The present invention relates to animal feed and to a process for preparing animal feed which yields improved feed quality and digestibility.

Carbohydrates and starch are used as feedstuffs in the farming industry and in this regard cereals, such as for example oats, used as a component of fish feed in the fish farming industry are a well-known carbohydrate source for farmed salmonids. Studies have shown that salmonids are capable of digesting starch from cereals and that cereals can be mixed in fish feed without adverse effects on digestibility or on the performance of the fish. However, in general, it is observed that the higher the level of cereal present in the feed, the lower the level of digestibility. As a typical example, an increase from 10% to 30% by weight of oats in fish feed typically reduces starch digestibility from about 61% to about 33%. Thus there is generally a need to increase starch digestibility in fish feed.

However, to date only small quantities of cereals have been used as a component of fish feed because of difficulty in extruding feed mixtures into pellets. Conventional feed pellets are also prone to contamination by bacterial strains such as Listeria monocytogenes. This poses a significant problem in the fish farming industry.

We have now found that pre-fermenting of cereal facilitates formation of firm feed pellets and produces feed which is more digestible than conventional feed containing non-fermented cereal, and moreover that where pre-fermentation is effected using a bacteriocin producing strain of lactic acid bacteria (lactobacilli) , the growth of undesirable bacterial strains such as food-borne pathogens or food spoilage bacteria, (for

example species of Listeria. eg. Listeria onoπyt-.oσensR) during the production and storage of the feed is suppressed. Furthermore, fish feed containing pre- fermented cereal has also been found to exhibit an immunostimulating effect.

Bacteriocins are peptides or proteins released by bacteria which show bactericidal or bacteriostatic activity towards both the producing strain and/or other bacteria. Generally, only bacteria closely related to the producing strain are inhibited although in some cases other, non related bacteria, including certain pathogens may also be affected.

Due to their potential use as antibacterial agents, bacteriocins have been the subject of intensive research. In recent years there has in particular been considerable interest in bacteriocins isolated from lactic acid bacteria (LAB) , in view of their potential utility in the food and brewing industries, in particular in the preservation of food products.

It is known that short chain (1-3) β-glucans have an immunostimulating effect in salmon. For example, salmon treated with short chain (1-3) β-glucan have been shown to have increased resistance to different types of bacteria including Yersina ruckeri (Enteric redmouth disease) , Vibro anguillarum (Clasical vibrose) and Vibro salmonicidia (Hitra disease) (see Robertsen B, Rørstad G, Engεtad R. and Raa J, Journal of fish diseases 13, 1990, 391-400) .

Thus viewed from one aspect the present invention provides an animal feed, preferably a fish (eg. salmon) feed or bird (eg. chicken) feed, comprising a fermented cereal, preferably selected from the group consisting of wheat, oats, rye and barley, more preferably from wheat, barley and oats, together with at least one animal foodstuff component.

Viewed from a further aspect the present invention provides a process for producing an animal feed, said

process comprising the steps of:

(a) fermenting a cereal, preferably with lactobacillus bacteria, especially preferably with a bacteriocin-producing lactobacillus strain, particularly especially preferably a Sakacin A or Sakacin P producing lactobacillus strain;

(b) admixing the fermented cereal with at least one animal foodstuff; and, optionally

(c) forming said admixture into feed units, eg. pellets .

Viewed from a still further aspect the invention provides the use of fermented cereal, preferably a bacteriocin containing fermented cereal, as a component in processed animal feed, preferably in fish or bird feed.

Viewed from a yet still further aspect the present invention provides a method of achieving an immunostimulatory response in an animal body by administering thereto an animal feed as defined hereinbefore and a method of reducing the level of cholesterol in an animal body by administering thereto an animal feed as defined hereinbefore together with a digestible amount of fibre capable of reducing the cholesterol level in said animal .

Viewed from an even further aspect the present invention provides the use of a fermented cereal in the manufacture of an animal feed for use in reducing the level of cholesterol in the animal to which the feed is fed or for use in a method of achieving an immunostimulatory response in an animal body.

Cereal fermentation has been found to have the advantage of changing the level of mono- and disaccharides present in the cereal and the solubility of polysaccharides in the cereal, allowing the cereal grains to swell and partly dissolve. The admixture may then more easily undergo extrusion and thereby become more digestible for the fish.

Moreover, fermentation yields a feed with a lower pH than identical feed containing unfermented cereal . This is particularly significant for fish feed in view of the fact that the main pigments used in salmonid feed, astaxanthins, are known to have increased stability under acidic conditions. The pH stabilisation of pigments (carotenoids) in pelletised fish feeds is, therefore, an important advantage of the invention, particularly in view of the high costs of pigments in feed.

