This invention relates to a composition of selected bacterial cells and to their use in the preservation of silage crops and the reduction of spoilage by yeasts and moulds. Background of the Invention

During the initial stage of the ensiling process, soluble carbohydrates in the plant tissues are oxidised and converted to carbon dioxide and water. Under these conditions, the growth of spoilage microorganisms, including yeasts and moulds, is observed. Anaerobic conditions, however, are normally established quickly, ensuring the dominance of heterofermentative and homofermentative lactic acid-producing bacteria. This phase of ensilage may be enhanced by the introduction of specific lactic acid-producing strains. The resultant increase in lactic acid levels reduces the pH of the silage to around 4, improving nutritional retention and initial aerobic stability. The bacterial inoculants convert sugars and other simple carbohydrates to lactate, in addition to molecules of acetate, ethanol or mannitol, in the presence of 5-carbon sugars. The availability of these carbon compounds in silage crops is enhanced by the addition of specific hydrolytic enzymes which mobilise the necessary sugars from the plant tissues.

Reduced stability, caused by the growth of invasive moulds, is generally limited to subsequent exposure of the silage to the air. Yeasts, which are primarily introduced during ensiling, may increase in numbers under certain conditions, and may also ferment the lactic acid to other redundant compounds. This decreased silage stability is also associated with increased levels of alcohol compounds.

Throughout the ensiling process, beneficial fermentation acids (also known as "volatile fatty acids") are formed. The amounts of such acids depend on the changing nutritional and physical characteristics of a

silo, on the particularly crop used, and on the method of ensilage.

Summary of the Invention

The present invention is based at least in part on the realisation that a beneficial effect may be achieved by maximising the possibility of a beneficial acid profile. Although such a profile may be achieved naturally, it would be more desirable to provide means whereby the farmer can expect to achieve patterns of acids, at the end of ensilage, which are nutritionally favourable and inhibitive to invasive or endemic spoilage microflora. The protective and ensiling qualities of lactic acid alone are recorded at levels of 80-150 g/kg CDM. The additional mould-inhibitive qualities of acetic and propionic acids are generally recorded at lower concentrations in silage, representing an overall ratio of between 20:4:1 and 60:10:1 (lactic acid:acetic acid:propionic acid, respectively) . These beneficial acids are known to be produced as primary and secondary metabolites by certain species of the bacterial genera such as Lactobacillus. Pediococcus. Acetobacter and Propionibacterium. Description of the Invention

In a preferred embodiment of the invention, a stable freeze-dried powdered blend contains viable cells of Lactobaci1lus plantarum _P Pediococcus pentsaceus. Acetobacter pasteurianus and Propionibacterium iensenii. Such bacteria may be used in association with specific enzymes such as xylanase, arabinofuranosidase and xylosidase activities, to release silage carbon sources to more easily assimilated acetate and lactate components.

Bacteria of the general type, genera or species described above may be for the controlled addition of specific stabilised bacterial cells (e.g. at 10 to 10 CFU per tonne) together with selected enzymes in order to preserve silage material during the initial anaerobic phase of the ensilage process. This prevents long-term storage

instability and microfloral spoilage after the silo is exposed to air.

Further, the addition of specific bacterial isolates and enzymes and 5-carbon sugars to silage material induces and extends a profile of volatile fatty acids. Greater silage stability is achieved by the establishment of a set profile of these beneficial metabolites from the early stages of ensilage to final use.

In order to prepare a formulation of the invention, the following steps may be adopted:

Preparation of viable cells

Viable cells of each of the bacterial strains are grown in each of four optimised liquid fermentations.

Selected cultures of P.. pentosaceus , L. plantarum and P. iensenii may be grown in liquid microaerophilic conditions, while A. pasteurianus may be grown in aerobic liquid conditions.

Preservation of viable cells

End fermentation broths of each of the 4 bacterial strains may be centrifuged to recover the viable cells.

Slurries are prepared by resuspending the cells in water before mixing in fixed proportions of sucrose and dried skimmed milk. The final solids content of the paste is suitably between 20% and 37.5% w/w, depending upon the cell type. Each cell formulation may then be freeze-dried, e.g. to an end moisture content of 3% or below.

Application of inoculum cells

Stabilised cells are milled and blended together.

This is, for example, at a total cell count ratio of 2:2:1:1 (£. pentosaceus. L.. plantarum. P. iensenii and A. pasteurianus. respectively) . After the addition of specific hydrolytic polysaccharides and nutrients, the final viable microorganism count may be 3.34 x 10 ⁷ CFU per g- Fermentation acid ratio

After resuspension in water and even application to the forage crop, the desired quantities and ratio of

lactic, acetic and propionic acids will be achieved between

7 and 14 days of ensilage and will extend over a 150 day period. The balance of these acids in silage material over the same time course will actively reduce the overall pH and inhibit any spoilage by filamentous moulds or yeasts.

The following Examples illustrate the invention.

