Field of the invention

The present invention relates to the production of compositions comprising natural vitamin B1 2.

Background of the invention

Vitamin B1 2 is an important vitamin for humans and animals. It is used to treat pernicious anaemia, and peripheral neuritis, and is used as a dietary supplement. Vitamin B1 2 is also an important animal feed supplement as growth enhancer.

The term vitamin B1 2 is used to describe compounds of the cobalt corrinoid family, in particular those of the cobalamin group. The most used compound of this group is cyanocobalamin and as such the term vitamin B1 2 is sometimes used to refer to cyanocobalamin. In this specification the term vitamin B1 2 should be attributed its broad meaning so as to include all the cobalt corrinoids of the cobalamin group, which include in particular cyanocobalamin, hydroxocobalamin, methylcobaiamin and 5'desoxyadenosylcobalamιn characterized by a cyano, hydroxyl, methyl or 5'-desoxyadenosyl radical respectively. The methylcobaiamin and 5'desoxyadenosylcobalamin compounds are known to be unstable to light in isolated form and are easily transformed to hydroxocobalamin in aqueous solution. For this reason, almost all commercial vitamin B1 2 preparations consist of the stable cyanocobalamin which as such is not the chemical form in which vitamin B1 2 can be found in nature. In this specification the term

natural vitamin B1 2 is defined so as to comprise all chemical forms of vitamin B1 2 naturally occurring in nature, cyanocobalamin thus being excluded.

Vitamin B1 2 is produced industrially by microbial fermentation, using almost exclusively Pseudomonas denitrificans and Propionibacterium species, then converting the natural vitamin B1 2 into the cyanocobalamin form by chemical processes including cyanidization followed by extraction and purification steps using organic solvents (as reviewed by Spalla et al, 1 989 "Microbial production of vitamin B1 2", In: Biotechnology of vitamins, pigments and growth factors, E.J. Vandamme ed., Elsevier, London, New York, pp 257-284) The chemical conversion step and any subsequent purification steps cause this production process to be expensive, unsafe to the operators and environmentally unfriendly.

Upon ingestion, animals and humans convert cyanocobalamin to one of the natural forms of vitamin B1 2 such as methylcobaiamin or 5'- desoxyadenosylcobalamin which are required to function as a coenzyme for several biochemical conversions (Ellenbogen, L. , in: Handbook of vitamins Nutritional, biochemical and clinical aspects; ed; L.J. Mach n, Marcel Dekker Inc , New York and Basel) . In addition efficient growth of animals requires the presence of a sufficient amount of vitamin B1 2 activity in the animal feed Vitamin B1 2 preparations are frequently sold as feed supplements as such or as part of a premix containing additional vitamins and other feed additives.

It is clear therefore that a direct supply of δ'-desoxyadenosylcobalamin or methylcobaiamin instead of cyanocobalamin would be of benefit. In view of the instability of these two compounds they are not produced in isolated form but are produced in dry formulation together with the biomass of the organism in which they are produced Such formulations are well suited for use as animal feed supplement. For the production of methylcobaiamin and 5'desoxyadenosylcobalamιn, bacteria of the genus Propionibacterium are

most preferred, because, unlike P. denitrificans, bacteria of this genus have obtained the GRAS (Generally Recognized As Safe) status from the U.S Food and Drug Administration and are not known to produce endotoxins. Propionibacterium species are aerotolerant, nonmotile, nonsporulatmg, Gram- 5 positive bacteria characterized further by the production of large amounts of propionic acid from carbohydrates, lactic acid and polyhydroxy alcohols. The genus Propionibacterium falls into the high "GC" subdivision of the Gram- positive bacteria (T.D. Brock, M.T. Madigan, J.M. Martmko and J Parker in: The Biology of Microorganisms 7th Edition, Prentice-Hall International Inc., o 1 994) .

Patent application RU 2001 953 (Antibiotics Enzymes Research Technical Institute) refers to a spray-drying method for Propionibacterium shermann which has been used to produce 5'-desoxyadenosylcobalamιn- 5 containing biomass. Methylcobaiamin in biomass from methanogenic bacteria is commercially available from the Gedeon Richter company. However, the vitamin B1 2 content in these preparations is limited by the level at which the vitamin B1 2 is produced during fermentation and, as a consequence, the vitamin B1 2 content in these preparations does not exceed 0.1 % (w/w) . The 0 concentration of vitamin B1 2 in a composition is customarily measured as the dry weight of vitamin B1 2 as a percentage of the dry weight of the composition.

