reasonably parallel beam was obtained, which passed through either of the cells used without touching the side.The cells K (9.2 and 29.2 cm. long, and 26.0 mm. diameter) were glass tubes with flanged ends and side tubes. The end plates were squares of photographic plate sealed on with Seccotine. The light, after passing through the cell fell on a selenium cell M, immediately in front of which was placed another Chance filter, 0 x 1 , N. A further filter P (Chance OV1, purple 8), was interposed in the beam before the cell. The two glasses together give a narrow trans- mission band in the 365 mp. region, as shown in Fig. 3 (manufacturers’ curves). SI’ECT‘tCIY Fig. 3. Manufacturers’ curves for transmission through Chance Glass Filters OX1 (ultra-violet) and OV1 i:purpIe) The solutions were prepared by precisely the same oxidation method as is used in normal procedure, but larger tubes were used, and all quantities multiplied by ten.The isobutanol extract (250ml.) was syphoned off from the aqueous layer by means of a syphon bent upwards at the tip to prevent contamination. The first attempts to determine the trans- mission of these solutions failed, as cloudiness caused a larger loss of light than the true absorption. The normal, or increased, amounts of ethyl alcohol added delayed the formation of cloud, but did not prevent it. The formation of minute crystals of sodium carbonate was found to be the main cause. Finally, a technique was adopted of syphoning the isobutanol extract into 350-ml. stoppered bottles and storing these in the refrigerator overnight. This threw out of solution a certain amount of water with dissolved sodium hydroxide. The next day the bottles were transferred to the instrument room (in darkness) and allowed to warm up without undue disturbance.Immediately before the determination of its transmission the isobutanol extract was transferred to a beaker containing 10 ml. of absolute alcohol and drawn up into the absorption cell. By thia means clear solutions that gave remarkably reproducible results were obtained. The extracts were later found to have ff uorescences indistinguishable from those of corresponding extracts prepared in the normal manner immediately before making the comparison. Two or more blanks, made by oxidising acid water only, and one or two standards made by oxidising solutions of pure aneurine only, were used in each run, and all determinations, including the extraction of any raw material, were repeated on several different days.The absorption coejicient of solutions derived from pure aneurine-From the results obtained with the isobutanol extracts prepared from standard solutions of pure aneurine,Dec., 19501 IN THE DETERMINATION OF ANEURINE (VITAMIN B ~ ) 641 used in each day's work, the absorption coefficient was calculated using the formula I loge I'" K,= - c, x L where the intensities I, and I were given by the galvanometer deflections obtained when the cell was filled with isobutanol extracts obtained by the oxidation of acid water and an aneurine solution respectively, C, was the concentration of the aneurine solution used in pg. per ml., and l was the length of the cell.The results obtained after the adoption of the technique described are shown in Table I. It was the agreement of the values of this coefficient which was used as the final check on the clarity of the solutions and the correctness of other dis- posit ions. It should be mentioned here that the aneurine used throughout this work, and in terms of which all aneurine concentrations are expressed, was a good commercial crystalline sample kept in a screw-topped bottle in the open laboratory. At a later date and for another purpose this was repeatedly and very carefully compared with the international standard aneurine by the thiochrome method and found to be equal to the international standard x 0.91. K, for the international standard would thus be 0.0087 per pg.per ml. per cm. TABLE I LIGHT ABSORPTION COEFFICIENTS OF SOLUTIONS DERIVED FROM PURE ANEURINE 7 Date 22.11.45 13.12.45 28.12.45 22. 