Furthermore, cereal fermentation has been found to improve firmness and to lower water solubility of the feed, both of which are highly advantageous properties when feeding fish kept in sea cages.

A further advantage of the process according to the invention is that it produces animal feed which may exhibit an increased immunostimulating effect in the animal to which it is fed. In this regard, it has been found that fish feed containing fermented barley has an immunostimulating effect in the fish to which it is fed. Without wishing to be bound by any theoretical considerations, it is thought that this effect may be due to the formation of short chain (1-3) β-glucan during cereal fermentation by degradation of polysaccharides such as β-glucan.

Another advantage of animal feed prepared in accordance with the process according to the invention is that fermentation of the cereal provides a fish feed which improved digestibility, particularly with regard to starch digestibility. This in turn allows a greater fibre content to be present in the feed than would otherwise be possible. This is important because fibre is an important dietary requirement for animals and iε known to have a cholesterol lowering effect. Typically up to 10% by weight of fibre will be digestible in fish feed according to the invention compared to an upper limit of 2.5% in conventional feed. Preferably the fibre content will be in the range 4-10% by weight.

The feed produced according to the invention may be processed to any conventional form, eg. by expansion, extrusion, heating, drying, or pelleting. By pellets, is meant any form of self-supporting feed unit, eg. tablets, cakes, bricks etc. Pelleting can be effected by compression but most conveniently will be by extrusion. While feeds in pellet form are preferred, the feeds according to the invention may be in other forms, eg. flakes or powders.

The cereal used in the present invention may be any feedstuff cereal. However for fish feed barley, wheat and oats are preferred and oats and barley are especially preferred. The digestibility in salmonids of cereal starch present in fish feed prepared in accordance with the process according to the invention is improved and it has been shown that for barley the increase in digestibility is significant (nearly 100%) and for wheat it is 20%.

The fermented cereal conveniently makes up from 1 to 95%, preferably 10 to 40% by weight of the animal feed. The remaining components may be conventional animal feedstuffs and feedstuff additives, such as unfermented cereals, fish meal, bacteriocins, fats, vitamins, minerals, water, colorantε, flavour-enhancers, therapeutic agents, stabilizers, antioxidants and pH regulators. Cereal components may be whole grain, grain fragment, crushed grain or milled grain (eg. flour) .

In the case of fish feed, the total cereal content of the feed is preferably no more than about 40% by weight and of this it is preferred that the major part be pre-fermented. Especially preferably substantially all the cereal component is pre-fermented.

The bacteriocins present as components of the fermented cereal may be produced during fermentation as mentioned hereinbefore but may also be produced separately and added pure or in a crude or partly purified form prior to proceεεing.

The bacteriocinε uεed may have one or εeveral inhibition spectra. Preferably bacteriocin having

inhibitative effects against pathogen-producing bacteria (εuch aε Liεteria) found in feedεtuffε will be uεed.

The cereal fermentation iε preferably effected using a lactic acid bacteria and in particular a bacteriocin producing Lactobacillus εtrain. Many Lactobacilli are known and are commercially available, for example for uεe in bread baking, eg. from Chr Hanεon, Denmark. Preferred fermentation agentε include the lactic acid bacteria containing εour dough cultivarε .

To inhibit bacterial strains such as Listeria monocytogenes in the animal feed, a bacteriocin producing lactobacilluε εtrain may be uεed. Several εuch εtrains are known, for example L. Sake strain Lb674 and Lb706 producing Sakacin P and Sakacin A respectively (L. Sake Lb674 iε available from Dr. Lothar Krδckel, Federal Inεtitute of Meat Reεearch, Kulmach, Germany and L. Sake Lb706 from MATFORSK, The Norwegian Food Research Institute, Osloveien, Norway) .

Lactic acid bacteria bacteriocins are usually small peptides, seldom containing more than 60 amino acids. Many bacteriocins from lactic acid bacteria have now been described (for a review see Nettles et al. , 1993) and include, moεt notably, the well-εtudied niεin, (εee for example Groεε, et al . , J. Am. Chem. Soc. 93: 4634- 4635, 1971 and Hurεt, Adv. Appl. Microbiol. 27: 85-123, 1981) .

Preferably a bacteriocin producing εtrain εhould εimultaneouεly carry out lactic acid fermentation and produce bacteriocin. Aε bacteriocins are peptides, they may be decomposed and inactivated by proteinases in the flour. In such cases it may be preferable to add bacteriocin prior to procesεing or to pre-heat the flour to inactivate any εuch proteinaεes . If the bacteriocin iε unable to survive procesεing (eg. extruεion) then further bacteriocin or bacteriocin producing εtrainε may be added after processing.