Concentrated (i.e. > 0.1 % w/w) products of cyanocobalamin are 5 commercially available. However, feed mills and premix manufacturers cannot use this material because as a consequence of its electrostatic properties cyanocobalamine segregates from its carrier upon processing in feed mills and premix manufacturers when formulated at concentrations > 0. 1 % w/w. Segregation is a particular problem during sieving, wind sifting or o allowing the composition to stand for a prolonged period of time

The present invention discloses a process by which it becomes possible to produce a composition which has a relatively high concentration ( > 0. 1 % w/w) of natural vitamin B1 2. These new concentrated products may be used for example in animal feed, human food or as an ingredient in cosmetics. These new concentrated products may be more convenient to use and to transport thus allowing a reduction in costs. They may advantageous¬ ly be used by feed mills and premix manufacturers as they may show reduced segregation of the vitamin B1 2 from the carrier during processing. The process can easily be carried out on an industrial scale and is relatively environmentally friendly as there is no need to use organic solvents or cyanidation.

Description of the invention

The present invention provides a process for the preparation of a composition comprising natural vitamin B1 2 which has a vitamin B1 2 concentration on dry matter in excess of 0.1 % (w/w). Said process compris¬ ing the steps of a) obtaining microbial cells containing natural vitamin B1 2, b) causing opening of the microbial cells such that at least part of the soluble content of the cells comprising vitamin B1 2 is released into a liquid in which the cells are contained, c) separating the opened cells and the liquid compris¬ ing the vitamin B1 2, and optionally d) preparing a mixture of the vitamin B12 and at least a part of the opened cells. The microbial cells containing vitamin B1 2 are preferably obtained in an industrial fermentation process using a microorganism known to produce vitamin B1 2. These include bacteria belonging to the genera of Acetobacterium, Aerobacter, Agrobacterium, Alcaligenes, Arthrobacter, Azotobacter, Bacillus, Clostridium, Corynebacterium, Escherichia, Eubac- terium, Flavobacterium, Methanobacillus, Methanosarcina, Mycobacterium, Propionibacterium, Proteus, Pseudomonas, Rhizobium, Rhodopseudomonas, Salmonella, Serratia, Streptococcus, Streptomyces and Xanthomonas.

Preferably the bacterium to be used in the process of the invention is safe for consumption by humans and/or animals and does not produce endo- or exotoxins. More preferably the bacterium has obtained the GRAS (Generally Recognized As Safe) status from the U.S. Food and Drug Administration. The most preferred bacterium for use in the process of the invention is of the genus Propionibacterium, the preferred species of which are P. freudenreichii or P.shermanu.

It is well known to persons skilled in the art that the industrial produc- tion of vitamin B1 2, like that of many other microbial metabolites, is carried out using strains resulting from programs designed to improve the desired qualities of any particular strain. These programs consist essentially of the treatment of the production strain with a mutagenic agent and the selection of mutants exhibiting an improved productivity or other advantages. Several techniques have been described for the rational selection of vitamin B1 2 overproducing mutant microorganisms (Spalla, C, Grein, A., Garofo, L and Ferni, G . Microbial production of vitamin B1 2 In" Biotechnology of vitamins, pigments and growth factors pp. 257-284 (E.J. Vandamme, ed.) Elsevier, London/NY, 1 989) . Such mutant strains are used in preferred processes of the invention and are capable of further increasing the vitamin B1 2 concen¬ trations in the compositions obtained.

In the invention the term " natural vitamin B1 2" is defined so as to comprise all chemical forms of vitamin B1 2 naturally occurring in nature, the cyanocobalamin form of vitamin B1 2 being excluded. In the preferred processes and compositions of the invention the natural vitamin B1 2 com¬ prises 5'-desoxyadenosylcobalamιn and/or methylcobaiamin.

In the process of the invention the microbial cells containing vitamin B1 2 are treated so as to cause lysis or other disruption of the cell membrane.