1.46 25. 1.46 Y9 1 ) 29.2 cm. cell Concentration, pg./ml. 0.8 2.0 1.0 2.0 1.0 2.0 0.8 A Mean 0.00814 0.00807 0.00745 0.00837 0.00755 0.00780 0.00794 0.00790 7 Date 30.11.45 2. 2.46 8. 2.46 15. 2.46 15. 3.46 11 Y Y 9.2 cm. cell Concentration, 4.0 2.0 1.2 2.0 2.0 1.6 2.0 A w / m l . Mean 0.00780 0.00793 0.00807 0.00797 0.00797 040790 0-00826 0.00796 Mean of all results 0.00793. Standard deviation 0.00026. Absorption in solutions derived from extracts of bran-Turning to solutions in which light was likely to be absorbed by substances other than thiochrome, bran was taken as a materia1 with a low recovery which varies with the concentration of bran in the extract.The first bran chosen, B.1045, was made up in concentrations of 2, 3, 4, 5, 6, 7, 8 g. in 50 ml., with,five Fig. 4. Bran B. 1045. Observed aneurine concentration642 RIDYARD : THEORETICAL AND PRACTICAL CONSIDERATIONS [Vol. 75 separate extracts a t each concentration. Two of these contained no added aneurine, the others 0.2, 0.4, 0.6 pg. per ml. respectively. On analysis these yielded the values shown in Fig. 4. The highest value at 8 g. bran concentrations with 0.6 pg. per ml. addition was off the normal scale of readings and hence was somewhat uncertain. Blank values were neglected as largely destroyed on oxidati~n.~ Absorption measurements were carried out on isobutanol extracts prepared as described above. These extracts were prepared twice for work on different days and with a different arrangement of the apparatus, two different cells employed and by different disposition of the diaphragms, etc., the light intensities were changed.A provisional value for the aneurine concentration was obtained from the 3 g. bran concentration as the most reliable, the mean value of 0.365 pg. per ml. at no addition divided by recovery interpreted from the graph as 87.5 per cent. giving a level of 0.4,17 pg. per ml. as the true value, or 0-139 pg. per ml. per g. per ml. I t was assumed for this purpose that the “recovery” at this bran concentration was the same for the “no addition” level as for that with additions; the justification for this assumption lying only in the reasonableness of the subsequent results as a whole.It was provisionally assumed also that the concentration of the absorbing substance would be proportional to the bran concentration, which was used for the term Cb in the expression I = 1,e - z(KaCu -t KbCb), where I, is the incident intensity and I the intensity of light leaving the cell, which has the length 1. K, is the absorption coefficient considered previously, C, the concentration of aneurine, and K,Cb the corresponding values relating to the component derived from bran. From this it follows that I loge f - K,C, x I C b x 1 K, = and values of this are given in Table 11. TABLE I1 LIGHT ABSORPTION COEFFICIENTS, BRAN B. 1045 Concentrations Cell length, cm. ----7 7- Bran in Added aneurine in ,---*------, extract aneurine extract Galvanometer AbsorDtion Galvanometer Absomtion A \ Total 29*2* 9.2t -.Tg./50 ml.) (pg./rnl.) (pg. /rnl.), deflection coeffikent deflection coeffiiient Ca Kb Kb 0 2 3 3 4 5 5 6 7 7 8 0 0.0 0.0 0-6 0.0 0.0 0.6 0.0 0.0 0.6 0.0 0 0.278 0.417 1.017 0.556 0.695 1.295 0.834 0.973 1.573 0.813 35.1 38.0 24.7 E;*O 0.0212 20.7 4:*6 0.0205 19-9 16.2 1:5 0.0204 13.6 l.5 0.0200 11.1 0.6 0.0197 9.0 0.4 0.0201 10.3 7.5 0.0222 0.0208 0-0206 0.02 19 0.021 1 0.0210 0.021 1 0.0160 0.0208 Means . . 0.0203 0.0206 * Date of determination, 22.11.45. t Date of determination, 30.11.45. The remarkably good agreement of these figures was encouraging, especially in view of the very small deflections obtained with the long cell. Hence notice was taken of the general decline in the absorption coefficient with increase in concentration of bran, and therefore of aneurine, and particularly the very low level recorded with 7 g.