The fermentation step in the procesε according to the invention iε conveniently effected in an aqueouε

medium, generally water, containing 40-60% by weight cereal, preferably in the form of finely ground whole grain flour, for a period of 10 to 48 hourε, preferably about 24-48 hourε, until an acid pH iε obtained, for example a pH of 1-6, eεpecially 3-5, and at ambient or above ambient temperature, for example 25 to 35°C, eεpecially about 30°C.

In the fermentation step, a starter culture may be used or alternatively one may uεe a spontaneous culture obtained from the cereal used, optionally one obtained in advance. Using a starter culture the fermentation period would typically be 10 to 16 hours, while with εpontaneous fermentation a longer period of 24 to 48 hours might be used. The fermentation will not generally be effected using an initially acid medium. The acid pH develops as a result of acid fermentation during fermentation.

The following Exampleε are intended to illuεtrate the invention in a non-limiting manner:

EXAMPLE 1

The raw materialε uεed to carry out the following experimentε were:

(a) Wheat - whole grain wheat flour,-

(b) Oats - finely ground whole grain oat flour;

(c) Fish meal - Norseamink, Norwegian fish meal with antioxidant, Egersund Sildoljefabrikk, Norway. Four mixtures were prepared by admixing fish meal with the following untreated or pre-treated cerealε:

(1) Untreated wheat (15%)

(2) Untreated oats (15%)

(3) Fermented oats (sour-dough) - (baεed on 15% flour weight)

(4) Scalded oatε (based on 15% flour weight) Fiεh feed mixtures containing cereals (1) and (2) were prepared by manually mixing the untreated wheat or oat flour (15%) with fish meal. The water content for both mixtureε was 8.1%.

For the pre-treatment of the fiεh feed mixture containing (3) , fermentation waε carried out by adding equal amountε of oatε flour and water to an equivalent amount of lactic acid bacteria (εour dough) culture. The mixture (20°C, pH 4.6) waε fermented for 24 hourε at 30°C to produce a preculture. The preculture was refreshed regularly with oat flour and water prior to extrusion a week later. The pH of the final fermentation mixture waε 4.0 and when admixed with fiεh meal yielded a mixture with a water content of 19.9%.

For the pre-treatment of the fiεh feed mixture containing (4) , boiling water was added to oats flour in a 1:1 ratio. Shortly after addition, the temperature waε 76°C and cooling to 30°C led to a weight loεs of 3%. The scalded flour/water mixture was added to fish meal yielding a mixture with a water content of 19.9%.

Extrusion was carried out on a Werner & Pleider, Continua 37 extruder. The temperature of the extruder was in the range 130-150°C and water was added at a rate of between 2 and 7.5 1/hour. Further variants included

the rotating screw (280-320 rpm) , the flour feeding rate (1.2-2.4 g/min) and the feeding temperature (130-145°C) . During the extruεion process, the presεure varied between 10 and 25 bar.

The pelletε were allowed to cool and dry at ambient temperature and then packed in closed drums . For each experiment at least two fractions were sampled and the results therefrom are compiled in Tables 1 and 2 below.

The % dry matter for each sample was determined shortly after production and again after 24 hours in air at ambient temperature. The pH was alεo meaεured. Breakage and cruεhing εtrength were measured in a Kramer cell in an Instron Universal teεting machine. The breakage strength was meaεured perpendicular to the length of the pellet and averaged over 12 pelletε. The width and length of each pellet were meaεured together with their water εolubility and rate of εinking. Aε uεed herein, the term solubility refers to the % weight of pellets dissolved in water during a certain period of soaking in water.

Resultε of the teεts carried out on the 4 types of fish feed mixture are presented in Tables 1 and 2 below.

Table i

Drv matter ( ) and pH of pellets from wheat and oat-.- ^***

Treatment Dry matter (%) after pH at production 24 h

Untreated wheat 81.9 89.9 6.1

Untreated oats 77.1 87.3 6.0

Fermented oats 75.6 87.7 5.8

Scalded oats 76.4 86.6 6.0

Table 2

Breakage and water solubility of pellets from wheat and oats

Treatment Breakage Crushing Water solubility (%) N N 2 min 20 h

Untreated wheat 16.4 213 8 62

Untreated oats 16.6 103 15 38

Fermented oatε 16.4 127 9 32

Scolded oatε 16.3 134 19 100

Notably from the resultε in Table 1, drying for 24 hours at ambient temperature led to an 8 to 12.1% increaεe in the dry matter content with the moεt extensive drying observed in the fermented oatε . All of the pelletε proved to be εtable toward mould growth or any other viεible deterioration over a 12 month period. In addition, pelletε containing fermented oats had a lower pH than the other pellet-types.