Lysis describes the opening of the microbial cells such that (at least part of) the soluble content of the cells comprising the vitamin B1 2 is released into a

liquid in which the cells are contained. Preferred treatments for opening the cells are heat treatments including pasteurization or heating in an autoclave; treatment with bacteriolytic enzymes such as lysozyme; mechanical disrup¬ tion of the cells including grinding or the use of shear forces; treatment with chemicals which cause cell lysis such as detergents or organic solvents; as well as combinations of these treatments. The lysis or other membrane disruption produces a lysate which can be separated into solid and liquid phases.

In a preferred process of the invention at least part of the solid phase

(opened cells) is added to the liquid phase containing the natural vitamin B1 2 so that a composition is obtained which is a mixture comprising biomass, i.e. preferably opened microbial cells, and natural vitamin B1 2. The vitamin B12 content or concentration of these compositions is expressed as weight- percentage of vitamin B1 2 based on the dry matter content of the composi¬ tion.

The process of the invention produces compositions having a vitamin B1 2 concentration in excess of 0.1 % (w/w) on dry matter. In preferred processes the composition has a natural vitamin B1 2 concentration in excess of 0.2% (w/w), more preferably in excess of 0.4% (w/w), still more preferab¬ ly in excess of 0.6%, still more preferably in excess of 0.8% (w/w) , and most preferably in excess of 1 .0% (w/w). It is preferable that vitamin B1 2 concentrations of the compositions of the invention do not exceed 1 0% (w/w), more preferably the vitamin B1 2 concentrations do not exceed 5%.

In the process of the invention the solid phase of the lysate comprising the cell debris resulting from opening the cells is separated from the liquid containing the released vitamin B1 2. A number of different solid-liquid separation techniques are available to the skilled person for performing this separation, including centrifugation and filtration techniques. A preferred

method for separating the solid cell debris from the vitamin B1 2 containing liquid is ultrafiltration.

In a preferred process of the invention the opened microbial cells are washed and the washings are combined with the vitamin B1 2 obtained after separation of the cells and the vitamin B1 2 containing liquid. Preferably the washing is effected by diafiltration with ion-free water being preferably used for washing the opened cells. The vitamin B1 2 containing diafiltrate is then combined with the vitamin B1 2 containing liquid phase.

In a preferred process of the invention further the natural vitamin B1 2 containing liquid phase is subjected to a drying treatment. Any suitable means of drying can be used such as e.g. spray-drying, fluid-bed drying, freeze drying, or drying in vacuum.

In a preferred process of the invention, the microbial cells containing vitamin B1 2 are washed prior to undergoing lysis, so as to further increase the vitamin B1 2 concentration on dry matter by removing medium compo¬ nents. Preferably said washing is performed using diafiltration whereby preferably using ion-free water.

In a further aspect invention provides compositions comprising vitamin B1 2, which are obtainable in a process of the invention. Preferably the majority of the vitamin B1 2 activity in these compositions is in the form of natural vitamin B1 2. Compositions of the invention are characterized in that the vitamin B1 2 is present in a concentration based on dry matter in excess of 0.1 % (w/w) . In a preferred composition of the invention the vitamin B1 2 concentration is in excess of 0.2% (w/w), preferably in excess of 0.4% (w/w), more preferably in excess of 0.6%, still more preferably in excess of 0.8% (w/w), and most preferably in excess of 1 .0% (w/w) . However, preferably the vitamin B1 2 concentrations of the compositions of the inven ^¬ tion do not exceed 1 0% (w/w), more preferably the vitamin B1 2 concentra-

tions do not exceed 5% . The compositions of the invention comprise vitamin B1 2 and a carrier which preferably comprises biomass e.g. whole cells and/or cell debris. The biomass comprised in these compositions is preferably derived from a microorganism capable of producing vitamin B1 2, such as the bacteria mentioned above, of which the Propionibacterium species are most preferred.

The compositions of the invention are preferably dry compositions, wherein dry is defined as having a water content of less than 1 5% by weight, more preferably less than 10%, most preferably less than 5%.

In preferred dry compositions of the invention the vitamin B1 2 activity is distributed substantially homogeneously throughout the powder or granu¬ lated powder containing the carrier and the vitamin B1 2. As result, the vitamin B1 2 activity in these compositions advantageously does not segre¬ gate to a large extent from the other constituents in the composition, especially when exposed to gravitational forces, sieving, wind sifting or electrostatical forces during processing of the compositions, even when used at concentrations in excess of 0.1 % (w/w) . Processing is herein understood to comprise processing of the compositions of the invention into feed and feed premixes.