+ 0.6, which actually gave a larger deflection than 7 g. + 0. To test if this were real, two pairs of solutions were prepared, 6 g. per 50 ml. + 0.0 and + 0.6, and 8 g. per 50 ml. + 0.0 and + 0.6. No more of the original bran was available, so some bran from a fresh milling of the same wheat was used, and in view of this the results agreed well with those obtained earlier. As theapparent fall in absorption might be caused by blue fluorescent light passing the OX1 filter in frontDec., 19501 IN THE DETERMINATION OF ANEURINE (VITAMIN B ~ ) 643 of the selenium cell, a second filter of the same type was interposed as an extra precaution, In spite of this the results were more definite than before (Table 111).TABLE I11 EFFECT OF ANEURINE CONCENTRATION r- -----7 LIGHT ABSORPTION COEFFICIENTS, BRAN SIMILAR TO B. 1045 Concentrations 13.12.45 Cell length, 9.2 cm. Galvanometer Absorption Bran, Added aneurine, deflection coefficient 6 0.0 13.7 0.0208 6 0-6 16.1 0.0190 8 0.0 10.3 0.0197 8 0.8 13.4 0.0163 g./50 nil. pg./ml. The probable value for the absorption coefficient Kb at each concentration was obtained by plotting the values obtained at no addition and at 0.6 pg. per ml. addition on a graph, drawing lines passing evenly between the points, and reading from these the values required at each bran concentration. The values for 0.2 and 0.4 pg. per ml. addition were obtained by symmetrical interpolation between the values for 0.0 and 0.6 pg.per ml. obtained from the curves. From these, the values of R given in column 4, Table IV, are derived (formula (IV a 4) above, Y taken as 1-5 cm., see Table XI). A calculation was then made of the apparent aneurine concentration C,R that would be given by each of the bran extracts if this light absorption were the only interference; and these values are given in column 5, TABLE IV EFFECT OF LIGHT ABSORPTION ON APPARENT ANEURINE CONCENTRATION IN BRAN EXTRACTS , AND OF LIGHT ABSORPTION + QUENCHING Concentrations r 3 Bran Aneurine Total extract, added, aneurine, g./50 ml. pg./ml. pg./ml. cb Ca 2.0 0.0 0.278 0.2 0.478 0.4 0.678 0.6 0.878 3.0 0.0 0.417 0.2 0.617 0-4 0.817 0.6 1-017 4.0 0.0 0.556 0.2 0.756 0.4 0-956 0.6 1.156 5.0 0.0 0.695 0-2 0.895 0.4 1.095 0.6 1.295 6.0 0.0 0.834 0.2 1434 0.4 1 *234 0.6 1.434 7-0 0.0 0-973 0.2 1-173 0.4 1.373 0.6 1.573 8.0 0.0 1.112 0.2 1-312 0.4 1.512 0.6 1.712 * h Factor, R 0.9175 0.9186 0.9193 0.9201 0.8804 0.8820 0.8837 0.8857 0.8457 0.8485 0.8 Fi 1 9 0.8545 0.8140 0.8181 0.8221 0.8262 0.7838 0.7894 0.7951 0.8008 0-7564 0- 7 6 2 6 0.7697 0.7768 0.7305 0.7388 0.7466 0.7557 Effect of light absorption only, CaK K./ml* 0.25.5 0.440 0.623 0.8 10 0.364 0*t546 0.721 0.900 0.471 0.642 0.824 0.987 0.566 0.7 3 3 0.900 1.070 0-653 0.816 0.982 1.150 0.736 0.896 1.057 1-222 0.814 0.970 1-130 1.297 .heurine content observed, pg. /ml. 0.25 0.17 0.43 0-60 0.80 0.38 0.35 0.53 0-69 0.90 0.48 0.45 0.63 0.77 0.95 0.56 0.54 0-72 0.85 1.02 0.65 0.64 0.78 0.91 1-06 0.69 0-69 0.79 0.93 1.06 0.75 0.73 0-91 1.01 1-18? Effect of light absorption + quenching Y (IV b 4) (IV b 6) Kc = 0.11 Ka = 0.0098 0.250 0.431 0.611 0.794 0.363 0-530 0.700 0.871 0-453 0-617 0.783 0.949 0.254 0.438 0.619 0.804 0.361 0.538 0.710 0.884 0.463 0.627 0.790 0.952 0.539 0.548 0-700 0-703 0.857 0.854 1.019 1.005 0.616 0.770 0.926 1.085 0.688 0-837 0.988 1.142 0.754 0.898 1.046 1.200 0-618 0.762 0-904 1.043 0.676 0.807 0.934 1.057 0.738 0.866 0.990 1.114644 RIDYARD : THEORETICAL AND PRACTICAL CONSIDERATIONS [Vol.75 Table IV. For comparison the values actually observed are given in colunin 6. Each observed aneurine concentration is the mean of three or four determinations on one extract. The duplicate values given are such means from two separate extracts.It will be seen from these figures that the observed values are in almost every result