In controlled laboratory trials, strains of

Propionibacterium iensenii and Acetobacter pasteurianus were selected for their abilities to actively inhibit the growth of filamentous moulds and yeasts, respectively. These inhibitive abilities were identified and quantified against selected spoilage microflora. The complete inhibition of filamentous moulds was observed in icroaerophilic infusion cultures challenged with cells of P. iensenii. The active bacterial metabolite was identified as propionic acid, effective at a concentration exceed 0.14 g/1. The concentration of propionic acid increased to 0.35 g/1 after 3 days' growth, when easily- assimilated sugar-based carbon compounds became depleted. Significant levels of acetic acid were also produced (1.44 g/1) as a primary metabolite during the initial 48 hours of fermentation, which then slowly decreased as it was assimilated to propionic acid by E. iensenii. This concentration of acetic acid is only partially inhibitive to selected silage spoilage yeasts in vitro. Complete eradication of yeasts in a grass infusion culture was obtained with an acetic acid concentration equivalent to 3.94 g/1.

Similar liquid cultures (aerated) containing cells of Acetobacter pasteurianus produced acetic acid levels of between 15 and 25 g/1 after 3 days incubation at 25°C. As described earlier, these concentrations are sufficient to kill endemic and secondary colonising yeast cells.

When all 4 isolates were combined under microaerophilic conditions on either grass or grass infusion broth, the lactic acid-producing strains dominated. Lactic acid levels were equivalent to those

produced on grass without the additional inoculation of Acetobacter and Propionibacterium. The Acetobacter bacterial numbers declined after 24 hours, due to the depletion of oxygen. The total concentration of acetic acid produced by this mixed culture within this period ranged between 7 and 35 g/1 CDM equivalent. These concentrations were confirmed sufficient to suppress introduced or invasive yeast strains on ensiled grass.

Propionibacterium cells were not detected in significant numbers until the 7th day of ensilage. This corresponded to increased concentrations of propionic acid, measured between 2.4 and 5.8 g/kg equivalent. These concentrations were adequate to suppress introduced or invasive filamentous mould strains. On a large-scale farm trial, grass was treated with a formulation containing L.. plantarum P. pentosaceus and £. iensenii at a combined loading rate of 1.2 x 10 viable organisms per tonne. After only 24 hours'' ensiling, the numbers of yeasts and moulds were significantly reduced, and virtually eliminated after a further 5 days. Mould numbers were minimal during ensiling but did not increase after repeated exposure of the silo to the air. Lactic acid levels increased significantly during the initial 7 days, while propionic acid levels did not rise until after 21 days of ensiling. The concentration of acetic acid did not rise to its maximum until 52 days after ensilage. Direct inoculation with the Acetobacter strain may result in this effective concentration at an earlier stage.

At the end of the trial (130 days) , a desired acid profile was recorded. As a result, the aerobic stability of exposed material was considered good, and was achieved at an earlier stage than silage treated with a lactic acid bacterial inoculant alone.

In further trials, various farmers (identified as H, P, L, W and S) applied the same formulation to their grass, for ensiling. Farm B did not apply the formulation to his grass, as a control. The results are tabulated below, in

terms of observed parameters at day 0 (immediately before addition of the formulation) and at day 30-45. Farmer H:

Grass Day 30-45

Acetic 0 19.86

Propionic 0.96 1.88

Lactic 8.7 71.66

LAB 2.05e5 l.lle7

Yeasts 1.18e6 3.6e5

Moulds 4e4 <50

Coliforms >5e5 <50

Farmer P:

Grass Day 30-45

Acetic 2.17 30.9

Propionic 0.09 1.74

Lactic 3.25 43.46

LAB 4.05e5 1.85e6

Yeasts le4 9e2

Moulds 1.48e4 <50

Coliforms 5.4e4 <50

Farmer L:

Grass Day 30-45

Acetic 6.21 31.53

Propionic 0.1 4.69

Lactic 12.69 73.88

LAB >5e6 8e6

Yeasts >5e6 1.07e4

Moulds 1.55e4 <50

Coliforms 5e5 <50

Farmer W:

Grass Day 30-45

Acetic 3.51 26.88

Propionic 0.66 1.3

Lactic 2.96 111.75

LAB 7.5e4 5.1e6

Yeasts 1.59e6 4.95e3

Moulds le4 <50

Coliforms 1.05e5 <50

Farmer S:

Grass Day 30-45

Acetic 0.38 17.49

Propionic 0.24 1.51

Lactic 0.77 114.83

LAB 1.7e5 2.5e6

Yeasts 4.35e5 2e3

Moulds le4 <50

Coliforms 5e5 <50

UNTREATED CONTROL Farmer B:

Grass Day 30-45

Acetic 1.3 4.43

Propionic 0.61 0

Lactic 0 28.72

LAB 3.26e6 2.4e6

Yeasts 6.25e4 >5e6

Moulds 3.75e3 <50

Coliforms >5e5 <50