In a further aspect of the invention, the composition of the invention is used as or in the production of growth promoting feed supplement for animals. To this end the compositions containing vitamin B1 2 are added to the other feed components, either directly or in the form of a premix which also contains other vitamins, minerals and/or bioactive ingredients. As shown in the specific example, feeding an animal a diet comprising a composition of the invention promotes its growth.

In a still further aspect of the invention the compositions comprising vitamin B1 2 are used as or in the production of a human food supplement

and/or are incorporated into cosmetic preparations, such as shampoos or body lotions.

Examples

Experimental

Vitamin B1 2 assay

Vitamin B1 2 activity was assayed using the turbidimetπc bioassay based on the growth response of Lactobacillus leichmanii ATCC 7830 as described in detail in. The United Stated Pharmacopoeia, The National

Formulary, 1 995, pp 1 71 9-1 721 , United Stated Pharmacopoeial Convention, Inc., Rockville MD.

Plasma B1 2 concentrations

Plasma B1 2 was measured using a commercially available radioassay kit (Bio-Rad Laboratories Ltd, Hemel Hempstead, UK.) The plasma sample is combined with vitamin B ₁₂ ( ⁵⁷ Co) in a solution containing dithiothreitol and cyanide. The mixture is boiled for 30 minutes to inactivate endogenous binding proteins and to convert the various forms of vitamin B ₁₂ to cyanocobalamin. The mixture is cooled and then combined with immobilised (bound to polymer beads), affinity-purified porcine intrinsic factor. This adjusts and buffers the pH of the reaction mixture to 9.35. The reaction mixture is then incubated for 60 minutes at room temperature. During incubation, the endogenous and labelled vitamins compete for the limited number of binding sites based on their relative concentrations. The reaction mixture is then centrifuged at 1 500g for 10 minutes. The labelled and unlabelled vitamins binding to the immobilised binding proteins are concen¬ trated at the bottom of the tube in the form of a pellet and the unbound vitamins remain in the supernatant. The supernatant is aspirated off and the radioactivity associated with the pellet is counted. Standard curves are prepared using vitamin B ₁₂ standards in a human serum albumin base. The concentration of vitamin B ₁₂ in the plasma sample is determined from the standard curves. Standard curves were plotted and unknowns determined

using the AssayZap universal assay calculator, version 2.32 (Biosoft, Cam¬ bridge, UK.) .

Plasma samples required to be diluted 1 0-50 fold to bring them within the most sensitive range of the assay.

Comparative Example 1

A fermentation was performed with Propionibacterium freudenreichii using a process known to those skilled in the art (see e.g. Spalla et al, 1 989 "Microbial production of vitamin B1 2" , In: Biotechnology of vitamins, pigments and growth factors, E.J. Vandamme ed., Elsevier, London, New York, pp. 257-284) . A suitable P. freudenreichii strain is available from the American Type Culture Collection under accession number ATCC 6207. A broth was obtained at the end of the fermentation which had a 5'- desoxyadenosylcobalamin potency of 10 mg/l and with a dry matter content of 7.5%. Spray-drying of this broth resulted in a product with a vitamin B1 2 concentration of 0.01 %.

In order to obtain a higher vitamin B1 2 concentration, a solid-liquid separation of the broth was attempted. Vitamin B1 2 is intracellular in P. freudenreichii, for which reason removal (extracellular) medium components and subsequent spray-drying of the biomass should result in a spray-dried product with a higher vitamin B1 2 concentration.

On lab scale a centrifugation step was performed at a maximal g-force of 5000-6000 *g. This g-force is comparable with the centrifugal forces of centrifuges that are used on industrial scale. Aliquots of 1 000 ml broth were centrifuged in a Beckmann JM/6E centrifuge during 10 minutes at 5000 rpm. However, no separation of biomass and broth was obtained under these conditions.

Because higher centrifugal forces are less attractive on an industrial scale for economic reasons, we tested whether it is feasible to separate the liquid and solids in the broth using ultrafiltration.