lower than those calculated. Hence, as would be expected, there is another form of inter- ference, possibly quenching. IV b. THE EFFECT OF QUENCHING IN ADDITION TO LIGHT ABSORPTION- It was assumed in the first instance that one molecule of thiochrome, activated by absorption of one quantum of ultra-violet light, is; de-activated by collision with one molecule of some substance derived from the bran whose concentration would be proportional to that of the bran itself. A stationary state will be reached when the rate of formation of activated thiochrome molecules will be equal to their rate of destruction. Now let [T] be the actual thiochrome Fig. 5 . Bran B.1045.Effect ot light absorption and quenching proportional t o bran C R concentration. Curves of the expression ---.-?--. Observed values in circles. KaC, -f 1 concentration, [TIB the apparent thiochrome concentration in presence of substances derived from bran, [T*l0 , [T*IB the stationary state concentration of activated thiochrome molecules in absence and presence of bran respectively; [labs]o and [I,ba]B the corresponding rates of absorption of light by the thiochrome. In addition, Ka KP Then T + hv --+ T*, T* -+ T + Izv'. a molecule of substance derived from bran. .. .. .. .. . . (IV b 1) .. .. .. .. . . (IV b 2 ) Since it has been assumed that [B] is proportional to C b , this may be transformed to give .. .. .. . . CaR K,Cb + 1 C, observed = . * (IV 64) and . .(IV b 5 ) C,R - C, observed K a = % = .. .. . . .. Kp C, observed x Cb . 'Dec., 19501 IN THE DETERMINATION OF ANEURINE (VITAMIN B ~ ) 646 The value of K8 will best be obtained from a high bran and aneurine concentration, and 5 g. per 50 ml. + 0-6 pg. per ml. was chosen as the most reliable high value. Substitution of the appropriate values for C,R and C, observed gave Ks = 0.0098. This value wasused for the calculation of the apparent aneurine concentrations given in column 7, Table IV, and plotted as the curves in Fig. 5 , on which the observed values are indicated by the circles. 1 4 - E z 4 z z V . ?I2 0 10 W 8 08 W d z 3 06 4 I- $ 0 4 4 0 U ' 02 (z ADDED ANEURINE Fig. 6. Bran B.1045. Curves of the expression CaR - 0.11 @ x Ca.>': 3- 1. Observed values in circles The differences between the observed and calculated results show means of +0.014 (16) and -0.019 (19).Although this could be regarded as satisfactory, the fact that all the results at 7 g. concentration are low is disturbing. This peculiar change in slope of the curves of observed aneurine concentration when passing from 6 to 7 g. of bran per 50ml. was too great to be due to normal experimental error, and could be a real effect--e.g., the concentration of some component could reach a maximum at 7 g. of bran, or all the molecules of thiochrome in the fluorimeter cell could be brought into some kind of association with molecules of a disturbing substance. Only a deviation proportional to the square of the aneurine concentra- tion would account for the changes in slope.Various provisional hypotheses were considered, and as a result of examining these it was found that the expression CUR- Kg $ x CUR (" >' .. . . (IV b 6) gave values in excellent agreement with the experimental results. the concentration 7 g. per 50 ml. + 0.6 pg. per ml., and found to be 0.11. Kt; was determined from Values for the TABLE V LIGHT ABSORPTION COEFFICIENTS, BRAN B. 1388 Concentrations Bran, Added aneurine, /ml. 0.0 0.6 0-0 0.0 0.6 0.0 0.0 0.6 Total aneurine; tLg. /ml. (C,) 0.54 1-14 1.08 1.62 2.22 1.89 2.16 2-76 8.2.46 9.2 cm. cell Absorption coefficient 0.0242 0.0252 0.0233 0-0246 0.0238 0.0242 0.0247 0.0245 Mean 0.0243646 RIDYARD : THEORETICAL AND PRACTICAL CONSIDERATIONS [Vol. 75 above expression are given in column 8, Table IV, and are plotted as curves in Fig.6, the observed values being indicated by circles. A second bran, B.1388, gave absorption coefficients very similar to those