Whilεt breakage strength εhowed no εignificant variation amongεt the pellet-typeε, a greater force waε required to crush the wheat-containing pellets than any of the oat-containing pellets.

Water solubility waε determined from the time taken for pellets to sink in water. All pellet types took several hours to sink but when finally wetted, εolubility roεe.

EXAMPLE 2

Experiments in Atlantic salmon fed with feed containing different amounts of fermented and unfermented wheat or barley flour

The following describes the utility of animal feed produced by the process according to the invention. Feed composition is stated alongεide digeεtibility

results, carbohydrate analysis and immunoεtimulating effects in εalmon. In all experiments, the salmon performed well .

The raw materials used to carry out the following experiments were:

(a) Wheat - whole grain wheat flour, finely ground;

(b) Barley - whole grain barley flour, coarεely ground.

Eight mixtureε were prepared by admixing fish meal with the following pre-treated (fermented) or untreated (unfermented) cereals:

(1) Pretreated wheat (12%) (HF 12)

(2) Pretreated wheat (24%) (HF 24)

(3) Pretreated wheat (36%) (HF 36)

(4) Untreated wheat (24%) (HU 24)

(5) Pretreated barley (12%) (BF 12)

(6) Pretreated barley (24%) (BF 24)

(7) Pretreated barley (36%) (BF 36)

(8) Untreated barley (24%) (BU 24)

The feed compositionε were choεen to give equal levels of energy and are shown in Table 3. Extrusion conditions were as described above in Example 1.

Ta le 3

Feed compositions (kg/100 kg feed)

Type of feed ingredient Fermented Unfermented

1 2 3 4

Coding H(B) F H(B)F H(B)F 36 H(B)U24 12 24

Wholemeal wheat/barley 12 24 36 24

Fish meal 66 57 48 57

Fish oil 22 19 16 19

Sum 100 100 100 100

Indicator (Ytriumoxide) 100 mg/kg feed

Mineral mix 100 mg/kg feed

Astaxanthin 75 ppm (937.5 mg Carophyll Pink/kg feed)

The feed was fed ad libi tum to Atlantic salmon weighing 0.5 kg. After an adaption period of 9 days, the experimental period lasted for 3 weeks. Faeces were collected on day 8 and day 16 according to the method of Austreng, E. (see Aquaculture, 13: 265-272, 1978) . Faeces were freeze-dried prior to analyεiε.

Chemical composition of feed

Chemical composition of feed aε revealed by analyεiε iε given in Table 4.

Table 4

Chemical composition of feed (g/lOQg feed)

Type of Dry Protein Fat Starch beta- sugars feed matter glucan ( ono- and di)

Wheat

HF 12 93.7 47.4 28.5 5.6 0.05 0.21

HF 24 93.7 43.0 25.8 11.5 0.08 0.48

HF 36 93.1 39.0 21.8 17.2 0.07 0.67

HU 24 94. C 42.2 26.2 11.1 0.08 0.20

Barley

BF 12 94.4 47.3 28.9 4.5 0.28 0.11

BF 24 94.1 42.5 24.8 10.2 0.53 0.38

BF 36 93.6 38.0 22.0 15.3 0.79 0.59

BU 24 94.2 42.8 26.2 8.2 0.40 0.00

Effects of fermentation on digestibility

Digeεtibility (aε %) of protein, fat, εtarch and energy were calculated from chemical analyεiε of feed and faeces collected from the fish. The resultε are given in Table 5.

Table 5 Digestibility (%) of fish feed fed to Atlantic salmon

Type of Protein Fat Starch Energy feed

Wheat

HF 12 83.3 88.7 70.9 83.6

HF 24 84.3 89.6 57.2 81.6

HF 36 85.5 84.5 38.4 75.8

HU 24 84.3 86.5 47.9 79.8

Barley

BF 12 84.4 85.4 68.9 81.7

BF 24 85.7 86.9 63.2 81.1

BF 36 85.8 84.4 48.0 75.4

BU 24 84.6 87.3 33.2 80.2

For protein, there was a tendency for increased digestibilty with increasing amountε of fermented wheat or barley flour in the fiεh feed. For feed containing barley, there was a slightly higher protein digeεtibility when the flour waε fermented (85.7) compared with non-fermented feed (84.6). These feedε were both formulated with 24 % barley flour.

For fats, no systematic variation in % digestibilty due to carbohydrate level or treatment were seen.