2000 ml of broth, with an initial concentration of 4.5 mg vitamin B1 2 /I (vitamin B1 2 on dry matter 0.02 %), was ultrafiltered by means of a 30 kD AMICON spiral wound module (0.09 m ² ) with a feed pressure of 1 bar. Under these conditions it was possible to concentrate the sample of vitamin B1 2 5 times. The average permeateflux was 30 l/m ² h . The vitamin B1 2 concentration in the concentrate was 20 mg/l, whereas in the permeate less than 0.2 mg/l vitamin B1 2 was detected. The vitamin B1 2 concentration in the concentrate, when expressed in terms of dry matter content, was 0.05 %, compared to 0.02% prior to ultrafiltration.

To further increase the titre of vitamin B1 2 on dry matter, we next tested diafiltration of the biomass.

The ultrafiltration concentrate as obtained in the above described experiment was washed six times with ion-free water, which resulted in a vitamin B1 2 potency in the concentrate of 20 mg/l. In the diafiltrate no vitamin B1 2 activity was detected. Subsequent spray-drying of the concen- trate resulted in a spray-dried product with a vitamin B1 2 concentration of 0.06%.

We conclude from the above experiments that it is not possible to obtain a product with a vitamin B1 2 concentration of more than 0.06% based on dry matter starting from a broth with a vitamin B1 2 titre of 5 - 10 mg/l.

Example 2

In order to prevent the presence of viable production organisms, i.e. live cells, in the dried product, a pasteurisation step of 65 °C for 30 minutes

was performed on a broth which was obtained as described in Example 1 . The treated broth was ultrafiltered as described above.

A pink coloured permeate was observed with a vitamin B1 2 concentra¬ tion (by dry matter) of 0.2 - 0.4%. The pasteurization is thought to cause lysis of (at least part of) the P. freudenreichii cells and release of the intracellular vitamin B1 2 into the medium. This can allow one to obtain dried products comprising vitamin B1 2 in concentrations ranging from 0.06% to 0.3% (based on dry matter), by combining the permeate and the pasteurised broth prior to spray-drying. In more detail 350 ml of the heat-treated broth was ultrafiltered under the conditions described above in Example 1 . The concentrate was diafiltered with 2500 ml ion-free water. The clear pink coloured permeate and diafiltrate were combined and mixed with 750 ml pasteurised ultrafiltration concen¬ trate. The mixture was spray- dried in a lab Bϋchi spray-dryer whereby the inlet temperature was set at 1 80°C and the outlet was set at a temperature of 1 04°C. The vitamin B1 2 concentration of the thus obtained spray-dried product was 0.1 3%

Example 3

1 200 I of broth with a vitamin B1 2 concentration of 8 mg/l was obtained in a pilot plant fermentation of P. freudenreichii and was ultrafiltrated in a DDS ultra filtration module M37 using 20 kD membranes and an initial waterflux of 38 l/m h. The temperature during ultrafiltration was kept ambient. The circulation flow was about 1 5 m ³ /h. The feed pressure was set at about 7 bar. The broth was concentrated 1 1 .5 times (average flux of 23 l/m ² h) . The concentrate was washed with 500 I of ion-free water at ambient temperature till a conductivity of about 3 mS/cm was reached.

The resulting concentrate was treated at 90°C for 6 hours. Part of the heat-treated concentrate was diafiltered, resulting in about 300 I of vitamin B1 2 containing permeate. The retentate was stored at 4°C.

The permeate was concentrated on a pilot scale in a glass evaporator at 40 °C under vacuum, resulting in about 6 I of concentrate.

By combining and mixing an appropriate amount of the retentate with the concentrated permeate, we composed a mixture which was spray-dried in a NIRO spray-dryer. The air temperature at the inlet was set at 1 60°C and the outlet temperature was regulated at about 90 - 95 °C. In 5 hours operation time about 45 I of the mixture was spray-dried (about 9 kg/hour) . This resulted in 1 307 g of spray-dried material with a vitamin B1 2 activity of 1 1 00 mg/kg, i.e. 0.1 1 %.

Example 4

In an experiment similar to the one as described in Example 3, a fermentation was performed using the improved P. freudenreichii strain CBS 929.97. The resulting vitamin B1 2 concentration in the fermentation broth was 40 mg/l. The ultrafiltration and diafiltration were essentialy performed as described in Example 3.

Using ultrafiltration, 100 I of the 40 mg/l broth was concentrated to about 40 I. Using diafiltration, the concentrate was washed with 200 I of ion- free water. The resulting washed concentrate had a vitamin B1 2 activity of 96.5 mg/kg and 6.72% of dry matter (spray drying of this washed concen¬ trate would have given a vitamin B1 2 activity of 0.1 4%) .