obtained from €3.1045 (Table V). In this material the aneurine content was greater than in B. 1045, which indicates the presence of more germ, whereas the drop in absorption coefficient with increase in aneurine concentration was not noticeable, possibly owing to the high initial aneurine level. The losses due to light absorption were again found to be insufficient to account for the low recoveries, and an additional factor of the same form as before was found to do this with great precision. C, apparent = C,R - 0.055 (2 x CaR)e 2 > 1 as is shown in the accompanying graph (Fig. 7). It will be noticed that the bran appears Fig.7. Bran 1388. Curves of expression C,R - 0.056 (5 x CaR)’ c_b 3- 1. 4 ’ 4 to exert its maximum effect on the solution at ,a concentration of 4 g. per 50 nil., while the constant KZ is one half of the value found with bran B.1045. An expression of this form is not easily accounted for in view of the very low aneurine concentrations involved, and further data are obviously required. The formation of “dimers” has been suggested as a cause of quenching at higher concentrations.14 Efect of acid concentration on absorption coefficient-The effect of the concentration of the acid used in extraction of the bran was also examined, with the following results- TABLE VI EFFECT OF ACID CONCENTRATION ON ABSORPTION COEFFICIENT Absorption coefficient uncorrected for aneurine concentration BRAN CONCENTRATION, 7 G .PER 50 ML. Hydrochloric acid Absorption coefficient 0.12 N 0.0 184 0.24 N 0.0220 0.36 N 0.0255Dec., 19501 IN THE DETERMINATION OF ANEURINE (VITAMIN BJ 647 Wheats-The absorption coefficients, I log, f - K,C, x I K, = I of a number of wheats were determined, and found to be surprisingly similar to one another (Table VII). Together with these, the eluates derived from sand adsorptiong were examined, and the results illustrated the remarkable degree of purification effected by this method. TABLE VII LIGHT ABSORPTION OF EXTRACTS DERIVED FROM WHEATS AND ELUATES OF WHEAT (Where the variety is mentioned more than once it was grow%at a different location) ' Corrected for aneurine concentration c, x I Concentrations L r- -7 Lab. Xo.€3.1054 1060 B. 1073 1076 1079 1303 1308 1400 1421 1441 H. 1390 7 7 ,f 39 B. 1320 Added Variety Wheat, aneurine, CW PQ. /ml. Little Joss 5 g./50 ml. Juliana, 9, 9, 7, ,Y 1, 79 7, Despres 80 S Steadfast 79 Conditioned Manitoba 9 9 Manitoba Y, Y, 5, Y Y 77 I f Y, Y, 10 g. /50 ml. 7, 5 g./50 ml. 3, 7, Mixed Grist >9 (70% Manitoba) 10 g. /50 ml. Y, Y, ?? 7, Ehates Manitoba 5 g./50 ml. 0.0 Mixed Grist 5 g./50 ml. Y9 I 9 10 g. /50 ml. Y Y 7, 10 g./50 ml. YY * In the eluate. Total aneurine, ca Pg. /ml. 0.35 0.35 0.35 0.35 0.39 0.3 8 0.44 0.40 0.43 0.42 0-46 0.86 0.88 0.47 1.07 0.375 0-75 0.75 0.375 0*305* 0*565* 0*278* 0*555* Cell length, 1 29.2 5 , Y Y 3, 9, 9, ,, 1) 9) 97 3, 1, Y Y 9-2 29.2 9, Y I 3, Y, 9, Y Y f Y n Absorption coefficient, % 0.00'742 0.00600 0.00548 0-00455 0.00573 0.00615 0.00564 0.00540 0.00534 0.00483 0.00555 15.2.46 0.00556 0.00548 ::::::} 27.2.46 0.00559 0.00558 0*00610 0.00617 } When it is considered that more than half of the total absorption of these eluates was due to the thiochrome, and that it had to be calculated from the separately determined aneurine absorption coefficient and concentrations and deducted before calcul&ion of the absorption coefficient of the eluate, the agreement is quite remarkable and leads to greatly increased confidence in the methods employed.The difference between the eluates from the two wheats is interesting in view of the similarity of the coefficients for the plain extracts. The difference is much too great to be error in galvanometer reading (I$ = 447/35.6 for log.per 50ml. of Manitoba; 43-0/36.6 for log. per 50ml. of mixed grist), is too great to be attributed to an error in the estimation of thiochrome concentration, and is unlikely to be caused by cloudiness