For εtarch digestibility, significant positive effects of fermentation were obtained for both wheat- and barley-containing feeds. In feeds containing 24% fermented wheat flour, starch digestibility waε 57.2% compared with 47.9% for unfermented feed. Thiε correεpondε to an increaεe in εtarch digeεtibility of 19%.

For barley the effect of- εtarch digeεtibility due to fermentation was even more pronounced; non-fermented feed with 24% flour had a digeεtibility of only 33.2% while that of the fermented feed was 63.2%. This is an

increase in starch digestibility of more than 90%.

As is known from previouε reεearch, starch digeεtibility decreaεes with increaεed levelε of cerealε. Thiε waε alεo seen in the present experimentε. However, for barley the fermentation process compensated for this so that the starch digestibility of feed containing 36% fermented barley (48.0%) was higher than the digeεtibility of feed containing the lower level of 24% unfermented barley (33.2%) .

The increaεed digeεtibility of starch due to fermentation was supported by the resultε for energy digeεtibility. The increaεed digeεtibility of fermented εamples caused higher energy digestibility than the unfermented samples. For wheat containing 24% flour the digestibilities were 81.6% and 79.8% for fermented and unfermented feed respectively. For barley, the corresponding figures were 81.1% and 80.2%.

Effects of fermentation on fish performance

The salmon performed well during the experimental period with no problems with appetite in the fish. Faeces were of normal consistency, even at the higher carbohydrate levels. This is reflected in normal dry matter content of the faeces, with no sign of diarrhoea (dry matter below 12) .

Type of feed Dry matter in faeces (g/100g)

Wheat

HF 12 15.0

HF 24 14.4

HF 36 15.7

HU 24 14.3

Barley

BF 12 14.9

BF 24 13.6

BF 36 17.2

BU 24 15.3

Effects of fermentation on carbohydrate composition

The εolubilty of carbohydrateε in the feed waε influenced by the fermentation procesε. Water soluble carbohydrate contents of wheat and barley flour prior to and after fermentation are given in Table 6.

Table 6

Carbohydrates (σ/lOOσ feed) in wheat and barley flour prior to and after fermentation

Sample Glucose Maltose Water soluble Total Total

(G) (M) polysaccharide G+M

Unfermented

Wheat 0.72 5.72 2.05 6.44 8.49

Barley 1.05 4.81 5.61 5.86 11.47

Fermented

Wheat 2.04 1.16 1.24 3.20 4.44

Barley 2.00 0.06 2.76 2.06 4.82

Aε theεe resultε εhow, the content of maltoεe decreased conεiderably, whilst the amount of glucose increased during fermentation. The amount of mono- and disaccharides was nearly half that of the unfermented flour for both wheat and barley. Alεo the amount of water εoluble polyεaccharideε decreased during the fermentation procesε. Approximately half aε much water εoluble polymeric compoundε waε detected after fermentation. The lowering in soluble polysaccharides was moεt extensive in barley flour where the initial level of 11.5% was lowered to 4.8%, a reduction of 59%.

Immunostimulating effects

Experimental fish feed was fed to salmon for 16 days and blood sampleε were teεted for εerum lyzoyme activity and uεed to evaluate the immunoεtimulating effect in the blood. The reεults are shown in Table 7.

Table Serum lysozyme activity in salmon fed feed containing fermented and unfermented barley

Type of feed Lysozyme activity ( % )

Barley

BF 12 28 . 445 ^abc *

BF 24 32 . 852 ^a

BF 36 28 . 480 ^abc

BU 24 23 . 004°

* figures denoted with the same letter are not significantly different (P<0.5)

Serum lysozyme activity waε determined by the Micrococcus lyεoplate aεεay. In this method the sample is placed in a small well stamped out in an agarose gel containing the bacteria Micrococcus lyπodeicticus . After incubation, the agarose gels are stained and

destained and the diameter of the lysed zones are measured. The resultε are compared with the diameter of a reference serum (100%) known to contain lysozyme activity and given aε a percent of the diameter of the reference. Actual levelε of lyεozymes may vary between groups of fish due to health εtatus, age etc. The effect in a special trial iε evaluated aε the relative lysozyme activities of fish within that trial.

Lysozyme is an enzyme which hydrolyεes cell wall components of bacteria. By degradation of the cell walls, the growth of the bacteria is inhibited. High lysozyme activity may therefore give an increased resistance toward pathogen bacteria and thereby enhanced health to the animal.

The results clearly demonstrate increased levels of lysozyme activity in salmon fed with fermented barley compared with salmon fed with unfermented feed. Lysozyme activity was εignificantly higher in εalmon fed with 24% fermented barley (32.85%) compared with salmon fed with the εame amount of unfermented barley (23.00%) .