The washed concentrate was subsequently lysed at 90°C for 5 minutes, followed by a separation of cell debris and liquid by ultrafiltration. The potential vitamin B1 2 concentration of the permeate reached of 0.87% at dry matter (claculated from a vitamin B1 2 activity of 52.1 mg/kg and a dry matter content of 0.6%) . After preconcentration of the permeate in vacuum at 65 °C, different amounts of permeate and concentrate were combined and

dried. The vitamin B12 potencies of the obtained dried products were as given below:

concentrate : permeate vitamin B12 activity by dry matter 1 : 1 0.5%

3 : 7 0.65%

2 : 8 0.72%

8 : 2 0.29%

P.freudenreichii CBS 929.97 was deposited 10 July, 1997 at the

Centraalbureau voor Schimmelcultures, Baarn, The Netherlands.

Example 5

Application of natural vitamin B1 2 in animal feed.

A trial was performed with broilers to compare the efficacy of natural vitamin B1 2 with cyanocobalamin. One day old male broilers were randomly assigned to cages. Eight animals were kept per cage. The cages were situated in an artificial heated, ventilated and illuminated broiler house The animals were vaccinated against New Castle disease at ages one and fourteen days. The animals received the experimental diets ad libitum from day 1 , the first 10 days in the form of crumbles, later in the form of pellets Water was available freely. The experiment lasted 28 days. At day 28 the animals were weighed, feed consumption determined and blood samples were taken. The blood samples were taken from 4 randomly selected broilers per cage. Blood was collected in heparinised tubes The tubes were centri¬ fuged and the plasma was frozen (-1 8 °C) until vitamin B1 2 analysis was carried out. This analysis was performed as described above under Experi¬ mental (Plasma B1 2 concentrations).

Three treatments were comprised in this experiment-

I. Basal diet without addition of vitamin B1 2

II. Basal diet with addition of cyanocobalamin (30 ppb active substance) III Basal diet with addition of natural vitamin B1 2 (30 ppb active sub¬ stance) Each treatment was replicated 5 times.

The composition of the basal diet is presented in table 1 , the results in table 2.

Table 1 . Composition of the basal diet (in %).

Maize 55 Tapioca 1

Soybeanmeal 30 Soybeans, heat treated 5 Feather meal 1

Soya oil 2

Animal fat 3

Vitamins * , minerals, ammo acids 3

Calculated contents:

Metabolic Energy 1 3.3 MJ/kg

Crude protein 22.0 %

Lysine (total) 1 .27 %

Methionine + Cysteine (total) 0.95 %

Without vitamin B1 2.

Table 2. Effect of cyanocobalamin and natural vitamin B1 2 on average growth, feed intake and feed conversion ratio between 1 and 28 days of age and on blood plasma vitamin B1 2 contents at day 28

Example 6

Segregation of vitamin B1 2 activity from a carrier

Segregation of the active compound from the carrier may be caused by particle shape, density and bulk density, flowability, adhesive and electro¬ static properties, moisture content, and hygroscopicity. These segregation processes may occur during manufacturing of premixes and compound feed and during subsequent transport and storage.

This segregation is particularly critical in the case of premixes which comprise microcomponents such as vitamins and trace elements, macrocomponents such as minerals and carriers.

The demixing properties of existing commercial cyanocobalamin preparations (obtained from Rhone Poulenc under the brandname Microvit ^® B1 2) has been compared with natural vitamin B1 2 preparation prepared as described in example 4.

Limestone is used as a carrier for the existing commercial cyanocobalamine preparations.

The test substances were run out of a storage vessel via a vibrating channel to form a pile as shown in Figure 1 .

Three samples were taken from the center and mixed. The procedure was repeated with samples from the base of the pile. The samples were analyzed for vitamin B1 2 activity as described above and compared withe the original homogeneous mix. Deviations in samples from the original homo¬ geneous mix are calculated as follows: •

Deviation at base (%) = {(concentration as base - concentration original mix)/ concentration original mix} * 100

Deviation at center (%) = {(concentration at center - concentration original mix)/ concentration original mix} * 100

Comparison of the deviations at the pile base and pile center of compositions of cyanocobalamin and natural vitamin B1 2 of the invention are shown in Table 3.

Table 3.

Form PCTιRO/U« (January 1M1)