in the solutions in view of the excellent agreement of the pairs of results, and of all other results after the first few days' work early in the investigation. The values for the thiochrome absorption coefficient were particularly good (0.00794 and 0.00790) on the two days on which these eluates were examined. The factor R for the Manitoba wheat was calculated and as with bran found to be inadequate to account for the low recovery found. The difference between C,R and observed aneurine concentration was accounted for well by a linear correction plus a square law correc- tion to allow for 15 per cent.of bran. Results are given in Fig. 8. Blank values have again been neglected.648 RIDYARD : THEORETICAL AND PRACTICAL CONSIDERATIONS [Vol. 76 IVC. ABSORPTION OF BLUE LIGHT- The absorption coefficients for blue light of isobutanol extracts derived from Manitoba wheats and from pure aneurine were determined, and are given in Table VIII. Since the fluorescent light passes only 0-7 cm. of the solution when leaving the cell, this factor has been neglected. TABLE VIII ABSORPTION COEFFICIENTS FOR BLUE LIGHT isoButano1 extract Absorption coefficient Aneurine (1.0 pg./ml.) . . .. .. . . 0.00037/pg./ml. Manitoba wheat, 10 g./50 ml. + 0.0 . . . . 0*000278/g./50 ml. Manitoba wheat, 10 g./60 ml.+ 0.4 . . . . O.O00219/g./50 ml. B Fig. 8. Manitoba wheat. Curves of the expression C,R - 0.056 C,K x b ”)*- 0.01 C, x C,R. Observed values in circles ( 4 x 100 Yeast-The ultra-violet absorption of isobutanol extracts prepared from digests of bakers’ yeast was also examined, with the results shown in Table IX. TABLE IX ABSORPTXON COEFFICIENTS OF iSOBUTANOL EXTRACTS DERIVED FROM YEAST, (FOR ULTRA-VIOLET LIGHT) Concentrations WL 7 Added Total Yeast, aneurine, aneurine, Buffer 1 0.0 0.22 Acetate 1 0.5 0.72 1, 2 0.0 0-44 1 0.0 0.22 Phosphate 1 0.5 0.72 I 9 2 0.0 0.44 18 g./50 ml. CLQ. /ml. Pg*/ml. (PH 4.5) ,, Absorption coefficient, per g./50 ml. 0.0 153 0.0145 0.0117 0.0141 0.0133 0.0138Dec., 1950j IX THE DETERMINATION OF ANEUKINE (VITAMIN BJ 649 The J1uorescence of solutions derived from pure arzeurine-The fluorescence of solutions derived from pure aneurine was compared with the light absorbed, which was calculated from the absorption coefficient relating to pure aneurine (0.00793 per pg.per ml.). Solutions of pure aneurine of various concentrations were oxidised and extracted with isobutanol in the usual manner, and the extract placed in the fluorimeter cuvette. The selenium cell receiving the fluorescent light was connected directly to the galvanometer with the same series-shunt arrangement as was used with the absorption measurements to diminish sensitivity, and the deflections given by the various solutions noted. The relative amount of light absorbed was calculated from the formula given above (IV a 3), r being the radius of a circle which includes about 100 per cent.of the light collected (1.5 cm., see Table XI). The factor relating deflection to light absorbed was calculated from the deflection at one concentration. The results are given in Table X, and it will be seen that there is good agreement between the observed and calculated figures a t concentrations much above those used elsewhere in this work. As was assumed before this study was undertaken, quenching due to interaction of activated thiochrome molecules can be neglected where C, is of the order TABLE X LIGHT ABSORBED RELATIVE TO FLUOKESCENCE For solutions derived from pure aneurine Galvanorne ter deflection (cm. ) Coilcentration o f aneurine, Relative light pf5 Iml. a.bsorbed 0.2 0.00158 0.4 0-003 15 0.6 0.00472 0.8 0.00626 1.0 0.00771 2.0 0.01538 4.0 0.0297% 6.0 0.04328 8.0 0.05594 10.0 0*0677!$ 20.0 0.1 1630 14.2.49 m- 7 Calculated Observed 0.379 0.44 0-755 0.88 1-130 1-17 1.502 1.45 1-85 1-83 3.54 3.62 7-13 7.08 10.39 10.18 13.40 13.32 16.30 16-33 (correlation factor a t 1.0 pg.) 22.2.49 1 Calculated Observed 0.346 0.36 0.690 0.69 1.032 1.03 1,372 1-35 1-69 1-73 3.37 3-38 6-52 6.49 9.48 9.38 12-25 12-15 14.84 14.84 25.44 24-95 (correlation factor at 10.0 pg.) Radius of circle from which is derived the light falling on the selenium cell in the SPekker juorimeter-To gain some idea of the radius of the circle from which the fluorescent light measured in the Spekker fluorimeter is derived, tinfoil diaphragms were placed in turn behind the fluorirneter cuvette, and the deflection of the galvanometer noted, the left-hand (com- parison) cell of the instrument being cut off frox'n the light source by a screen.The following results were obtained (Table XI). TABLE XI LIGHT PASSING DIAPHRAGMS PLACED IMMEDIATELY BEHIND THE FLUORIMETER CUVETTE Diaphragm radius, cm. 0.500 0.625 0.750 0.900 0.975 1.500 30 diaphragm Deflection, Percentage passed cm. 0-5 22 0.8 35 1.0 45 1.4 61 1.8 78 ".3 100 2-3 DISCUSSION OF RESULTS I t will be seen from the foregoing considerations that the determination of aneurine involves an exceedingly complex series of chemical reactions and physical processes. The addition of pure aneurine to extracts under examination is a valuable method of investigation if carried out with meticulous care, but neither recovery of added aneurine nor consistency of results is an adequate measure of the agreement of a result with the amount of aneurine650 RIDHARD [Vol.76 actually present in the sample. Nor is agreement with biological estimation a fully satis- factory criterion, since utilisation may very well be affected by other substances present (cf. vitamins A and E153), and important biochemical effects may easily be obscured. The fullest possible theoretical understanding of the determination is a great safeguard. It will be seen from the present study that factors affecting recovery of added aneurine are not, strictly speaking, linearly related to the total concentration of aneurine present in an extract, but that the deviations from linearity may be very small owing to the very low concentrations under consideration. Nevertheless, in bran and probably other materials, part of each deviation appears to be proportional to the square of the aneurine concentration, but this needs confirmation, and if confirmed, satisfactory theoretical explanation. Some support for this is given by the common experience in this laboratory that the quotient- amarent aneurine concentration recovery factor increases with the factor and therefore with concentration of + starting material. Base- exchange methods of purifying extracts do not completely remove the need for this kind of study, for, although as a rule they greatly diminish optical interferences, they also add some new complications. The writer’s thanks are due to Mr. G. G. Grindley for assistance in the absorption measurements and aneurine determinations. 1. 8. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. REFERENCES Wang, Y. L., and Hams, L. G., Biochem. J., 1939, 33, 1366. Booth, R. G.. J . Soc. Chem. Ind., 1940, 59, 181. Dawson, E. R., and Martin, G. W., Ibid., 1942, 61, 14. Vitamin B, Sub-committee of the Medical Research Council and the Lister Institute, Biochem. J., Wokes, F., and Organ, J. G., Ibid., iii. Wood, E. C., Analyst, 1946, 71, 1. Williams, R. R., and Spies, T. D., “Vitamin B1,” Macmillan Co., New York, 1939, p. 218. Ridyard, H. N., Analyst, 1949, 74, 18. -, J . SOC. Chem. Ind., 1946, 65, 93. Bouman, J., 2. V’itaminforsch., 1948, 19, 391. Weil-Malherbe, H., Biochem. J., 1939, 33, 1997: Wegescheider, R., 2. physikal. Chem., 1902, 41, 56. Herd, C. W., Mundy, L. M., and Ridyard, H. N., Analyst, 1943, 68, 174. Weil-Malherbe, H., and Weiss, J., Nature, 1942, 167, 471. Moore, T., Biochem. J . , 1940, 34, 1321. Davies, A. W., and Moore, T., Nature, 1941, 147, 794. 1943, 37, 438. THE RESEARCH ASSOCIATION OF BRITISH FLOUR MILLERS CEREALS RESEARCH STATION, ST. ALBANS July. 1949